Spectroscopy at the Interface

A wide variety of processes occur at biological interfaces, such as those between drugs and membranes, metal ions and membranes, and water and membranes. Paul S. Cremer, the J. Lloyd Huck Chair in Natural Sciences in the Department of Chemistry at Penn State, is the recipient of the 2016 ANACHEM Award, and he and his group study these processes using various novel spectroscopy and microfluidic approaches..

A wide variety of processes occur at biological interfaces, such as those between drugs and membranes, metal ions and membranes, and water and membranes. A better understanding of the molecular-level mechanisms occurring at such interfaces can be useful in a many fields, such as the study of oxidative damage in neurodegenerative diseases. These processes can be investigated using a combination of microfluidic and nonlinear optical techniques. Paul S. Cremer, the J. Lloyd Huck Chair in Natural Sciences in the Department of Chemistry at Penn State, is the recipient of the 2016 ANACHEM Award, and he and his group study these processes using various novel spectroscopy and microfluidic approaches. He recently spoke to us about this work. This interview is part of a series of interviews with the winners of awards presented at the 2016 SciX conference.

One of your group’s publications from 2012 describes the use of a surface-specific nonlinear optical technique, vibrational sum frequency spectroscopy (VSFS), to investigate the effects of five monovalent anions on water structure adjacent to quartz and titanium dioxide substrates (1). Can you briefly describe this method? What makes VSFS a good technique for studying these effects?

VSFS, also called sum frequency generation (SFG) spectroscopy, is one of the most powerful techniques in use today to explore interfacial water structure. This is a three-wave mixing technique that combines tunable infrared (IR) and visible laser pulses to create light and the sum of the two frequencies. The technique is largely surface specific at the interface between media lacking a center of inversion symmetry (for example, an air–water or oxide–water interface) because both the infrared and Raman selection rules need to be obeyed simultaneously, which requires the breaking of inversion symmetry. Aqueous interfaces lack inversion symmetry, while the bulk solution is isotropic. As such, one can obtain a vibrational spectrum from water molecules at these interfaces even in the presence of a large amount of bulk water, which does not give a strong VSFS response. Surface specificity has traditionally been much more difficult to achieve with corresponding linear techniques, like infrared and vibrational Raman spectroscopy, because the signal from the interface is usually buried under a far stronger signal from the adjacent aqueous solution.

In the paper that you mention, we took advantage of the surface specificity of VSFS to probe how interfacial water structure at the silica–water and TiO₂–water interfaces is influenced by the introduction of various salts into the aqueous solution. In the absence of salt, the alignment of water structure is quite high because oxide surfaces bear a negative charge. When sodium salts are added, however, the water structure is attenuated by formation of a double layer that screens the surface charge. Also, cations can adsorb directly onto the surface and partially cancel the interfacial charge. Curiously, what we found in this study was that the identity of the anion is also crucial for understanding interfacial water structure. Namely, better hydrated anions like Cl^- partition to the least extent to the interface. As such, the attenuation effect by Na⁺ on the interfacial water structure is strongest. By contrast, more weakly hydrated anions like ClO₄^- or SCN^- can partition into the double layer and adsorb to the surface, helping to maintain the original negative charge. Such results are also reminiscent of anion binding found at protein–water and polymer–water interfaces.

In a second study performed by your group, attenuated total reflection Fourier transform infrared (ATR-FT-IR) and VSFS were used to examine salt interactions with butyramide as a mimic of cation interactions with protein backbones (2). What types of information do the two techniques provide with respect to this study, and how are they complementary?

Understanding whether cations from solution can bind to the carbonyl oxygen on protein backbones had been a subject of long-standing interest in the biophysical community before our spectroscopic studies. Indeed, cation binding can lead to protein denaturation and the salting-in of these macromolecules into solution. We wanted to know which cations bound and whether such interactions involved contact pair formation between the amide and the cation or if this interaction was solvent separated. ATR-FT-IR allowed us to probe contact pair formation. In this case, there is a single peak in the amide I band around 1620 cm^-1 for butyramide in neat water. When a sufficient concentration of a strongly hydrated metal cation is added to solution (for example, Ca²⁺, Mg²⁺, or Li⁺), a new peak round 1645 cm^-1 clearly shows that a water hydrogen bond to the amide has been replaced by metal ion binding. Such binding is quite weak; however, more weakly hydrated metal ions like Na⁺ and K⁺ hardly interact with them at all.

Although FT-IR provides significant insight into metal ion contact pair binding, it does not provide clear information about solvent separated interactions. VSFS, however, can help with this. In this case, experiments are conducted at the air–water interface in the presence of a monolayer of adsorbed butyramide. This molecule does not possess a net charge, so in the absence of salt, the associated interfacial water structure is quite weak. When CaCl₂, MgCl₂, or LiCl is introduced into the water phase beneath the monolayer, the interfacial water peaks are strongly enhanced by the binding of the cations. This enhancement occurs because the interface takes on a net positive charge, which aligns the adjacent water molecules. Such VSFS data nicely matched the results obtained by FT-IR. More significantly, the experiments could be repeated with NaCl and KCl. In this case, the infrared data could only prove that Na⁺ and K⁺ did not form contact pairs with the amide, but provided no information on solvent separated binding. The VSFS data showed no increase in water structure, which indicates that these ions do not bind either through contact pair formation or by solvent separated binding.

Your group has used a fluorescence quenching assay performed inside microfluidic devices to measure the binding affinity of Cu²⁺ to phosphatidylserine (PS) lipids (3). Can you please briefly describe the experimental setup and discuss how this approach provides information about PS and the physiological effects of Cu²⁺ binding?

This is a new analytical assay developed by coating lipid bilayers inside polydimethylsiloxane–glass microfluidic devices. Initially, lipid vesicles are introduced into a nascently formed microfluidic channel and fuse spontaneously to the glass and polymer surfaces. This forms a 4-nm-high supported lipid bilayer architecture, which is two-dimensionally fluid. A trace concentration of a dye-conjugated lipid can be embedded into this bilayer. When that is done, the bilayers can be easily visualized by fluorescence microscopy. Moreover, such dye molecules are quenched upon Cu²⁺ binding to PS lipids. The quenching occurs because of resonance energy transfer (RET) from the excited state of the dye molecule to the ground state of the Cu²⁺-2PS complex. This assay demonstrated that the dissociation constant of the metal to the binding site is very tight (low picomolar range) when 20 mol% PS is present in the membrane. Such results are highly significant, because it had not been previously known that lipid membranes have high affinity sites for any type of metal ion.

In addition to probing the binding by RET, the interfacial water structure could also be monitored by VSFS. In this case, the result is quite fascinating. The interfacial water spectrum from membranes containing phosphatidylserine and phosphatidylcholine has peaks at 3200 cm^-1 and 3400 cm^-1. The water is aligned by the net negative charge on the bilayer, since PS is a negatively charged lipid. The 3400 cm^-1 resonance results from water directly bonded to the membrane, and the 3200 cm^-1 peak also involves water molecules aligned by the interfacial potential. Significantly, the introduction of Cu²⁺ above the membrane strongly attenuates the 3400 cm^-1 peak, while leaving the one at 3200 cm^-1 nearly intact. The lack of change in the second resonance is the result of the deprotonation of two amine groups on PS for each Cu²⁺ binding event. By contrast, Ca²⁺ binding to PS lipids attenuates both the 3200 cm^-1 and 3400 cm^-1 resonance because the binding occurs without deprotonation of the amine group. As such, the net negative charge on the membrane is strongly attenuated or even reversed by Ca²⁺. VSFS reveals that these two metals bind events occur in substantially different ways. Namely, the Cu²⁺ binds to the amine and carboxylate groups, while Ca²⁺ interacts primarily with the phosphate and carboxylate groups. This binding difference leads not only to very different changes in interfacial water structure, but also is associated with very different dissociation constants to the lipid headgroup. Indeed, Cu²⁺ binds extremely tightly to the membrane with K_d values in the nanomolar or even picomolecular range depending on PS concentration, while the K_d value for Ca²⁺ is in the millimolar range.

A fluorescence quenching assay was also used by your group in a second Cu²⁺ binding study, in this case investigating its binding to phosphatidylethanolamine (PE) (4). How was the assay used in this study different from the assay used in the previous Cu²⁺ binding study (3)? What are the detrimental effects of Cu²⁺ binding to phosphatidylethanolamine in cell membranes? What are the implications of the results of this study for the influence of metal ion complexes with membranes?

The same type of RET assay that was used to monitor the binding of Cu²⁺ to PS could also be used to monitor Cu²⁺ to PE. However, we also used a second type of dye in this particular case that demonstrated free radical oxidation was occurring at the membrane surface. Copper is a redox active ion and can be cycled back and forth between Cu⁺ and Cu²⁺ by a combination of Haber-Weiss and Fenton chemistry in the presence of an oxidant, like hydrogen peroxide. It is well known that this process produces free radicals such as hydroxyl radical (×OH), which are highly reactive and can lead to lipid peroxidation. Nevertheless, hydroxyl radicals are very short lived and their ability to damage lipid membranes should depend on their proximity to the membrane surface. Our hypothesis was that that the presence of Cu²⁺ binding sites, like those on PE lipids, would greatly increase the rate of lipid oxidation compared to conditions where PE was not present in the membrane. Indeed, we were able to demonstrate that the rate of lipid oxidation increased by almost an order of magnitude in the presence of both Cu²⁺ and H₂O₂, when sufficient PE was also present in the membrane.

Together, our results with PE and PS lipids speak to the presence of specific transition metal ion binding sites in lipid membranes that are far tighter than those that had been previously expected to exist based upon electrostatic considerations alone. In fact, the binding of divalent cations to negatively charged lipids would only be expected to lead to dissociation constants of approximately 1 mM. The formation of coordination complexes for transition metal ion binding to the lone pairs on amines is far tighter because of charge transfer from the amine to the unfilled d-orbitals on the transition metal ions. Such binding for Cu²⁺ could be especially significant because metal ion dyshomeostasis in the body may lead to membrane binding and oxidative damage in neurodegenerative diseases such as Alzheimer’s and Parkinson’s. Oxidative lipid membrane damage is probably also of importance in autism. The presence of transition metal ion binding sites may therefore play a significant role in such diseases.

What are the next steps in your research?

Our research strategy is to combine spectroscopic information on interfacial water structure as well as information on small molecule, lipid membrane, or protein structure with thermodynamic measurements to unravel molecular-level mechanisms of biophysical processes at interfaces. In the past decade, we have been able to learn quite a bit about the interactions of ions with model biological systems. Going forward, we are interested in more-complex binding processes including protein aggregation and the binding of cations at the sites of high charge density. New results in these fields may help unravel some of the fundamental chemical causes of pathology in neurodegenerative diseases and those associated with protein aggregation. On the spectroscopy side, we have begun to use Raman multivariate curve resolution (MCR) spectroscopy for exploring hydration shells of macromolecules. This technique, originally developed for probing the hydration shells of small molecules in aqueous solution by Ben-Amotz and coworkers, may prove to be the first new really useful spectroscopic tool for probing interfacial water structure since the development of SFG.

References

• S.C. Flores, J. Kherb, and P.S. Cremer, J. Phys. Chem C116, 14408–14413 (2012).
• H.I. Okur, J. Kherb, and P.S. Cremer, J. Am. Chem. Soc. 135, 5062–5067 (2013).
• X. Cong, M.F. Poyton, A.J. Baxter, S. Pullanchery, and P.S. Cremer, J. Am. Chem. Soc. 137, 7785–7792 (2015).
• M.F. Poyton, A.M. Sendecki, X. Cong, and P.S. Cremer, J. Am. Chem. Soc. 138, 1584–1590 (2016).