Thermal stability of outer membrane protein porin from Paracoccus denitrificans: FT-IR as a spectroscopic tool to study
lipid–protein interaction
Suja Sukumaran, Karin Hauser, Annette Rauscher, Werner Ma¨ntele
*Institut fu¨r Biophysik, Johann Wolfgang Goethe Universita¨t, Theodor Stern Kai 7, Haus 74/75, 60590 Frankfurt am Main, Germany
Abstract Lipid protein interactions play a key role in the stabil- ity and function of various membrane proteins. Earlier we have reported the extreme thermal stability of porin fromParacoccus denitrificansreconstituted into liposomes. Here, we used Fourier transform infrared spectroscopy for a label free analysis of the global secondary structural changes and local changes in the tyrosine microenvironment. Our results show that a mixed lipid system (non-uniform bilayer) optimizes the thermal stability of porin as compared to the porin in pure lipids (uniform bilayer) or detergent micelles. This is in line with the fact that the bacte- rial outer membrane is a dynamic system made up of lipids of varying chain lengths, head groups and the barrel wall height contacting the membrane is uneven.
Keywords:Porin; Membrane protein stability; Lipid protein interaction; Fourier transform infrared spectroscopy
1. Introduction
Porins are the channel forming proteins found in the outer membrane of gram negative bacteria, mitochondria and chlo roplasts. They facilitate the transport of hydrophilic molecules of a molecular mass below600 Da in the case of bacterial porins [1,2]. According to the X ray crystal structure of the porin fromParacoccus denitrificans, the functional unit is a tri mer. Each monomer is composed of 16 anti parallelbstrands [3]. The length of the strands varies between 7 and 16 amino acid residues. The height of the barrel wall is 27 A˚ at the intermonomeric contact surface and33 A˚ at the membrane interface.
It is known for the smallbbarrel proteins like outer mem brane protein (Omp) A and OmpX that they contain girdles
of aromatic residues. In Paracoccusporin a similar arrange ment of amino acid residues as other non specific porins is ob served [3]. There is a central belt of non polar residues bordered at each edge by girdles of aromatic amino acids (Fig. 1). The girdles are defined predominantly by Tyr residues with their hydroxyl groups pointing towards the aqueous phase. A porin monomer has 18 tyrosine residues. Various groups have concluded that the hydroxyl groups of the Tyr residues are located in the glycerol backbone of the bilayer, resulting in a thickness of 24 28 A˚ for the hydrophobic core of the protein (for a review see[4]). Two girdles of aromatic residues run across the outer face of the barrel, with a vertical separation of 20 25 A˚ . The aromatic girdles appear to mark the boundaries of the barrel surface interacting with the hydro phobic core of the membrane.
We have earlier reported the details on the thermal stability of porin from Paracoccus denitrificans[5]. It was found that porin undergoes temperature induced aggregation above the transition temperature of 86.2C if the protein is in detergent micelles, whereas no change in its secondary structure is ob served up to 95C if the protein is reconstituted into lipo somes. The reason for this extreme difference in stability of the protein can be attributed to the difference in the interaction of the protein with lipids and detergents. The result onPara coccusporin is not surprising as the thermal stability for most of the membrane proteins like bacteriorhodopsin and other porins is high[6,7]. The three major factors reported for the stability of porin are the extensive hydrogen bonding in the bbarrel, the rigid structured loops on both sides of the mem brane and the strong interaction between the monomers [7].
The interaction between protein and lipids enhances the stabil ity significantly in comparison to the detergent micelle system.
Therefore, the incorporation of a protein into the lipid bilayer provides a means of achieving a remarkably thermostable structure in organisms that are not thermophiles likeParacoc cusporin[5,7]. This study aimed to investigate the molecular basis for such extreme stability by Fourier transform infrared (FT IR) spectroscopy what in turn could provide insights into understanding the most appropriate environment for mem brane proteins in vitro.
Temperature induced unfolding of proteins does not only lead to the breakdown of various secondary and quaternary structural elements, but also involves the changes in the micro environment of specific side chain groups. A side chain vibra tion which can be distinguished very well in the mid infrared spectra of proteins is the C C stretching mode of the tyrosine Abbreviations: LLaPC, LLa phosphatidylcholine (lecithin); DOPC,
dioleoylphosphatidylcholine (di(C18:1)PC); DOPE, dioleoylphospha tidylethanolamine (di(C18:1)PE); DOPG, dioleoylphosphatidylglycerol (di(C18:1)PG); DLPC, dilauroylphosphatidylcholine (di(C12:0)PC);
DNPC, dinervonylphosphatidylcholine (di(C24:0)PC); Omp, outer membrane protein; Tyr, tyrosine; FT IR, Fourier transform infrared
*Corresponding author. Present address: Max von Laue Str. 1, 60438 Frankfurt am Main, Germany. Fax: +49 69 6301 5838/798 47423.
E mail address:maentele@biophysik.org(W. Mantele).
Konstanzer Online-Publikations-System (KOPS) URL: http://nbn-resolving.de/urn:nbn:de:bsz:352-175524
aromatic ring at 1515 cm 1[8,9]. Since most of the tyrosines are located in the protein lipid interface they provide an excel lent reporter group as changes in their environment may result from both changes in the secondary as well as quaternary structure.
2. Materials and methods
Porin gene cloned in vector pJC 40 was obtained from B. Ludwig, Institut fur Biochemie, Universitat Frankfurt, Germany[2,10]. The structure coordinates of P. denitrificansporin was provided by W.
Welte, Universitat Konstanz, Germany.LaPhosphatidylcholine (LL
aPC) was purchased from Sigma, Germany. All other pure lipids were bought from Avanti Polar Lipids, Inc. USA.
2.1. Protein purification
Purification of protein was carried out according to Saxena et al.
[10].
2.2. Lipid mixtures
Commercially availableLaPC is a mixed lipid having a composi tion 17% (C16:0), 4% (C18:0), 9% (C18:1), 60% (C18:2) and 7%
(C18:3) according to the product information sheet available from Sigma.
In this study, the following lipid mixture combinations have been used:
Mix1 (DOPC 30%, DLPC 35 %, DNPC 35%).
Mix2 (DOPE 60%, DOPG 20%, DLPC 10%, DNPC 10%).
Mix3 (DOPE 20%, DOPG 20%, DLPC 30%, DNPC 30%).
2.3. Reconstitution into liposomes
Reconstitution of the protein into liposomes was carried out by a modified freeze thaw method. 2 mg ofLLaPC was dissolved in chloro form and dried overnight. The dried lipids were resuspended in 2 mL of protein solution (0.5 mg/mL). The mixture was incubated at 37C for 30 minutes under agitation. Detergent was removed using Bio Rad bio beads (pre treated with methanol and water). This was fol lowed by four freeze thaw cycles. After removing the last traces of bio beads the lipid vesicles were collected by ultra centrifugation at 80 000·gfor 3 h. To remove the residual bio beads the pellet was washed with buffer (Tris 50 mM, EDTA 10 mM, and NaCl 100 mM) followed by centrifugation at 80 000·gfor 20 min. The liposomes col lected were directly used as IR sample.
2.4. FT IR transmission spectroscopy
The measurements were carried out in transmission mode. For the protein in detergent, the sample (20lL of 2 mg/mL) was dried in a gen tle stream of N2and re suspended in 3lL D2O. Approximately, 1.5lL of sample was loaded in the center of a demountable CaF2microcell
with an optical path length of 7.4lm[11,12]. The cuvette was sealed by coating the outer ring of one of the windows with a thin layer of an ethanol oil mixture, to prevent loss of sample by evaporation upon heating. IR spectra were recorded with a Bruker VECTOR 22 FT IR spectrometer equipped with a DTGS (deuterated triglycine sulfate) detector. For each spectrum 20 interferograms were averaged, apod ized with a Blackman Harris 4 term function, zero filled and Fou rier transformed to yield a nominal spectral resolution of 2 cm 1 with a data point interval of 1 cm 1. The spectra were processed and visualized using windows OPUS version 3.1 software. The built in pro gram for calculating second derivatives was used to identify the minute changes.
3. Results
3.1. Thermal stability of porin in detergent micelles and lecithin The second derivative spectra not only provide information about the minute secondary structure details but also yield an appreciable amount of information about the amino acid side chains. Tyrosine side chain C C mode is easily distinguishable as it is not overlapped by other bands. Since changes in the tyrosine environment may result from changes in both second ary and quaternary structure, it provides a reporter group to observe such changes. The strong and narrow IR band at 1515 cm 1, clearly seen in the second derivative spectra, can be assigned to the tyrosine side chains (Fig. 2).
In order to follow the unfolding process, the temperature dependence of the position of the amide I band and the tyro sine band has been studied. Spectra obtained from 2 to 3 independent measurements were analyzed, but not averaged.
Fig. 1. Paracoccusporin monomer (structure coordinates were pro vided by W. Welte). Residues highlighted as spacefilling model are tyrosines.
Fig. 2. Second derivative spectra of porin at 25C (solid line) and 95C (dashed line): (A) in detergent LDAO, (B) in lecithin.
Figs. 2A and B present the second derivative spectra of wild type protein in detergent micelles and liposomes (lecithin), respectively. The amide I bands at 1628 and 1695 cm 1, respectively, indicate thebsheet secondary structure [13,14].
The shift in the peak position to 1619 and 1685 cm 1 for the protein in detergent micelles shows that the protein aggre gates upon heating[15,16]. The amide I band position with respect to increasing temperature was plotted for protein in detergent micelles and in lecithin (Fig. 3). The plot clearly indicates that the protein in detergent micelles undergoes aggregation with a transition temperature of 86.2C, whereas for the protein reconstituted into lecithin, there is no change in their global secondary structure[5].Fig. 4shows the plot for tyrosine band position with respect to temperature. It is clearly observed that for the protein reconstituted into deter gent micelles the frequency of this band gradually shifts by almost 1 cm 1towards lower frequencies between room tem perature and 95C. The band position for the protein recon stituted into lecithin shows that up to 75C there is no change in the tyrosine band position. At higher temperatures an abrupt downshift is observed not only for the wild type but also for three site directed mutants (E81Q, W74C, and E81Q/D148N) (data not shown) [5]. This is attributed to the loss of water in the sample (see Section 4).
3.2. Tyrosine microenvironment of porin in liposomes formed from pure lipid
The question arose whether reconstituting the protein into various pure lipids would have any effect on the protein stabil ity. Using lipids of three different head groups and various chain lengths onEscherichia coliporin OmpF, OÕKeefe et al.
[17]suggested the difference in the binding constants. It was shown that phosphatidylglycerol exhibits half the binding con stant if compared to phosphatidylethanolamine and phospha tidylcholine.
Using similar lipids in our studies we have observed that reconstitution of porin in pure lipids of dioleoylphosphat idylcholine (di(C18:1)PC) [DOPC], dioleoylphosphatidyletha nolamine (di(C18:1)PE) [DOPE], dioleoylphosphatidylglycerol (di(C18:1)PG) [DOPG], dilauroylphosphatidylcholine (di(C12:0)PC) [DLPC], and dinervonylphosphatidylcholine (di(C24:0)PC) [DNPC] did not have any effect in the global stability of the protein as no shift in the amide I band position was detected (Fig. 3). To further analyze the changes in the microenvironment of the protein in the pure lipid, the tyrosine mode at 1515 cm 1was analyzed with respect to temperature (Fig. 5A D). Data for porin reconstituted into DOPE are not reproducible due to technical difficulty in hydrating the li pid[18]. It can be observed fromFig. 5A D that the tempera ture dependence of the tyrosine band is unlike that for protein reconstituted in lecithin (Fig. 4). It is evident that there is a temperature dependent downshift in the band position if porin is reconstituted into these pure lipids. Therefore, the tyrosine microenvironment of porin is different for protein reconsti tuted in lecithin and in pure lipid systems. We thus conclude that a non uniform lipid bilayer appropriately mimics the bac terial outer membrane.
3.3. Thermal stability of porin in liposomes formed from lipid mixture
To test the above conclusion, pure lipids were mixed in dif ferent ratios and the porins were reconstituted into such mixed lipid systems. It was observed that for the three mixtures tested, the global stability of the protein remains unchanged (Fig. 3). Fig. 6depicts the change in position of the tyrosine band with respect to temperature. It is evident from the figure that the position of the tyrosine band remains approximately constant up to 70C which is similar to that observed for lec ithin. This data supports that porins are more stable in a mixed lipid system (non uniform bilayer) than in pure lipids (uniform bilayer).
4. Discussion
It is evident from the temperature dependent IR spectra pre sented here that porin is highly stable in the lipid environment and undergoes aggregation in the detergent surrounding.
Moreover, tyrosine side chain absorption represents the micro environment of the protein under various conditions. The Tyr band shift appears to be very small, but it is significant if com pared to the variation of the values given for totally different proteins in the literature (1513 1517 cm 1)[9]. Analogous tem perature dependent frequency shifts of the tyrosine C C mode have been reported for RNase T1[19,20], concanavalin A[21]
and tendamistat[22].
Fig. 3. Temperature dependent peak of amide I in various lipid composition compared to that of the protein in detergent. (·) Detergent LDAO, (j) lecithin, (h) DOPG, (s) DOPE, (n) DOPC, (,) NPC, (%) DLPC, (m) Mix 1, (.) Mix 2, (d) Mix 3.
Fig. 4. Change in tyrosine band position of the protein on heating. (j) in detergent micelles, (d) in lecithin.
It was earlier suggested by us and other groups that the shift in the amide I band of porin occurs due to intermolecular hydrogen bonding[5,15]. The shift in the tyrosine band posi tion can be explained by the fact that the distance between the tyrosine oxygen and hydrogen is modulated by hydrogen bonding, which influences the electron density and thus the strength of the bonds between the carbons of the aromatic ring. The frequency of the corresponding C C ring stretching vibration reflects the bond strength[22]. Therefore, the down shift of the tyrosine band due to aggregation of the protein can be explained by stronger hydrogen bonds of the tyrosine O H with acceptors in the protein aggregates. This explains a change in the environment of one or more tyrosine residues upon denaturation. As inFig. 4, it is evident that the Tyr band position for porin reconstituted in lecithin does not change up to a temperature of 70C. Above this temperature, changes are seen which can be confidently assigned to loss of water
molecules, since the major secondary structure does not change up to 95C.
What might be the reason for the difference in stability?
Neutron diffraction studies have shown that in crystals grown from bOG or LDAO the hydrophobic region of OmpF is covered by detergent, with the two girdles of aromatic resi dues coinciding with the boundary between the polar and non polar regions of the detergents [23]. A location of the aromatic residues is also seen in molecular dynamic simula tions of OmpF in bilayers of phophatidylcholine (C14:0) [24]. Detergents may cover the protein but there are no spe cific interactions. However, in the case of lipids significant interaction between protein residues and lipids should cause the enormous increase in stability. Crystallography of mem brane proteins has also provided insights into the roles of specific membrane lipids in the key biophysical functions within the protein lipid environment (for reviews see [25 27]). It can be concluded that if the protein is in detergent mi celles, its residues are in a less protected state as compared to the protein in liposomes. The processes contributing to the stability and unfolding are thus completely different for porin in detergent and in liposomes.
The amide I band position of the protein reconstituted into various pure lipids does not change when the protein is heated up to 95C. This clearly indicates that even though the binding of these lipids to the protein is different for each one, it does not bring about changes in the global stability of the protein. However, the tyrosine microenvironment changes with temperature when the protein is even embedded in pure lipids (Fig. 5A D). Interestingly, the overall secondary struc ture of the protein does not change but significant changes are detected in the microenvironment of tyrosines. It is known that a bacterial cell membrane is made up of both bi layer forming and non bilayer forming lipids. Among bilayer Fig. 5. Shift in tyrosine band position with respect to temperature for pure lipids. (A) DNPC, (B) DOPG, (C) DOPC, (D) DLPC.
Fig. 6. Shift in tyrosine band position with respect to temperature for lipid mixtures.
forming lipids, there is generally the presence of lipids of var ious different chain lengths. In the case of lecithin which is made up of phosphatidylcholine with chain lengths from C:16 to C:18 with varying unsaturations, the microenviron ment of the protein is much more protected than for the pro tein in pure lipids. According to the 3D structure of Paracoccusporin (Fig. 1) it is clear that because of the porin asymmetry thebstrands are of various sizes (see Section 1) [3]. The resulting distances between the aromatic residues at the interface are hence different, thus necessitating the need of lipids of varying chain lengths. Consequently, it can be ar gued that the protein tends to be more stable if it is in a mixed lipid surrounding as compared to a pure lipid sur rounding. According to the earlier explanation, a tyrosine band downshift would indicate aggregation or stronger hydrogen bond formation due to instability. However, here we see a downshift without change in the global structure.
Therefore these changes in tyrosine microenvironment are not sufficient to reduce the global stability of the protein.
Testing of the hypothesis with the mixed lipid systems proves that it is preferable to reconstitute protein into bilayer formed of lipid mixtures rather than pure lipids.
The combined analysis of the amide I band profile as a glo bal parameter and the tyrosine C C mode as a local parameter has clearly shown the interaction of the hydrophobic part of porin with the lipids. Moreover, thermal stability was shown highest for lipid mixtures, suiting the asymmetry of the mem brane barrier for porin. The use of FT IR for studying lipid protein interaction is not new but our method of combined analysis of the amide I bands and the tyrosine microenviron ment makes it a new approach to analyze not only the stability and unfolding but also for structure and function correlation for other membrane proteins.
Acknowledgements:We thank B. Ludwig and K. Saxena, Institut fur Biochemie, Universitat Frankfurt, Germany for providing the wild type porin gene. We thank W. Welte for the porin structure coordi nates. Financial assistance from International Max Planck Research school and SFB 472 to S.S. is deeply acknowledged. M. Susse is thanked for help with the experiments.
References
[1] Schulz, G.E. (2000) Barrel membrane proteins. Curr. Opin.
Struct. Biol. 10, 443 447.
[2] Saxena, K., Richter, O. M.H., Ludwig, B. and Benz, R. (1997) Molecular cloning and functional characterization of thePara coccus denitrificansporin. Eur. J. Biochem. 245, 300 306.
[3] Hirsch, A., Breed, J., Saxena, K., Richter, O. M.H., Ludwig, B., Diederichs, K. and Welte, W. (1997) The structure of porin from Paracoccus denitrificansat 3.1 A˚ resolution. FEBS Lett. 404, 208 210.
[4] Lee, A.G. (2003) Lipid protein interaction in biological mem branes. A structural perspective. Biochim. Biophys. Acta 1612, 1 40.
[5] Sukumaran, S., Zscherp, C. and Mantele, W. (2004) Investigation of the thermal stability of porin fromParacoccus denitrificansby site directed mutagenesis and FT IR. Biopolymers 74, 82 86.
[6] Heinz, C., Engelhard, H. and Niederweis, M. (2003) The core of the tetrameric mycobacterial porin MspA is an extremely stable sheet domain. J. Biol. Chem. 278, 8678 8685.
[7] Haltia, T. and Freire, E. (1995) Forces and factors that contribute to the structural stability of membrane proteins. Biochim.
Biophys. Acta 1241, 295 322.
[8] Chirgadze, Y.N., Fedorov, O.V. and Trushina, N.P. (1975) Estimation of amino acid residue side chain absorption in the infrared spectra of protein solutions in heavy water. Biopolymers 14, 679 694.
[9] Barth, A. (2000) The infrared absorption of amino acid side chains. Prog. Biophys. Mol. Biol. 74, 141 173.
[10] Saxena, K., Drosou, V., Maier, E., Benz, R. and Ludwig, B.
(1999) Ion selectivity reversal and induction of voltage gating by site directed mutations in the Paracoccus denitrificans porin.
Biochemistry 38, 2206 2212.
[11] Abrecht, H., Goormaghtigh, E., Ruysschaert, J.M. and Homble, F. (2000) Structure and orientation of two voltage dependent anion selective channel isoforms. An attenuated total reflection Fourier transform infrared spectroscopy study. J. Biol. Chem.
275, 40992 40999.
[12] Fabian, H. and Mantele, W. (2002) Infrared spectroscopy of proteins, in: Handbook of Vibrational Spectroscopy (Calmers, J.M. and Griffiths, P.R., Eds.), Vol. 5, pp. 3399 3426. Wiley, Chichester.
[13] Goormaghtigh, E., Cabiaux, V. and Ruysschaert, J. M. (1994) Physicochemical methods in the study of biomembranes, in:
Subcellular Biochemistry (Hilderson, H.J. and Ralston, G.B., Eds.), Vol. 23, pp. 329 403. Wiley, Chichester.
[14] Arrondo, J.L., Muga, A., Castresana, J. and Goni, F.M. (1993) Quantitative studies of the structure of proteins in solution by Fourier transform infrared spectroscopy. Prog. Biophys. Mol.
Biol. 59, 23 56.
[15] Byler, D.M. and Susi, H. (1986) Examination of the secondary structure of proteins by deconvoluted FT IR spectra. Biopoly mers 25, 469 487.
[16] Jackson, M. and Mantsch, H.H. (1995) The use and misuse of FT IR spectroscopy in the determination of protein structure. Crit.
Rev. Biochem. Mol. Biol. 30, 95 120.
[17] OÕKeeffe, A.H., East, J.M. and Lee, A.G. (2000) Selectivity in lipid binding to the bacterial outer membrane protein OmpF.
Biophys. J. 79, 2066 2074.
[18] Gunstone, F.D., Harwood, J.L and Padley, F.B. (1994) The Lipid Handbook, Chapman & Hall, London; New York.
[19] Fabian, H., Schultz, C., Naumann, D., Landt, D.O., Hahn, U.
and Saenger, W. (1993) Secondary structure and temperature induced unfolding and refolding of ribonuclease T1 in aqueous solution. J. Mol. Biol. 232, 967 981.
[20] Fabian, H., Schultz, C., Backmann, J., Hahn, U., Saenger, W., Mantsch, H.H. and Naumann, D. (1994) Impact of point mutations on the structure and thermal stability of ribonuclease T1 in aqueous solution probed by Fourier transform infrared spectroscopy. Biochemistry 33, 10725 10730.
[21] Arrondo, J.L.R., Young, N.M. and Mantsch, H.H. (1988) The solution structure of concanavalin A probed by FT IR spectros copy. Biochim. Biophys. Acta 952, 261 268.
[22] Zscherp, C., Aygun, H., Engels, J.W. and Mantele, W. (2003) Effect of proline to alanine mutation on the thermal stability of the allbsheet protein tendamistat. Biochim. Biophys. Acta 1651, 139 145.
[23] Pebaypeyroula, E., Garavito, R.M., Rosenbusch, J.P., Zulauf, M.
and Timmins, P.A. (1995) Detergent structure in tetragonal crystals of OmpF porin. Structure 3, 1051 1059.
[24] Im, W. and Roux, B. (2002) Ions and counterions in a biological channel: a molecular dynamics simulation of OmpF porin from Escherichia coliin an explicit membrane with 1 M KCl aqueous salt solution. J. Mol. Biol. 319, 1177 1197.
[25] Seddon, A.M., Curnow, P. and Booth, P.J. (2004) Membrane proteins, lipids and detergents: not just a soap opera. Biochim.
Biophys. Acta 1666, 105 117.
[26] Palsdottir, H. and Hunte, C. (2004) Lipids in membrane protein structures. Biochim. Biophys. Acta 1666, 2 18.
[27] Lee, A.G. (2004) How lipids affect the activities of integral membrane proteins. Biochim. Biophys. Acta 1666, 62 87.
