2D-IR studies of cyanamides (NCN) as spectroscopic reporters of dynamics in biomolecules: Uncovering the origin of mysterious peaks

Abstract

Cyanamides (NCN) have been shown to have a larger transition dipole strength than cyano-probes. In addition, they have similar structural characteristics and vibrational lifetimes to the azido-group, suggesting their utility as infrared (IR) spectroscopic reporters for structural dynamics in biomolecules. To access the efficacy of NCN as an IR probe to capture the changes in the local environment, several model systems were evaluated via 2D IR spectroscopy. Previous work by Cho [G. Lee, D. Kossowska, J. Lim, S. Kim, H. Han, K. Kwak, and M. Cho, J. Phys. Chem. B 122(14), 4035–4044 (2018)] showed that phenylalanine analogues containing NCN show strong anharmonic coupling that can complicate the interpretation of structural dynamics. However, when NCN is embedded in 5-membered ring scaffolds, as in N-cyanomaleimide and N-cyanosuccinimide, a unique band structure is observed in the 2D IR spectrum that is not predicted by simple anharmonic frequency calculations. Further investigation indicated that electron delocalization plays a role in the origins of the band structure. In particular, the origin of the lower frequency transitions is likely a result of direct interaction with the solvent.

INTRODUCTION

Fluorescence and infrared (IR) spectroscopy have greatly benefitted from specific probes which have allowed for the monitoring of functional events of interest in biological systems.^1–9 Discovering vibrational reporters with novel properties is important to potentially improve the ability to investigate site-specific structural dynamics and gain insight into which regions of a biomolecule are responsible for functionally relevant motions. The process of evaluating the properties of novel vibrational reporters is best done by the characterization of their vibrational signatures via two-dimensional infrared (2D IR) spectroscopy. Important spectroscopic features, such as sensitivity to electric field strength, chemical exchange, and the ability to monitor degrees of hydration, have been observed for many infrared probes, e.g., azides, nitriles, thiocyanates, selenocyanates, and isonitriles.¹⁰ In particular, nitriles have been incorporated in large protein systems¹¹ and HIV drugs¹² to capture the conformational dynamics and the changes in the local environment during biological events such as protein activity (folding, unfolding, aggregation, and peptide–membrane interaction) and drug binding. Attaching the CN group to heavier atoms, such as S or Se, has been shown to increase the vibrational lifetime of the reporter which extends the observation windows to time scales of interest for larger biological systems. For example, the vibrational lifetime can be increased to hundreds of picoseconds by utilizing SeCN reporters.^13–15 However, the significant decrease in the transition dipole strength of these reporters compared to CN can be a limitation when they are used for protein or RNA/DNA dynamics. Non-invasive vibrational probes have also been utilized in prior studies to map structural dynamics in biological systems such as DNA and RNA.^16–20 For example, backbone modes at the DNA–water interface were used to detect variations in the hydration shell surrounding DNA.²⁰ In another study, the vibrational coupling signature between backbone residues was utilized to identify differences in the hydration of RNA and DNA.¹⁶

Although the IR signature of several of these probes is believed to be quite localized, others, such as cyanates, aryl azides, and cyanamides (NCN), are known to have accidental Fermi resonances which can make structural analysis more challenging.^10,21–23 For example, despite the expansion to multiple dimensions via 2D IR, these vibrational transitions and their Fermi resonances often remain overlapped, masking important spectral features related to structure (i.e., cross peaks). Thus, electron delocalization is of key importance in the transition energies as a result of the distribution of charges in conjugated molecules. For example, variations in electron delocalization may result in the appearance of additional absorption bands, changes in the vibrational frequency, and even cause significant changes in the transition dipole strength.^10,24

Since the transition dipole strength of cyanamide has been suggested to be three-fold larger than a cyano-probe and comparable to that of isonitrile and azido-probes, Lee and co-workers¹⁰ studied cyanamides in several amino acid analogues. For example, cyanamide-derivatized alanine and phenylalanine were synthesized and characterized via pump–probe IR spectroscopy. While the alanine cyanamide analogue showed a single transition in the FTIR spectrum, multiple transitions were observed for the FTIR spectrum of the phenylalanine cyanamide analogue in various solvents. The cyanamide transition is in an isolated region of the IR spectrum, 2220–2243 cm⁻¹, similar to cyano- and azido-transitions. NCN has structural similarity to an azido transition (RN—N≡N ↔ RN=N=N), which has been shown to cause limited perturbation in biological systems.^4,25 Minimal perturbation in biological systems have been determined experimentally and computationally for these commonly incorporated vibrational probes.^8,26–31 Despite the increased dipole strength relative to nitriles, the cyanamide vibrational lifetime is similar to azido-probes (∼1 ps).¹⁰

The cyanamide vibrational probes studied by Cho all had a hydrogen bond donor, i.e., NHCN. We were interested in studying cyanamide vibrational probes without NH while also exploring the effects of delocalization using a variety of scaffolds. For this purpose, several conjugated and non-conjugated cyanamide scaffolds were designed and synthesized. The number of transitions, peak positions, and vibrational lifetimes were then investigated. Herein, we demonstrate from our 2D IR measurements of NCN that aliphatic scaffolds typically avoid Fermi resonance unlike the aromatic scaffold examined. Moreover, scaffolds containing NCN embedded in 5-membered rings show an unusual vibrational band structure, not easily explained by simple computational analyses.

MATERIALS AND METHODS

Cyanamides 1–3 (Fig. 1) were synthesized following known procedures, and full details are provided in the supplementary material. In short, N-cyanomaleimide (1, CMI) and N-cyanosuccinimide (2, CSI) were prepared in modest yields by treating the maleimide and succinimide, respectively, with cyanogen bromide and triethylamine following published methods.^32,33 N-(4-tolyl)cyanamide (3) was prepared in 66% yield by treating 4-methylaniline with cyanogen bromide and sodium bicarbonate in toluene following a literature procedure.³⁴

FIG. 1.

Chemical structures of the cyanamides studied.

The cyanamides were dissolved in DMSO (dimethyl sulfoxide), THF (tetrahydrofuran), and water for infrared measurements. The extinction coefficient at 2117 cm⁻¹ for diisopropylcarbodiimide (DIC), a related cyanamide, was determined from the FTIR spectrum to be 2900 ± 300 M⁻¹ cm⁻¹. All experiments were performed at concentrations of ∼5 mM. A Harrick sample cell with two CaF₂ and a 50 µm spacing was utilized to collect the infrared spectrum on a Nicolet 6700 FTIR spectrometer.

Heterodyned spectral interferometry was utilized for obtaining the spectra. Fourier-transform limited 80 fs pulses with a central wavelength ranging from 4500 to 4700 nm were employed in the 2D IR experiments. Three ∼1 µJ laser pulses with wave vectors k₁, k₂, and k₃ were incident to the sample generating a signal in the direction k_s = −k₁ + k₂ + k₃ with the ordering 123 (rephasing) and 213 (nonrephasing). To obtain absorptive spectra, the rephasing and nonrephasing 2D frequency spectra were properly phased and combined. To observe any changes in the spectral characteristics, the waiting time, T, between the second and third pulse was varied from 0 ps to 2 ps. After appropriate Fourier transforms along the coherence, τ, and detection, t, axes, the 2D IR spectra were plotted as ω_τ vs ω_t. The vibrational relaxation time, T₁₀, for the two main transitions were estimated from the signal strength of the diagonal peak within the 2D IR at different T values and fitted to $A=A_0{e}^{-T/T_{10}}$.

COMPUTATIONAL METHODS

The geometries of CMI and its saturated analogue, CSI, were optimized in solution (DMSO). In all cases, we added three explicit molecules of DMSO and also utilized the implicit solvation self-consistent reaction field polarizable continuum model (PCM). We used the B3LYP-D3BJ density functional in combination with def2-TZVP basis sets.^35–39 To obtain vibrational frequencies, we used the harmonic approximation. We also computed anharmonic corrections to these frequencies. The Gaussian 16 package was used for all calculations.⁴⁰ In order to analyze the potential energy curves of these molecules, we also carried out constrained geometry optimizations where we constrained the bond distances between the first N and the CN group of the NCN moiety.

Ab initio molecular dynamics (AIMD) simulations were used to probe the thermal broadening of the density functional theory (DFT)-calculated IR spectra. These simulations were carried out on isolated molecules at 300 K for 15 ps using a time step of 0.5 fs in the NVT ensemble while employing the PBE-D3BJ functional. We used def2-SVP Gaussian basis sets and the NWChem⁴¹ code for these simulations.⁴²

RESULTS AND DISCUSSION

In order to evaluate and characterize the 2D IR spectral properties of a cyanamide that does not contain a NH, N,N′-diisopropylcarbodiimide (DIC) was chosen as a model compound. A large absorption cross section (${\epsilon }_{2117\text{\hspace{0.17em}c}{\text{m}}^{-1}}^{THF}\sim 2900{\text{M}}^{-1}\text{\hspace{0.17em}c}{\text{m}}^{-1}$) was measured. The large transition dipole strength of cyanamides is a characteristic of electron delocalization between the first N and the CN of the NCN moiety within the organic framework.¹⁰ The measured extinction coefficient of DIC in THF is five-fold larger than previously observed for NHCN stretches in two cyanamide-derivatized amino acids in various solvents.¹⁰ A single narrow transition located at 2117 cm⁻¹ is observed in the infrared spectrum as a result of the NCN stretch. Additionally, the extinction coefficient, frequency, and bandwidth of this vibrational transition are sensitive to the local environment.

The 2D IR spectrum of DIC in THF at T = 200 fs reveals two transitions along the diagonal for the NCN stretch. The positive peak, located at ω_τ = ω_t = 2117 cm⁻¹, corresponds to the υ = 0 → υ = 1 transition (Fig. 2). The negative peak, arising from the υ = 1 → υ = 2 transition, is red shifted along the ω_t axis by 23 ± 1 cm⁻¹. As a result of the population relaxation time, T₁₀, the positive diagonal signal decays during the waiting time, T, with a time constant of 989 ± 63 fs. This lifetime is comparable to azide lifetimes. Azides have been widely utilized as vibrational reporters in many biomolecular studies to uncover dynamics and structure.^10,43,44

FIG. 2.

FTIR spectra of DIC in (a) water and in (c) THF and 2D IR spectra of DIC in (b) water at T = 200 fs and in (d) THF at T = 200 fs.

In water, the NCN transition is observed as the positive peak at position ω_τ = ω_t = 2111 cm⁻¹ red shifted by ∼5 cm⁻¹ compared to the transition in THF. The negative peak is anharmonically shifted by 23 ± 1 cm⁻¹ from the positive peak. In contrast to the behavior in THF, the 2D IR of the NCN region in water changes significantly with the waiting time. At T = 200 fs, the inhomogeneously broadened transitions are tilted and elongated along the diagonal of the ω_τ vs ω_t plot. The elliptically shaped transitions become less tilted as the waiting time increases, which is quantified by the change in the inverse slope as a function of time. This spectral behavior is related to the loss of correlation (spectral diffusion) between components of the inhomogeneous distribution of frequencies of the NCN mode in water. The correlation decays were determined to be 1.2 ± 0.1 ps and 2.4 ± 0.3 ps in water and THF, respectively. The positive diagonal signal strength in water decays exponentially with T, yielding a population relaxation time constant similar to THF, 1.05 ± 0.080 ps. The NCN stretch significantly overlaps with the water combination band at 2127 cm⁻¹.²¹ The large thermal grating signal generated at all frequencies by the water combination band often increases the spectral noise at waiting times larger than 1 ps.⁴³

Prior studies have observed that the infrared band structure of the NCN vibrational mode is significantly affected by the presence of a scaffold with increased electron delocalization. In particular, phenylalanine derivative containing cyanamide has shown extended vibrational signatures in the IR spectrum, likely due to Fermi resonances as suggested from quantum chemical calculations.¹⁰ To ascertain the influences of different structural scaffolding on the cyanamide vibrational mode and its unique vibrational signature, it was incorporated into a 5-membered ring and attached to a 6-membered ring and investigated via 2D IR spectroscopy. Due to the low solubility of these aromatic scaffolds, the experiments were performed in dimethyl sulfoxide (DMSO). DIC was also measured in DMSO for direct comparison with other scaffolds, showing a single transition at 2111 cm⁻¹ (Fig. S10).

The molecular scaffold 4-methylphenylcyanamide (4MPC, Fig. 3) has the cyanamide attached to a 6-membered aromatic ring. Both the linear and 2D IR spectra show multiple transitions within the region of the NCN transition. Density functional theory (DFT) calculations of the harmonic vibrational frequencies predict only one transition in this spectral region. The linear spectrum shows that the strongest transition is at 2236 cm⁻¹, which we will refer to as the fundamental NCN stretching band. There is another band at 2201 cm⁻¹ of moderate intensity and a weaker one at 2222 cm⁻¹ (Fig. 3). The 2D IR spectrum of 4MPC in the NCN region shows the presence of both diagonal and off-diagonal peaks at early waiting times which indicates anharmonic coupling between the transitions.²¹ The multiple transition pattern observed is consistent with near accidental resonances between a small number of overtone states, i.e., Fermi resonances with low frequency ring modes, previously detected with both OCN and N₃ transitions in aromatic compounds.^21–23 The 2D IR spectrum of 4MPC in DMSO has three peaks along the diagonal shown as positive peaks located at ω_τ = ω_t = 2236 cm⁻¹, 2219 cm⁻¹, 2201 cm⁻¹ and shows no significant elongation that is typically seen in inhomogeneously broadened transitions. The negative diagonal peaks are anharmonically shifted by 40 cm⁻¹. Coupling between the three transitions is observed within the cross peak region located at {ω_τ, ω_t} values of {2197, 2220} cm⁻¹, {2220, 2237} cm⁻¹, and {2200, 2237} cm⁻¹. The intensity ratio of the 2201 cm⁻¹ and 2236 cm⁻¹ diagonal transitions is 1:0.84. Significant overlap of the transitions in 4MPC in the 2D IR spectrum makes the analysis of dynamic studies somewhat challenging. The vibrational lifetimes of the NCN modes was measured to be 0.40 ± 0.03 ps. While this is shorter than the lifetimes measured by Lee and co-workers and that measured in DIC, the presence of the Fermi resonance is known to lead to shorter vibrational lifetimes.¹⁰

FIG. 3.

(a) Molecular structure of 4-methyl-phenylcyanamide (4MPC), (b) FTIR spectrum of 4MPC in DMSO, and (c) 2D IR spectrum at T = 200 fs of 4MPC in DMSO.

To further identify the modes responsible for Fermi resonance, the frequency distributions were calculated from DFT. 4-MPC has calculated frequencies at 2305 (fundamental NCN stretch), 2279 (overtone), 2300 (combination), and 2296 (combination) cm⁻¹. The combination transition at 2300 cm⁻¹ is near-degenerate with the fundamental band at 2305 cm⁻¹ (Table I). The energy separation between these peaks favors the Fermi resonance effect, as observed experimentally (Fig. 3). Therefore, the transitions located at 2201–2222 cm⁻¹ in the experimental spectrum are likely due to Fermi resonance. Based on prior work, quantum chemical calculations for aromatic cyanamide suggest Fermi resonance is the cause for the complicated lineshapes seen in the IR spectra.¹⁰ Fermi resonance often complicates the spectral analysis by showing significant overlap with the fundamental transition.

TABLE I.

Calculated dominant peaks (cm⁻¹) and their intensities, ν (I), obtained from anharmonic corrections to the IR spectrum of 4-MPC.

	Peak	Intensity		Peak	Intensity		Peak	Intensity
Fundamental	2305	113.1	Overtone	2279	6.9	Combination	2245	4.2
							2296	12.7
							2351	13.1
							2300	93.4
							2347	0.3

In an attempt to remove the Fermi resonance, a 5-membered ring scaffold was adopted with the NCN embedded within the ring. This scaffold also removes the NH hydrogen-bond donor from the cyanamide. Two 5 membered rings, cyanomaleimide (CMI) and cyanosuccinimide (CSI), were analyzed by FTIR and 2D IR spectroscopy. This cyanamide vibrational reporter can easily be incorporated into a peptide or protein through conjugate addition of a thiol (i.e., reaction with a cysteine side chain; see Fig. S6 of the supplementary material).

The linear IR spectra of CMI in DMSO reveal multiple transitions around 2100–2200 cm⁻¹(Fig. 4). The vibrational transitions of CMI, a conjugated scaffold, are located at 2251 cm⁻¹, 2154 cm⁻¹, 2175 cm⁻¹, and 2184 cm⁻¹. Similarly, the linear IR spectra of CSI, a more saturated molecule with less conjugation, also have multiple transitions located at 2258 cm⁻¹, 2245 cm⁻¹, 2192 cm⁻¹, and 2144 cm⁻¹ (Fig. 5). The IR intensity ratio however is different between the CMI and CSI spectra. The 2D IR spectra of both compounds show similar band structures with the peaks indicated for the FTIR along the diagonal with significant inhomogeneous broadening. Unlike the 4MPC (above) and the phenylalanine cyanamide,¹⁰ no cross-peaks are present in the off diagonal in the 2D IR spectra for these molecules. Based on DFT harmonic and anharmonic vibrational frequency analysis with the implicit DMSO solvent, only one peak is expected in the 2100–2200 cm⁻¹ region of the IR spectrum. Thus, with the presence of no anharmonically coupled cross peaks in the 2D IR spectra, the physical origin of the multiple transitions was an open question that required further computational analysis. It should be noted that the overall band shape is different between the two compounds, suggesting that the extent of conjugation by the ring plays a role in the peak distribution. Furthermore, analysis shows that the vibrational lifetime of the CSI scaffold is 0.70 ± 0.03 ps for the most intense transition (2148 cm⁻¹) and 0.80 ± 0.01 ps for the transition located at 2144 cm⁻¹. Whereas, the vibrational lifetime was ∼2 times longer (1.20 ± 0.03 ps) for the CMI scaffold for the higher frequency band at 2251 cm⁻¹ and 0.80 ± 0.01 ps for the transition located at 2184 cm⁻¹ along the diagonal. The longer lifetime for the CMI scaffold likely results from the additional electron delocalization of the highly conjugated ring system reducing vibrational damping. However, the origin of the extra vibrational transitions observed was still unknown and several approaches were explored to uncover it.

FIG. 4.

(a) Molecular structure of N-cyanomaleimide (CMI), (b) FTIR spectrum of CMI in DMSO, and (c) 2D IR spectrum at T = 200 fs of CMI in DMSO.

FIG. 5.

(a) Molecular structure of N-cyanosuccinimide (CSI), (b) FTIR spectrum of CSI in DMSO, and (c) 2D IR spectrum at T = 200 fs of CSI in DMSO.

As a first experimental approach, the temperature dependence was measured for the FTIR. However, no significant change in the vibrational spectra of either compound was observed with temperature, suggesting that the additional peaks were not due to simple population distribution between states. As a second experimental approach, the solvent was changed to determine the presence of any bathochromic effects within DMSO. The FTIR spectrum of CMI measured in acetone demonstrated that the solvent did have a significant effect on the vibrational band structure (Fig. 6). For example, the location of the vibrational transitions (2259 cm⁻¹, 2188 cm⁻¹, and 2178 cm⁻¹) and a broad transition (around 2150 cm⁻¹) is shifted to higher frequencies and their relative intensities are significantly different. These experimental findings suggest that the spectral features were, indeed, sensitive to these changes in solvent due to the differences in the electron donor ability of the solvents. However, in order to more fully understand the frequency distribution, several quantum mechanical computations were undertaken.

FIG. 6.

FTIR spectrum of CMI in acetone showing solvent dependence on the band structure.

DFT calculations and AIMD simulations were performed on CMI and CSI. For the frequency analysis based on the equilibrium structures from DFT calculations in a vacuum, an implicit solvent or an explicit solvent, the band structures seen in the experimental spectra were not reproduced. All these calculations predicted only one peak in the 2100–2300 cm⁻¹ region, with little possibility for anharmonic peaks. We, therefore, proceeded to analyze the potential energy curves associated with the N—C bond distance of the N—C≡N moiety in CMI and CSI.

Overall, we find that the cubic contributions to the potential are similar for both molecules (Fig. 7). We, therefore, conclude that empirical peaks at 2144 cm⁻¹ and 2175 cm⁻¹ are likely not due to anharmonic effects, in agreement with the 2D IR data (Figs. 4 and 5). Further examination illustrates that the potential energy curves are quite flat near equilibrium. To illustrate, CMI with a N—C bond distance of 1.13 Å is only about 0.9 kcal/mol less stable than the equilibrium structure (1.152 Å).

FIG. 7.

Potential energy curves associated with the (ring) N—C distance of CMI and CSI. The inset is focused on the PEC about 1 kcal/mol from the equilibrium structures.

As expected, structures with longer N—C distances have lower frequencies for the NCN stretch as well as higher intensities. We see a similar situation for the saturated compound CSI (Fig. 7). We conclude that thermal averaging will result in rather broad peaks for the fundamental frequencies of both compounds. These peaks are likely to be more asymmetric toward the low-frequency portion of the IR spectrum (Fig. 7). We investigate these effects by performing AIMD simulations over a period of 12.1 ps (Fig. 8). The peaks are broad and asymmetric, as expected. However, the AIMD simulations on isolated molecules were not able to reproduce the extra vibrational transitions observed in the FTIR and 2D IR experiments for the CMI and CSI scaffolds in DMSO.

FIG. 8.

IR spectrum of CMI (conjugated) and CSI (saturated) obtained from AIMD simulations at 300 K.

Since the data indicated that the solvent had a significant effect on the band structure of the CMI scaffold, we decided to computationally investigate the effect of a repulsive interaction of the solvent with the conjugated ring by varying the distance between a DMSO molecule and CMI. The calculations showed a significant change in the vibrational frequencies as the distance between DMSO and CMI is reduced. At a distance of 2.3 Å (NCN to O of DMSO) between the CMI and solvent, the NCN stretch is reduced to 1961 cm⁻¹, which is significantly different than the 2350 cm⁻¹ frequency for the equilibrium structure of CMI. Additionally, if the DMSO and CMI are optimized with a constraint of 2.1 Å (between the S atom of DMSO and the NCN of CMI), the NCN stretch is calculated to be 2180 cm⁻¹; see the supplementary material. With appropriate scaling factors, this NCN stretch would be around 2134 cm⁻¹, a region that coincides with the experimentally observed mysterious peaks. In a similar manner, repulsive interactions between CMI and the O atom of acetone results in a shift of the NCN stretch to 2260 cm⁻¹. This occurs when the distance between the two molecules is constrained to about 1.7 Å. This intermolecular distance is shorter than that of CMI and DMSO, suggesting the existence of subtle differences in the degree of repulsive interactions between CMI and the lone pair electrons of the two solvents. This likely explains the differences in the energies and bandwidths of the mysterious peaks (Figs. 4 and 6). On the whole, these results suggest the existence of the electron repulsion effect of DMSO on the ring mode of CMI. This has previously been seen before with the tyrosine ring mode.²

CONCLUSIONS

A combination of 2D IR spectroscopy and computational analysis was used to evaluate the properties of the cyanamide vibrational reporter using different structural scaffolds. The FTIR and 2D IR spectra of the different cyanamides in DMSO showed significant variations in the number of transitions, peak positions, and vibrational lifetimes. In DIC, a simple single transition was observed with a vibrational lifetime of ∼1 ps. The correlation decay was measured through the center line slopes of the transition yielding lifetimes of 1.2 ± 0.1 ps and 2.4 ± 0.3 ps in water and THF, respectively. The 2D IR spectrum of 4MPC (conjugated to an aromatic ring) showed strong anharmonic coupling, i.e., an accidental Fermi resonance, resulting in significant overlapping cross peaks within the spectrum. The overlap of the transitions and the significantly shorter vibrational lifetime (0.40 ± 0.03 ps) render this scaffolding less useful than DIC as a vibrational reporter.

Most interestingly, the 2D IR spectra of the CMI and CSI scaffolds, where the NCN is embedded in a 5-membered ring, exhibited distinct separation between multiple peaks along the diagonal and no cross peaks, indicating no evidence of strong anharmonic coupling, i.e., Fermi resonance. Significant differences between the CMI (conjugated) and CSI (saturated) compounds indicated that electron delocalization plays a role in the band structure. To uncover the origins of these transitions, several experimental and computational approaches were undertaken. First, DFT vibrational analysis predicted a single transition in this region. Further analysis with variable temperature experimental data, DFT, and ab initio molecular dynamics simulations suggest that these transitions are not due to population distribution within the thermal level. However, the vibrational band structure was affected by the solvent. This experimental observation, in conjunction with the computational results of the frequency dependence on the distance of a DMSO solvent molecule from the ring, revealed that the origin of the lower frequency transitions is likely a result of electron repulsion between the conjugation of the 5-membered rings and the lone pairs of electrons on the DMSO.

Thus, cyanamide is a vibrational reporter with a large absorption cross section that has a vibrational frequency between the azido- and cyano-transitions in an open window in the IR spectrum with respect to absorbances in biological systems. The sensitivity of its IR transition to variations in the local environment is beneficial for monitoring structural dynamics by observing changes in the spectral lineshape. Finally, this probe can be incorporated into a peptide or protein by a conjugate addition reaction with a cysteine side chain (see Fig. S6 of the supplementary material), thereby providing a new way to determine structural dynamics in important biomolecules using 2D IR spectroscopy.

SUPPLEMENTARY MATERIAL

The synthetic and computational methods for the different molecular scaffolds are explained in more detail. The optimized geometries, sample input files for DFT, and AIMD for the calculations are shown and described. An example of how these systems can be applied to larger protein systems through conjugated addition is illustrated. The second derivative IR spectra for each scaffold is presented to help identify the different peaks contributing to complete spectrum. The frequencies for the various compounds are listed. 2D IR spectra of N,N′-diisopropylcarbodiimide in DMSO is included for comparison with other scaffolds spectra in DMSO. Finally, the normalized correlation decay of N,N′-diisopropylcarbodiimide (DIC) in water and THF is shown in the supplementary material.