pubs.acs.org/Langmuir © 2009 American Chemical Society
Orientation of 1-Palmitoyl-2-oleoyl-sn-glycero-3-phosphocholine at the Surface of Formamide and Hydroxypropionitrile Hartwig Pohl,*,† Thomas Krebs,‡ and Harald Morgner† †
Wilhelm-Ostwald-Institute for Physical and Theoretical Chemistry, University Leipzig, Linn estrasse 2, D-04103 Leipzig, Germany and ‡Department of Chemistry, University of Wisconsin;Madison, 1101 University Avenue, Madison, Wisconsin 53706 Received July 23, 2009. Revised Manuscript Received September 28, 2009
Phospholipids are a main component of cell membranes. Therefore, the experimental investigation of the selforganization of phospholipids is of great interest. Here we present results concerning the orientation of the phospholipid 1-palmitoyl-2-oleoyl-sn-glycero-3-phosphocholine (POPC) at the surface of the polar solvents formamide and 3-hydroxypropionitrile (HPN), which we investigated by means of neutral impact collision ion scattering spectroscopy. It is shown that, in HPN, at low POPC surface excesses, the phospholipids are oriented with the polar headgroups pointing out of the solution. It is concluded that the behavior of lipids at the surface of liquids is more complex than expected and to a great extent dependent on the solvent. Comprehension of the behavior of POPC in the applied solvents might contribute to the understanding of the self-organization of phospholipids in water.
Introduction The understanding of the self-organization of lipids at the surface of polar liquids (e.g., to form membranes) is of great interest in biochemical as well as in technical topics. We study monolayers on a liquid substrate which is suitable for investigation by several surface analytical techniques. As the principles of self-organization of the lipids should be similar for monolayers and double layers, some information about the relation between surface and bulk might be relevant for understanding the relation between a membrane and the surrounding solvent. In order to gain information about the surface structure of these systems, we have previously performed investigations using various particle spectroscopies, such as metastable induced electron spectroscopy (MIES),1,2 X-ray photoelectron spectroscopy (XPS),4 and ion scattering spectroscopy (ISS). Neutral impact collision ion scattering spectroscopy (NICISS) is a method particularly appropriate for the investigation of the surface structure, as it yields concentration depth profiles of the elements contained in the sample.6 Recent XPS data suggest that, in HPN, the polar headgroup of phospholipids in lecithin extracted from eggs resides closer to the surface than the hydrophobic chains, that is, the chains point into the liquid, if the surface concentration is low.3 In those earlier experiments, however, the spectral resolution was not good enough to back up this interpretation beyond any doubt. Motivated by the XPS results, we performed NICISS measurements on solutions of POPC in formamide and HPN to endorse those findings. As the mentioned methods require high-vacuum conditions, water is not accessible in our present experimental setup due to its high vapor pressure. Therefore, we have dissolved the phospho*To whom correspondence should be addressed. E-mail:
[email protected]. (1) Morgner, H. Adv. At., Mol., Opt. Phys. 2000, 42B, 387. (2) Morgner, H.; Oberbrodhage, J.; Richter, K. Mol. Phys. 1992, 76, 813. (3) Knoll, F. Ph.D. Thesis, University Witten, 2001. (4) Eschen, F.; Heyerhoff, M.; Morgner, H.; Vogt, J. J. Phys.: Condens. Matter 1995, 7(10), 1961. (5) Niehus, H.; Comsa, G. Nucl. Instrum. Methods Phys. Res., Sect. B 1986, 13, 213.
Langmuir 2010, 26(4), 2473–2476
lipid 1-palmitoyl-2-oleoyl-sn-glycero-3-phosphocholine (POPC) in formamide and 3-hydroxypropionitrile (HPN), respectively. These solvents have some solution properties similar to those of water (e.g., dipole moment; H2O, 1.85 D;7 formamide, 3.73 D;7 HPN, ∼3.5 D8) and form hydrogen bonds in the condensed phase. The surface tension of formamide is 56.9 mN m-1 at 293 K, that of HPN is 48.0 mN m-1 at 293 K, and water has a surface tension of 72.8 mN m-1 at 293 K.7
Experimental Section Materials. The phospholipid POPC was purchased from Avanti Polar Lipids Inc. and was dried in a vacuum to weight constancy before the solutions were prepared. HPN was obtained from Merck (98%) and was purified by vacuum distillation. The boiling temperature was ∼350 K at a pressure of ∼2 mbar. Formamide was purchased from Acros Organics (>99.5%) and was used as obtained. The concentration of the POPC solutions in HPN ranged from 1.20 10-5 to 1.48 10-3 mol kg-1, and for POPC solutions in formamide the concentration range was 10-6-10-5 mol kg-1. Figure 1 shows the structural formulas of the solute and the solvents. NICISS Measurements. In the NICISS technique, helium ions are accelerated onto the surface of the sample. Ions that are backscattered from the top atom layer of the sample undergo elastic collisions to a good approximation. If the ions are scattered at a layer below the surface, almost all of them are neutralized and an additional inelastic energy loss, dependent on the penetration depth and the stopping power of the sample,9 is observed. The extent of inelastic energy loss depends on the depth of the hit atom below the surface. The energy of the backscattered projectiles is determined from time-of-flight (TOF) measurements. A scattering angle of 168° and a TOF length of ∼124 cm are employed in our experimental setup. A detailed description of the apparatus can be found in ref 6; hence, only a short description follows here. (6) Andersson, G.; Morgner, H. Surf. Sci. 1998, 405, 138. (7) Lide, D. R. Handbook of Chemistry and Physics, 80th ed.; CRC Press LLC: Cleveland, OH, 1999. (8) Wulf, M. Ph.D. Thesis, Witten, 1994. (9) K€uhrt, E.; Wedell, R. Phys. Lett. 1983, 96A, 347.
Published on Web 11/18/2009
DOI: 10.1021/la902706x
2473
Article
Figure 1. Structural formulas of POPC and the solvents HPN and formamide.
The design of the NICISS setup is similar to the one used by Niehus and Comsa5 and has been described in detail before.6 The liquid surface is prepared by rotating a disk that is immersed into a liquid reservoir so that a film is established on the disk. A so-called “skimmer” scrapes the surface before the ion beam strikes the surface, so that fresh bulk material is exposed to the surface. Variation of the rotation speed permits adjustment of the surface age of the solution. Furthermore, broadening of the signal occurs due to the socalled “straggling”.10-13 For the purpose of the evaluation of the orientation, the importance of these phenomena is very small. The positions of the signal maxima of the depth profiles should not be affected at all, as the broadening caused by straggling is much smaller than the width of the obtained signals. If molecules contain prominent heteroatoms in different parts of the molecule, it is possible to derive the preferential mean orientation of the molecule from the element depth profiles. This feature of NICISS is exploited in the following. We have determined the profile of phosphorus, which is present only in the headgroup of POPC, and the profile of carbon pertaining to POPC, which is mainly present in the hydrophobic chains.
Results and Discussion The concentration depth profiles of phosphorus and carbon in the solutions were determined from time-of-flight He backscattering spectra as described previously.6 The depth profile of phosphorus can be regarded as representative for the depth profile of the POPC headgroup. The depth profile of carbon, however, is composed of the POPC and HPN contributions to the total carbon signal. The HPN carbon depth profile can be derived from the HPN nitrogen depth profile, which was also measured. The HPN carbon depth profile cC,HPN(z) can be obtained from cC,HPN(z) = 3[cN(z) - cP(z)], where cN(z) denotes the measured nitrogen depth profile and cP(z) denotes the measured phosphorus depth profile. The factor of 3 in the equation is the stoichiometric ratio between carbon and phosphorus in the HPN molecule. The phosphorus depth profile is subtracted from the nitrogen depth profile because the nitrogen signal also contains the contribution from the nitrogen atom in the choline group of POPC. Since the phosphorus and nitrogen atoms in POPC are in proximity, the phosphorus depth profile is assumed to be approximately equal to the depth profile of nitrogen in the choline group. The POPC carbon depth profile cc,POPC(z) can then be obtained from cC,POPC(z) = cC(z) - cC,HPN(z), where cC(z) denotes the total carbon depth profile as measured by (10) Krebs, T. Ph.D. Thesis, University Leipzig, 2007. (11) Andersson, G.; Morgner, H.; Pohl, H. Phys. Rev. A 2008, 78, 032904. (12) Briere, M. A.; Biersack, J. P. Nucl. Instrum. Methods Phys. Res., Sect. B 1992, 64, 693. (13) Szilagyi, E. Nucl. Instrum. Methods Phys. Res., Sect. B 2000, 161-163, 37.
2474 DOI: 10.1021/la902706x
Pohl et al.
Figure 2. Determination of the POPC carbon signal of a 1.05 mmol kg-1 POPC solution in HPN (see Experimental Section for further details). The POPC carbon depth profile (0) is equal to the difference between the NICISS total carbon signal (O) and the HPN carbon signal (4).
NICISS. Figure 2 shows the three carbon depth profiles cC(z), cC,HPN(z), and cC,POPC(z) of a 1.05 mmol kg-1 POPC solution in HPN as an example. Figure 3 shows the comparison of the depth profiles of POPC carbon and phosphorus of a 1.05 mmol kg-1 POPC solution in HPN (panel a) and a 10 μmol kg-1 POPC solution in formamide (panel b) as examples. The surface excesses Γe of phosphorus and carbon in POPC have been determined from the experimentally obtained depth profiles of phosphorus and carbon via14 Γe ¼
Z b a
csolute ðzÞ -csolvent ðzÞ
xsolute dz xsolvent
ð1Þ
In eq 1, csolute(z) denotes depth profile of the solute, csolvent(z) is the depth profile of the solvent, and xsolute and xsolvent denote the mole fraction of the solute and the solvent, respectively, of the solution. The variables a and b are reasonable lower and upper boundaries for integration of the depth profile, for example, -20 and 60 A˚. Due to the small bulk concentrations of POPC that were used to prepare the solutionsR(e1 mmol kg-1), xsolute , xsolvent. Hence, eq 1 reduces to Γe ≈ ba csolute(z) dz to a very good approximation. The surface excess was determined from the phosphorus depth profile using eq 1. Surface excesses have been determined for different surface ages. The data points shown in Figure 3 are those corresponding to the highest surface age (∼1500 ms). The distance between the centers of mass of the profiles are plotted versus surface excess in Figure 4. The distance between the centers of mass is correlated with the orientation of the lipids at the surface: The carbon depth profile does not exactly represent the distribution of the hydrophobic chains in POPC but also includes the eight carbon atoms of the headgroup. Assuming that the phosphorus depth profile represents the distribution of the headgroup, the depth distribution of the chains can be calculated as cc,POPC(z) - 8cP(z). This correction did not cause a noticeable shift in the center of mass of the carbon depth profile, however, and is therefore omitted here. The distance between the centers of mass of P and C of POPC in HPN as well as in formamide is found to be dependent on the bulk concentration of the lipid as well as on the surface age. Plotting the orientation versus the surface excess of both dynamic and equilibrium data reveals that the orientation is independent of the combination of concentration and surface age leading to a certain surface excess.
(14) Adamson, A. W.; Gast, A. P. The Physical Chemistry of Surfaces; WileyInterscience: New York, 1997.
Langmuir 2010, 26(4), 2473–2476
Pohl et al.
Figure 3. Comparison of the depth profiles of excess carbon (left scale) and phosphorus (right scale) in the case of a 1.05 mmol kg-1 POPC solution in HPN at 267 K (a) and a 10 μmol kg-1 POPC solution in formamide at 279 K (b).
Figure 4. (Diagram on left side) Orientation of POPC at the surface of formamide (2) and HPN (O). The shown data include dynamic as well as equilibrium measurements. The distance between the centers of mass of carbon and phosphorus distribution is plotted versus the surface excess determined from NICISS data using eq 1. If the orientation of POPC was vertical, the distance should be about 24 A˚. (Drawing on right side) The distance between the centers of mass in the stretched-out molecule is calculated as 24 A˚. The panel below the diagram schematically displays different regimes of the mean molecular orientation as a function of surface excess.
The behavior of the lipid molecules depends remarkably on the polar solvent. In formamide solutions, we find the center of mass of the phosphorus profile always deeper below the surface than the center of mass of the POPC related carbon. The distance between the centers of mass of both profiles increases with increasing surface excess; see Figure 4. The plot of the orientation of the lipid versus surface excess can be regarded as monotonous. Thus, a slight reorientation from “lying” to “standing” lipids occurs in this system, that is, the surfactant reaches the surface with the hydrophobic chains pointing toward vacuum, even at Langmuir 2010, 26(4), 2473–2476
Article
low concentrations. This orientation of a phospholipid at a polar surface is commonly accepted in biochemistry (e.g., ref 15). The measured distance divided by the nominal distance between the centers of mass in the stretched out molecule can be regarded equal to the cosine of the mean angle of orientation the molecule with respect to the surface normal. However, the behavior of POPC in HPN seems to be fundamentally different. Previous angular resolved XPS data suggest that, at low concentrations, the polar headgroups are located at the surface while the hydrophobic chains appear to be immersed into the solvent.3 This is supported by the present investigations with NICISS. Plotting the depth difference between the centers of mass of phosphorus and carbon versus surface excess, we find a continuous reorientation of the POPC molecules with a change in sign at a surface excess of approximately 10-10 mol cm-2. Figure 4 demonstrates that at low concentration the center of mass of the headgroup resides closer to the surface than the center of mass of the hydrophobic chains. With increasing surface excess, the relative position between both elements approaches the expected order. This does not necessarily mean that the lipid is oriented “upside-down”, as the onset of the profiles of carbon and phosphorus are almost at the same depth. Nevertheless, it can be concluded that, at low concentrations, the phosphorus containing group is exposed to the gas phase or vacuum. At this time, we have no clear explanation for the orientation of POPC at the surface of HPN at low POPC bulk concentrations. At low POPC surface concentrations, the “inverted” orientation of the POPC molecule at the surface of the liquid seems to indicate that the POPC chains are more soluble in HPN than the POPC headgroup. The supposedly weaker solvation of the POPC headgroup in HPN as compared to formamide may be supported by the finding that POPC micelles are relatively small aggregates16 of approximately 20 POPC molecules. A recent molecular dynamics study17 shows that micelles of this size show a disordered structure, where the chains and headgroups are intermixed with each other. The different behavior of POPC in HPN and formamide, however, may be correlated to the fact that HPN forms less hydrogen bonds per unit volume than formamide.18 Hence, the POPC chains have a less perturbing effect on the hydrogen-bonding network in HPN as compared to formamide in the bulk and at the surface. The comparison of the behavior in HPN and formamide seems to indicate that the degree of this orientation as well as the surface excess depends on this solvent property if the concentration approaches the maximum solubility. It can be speculated that, in water, with more hydrogen bonds per unit volume than formamide, the maximum excess is significantly higher.
Conclusion and Outlook Comparison of data achieved by dynamic and equilibrium measurements show that the mean orientation of POPC at the surface of the investigated solvents depends on the surface excess, rather than on the surface age.10 Therefore, it can be concluded that the orientation relaxation is much faster than the diffusion toward the surface. In formamide, POPC displays the expected behavior, that is, the polar headgroups are oriented toward the polar solvent, whereas the hydrocarbon chains are exposed to the vacuum at all concentrations. In HPN, surprising behavior of POPC is (15) (16) (17) (18)
Voet, D.; Voet, J. G. Biochemistry; Wiley: New York, 1990. Morgner, H. J. Phys. Chem. B 2008, 112(5), 1383. Schenk, A. Master Thesis, University Leipzig, 2007 Luck, W. A. P. Colloid Polym. Sci. 2001, 279, 554.
DOI: 10.1021/la902706x
2475
Article
found; the phosphorus containing headgroup is exposed to vacuum at low POPC surface concentrations. In order to elucidate the surface structure that leads to the observed phenomenon, further experiments are planned. Current experiments in our group utilize the surface sensitive electron spectroscopy MIES in order to characterize the topmost layer of the liquid. It will be of great interest to investigate the adsorption of POPC at a
2476 DOI: 10.1021/la902706x
Pohl et al.
formamide surface by means of MIES to confirm the discussed phenomena. Finally, the orientation of POPC at an aqueous surface has to be observed by NICISS which requires special construction details.11 Acknowledgment. Financial support by the German Science Foundation (DFG, MO288/34-1) is gratefully acknowledged.
Langmuir 2010, 26(4), 2473–2476
