Fluorescence spectroscopic characterization of Humicola lanuginosa lipase dissolved in its substrate

Biochimica et Biophysica Acta 1702 (2004) 181 – 189 http://www.elsevier.com/locate/bba

Fluorescence spectroscopic characterization of Humicola lanuginosa lipase dissolved in its substrate Arimatti Jutila*, Keng Zhu, Esa K.J. Tuominen, Paavo K.J. Kinnunen Helsinki Biophysics and Biomembrane Group, Institute of Biomedicine/Biochemistry, P.O. Box 63 (Haartmaninkatu 8), FIN-00014 University of Helsinki, Finland Received 17 February 2004; received in revised form 4 August 2004; accepted 18 August 2004 Available online 8 September 2004

Abstract The conformational dynamics of Humicola lanuginosa lipases (HLL) and its three mutants were investigated by steady state and timeresolved fluorescence spectroscopy in two different media, aqueous buffer and the substrate triacetin. The fluorescence of the four Trps of the wild-type HLL (wt) reports on the global changes of the whole lipase molecule. In order to monitor conformational changes specifically in the a-helical surface loop, the so-called dlidT of HLL comprised of residues 86–93, the single Trp mutant W89m (W117F, W221H, W260H) was employed. Mutants W89L and W89mN33Q (W117F, W221H, W260H, N33Q) were used to survey the impact of Trp89 and mannose residues, respectively. Based on the data obtained, the following conclusions can be drawn. (i) HLL adapts the dopenT conformation in triacetin, with the a-helical surface loop moving so as to expose the active site. (ii) Trp89 contained in the lid plays an unprecedently important role in the structural stability of HLL. (iii) In triacetin, but not in the buffer, the motion of the Trp89 side chain becomes distinguishable from the motion of the lid. (iv) The carbohydrate moiety at Asn33 has only minor effects on the dynamics of Trp89 in the lid as judged from the fluorescence characteristics of the latter residue. D 2004 Elsevier B.V. All rights reserved. Keywords: Humicola lanuginosa lipase; Triacetin; Fluorescence spectroscopy

1. Introduction Lipases (triacylglycerol hydrolases, E.C. 3.1.1.3) constitute a large family of enzymes and are widely distributed in Nature. True lipases are distinguished from esterases by their characteristic interfacial activation at lipid–water interfaces [1,2]. Several mechanisms have been forwarded to explain this property [3,4]. These include an increased substrate Abbreviations: EDTA, ethylenediaminetetraacetic acid; HEPES, N-(2hydroxyethyl) piperazine-NV-2-ethanesulfonic acid; HLL, Humicola lanuginosa lipase; RFI, relative fluorescence intensity; W89m, single Trp HLL mutant with substitutions W117F, W221H, and W260H; W89mN33Q, single Trp HLL mutant with substitutions W117F, W221H, W260H, and N33Q; wt, wild type; s, fluorescence lifetime; /, rotational correlation time; r l, residual anisotropy * Corresponding author. Tel.: +358 9 19125427; fax: +358 9 19125444. E-mail addresses: [email protected] (A. Jutila)8 [email protected] (P.K.J. Kinnunen). 1570-9639/$ - see front matter D 2004 Elsevier B.V. All rights reserved. doi:10.1016/j.bbapap.2004.08.011

concentration at the interface [5], a better orientation of the scissile ester bond [6], a reduction in the water shell around the ester molecule in water [7], as well as conformational changes of the enzyme leading to an optimized active site geometry and enhanced catalytic activity [1,8]. The amino acid sequence of Humicola lanuginosa lipase (HLL) consists of 269 residues, including four tryptophans [9]. The crystal structure of HLL has been solved at 1.8-2 resolution [10]. Accordingly, HLL consists of a single, roughly spherical domain containing a central eightstranded, predominately parallel h-pleated sheet, with five interconnecting a-helices, compacted to a volume of approximately 9.7
103 23. The active site contains a Ser(146)–Asp(201)–His(258) catalytic triad, closely reminiscent to that seen in serine proteases [11]. The active site Ser is additionally involved in maintaining the structure of HLL and its substitution (S146A) leads to substantial conformational alterations as well as different substrate binding affinities [12]. The a-helical surface loop (amino

182

A. Jutila et al. / Biochimica et Biophysica Acta 1702 (2004) 181–189

acids 86–93) constituting the so-called dlidT lies directly over the active site S146 [11,13], and is highly mobile in the crystals [14]. The lid of HLL contains a single Trp at position 89. Site-directed mutagenesis [15,16] showed that Trp89 is critical for efficient hydrolysis of tributyrin and that it plays a role in the catalytic steps subsequent to the absorption of the lipase to the substrate interface. Trp89 has been shown to influence in a qualitative manner the binding of substrates into the active site [16]. Two distinct conformations, dclosedT and dopenT, inactive and active, respectively, have been proposed for HLL [17]. Accordingly, in an aqueous medium, access to the catalytic triad is blocked by the lid and the conformation is closed. Martinelle et al. [9] reported that HLL displays a pronounced interfacial activation in the presence of the substrate p-nitrophenyl butyrate (PNPB). This would be associated with a significant conformational change of the lid, its opening generating a large hydrophobic surface, which also includes the active site. The ability of HLL to efficiently form the acyl-enzyme intermediate at the interface thus requires a number of separate steps: (i) the enzyme must adapt the active, lid open conformation, (ii) the lid and the hydrophobic surface has to attach to the interface, and (iii) the active site has to be saturated with the substrate [15]. Importantly, in complexes with substrate analogues or serine protease inhibitors the lid helix is displaced and the active site is exposed, yielding the open conformation [18–20]. To this end, not only lipids [13,21] but also detergents [22] and i-propanol, an organic solvent [23], can induce the opening of the lid of HLL. It is generally difficult to obtain conditions where the chemical equilibrium for the binding of lipase to the substrate would be shifted quantitatively to the presence of the lipase–substrate complex. In the present study we utilized fluorescence spectroscopy to compare the conformations of HLL in an aqueous buffer and dissolved in its substrate, triacetin. Under the latter conditions we can expect the enzyme to be maximally saturated with the substrate. Moreover, triacetin is optically transparent and does not impede the use of fluorescence spectroscopy techniques. Both steady state and time-resolved analyses were employed to monitor the changes in the fluorescence of lipases. HLL contains approximately 20 mannose residues, glycosylated to Asn33. The variant with four mutations W89mN33Q lacks the glycosylation site and thus also the carbohydrate moiety (S.A. Patkar, personal communication). Accordingly, W89m and W89mN33Q were made to study the impact of the carbohydrate moiety on the dynamics of the lid.

2. Materials and methods

(W117F, W221H, W260H, N33Q) were obtained from Novo Nordisk (Bagsv&rd, Denmark). Their concentrations were determined as described previously [24]. Specific activities of wt HLL, W89L, W89m, and W89mN33Q were 2158, 1587, 1130 and 1016 Amol/mg/min, respectively [17]. The buffer used in the experiments was 20 mM HEPES, 0.1 mM EDTA, pH 7.0 (adjusted with NaOH), prepared in freshly deionized (Milli RO/Milli Q, Millipore, Bedford, MA) water. Fluorescence intensity of HLL in triacetin increased linearly with the lipase concentration indicating ideal solubility. 2.2. Stationary fluorescence spectroscopy Steady-state fluorescence measurements were carried out using a Perkin Elmer LS 50B spectrofluorometer equipped with a magnetically-stirred, thermostated cuvette compartment. The excitation wavelength for Trp was 295 nm [25,26] with both excitation and emission bandpasses set at 5 nm. To improve signal-to-noise ratio in the steady-state anisotropy measurements, excitation and emission bandpasses were 10 nm, and the emission wavelength was varied between 332 and 346 nm. 2.3. Time-resolved fluorescence spectroscopy Commercial laser spectrometer (Photon Technology International, Ontario, Canada) was used to measure fluorescence lifetimes, rotational correlation times, and residual anisotropies. The minimum lifetime accessible to the instrument is 200 ps. A train of 500 ps excitation pulses at 337 nm at a repetition rate of 10 Hz was produced by a nitrogen laser, pumping 5 mM solution of rhodamine 6G (Merck, Darmstadt, Germany) in methanol. Pulses from the lasing dye solution emitting at 590 nm were frequency doubled for Trp fluorescence measurements at excitation wavelength of 295 nm. Depending on the sample, the emission wavelength was varied between 332 and 346 nm using a monochromator. For the determination of the fluorescence lifetimes, the averages of five emission decay curves were analysed using the software provided by the instrument manufacturer. Instrument response functions were measured separately using aqueous glycogen solution. The validity of the fit was judged by the value of the reduced chi-square, v 2 [27,28] which varied in the range of 0.9–1.2. Fractional intensities I(t) were calculated according to the equation:
X N I ð t Þ ¼ ai s i ai s i ð1Þ i¼1

2.1. Materials

where a i is amplitude and s i is lifetime. For polarized light, the decay of fluorescence intensity, F(t) was calculated from the raw data according to:

H. lanuginosa lipase (HLL) and its mutants W89L, W89m (W117F, W221H, W260H), and W89mN33Q

F ðt Þ ¼ Ijj ðt Þ þ 2
G
I8 ðt Þ

ð2Þ

A. Jutila et al. / Biochimica et Biophysica Acta 1702 (2004) 181–189

where I t (t) is the intensity of light detected with both the excitation and emission polarizer being vertical (i.e., parallel polarizers) and I 8(t) is the intensity of light detected with vertical excitation polarizer and horizontal emission polarizer (i.e., perpendicular polarizers), and G is the correction term for the relative throughput of the respective polarized light component through the emission optics. Time-resolved anisotropy r(t) is defined as:     rðt Þ ¼ Ijj ðt Þ  G
I8 ðt Þ = Ijj ðt Þ þ 2
G
I8 ðt Þ

ð3Þ

where G, I t and I 8 are as above. The fluorescence and anisotropy decays are described by a sum of exponentials, as follows: X F ðt Þ ¼ Ai et=si ð4Þ i

r ðt Þ ¼

X

ri et=/i

ð5Þ

i

In the above A i and r i are the normalized pre-exponential initial fluorescence intensities and initial anisotropies, respectively, and s i and / i are the corresponding fluorescence lifetimes and rotational correlation times. When the angular range of the rotational motion of fluorophore is limited rðt Þ ¼ ðr0  rl Þet=/ þ rl

183

wild-type HLL contains four Trps residing in positions 89, 117, 221, and 260, which combined report on more global changes in this lipase. Regarding the interaction of HLL with its substrate and the conformational change causing enzyme activation, the recently introduced single Trp mutant W89m is of particular interest [22,23,29,30]. Furthermore, mutants W89L and W89mN33Q enable studies on the importance of Trp 89 and mannose residues, respectively, for the structure of HLL. Fluorescence spectra for these lipases are shown in Fig. 1 and the measured photophysical parameters are compiled in Figs. 2–4. 3.2. Emission intensity and k max The wavelength of the emission maximum k max and the fluorescence quantum yield for Trp depend on its microenvironment. In brief, with increase in polarity k max shifts to longer wavelength and quantum yield decreases [31,32]. Notably, the fluorescence quantum yield of W89m is high and the emission of its single Trp is approximately 60% of that measured for the wild-type HLL with four Trps [22].

ð6Þ

where r 0 stands for the anisotropy in the absence of rotational diffusion. Eq. (6) was used only to fit the data so as to obtain the value of the residual anisotropy. Rotational correlation times (/ i) and residual anisotropies (r l) were derived from the fitted curves. The decays I(t)t and I(t)8 were fitted before calculating the respective r(t) decays from these data. Accordingly, r(t) contains no experimental noise and therefore criteria such as the v 2 value do not apply. For the determination of rotational correlation times and residual anisotropies, each emission decay was measured 10 times.

3. Results 3.1. General considerations Trp residues provide intrinsic fluorescent probes for protein structure, allowing for detailed studies on their conformational dynamics. We used both steady state and time-resolved fluorescence spectroscopy to compare HLL dissolved in an aqueous buffer and in its substrate triacetin. The pH of triacetin is approximately 5.5, whereas that of the buffer used in this study is 7.0. To check the pH effect we recorded steady state fluorescence spectra at these two pHs in buffer. These spectra were identical verifying that the differences between the spectra recorded in buffer and triacetin are due to the conformational changes in HLL. The

.

Fig. 1. Fluorescence spectra of 2 AM wt HLL ( ), W89L (4), W89m (n), and W89mN33Q (5) in 20 mM HEPES, 0.1 mM EDTA, pH 7.0 (panel A) and in triacetin (panel B) at 25 8C. The concentration of water in the latter medium was approximately 0.5 vol.%. The excitation wavelength was 295 nm, both excitation and emission bandpasses were 5 nm.

184

A. Jutila et al. / Biochimica et Biophysica Acta 1702 (2004) 181–189

This is likely to be explained by the low quantum yields of the other three Trps, Trp117 and Trp221 in particular. More specifically, the latter two reside in locations with both anionic and cationic residues in their immediate vicinity, which could be sufficient to perturb the electronic configuration of the Trps so as to cause efficient quenching of their fluorescence [30]. For wt HLL, W89L, W89m, and W89mN33Q in buffer values for k max are 342, 332, 345, and 346 nm, respectively (Fig. 2A). As expected, in the less polar triacetin a blue shift by 2, 8, and 8 nm in k max was observed for wt, W89m, and W89mN33Q, respectively. On the other hand, fluorescence intensity was significantly lower in triacetin for these three lipases (Fig. 1), in keeping with reduced quantum yield of Trp89 in triacetin. Accordingly, the opposite was true for W89L as a 6-nm red shift and an increase in emission intensity were observed, in keeping with decreased interactions between Trps and their vicinal charged amino acids in triacetin. 3.3. Stationary fluorescence anisotropy

Fig. 2. Wavelength of the maximum emission intensity (panel A) and steady state fluorescence anisotropy (panel B) for the studied lipases in buffer (open columns) and in triacetin (striped columns).

Rotational diffusion of fluorophores is the dominant cause of fluorescence depolarization. In proteins the mobility of fluorescent amino acid residues bears a close relationship with the overall state of a protein, and any factor which affects its size, shape, state, or segmental flexibility will also affect the observed anisotropy value

Fig. 3. Fluorescence lifetimes s 1 (panel A) and s 2 (panel B) and the contribution of the latter (panel C) for the lipases in buffer (open columns) and triacetin (striped columns). Panel D shows the average fluorescence lifetimes in these two media.

A. Jutila et al. / Biochimica et Biophysica Acta 1702 (2004) 181–189

185

Fig. 4. Rotational correlation times / 1 (panel A), / 2 (panel B), and / 3 (panel C) and residual anisotropy r l (panel D) for the studied lipases in buffer (open columns) and in triacetin (striped columns).

[33]. Accordingly, anisotropy reports on both conformational changes and the state of association of a protein. When the medium was changed from the aqueous buffer to the more viscous triacetin, the steady state anisotropy (r) values increased for wt, W89m, and W89mN33Q, with most pronounced change for the latter, from 0.124 to 0.188 (Fig. 2B). Interestingly, W89L stands again apart from the other studied lipases as the value of r decreased from 0.180 to 0.147 in triacetin. Accordingly, W89L is least compactly folded in this solvent with the neighboring amino acids being less restrictive to the motions of the three Trps.

respectively, emerging in the latter medium. For wt, W89m, and W89mN33Q the longer fluorescence lifetime (s 2) component decreased in triacetin (Fig. 3B). Instead, for W89L the value for s 2 increased from 2.15 to 3.76 ns in this solvent. The fractional intensity of the longer fluorescence lifetime (s 2%) of the wt and W89L increased from 82% and 58%, respectively, to 91% for both (Fig. 3C). For W89m and W89mN33Q, s 2% decreased from 100% to 98% and 91%, respectively.

3.4. Fluorescence lifetimes

Time-resolved fluorescence anisotropy can provide information on the diffusive motions of a fluorophore during its excited lifetime. More specifically, these data can reveal whether a fluorophore is free to rotate over all angles, or if the surroundings of the fluorophore restrict its angular Brownian motion. Moreover, these measurements allow to distinguish between decays of anisotropy due to a single process, and those involving multiple modes of rotations [33]. We measured fluorescence anisotropy decays of these lipases in the aqueous buffer and triacetin, and calculated the rotational correlation times (/) and residual anisotropies (r l) from these data (Fig. 4). In the aqueous medium, two rotational correlation times were required for all four proteins for satisfactory fit of the data. The long rotational correlation time (/ 3) varied between 20 and 24 ns (Fig. 4C), while the short rotational correlation time (/ 1) was b1 ns (Fig. 4A). However, in triacetin the anisotropy decays

Not only the spectral features but also the fluorescence lifetimes of Trp depend on its microenvironment, thus allowing to obtain further insight into the conformational changes of a protein. Average fluorescence lifetimes (s¯ ) decreased from 3.95, 5.40, and 6.37 to 2.84, 3.58, and 3.51 ns for the wt HLL, W89m, and W89mN33Q, respectively, whereas for W89L s¯ increased from 1.58 to 3.43 ns (Fig. 3D). Two-exponential fluorescence decays were measured for wt and W89L both in buffer and triacetin. The shorter fluorescence decay (s 1) was significantly faster in triacetin than in the buffer (Fig. 3A). Yet, the measured values (0.12 and 0.09 ns, respectively) are already beyond the resolution of the instrument. W89m and W89mN33Q exhibited singleexponential decays in buffer and two-exponential in triacetin, with the shorter component of 0.33 and 0.76 ns,

3.5. Rotational correlation times and residual anisotropy

186

A. Jutila et al. / Biochimica et Biophysica Acta 1702 (2004) 181–189

became three-exponential for wt, W89m, and W89mN33Q, whereas for W89L a two-exponential anisotropy decay was measured also in triacetin. The long rotational correlation times (/ 3) were increased in triacetin by 3.3, 23.9, 3.4, and 1.8 ns for wt, W89L, W89m, and W89mN33Q, respectively (Fig. 4C). In triacetin the short rotational correlation times (/ 1) for the wt, W89m, and W89mN33Q were less than 1 ns, and the intermediate rotational correlation times (/ 2) were in the range of 3.9–6.6 ns (Fig. 4B). For W89L, the value for / 1 was 3.33 ns. The residual anisotropies (r l) for all these lipases revealed the same behaviour, the measured values decreasing upon the transfer of the lipases from the aqueous buffer into triacetin (Fig. 4D).

4. Discussion A major problem in studying substrate induced changes in an enzyme such as lipase is due to the difficulty in obtaining the enzyme to be quantitatively associated with the substrate. This problem was alleviated in the present study by dissolving HLL in its substrate, triacetin. Moreover, the concentration of water is very low, c0.28 M (approximately 0.5 vol.%), allowing maximally about 6.5% of the total triacetin present to be hydrolyzed. Accordingly, we can expect most of the enzyme to be in the form of the acyl-enzyme intermediate. The latter state should be associated with significant conformational changes especially in the lid of HLL. Wt HLL contains four Trps two of which, Trp89 and Trp117, are located in two different ahelices and the other two, Trp221 and Trp260, in two different h-pleated sheets. Changes in the intrinsic fluorescence of the wt thus include contributions from all four Trps and therefore report on the global conformational changes of this protein. Accordingly, particular interest was in the single Trp mutant W89m, reporting on the changes in the dlidT. In order to survey possible effects of the carbohydrate moiety on the conformational dynamics of the lid, the mutant W89mN33Q was employed. Tryptophan emission is sensitive to its microenvironment. Comparison of k max in triacetin and aqueous buffer revealed blue shifts of 2, 8, and 8 nm for HLL, W89m, and W89mN33Q, respectively, in keeping with triacetin being less polar than water [31]. On the other hand, fluorescence intensity was significantly lower in triacetin for these three lipases (Fig. 1). This finding is in apparent contradiction with the quantum yield of Trp fluorescence being inversely proportional to the polarity of microenvironment [32]. We attribute this decrement to quenching of the Trp89 fluorescence in triacetin. Fluorescence lifetime of Trp is related to its microenvironment and reduced polarity of the medium is associated with longer lifetimes [34]. Interest¯ ingly, the average fluorescence lifetimes s of wt, W89m, and W89mN33Q decreased in triacetin, thus indicating the quenching of Trp89 in triacetin to be dynamic [33]. In buffer, k max of W89L is at 332 nm suggesting the three Trps

of this lipase to reside in hydrophobic microenvironments [31]. In contrast to the other studied lipases, there is a red shift by 6 nm, a threefold increase in fluorescence intensity, ¯ and an increased s for the emission of W89L in triacetin. These effects are probably due to decreased interactions between Trps and their vicinal charged amino acids such as His, Lys, Arg, Glu, and Asp [30] upon increase in hydrodynamic volume of the lipase. This would result in decreased quenching and prolonged fluorescence lifetime for this mutant. Time-resolved fluorescence studies provide further insight into the changes in conformational dynamics induced by triacetin. Intensity decay of W89m and W89mN33Q was one-exponential in the aqueous buffer, whereas two-exponential fitting was required in triacetin. Two populations of Trp(s) with different fluorescence lifetimes have been ascribed to different Trp rotational isomers [35]. In proteins, the shorter fluorescence lifetime component s 1 has been further suggested to reflect interactions of the surface-exposed Trp(s) with the solvent [36]. The crystal structure of the open conformation of HLL revealed Trp89 to be exposed on the protein surface [37], in accordance with the appearance of s 1 in triacetin. Similar behaviour has been observed for protein tyrosine phosphatase (PTPase) which exhibits two-exponential emission decay in the dopenT conformation [38]. Decrease in steady state fluorescence anisotropy of a fluorophore indicates augmented mobility and vice versa [33]. Judged from these data on W89m and W89mN33Q, the movements of Trp89 became more restricted in triacetin. As Trp89 locates on the surface of HLL in the open conformation [29], the increased r is probably due to the increased viscosity of the medium. An opposite change is observed for W89L indicating that the average movements of Trps 117, 221, and 260 became less restricted. It can be concluded that in W89L the interactions of these three Trps with their vicinal amino acids are weaker in triacetin. Time-resolved fluorescence anisotropy yields information on the rapid local motions of the Trp side chains as well as the overall rotational diffusion of the entire protein [39]. Two rotational correlation times (/ 1 and / 3) were measured for these four lipases in buffer. However, three rotational correlation times (/ 1, / 2, and / 3) were required for a satisfactory fit for wt, W89m, and W89mN33Q in triacetin. In buffer, the short correlation time / 1 can be assigned to the local segmental motions of Trp(s), [40], i.e., movements of side chains, a-helices and h-pleated sheets containing Trp(s) [38]. As the lid of HLL is highly mobile in the crystals [14], it can be expected that the rates of motion of the Trp89 side chain and the lid are relatively close. Consequently, these cannot be distinguished by timeresolved anisotropy measurement, and in the time range shorter than 10 ns only one rotational correlation time is likely to be present. As indicated by steady state anisotropy measurements the three Trps of W89L in buffer are located inside the protein, in surroundings more restricting than that

A. Jutila et al. / Biochimica et Biophysica Acta 1702 (2004) 181–189

for Trp89. Accordingly, the local motions of Trps 117, 221, and 260 would be slower than Trp89, as revealed by timeresolved anisotropy measurements showing the rotational correlation time / 1 for W89L to be longer than that for W89m or W89mN33Q. In a medium of higher viscosity, e.g., triacetin, the motions of the lid can be expected to be slower, and these two different modes of local motion thus become separable. Under these conditions it is feasible to attribute the short and the medium correlation times / 1 and / 2 to the motion of Trp89 side chain and the lid, respectively. In triacetin the motions of Trps and a-helices and h-pleated sheets of W89L could not be distinguished, and only one short rotational correlation time was required for a satisfactory fit. In triacetin the motions of both Trp89 and lid are slower for W89mN33Q than for W89m, as indicated by the somewhat larger values for / 1 and / 2, respectively. This shows that also the carbohydrate moiety affects the conformational dynamics of HLL. Non-zero values of residual anisotropy r l have been interpreted to result from an energy barrier preventing rotational diffusion of the fluorophore beyond a certain angle within the fluorophore lifetime [33]. For all the lipases in the present study, the value of r l was lower in triacetin than in the aqueous buffer, indicating diminished constraints for mobility of the fluorophores. These data would be compatible with Trp89 to be located on the protein surface in the open conformation [29]. The long rotational correlation time / 3 reflects the global motions of the entire lipase, and thus it is related to the hydrodynamic volume of the molecule by the Einstein– Stoke’s equation /3 ¼ gM ðv þ hÞ=RT

187

dynamics of the lipase as shown by the emission of Trp89. Comparison of the single Trp mutants W89m and W89mN33Q reveals that the absence of the mannose residues causes the microenvironment of Trp89 to become more hydrophilic, indicated by the decrease in emission intensity (Fig. 1) and a red shift in k max (Fig. 2A). This could result from more loose packing of HLL and enhanced access of water to Trp89. This conclusion is supported by the lower steady state anisotropy value (Fig. 2B) and shorter rotational correlation time / 1 (Fig. 4A) measured for W89mN33Q in aqueous buffer. In the more viscous medium triacetin the observed steady state anisotropy value is likely to be dominated by the global motions of the lipase. Thus, it is rational that the glycosylated mutant W89m with tighter packing and smaller radius of hydration has a somewhat lower anisotropy value. The long rotational correlation time / 3 (Fig. 4C) obtained from time-resolved anisotropy measurements should measure directly the global motion of the lipase. However, no significant changes in this parameter were observed in either medium. This contradiction may be apparent only as changes in steady state anisotropy are very small, and, as the global motions of HLL are likely to contribute to / 2 (Fig. 4B), the value of the latter being decreased upon mutation N33Q. In buffer the average fluorescence lifetime for W89mN33Q is longer than that for W89m (Fig. 3D). As a more hydrophilic microenvironment of Trp usually decreases its fluorescence lifetime, the explanation for this difference remains unknown at present. However, the longer fluorescence lifetime provides a likely reason for the lower residual anisotropy value measured for W89mN33Q (Fig. 4C).

ð7Þ

where g is the viscosity of the medium at temperature T, M is the molecular mass of protein, v is the specific volume of protein, h is the hydration factor, and R is the gas constant. Using a value of 0.89 cp for the viscosity of water at 25 8C [41], / 3 of 20.11 ns measured for W89m in buffer corresponds to a hydrodynamic radius of 27.8 2. This is in excellent agreement with the value of 28 2 (=the maximum distance from the center of mass) obtained from crystal structure [42], thus excluding the possibility of oligomerization of HLL. Viscosity of the medium slows down the rotations of molecules, and, as expected, the long rotational correlation times / 3 for all the lipases were augmented in triacetin. For W89L the value for / 3 was increased twofold, indicating a very dramatic increase in the hydrodynamic volume of the rotating unit. This can be attributed to either aggregation or increase in volume of single protein. The latter seems more likely, as Trp89 is important for the structural stability of HLL [30]. HLL contains a single carbohydrate moiety linked to Asn33. The function of the sugar remains unknown. Deletion of the carbohydrate residues of HLL by the mutation N33Q has only minor effects on the structural

5. Conclusions The data reported here demonstrate open conformation of HLL to be induced by its substrate triacetin. In triacetin, W89L increased in volume significantly revealing Trp89 to play an important role in the structural stability of HLL. In an aqueous medium, the motions of the W89 side chain and the lid are not separable by time-resolved fluorescence anisotropy measurements. However, the difference between the frequencies of these two modes of motion is augmented in triacetin, allowing to distinguish between these two modes. Accordingly, the rotational correlation times for Trp89 side chain and the lid, approximately 0.2 and 5 ns, respectively, could be resolved by fluorescence anisotropy. The carbohydrate moiety attached to Asn33 also contributes on the conformational dynamics of the lid, although only to a minor extent.

Acknowledgments We thank Drs. Allan Svendsen and Shamkant Anant Patkar (Novo Nordisk, Bagsv&rd, Denmark) for providing

188

A. Jutila et al. / Biochimica et Biophysica Acta 1702 (2004) 181–189

the lipases. This study was supported by the Finnish Medical Research Council and by the EU Structural Biology Programme (BIO4 CT972365).

References [1] L. Sarda, P. Desnuelle, Action de la lipase pancreatique sur les esters en emulsion, Biochim. Biophys. Acta 30 (1958) 513 – 521. [2] W.A. Pieterson, J.C. Vidal, J.J. Volwerk, G.H. de Haas, Zymogencatalyzed hydrolysis of monomeric substrates and the presence of a recognition site for lipid–water interfaces in phospholipase A2, Biochemistry 13 (1974) 1455 – 1460. [3] J.W. Simons, J.W. Boots, M.P. Kats, A.J. Slotboom, M.R. Egmond, H.M. Verheij, Dissecting the catalytic mechanism of staphylococcal lipases using carbamate substrates: chain length selectivity, interfacial activation, and cofactor dependence, Biochemistry 36 (1997) 14539 – 14550. [4] J.J. Volwerk, G.H. de Haas, Pancreatic phospholipase A2: A model for membrane-bound enzymes? in: O.H. Griffith, P.C. Jost (Eds.), Molecular Biology of Lipid–Protein Interactions, Wiley, New York, 1982, pp. 69 – 149. [5] H.L. Brockman, J.H. Law, F.J.J. Kezdy, Catalysis by adsorbed enzymes. The hydrolysis of tripropionin by pancreatic lipase adsorbed to siliconized glass beads, J. Biol. Chem. 248 (1973) 4965 – 4970. [6] M.A. Wells, The mechanism of interfacial activation of phospholipase A2, Biochemistry 13 (1974) 2248 – 2257. [7] H. Brockerhoff, Substrate specificity of pancreatic lipase, Biochim. Biophys. Acta 159 (1968) 296 – 303. [8] B. Entressangles, P. Desnuelle, Action of pancreatic lipase on monomeric tripropionin in the presence of water-miscible organic compounds, Biochim. Biophys. Acta 341 (1974) 437 – 446. [9] M. Martinelle, M. Holmquist, K. Hult, On the interfacial activation of Candida antarctica lipase A and B as compared with Humicola lanuginosa lipase, Biochim. Biophys. Acta 1258 (1995) 272 – 276. [10] U. Derewenda, L. Swenson, R. Green, Y. Wei, G.G. Dodson, S. Yamaguchi, M.J. Haas, Z.S. Derewenda, Current progress in crystallographic studies of new lipases from filamentous fungi, Protein Eng. 7 (1994) 551 – 557. [11] L. Brady, A.M. Brzozowski, Z.S. Derewenda, E. Dodson, G. Dodson, S. Tolley, J.P. Turkenburg, L. Christiansen, B. HugeJensen, L. Norskov, L. Thim, U. Menge, A serine protease triad forms the catalytic centre of a triacylglycerol lipase, Nature 343 (1990) 767 – 770. [12] G.H. Peters, A. Svendsen, H. Langberg, J. Vind, S.A. Patkar, S. Toxvaerd, P.K.J. Kinnunen, Active serine involved in the stabilization of the active site loop in the Humicola lanuginosa lipase, Biochemistry 37 (1998) 12375 – 12383. [13] O.G. Berg, Y. Cajal, G.L. Butterfoss, R.L. Grey, M.A. Alsina, B.Z. Yu, M.L. Jain, Interfacial activation of triglyceride lipase from Thermomyces (Humicola) lanuginosa: kinetic parameters and a basis for control of the lid, Biochemistry 37 (1998) 6615 – 6627. [14] U. Derewenda, L. Swenson, R. Green, Y. Wei, G.G. Dodson, S. Yamaguchi, M.J. Haas, Z.S. Derewenda, An unusual buried polar cluster in a family of fungal lipases, Structural Biology 1 (1994) 36 – 47. [15] M. Holmquist, M. Martinelle, I.G. Clausen, S.A. Patkar, A. Svendsen, K. Hult, Trp89 in the lid of Humicola lanuginosa lipase is important for efficient hydrolysis of tributyrin, Protein Eng. 29 (1994) 599 – 603. [16] M. Holmquist, I.G. Clausen, S.A. Patkar, A. Svendsen, K. Hult, Probing a functional role of Glu87 and Trp89 in the lid of Humicola lanuginosa lipase through transesterification reaction in organic solvent, J. Protein Chem. 14 (1995) 217 – 224.

[17] A. Svendsen, I.G. Clausen, S.A. Patkar, B. Kim, M. Thellersen, Protein engineering of microbial lipase of industrial interest, Methods Enzymol. 284 (1997) 317 – 340. [18] D.M. Lawson, G.G. Dodson, R.E. Hubbard, B. Huge-Jensen, E. Boel, A.M. Brzozowski, Z.S. Derewenda, The three-dimensional structures of two lipases from filamentous fungi, in: P. Woolley, S. Petersen (Eds.), Lipase—Their Biochemistry, Structure and Application, Cambridge University Press, UK, 1994, pp. 77 – 94. [19] Z.S. Derewenda, U. Derewenda, G.G. Dodson, The crystal and molecular structure of the Rhizomucor miehei triacylglyceride lipase at 1.9 2 resolution, J. Mol. Biol. 227 (1992) 818 – 839. [20] A.M. Brzozowski, U. Derewenda, Z.S. Derewenda, G.G. Dodson, D.M. Lawson, J.P. Turkenburg, F. Bjfrkling, B. Huge-Jensen, S.A. Patkar, L. Thim, Model for interfacial activation in lipases from the structure of a fungal lipase–inhibitor complex, Nature 351 (1991) 491 – 494. [21] Y. Cajal, J. Prat, A. Svendsen, J. de Bolos, M.A. Alsina, Spectroscopic studies of the interfacial binding of Humicola lanuginosa lipase, Analyst 123 (1998) 2229 – 2233. [22] A. Jutila, K. Zhu, S.A. Patkar, J. Vind, A. Svendsen, P.K.J. Kinnunen, Detergent-induced conformational changes of Humicola lanuginosa lipase studied by fluorescence spectroscopy, Biophys. J. 78 (2000) 1634 – 1642. [23] K. Zhu, A. Jutila, E.K.J. Tuominen, P.K.J. Kinnunen, Effects of ipropanol on the structural dynamics of Thermomyces lanuginosa lipase revealed by tryptophan fluorescence, Protein Sci. 10 (2001) 339 – 351. [24] M.M. Bradford, A rapid and sensitive method for the quantitation of microgram quantities of protein utilizing the principle of protein–dye binding, Anal. Biochem. 72 (1976) 248 – 254. [25] T.K. Das, S. Mazumdar, pH-induced conformational perturbation in horseradish peroxidase. Picosecond tryptophan fluorescence studies on native and cyanide-modified enzymes, Eur. J. Biochem. 227 (1995) 823 – 828. [26] K.B. Jennifer, S.M. Kathleen, Ligand-induced conformational changes in lactose repressor: a fluorescence study of single tryptophan mutants, Biochemistry 36 (1997) 15632 – 15642. [27] E. Gratton, D.M. Jameson, R.D. Hall, Multifrequency phase and modulation fluorometry, Annu. Rev. Biophys. Bioeng. 13 (1984) 105 – 124. [28] D.M. Jameson, T.L. Hazlett, Timeresolved fluorescence in biology and biochemistry, in: G. Dewey (Ed.), Biophysical and Biochemical Aspect of Fluorescence Spectroscopy, Plenum Press, New York, 1991, pp. 105 – 133. [29] K. Zhu, A. Jutila, P.K.J. Kinnunen, Steady state and time-resolved effects of guanidine hydrochloride on the structure of Humicola lanuginosa lipase revealed by fluorescence spectroscopy, Protein Sci. 9 (2000) 598 – 609. [30] K. Zhu, A. Jutila, E.K.J. Tuominen, S.A. Patkar, A. Svendsen, P.K.J. Kinnunen, Impact of tryptophan residues of Humicola lanuginosa lipase on its thermal stability, Biochim. Biophys. Acta 1547 (2001) 329 – 338. [31] E.A. Brustein, N.S. Vedenkina, M.N. Irkova, Fluorescence and the location of tryptophan residues in protein molecules, Photochem. Photobiol. 18 (1973) 263 – 279. [32] M. She, W.J. Dong, P.K. Umeda, H.C. Cheung, Tryptophan mutants of troponin C from skeletal muscle—an optical probe of the regulatory domain, Eur. J. Biochem. 252 (1998) 600 – 607. [33] J.R. Lakowicz, Principles of Fluorescence Spectroscopy, 2nd edition, Plenum, New York, 1999. [34] A.J. Kungl, N.V. Visser, A. van Hoek, A.J. Visser, A. Billich, A. Schilk, H. Gstach, M. Auer, Time-resolved fluorescence anisotropy of HIV-1 protease inhibitor complexes correlates with inhibitory activity, Biochemistry 37 (1998) 2778 – 2786. [35] R. Chen, J. Knutson, H. Ziffer, D. Porter, Fluorescence of tryptophan dipeptides: correlations with the rotamer model, Biochemistry 30 (1991) 5184 – 5195.

A. Jutila et al. / Biochimica et Biophysica Acta 1702 (2004) 181–189 [36] M. Graupner, L. Haalck, F. Spener, H. Linder, O. Glatter, F. Paltauf, A. Hermetter, Molecular dynamics of microbial lipases as determined from their intrinsic tryptophan fluorescence, Biophys. J. 77 (1999) 493 – 504. [37] A.M. Brzozowski, H. Savage, C.S. Verma, J.P. Turkenburg, D.M. Lawson, A. Svendsen, S. Patkar, Structural origins of the interfacial activation in Thermomyces (Humicola) lanuginosa lipase, Biochemistry 39 (2000) 15071 – 15082. [38] L.J. Juszczak, Z.Y. Zhang, L. Wu, D.S. Gottfried, D.D. Eads, Rapid loop dynamics of Yersinia protein tyrosine phosphatases, Biochemistry 36 (1997) 2227 – 2236.

189

[39] R.F. Steiner, in: J.R. Lakowicz (Ed.), Topics in Fluorescence Spectroscopy, vol. 2, Plenum Press, New York, 1991. [40] W.S. Davidson, K. Arnvig-McGuire, A. Kennedy, J. Kosman, T.L. Hazlett, A. Jonas, Structural organization of the N-terminal domain of apolipoprotein A-I: studies of tryptophan mutants, Biochemistry 38 (1999) 14387 – 14395. [41] R.C. Weast (Ed.), Handbook of Chemistry and Physics, CRC Press, Cleveland, OH, 1972–1973. [42] J. Navaza, Amore—an automated package for molecular replacement, Acta Crystallogr., A 50 (1994) 157 – 163.