Mobility of Thermomyces lanuginosus Lipase on a Trimyristin

Jan 30, 2007 - trimyristin substrate surface using fluorescence recovery after ... in adsorbed amount when the trimyristin surface was pretreated with...
0 downloads 0 Views 235KB Size
Download PDF
2706

Langmuir 2007, 23, 2706-2713

Mobility of Thermomyces lanuginosus Lipase on a Trimyristin Substrate Surface Andreas W. Sonesson,*,†,‡ Hjalmar Brismar,‡ Thomas H. Callisen,§ and Ulla M. Elofsson† YKI, Institute for Surface Chemistry, Stockholm, Sweden, Department of Cell Physics, Royal Institute of Technology, Stockholm, Sweden, and NoVozymes A/S, BagsVaerd, Denmark ReceiVed July 11, 2006. In Final Form: NoVember 27, 2006 We have studied the mobility of active and inactive Thermomyces lanuginosus lipase (TLL) on a spin-coated trimyristin substrate surface using fluorescence recovery after photobleaching (FRAP) in a confocal microscopy setup. By photobleaching a circular spot of fluorescently labeled TLL adsorbed on a smooth trimyristin surface, both the diffusion coefficient D and the mobile fraction f could be quantified. FRAP was performed on surfaces with different surface density of lipase and as a function of time after adsorption. The data showed that the mobility of TLL was significantly higher on the trimyristin substrate surfaces compared to our previous studies on hydrophobic model surfaces. For both lipase variants, the diffusion decreased to similar rates at high relative surface density of lipase, suggesting that crowding effects are dominant with higher adsorbed amount of lipase. However, the diffusion coefficient at extrapolated infinite surface dilution, D0, was higher for the active TLL compared to the inactive (D0 ) 17.9 × 10-11 cm2/s vs D0 ) 4.1 × 10-11 cm2/s, data for the first time interval after adsorption). Moreover, the diffusion decreased with time after adsorption, most evident for the active TLL. We explain the results by product inhibition, i.e., that the accumulation of negatively charged fatty acid products decreased the diffusion rate of active lipases with time. This was supported by sequential adsorption experiments, where the adsorbed amount under flow conditions was studied as a function of time after adsorption. A second injection of lipase led to a significantly lower increase in adsorbed amount when the trimyristin surface was pretreated with active TLL compared to pretreatment of inactive TLL.

Introduction Lipases (triacylglycerol hydrolases, E.C. 3.1.1.3) catalyze the hydrolysis of glycerides into free fatty acids and glycerol. Almost all lipases are interfacially activated at the lipid-water interface.1 The catalytic hydrolysis by a secreted lipase often involves adsorption from the bulk aqueous phase to a substrate surface, followed by catalytic turnover of substrate at the interface. How the catalytic turnover of lipases is related to interfacial mobility is not well understood. For phospholipase A2, the enzyme dynamics at an interface has been divided into two types of interactions, the hopping or the scooting mode.2 In the scooting mode, adsorbed enzyme remains at the surface due to multiple noncovalent interactions and must therefore migrate laterally to explore and hydrolyze the substrate, if not the substrate at the interface is mobile. The hopping mode, on the other hand, involves desorption and readsorption to the substrate. Triglycerides are a main constituent of vegetable oil and animal fats and play a crucial role as an energy source in metabolism from simple to higher organisms. The triglyceride lipase from the fungus Thermomyces lanuginosus (TLL) is an enzyme with pronounced interfacial activation.3,4 We have earlier studied the mobility of irreversibly adsorbed TLL at model surfaces,5 including conditions of competition with surfactants.6 This has revealed that the mobility is modulated by the hydrophobicity * Corresponding author: Institute for Surface Chemistry, Box 5607, SE114 28 Stockholm, Sweden; e-mail, [email protected]. † YKI, Institute for Surface Chemistry. ‡ Department of Cell Physics, Royal Institute of Technology. § Novozymes A/S, Bagsvaerd, Denmark. (1) Verger, R. Trends Biotechnol. 1997, 15, 32-38. (2) Berg, O. G.; Gelb, M. H.; Tsai, M. D.; Jain, M. K. Chem. ReV. 2001, 101, 2613-2653. (3) Brzozowski, A. M.; Savage, H.; Verma, C. S.; Turkenburg, J. P.; Lawson, D. M.; Svendsen, A.; Patkar, S. Biochemistry 2000, 39, 15071-15082. (4) Hedin, E. M. K.; Hoyrup, P.; Patkar, S. A.; Vind, J.; Svendsen, A.; Fransson, L.; Hult, K. Biochemistry 2002, 41, 14185-14196.

of the surface, where the mobility decreased as a function of time after adsorption and degree of hydrophobicity. Moreover, the presence of surfactants boosted the lipase mobility, explained by a surfactant-induced desorption-rebinding migration on the surface. On a hydrophobic C18-terminated silica surface, used to mimic a triglyceride substrate, the lipases were found to have very low mobility. The aim of this work was to investigate TLL diffusion on a substrate surface of trimyristin. There are to our knowledge no earlier diffusion measurements of lipases on triglyceride surfaces and only a few on enzyme surface mobility in other enzymesubstrate systems.7-11 To quantify surface diffusion, fluorescence recovery after photobleaching (FRAP) was used in a confocal microscope. FRAP is a well-established technique commonly used in cell biology12,13 and pharmaceutical research.14 In short, a high-intensity laser is used to bleach a predefined region of a fluorescent sample. The subsequent fluorescence recovery of the bleached area is monitored and fitted to the solution of the diffusion equation. In this way both the diffusion coefficient D and mobile fraction f of the studied species can be quantified. (5) Sonesson, A. W.; Callisen, T. H.; Brismar, H.; Elofsson, U. M. Langmuir 2005, 21, 11949-11956. (6) Sonesson, A. W.; Elofsson, U. M.; Brismar, H.; Callisen, T. H. Langmuir 2006, 22, 5810-5817. (7) Jervis, E. J.; Haynes, C. A.; Kilbourn, D. G. J. Biol. Chem. 1997, 272, 24016-24023. (8) Roy, S.; Thomas, J. M.; Holes, E. A.; Kellis, J. T.; Poulose, A. J.; Robertson, C. R.; Gast, A. P. Anal. Chem. 2005, 77, 8148-8150. (9) Gaspers, P. G.; Robertson, C. R.; Gast, A. P. Langmuir 1994, 10, 26992704. (10) Collier, I. E.; Saffarian, S.; Marmer, B. L.; Elson, E. L.; Goldberg, G. Biophys. J. 2001, 81, 2370-2377. (11) Henis, Y. I.; Yaron, T.; Lamed, R.; Rishpon, J.; Sahar, E. O ¨ .; KatchalskiKatzir, E. Biopolymers 1988, 27, 123-138. (12) Reits, E. A. J.; Neefjes, J. J. Nat. Cell Biol. 2001, 3, E145-E147. (13) White, J.; Stelzer, E. Trends Cell Biol. 1999, 9, 61-65. (14) Meyvis, T. K. L.; De Smedt, S. C.; Van, Oostveldt, P.; Demeester, J. Pharm. Res. 1999, 16, 1153-1162.

10.1021/la062003g CCC: $37.00 © 2007 American Chemical Society Published on Web 01/30/2007

Lipase Mobility on Trimyristin

Two fluorescently labeled variants of TLL were used in this work, one active and one inactive variant, so the effect of the hydrolysis reaction on the mobility could be estimated. It has previously been shown that these two lipase variants have similar slow diffusing behavior on hydrophobic model surfaces.5,6 Diffusion coefficients and mobile fractions were measured at different relative surface densities of enzyme. Therefore, the diffusion coefficient at infinite surface dilution, D0, could be extrapolated. D0 is the best way of comparing different systems since it effectively removes crowding effects from the diffusion process. Materials and Methods Materials. Trimyristin (C45H8606, Mw ) 723.18, 97% (GC) catalog no. 92780) and toluene (g99.5% (GC)) were from Fluka AG (Buchs, Switzerland). Alexa Fluor 488 protein labeling kits were purchased from Molecular Probes Europe BV (Leiden, The Netherlands). Two lipase variants from Thermomyces lanuginosus (TLL) were provided by Novozymes A/S (Bagsvaerd, Denmark). The buffer used in adsorption and diffusion experiments was glycine, pH 9.0 (10 mM NaCl, 0.05 mM EDTA, 50 mM glycine, and 1 mM NaN3), and all water was of Milli-Q grade. Preparation of Trimyristin Films. Poly(vinyl chloride) (PVC) surfaces, cut in appropriate size for the CLSM flow cell (13 × 18 mm), were treated with detergent solution, Milli-Q water, and ethanol and finally blown dry with N2. Cleaned PVC surfaces were spincoated with trimyristin using a PWM32 photoresist spinner (Headway Research, Inc., Garland, TX). A 50 µL drop of 15% (w/w) trimyristin in toluene was placed in the center of the surface. The surface was rotated at a speed of 4000 rpm for 4 min, with an acceleration of 100 rpm/s. This led to a thin trimyristin film on the PVC surface, with a contact angle with water of 121 ( 4°. This procedure has also earlier been found to give smooth triglyceride surfaces.15 The root mean square (rms) surface roughness parameter was measured with white light profilometry to Sq ) 0.302 ( 0.04 µm and an effective area which was approximately 30% larger than the planar geometry. Protein Labeling. The two lipase variants used in this study are denoted TLL-R and TLL-β, respectively. The difference between them was that TLL-R is more charged at the studied pH (pH 9) and modified to have fewer specified labeling sites for a fluorophore. The lipases were labeled with Alexa488. Labeling of TLL-β was performed using the standard protocol from the manufacturer (http:// www.probes.com) while labeling of TLL-R was done by incubating 0.5 mL of 0.33 mg/mL TLL-R to one vial of reactive Alexa488 for 8 h in room temperature followed by 14 h at +8 °C. Labeled TLL-R and TLL-β were separated from free dye using a Bio-Rad BioGel P-30 fine exclusion purification resin with glycine buffer as elution buffer. To determine protein and fluorophore concentrations, the absorbances at 280 and 494 nm were measured, using TLL,280nm ) 38 440 cm-1 M-1 and Alexa488,494nm ) 71 000 cm-1 M-1, which led to a degree of labeling of 0.95 mol of dye per mole of enzyme for TLL-R and 0.6 mol of dye per mole of enzyme for TLL-β. Labeled lipase was stored in aliquots at -20 °C. Lipase Activity. Lipase activity was measured using 4-nitrophenylvalerate (pNP-valerate), a fatty acid ester, as substrate in a similar protocol described elsewhere.16,1717 The amount of product (pNP) produced after incubation with lipase was monitored spectrophotometrically at 405 nm against a blank without enzyme. Lipase Adsorption onto Trimyristin. Adsorption and subsequent surface diffusion of fluorescently labeled TLL was studied in a Zeiss LSM 510 Meta confocal microscope, using a 40 × 1.3 oil immersion objective and the 488 nm line of an Ar-ion laser for excitation. PVC surfaces of 13 × 18 mm, coated with trimyristin, were mounted in a flow cell of about 15 µL, defined by a silicone rubber gasket and a coverslip glass. Lipase samples were loaded in a 50 µL loop and injected to the flow cell via a six-port injection (15) Engstro¨m, S.; Ba¨ckstro¨m, K. Langmuir 1987, 3, 568-574. (16) Gupta, N.; Rathi, P.; Gupta, R. Anal. Biochem 2002, 311, 98-99. (17) Kademi, A.; Ait-Abdelkaber, N.; Fakhreddine, L.; Baratti, J. C. Enzyme Microb. Technol. 1999, 24, 332-338.

Langmuir, Vol. 23, No. 5, 2007 2707 valve (Skandinaviska Genetec AB, Go¨teborg, Sweden). A syringe pump (Harvard Apparatus) was used to pump buffer and sample through the system. Lipase samples were injected and adsorbed to the trimyristin surface for 600 s under a constant flow rate of 10 µL/min, followed by a rinse with buffer for 400 s. The adsorbed amount (a.u.) was calculated as the mean surface fluorescence intensity of ten 220 × 220 µm squares after the rinsing. An InSpeck Green(505/515) Microscope Image Intensity Calibration kit with 6 µm microspheres purchased from Molecular Probes Europe BV (Leiden, The Netherlands) was used to control the laser and PMdetector gain so absolute intensity values were comparable between experiments. We have previously established a linear correlation between surface fluorescence intensity and surface excess as measured by ellipsometry.5,6 Fluorescence Recovery after Photobleaching (FRAP). The general idea of using FRAP to monitor lateral diffusion is to photobleach one region of a fluorescent sample and monitor the time evolution of the fluorescence recovery of the illuminated region. The recovery is thus a measure of the lateral diffusion of the fluorescent species. If a portion of the fluorescent molecules is immobile or very slow diffusing, the recovery profile will not reach its prebleach value but level out on a plateau level of lower fluorescent intensity. When the diffusion equation for the geometry of the bleached region is solved, the mobile fraction f and the diffusion coefficient D of that fraction can be estimated from the recovery profile. Normalizing a recovery profile to the maximum recovery gives the fractional fluorescence recovery profile, f(t). The theoretical solution for f(t) when bleaching a circular geometry has been determined, e.g., by Axelrod et al.18 f(t) ) 1 - (τD/t) exp(-2τD/t)[I0(2τD/t) + I2(2τD/t)] 2

∞

(-1)k(2k + 2)!(k + 1)!(τD/t)k+2

k)0

(k!)2[(k + 2)!]2

∑

(1)

where I0 and I2 are modified Bessel functions and τD is the characteristic diffusion time. The diffusion coefficient D can be calculated using the relationship in eq 2

( )

D ) γD

ω2 4τ1/2

(2)

where ω is the radius of the bleached circle, τ1/2 is the time when the recovery profile has reached half its maximum, and γD is a constant defined as τ1/2/τD. For a circular beam, γD ) 0.88.18 In this work, the bleaching software of the Zeiss 510 Meta confocal microscopy with a 40 × 1.3 oil immersion objective was used to define circles with radius of 2-3 µm on flat, uniform regions on the trimyristin film. These regions were photobleached for 10 s using the 488 laser at maximum intensity, and images were collected every 10th second after bleaching for 1700 s. The fluorescence from adsorbed lipases was detected using the 488 nm line of an Ar-ion laser and a 505 nm long-pass emission filter. The pinhole was fully opened during the recovery to compensate for any drift in focus. The recovery profile as well as the radius of the bleached circles was then calculated using the built-in software. The recovery profile F(t) was normalized to the intensity of the prebleach intensity, F0, so that the mobile fraction f could be estimated as the average value during the last 200 s (F∞) and τ1/2 as the time when the intensity had recovered to F∞/2 (Figure 1.)

Results Lipase Activity. The activity of unlabeled and Alexa488labeled TLL-R and TLL-β was measured after 25 min incubation with pNP-valerate. The results are displayed in Table 1, where the absorbance of the blank with no enzyme has been subtracted and the resulting values normalized to the lipase of highest activity (18) Axelrod, D.; Koppel, D. E.; Schlessinger, J.; Elson, E. L.; Webb, W. W. Biophys. J. 1976, 16, 1055-1069.

2708 Langmuir, Vol. 23, No. 5, 2007

Sonesson et al.

Figure 2. (A) Adsorption isotherms, i.e., surface intensity after adsorption vs concentration in bulk solution for TLL-R (b) and TLL-β (4). Intensity values are normalized to the maximum intensity found with TLL-β. Error bars are ( standard deviation. (B) Surface intensity after adsorption with 500 nM TLL-β with different fractions of the labeled stock population (degree of labeling ) 0.6). Dashed line is a linear fit to the data. All intensity values are corrected for the difference in degree of labeling of the two enzymes. Figure 1. (A) Recovery profile of a 4 µm bleached spot on trimyristin (TLL-R, θ/θmax ) 0.12) with a mobile fraction, f ) F∞ ) 0.67 and τ1/2 ) 85 s. (B) Confocal micrographs of the recovery, bleaching performed prior to t ) 0 s. Table 1. Relative Activity of the Lipase Variants enzyme

activity (A405)

TLL-R TLL-R-Alexa488 TLL-β TLL-β-Alexa488

0.49 0.26 1.00 0.02

(unlabeled TLL-β). The activity after labeling decreased for both lipase variants to about 50% of the initial activity for TLL-R and only 2% of the initial activity for TLL-β. Lipase Adsorption to Trimyristin. TLL-R and TLL-β samples of 100-3000 and 50-2000 nM, respectively, were adsorbed according to the usual protocol. Each molecule of the two lipase variants was assumed to occupy the same surface area when adsorbed on the triglyceride surface. The surface fluorescence intensity values were corrected for the different degrees of labeling (0.95 for TLL-R and 0.60 TLL-β) and could be used as a relative measure of the adsorbed amount. The resulting adsorption isotherms can be seen in Figure 2A. The active TLL-R isotherm saturated more slowly and to a lower maximum amount, about 0.6 of the maximum amount found with TLL-β. Due to the large fraction of unlabeled TLL-β in the labeled TLL-β population, samples with different fractions of the labeled

population were adsorbed to the trimyristin surface in order to ensure that the affinity of the unlabeled and labeled population was equivalent. The stock population of labeled TLL-β, with a degree of labeling of 0.6, was diluted with unlabeled TLL-β down to a minimum of 10% of labeled stock population. All samples had the total TLL-β concentration of 500 nM. The results of relative surface intensity vs fraction of labeled population in bulk solution are displayed in Figure 2B and fitted to a linear function. Sequential Adsorptions of Lipase onto Trimyristin. In order to monitor differences in film properties with time after adsorption, sequential adsorptions of the two lipase variants were performed. TLL-β (250 nM) or TLL-R (1000 nM) was adsorbed according to the usual protocol, and after rinsing, the surface intensity after adsorption was monitored for 3 h under flow (10 µL/min); see Figure 3. In the case of TLL-β, the surface intensity was more or less constant after the first adsorption (Figure 3A). A second addition of 250 nM TLL-β 50 min after the first adsorption of 250 nM TLL-β stabilized during flow to an approximate 70% increase in surface intensity, the signal being relatively constant throughout the remaining time (Figure 3A). When TLL-R was initially adsorbed instead (Figure 3B), the signal during constant flow decreased 25-30% during 3 h after the adsorption. When a sequential addition of TLL-β to the surface pretreated with TLL-R was performed after 50 min, the adsorption signal stabilized at an approximate 20% increase compared to the surface

Lipase Mobility on Trimyristin

Figure 3. Stability of adsorbed layer during flow (solid lines) for labeled TLL-β (A) and labeled TLL-R (B). Sequential adsorption (O) of 250 nM labeled TLL-β, as indicated by an arrow, to the initially adsorbed layers of TLL-β (A) and labeled TLL-R (B), respectively.

intensity when no sequential addition was made (Figure 3B). The same behavior was observed when the sequential adsorption was performed 2 h instead of 50 min after the initial adsorption (data not shown). The concentrations of the two enzymes, 250 nM TLL-β and 1000 nM TLL-R, respectively, were chosen so that the adsorbed amount would be similar and equal to θ/θmax ) 0.5 in the TLL-β adsorption isotherm (Figure 2A). The intensity values were not corrected for the different degrees of labeling for TLL-R and TLL-β, and the intensity during all adsorption steps was excluded for clarity since this had a large contribution from lipases in bulk. Lipase Diffusion on Trimyristin. FRAP was performed under no-flow conditions on trimyristin surfaces with TLL-R or TLL-β adsorbed according to the adsorption protocol. The rinsing step after adsorption made sure that no lipase was present in the bulk solution to minimize contribution from exchange of lipases from bulk to the fluorescence recovery. Bleaching of two to three spots was performed every 30 min up to 3 h after adsorption so that the time evolution of the diffusion coefficient D and mobile fraction f could be studied. A representative recovery profile can be seen in Figure 1. This procedure was repeated on different trimyristin surfaces with different relative adsorbed amount of the two types of TLL. Assuming that Imax,TLL-β corresponded to the maximum surface density of TLL on trimyristin, θmax, the relative surface densities after adsorption, θ/θmax, could be calculated for each diffusion experiment. In Figure 4, the time evolutions of D and f are plotted for both TLL-R (Figure 4A) and TLL-β (Figure 4B). The results are displayed from the surfaces of lowest and highest θ/θmax. For

Langmuir, Vol. 23, No. 5, 2007 2709

Figure 4. (A) Time evolution of the diffusion coefficient D and mobile fraction f of TLL-R on trimyristin: (b) D for θ/θmax of 0.12, (O) f for θ/θmax of 0.12, ([) D for θ/θmax of 0.65, (]) f for θ/θmax of 0.65. D and f are plotted as functions of time after adsorption for surfaces of low surface density of TLL-R (circles) and high surface density (diamonds). (B) Same as (A) but for TLL-β: (b) D for θ/θmax of 0.18, (O) f for θ/θmax of 0.18, ([) D for θ/θmax of 1, (]) f for θ/θmax of 1. All data points are the average of at least two different surfaces, and the error bars represent the standard deviations.

TLL-R of θ/θmax) 0.12, the diffusion is very dependent on the time after adsorption, dropping from 13.7 × 10-11 cm2/s in the interval 0-15 min after adsorption to 4.4 × 10-11 cm2/s in the interval 165-180 min after adsorption (Figure 4A). This is accompanied with a decrease in mobile fraction, from f ) 0.61 to f ) 0.49. At the highest surface density of TLL-R (θ/θmax ) 0.65) the trend is similar but less pronounced, D decreasing from 2.2 × 10-11 to 1.1 × 10-11 cm2/s and f from 0.50 to 0.38. For all studied θ/θmax of TLL-R, D decreased with time after adsorption and f either decreased or remained more or less constant (all data not shown). There was also a time dependence on the diffusion of TLL-β (Figure 4B), however not as distinct as compared with TLL-R. On the surfaces of lowest TLL-β density (θ/θmax ) 0.18), D decreased from 4.6 × 10-11 to 1.6 × 10-11 cm2/s when comparing the first and the last time interval after adsorption. On the surfaces saturated with TLL-β (θ/θmax ) 1) the decrease was only from 1.8 × 10-11 to 0.9 × 10-11 cm2/s. Although, there was no time dependence of the mobile fraction, f was fluctuating between 0.4 and 0.6 (Figure 4B). The observed time dependence of D and irregularity of f was observed with all different θ/θmax of TLL-β. Plots of the surface density dependence of TLL diffusion are shown in Figure 5. For comparison with our earlier work on TLL diffusion, D measured at different surface densities has been averaged within two time intervals after adsorption, 0-90 and

2710 Langmuir, Vol. 23, No. 5, 2007

Figure 5. (A) The diffusion coefficient on trimyristin plotted vs relative surface density, θ/θmax, of TLL-R. The diffusion has been averaged in two time intervals after adsorption, 0-90 min (b) and 90-180 min (O). Solid and dashed lines are fits using an exponentially decaying function. Error bars are ( standard deviation. (B) Same as in (A) but for TLL-β.

90-180 min, denoted Dt)0-90 and Dt)90-180, respectively. Results for TLL-R had a strong dependence on surface coverage: Dt)0-90 decreasing from 10.6 × 10-11 to 1.5 × 10-11 cm2/s when θ/θmax increases from 0.12 to 0.65 (Figure 5A). For Dt)90-180 the trend was similar, D decreasing from 6.5 × 10-11 to 1.0 × 10-11 cm2/s comparing the lowest and highest θ/θmax. From this graph, it is also evident that the decrease in TLL diffusion with time after adsorption is more pronounced at low surface densities; i.e., the difference between Dt)0-90 and Dt)90-180 is more noticeable the lower the θ/θmax. For TLL-β, the magnitude of the diffusion is much lower compared to TLL-R (Figure 5B). Although, the trend was similar for this protein, i.e., Dt)0-90 decreasing from 3.2 × 10-11 to 1.2 × 10-11 cm2/s and Dt)90-180 from 1.9 × 10-11 to 0.8 × 10-11 cm2/s. For all plots in Figure 5, exponentially decaying functions were fitted to the data in order to estimate the diffusion coefficient at infinite surface dilution, D0. An exponential relation between diffusion and surface density is supported by theoretical predictions.19 Data were well fitted to an exponential function, and the results for D0 are displayed in Table 2. In Figure 6, the mobile fraction is plotted as function of surface coverage for both TLL-R and TLL-β and f is averaged in the same way as used for D in Figure 5. For both lipases, f was found to be in the range of 0.4-0.6. There was a small trend for TLL-R, where ft)0-90 and ft)90-180 decrease slightly from surfaces of low θ/θmax to high θ/θmax (Figure 6A). No such trend was visible in (19) Minton, A. P. Biophys. J. 1989, 55, 805-808.

Sonesson et al.

Figure 6. (A) The mobile fraction f on trimyristin plotted vs relative surface density, θ/θmax, of TLL-R. The mobile fraction has been averaged in two time intervals after adsorption, 0-90 min (b) and 90-180 min (O). Error bars are ( standard deviation. (B) same as in (A) but for TLL-β. Table 2. Diffusion Coefficients at Infinite Surface Density, D0, on Trimyristin lipase

D0,t)0-90 (cm2/s × 10-11)

D0,t)90-180 (cm2/s × 10-11)

f

TLL-R TLL-β

17.9 4.1

10.4 3.0

0.4-0.6 0.4-0.6

the plot for TLL-β. The mobile fraction was more or less constant with θ/θmax and with no pronounced deviations between ft)0-90 and ft)90-180 (Figure 6B).

Discussion The aim of this work was to analyze the mobility of active and inactive variants of Thermomyces lanuginosus (TLL) on a substrate surface using fluorescence recovery after photobleaching (FRAP). In our previous work, we have developed and presented a simple way of conducting FRAP on slow-diffusing proteins at surfaces using a rudimentary confocal microscope limited to bleach lines during the scan. The method was based on bleaching a large rectangular area and analyzing the process as a onedimensional relaxation of a step function.5 This method proved to be simple and averaged the diffusion coefficient over 220 µm regions of the model surfaces. However, using the rougher substrate surfaces of trimyristin, bleaching to a satisfying step function proved to be difficult. Therefore, we used the confocal microscope to conduct spot-FRAP experiments with a data analysis developed by, e.g., Axelrod et al.18 The analysis is based on the assumption that the bleached spots are uniform circles. However, since the scanning laser has a Gaussian bleach profile,

Lipase Mobility on Trimyristin

Langmuir, Vol. 23, No. 5, 2007 2711 Table 3. Reports on Surface Diffusion of Enzymes on Substrate Surfaces

the resulting edges of the bleached circle will have a Gaussianlike shape.20 The radial resolution of the laser is about 250 nm, which is similar to the selected pixel size of 200 nm. Therefore, this will not be visible in the microscopic images. However, since the radius of the circle (2-3 µm) is much larger than the resolution of the bleaching beam, this will have a minimal effect on the recovery process. It has been shown that the assumption of a circular region is valid if the radius of the bleached region is more than five times larger than that of the bleaching laser.20 As a substrate surface for TLL we used spin-coated films of trimyristin. Since the spin-coated films had a certain roughness, the surface diffusion coefficient estimated with FRAP is an underestimation of the true surface diffusion coefficient. The effective surface area of the trimyristin was about 30% larger compared to a theoretically two-dimensional smooth surface, when measured over 150 × 150 µm areas with white light profilometry. Thus, the area of the bleached circles calculated from the micrographs is underestimated. Due to the direct proportionality between area and recovery time, the calculated diffusion coefficients in this work should be higher in order for it to represent true two-dimensional surface diffusion coefficients. We have chosen not to compensate for this since we do not know exactly how much the area enlargement is for the bleached regions. In order to reduce this effect, the confocal microscope was used to select areas for bleaching that were subjectively smooth, i.e., that displayed a mainly homogeneous intensity signal. The structure of the wild-type lipase from the fungus Thermomyces lanuginosus (TLL) is well characterized and has been solved crystallographically by Brzozowski et al.3 The active site is defined by a catalytic triad (Ser, Asp, His) that attacks the ester bonds between the glyceride backbone and the acyl chains. Two variants of TLL were used in this study, denoted TLL-R and TLL-β. After being labeled with the Alexa488 dye, the activity of the two lipases differed significantly (Table 1). TLL-β lost almost all activity while TLL-R maintained 50% of its activity. Thus, labeled TLL-β was considered as inactive in this work. Adsorption of lipase to trimyristin was studied in the confocal microscope, and therefore absolute adsorbed amounts could not be determined. However, relative adsorption isotherms after 10 min adsorption were created using the fluorescence intensity values (Figure 2A). Both TLL-R and TLL-β adsorbed readily to the trimyristin surface, but with different affinity and to different maximum surface densities. By use of ellipsometry and SPR, these lipases have been found to adsorb to different maximum surface densities on hydrophobic C18-terminated surfaces:5,6 TLL-R to 1.75 mg/m2 and TLL-β to 2.20 mg/m2. These differences in affinities and maximum surface densities of the lipases could be explained by a different net charge of the enzymes. Thus, adsorbed TLL-R is expected to have higher lateral repulsion and adsorbed molecules of TLL-R and TLL-β were assumed to occupy the same surface area. The adsorption isotherms were used to derive the relative surface density, θ/θmax (Figure 2), on every surface before performing FRAP. In order to easily compare the results from the two lipases, all surface intensities with TLL-R or TLL-β were normalized to the maximum intensity found with TLL-β; i.e., the plateau value in the TLL-β isotherm was considered as θmax. Therefore, the absolute amount of lipase should be similar on TLL-R and TLL-β surfaces with the same θ/θmax. Moreover, since the Alexa488-labeled TLL-β population also consisted of a relatively large portion of unlabeled TLL-β (degree of labeling ) 0.6), a control with 500 nM samples with different fractions of the labeled population was performed (Figure

2B). The graph decreased linearly down to the origin of coordinates; i.e., at the studied concentration the affinity of labeled and unlabeled TLL-β to the trimyristin surface appeared similar. When the adsorbed amount after adsorption under flow conditions was studied, it was observed that some TLL-R was rinsed off the trimyristin surface, contrary to the inactive TLL-β. The signal from adsorbed TLL-R decreased almost 30% over 3 h after adsorption whereas the signal was more or less constant for TLL-β (Figure 3). This might imply that the active TLL-R binds with lower affinity to the trimyristin surface or that the catalytic activity includes desorption from the substrate. Another explanation to this might be related to the production of fatty acids in the interface region. Negatively charged myristic acid residues may accumulate at the triglyceride surface and increase the negative charge of the surface. This would increase the electrostatic repulsion between the myristic acid and negatively charged lipase. Further, preadsorption with the more negatively charged TLL-R or production of fatty acids could also explain, as compared to when the triglyceride surface was pretreated with the inactive TLL-β, the lower level of adsorbed TLL-β found after the second injection in the sequential adsorption (Figure 3). Since the signal after the second addition (Figure 3B) decreased in the same rate compared to when only TLL-R was adsorbed to the surface, it is believed that the decrease was mainly due to desorption of the more weakly bound TLL-R. Compared to other types of protein diffusion, lateral diffusion of proteins on solid surfaces is a slow process. A globular molecule of similar size as TLL (radius 23 Å) diffuses in free solution with D about 10-6 cm2/s, according to the Stokes-Einstein relation, and proteins bound to lipid membranes, such as receptors in a cell membrane, have been found to diffuse with D from 10-8 to 10-7 cm2/s .21 The lipase diffusion coefficients on the trimyristin surface reported in this study were between 10-11 to 10-10 cm2/s. Other reports on the diffusion of interfacially active enzymes on a substrate surface are sparse. There are to our knowledge only reports of a few other enzyme-substrate systems, which include FRAP measurements: cellulase on cellulose,7 subtilisin protease on a monolayer of BSA,8 collagenase on the peptide FALGPA,9 gelatinase on gelatine,10 and β-amylase on a starch gel.11 The studies report varying diffusion rates in the range of 10-9 to 10-12 cm2/s and are summarized in Table 3. The lipase surface diffusion rate in this study was in the same range as has been found earlier with cellulase and protease (Table 3). In our previous work, TLL-R and TLL-β have been found to have similar diffusion rate on silica surfaces modified with octadecyltrichlorosilane (OTS).5,6 On those hydrophobic model surfaces, the diffusion at infinite surface dilution, D0, was quantified and averaged over time spans of 0-90 min and 90-180 min after adsorption. D0,t)0-90 and D0,t)90-180 for TLL-R were both estimated to 0.8 × 10-11 cm2/s and for TLL-β, D0,t)0-90 and D0,t)90-180 were estimated to 0.8 × 10-11 and 0.5 × 10-11 cm2/s, respectively. Furthermore, the lateral mobility was found to depend on the hydrophobicity of the surface, leading

(20) Braeckmans, K.; Peeters, L. N. S. N.; De Smedt, S. C.; Demeester, J. Biophys. J. 2003, 85, 2240-2252.

(21) Vaz, W. L. C.; Goodsaid-Zalduondo, F.; Jacobson, K. FEBS Lett. 1984, 174, 199-207.

enzyme

substrate

D (cm2/s × 10-11)

f

cellulase7 subtilisin protease8 collagenase9 gelatinase A10 β-amylase11

cellulose BSA-monolayer FALGPA-peptide gelatin starch gel

2.0-12 14 20-98 110-310 ∼100

∼0.7 NA ∼0.85 ∼0.10 ∼0.55

2712 Langmuir, Vol. 23, No. 5, 2007

to higher diffusion rates on more hydrophilic surfaces. In this work the diffusion was found to be significantly faster for both lipases on the trimyristin surface compared to the model hydrophobic surface under the identical conditions (Table 2), even though the surface was apparently more hydrophobic (the contact angle of trimyristin was 120° compared to 105° of the OTS surface). The increase in diffusion for the inactive TLL-β must arise primarily from surface properties; on the basis of lack of activity toward pNP-valerate, we assume that this enzyme has very low activity toward trimyristin. The diffusion at maximum lipase surface density on the triglyceride film was for both lipase variants 1.0 × 10-11 cm2/s (Figure 4). On hydrophobic model surfaces, lipases adsorbed to maximum surface density has been found to be more or less immobile (1.0 × 10-12 cm2/s).5,6 A tentative explanation is therefore that the triglycerides themselves are mobile in the triglyceride film and thus facilitate enzyme mobility as the lipase move along with them. Another possibility would be that the spot bleaching raises the temperature from room temperature to the melting point of trimyristin (56 °C). However, the short bleach period of 10 s and the large volume of the surrounding bulk solution should ensure that the physical state of the substrate was not altered. A general trend in the diffusion behavior with both types of TLL is that the diffusion decreased with increasing relative surface density of protein (Figure 5). This has been observed with both types of lipases on model surfaces5,6 and with collagenase on collagen9 and β-amylase on starch gels.11 This phenomenon is supported by theoretical predictions19,22 and Brownian dynamics modeling23 and is explained by sterical hindrance of adsorbed molecules at high surface densities of protein. Further, the activity of the enzyme could be of importance for the diffusion. For TLL-R, D0,t)0-90 and D0,t)90-180 on trimyristin was estimated to 17.9 × 10-11 and 10.4 × 10-11 cm2/s, respectively, about four times higher compared to the diffusion rates of the inactive TLL-β (Figure 5 and Table 2). Apparently, the active lipase that is able to hydrolyze the underlying substrate can migrate faster across the surface compared to the inactive variant. The same trend was found in the cellulase and β-amylase systems but is contrary to the results found with collagenase on FALGPA.9 The mobility of the active lipase could involve desorption and rebinding to the trimyristin surface. Desorption was confirmed by studying the adsorbed amount after adsorption under flow conditions where a portion of the adsorbed lipases was rinsed of the surface (Figure 3B). During no-flow conditions, desorbed molecules could be expected to rebind to the surface. Thus, active lipases might desorb and rebind from the surface while exploring the substrate. Another possible explanation could be changes in film properties. QCM-D experiments on triolein surfaces24 have shown an increase in film thickness and decrease viscosity at high lipase activity on the surface. This was interpreted as swelling of the triolein film due to product accumulation of negatively charged oleic acid, which made the film more accessible to water. Thus, water incorporation due to myristic acid accumulation could change the film properties and thereby facilitate lipase migration on the surface. However, the low diffusivity of TLL-β could also be due to the higher affinity for the surface, as seen in the initial slopes in the adsorption isotherms in Figure 2A. On surfaces saturated with lipase, the diffusion of the active and inactive TLL is similar, about 1.0 × 10-11 cm2/s (Figure 4). This suggests that crowding effects dominate the diffusion rate at high surface density. (22) Scalettar, B. A.; Abney, J. R.; Owicki, J. C. Proc. Natl. Acad. Sci. U.S.A. 1988, 85, 6726-6730. (23) Ravichandran, S.; Talbot, J. Biophys. J. 2000, 78, 110-120. (24) Snabe, T.; Petersen, S. B. Chem. Phys. Lipids 2003, 125, 69-82.

Sonesson et al.

For both lipases, the diffusion decreased as a function of time after adsorption, especially on surfaces with low density of lipase (Figure 4). This has earlier been observed with TLL on two different hydrophobic model surfaces5 and was attributed to a rotation-diffusion equilibration of the adsorbed lipases. As a function of time, larger portions of the lipases are expected to be oriented in the open and more hydrophobic conformation state with the active site region toward the surface, thereby decreasing the observed diffusion rate. In the case of the active TLL-R an explanation for the decrease in diffusion rate might be product inhibition. The hydrolysis products generated at the surface can either be released in solution or remain in the substrate film.25 The myristic acid formed by TLL hydrolysis of trimyristin (tri-C14) has low solubility in water (0.2 ng/µL). The fatty acids are therefore expected to remain at the substrate surface considering the small chamber volume of about 15 µL. Accumulation of fatty acids at the substrate surface may induce molecular reorganizations of the surface that alters the hydrolysis rate of the enzyme.26,27 Since the TLL surface diffusion was measured under no-flow conditions, myristic acid formed during catalysis would not be transported from the surface by other means than diffusion. Hence, product inhibition may contribute to the decrease in diffusion rate with time seen with the active lipase, TLL-R, at low θ/θmax. The pKa values of monomeric fatty acids in water have been reported to 4.8;28however the apparent pKa values can change dependent on the conditions. As an example, fatty acids in lipid bilayers have been found to have pKa between 7.0 and 7.5,29,3030 primarily explained by the dielectric effect of the nonpolar environment. Since the diffusion was measured in a calcium-free buffer at pH 9, it could be assumed that the produced myristic acid is deprotonated and negatively charged. This is in line with the sequential adsorption experiments (Figure 3) as discussed above. While the diffusion coefficient is a measure of the mobility of the enzyme, the mobile fraction is a reflection of the different adsorption states on the surface. The mobile fractions of the diffusing TLL-R and TLL-β on trimyristin surfaces were similar, varying between 0.4 and 0.6. Similar results were found previously on the model hydrophobic surfaces.5,6 For other enzymesubstrate systems there seems to be no correlation between mobile fraction and diffusion rate (Table 3) or mobile fraction and activity. The nature of the immobile molecules of the studied lipases is unclear. One possibility could lie in the nature of the substrate; i.e., lipases could after adsorption be trapped in discontinuities of the trimyristin surface that prevents lateral diffusion. This would explain that the immobile fraction was almost invariable with θ/θmax for both lipase variants (Figure 6). The immobile fraction could also reflect lipases adsorbed in an orientation that minimizes the free energy of adsorption, which would render them immobile. However, in that case one would have expected a more distinct decrease in mobile fraction with time after adsorption, which was not seen (Figures 4 and 6). Another possible explanation could be that the immobile fraction is a subpopulation (25) Flipsen, J. A. C.; van der Hijden, H. T. W. M.; Egmond, M. R.; Verheij, H. M. Chem. Phys. Lipids 1996, 84, 105-115. (26) Balashev, K.; Gudmand, M.; Iversen, L.; Callisen, T. H.; Svendsen, A.; Bjornholm, T. Biochim. Biophys. Acta 2003, 1615, 93-102. (27) Peters, G. H.; Dahmen-Levison, U.; de Meijere, K.; Brezesinski, G.; Toxvaerd, S.; Mo¨hwald, H.; Svendsen, A.; Kinnunen, P. K. J. Langmuir 2000, 16, 2779-2788. (28) Cistola, D. P.; Hamilton, J. A.; Jackson, D.; Small, D. M. Biochemistry 1988, 27, 1881-1888. (29) Ptak, M.; Egret-Charlier, M.; Sanson, A.; Bouloussa, O. Biochim. Biophys. Acta 1980, 600, 387-397. (30) Hamilton, J. A.; Cistola, D. P. Proc. Natl. Acad. Sci. U.S.A. 1986, 83, 82-86.

Lipase Mobility on Trimyristin

of the lipases that is unstable, possibly caused by labeling at unfavorable sites that leads to unfolding or aggregation on the surface.

Conclusions We have shown that the mobility of Thermomyces lanuginosus lipase (TLL) is significantly enhanced on a substrate surface compared to our earlier work on hydrophobic model surfaces. The mobility of fluorescently labeled active and inactive TLL variants on trimyristin was quantified with FRAP. The diffusion coefficient and mobile fraction were measured at different surface densities and time after adsorption. When extrapolated to infinite surface dilution, D0, the TLL diffusion was estimated in the range of (3.0-17.4) × 10-11 cm2/s, in the same range as work

Langmuir, Vol. 23, No. 5, 2007 2713

with other enzyme systems. The active lipase variant was found to be more mobile on the substrate compared to the inactive variant, suggesting that the catalytic activity increased the ability for a lipase to explore the substrate surface. The diffusion rate for the inactive and in particular the active TLL variants was affected by both relative adsorbed amount and time after adsorption. For the active TLL, the latter could be due to product inhibition or electrostatic repulsion when negatively charged myristic acid residues accumulate at the surface. Acknowledgment. Novozymes A/S is acknowledged for financial support of the project. LA062003G