Terahertz Spectroscopic Analysis in Protein Dynamics: Current Status

Proteins play a key role in living organisms. The study of proteins and their dynamics provides information about their functionality, catalysis and potential alterations towards pathological diseases. Several techniques are used for studying protein dynamics, e.g., magnetic resonance, fluorescence imaging techniques, mid-infrared spectroscopy and biochemical assays. Spectroscopic analysis, based on the use of terahertz (THz) radiation with frequencies between 0.1 and 15 THz (3–500 cm−1), was underestimated by the biochemical community. In recent years, however, the potential of THz spectroscopy in the analysis of both simple structures, such as polypeptide molecules, and complex structures, such as protein complexes, has been demonstrated. The THz absorption spectrum provides some information on proteins: for small molecules the THz spectrum is dominated by individual modes related to the presence of hydrogen bonds. For peptides, the spectral information concerns their secondary structure, while for complex proteins such as globular proteins and viral glycoproteins, spectra also provide information on collective modes. In this short review, we discuss the results obtained by THz spectroscopy in the protein dynamics investigations. In particular, we will illustrate advantages and applications of THz spectroscopy, pointing out the complementary information it may provide.


Introduction
Amino acids/proteins participate in the formation of cells and tissues, and they are responsible for life activities in living organisms, such as enzymes, antibodies, signaling and transporting molecules, cell membrane, metabolic or catalytic functions, etc. [1][2][3][4].

Figure 1.
THz region and applications in the protein field of THz radiation. The THz domain typically lies between microwave and infrared regions, which refers to the frequency band spanning 0.1-15 THz (from 3 cm −1 to 500 cm −1 ). The generation of THz radiation was successfully developed with both electronics and photonics, including the up-conversion of electronic radiofrequency sources and the down-conversion of optical sources. THz technology detects proteins, illustrated by their structure hierarchy.

Consolidate Technologies for Amino Acid/Protein Detection
There are two well-established methodological approaches to protein characterization: total protein nitrogen measurements and chemical and structural identification of a specific protein, schematically summarized in Figure 2. The determination of total protein nitrogen is a proximity tool for the early and rapid quantification of the protein content. The Kjeldahl [10] and Dumas methods are widely used for nutrition labeling and quality control [15][16][17]. Other techniques are used and preferred for their accuracy, low cost and great applicability in many fields where the identification of proteins is required, such as the bicinchoninic acid, biuret and Lowry methods as well as UV absorption. However, many of these are time-consuming and can only measure the total organic nitrogen present in the sample.
The most common immunoassays are enzyme-linked immunosorbent assays, lateral flow immunoassay and Western blotting, but other analytical technologies [16,[18][19][20][21][22][23][24][25] have been adopted in biomedical and clinical studies. Some analytical methods for monitoring and detection are based on the ability to propagate/amplify biomolecules in cell cultures or on the detection of a specific antibody. However, these laboratory diagnostic techniques require time to detect the specific immune response and its quantification. Especially, when used for virus recognition or metabolic evaluation the task is challenging, and some immunoassays may give false positive results. Other techniques, such as microscopy, fluorescence and spectroscopy, are preferred in describing the protein dynamics. Here we briefly recall the most used techniques: X-ray crystallography, nuclear Here, we review results obtained by THz spectroscopy and imaging in the study of amino acids and protein dynamics. In the first sections of this review, we discuss the experimental techniques commonly used for protein detection, their limitations and some critical issues. Then, we illustrate the advantages and applications of THz radiation providing complementary information to existing techniques. Finally, we discuss perspectives and other potential THz contributions in the field of biomolecules.

Consolidate Technologies for Amino Acid/Protein Detection
There are two well-established methodological approaches to protein characterization: total protein nitrogen measurements and chemical and structural identification of a specific protein, schematically summarized in Figure 2. The determination of total protein nitrogen is a proximity tool for the early and rapid quantification of the protein content. The Kjeldahl [10] and Dumas methods are widely used for nutrition labeling and quality control [15][16][17]. Other techniques are used and preferred for their accuracy, low cost and great applicability in many fields where the identification of proteins is required, such as the bicinchoninic acid, biuret and Lowry methods as well as UV absorption. However, many of these are time-consuming and can only measure the total organic nitrogen present in the sample.
The most common immunoassays are enzyme-linked immunosorbent assays, lateral flow immunoassay and Western blotting, but other analytical technologies [16,[18][19][20][21][22][23][24][25] have been adopted in biomedical and clinical studies. Some analytical methods for monitoring and detection are based on the ability to propagate/amplify biomolecules in cell cultures or on the detection of a specific antibody. However, these laboratory diagnostic techniques require time to detect the specific immune response and its quantification. Especially, when used for virus recognition or metabolic evaluation the task is challenging, and some immunoassays may give false positive results. Other techniques, such as microscopy, fluorescence and spectroscopy, are preferred in describing the protein dynamics. Here we briefly recall the most used techniques: X-ray crystallography, nuclear magnetic resolution (NMR), Cryo-electron microscopy (Cryo-EM), fluorescence, circular dichroism (CD) and vibrational spectroscopies, including infrared (IR) and Raman ones.
features of an evanescent field, probing the sample with the Attenuated Total Reflection (ATR) technique. ATR is frequently used for small amounts of solid and liquid samples, avoiding difficult preparation. Instead, time-resolved FTIR enables the observation of protein reaction at the sub-millisecond timescale, such as the study of the light-driven proton pump bacteriorhodopsin [53]. Nevertheless, linear Raman spectroscopy and microscopy are limited to weak signals and very long times for signal collection. Thus, only recently with the accessibility of ultrafast lasers, coherent Raman Scattering (CRS) techniques are going to assert themselves thanks to the sensitivity at the same molecular vibrations probed in spontaneous Raman spectroscopy and to the nonlinear dependence on the incoming light fields [30][31][32][33].
These systems have promoted the diffusion of THz radiation for spectroscopy applications. Thus, it was applied with success in various scientific fields e.g., gas sensing [75][76][77][78][79][80], chemical and pharmaceutical analysis [81,82], condensed matter [83][84][85], identification of crystalline polymorphs, microelectronics and security [86][87][88][89], agri-food industry [90] and cultural heritage [91], including the emerging support in the field of biomedicine and bioimaging [92][93][94][95][96][97][98][99], etc. Thus, THz spectroscopy has become a valuable tool for rapid and non-invasive detection thanks to many advantages that make it particularly appealing for probing the intermolecular structure and dynamics of biomolecules [95,100,101]. In fact, THz radiation lies between microwave and infrared, being characterized by low photon energy (4 meV @ 1 THz), and it provides energy levels related to rotational and vibrational molecular modes and intermolecular vibrations, such as hydrogen bonds [102,103]. For example, external lattice vibrations typically dominate the low-frequency region, and these modes are influenced by the crystalline arrangement of molecules. This occurs for different polymorphs where the unique THz spectral features can be used as identifying fingerprints [104]. THz waves are sensitive to the conformation and structure of proteins [105,106], and they can be used for providing useful information in combination with quantum-mechanical theory [101,107]. Moreover, the low photon energy is too low to heat materials or to induce atoms/molecules ionization. For this reason, it is a non-ionizing X-ray crystallography, NMR and Cryo-EM mainly concern investigations of the threedimensional structure of molecules. X-ray crystallography exploits the diffraction lattice for collecting an electron density map of the crystal. The need for a solid crystalline sample and the ionizing properties of X-rays represent limitations in the study of proteins in the physiological environment [39][40][41][42].
Most studies in organic chemistry involve the probing of hydrogen atoms with NMR [43][44][45]. The process of relaxation of nuclear magnetic spins from higher to lower energy levels after a certain time-span is the principle of NMR to achieve structural atomic information.
Cryo-EM provides a 3D structure of cells and macromolecules visualized with 5-8 nm spatial resolution. Although Cryo-EM allows high spatial resolutions, suitable for studying the cytoskeleton and the complex surface structures of viruses, the need of a frozen sample, although useful since it fixes the protein in a well-defined configuration, is not ideal to the investigate protein dynamics.
Fluorescence spectroscopy is an alternative method that works most accurately at very low concentrations of fluorophores, giving structural information about a single molecule or protein complex interactions. It exploits mainly the intrinsic and extrinsic protein fluorescence. It may contribute to reconstruct structural effects induced by pH and solvent composition, to measure lifetimes and temporal kinetics. Few proteins possess intrinsic fluorophores, such as tryptophan and tyrosine, but most of them are non-fluorescent. In many of these cases, an external fluorophore can be introduced by chemical coupling or non-covalent binding. This labelling technique allows the development of new fluorescent dyes with different excitation and emission properties. This has promoted the emergence of new techniques for the study of protein dynamics: quenching, fluorescence resonance energy transfer (FRET) [46,47], bioluminescence resonance energy transfer (BRET) [48], fluorescence recovery after photobleaching (FRAP) [49], fluorescence polarization and cross-correlation, and on all microscopy.
However, all spectroscopic/microscopic techniques have clear limitations. Indeed, requiring chemical labels, the biological functionalities can interfere with the protein dynamics; additionally, photo-bleaching can introduce artefacts and limit the measurement repeatability in the microscopic approach. Therefore, it may be desirable to implement real-time, three-dimensional imaging with high spatial resolution, high sensitivity and high chemical selectivity for unlabeled living cells.
Circular dichroism (CD) involves the measurements of small differences in the UV-vis absorption values due to the chirality of biomolecules. The left-and right-handed polarized components of the incident light are absorbed differently by the optically active molecules, which yields a difference in the absorption parameters [50][51][52]. CD spectroscopy is largely adopted to verify the natural secondary structure of proteins and their changes induced by chemical-physical conditions. Vibrational spectroscopy, including Fourier Transform Infrared (FTIR) and Raman spectroscopy, provides selective, non-destructive identification of the molecules, through the interrogation of molecular roto-vibrational and vibrational modes [53][54][55]. Both FTIR and Raman spectroscopy give similar information about a molecule: FTIR is associated with the change of dipole momentum and Raman with a change of the polarizability, respectively, as a consequence of the matter-light interaction. Their use is typically in chemical and biochemical research of small-molecule compounds and in the identification of synthesized compounds due to their intrinsic chemical selectivity. The nine characteristic vibrational bands (named amide A, B, I, II, III, . . . , VII, where amide I and II, between 1600 and 1700 cm −1 as well as 1500 and 1600 cm −1 respectively, are the major contributors in the IR spectrum) exhibited by the peptide bonds in the infrared region ensure the recognition of the protein backbone conformation and the assessment of the secondary structure of peptides and proteins [53]. Novel instruments combine the FTIR technique and the features of an evanescent field, probing the sample with the Attenuated Total Reflection (ATR) technique. ATR is frequently used for small amounts of solid and liquid samples, avoiding difficult preparation. Instead, time-resolved FTIR enables the observation of protein reaction at the sub-millisecond timescale, such as the study of the light-driven proton pump bacteriorhodopsin [53]. Nevertheless, linear Raman spectroscopy and microscopy are limited to weak signals and very long times for signal collection. Thus, only recently with the accessibility of ultrafast lasers, coherent Raman Scattering (CRS) techniques are going to assert themselves thanks to the sensitivity at the same molecular vibrations probed in spontaneous Raman spectroscopy and to the nonlinear dependence on the incoming light fields [30][31][32][33].
These systems have promoted the diffusion of THz radiation for spectroscopy applications. Thus, it was applied with success in various scientific fields e.g., gas sensing [75][76][77][78][79][80], chemical and pharmaceutical analysis [81,82], condensed matter [83][84][85], identification of crystalline polymorphs, microelectronics and security [86][87][88][89], agri-food industry [90] and cultural heritage [91], including the emerging support in the field of biomedicine and bio-imaging [92][93][94][95][96][97][98][99], etc. Thus, THz spectroscopy has become a valuable tool for rapid and non-invasive detection thanks to many advantages that make it particularly appealing for probing the intermolecular structure and dynamics of biomolecules [95,100,101]. In fact, THz radiation lies between microwave and infrared, being characterized by low photon energy (4 meV @ 1 THz), and it provides energy levels related to rotational and vibrational molecular modes and intermolecular vibrations, such as hydrogen bonds [102,103]. For example, external lattice vibrations typically dominate the low-frequency region, and these modes are influenced by the crystalline arrangement of molecules. This occurs for different polymorphs where the unique THz spectral features can be used as identifying fingerprints [104]. THz waves are sensitive to the conformation and structure of proteins [105,106], and they can be used for providing useful information in combination with quantum-mechanical theory [101,107]. Moreover, the low photon energy is too low to heat materials or to induce atoms/molecules ionization. For this reason, it is a non-ionizing radiation, suitable and attractive for noninvasive biological applications and biomedical imaging [97]. At variance, other biochemical/molecular techniques, such as ultraviolet or X-rays using higher-energy photons (from few eV to~keV), may damage any biological sample [81,101,108,109].
Furthermore, THz radiation strongly interacts with polar molecules [74,110,111], such as water. Especially in THz imaging, the intense absorption, due to water molecules [95], represents a limiting condition for the THz penetrability inside fresh tissues (from tens to hundreds of microns) [94], reducing the diagnostic capability in vivo only to superficial layers. On the contrary, nonpolar materials (such as paper, cloths and plastic) are usually transparent in the THz range [89,112,113].
Many layouts and materials can be used for THz signal collection in THz spectroscopy [114][115][116], exhibiting high performance in terms of the signal-to-noise ratio with coherent detection mode [116]. Because THz spectroscopy is insensitive to the thermal background, it shows a high signal-to-noise Ratio (SNR), and it does not require the use of cooled detectors [116,117]. Concerning the coherent detection mode, the THz electric field is directly assessed. Therefore, both amplitude and phase of the THz pulse electric field can be simultaneously measured, and the optical parameters, including sample absorption coefficient and refractive index, can be extracted without using Kramers-Kronig relations [101,116].

Current Status of THz Spectroscopy for Research on Amino Acids and Short-Chain Peptides
Amino acids are the backbones of proteins as well as fundamental elements to assemble cells and tissues. They are involved in all cellular metabolic and repairing processes and are responsible for various pathological genesis. The structural differences between amino acids depend on the side chain R and its polarity. Most THz biomolecule investigations have been focused on amino acids in the low-frequency spectral interval [118][119][120][121][122][123][124][125][126][127][128][129][130][131], see Table 1. All 20 standard α-amino acids under the same conditions revealed a correlation between molecular structures and spectral peaks between 0.2 and 3.0 THz (6.67-100 cm −1 ) [122]. The amino acids were studied in the solid phase, mixing them with a polyethylene powder in various proportions. For example, high proportions of amino acids to polyethylene (1:1 and 2:1) [123,124], as well as low proportions (1:15 and 1:9) [125,126], are used. From these, it emerged that the proportion of amino acids to THz transparent material in the production of the pellet influences the spectrum, especially in the low-frequency region [123,127]. Taday et al. [127] characterized the temperature-dependent L-glutamic acid spectrum and compared it with the theoretical model proposed by Chiba et al. [132], but poor agreement was achieved. Actually, theoretical simulations are not able to model the complexity of hydrogen bonding and crystalline structure occurring at low frequency in biological molecular systems.  [118][119][120][121][122][123][124][125][126][127][128][129][130][131][132][133][134].

Amino Acids
THz Absorption Frequencies (THz) the solute, solvent and water of hydration. The departure of the absorption coefficient from the ideal behavior, ∆α, is associated with the section of the hydrated water molecules around the amino acid molecules, which differ from the absorption of water in mass up to 5-7 Å from a single solute molecule. Interaction with water molecules is mainly due to -COO-and -NH 3+ . They found a negative value of ∆α for all amino acids, with the exception of Gly, meaning this is a structure breaker. Arg, Ser, Trp, Lys and Asp are instead structure makers, leading to the formation of an ordered water molecule network around single amino acids. Furthermore, the anomalous behavior of Gly was described and explained. Fitting real and imaginary parts of the solution dielectric constant with a multiple Debye model, three time scales were obtained (8.7 ps, 80 and 200 fs). They were associated with cooperative rearrangement of the hydrogen bonded network for 8.7 ps and to competitive quick jump of under-coordinated water and small angular rotation preceding a large angle jump for 200 fs. Finally, the contribution at 80 fs was due to hydrogen-bond bending and the related transverse acoustic phonons, which propagate in a direction normal to the hydrogen bonds in between two neighboring water molecules. The first two are found to depend on both hydrophobic and hydrophilic residues of the considered amino acid, and their trends were analyzed for all molecules. Low-frequency vibrational modes are strongly influenced by size and long-range order of the molecule. As consequence, small biological molecules, such as amino acids, tend to have distinct, relatively isolated features compared to polymers, such as short chain of peptides, with a larger number of modes. Thus, the THz spectral characteristics of the peptide reflect their amino acid composition, permutation sequence, intermolecular hydrogen bond and crystal structure. For example, the simplest molecule with a peptide bond, crystalline N-methyl acetamide, shows bands at 3.6 THz (120 cm −1 ) and 6.0 THz (200 cm −1 ) [159]. Yamamoto et al. used THz-TDS to study the absorption coefficient and refractive index of simple amino acids structures, glycine and L-alanine, compared with their corresponding polypeptides in the 0.21-1.65 THz (7-55 cm −1 ) domain [160]. Polyglycine was observed to have peaks at 1.365 THz (45.5 cm −1 ) induced by the interaction of the chain different from the poly-L-alanine. The differences indicate that poly-glycine has a longer-range, ordered structure of solid-state crystals than poly-L-alanine does.
Kutteruf et al. [107] also demonstrated the great potential of THz spectroscopy for solid-phase short-chain peptide sequences, in the spectral region 1-15 THz. Pure solid samples of low molecular weight protein fragments pressed into polyethylene pellets were studied at 77 K and 298 K. Highly structured THz absorption spectra were typically observed with sharp spectral features determined by molecular symmetries and structure. In the same work, authors proved how quantum mechanical calculations, applied for the isolated diglycine species, can match experimental results if the description includes solidstate hydrogen-bonding and intermolecular interactions.
Likewise, Neu et al. [161] studied six polypeptide chains with very similar primary and secondary structures in the low-frequency region up to 2.5 THz. The result modeled with DFT calculations was used to calculate the dynamic motions of several peptides and to visualize the corresponding displacements of the strongest resonances.
Notably, the difference is clear between amino acids and polypeptides existing in the THz frequency dynamics range. The physical quantities show a different dependence by the length of the peptide chain, indicating that they reflect different interactions.  [118,119,120,121,122,123,124,125,126,127,128,129,130,131,132,133,134].

Current Status of THz Spectroscopy for Protein Research
As mentioned above, it is clear that molecular conformation affects the biological activity and functionality of proteins [162], so it is strategic to characterize the native protein conformation. With the advance in THz technologies, research on proteins and their conformation in the THz spectral region has been extensively carried out [163,164,165,166]. All findings demonstrated that the collective vibration mode related to the protein molecular conformation is optically active in the THz range [167,168]. Markelz and co-authors collected the THz spectrum of lyophilized powder bovine serum albumin (BSA) of pressed pellets in the range between 0.06 and 2.00 THz (2-66.6 cm −1 ). Paciaroni et al. focused the attention on vibrational collective dynamics of a dry perdeuterated maltose-binding protein in the THz domain [169]. Sun et al. used THz spectroscopy to study the binding of hemagglutinin protein and broadly neutralizing monoclonal antibodies in the liquid environment [170].
THz spectroscopy is, among others, a convenient method for probing conformational changes in photoactive proteins [171,172,173]. Castro-Camus and Johnston [171] investigated conformational changes in photoactive yellow protein (PYP) suspended in a physiological buffer solution. The sample was uniformly illuminated by high-intensity blue (450 nm) light emitting diodes to trigger a conformational change from the ground state of the photoactive yellow protein (called pG) to the photo-intermediate (pB). Changes in the protein structure were probed by THz radiation, which is far from triggering any modifications in the protein itself. Absorption differences were observed, such as an increase in the absorption of the pB configuration compared to that in the pG state in the range 0.25-2 THz (8-66.6 cm −1 ).
An interesting study was conducted by Han et al. using both THz spectroscopy and imaging [174]. These attempted to identify proteins and quantify their content. The frequency-dependent refractive index of six proteins (rn21, rn22, rn28, n42, n43 and n53 with concentration of 2.0 μg/μL and quantity of 8 μg) were initially evaluated with THz-TDS in the spectral region 0.1-3.5 THz (3.33-117 cm −1 ). In Figure 3, the refractive index of these six proteins is reported in the sub-frequency band.

Current Status of THz Spectroscopy for Protein Research
As mentioned above, it is clear that molecular conformation affects the biological activity and functionality of proteins [162], so it is strategic to characterize the native protein conformation. With the advance in THz technologies, research on proteins and their conformation in the THz spectral region has been extensively carried out [163,164,165,166]. All findings demonstrated that the collective vibration mode related to the protein molecular conformation is optically active in the THz range [167,168]. Markelz and co-authors collected the THz spectrum of lyophilized powder bovine serum albumin (BSA) of pressed pellets in the range between 0.06 and 2.00 THz (2-66.6 cm −1 ). Paciaroni et al. focused the attention on vibrational collective dynamics of a dry perdeuterated maltose-binding protein in the THz domain [169]. Sun et al. used THz spectroscopy to study the binding of hemagglutinin protein and broadly neutralizing monoclonal antibodies in the liquid environment [170].
THz spectroscopy is, among others, a convenient method for probing conformational changes in photoactive proteins [171,172,173]. Castro-Camus and Johnston [171] investigated conformational changes in photoactive yellow protein (PYP) suspended in a physiological buffer solution. The sample was uniformly illuminated by high-intensity blue (450 nm) light emitting diodes to trigger a conformational change from the ground state of the photoactive yellow protein (called pG) to the photo-intermediate (pB). Changes in the protein structure were probed by THz radiation, which is far from triggering any modifications in the protein itself. Absorption differences were observed, such as an increase in the absorption of the pB configuration compared to that in the pG state in the range 0.25-2 THz (8-66.6 cm −1 ).
An interesting study was conducted by Han et al. using both THz spectroscopy and imaging [174]. These attempted to identify proteins and quantify their content. The frequency-dependent refractive index of six proteins (rn21, rn22, rn28, n42, n43 and n53 with concentration of 2.0 μg/μL and quantity of 8 μg) were initially evaluated with THz-TDS in the spectral region 0.1-3.5 THz (3.33-117 cm −1 ). In Figure 3, the refractive index of these six proteins is reported in the sub-frequency band.

Current Status of THz Spectroscopy for Protein Research
As mentioned above, it is clear that molecular conformation affects the biological activity and functionality of proteins [162], so it is strategic to characterize the native protein conformation. With the advance in THz technologies, research on proteins and their conformation in the THz spectral region has been extensively carried out [163,164,165,166]. All findings demonstrated that the collective vibration mode related to the protein molecular conformation is optically active in the THz range [167,168]. Markelz and co-authors collected the THz spectrum of lyophilized powder bovine serum albumin (BSA) of pressed pellets in the range between 0.06 and 2.00 THz (2-66.6 cm −1 ). Paciaroni et al. focused the attention on vibrational collective dynamics of a dry perdeuterated maltose-binding protein in the THz domain [169]. Sun et al. used THz spectroscopy to study the binding of hemagglutinin protein and broadly neutralizing monoclonal antibodies in the liquid environment [170].
THz spectroscopy is, among others, a convenient method for probing conformational changes in photoactive proteins [171,172,173]. Castro-Camus and Johnston [171] investigated conformational changes in photoactive yellow protein (PYP) suspended in a physiological buffer solution. The sample was uniformly illuminated by high-intensity blue (450 nm) light emitting diodes to trigger a conformational change from the ground state of the photoactive yellow protein (called pG) to the photo-intermediate (pB). Changes in the protein structure were probed by THz radiation, which is far from triggering any modifications in the protein itself. Absorption differences were observed, such as an increase in the absorption of the pB configuration compared to that in the pG state in the range 0.25-2 THz (8-66.6 cm −1 ).
An interesting study was conducted by Han et al. using both THz spectroscopy and imaging [174]. These attempted to identify proteins and quantify their content. The frequency-dependent refractive index of six proteins (rn21, rn22, rn28, n42, n43 and n53 with concentration of 2.0 μg/μL and quantity of 8 μg) were initially evaluated with THz-TDS in the spectral region 0.1-3.5 THz (3.33-117 cm −1 ). In Figure 3, the refractive index of these six proteins is reported in the sub-frequency band. Yi et al. [133] employed THz-TDS and IR spectroscopy to study all 20 amino acids, accurately collecting their spectra over a wide range from 0.2 to 6 THz (6.67 to 200 cm −1 ). Mixtures in different proportions of amino acids with polyethylene powder, transparent to THz radiation, were made. Their absorbance spectra were characterized to evaluate the optimal proportion for each amino acid. Among the fifteen aliphatic amino acids, Lglycine (L-gly), L-alanine (L-ala), L-valine (L-val), L-isoleucine (L-iso) and L-leucine (L-leu) have a similar structure, the only structural difference being the side chain. Nevertheless, they exhibit differences in the THz absorption spectra: L-val has prominent absorption at 1.11 THz and 4.39 THz; for L-iso, the main absorption peaks were shifted to lower frequencies at 0.85 and 4.27 THz; and finally for L-leu, at 0.66 and 3.68 THz. The authors correlated the differences due to the longer side chains that shift the absorption peaks to lower frequencies [133]. Moreover, the aromatic L-Tyrosine (L-tyr), L-phenylalanine (L-phe) and L-tryptophan (L-try) present a phenyl group on the side chain, and although they have similar structures, the THz spectra are different. A broad absorption band at 3.22 THz is observed in the L-try spectrum, and three absorption bands at 2.02, 2.76 and 4.16 THz appear in the L-phe absorption spectrum. L-tyr includes two symmetric peaks around them, with splitting attributable to the -OH group of L-tyr. For L-ala and L-threonine (L-thr) the main absorption peaks are located at 2.91 THz and 3.33 THz, respectively. The accurate study by Yi et al. is in a good agreement with previous works [118][119][120][121][122][123][124][125][127][128][129][130][131] ranging between 0.1 and 3 THz. Thus, they established an absorption spectral database of protein constituents in the solid phase and demonstrated that THz spectroscopy is able to distinguish different spectral features of amino acids with a very similar structure.
Another interesting aspect regards enantiomers, chiral molecules which are nonsuperimposable mirror images of each other, isotopologues and polymorphism [134]. In this context, THz spectroscopy is sensitive to crystalline structures. In pharmaceutical and pharmacological science, the use of amino acids and small peptide chains is required; the knowledge of their crystallinity is a topic issue from studies of changes in medicinal aging and to the detection of illegal substances. In particular, drugs bind to the human body thanks to receptors which are protein complexes [135]. In addition, the THz spectra of DL-leu and the two polymorphs of DL-val have been measured, and computer calculations have been implemented to reliably distinguish the nuances of THz spectra similar to solid-state systems [136,137]. Yamaguchi et al. [92] demonstrated THz sensitivity in enantiomers recognition, showing how the THz absorption spectra of L-(D-) and DL-ala are quite different; L-and D-ala show two absorption bands located at 74.4 and 85.7 cm −1 [92], and racemic compound DL-ala exhibits one absorption band at 41.8 cm −1 . Therefore, THz spectroscopy is considered an off-line tool for assessing crystallinity, also in co-lyophilized amino acid/gelatin mixtures (L-ala, serine and valine) [128]. The physical/chemical conditions are critical for amino acids, and in particular, amino acid medicine is strongly susceptible to the storage and transportation states. Herein, the induced changes affect several properties of amino acids, e.g., stability, solubility, mechanical and physic-chemical properties, and alter the protein dynamics. Several experimental as well as theoretical studies carried out addressing those issues [138][139][140][141][142][143]. Aqueous solutions of amino acids have been extensively studied through different experimental techniques [144][145][146][147][148][149] and molecular dynamics simulations [150]. On the other side, there have been a few dielectric relaxation (DR) studies, covering MHz-GHz frequency domain, to probe the behavior of hydrated amino acids [151][152][153].
One of the amino acids extremely susceptible to change is Lysine (Lys), which has a key role in human metabolism and various physiological functions, such as promotion of human growth and enhancement of immunity. With the addition of water molecules, the surrounding environment of Lys is changed, affecting its collective molecular vibrations. Exploring the molecular collective modes at low frequency, Bian et al. [154] found that Lys and Lys·H 2 O have distinct spectral features in the range between 0.3 and 2.5 THz. The interpretation of the results was entrusted by theoretical calculations based on Density Function Theory (DFT) [155,156]; the water molecule is recognized to locate the carboxyl group and the amino group of hydrated Lys molecule, while its vibrational modes are produced by a dihedral torsion or bond angle bend of molecular chains. An equivalent description was also achieved for differences shown in THz spectral features of L-phe and for L-phe·H 2 O [157], associating them to different crystalline structures and to the presence of hydrogen bond interaction in the case of L-phe·H 2 O.
The hydration process of amino acids is the basis for a further understanding of the behavior of proteins in the aqueous environment. Samanta et al. [158] studied five amino acids dissolved in phosphate buffers. The optical absorption parameters dependent on the frequency, coefficient and refractive index were distinguished from the contribution of the solute, solvent and water of hydration. The departure of the absorption coefficient from the ideal behavior, ∆α, is associated with the section of the hydrated water molecules around the amino acid molecules, which differ from the absorption of water in mass up to 5-7 Å from a single solute molecule. Interaction with water molecules is mainly due to -COO-and -NH 3+ . They found a negative value of ∆α for all amino acids, with the exception of Gly, meaning this is a structure breaker. Arg, Ser, Trp, Lys and Asp are instead structure makers, leading to the formation of an ordered water molecule network around single amino acids. Furthermore, the anomalous behavior of Gly was described and explained. Fitting real and imaginary parts of the solution dielectric constant with a multiple Debye model, three time scales were obtained (8.7 ps, 80 and 200 fs). They were associated with cooperative rearrangement of the hydrogen bonded network for 8.7 ps and to competitive quick jump of under-coordinated water and small angular rotation preceding a large angle jump for 200 fs. Finally, the contribution at 80 fs was due to hydrogen-bond bending and the related transverse acoustic phonons, which propagate in a direction normal to the hydrogen bonds in between two neighboring water molecules. The first two are found to depend on both hydrophobic and hydrophilic residues of the considered amino acid, and their trends were analyzed for all molecules.
Low-frequency vibrational modes are strongly influenced by size and long-range order of the molecule. As consequence, small biological molecules, such as amino acids, tend to have distinct, relatively isolated features compared to polymers, such as short chain of peptides, with a larger number of modes. Thus, the THz spectral characteristics of the peptide reflect their amino acid composition, permutation sequence, intermolecular hydrogen bond and crystal structure. For example, the simplest molecule with a peptide bond, crystalline N-methyl acetamide, shows bands at 3.6 THz (120 cm −1 ) and 6.0 THz (200 cm −1 ) [159]. Yamamoto et al. used THz-TDS to study the absorption coefficient and refractive index of simple amino acids structures, glycine and L-alanine, compared with their corresponding polypeptides in the 0.21-1.65 THz (7-55 cm −1 ) domain [160]. Polyglycine was observed to have peaks at 1.365 THz (45.5 cm −1 ) induced by the interaction of the chain different from the poly-L-alanine. The differences indicate that poly-glycine has a longer-range, ordered structure of solid-state crystals than poly-L-alanine does.
Kutteruf et al. [107] also demonstrated the great potential of THz spectroscopy for solid-phase short-chain peptide sequences, in the spectral region 1-15 THz. Pure solid samples of low molecular weight protein fragments pressed into polyethylene pellets were studied at 77 K and 298 K. Highly structured THz absorption spectra were typically observed with sharp spectral features determined by molecular symmetries and structure. In the same work, authors proved how quantum mechanical calculations, applied for the isolated diglycine species, can match experimental results if the description includes solid-state hydrogen-bonding and intermolecular interactions.
Likewise, Neu et al. [161] studied six polypeptide chains with very similar primary and secondary structures in the low-frequency region up to 2.5 THz. The result modeled with DFT calculations was used to calculate the dynamic motions of several peptides and to visualize the corresponding displacements of the strongest resonances.
Notably, the difference is clear between amino acids and polypeptides existing in the THz frequency dynamics range. The physical quantities show a different dependence by the length of the peptide chain, indicating that they reflect different interactions.

Current Status of THz Spectroscopy for Protein Research
As mentioned above, it is clear that molecular conformation affects the biological activity and functionality of proteins [162], so it is strategic to characterize the native protein conformation. With the advance in THz technologies, research on proteins and their conformation in the THz spectral region has been extensively carried out [163][164][165][166]. All findings demonstrated that the collective vibration mode related to the protein molecular conformation is optically active in the THz range [167,168]. Markelz and co-authors collected the THz spectrum of lyophilized powder bovine serum albumin (BSA) of pressed pellets in the range between 0.06 and 2.00 THz (2-66.6 cm −1 ). Paciaroni et al. focused the attention on vibrational collective dynamics of a dry perdeuterated maltose-binding protein in the THz domain [169]. Sun et al. used THz spectroscopy to study the binding of hemagglutinin protein and broadly neutralizing monoclonal antibodies in the liquid environment [170].
THz spectroscopy is, among others, a convenient method for probing conformational changes in photoactive proteins [171][172][173]. Castro-Camus and Johnston [171] investigated conformational changes in photoactive yellow protein (PYP) suspended in a physiological buffer solution. The sample was uniformly illuminated by high-intensity blue (450 nm) light emitting diodes to trigger a conformational change from the ground state of the photoactive yellow protein (called pG) to the photo-intermediate (pB). Changes in the protein structure were probed by THz radiation, which is far from triggering any modifications in the protein itself. Absorption differences were observed, such as an increase in the absorption of the pB configuration compared to that in the pG state in the range 0.25-2 THz (8-66.6 cm −1 ).
An interesting study was conducted by Han et al. using both THz spectroscopy and imaging [174]. These attempted to identify proteins and quantify their content. The frequency-dependent refractive index of six proteins (rn21, rn22, rn28, n42, n43 and n53 with concentration of 2.0 µg/µL and quantity of 8 µg) were initially evaluated with THz-TDS in the spectral region 0.1-3.5 THz (3.33-117 cm −1 ). In Figure 3, the refractive index of these six proteins is reported in the sub-frequency band. The refractive index vs. frequencies in the THz range of these six different species of proteins are different and well-separated, allowing to distinguish each protein. Although the error bars overlap, the curve trend points out that the protein with a higher molecular mass has a stronger refraction capacity, namely a greater refractive index at every frequency. Authors have investigated the ability of THz imaging to differentiate and quantify the content of different kinds of specific proteins. The intensity or the level of shade of the image is directly related to the number of proteins participating in the absorption, see Figure 4. For this purpose, a nitrocellulose blotting membrane (NC) was dropped with liquid solutions of the six proteins, formulated inside the five solutions with corresponding concentrations of 2 µg/µL, 1 µg/µL, 0.5 µg/µL, 0.25 µg/µL and 0.125 µg/µL, dissolved in solution of 1 × PBS at pH 7.2. Observing Figure 4b, a variation of the average colorscale of proteins rn21, rn22 and n43 was evident; firstly, it increases at 0.5 µg, and then it decreases at 1 µg, 2 µg, 4 µg and 8 µg as a function of protein quantity (except for the 8 µg of Protein n43). At variance, the average color scales of proteins rn28 and n42 exhibit the opposite trend (see Figure 4b). For the six kinds of proteins with the same amount, the intensity of gray levels behaves similarly to the refractive index absorption. The THz experimental results of Han et al. match exactly the results of the dot blot (see Figure 4a) method. Besides, the application of THz spectroscopy does not merely distinguish the different proteins, it precisely acquires the protein quantity distribution, but it is able to reduce the time procedure compared to conventional immunoassays in the label-free and non-destructive mode. The structural flexibility of the protein allows to skillfully change its conformation without modifying the function it has to perform [175]. The frequency dependence of the density of the vibrational modes in the biological temperature range can quantify flexibility [163]. One example involves the transfer of cytochrome c (CytC) protein electrons. The CytC constituent amino acids have a covalently linked heme group with two oxidation states, ferri-and ferro-CytC. Ferri-CytC is less thermally stable than ferro-CytC and has also a higher hydrogen exchange and proteolytic digestion rate [169]. The THz spectroscopy was revealed as a powerful investigation tool because the collective vibrational modes fall into the THz domain, ranging from 0.03 to 6 THz (1-200 cm −1 ) [176]. Chen and coworkers [175] examined the complex THz dielectric response of CytC films in the two oxidation states and found that absorption and refractive indexes tend to increase with the degree of oxidation. A sharp increase in the density of the vibrational modes and/or the dipole coupling in the interval is observed in the range 0.15-2.4 THz (5-80 cm −1 ).
The influence of various physical and/or chemical factors [177][178][179][180][181][182] can induce protein denaturation with a relative change of the internal structure and properties of protein molecules. The structural change induced in the protein consequently alters its functionality, a delicate aspect in food processing, protein purification, nutrition, biomedicine and the food industry. Several examples underlined the advantage of THz technology in the study of protein denaturation and conformational changes. Studies with native sperm-whale myoglobin indicated that loss of secondary and tertiary structures influences the far-IR spectra as spectral changes in the region between 11.1 and 15.6 THz (370-520 cm −1 ) along with broader changes around 6.0 THz (200 cm −1 ) [183]. Firstly, the denaturation of the PsbO protein [184] and two photosynthesis chlorophyll proteins CP43 and CP47 [185][186][187] were examined though THz-TDS. Recently, a consistent, strong increase in the <3 THz (100 cm −1 ) absorbance was shown in the measurements of non-native hen egg white lysozyme (HEWL) [188] and human serum albumin (HSA) [189]. Yoneyama et al. [162], using THz spectroscopy, measured the thermal denaturation of BSA protein held in a membrane device, observing a higher THz transmittance of the thermal denatured BSA sample compared to that of the native-conformation sample. Temperature-dependent THz analysis was also led by George et al. [190] on frozen-solution-phase samples of hen HEWL and CytC proteins, in order to study temperature-dependent conformation of proteins. The estimation of the imaginary part of the dielectric function as a function of T and fitting with Arrhenius model activation energies explained the behavior of HEWL being free, bonded to 3NAG and denatured; CytC was oxidized, reduced and denatured. The thermal denaturation of HSA in an aqueous buffer solution revealed changes in the absorption coefficient and refractive index as a function of the temperature [189].
Many of these experiments proved the feasibility of both protein and denaturation detection, but on the other hand they drew attention to some problems such as low signalto-noise ratio and poor sensitivity, especially in the liquid environment because of water absorption. To overcome these limitations, different approaches can be adopted to enhance the THz response of biological samples based on graphene [191], on microfluidic chips [192,193], novel materials such as meta-and nano-materials [109,[194][195][196], on gold nanoparticles [197] and on nanoantennas [198].
A promising further step forward in development of THz technology was given recently by Zhang's group [199]. They utilized the reflective THz time-domain polarization spectroscopy (THz-TDPS) method for protein sensing in the liquid environment in the spectral region 0.1-2.5 THz (3.33-83.3 cm −1 ). The system, reported in Figure 5a,b, allows to obtain a polarization sensing, measuring the reflective polarization spectra for a liquidphase sample. To this aim, a traditional transmission THz-TDS [97,98] was modified, adding a reflection module (Figure 5b) to the bottom of the 3D printing sample cell and two rotatable THz polarizers at the emitting end and detection ports of the spectroscopic system (Figure 5a). The advantage is to measure the states of signals without the strong THz water absorption in reflection mode. In addition, authors used a flexible twisted dual-layer metasurface structure with geometric chirality (Figure 5c-e), as a sensor, to enhance the polarization response of three protein samples: BSA, whey protein (WP) and ovalbumin (OVA). Generally, circular polarization (CP) sensing shows an improvement in the sensitivity and more information sensing, if compared with the traditional linear polarization (LP) spectra. The influence of a double-layer metasurface sensor is proven to enhance also the polarization response of the sample. In fact, nature-conformation proteins are clearly distinguishable, and denaturation induced by temperature treatments can be detected through variations of CP reflection spectra, both right-(RCP) and left-handed (LCP). Compared with the traditional LP spectra, the CP sensing sensitivity is improved. The detection sensitivity achieved with thermal denaturation measurements is established to be Sd = 6.30 dB%, but THz-TDPS reaches a detection sensitivity of Sc = 52.9 dB mL/g for concentration estimations, reported in Figures 6 and 7.   [199].
Proteins require an environment as close as possible to the living environment, i.e., an aqueous solution. The role of water in protein dynamics has been, and still is, debated. Although water is a limiting factor in THz spectroscopy due to strong absorption in the spectral region of interest, this does not limit protein hydration studies. Combining THz spectroscopy and molecular simulations, Meister's group [200] looked at the behavior of long-range protein-water dynamics in hyperactive insect antifreeze proteins. THz measurements are highly dependent on relative humidity, as protein films have a high affinity for adsorbed water [201]. Heyden and Havenith [202] engaged in both experimental and theoretical study of protein-hydration coupling. Xu et al. [203] measured the absorption spectrum of solvated BSA between 0.3 and 3.72 THz (10-124 cm −1 ) in order to monitor its collective vibrational dynamics and obtain information in the low-frequency region. They successfully estimated the THz molar absorption of solvated BSA from the much stronger attenuation of water. They deduced that the vibrational modes of solvated proteins lead to a dense and overlapping spectrum monotonously increasing with frequency. The lack of a distinct and spectrally structured spectrum suggested the lack of a specific dominant collective vibration foreseen by molecular dynamics simulations and normal mode analysis of a series of small proteins.
Furthermore, the study of the protein-water interaction can be reported at the cellular level, which turns out to be highly fashionable. In fact, membrane proteins are those that in this context are most involved in the interaction with water, with possible changes in the structure and the ability to create a network of hydrogen bonds, which can be explored from 1 to 6 THz. Pal and Chattopadhyay [204] probably provided the first and accurate overview of the potential of THz spectroscopy in this area. They demonstrated how THz-TDS and some optical parameters (e.g., the dielectric constant) can describe specific changes in the membrane microenvironment and lipid concentration. The references listed since up to now only emphasize the potential of THz radiation in the field of protein biochemistry. Specifically, the ability of THz spectroscopy to detect the presence of types of proteins, to quantify their content and to be sensitive to protein conformation paves the way for its use in the biomedical and bioclinical fields. Many pathologies are in fact due to an alteration of the aminoacetic sequences and/or functional protein alterations. Recently, Wang et al. [205] exploited time-resolved THz spectroscopy to analyze serum and cerebrospinal fluid (CSF) extracted from rats at different times after blast-induced head injury (bTBI), from both the hypothalamus and the hippocampus.
THz spectra change as a function of time, with different trends for proteins of the hypothalamus and hippocampus. Specifically, Figure 8a,b shows the results of Wang et al. [205], relating to absorption spectra and refractive index between 0.2 and 2.0 THz (6.66-66.6 cm −1 ) of the total protein in the hypothalamus. The panels showed a magnified area in the range 1-1.6 THz (33.3-53.3 cm −1 ). Overall, the absorption coefficient and refractive index of total proteins in the hypothalamus rose and fell monotonously as a function of frequency, respectively. To describe the temporal evolution in a clear way, normalized values of optical parameters were selected at 1.6 THz, in Figure 8c. Differences were observed just at 3 h from blast exposure; the absorption coefficients relative to mild and moderate insults were significantly higher than those in the sham group. Subsequently, increasing the time (6 h and 24 h), these values decreased gradually, returning to the normal level, although differences persist between the two parameters also 24 h after the blast exposure. In the same way, THz spectra of the total protein in the hippocampus and its temporal changes are evaluated and are shown in Figure 9a,c. Both the THz absorption and refractive index showed no significant changes 3 h after blast exposure, unlike what happens for the hypothalamus. However, after an additional 3 h, the THz absorption coefficient of the total protein in the hippocampus increases accompanied by a decrease in the refractive index THz, with the increase in the traumatic degree. These changes were more noticeable with the increase in traumatic degree. The bTBI caused some changes in both the hypothalamus and hippocampus total proteins of rats, a behavior associated with the decrease or the increase in the number of some biomolecules with effects on the functioning of brain areas and symptoms of neurological damage. THz technology then appears as a powerful tool for early recognition and diagnosis of bTBI.

Conclusions
Proteins play a fundamental role in biology and, in particular, in living systems. The study of the dynamics of proteins provides information on their functionality, catalysis and potential alteration towards pathological diseases and, therefore, are of interest for various research fields.
Various techniques are currently used for their quantification, identification and evolutionary study; still, the demand of new techniques to complement the existing ones is growing. Although various techniques for protein investigation are currently used in the industrial, pharmaceutical and diagnostic fields, limitations, complexity and cost represent serious drawbacks. In this framework, emerging spectroscopic analytical methods, based on the use of THz radiation in the range 0.1-15 THz, are now competitive methods in the biochemical community. Thanks to new technology, the potential of THz spectroscopy has been affirmed in the analysis of both simple structures, such as polyamide molecules, and complex structures, such as protein complexes.
This short review highlights the possibility of identifying amino acids and proteins and revealing protein dynamics. In the first part we recalled the most common and used techniques from those for the evaluation of protein concentration, the differentiation of proteins and the study of conformational dynamics and modifications. In the second part we focused on THz radiation opportunities outlining issues and applications in which it offers real advantages.
THz spectroscopy provides a unique perspective on the chemical structure, rotational and vibrational molecule modes and intermolecular vibrations, such as hydrogen bonds [102,103]. Furthermore, it is sensitive to the crystalline structure, therefore able to distinguish enantiomers, isotopologists and polymorphisms [134]. Most amino acids have a chiral shape, and THz spectroscopy supports their detection, being particularly useful in the field of pharmaceuticals, where it is able to clearly differentiate between the amino acids and the polypeptide and to monitor protein-ligand interactions.
Protein dynamics analysis shows great potentialities, although there are many studies on the spectral change associated with spontaneous and physico-chemical-induced conformational changes. Some caution must be considered in interpretation of these results. The use of chemical simulations and theoretical approaches based on DFT calculations have limitations, but learning algorithms, neural models and deep learning approaches are viable future approaches, as showed by Sun et al., 2018 [206], which combine machine learning and THz analysis for quantitative protein analysis of BSA, deposited at different concentrations on thin films. The information content of the THz absorption spectra was analyzed with principal component analysis (PCA), spectrum regression analysis (SVR) and maximal information coefficient (MIC) to discriminate frequencies, and machine learning methods proved efficient for the recognition of spectral features. Although different applications of this technology showed great potential, scientific and technological issues still need to be addressed, such as the methodological approach to the analysis of samples in aqueous solutions and the need to increase the sensitivity of the THz signal in the presence of weakly active THz materials.