Exploring the fluorescence quenching interaction of amino acids and protein with natural organic matter by a multi-spectroscopic method Open Access

The successful determination of disinfection by-products in water and wastewater analysis using the fluorescence method is dependent on the fluorescence intensity generated by the targeted compound. The widespread presence of natural organic matter (NOM), such as dissolved organic matter (DOM), proteins and amino acids in water have the potential to interact with each other, leading to the suppression or quenching and enhancement of the fluorescence intensity (Wang et al. 2015). In natural water, due to the natural or anthropogenic presence of other organic compounds, such as protein, amino acids, pollutants, or surfactants, the interaction between DOM and natural/anthropogenic organic compounds is inter-molecule (amino acid and amino acid) or inter-component (amino acid, proteins and other components) (Hernandez-ruiz et al. 2011; Maqbool & Hur 2016). The DOM–DOM or DOM–other organic compound interactions have the potential to modify the indigenous fluorescence property of each individual constituent, which results in the interference in quantification of these compounds by the fluorescence mechanism. In fact, the intercomponent interactions of fluorescent DOM have been found to significantly reduce fluorescence dynamics (Wang et al. 2015), as a result, the classical approach of the fluorescence method to determine proteinaceous components in DOM is underestimated.

An excitation emission matrix (EEM) is one of the most widely used methods for the characterization of DOM (Yu et al. 2020). The contour plot of wavelength of excitation and emission vs. fluorescence intensity, and the information from it, such as peak intensity and position, wavelength zone, and spectral deconvolution, give important information on the molecular fingerprint (physicochemical properties) of different types of DOM, for example DOM dynamics (Fellman et al. 2010), disinfection by-products (Peleato & Andrews 2015), membrane fouling (Teychene et al. 2018), and toxicity change (Khanal et al. 2019). EEM is based on the principle that when compounds with fluorophores after energy absorption are transferred to an excited state, and when these excited compounds release the energy, fluorescence is observed. Based on the intensity of fluorescence, the characterization of DOM is done by EEM. However, when inter-molecule or intercomponent interaction occurs, there is a possibility that fluorescence might be suppressed. Such a suppression of fluorescence is referred to as fluorescence quenching. Despite the widespread application of EEM, the mechanism of fluorescence quenching is still a matter for further investigation.

Most previous studies have focused on the investigation of the mechanism of the interaction between DOM and other compounds such as metals (Zhang et al. 2010; Fu et al. 2016) and pharmaceuticals (Ruiz et al. 2013; Wang et al. 2016). There are limited studies on the mechanism of the interactions between DOM constituents involving either amino acid-fulvic acid (FA) or protein–humic acid, which may not well represent the actual situation in natural aquatic systems (Wang et al. 2015; Lin et al. 2017). In most cases, the concentration of proteins or peptides present in experiments is much higher than that in the natural water environment (Aiken 2014). In addition, the acidic fractionation in the FA isolation procedure is known to potentially alter its parental compound structure (Pan et al. 2016). Thus, the amino acids–FA interaction may not clearly reveal the mechanism of fluorescence loss of protein-related compounds. Further investigation is required to elucidate the interaction mechanism between protein-related compounds and humic substances. In previous research, the quenched fluorescence intensity was found to influence the prediction accuracy of trihalomethane prediction, both trichloromethane and tribromomethane, by EEM-based models (Saipetch & Yoshimura 2019; Saipetch et al. 2021).

Considering the disadvantages of fluorescence quenching in the accurate characterization of DOM, this research has been designed to elucidate the fluorescence quenching mechanism of amino acids and protein compounds interacting with natural organic matter. Tryptophan and tyrosine were selected as amino acids. Bovine serum albumin (BSA) was selected to represent protein-related compounds because the secondary structure of BSA protein (∝-helix type) was found to possess the amino acid residues of tyrosine and tryptophan, which are the main fluorescence sources of the protein molecule. In addition, tryptophan, tyrosine, and BSA can also be considered to represent autochthonous DOM (Pivokonsky et al. 2015). The natural organic matter was represented by Suwannee River natural organic matter (SWNOM), one of the most widely characterized NOMs, and can also be considered to represent allochthonous DOM (Ateia et al. 2017). The mechanism of fluorescence quenching was elucidated using multi-spectroscopic methods (i.e., EEM, UV–Vis absorbance and Fourier-transform infrared (FTIR) spectroscopy). The complexity of peaks in the EEM data, excitation, emission and intensity, was decoded to provide quantitative information on the peak intensities of fluorescent components (i.e., component scores) of each component in the mixture by Parallel Factor Analysis (PARAFAC) (Yang et al. 2015). Investigation of the binding force in the case of the formation of a quenched complex from the interaction between protein and DOM was calculated by thermodynamic calculations.

L-tryptophan (≥99%, Kanto Chemical), L-tyrosine (≥98%, Sigma–Aldrich), and BSA (≥96%, Sigma–Aldrich) were obtained as representatives of protein-related compounds (Appendix 1). SWNOM (2R101N) was obtained from the International Humic Substances Society (IHSS) to represent a humic substance. Tryptophan was selected as a representative fluorescence quencher because tryptophan-like fluorescence has been reported to be the strongest protein-like fluorescence in freshwater. Amino acid and protein concentrations were selected on the basis of equivalent strength of fluorescence intensities of both the humic-like and protein-like substances in Tama River, Tokyo, Japan. The synthetic water of SWNOM (5 mg/L) and tryptophan (0.5 μM) showed similar fluorescence intensities to both humic-like and protein-like substances in the Tama River. For better comparison amongst tryptophan, tyrosine, and BSA, the same molar concentration as tryptophan was selected for both tyrosine and BSA to compare the quenching mechanism across different molecular weights. NOM was prepared in the range 0–10 mg/L, which was similar to the range of 0–10 mgC/L of DOC concentration in freshwater (Chen et al. 2003; Williams et al. 2016; Raeke et al. 2017; Tungsudjawong et al. 2018). Based on our unpublished experience and detailed elemental composition of SWNOM (2R101N – 50.70% C – Source, IHSS (IHSS, 2020), 1 mg 2R101N NOM/L has approximately 0.5 mgC/L of DOC.

The titration of amino acids or protein solution by the SWNOM solution is considered to be representative of the interaction between autochthonous DOM and allochthonous DOM in an aquatic system. Tryptophan, tyrosine, and SWNOM were separately dissolved in 0.1 M NaOH to obtain the concentrated stock solutions of 10 mM, 10 mM, and 10 g/L, respectively. The pH of these stock solutions was adjusted to 7.0 ± 0.2, representing the pH of the neutral water in natural environment, using 1 M HCl and 1 M NaOH. Each of the solutions was diluted with ultrapure water of resistivity of 18.3 M Ohm cm (Milli-Q (MQ) water, Millipore Corporation, Japan) for the preparation of tryptophan, tyrosine and SWNOM at 100 μM, 100 μM and 200 mg/L, respectively. A stock solution of BSA (6 μM) was prepared by dissolving BSA into MQ water at 4 °C for 10 min. All the stock solutions were passed through a 0.45 μM polyethersulfone membrane filter before fluorescence titration.

The mechanism of fluorescence quenching was explored by two types of fluorescence titrations of tryptophan, tyrosine and BSA with SWNOM, as reported above (Wang et al. 2015). In the first type of fluorescence titration, a fixed concentration of amino acids or protein (0.5 μM) was titrated with SWNOM (0 to 10 mgC/L) in phosphate buffer solution (PBS) (pH 7.0± 0.2) at a control temperature of 298, 308, 318 and 328 K using a water bath. In the second type, fluorescence titration was conducted by titrating amino acids or protein (0–1.25 μM) with SWNOM (0–10 mgC/L) in PBS (pH 7.0± 0.2) only at 298 K. The ionic strength of titrated, control and blank (PBS) samples was adjusted to 0.01 M using NaCl to avoid its effect on the fluorescence measurement. Finally, the titrated solution was analyzed by EEM, UV–Vis absorbance and FTIR spectroscopy.

EEM spectra were measured at the controlled room temperature (298 K) using a spectrofluorometer with a xenon lamp (RF5300, Shimadzu, Japan) and a quartz cuvette (1-cm path length). In the first type of fluorescence titration, prior to EEM analysis, an inner filter effect (IFE) of the titrated samples was removed by diluting stock solution with MQ water so that the UV absorbance of the sample became less than 0.1 cm⁻¹ (Murphy et al. 2013). Then, the EEM spectra of the samples and blanks (PBS) were scanned over an excitation (Ex) range of 275–280 nm with a 5-nm increment and an emission (Em) range of 280–550 nm with a 2-nm increment. A small wavelength increment was performed to minimize the reduction of change in fluorescence due to the change in temperature, if any, during the EEM analysis. For the second type of fluorescence titration, the EEM spectra of the samples and blanks were scanned over an Ex and Em range of 240–500 nm with a 5-nm increment and 280–550 nm with a 2-nm increment, respectively. Then, the raw EEM data were pretreated following the procedure reported earlier (Murphy et al. 2013) to remove the IFE and the fluorescence signal fluctuation (Raman unit (RU) calibration). Pretreatment of the EEM data was performed on MATLAB software using the drEEM toolbox (Murphy et al. 2013).

Besides the EEM analysis, UV–Vis and FTIR spectra were also analyzed to obtain supplementary information in understanding the mechanism of fluorescence quenching. The UV–Vis spectra were scanned over the wavelength of 190–800 nm using a spectrophotometer (UV-1800, Shimadzu, Japan) with a quartz cuvette (1-cm path length). Prior to the FTIR analysis, solutions of BSA and SWNOM, with mass ratio (BSA:SWNOM) identical to the condition in the fluorescence titration, were freeze-dried for 24 h using a freeze dryer (FDU-1200, EYELA, Japan). Finally, the freeze-dried samples were analyzed using a FTIR spectrometer (FTIR-4600, Jasco, Japan) with a mini KBr plate (3 × 3 × 0.5 mm) and clear disk (CD-05, Jasco, Japan) at a resolution of 4 cm⁻¹ in transmittance mode.

The mechanism of fluorescence quenching of amino acids and protein by SWNOM was investigated by interpreting the results of multi-spectroscopic techniques namely, EEM, UV–Vis absorbance and FTIR spectroscopies. The analysis of fluorescence quenching using EEM data is challenging because of the large volume of data which tends to overlap with the spectra from the fluorophores containing similar structures. Therefore, the overlapping of the EEM spectra was minimized by the application of PARAFAC. This analysis decomposed the mixed EEM spectra of fluorescence titration into a limited number of independent components with the information of their peak locations and intensities. The appropriate number of PARAFAC-derived components was determined by the criteria of having a low sum of square error and passing the core consistency of 80% (Stedmon et al. 2003). It should be noted that in this study, the PARAFAC components could not be derived from the dataset, which included tryptophan–SWNOM, tyrosine–SWNOM, and BSA–SWNOM, due to the large intensity difference of tyrosine relative to tryptophan (3-folds) and BSA (19-folds) at the same molar concentration. Therefore, each mixture of the 144 samples of tryptophan–SWNOM, tyrosine–SWNOM and BSA–SWNOM was separately analyzed using PARAFAC.

In Equation (1):

• F₀ and F are the fluorescence intensities of tryptophan, tyrosine, or BSA in an absence and presence of SWNOM, respectively;

• k_q is the bimolecular quenching constant (L mol C⁻¹ s⁻¹);

• τ₀ is the lifetime of the protein-related compound fluorophore in the absence of SWNOM (s);

• Q is the concentration of SWNOM (mmol C/L); and

• k_qτ₀ denotes KD in the case of collisional quenching and Ksv in the case of static quenching. Both KD and Ksv are Stern–Volmer quenching constants, and were derived from the Stern–Volmer plot using linear regression (Lackowicz 2006).

In Equation (2):

• f_a is a fraction of initial fluorescence which is accessible to the quencher, and

• K_a is a modified Stern-Volmer quenching constant (L mol C⁻¹).

In Equation (3):

• K is the binding constant (L mol C⁻¹), and

• n is the number of binding sites.

The experimental approach is shown in Figure 1.

Two independent peak intensities of fluorescent components (i.e., component scores) of each component from the EEM spectra of each fluorescence titration mixture was extracted by PARAFAC analysis with a core consistency of 80% (Figure 2). Two PARAFAC components of each mixture of tryptophan-SWNOM, tyrosine–SWNOM and BSA–SWNOM are shown in Figures 2(a)–2(b), 2(a)–2(c), and 2(a)–2(d), respectively. The locations of the peak of tryptophan, tyrosine, and BSA from PARAFAC were at Ex/Em 280/356 nm (peak T), 275/302 nm (peak B), and 280/344 nm (peak T), respectively. For SWNOM, two peaks appeared at Ex/Em of 240/448 nm (peak A), and 350/450 nm (peak C). All the PARAFAC component peaks located in the reference region respective to the characteristic of their parent compounds were consistent to earlier studies (Coble 1996; Chen et al. 2003; Saipetch et al. 2021).

The fluorescence titration with a fixed concentration of tryptophan by SWNOM increased the fluorescence intensity of peak T. SWNOM had fluorescence in an overlapping location with peak T (Figure 3(a) and 3(b)). After the elimination of the overlapped SWNOM fluorescence by PARAFAC, the fluorescence intensity of peak T did not decrease with the addition of SWNOM. Moreover, the addition of SWNOM did not decrease the fluorescence intensity of peak B of tyrosine (Figure 3(c) and 3(d)), while a substantial decrease in the fluorescence intensity of peak T of BSA was observed (Figure 3(e) and 3(f)).

The negligible effect of SWNOM on the fluorescence intensities of tryptophan and tyrosine was contrary to a previous observation (Wang et al. 2015). The non-quenching patterns of tryptophan and tyrosine fluorescence were found to be the same at all levels of initial concentrations of tryptophan and tyrosine, at 0.15, 0.25, 0.75, 1 and 1.25 μM (Figure 3(b) and 3(d)), whereas the fluorescence intensity of BSA was quenched at all levels of initial concentration of BSA (Figure 3(f)). The increase in the SWNOM concentration resulted in higher reduction of the fluorescence intensity of BSA (Figure 3(f)). The Kruskal–Wallis test confirmed the non-quenching pattern of the fluorescence intensity of peaks T and B of tryptophan and tyrosine with the addition of SWNOM (p > 0.95), and it also confirmed the significant reduction of peak T of BSA (p < 0.05). Thus, only the fluorescence intensity of BSA was quenched by the interaction with SWNOM among the three types of protein-related compounds. Such a non-quenching phenomenon of tryptophan fluorescence and quenching of BSA fluorescence were also previously observed in the case of interaction with iodide (Moller & Denicola 2002).

Iodide, SWNOM and target compounds (amino acids and BSA) all have negative charge at a given pH of 7 (Table 1). Among them, amino acids were not quenched, whereas BSA was quenched. The possible reason is that the target compounds were negatively charged, and their coupling with humic substances cannot be expected due to electrostatic repulsion which does not lead to quenching. However, in spite of BSA having negative charge, it was quenched. The difference in the quenching of amino acids and BSA could be due to the different hydrogen bonding and van der Waals force between the BSA–SWNOM and amino acid–SWNOM complexes.

As there was no fluorescence quenching of tryptophan and tyrosine with the addition of SWNOM, analysis of the Stern–Volmer plot at various temperatures on tryptophan–SWNOM and tyrosine–SWNOM interactions was deemed unnecessary. For the BSA–SWNOM interaction, an increase in temperature decreased the Stern–Volmer quenching constant (Figure 4).

The Stern–Volmer quenching constants between BSA and SWNOM were found to be 9.99 × 10³ L mol C⁻¹ at 298 K (R² = 0.88, standard error (SE) = 1.15 × 10³ L mol C⁻¹), 8.34 × 10³ L mol C⁻¹ at 308 K (R² = 0.78, SE = 1.40 × 10³ L mol C⁻¹), 5.56 × 10³ L mol C⁻¹ at 318 K (R² = 0.80, SE = 0.87 × 10³ L mol C⁻¹), and 3.05 × 10³ L mol C⁻¹ at 328 K (R² = 0.78, SE = 0.51 × 10³ L mol C⁻¹), respectively. The correlation coefficient (r) between the Stern–Volmer quenching constant and temperature was negative (r = –0.99, p < 0.05), which indicated a static quenching mechanism (Lackowicz 2006). Thus, all the calculated quenching constants from the Stern–Volmer equation were denoted as Ksv. The rising temperature in a static type of fluorescence quenching typically leads to the dissociation of the ground-state complex, which decreases Ksv with the increase in temperature. On the other hand, the rising temperature in a collisional type of fluorescence quenching leads to a larger diffusion coefficient and faster molecular motion, which results in an increase in its quenching constant (KD).

The Stern–Volmer plots in Figure 4 show upward curves along the x-axis, which ideally differs from the theory of linear relationship. Nevertheless, the results from the modified Stern–Volmer and double log analyses (Figure 5) confirmed the static type of fluorescence quenching for the BSA–SWNOM interaction. In the modified Stern–Volmer analysis, the rising temperature decreased the binding constant (K), as a result of the dissociation of the ground state complex at increasing temperature, which was in accordance with a static quenching mechanism (Figure 5 and Table 2).

It is evident from Figure 4 (and also Figure 5) that fluorescence quenching of BSA–SWNOM is a function of both the SWNOM concentration and temperature. For a fixed temperature, when the SWNOM concentration increased, the ratio of fluorescence intensity in the absence and presence of SWNOM (F₀/F) in Figure 4 (or log ((F₀-F)/F in Figure 5)) also increased, which means the fluorescence intensity of BSA in the presence of SWNOM (F) decreased. Since, the fluorescence intensities of BSA in the absence of SWNOM were the same for all temperatures (Figure 4), the increase in F₀/F can be interpreted as a decrease in F, i.e., the fluorescence intensity of BSA in the presence of SWNOM decreases. Similarly, it can also be interpreted that, for a fixed SWNOM concentration, an increase in temperature decreases F₀/F, which means quenching of BSA by SWNOM decreases with increase in temperature.

The values of k_q from the double log analysis ranged from 2.40 × 10¹¹ L mol C⁻¹s⁻¹ to 4.11 × 10¹² L mol C⁻¹s⁻¹ at all levels of initial BSA concentration (Table 3), which were higher than the largest possible value of k_q of collisional quenching in an aqueous solution of 1 × 10¹⁰ M⁻¹s⁻¹ (Lackowicz 2006; Kandagal et al. 2008; Ghosh et al. 2015). Thus, the collisional type of fluorescence quenching was not a possible option for the quenching of the BSA–SWNOM interaction. The integration of information from the Stern–Volmer, modified Stern–Volmer and double log analyses indicated that the formation of the ground-state complex between BSA and SWNOM is the cause of the quenching of the fluorescence intensity of BSA.

The peak absorption spectrum for BSA was seen at 280 nm, which was mainly due to the presence of tryptophan residue in BSA (Figure 6(c), and Figure 2, also as discussed above, under Quenching in fluorescence spectra). There was no any specific peak absorption spectrum of SWNOM, and the spectrum was seen over a wide absorbance without any specific peaks (Figure 6(a)–6(c)). The absorption spectrum of BSA–SWNOM had a peak at the same location as the absorption spectrum of BSA (280 nm) (Figure 6(c)). The difference in the absorption spectrum of BSA–SWNOM, and the BSA spectrum was the same as the absorption spectrum of BSA (Figure 6(c)). There was no deviation in the absorption spectrum of BSA–SWNOM, and BSA with the addition of SWNOM, to BSA. Thus, the addition of SWNOM to BSA can be said not to have altered the bonding arrangement of both the π-bond and a lone pair in the BSA molecule because of the same energy requirement for the π-π* transition, and n-π* transition for both the cases of before and after SWNOM addition (i.e., no change in the absorption spectrum). In addition, no deviation of the absorption spectrum of tryptophan and tyrosine was observed with the addition of SWNOM (Figure 6(a) and 6(b)), which was opposite to the interaction between tryptophan or tyrosine with a fulvic acid fraction of SWNOM in earlier research by Lin et al. (2017), due to no bonding or quenching of tryptophan and tyrosine humic acid fraction of SWNOM in this research. In their research, Lin et al. (2017) found the formation of non-covalent bonding or static quenching between tryptophan or tyrosine with a fulvic acid fraction of SWNOM. The deviation of UV–Vis absorbance spectrum of BSA–SWNOM in this research implies that the stabilization force of the BSA–SWNOM complex is non-covalent bonding, which leads to static quenching.

The FTIR spectra of BSA contained one broad peak at a wavenumber of 3,304 cm⁻¹, and three sharp peaks at wavenumbers of 1,656, 1,545, and 1,451 cm⁻¹ (Figure 7). The broad peak at 3,304 cm⁻¹ in BSA was caused by stretching of the N–H bond of the amino group. Three sharp peaks were observed at 1,656 cm⁻¹ (amide I), 1,545 cm⁻¹ (amide II), and 1,451 cm⁻¹, which were caused by the C = O stretching vibration of the peptide bond, stretching of the N–H bond of the amino group in the peptide chain, and presence of primary amine, respectively. In the case of SWNOM, its spectrum contained one broad peak with two sharp peaks. The broad peak was at 3,431 cm⁻¹ due to the O–H of alcohol and phenol. The two sharp peaks at 1,582 cm⁻¹ and 1,400 cm⁻¹ were due to the C–O symmetric stretching of carboxyl moieties (Fu et al. 2016). Even though the pattern and peaks of the FTIR spectrum of SWNOM in this research and that in IHSS were similar, the peak at 1,669 in IHSS compared with 1,660 in this research, was probably caused by the freeze-dried nature of the sample and slightly lower resolution of the FTIR in this research compared with that by IHSS (IHSS 2020).

The addition of SWNOM to BSA shifted the peak position of amide I and amide II from 1,656 cm⁻¹ to 1,659 cm⁻¹, and from 1,545 cm⁻¹ to 1,548 cm⁻¹, respectively. The amide I peak is used to indicate conformational change of the protein structure because the amide I peak observes the stretching frequency of the C=O hydrogen bonded to N–H moieties, which is dependent on the secondary structure adopted by the peptide chain (Kandagal et al. 2008). The shift of the amide I peak indicates the disturbance of hydrogen bonding, which is the holding force for the formation of the secondary structure of protein from polypeptide. As a result, the addition of SWNOM led to the change in the structure of BSA (Kandagal et al. 2008). The SWNOM–BSA complex possibly caused the reduction of the polypeptide carbonyl hydrogen bonding network and probably reduced the ∝-helix structure of BSA. BSA is highly likely to be unfolded by the interaction with SWNOM as indicated by the highest adiabatic compressibility of 6.6 × 10⁻⁶ cm³ g⁻¹bar⁻¹ compared with other types of protein, such as pepsin 5.2 × 10⁻⁶ cm³ g⁻¹bar⁻¹ and lysozyme 1.3 × 10⁻⁶ cm³ g⁻¹bar⁻¹ (Zhao et al. 2014). Previous studies have also reported unfolding of BSA by the interaction with organic matter such as humic acid (Zhao et al. 2014; Guan et al. 2018).

The estimated ΔH, ΔS, and ΔG indicated two possible binding forces of the BSA–SWNOM complex, which are van der Waals force and hydrogen bonding, according to the negative values of both ΔH and ΔS, as shown in Table 2. The binding force is hydrophobic if both ΔH and ΔS are positive, whereas it is electrostatic interaction if ΔH is negative and ΔS is positive. In addition, the negative value of ΔG indicated that the interaction process was spontaneous (Kandagal et al. 2008; Lin et al. 2017). It was also evident that electrostatic interaction was not likely to occur as the binding force in the BSA–SWNOM complex because both of them had the same charge (negative) at the working pH according to their isoelectric points (Table 1).

By combining all the results, it is confirmed that the fluorescence intensity of BSA is quenched by the formation of the BSA–SWNOM complex by hydrogen bonding or van der Waals force. The mechanism of quenching is illustrated in Figure 8. Ionic interaction, hydrogen bonding between R side chain, van der Waals force, a salt-bridge between two non-covalent interactions of hydrogen bonding, and ionic bonding in the SWNOM–BSA mixture are the principle sites and mechanisms of binding of SWNOM and BSA (Figure 8).

The presence of a fluorescence quencher, such as natural organic matter represented by SWNOM in this study, reduces the fluorescence intensity of a targeted compound, such as BSA, in a sample, thereby masking the actual concentration of the targeted compound. Prior knowledge of the concentration of the quencher, targeted compound, the quenching mechanism, and bonding sites of quenching, as elucidated in this study for a typical BSA–SWNOM complex, helps in accurately predicting the concentration of the targeted compound in real water samples, such as wastewater and natural water. The possibility of the presence of a quencher in both wastewater and natural water cannot be overlooked; hence, the results from this study are expected to contribute to the determination of the actual concentration of the targeted compound in a real water sample. As the quenching mechanism is dependent on a number of factors, for better elucidation of a natural water sample, it is recommended that future studies be done with characterization of the DOC in the water sample. In addition, studies on the integrated impact of other influent factors such as pH, temperature, dissolved oxygen, oxygen reduction potential and temperatures in fluorescence quenching are also recommended. Finally, further studies on other types of protein-related compounds such as peptide- and microbial-derived organic matter would help in the generalization of knowledge on the quenching mechanism of the fluorescence intensity of protein-related compounds by natural organic matter.

This study investigated the mechanism of the interaction between protein-related compounds (tryptophan, tyrosine and BSA) with natural organic matter (SWNOM) by fluorescence titration and multi-spectroscopic methods. The experimental results indicated non-quenching of tryptophan and tyrosine and quenching of BSA by interaction with SWNOM. The bonding between BSA and SWNOM was found to be the cause of fluorescence quenching of BSA. Hydrogen bonding and van der Waals force were likely to be the binding forces between BSA–SWNOM, possibly by altering the BSA structure.

This study was financially supported by the Ministry of the Environment of Japan [S-13-2-3] and JSPS KAKENHI [grant number 18H01566]. The first author would also like to thank Dr Manabu Fujii for his valuable suggestions in improving the manuscript.

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