Effects of solvent environment on the spectroscopic properties of tylosin, an experimental and theoretical approach

Antibiotic resistance has emerged as a topic of concern for the scientific community. A particular area of concern is the use of tylosin to maintain livestock health, as up to 80% of the orally administered dosage can be excreted back into the environment by the animal (Stromer et al. 2019). Once released into the environment natural processes can transport tylosin through the ecosystem. For example, Dinh et al. (2011) detected a concentration of 2.8 ng L⁻¹ of tylosin in the Marne River. The authors attributed the presence of tylosin in the river to either the incomplete removal of antibiotics by treatment plants or runoff contamination from nearby agricultural land. Moreover, out of the 23 contaminants tested by the authors, tylosin was one of only three present in the river at a high enough concentration to be detected. Likewise, Kim et al. (2012) reported a concentration of tylosin in agricultural soil bordering animal feeding operations at 19.6 μg kg⁻¹. The presence of tylosin in the environment has made it necessary to develop treatment methods that can remove it from the ecosystem. One such method is the usage of light-based technologies.

The use of UV irradiation to sterilize an environment is a well-known concept. When exposed to UV light a molecule can undergo photodegradation, breaking down into simpler molecules, with aromatic compounds especially vulnerable to this decay pathway (Wicks et al. 2007). Using light to break down molecules has several advantages. First, since the process uses light as the instrument of degradation, it is easy to use and can be initiated at room temperature. Second, in theory, light-induced reactions can achieve high rates of conversion as it is easy for irradiated photons to reach most of the sample molecules. Third, light irradiation is most efficient for flat surfaces, such as water. However, this technique has several notable disadvantages, primarily the requirement of specialized UV lamps to achieve optimal results. Indeed, work by Voigt & Jaeger (2017) showed that tylosin could be eliminated via photodegradation in 4 min at pH 3 and pH 7 and 6 min at pH 9 when irradiated by 4.88 eV (254 nm) light at a photon fluence rate of 2 mmol min⁻¹ L⁻¹. Therefore, it is possible to efficiently remove antibiotics from water using high-energy UV light. However, UV lamps are energy-intensive, require special safety requirements to operate as many utilize mercury, and require frequent replacement to maintain efficiency (Shin et al. 2016). Also, the penetration depth of a photon is inversely proportional to its wavelength; that is, a high-energy, short-wavelength UV photon will not be able to penetrate as deep as a longer, lower energy photon (Saleh & Teich 2007; Wicks et al. 2007). Many of the issues of cost and safety concerns could be overcome if sunlight could be harvested as the light source used to degrade antibiotics in wastewater. However, the spectral power density of sunlight falls considerably above 3.1 eV (Pettit 1940). Moreover, the spectrum of sunlight varies with the seasons, with the northern hemisphere exposed to a higher amount of UV irradiation around the summer solstice than the winter solstice (Forsythe & Christison 1930).

While using light to degrade unwanted molecules is a well-established technique, it comes with the caveat that a molecule's UV absorption spectra is dependent on the surrounding environment. A molecule's absorbance may blue shift to a higher energy, red shift to a lower energy, or be heavily suppressed based on the surrounding solvent molecules. In simple terms, a red shift would occur when the solvent molecules act to stabilize the excited state potential energy surface (PES). This may occur if the excited state is more polar than the ground state and is stabilized by a polar solvent (Kelley 2012). The molecule's environment can also influence the lifetime of the excited state. For example, while acting to stabilize the PES of a molecule a solvent can impact the Franck–Condon overlap between the vibrational wavefunctions (Kelley 2012). Because the surrounding molecular environment impacts a molecule's UV spectra, it would, in turn, impact the effectiveness of lamps and photocatalysts used to degrade antibiotics by shifting the absorption spectra of a molecule and reducing the lamp's effectiveness. Therefore, understanding how the spectrum of antibiotic contaminants such as tylosin is impacted under various solvent conditions is imperative in the successful design of effective photodegradation mechanisms.

Porphyrins are one class of molecules that can serve as photocatalysts to initiate the photodegradation of antibiotics, such as tylosin, under lower energy conditions such as sunlight due to the presence of the Soret band, which is a strong absorptive region the molecule possesses in the near UV region. While the spectra of porphyrins have been studied since the 1950s, with the ‘Four Orbital’ model explaining the behavior of porphyrin spectra, few spectroscopic studies have been conducted on tylosin to analyze its behavior in different solvent conditions (Gouterman 1959). By examining the UV spectra of tylosin, we can approximate its excited state lifetime and infer the molecule's energetic behavior when exposed to different environments. Moreover, it is possible to infer the molecule's structure and PES using computational tools. As a molecule's photoreactivity is determined by how it travels along its PES, understanding it is critical to designing proper photocatalysts (Feist et al. 2017).

Tylosin A was dissolved in six different solutions to measure its spectral properties. Pharmaceutical standard grade Tylosin was sourced from Sigma-Aldrich and prepared in pH 2.2 deionized (DI) water (15 MΩ) obtained via a Purelab flex purification system to create a 0.0970 mM solution. The acidic DI water was created by titrating 5 mL of acetic acid (Fisher Scientific, 99.5%) with 40 mL DI water. Next, we prepared a neutral DI water solution by dissolving tylosin in a water sample with pH 7.44 at a concentration of 0.0917 mM. A basic solution with pH 11.62 was prepared by titrating 1 mL NaOH (Sigma-Aldrich, 1 N) with 49 mL DI water, and tylosin was added to give a final concentration of 0.0939 mM. A sample of tylosin was then dissolved in toluene to give a final concentration of 0.0982 mM to simulate a non-polar environment. Samples in acetone and acetonitrile were created at a concentration of 0.0917 mM tylosin. Finally, a sample of tylosin was dissolved in 2-propanol to give a final concentration of 0.0939 mM. The toluene was purchased from Sigma-Aldrich, while the acetone, acetonitrile, and 2-propanol was purchased from Fisher Scientific, the solutions were at >99.5% purity. The pH measurements were taken by the Fisher Scientific Orion pH meter.

The steady-state UV–Vis spectra of the created solutions were measured at room temperature (22.5 °C) using a SpectraMax M5 monochromator and a 1 cm cuvette. Density functional theory (DFT) calculations provided in the Gaussian 16 package were used to calculate the molecule's molecular orbitals and vibrational energies. The B3LYP correlation function, along with the 6–31 g(d,p) basis set, was employed for this calculation (Frisch et al. 2016). Based on the data obtained from the Gaussian 16 calculations, the observed UV–Vis spectra were modeled to a set of two Gaussian functions using MATLAB software to account for the two transitions that were responsible for the absorbance of the tylosin peak in the UV region. The full width half maximum (FWHM) of the modeled peaks was then extracted from which the lifetime of the excited state was inferred. The Gaussian simulating the lower energy transition is referred to as the first excited state (E₁), while the higher energy transition is dubbed the second excited state (E₂). To compare the vibrational results of the Gaussian calculation, a vibrational spectra of a tylosin powder was obtained using a Nicolet Summit FTIR spectrometer. The Raman data was collected using a Horiba/Jobin Yvon LabRAM HR800 Raman Microscope employing a 633 nm laser source, two averages of 30 s were used to acquire the data. Lastly, a MATLAB model of a polariton was constructed using the wave transfer matrix method of approximating optical spectra, describing how a polariton can be used to degrade a photoactive molecule such as tylosin.

Observations of the FWHM of tylosin in the UV region revealed that the FWHM of E₁ decreased as the pH of DI water increased, with the opposite phenomenon occurring in E₂. A decrease in FWHM suggests an increase in the lifetime of the excited state (Saleh & Teich 2007). Therefore, the increased H⁺ concentration at lower pH act to stabilize the E₂ energy level. On the contrary, a decrease in the H⁺ concentration at higher pH seems to stabilize the E₁ energy level, as seen in Figure 3. The results suggest that the high polarity of acetonitrile stabilizes the excited state in comparison to 2-propanol. However, it seems that steric hindrance must also play a role since the FWHM of both excited states were calculated to be higher when tylosin was dissolved in 2-propanol as compared to water. When dissolved in the non-polar solvent toluene, the lifetimes of the excited states become inverted, with the FWHM of the second excited state calculated to be smaller than that of the first excited state.

When the molecule was dissolved in toluene, a drastic decrease in the attenuation coefficient was observed, as seen in Table 1. Several factors might contribute to such an outcome. First, toluene possesses its own conjugate bond system and absorbs strongly in the same UV region as tylosin (Ginsburg et al. 1946). Therefore, it may have effectively photobleached the system, outcompeting the tylosin for the available photons. Second, because tylosin absorbs in the same UV region, the excited state PES of the two molecules can form conical intersections, allowing for the excited tylosin to quickly disperse energy into the PES of the toluene (Kelley 2012).

It is also noted that while the maximum peak energy of tylosin does not shift in most solvents, there is a considerable blue shift of 154 meV when tylosin is dissolved in acetonitrile, accompanied by an increase in the attenuation coefficient. A blue shift indicates an increase in the energy gap between the ground and excited states, implying that the ground state of tylosin is more polar than the excited state (Reichardt 1994). The polar acetonitrile molecules interact with the electronegative conjugated region of tylosin to stabilize the molecule, allowing the ground state to relax into a lower energy state. Tylosin molecules would require slightly larger light energy to reach the excited states in a more relaxed configuration. However, for a molecular electron transition to occur, there must be some overlap between the ground and excited state wave functions. As seen in Table 1, this phenomenon is also observed to as smaller degree in 2-propanol, in which the peak is shifted by 91 meV. 2-Propanol possesses a polarity intermediate of water and acetonitrile, which likely explains why the tylosin dissolved in propanol has an absorption maximum intermediate between the two solvents (Kosower 1957).

The probability of making a transition between states depends on the complex conjugate of the transition dipole moment = ⁠, where r is the dipole operator, which is a vector dependent on the position and charge of electrons and nuclei within the molecule (Gieseking et al. 2014). Meanwhile, and are the two wavefunctions describing the eigenstates of the molecule involved in the transition. Because eigenstates are typically described in Hilbert (vector) space if and are orthogonal to each other, there will be no transition (Mukamel 1995). For this reason, a molecule stabilized within its ground state configuration can have a better overlap with the excited state orbitals, resulting in a larger attenuation coefficient.

Lastly, we propose a theoretical model in which a porphyrin-based cavity polaritons can be used as a tool to degrade tylosin. Voigt & Jaeger (2017) demonstrated that tylosin quickly degrades when exposed to high-energy UV radiation. Several issues of using high-energy UV light for such purposes are discussed in the introduction, such as safety concerns and its short penetration depth. However, it may be possible to use lower energy light to form cavity polaritons, which will then initiate the decay of the tylosin molecule. Cavity polaritons are formed when a molecule is placed inside an optical cavity whose photon is tuned to the excitation wavelength of the molecule. If the molecule and the photon exchange energy faster than the decay of either state, they form a polariton state, a quasi-particle with hybrid characteristics of both light and matter. Two polariton states will form when this process occurs, a lower and an upper polariton (Ebbesen 2016). As indicated by its name, the upper polariton energy will be higher than that of the molecule in free space.

In this study, steady-state spectroscopy and computational tools were used to analyze the spectra of tylosin under different solvent conditions, concluding that the UV peak in tylosin is a result of two type transitions of similar intensity. The orbitals involved in these transitions are centered around an electronegative portion of the molecule, making them particularly sensitive to solvent polarization effects. The spectroscopic activity of the molecule was also greatly reduced when dissolved in a solvent that is also optically active in the same region as itself, such as toluene. Finally, an optical model of a porphyrin polariton was constructed, which indicates that the partially photon-like nature of the upper branch of the upper polariton may be able to initiate the photodegradation of tylosin. This fundamental information can further advance light-based techniques to break down tylosin.

The spectral measurements were performed at the U.S. Meat Animal Research Center, Animal Health Genomics Research Unit in Clay Center, NE. Density functional theory calculations were performed using Wayne State University's High-Performance Computing system. The vibrational measurements were carried out with the help of the University of Nebraska-Kearney's Department of Chemistry. Mention of a trade name, proprietary product, or specific agreement does not constitute a guarantee or warranty by the USDA and does not imply approval of the inclusion of other products that may be suitable. USDA is an equal opportunity provider and employer.

All relevant data are included in the paper or its Supplementary Information.

The authors declare there is no conflict.