Detecting Interplay of Chirality, Water, and Interfaces for Elucidating Biological Functions

Conspectus

Chemists have long been fascinated by chirality, water, and interfaces, making tremendous progress in each research area. However, the chemistry emerging from the interplay of chirality, water, and interfaces has been difficult to study due to technical challenges, creating a barrier to elucidating biological functions at interfaces. Most biopolymers (proteins, DNA, and RNA) fold into macroscopic chiral structures to perform biological functions. Their folding requires water, but water behaves differently at interfaces, where the bulk water hydrogen-bonding network terminates. A question arises as to how water molecules rearrange to minimize free energy at interfaces while stabilizing macroscopic folding of biopolymers to support biological function. This question is central to solving many research challenges, including the molecular origin of biological homochirality, folding and insertion of proteins into cell membranes, and design of heterogeneous biocatalysts. Researchers can resolve these challenges if they have theoretical tools to accurately predict molecular behaviors of water and biopolymers at various interfaces. However, developing such tools requires validation by experimental data. These experimental data are scarce because few physical methods can simultaneously distinguish chiral folding of the biopolymers, separate signals of interfaces from the overwhelming background of bulk solvent, and differentiate water in hydration shells of the polymers from water elsewhere.

We have recently illustrated these very capacities of chirality-sensitive vibrational sum frequency generation spectroscopy (chiral SFG). While chiral SFG theory dictates that the method is surface-specific under the condition of electronic non-resonance, we show the method can distinguish chiral folding of proteins and DNA, and probe water structures in the first hydration shell of proteins at interfaces. Using amide I signals, we observe protein folding into β-sheets without background signals from α-helices and disordered structures at interfaces, thereby demonstrating the effect of two-dimensional crowding on protein folding. Also, chiral SFG signals of C-H stretches are silent from single-stranded DNA, but prominent for canonical antiparallel duplexes as well as non-canonical parallel duplexes at interfaces, allowing for sensing DNA secondary structures and hybridization. In establishing chiral SFG for detecting protein hydration structures, we observe an H₂¹⁸O isotopic shift that reveals water contribution to the chiral SFG spectra. Additionally, the phase of the O-H stretching bands flips when the protein chirality is switched from L to D. These experimental results agree with our simulated chiral SFG spectra of water hydrating the β-sheet protein at the vacuum-water interface. The simulations further reveal that over 90% of the total chiral SFG signal comes from water in the first hydration shell. We conclude that the chiral SFG signals originate from achiral water molecules that assemble around the protein into a chiral supramolecular structure with chirality transferred from the protein. As water O-H stretches can reveal hydrogen-bonding interactions, chiral SFG shows promise in probing structures and dynamics of water-biopolymer interactions at interfaces. Altogether, our work has created an experimental and computational framework for chiral SFG to elucidate biological functions at interfaces, setting the stage for probing the intricate chemical interplay of chirality, water, and interfaces.

Graphical Abstract

Introduction

Chirality, water, and interfaces are linked to the origin of life.⁵ While each has long been a subject of intense interest in chemistry, how they interact with one another at the molecular level has remained largely unexplored due to technical challenges that hinder elucidation of biological functions at interfaces. Biological functions are encoded at the molecular level in genomes and carried out by biopolymers, proteins, DNA, and RNA. These polymers are built by homochiral monomers, i.e., (L-) amino acids and (D-) nucleic acids, and fold into macroscopic chiral structures that are also homochiral.^{6, 7} For example, (L-) proteins fold into right-handed but not left-handed α-helices (Figure 1a). Correct folding of biopolymers is essential for their biological functions but requires the mediation of water.⁸ Water has been a central research area in chemistry. Its extensive hydrogen-bonding networks (Figure 1b) generate many fascinating chemical and physical properties that support life processes,⁹ including an enormous dielectric constant for modulating electrostatic interactions and a large specific heat capacity for regulating temperature. These types of properties can change drastically at an interface, where two media meet to create a boundary. Across the boundary, the asymmetric chemical environment aligns molecules (e.g., air-water, Figure 1c), manifesting new properties, such as catalyzing reactions that cannot take place in bulk solutions.

Fig. 1. Chirality, water, and interfaces: three important research areas in chemistry.

(a) A native protein formed by (L-) amino acids folds into a right-handed α-helix and its mirror-image non-native (D-) protein. (b) A hydrogen-bonding network in liquid water. (c) Asymmetric molecular environment at interfaces. Inset: a long-chain alcohol is oriented at the air-water interface.

Elucidating biological functions at interfaces requires fundamental understanding of how the asymmetric environment regulates water molecules in stabilizing correct folding of biopolymers. This is a complex question. One can grasp the complexity by imagining a biopolymer (e.g., protein) diving into water. Water molecules surrounding the polymer can stay in the bulk hydrogen-bonding network (Figure 1b) or break away from it to stabilize a particular fold of the biopolymer. Either case contributes to a delicate balance of free energy. Predicting such a balance to yield a hydrated three-dimensional structure of the biopolymer has been a grand research challenge in biophysics.⁸ Now, the biopolymer moves to an interface, where the water hydrogen-bonding networks terminate suddenly.^10–13 A layer of complexity is then added as to how water molecules rearrange at the interface to minimize their free energy while stabilizing the folded structure of the biopolymer to support biological function. Answering this question is key to many research problems, such as searching for the molecular origin of homochirality in the biological world, understanding the molecular mechanism of protein folding and simultaneously inserting into cell membranes, designing biocompatible materials for industrial applications, and innovating biocatalysts for generating renewable energy.

Researchers can solve these problems if they can count on theoretical tools to predict structures and dynamics of biopolymers and water at various interfaces. Nonetheless, these theoretical tools need experimental benchmarks for validation. High-quality experimental benchmarks are limited due to a series of stringent demands on experimental methods. Ideally, the methods do not rely on readouts from molecular labels that can perturb water and biopolymers. They can operate under ambient conditions for in situ and real-time studies. They can also distinguish different chiral macroscopic folding of the biopolymers, such as protein β-sheet versus α-helix. Moreover, they should have surface-selectivity in detecting molecules only at interfaces without the overwhelming background signals of bulk solution. Finally, they need to separate signals from water molecules in the hydration shell of the biopolymers from the vast population of water molecules elsewhere. These demands have posed a technical barrier in elucidating biological functions at interfaces.

Our recent work has shown that chirality-sensitive vibrational sum frequency generation (chiral SFG) spectroscopy can potentially overcome this technical barrier.^{1–4, 14} Based on chiral SFG theory, the method is surface specific under the condition of electronic non-resonance.^15–17 We show that chiral SFG can distinguish chiral folding of various secondary structures of proteins and DNA at interfaces.^{1, 2, 18–20} Combining experiments and computations, we demonstrate that chiral SFG can detect water vibrational structures of the first hydration shell of a β-sheet protein at an interface.^{3, 4, 14, 21} Our work has established an experimental and computational framework of chiral SFG to investigate structural and dynamical properties of water-biopolymer interactions at interfaces, introducing an approach to explore new chemistry emerging from the molecular interplay of chirality, water, and interfaces for elucidating biological functions.

Selectivity to Interfaces

Surface-selectivity.

Chiral SFG has been developed as a subtype of second-order vibrational sum frequency generation (SFG) spectroscopy, referred to as conventional SFG^22–31 or achiral SFG here. Studies of interfaces using SFG require two ultrafast laser beams, one in the visible frequency range (ω_vis) and one in the infrared frequency range (ω_IR). These two beams temporally and spatially overlap at an interface and generate second-order optical signals at the sum frequency (ω_vis + ω_IR) (Figure 2a). Vibrational SFG keeps the frequency of the visible beam constant while varying the infrared frequency to resonate with vibrations of molecules at interfaces. Thus, SFG is a vibrational tool.

Figure 2. Experimental and theoretical background of chiral SFG.

(a) An SFG experiment for studying interfaces is set up by overlapping a visible beam and an infrared beam at an interface and detecting SFG signals in the reflection geometry. The visible and infrared beams are often linearly polarized in s- or p-polarization, and the s- and p-components of the SFG signal can be isolated and measured. The setting of polarization can vary, allowing for measurements of individual or a subset of the 27 second-order susceptibility (χ⁽²⁾) elements (equations 1 and 2). (b) An achiral interface has C_∞v symmetry whereas a chiral interface has C_∞ symmetry. Group theory dictates that seven (blue) out of the 27 elements are non-zero for both achiral and chiral surfaces, but an additional six orthogonal ${\chi }_{IJK (I\ne J\ne K)}^{\left(2\right)}$ elements (red) are non-zero only for a chiral interface. Under the condition of electronic non-resonance, the psp polarization setting shown in (a) can be used to measure the orthogonal ${\chi }_{zyx}^{\left(2\right)}$ element in chiral SFG experiments. Adapted with permission from ref 20. Copyright 2014 American Chemical Society.

To differentiate conventional SFG and chiral SFG,^{15–17, 20, 32} we need to understand the formulation of the SFG optical response in terms of second-order susceptibility (χ⁽²⁾) and the physical meaning of χ⁽²⁾. An SFG experiment measures the SFG intensity (I_SFG) or optical field (E_SFG) (Figure 2a). Any light intensity is the square of the magnitude of the optical field of electromagnetic radiation, and the optical field can be expressed using unit vector ($\widehat{\mathit{x}}$, $\widehat{\mathit{y}}$, and $\widehat{\mathit{z}}$) in laboratory coordinates (x, y, z) (Equation 1 and Figure 2a). Under the electric dipole approximation, the SFG optical field is related to the incident visible optical field (E_vis) and infrared optical field (E_IR) via the second-order susceptibility, χ⁽²⁾:

$${\mathit{E}}_{\mathit{\text{SFG}}}=E_{SFG}^x\widehat{\mathit{x}}+E_{\mathit{\text{SFG}}}^y\widehat{\mathit{y}}+E_{\mathit{\text{SFG}}}^z\widehat{\mathit{z}}$$ (1)

$$E_{\mathit{\text{SFG}}}^I\propto {\sum }_{JK}{\chi }_{IJK}^{\left(2\right)}{E}_{vis}^J{E}_{IR}^K$$ (2)

The second-order susceptibility (χ⁽²⁾) is a third-rank tensor containing 27 elements, ${\chi }_{IJK}^{\left(2\right)}$, where the indices I, J, and K indicate the direction of the three optical fields and can be equal to x, y, or z in the laboratory frame. For a particular vibrational normal mode, namely the q^th normal mode, each χ⁽²⁾ element is an ensemble average of the molecular SFG response that is the hyperpolarizability (β):

$${\chi }_{IJK,q}^{\left(2\right)}\propto {\sum }_n{\sum }_{ijk}{R}_{Ii}^n{R}_{Jj}^n{R}_{Kk}^n{\beta }_{ijk,q}^n$$ (3)

where ${\beta }_{ijk,q}^n$ are the hyperpolarizability tensor elements for the n^th moiety (such as a water molecule in a molecular system or an amide group in a protein); $R_{Ii}^n$, $R_{Jj}^n$, and $R_{Kk}^n$ are elements of the rotational transformation matrix projecting the SFG response from the molecular frame to the laboratory frame; I, J, K indicate directions in laboratory coordinates (x, y, z); and i, j, and k indicate directions in molecular coordinates (a, b, c) with i, j, k = a, b, or c. Equation 3 describes a general case, where the hyperpolarizability (β) for the q^th vibrational mode of each moiety being analyzed can be different due to local interactions and vibrational coupling. The β element for the n^th moiety is related to the Raman and infrared vibrational responses of the q^th vibrational normal mode:

$${\beta }_{ijk,q}^n\propto \frac{\partial {\alpha }_{ij}^n}{\partial Q_q}\frac{\partial {\mu }_k^n}{\partial Q_q}$$ (4)

Where $\partial {\alpha }_{ij}^n/\partial Q_q$ is the transition polarizability derivative (Raman) tensor and $\partial {\mu }_k^n/\partial Q_q$ is the infrared transition dipole derivative with respect to the q^th normal mode coordinates (Q_q). Thus, a vibrational mode must be both Raman active (i.e., $\partial {\alpha }_{ij}^n/\partial Q_q\ne 0$) and infrared active (i.e., $\partial {\mu }_k^n/\partial Q_q\ne 0)$ in order to be SFG active. Hence, SFG can assess the microscopic vibrational response of molecules (β_ijk,q) by measuring the macroscopic second-order susceptibility tensor elements (${\chi }_{IJK,q}^{\left(2\right)}$).

Chiral SFG is surface sensitive but only under the condition of electronic non-resonance. This condition is fulfilled when the visible beam used in an SFG experiment is at a frequency far from any electronic resonance of the molecular systems.^{15–17, 32} Nonetheless, Belkin et al. reported chiral SFG signals from bulk pure chiral liquid even under the electronic non-resonance condition,^{33, 34} and many researchers in the field of nonlinear spectroscopy had long considered chiral SFG to be lacking surface specificity until Simpson explained how chiral SFG signals arise from the interface.^15–17 Under the condition of electronic non-resonance, Simpson considered a chiral interface with C_∞ symmetry and an achiral surface with C_∞v symmetry (Figure 2b).^15–17 These two types of surfaces share common symmetry characteristics: (1) isotropic molecular arrangement on the surface plane and (2) anisotropic molecular arrangement across the interface (Figure 1c). These two characteristics lead to seven non-zero χ⁽²⁾ elements, shown in blue in Figure 2b. Conventional SFG measures these non-zero elements and offer surface selectivity. A chiral surface, however, differs from an achiral surface in lacking reflection planes perpendicular to the interface. This lowering of symmetry results in six additional non-zero χ⁽²⁾ elements (red, Figure 2b), which are orthogonal with the characteristics of ${\chi }_{IJK (I\ne J\ne K)}^{\left(2\right)}$. These orthogonal elements are zero for achiral interfaces (C_∞v) but non-zero for chiral interfaces (C_∞). They are measured in chiral SFG experiments.

Further derivations using equations (3) and (4), as detailed elsewhere,^{15–17, 20} showed that the orthogonal χ⁽²⁾ elements, which are the source of the chiral SFG signal from a chiral interface (C_∞), originate from the symmetric components of the Raman polarization tensor (α). As Belkin et al.³³ pointed out, however, the chiral SFG signals from bulk media are contributed only by the antisymmetric tensor components. Under the condition of electronic non-resonance, the antisymmetric components of the Raman tensor are orders of magnitude smaller than the symmetric components.³⁵ Therefore, the bulk chiral SFG signals due to the antisymmetric components are negligible relative to the surface chiral SFG signal due to symmetric components. Hence, any significant chiral SFG signal measured from a molecular system with an anisotropic chiral interface in contact with an isotropic chiral bulk medium must come from the chiral interface with C_∞ symmetry. Under these circumstances, chiral SFG is surface-sensitive.

When using chiral SFG to study biopolymers at interfaces, we and many other researchers use outputs from Ti:sapphire lasers as the visible beam in the near-infrared region at roughly 800 nm.²⁰ This wavelength is far away from electronic transitions of most biopolymers, which are in the ultraviolet region (e.g., protein at ~280 nm and DNA at ~260 nm),^{20, 31} thus satisfying the condition of electronic non-resonance. However, in the cases when the visible beam used in an SFG experiment is in the visible range (e.g., the doubling of the Nd:YAG laser with an output at 532 nm) or when biopolymers contain prosthetic chromophores (e.g., retinal proteins) that absorb in the visible region (e.g., 400–600 nm), the resonance or pre-resonance effect can amplify the antisymmetric components of the Raman tensor and thereby enhance bulk chiral SFG signals. These enhanced bulk signals can potentially mix with the chiral SFG signal generated from a chiral interface, jeopardizing the surface sensitivity of chiral SFG. Nonetheless, the use of the 800-nm beam in studies of biopolymers without prosthetic chromophores meets the condition of electronic non-resonance.^{20, 36} Therefore, chiral SFG is expected to be useful in probing biopolymers at interfaces and address many important problems, such as adsorbing protein on a polymer surface, detecting DNA using biosensor, and self-assembling biomolecules to make functional materials.

Guided by theory, our chiral SFG experiments and computations assess the orthogonal ${\chi }_{zyx}^{\left(2\right)}$ elements. We obtain chiral SFG spectra by measuring ${\chi }_{zyx}^{\left(2\right)}$ using the psp polarization setting (p-polarized SFG, s-polarized visible, and p-polarized infrared) (Figure 2a).²⁰ We simulate chiral SFG spectra by calculating ${\chi }_{zyx}^{\left(2\right)}$ for direct comparison with experimental data. Combining experimental and computational approaches, we are developing a framework of using chiral SFG for investigating vibrational structures and dynamics of biopolymers and their hydration at aqueous interfaces.^{3, 4}

Selectivity to Chiral Macroscopic Folding of Biopolymers

Since Simpson developed chiral SFG theory,^15–17 researchers have been using chiral SFG to study proteins and DNA at interfaces. Chen and coworkers reported the first chiral SFG spectrum of amide I modes of antiparallel β-sheets at interfaces.³⁷ Yet, without knowledge about the chiral SFG response from other protein secondary structures, there had been limited applications of chiral SFG for protein characterization. We studied a series of protein secondary structures at aqueous interfaces and obtained spectra of not only the amide I modes but also the N-H stretches of the protein backbone.^{19, 38, 39} We found that different protein secondary structures give distinctive chiral SFG responses: disordered structures are silent for both the amide I and N-H stretching vibrational modes, α-helical structures are only active for N-H stretching but not amide I, and β-sheet structures are active for both N-H stretching and amide I.^{18, 19} The differences in selectivity of the same vibrational modes for various secondary structures is unprecedented in other vibrational tools (e.g., infrared, Raman, vibrational circular dichroism spectroscopy, etc.). This unique selectivity allows chiral SFG to simplify the deconvolution of vibrational spectra for studying protein secondary structures, which is a constant challenge in vibrational studies of proteins due to spectral overlap.^{40, 41} Moreover, infrared studies of proteins often require D₂O as solvent to suppress water background because the O-H bending modes (~1600 cm⁻¹) overlap with protein amide I modes, and the O-H stretches (~3400 cm⁻¹) overlap with protein N-H stretches. Chiral SFG does not require D₂O because chiral SFG is sensitive only to chiral surfaces with C_∞ symmetry (Figure 2b) and is not sensitive to the majority of water molecules in the systems under study. We then applied chiral SFG to study amyloid aggregation on lipid surfaces,^{19, 38} correlate orientations of protein aggregates at interfaces with disease mechanisms,⁴² examine biological functions of a native surface-active protein,⁴³ and develop methods for probing proton exchange of proteins at interfaces.⁴⁴ Other researchers have also used chiral SFG to characterize protein structures, study ultrafast dynamics of proteins, and probe other amide modes at interfaces.^{31, 45–47}

More recently, we used chiral SFG to demonstrate the crowding effect of protein folding at interfaces.¹ We designed a molecular construct by conjugating a lipid to a protein (Figure 3a). The lipid part anchors the construct at the air-water interface while the protein part is a zinc finger. The zinc finger exhibits α-helical and disordered structures in the absence of zinc ions but folds into the ββα structure comprising an antiparallel β-hairpin and an α-helix upon binding to zinc ions. Because the α-helical and disordered structures do not contribute to the amide I signals, chiral SFG can detect zinc-induced folding of the β-hairpin structures without background from the α-helical and disordered structures (Figure 3b). This background-free detection makes the interpretation of chiral SFG spectra much simpler than that of achiral SFG spectra (Figure 3c). A comparison of Figures 3b and 3c shows that the chiral amide I signals can serve as binary readouts to report zinc finger folding (Figure 3b). Our study shows that steric repulsion can inhibit folding of zinc fingers at the interface, whereas releasing the steric repulsion by adding phospholipid as a spacer can restore the folding (Figure 3a), demonstrating the two-dimensional crowding effect on a protein folding at the interface.

Figure 3. Two-dimensional crowding effect on protein folding.

(a) Zinc finger covalently linked to palmitic acid for surface anchoring; zinc ion induces folding of the N-terminus from the disordered structure to the β-hairpin; 1,2-dipalmitoyl-sn-glycerol (DG) is used as a spacer to control the crowding at the air/water interface. (b) Chiral and achiral SFG spectra in the amide I region under three conditions: (i) without Zn²⁺ and without DG spacer lipid (red); (ii) with Zn²⁺ and without DG (blue), (iii) with Zn²⁺ and with DG (green). Adapted with permission from ref 1. Copyright 2018 the Royal Society of Chemistry.

Aside from proteins, chiral SFG can also probe DNA at interfaces. Geiger and coworkers reported C-H stretches of duplex DNA,^48–50 attributing a major vibrational band to methyl (CH₃) stretches. However, Simpson’s chiral SFG theory predicts that methyl groups with three-fold symmetry (C_3v) cannot generate chiral SFG signals.^{16, 17, 32} Looking into this discrepancy, we showed that intermolecular and/or intramolecular vibrational coupling can reduce symmetry, thereby breaking the C_3v symmetry and making the methyl stretches chiral-SFG active.²¹ We then studied C-H stretches from DNA duplexes at various interfaces.² We established three sample platforms for probing DNA at interfaces (Figure 4a). Different from previous studies, we probed the phase of the chiral SFG response and used single-stranded DNA as controls. With these strategies, we made several new observations.² First, chiral SFG is silent to single-stranded DNA but provides strong signals from double-stranded DNA (Figure 4b), offering simple binary readouts for sensing DNA hybridization. Second, the phase of chiral SFG spectra reverses when the DNA duplex is switched from the native (D-) to the non-native (L-) form (Figure 4c), indicating that chiral SFG can distinguish the absolute chirality of DNA duplexes. Third, chiral SFG captures the aromatic C-H stretching modes of nucleobases in the DNA duplex at >3000 cm⁻¹ whereas these modes are silent in the achiral SFG spectra (Figure 4b), demonstrating that chiral SFG can probe highly specific local structures for revealing DNA-drug interactions and DNA photodamage. Finally, chiral SFG detects C-H stretches not only from the canonical antiparallel DNA duplex but also from a non-canonical parallel duplex (Figure 4d), suggesting the promise of chiral SFG for characterizing DNA secondary structures. Overall, our results suggest that chiral SFG can be potentially useful to build high-throughput DNA sensors, probe various modes of drug binding to DNA, and beyond.

Figure 4. Characterization of DNA at interfaces.

(a) Surface platforms for immobilizing DNA at the air-glass interface using a covalent linkage (anchor), at the air-glass or crystalline quartz interface (island), and at the air-water interface using a stearyl moiety (buoy). The structure of an adenine—thymine base pair. (b) Chiral (psp) and achiral (ssp) SFG spectra of single- and double-stranded DNA at the air-glass interface. (c) Heterodyne phase-resolved chiral SFG spectra of (L-) and (D-) (dA)₁₂ · (dT)₁₂ double-stranded DNA on quartz. Chemical structures of native (D-) DNA versus non-native (L-) DNA. (d) Chiral SFG spectra of stearyl-(dA)₁₂·(dT)₁₂ double-stranded DNA, stearyl-(dA)₁₂, and stearyl-(dT)₁₂ at the air-water interface. The chiral SFG signal of stearyl-(dA)₁₂ is due to a non-canonical parallel double helix of (dA)₁₂· (dA)₁₂. Adapted with permission with permission from ref 2. Copyright 2022 American Chemical Society.

Selectivity to Hydration of Biopolymers at Interfaces

We have recently performed experiments and computations to investigate water structures in the hydration shell of a β-sheet protein at an interface.^{3, 4, 14, 21} In 2017, Petersen and coworkers reported the chiral SFG signal from a double-helix DNA at a frequency around 3200 cm⁻¹ and attributed the signal to O-H stretches of water molecules embedded in the minor groove of DNA,⁵¹ revealing the potential of chiral SFG to probe hydration structures of biomolecules at interfaces. Nonetheless, before using chiral SFG to solve any problem, we need to understand how it is possible for achiral water molecules to generate a chiral SFG response. Building on our experience of studying proteins, we asked whether we could observe chiral SFG signals of water hydrating proteins, and if yes, what part of the system generates the signals.

To address these questions, we examined a model peptide (LK₇β) that forms antiparallel β-sheets at interfaces (Scheme 1).⁵² We deposited the peptide on a glass surface and obtained chiral SFG spectra at around 3200 cm⁻¹.¹⁴ This spectral region encompasses protein N-H and water O-H signals. Thus, we must differentiate the water signals from the protein signals. Because only the water O-H stretches can be red-shifted by ¹⁸O substitution, we used isotopically labeled H₂¹⁸O or D₂¹⁸O as solvent. We observed selected vibrational bands shifting to the red by around 10 cm⁻¹ (Figure 5), thus confirming chiral SFG signals from water (Figure 5).¹⁴ Moreover, using heterodyne detection, we obtained phase-resolved chiral SFG spectra of LK₇β on a crystalline quartz surface.³ When switching the protein chirality from the native (L-) to the non-native (D-) form, we observed flipping of the phase for not only the protein N-H stretches but also the water O-H stretches (Figure 6).³ This observation has profound implications because not only biopolymers but also their hydration structures are chiral, raising a question about the role of water in the molecular evolution of biological homochirality, where native proteins are in the (L-) form and nucleic acids are in the (D-) form.

Scheme 1.

Structure of antiparallel β-sheet LK₇β (LKLKLKL).

Figure 5. Chiral SFG spectra of LK₇β deposited on glass slides.

The samples are prepared using solvents of (a) H₂O, (b) D₂O, (c) H₂¹⁸O, and (d) D₂¹⁸O. Each spectrum exhibits a minor and a major vibrational band. The ¹⁸O substitution shifts the minor peak in the H₂O spectrum and both major and minor peaks in the D₂O spectrum by about 10 cm⁻¹, confirming water contributions to the chiral SFG spectra. Adapted with permission from ref 14. Copyright 2019 American Chemical Society.

Figure 6. Phase-resolved chiral SFG spectra of LK₇β on quartz surface.

(a) (L-) LK₇β and (b) (D-) LK₇β. Chiral SFG spectra in the O-H/N-H region (~3200 cm⁻¹) of LK₇β prepared in H₂O and H₂¹⁸O and chiral SFG spectra in the O-D/N-D region (~2400 cm⁻¹) of LK₇β prepared in D₂O and D₂¹⁸O. All spectra are fit to Lorentzian lineshapes. Individual vibrational bands based on the fitting are shown in gray, and the sum of these peaks are shown in solid lines. Red or blue dashed lines indicate perturbation by ¹⁸O substitution for the highest (ω₁) and lowest (ω₄) vibrational bands that are assigned to water O-H stretching. The two middle (ω₂ and ω₃) bands are not impacted by ¹⁸O substitution and are assigned to protein N-H or N-D stretches. The parameters fitting the four spectra of (L-) LK₇β are comparable to those of (D-) LK₇β, except the phase of each vibrational band reverses. Adapted with permission from ref 3. Copyright 2020 National Academy of Science.

To understand our experimental results at the theoretical level, we computed the chiral SFG response of water around the β-sheet protein.^{3, 4} Although chiral SFG signals of water were first reported from DNA,⁵¹ simulation of the chiral SFG signals of water around any biomolecules had yet to be reported. We applied Skinner’s electrostatic map^53–55 to simulate the chiral SFG spectra of water around the β-sheet protein, which is modeled by five peptide strands of LK₇β in either the (L-) or (D-) forms at the vacuum-water interface (Figure 7a). Because these peptide strands are highly amphiphilic with hydrophobic leucine sidechains pointing up to the air phase and hydrophilic positively charged lysine sidechains pointing down to the aqueous phase, the orientation of this protein relative to the interface is highly restricted, making the β-sheet lie flat at the interface.⁵⁶ The simulated spectra show strong water O-H stretching bands, and the phase of these bands reverses with the protein chirality (Figure 7b), agreeing with experimental observations (Figures 6 and 7c).^{3, 4} The combined computational and experimental studies (Figures 5–7) have established that chiral SFG signals of achiral water molecules originate from water assembling around the protein into a chiral supramolecular structure with the chirality imprinted from the protein.

Figure 7. Computational model and simulated chiral SFG spectra and comparison to experimental results.

(a) Molecular model for the simulation: an LK₇β pentamer (black) at the vacuum-water interface forming an antiparallel β-sheet surrounded by water molecules (yellow). (b) Simulated chiral SFG spectra of (L-) LK₇β and (D-) LK₇β comprising only the water O-H stretching responses. (c) Experimental heterodyne phase-resolved spectra in the N-H/O-H region of (L-) LK₇β and (D-) LK₇β showing signals with opposite signs. Adapted with permission from ref 4. Copyrights 2022 American Chemical Society.

Further computational analyses show that chiral SFG can probe vibrational structures of water hydrating the β-sheet protein at the air-water interface with selectivity to the first hydration shell.⁴ Using a Voronoi tessellation algorithm, we identified immediate neighbors of the protein as water molecules in the first hydration shell and immediate neighbors of the first hydration shell as water molecules in the second hydration shell. We could then simulate chiral SFG spectra of the first and second hydration shells. The simulated spectra reveal that almost all (>92%) of the chiral SFG signal comes from the first hydration shell, and the signal from the second hydration shell drops to almost zero (Figure 8).

Figure 8. Selectivity of first hydration shell.

(a) The model system contains the LK₇β protein (yellow) with all water (black), but the spectra can be computed from select subsets of water, such as only the first hydration shell (red) or only the second hydration shell (blue). (b) Calculated spectra of the O-H stretches from all water molecules (black), the first hydration shell (red), and the second hydration shell (blue). Adapted with permission from ref 4. Copyright 2022 American Chemical Society.

To understand the remarkable selectivity to the first hydration shell, we analyzed water dipoles around the β-sheet protein at the interface (Figure 9).⁴ We found that water dipoles generally orient toward the protein because of the positively charged lysine residues. The organization outside the first hydration shell is quite symmetric (Figures 9a–c), giving rise to at least two reflection planes (σ_// and σ_⊥) perpendicular to the interface. Hence, the organization of water molecules outside the first hydration shell is in C_∞v symmetry and thus not chiral-SFG active according to chiral SFG theory (see Figure 2b). Inside the first hydration shell, however, local interactions with the protein break the pattern of symmetry and remove the reflection planes (Figures 9d and 9e). Thus, water molecules inside the first hydration shell exhibit C_∞ symmetry, thereby generating chiral SFG signals. This analysis has established an intuitive but rigorous molecular explanation based on chiral SFG theory (Figure 2b) for the selectivity of chiral SFG to first hydration shell structures of proteins at interfaces.⁴

Figure 9. Water dipole arrangements around LK₇β at the vacuum-water interface.

(a) Top view of water dipoles at the vacuum-water interface. Each arrow is a unit vector of a water dipole averaged in a 1-ų grid box with protein (yellow), first hydration shell (red), and second hydration shell (blue). (b), (c), and (d) Zoomed-in details in the blue, red, and orange boxes. Water dipoles in panel (c) point toward LK₇β due to the positively charged lysine residues. (e) Zoomed-in depiction of the orange box showing the magnitude of each water dipole. Water arrangement outside the first hydration shell constitutes at least two reflection planes (σ_// and σ_⊥) perpendicular to the interface, and thus the water molecules are arranged in C_∞v symmetry. Water arrangement within the first hydration shell is less symmetric without a reflection plane and thus has C_∞ symmetry. Therefore, chiral SFG theory (Figure 2b) implies that only water molecules in the first hydration shell are chiral-SFG active. Adapted with permission from ref 4. Copyright 2022 American Chemical Society.

This work has developed a synergistic experimental and computational framework that can be further used to understand molecular structures and dynamics of water in hydration shells of biopolymers at interfaces.^{3, 4, 14, 21} In this framework, experimental chiral SFG vibrational spectra can serve as quantitative benchmarks for validation of computational tools developed for predicting structural evolution and dynamics of water-biopolymer interactions at aqueous interfaces.

Conclusion and Outlook

Our recent experiments and computations have established unique selectivity of chiral SFG for probing the intricate interplay of chirality, water, and interfaces. This method can simultaneously differentiate (1) surface signals from bulk signals, (2) various secondary structures of biopolymers, and (3) water molecules in hydration shells of the biopolymers from water molecules elsewhere. Because chiral SFG offers surface-selectivity and works under ambient conditions for in situ and real-time studies, it can complement NMR, X-ray crystallography, neutron scattering, and terahertz spectroscopy in probing hydration structures of biomolecules. The unique selectivity of chiral SFG will allow it to address problems related to structure-function correlations of biomolecules at interfaces.

Nonetheless, there are still several remaining challenges. One of these challenges is the difficulty in interpreting water spectra due to the coupling of water O-H stretches with N-H stretches of biopolymers. To tackle this problem, our group developed an electrostatic-frequency map that allows for simulating N-H spectra of the protein backbone and thereby examining the effect of the vibrational coupling between the protein backbone and water.⁵⁷ We also need to develop new methods to simulate the N-H stretching response from nucleic acids. Another challenge is for spectroscopists to obtain high-quality biological samples and identify problems with genuine biological relevance, which can be most effectively overcome by collaboration with researchers in the biological and biomedical fields. In addition, to broaden the scope of applications, chiral SFG needs to be combined with advanced spectroscopic methods, such as pump-probe and multi-dimensional techniques. This method development will enable studies of ultrafast dynamics of important biomolecular events, such as photodamage of DNA, solvation dynamics of photoactive proteins, and intramolecular vibrational energy redistribution of biocatalysts.

Despite these challenges, our case study of the LK₇β protein at the air-water interface has produced an experimental and computational framework for investigating structural and dynamical properties of biomolecules at aqueous interfaces. This framework can serve as the basis for future developments to elucidate biological functions at interfaces for attacking fundamental and engineering problems. The unique selectivity of chiral SFG should empower researchers to explore new chemistry emerging from the molecular interplay of chirality, water, and interfaces.

Biographies

Elsa C. Y. Yan is a professor in the Chemistry Department at Yale University. She obtained her B.S. in Chemistry from the Chinese University of Hong Kong. Working with Kenneth Eisenthal at Columbia University, she developed second-order spectroscopy to probe colloidal surfaces. Then, she became a joint postdoctoral fellow with Richard Mathies at UC Berkeley and Thomas Sakmar at The Rockefeller University, exploring functional mechanism of transmembrane receptors. At Yale, she develops and applies optical spectroscopy to study biomolecular structures and functions. She focuses on fundamental studies using chiral SFG spectroscopy to probe chiral biomacromolecular structures and their hydration at interfaces.

Ethan A. Perets is a postdoctoral researcher at Harvard University working to develop novel technologies for neurobiological imaging. Ethan received his B.A. in biochemistry from Columbia University, his M.A. in the history of art from the Courtauld Institute of Art, and his Ph.D. in chemistry from Yale University. Working alongside Professors Elsa C. Y. Yan and Sharon Hammes-Schiffer at Yale, Ethan applied chiral sum frequency generation spectroscopy to study the interplay of secondary structure and hydration of proteins and DNA at molecular interfaces.

Daniel Konstantinovsky is a PhD student at Yale University working on modeling chiral sum frequency generation spectroscopy responses of proteins, nucleic acids, and the water around them with the goal of understanding the physics of the macromolecular solvation shell. Daniel received his B.S. in Chemistry from Haverford College and subsequently worked in Dr. Alexander Rudensky’s group at Memorial Sloan Kettering, studying the effect of bile acids on regulatory T cells, before turning to computational work at Yale as a joint student in the Hammes-Schiffer and Yan groups.

Sharon Hammes-Schiffer is a Sterling Professor of Chemistry at Yale University. She received her B.A. from Princeton University and her Ph.D. from Stanford University, followed by two years at AT&T Bell Laboratories. Her academic career progressed from the University of Notre Dame to Pennsylvania State University to the University of Illinois to Yale University. Her work encompasses the development of analytical theories and computational methods, as well as applications to a wide range of chemical, biological, and interfacial systems.