Adsorption Separation of l-Tryptophan Based on the Hyper-Cross-Linked Resin XDA-200.

l-Tryptophan (l-Trp) was separated from its aqueous solution by hyper-cross-linked resins. The adsorption and desorption performances of l-Trp on different resins were compared. The weakly polar resin XDA-200 was selected as an excellent adsorbent with high adsorption amount and easy elution. The resin has a high adsorption selectivity and strong salt resistance. The adsorption mechanism of l-Trp on resin XDA-200 was elucidated based on adsorption thermodynamics experiments, molecular dynamics simulations, and adsorption kinetics experiments. The dynamic separation process of l-Trp was finally studied. The adsorption of l-Trp on resin XDA-200 is a spontaneous process driven by adsorption enthalpy. l-Trp± is the most favorable form for l-Trp adsorption on resin XDA-200 because of the strongest affinity of l-Trp± to the resin and relatively low water solubility. The adsorption of l-Trp is mainly based on π-π and hydrophobic interactions. Surface diffusion is the sole rate-limiting step of l-Trp mass transfer on resin XDA-200. l-Trp was separated satisfactorily from l-glutamic acid (l-Glu) and NaCl with both the recovery rate and purity of l-Trp higher than 99% in the fixed bed packed with resin XDA-200.

Introduction

l-Tryptophan (l-Trp) is an essential amino acid for humans and animals. It has been widely used in the food, pharmaceutical, and animal feed industries.[1] In recent years, with the development of l-Trp and its derivatives in new typical functional foods and antiviral and anticancer drug fields, the market demand for l-Trp has been increasing rapidly. Currently, the total market demand for l-Trp globally is more than 50 000 tons, with an annual growth of more than 10%. However, the yield of l-Trp every year is about 14 000 tons.[2] The output is far from meeting the market demand.

The industrial production of l-Trp is mainly achieved by microbial fermentation with Escherichia coli.[3] The separation and purification are one of the key steps restricting the industrial development of l-Trp.[4,5] After removing the microbial cells, proteins, and pigments in the fermentation broth, there are still multiple amino acid impurities including l-Glu, glycine, and aspartic acid apart from the high concentration of soluble salts. l-Glu is the main amino acid impurity in the preprocessed fermentation broth.[6]

The ion exchange technique is the widely used method to separate l-Trp from its preprocessed fermentation broth.[4,7,8] However, a lot of acids and alkalis are needed to regenerate the ion changer. Moreover, high concentrations of soluble salts can reduce the ion exchange capacity and separation efficiency of the ion changer.[9] Subsequent desalination is still needed to remove the salt in the l-Trp product obtained from the ion exchange system. Therefore, a novel, efficient separation method needs to be developed to separate l-Trp.

Hyper-cross-linked resins have a lot of merits including a large Brunauer–Emmett–Teller (BET) specific surface area, high adsorption capacity, easy elution and regeneration, high mechanical strength, and long service life.[10] The resins with different polarities can be obtained with different functional groups introduced into the resins through chemical modifications.[11] Currently, hyper-cross-linked resins are used to separate a lot of biochemical products.[12−14]l-Trp is a small amphoteric molecule with an indolyl. group. The strong hydrophobicity of indolyl leads to the relatively low solubility of l-Trp in water. l-Trp molecules are expected to have a relatively strong affinity to nonpolar and weakly polar hyper-cross-linked resins. The solution pH affects the degree of dissociation and integral hydrophobicity of l-Trp molecules. Therefore, l-Trp has different affinity strengths and adsorption amounts on the resins at different solution pH values. When the hyper-cross-linked resins are used to separate l-Trp, the efficient adsorption and desorption are realized easily through reasonably regulating the pH of the feed solution and eluants. The consumption of a lot of acids and alkalis is avoided. In addition, l-Glu has strong hydrophilicity, which leads to relatively low adsorption amount of l-Glu on nonpolar and weakly polar hyper-cross-linked resins. There are no dissociable groups on the resins. No adsorption of salts on the resins is expected. Therefore, high concentration of soluble salts in the fermentation broth of l-Trp will have little influence on the adsorption of l-Trp on the resins. Synchronous separation of l-Trp from l-Glu and soluble salts based on hyper-cross-linked resins can be realized. In conclusion, the separation of l-Trp based on hyper-cross-linked resins will be one of the alternative methods to ion exchange. To the best of our knowledge, the separation of l-Trp based on hyper-cross-linked resins has never been reported in the literature.

In this work, hyper-cross-linked resins were used to separate l-Trp. The adsorption and desorption performances of different resins were compared. The resin with high adsorption amount to l-Trp and easy elution was selected. The adsorption selectivity and salt tolerance of the resin were studied by static adsorption equilibrium experiments. The adsorption mechanism of l-Trp on the resin was elucidated based on adsorption thermodynamics experiments, molecular dynamics (MD) simulations, and adsorption kinetics experiments. The dynamic separation process of l-Trp was studied in the fixed bed packed with the hyper-cross-linked resin.

Results and Discussion

Screening Resin for Separation of l-Trp

The adsorption and elution performances of l-Trp on different types of resins were compared. The adsorption amounts of l-Trp on different resins are shown in Figure . As can be seen in Figure , the adsorption amount of l-Trp on resin XDA-200 is higher than that on other resins. The adsorption amount of a certain adsorbate is mainly affected by specific surface area and the surface properties, especially the polarity of adsorbents.[15] Generally, a large specific surface area leads to the high adsorption capacity of the resins.[16] The specific surface areas of the resins XDA-200, XDA-1, LX-3020, and LX-68M are higher than those of other resins. Therefore, the adsorption amounts of the four resins are higher than those of other resins. The adsorbents without dissociable groups adsorb adsorbates mainly via the hydrophobic interaction, hydrogen bonding interaction, van der Waals force, and so on.[17,18] The resins XDA-200 and LX-68M are both weakly polar resins, while the resin LX-3020 is a strongly polar resin. The hydrophobic interaction between the two weakly polar resins and indole group in l-Trp molecules is stronger than that between resin LX-3020 and the indole group. Therefore, the adsorption amounts of resins XDA-200 and LX-68M to l-Trp are higher than that of resin LX-3020. Although the resins XDA-200 and LX-68M are both weakly polar resins, the hydrogen bonding interaction between l-Trp molecules and the carbonyl on resin XDA-200 leads to a higher adsorption amount of resin XDA-200. The resin XDA-1 has a higher specific surface area than XDA-200. Moreover, the resin XDA-1 is a nonpolar resin. The hydrophobic interaction between resin XDA-1 and l-Trp is stronger than that between resin XDA-200 and l-Trp. However, the adsorption amount of resin XDA-200 to l-Trp is slightly higher than that of the resin XDA-1. The reason may be that a few adsorption sites in pores of resin XDA-1 are inaccessible for l-Trp molecules because the hydration of nonpolar resin XDA-1 is not favorable in energy.[19]

Figure 1

Adsorption amounts of l-Trp on different resins.

The adsorption amounts of resins XDA-1, XDA-200, LX-3020, LX-68M, and LX-T81 are relatively higher than those of other resins (see Figure ). Therefore, the elution performances of the five resins were further studied. The recovery rate of l-Trp from resin XDA-200 is slightly higher than that from the other four resins (see Figure ). The result indicates that resin XDA-200 can be eluted more easily than other resins. In conclusion, resin XDA-200 was selected as the adsorbent for our subsequent experiments because of its higher adsorption amount and easier elution. The effects of soluble salts on the adsorption amount and adsorption selectivity of the resin to l-Trp relative to l-Glu were further studied.

Figure 2

Recovery rate of l-Trp from different resins.

NaCl was selected as the representative of soluble salts in the preprocessed fermentation broth of l-Trp. The adsorption amount of l-Trp on resin XDA-200 at different pH values when NaCl or no NaCl exists in the solutions is shown in Figure . As can be seen in Figure , the adsorption amount of l-Trp is higher when NaCl exists in the solution. The phenomenon may be explained by the “salting-out” effect. Wang et al. found the same phenomenon in the research about phenol adsorption on ethylenediamine-modified hyper-cross-linked polystyrene resins HJ-D33.[12] The adsorption amounts of resin XDA-200 to l-Glu at different pH values are also shown in Figure . As can be seen in Figure , the adsorption amounts of l-Trp on resin XDA-200 are significantly higher than those of l-Glu in the pH range of 2–10. There is almost no l-Glu adsorption on resin XDA-200. In conclusion, the resin XDA-200 shows excellent salt resistance and high adsorption selectivity to l-Trp.

Figure 3

Adsorption amount of l-Trp and l-Glu on resin XDA-200 at different pH values.

Effect of pH on the Adsorption Amount of l-Trp

As can be seen in Figure , the adsorption amount of l-Trp on resin XDA-200 increases first and then decreases with the increasing pH from 1.5 to 12. The adsorption amounts of l-Trp in the pH range of 4–8 are higher than those at other pH values. The phenomenon occurs because l-Trp molecules exist mainly as zwitter-ions in the pH range of 4–8 (see Figure ). The solubility of l-Trp± is lower than those of l-Trp+ and l-Trp–. The low solubility of adsorbates in the solution promotes adsorption of the adsorbates on hyper-cross-linked resins.[20] Therefore, low solubility benefits the adsorption of l-Trp. There is almost no adsorption of l-Glu on the resin XDA-200 in the pH range studied. The result is probably due to the strong hydrophilicity of l-Glu molecules. The affinity between resin XDA-200 and l-Glu is very low, and l-Glu has a relatively large solubility. Based on the relationship between the adsorption amount of the two amino acids and solution pH, the pH of the feed solution in the subsequent dynamic separation process can be set in the range of 4–8, and the pH of the eluant can be set at about pH 12.

Figure 4

Theoretical concentration distribution of l-Trp species at different pH values.

Thermodynamics of l-Trp Adsorption on the Resin XDA-200

The adsorption isotherms of l-Trp on resin XDA-200 at different temperatures are shown in Figure . As can be seen in Figure , the temperature has a negative effect on the adsorption of l-Trp. The adsorption isotherm data of l-Trp were fitted by Langmuir and Freundlich adsorption isotherm models. The adsorption isotherm model parameters are listed in Table . As can be seen in Table , the values of R2 for the Langmuir adsorption isotherm model are larger than those for the Freundlich isotherm model. The result indicates that the adsorption isotherms of l-Trp can be fitted better by the Langmuir adsorption isotherm model. The values of RL at 298, 308, and 318 K are all in the range of 0–1 because the values of KL are all positive. The results indicate the favorable adsorption of l-Trp on resin XDA-200.

Figure 5

Adsorption isotherms of l-Trp on resin XDA-200 at different temperatures.

Table 1

Adsorption Isotherm Parameters of l-Trp

	Langmuir isotherm model			Freundlich isotherm model
T (K)	KL (L/mol)	qm (mmol/g)	R2	Kf (L/mol)	n	R2
288	277.99	0.59	0.997	4.46	2.04	0.988
298	198.40	0.62	0.982	5.11	1.86	0.977
308	125.95	0.72	0.999	6.98	1.63	0.995

The adsorption equilibrium depends not only on the adsorption enthalpy change but also on the adsorption entropy change. Therefore, the adsorption process can be divided into enthalpy promotion and entropy promotion processes. The Gibbs free energy change determines the spontaneity of adsorption processes and the negative value of the Gibbs free energy change indicates that the adsorption of adsorbates proceeds spontaneously.[21] The adsorption thermodynamic parameters were calculated by the following equations[22]where ΔG is adsorption Gibbs free energy change (kJ/mol), ΔH is the adsorption enthalpy change (kJ/mol), ΔS is the adsorption entropy change (J/mol·K), R is the universal gas constant (8.314 J/mol·K), T is the temperature (K), and Ke0 is the thermodynamic equilibrium constant of adsorption process (dimensionless).

Ke0 can be calculated by the following equation[23]where γ is the activity coefficient (the value of the activity coefficient is 1 because of the relatively low concentration of l-Trp solutions) and [adsorbate] is the standard concentration of the adsorbate (1 mol/L).

ΔG was calculated by eq . ΔH and ΔS were calculated from the slope and intercept of the plot ln Ke0 versus 1/T shown in Figure S1. The calculated adsorption thermodynamic parameters are listed in Table . The values of ΔG at different temperatures are all negative. The result indicates that the adsorption of l-Trp onto resin XDA-200 is spontaneous. The absolute value of ΔG increases as the temperature decreases. The result indicates that the adsorption of l-Trp is more favorable at low temperatures. The negative ΔH (−29.13 kJ/mol) means that the adsorption of l-Trp is an exothermic process. The absolute values of ΔH are lower than 40 kJ/mol, indicating that the adsorption of l-Trp belongs to physical adsorption. The negative ΔS (−54.21 J/mol·K) suggests that the adsorption system is more ordered after the adsorption of l-Trp. The negative ΔH and ΔS indicate that the adsorption of l-Trp is an enthalpy promotion process.

Table 2

Adsorption Thermodynamic Parameters of l-Trp

T (K)	ΔH (kJ/mol)	ΔS (J/mol)	ΔG (kJ/mol)
288	–29.13	–54.21	–13.47
298			–13.11
308			–12.38

MD Simulations of XDA-200-l-Trp and XDA-200-l-Glu Adsorption Systems

l-Trp and l-Glu molecules exist at different ionization states under different solution pH values. Thus, the adsorption behaviors of l-Trp and l-Glu at different protonation states were all investigated by MD simulations. The MD simulation results of the l-Trp adsorption process on resin XDA-200 are shown in Figure . The fluctuation of root-mean-square deviation (RMSD) reveals that the adsorption systems attain equilibrium quickly. The MD trajectories show that l-Trp± and l-Trp+ diffuse quickly and then adsorb onto the resin surface. Apparently, more l-Trp± than l-Trp+ is adsorbed on the resin surface. In contrast, there is almost no l-Trp– adsorbed on the resin surface. It is noteworthy that some l-Trp– adsorbs on the resin surface when more l-Trp– is added into the adsorption system (see Figure S2). These phenomena can be explained by the relatively weak affinity between the resin and l-Trp–. The interaction analysis of the XDA-200-l-Trp system shows that the dominating interactions between l-Trp and resin are nonelectrostatic interactions (see Table ). Further conformational analysis indicates that the main interactions between the resin and l-Trp are the nonelectrostatic interaction between hydrophobic indole rings and benzene rings in the resin skeleton mainly including π–π and hydrophobic interactions (see Figure ). The potential of the XDA-200-l-Trp± system is significantly higher than that of the XDA-200-l-Trp+ system (see Table ), indicating that l-Trp± has a stronger affinity to resin XDA-200 than l-Trp+. As can be seen in Figure , l-Trp mainly exists as l-Trp± in the pH range of 4–8. Therefore, the higher adsorption amounts of l-Trp in the pH range of 4–8 than at other pH values can be attributed to the common effect of strong affinity and low solubility of l-Trp±. Although l-Trp– shows a weak affinity to resins in MD simulations, l-Trp– still shows a certain amount of adsorption under higher concentrations. This result explains that a few l-Trp molecules are adsorbed at around pH 12 (see Figure ).

Figure 6

MD simulations of XDA-200 and l-Trp adsorption systems. The nonelectrostatic interaction is indicated in purplish red.

Table 3

Interactions and Energy Analysis of XAD-200 and l-Trp Adsorption Systems

	resin-l-Trp+	resin-l-Trp±
electrostatic interaction (kJ/mol)	11.72	–57.31
nonelectrostatic interaction (kJ/mol)	–149.78	–201.19
potential (kJ/mol)	–78.08	–287.26

MD simulation results of the XDA-200-l-Glu system are shown in Figure S3. The theoretical concentration distribution of different species of l-Glu at different pH values is shown in Figure S4. As can be seen in Figure S3, there is a very small amount of l-Glu adsorbed on the resin surface. The results indicate that l-Glu has a lower affinity to the resin than l-Trp. Obviously, the MD simulation results are consistent with the adsorption equilibrium experimental results of l-Glu (see Figure ).

Adsorption Kinetics of l-Trp on Resin XDA-200

The adsorption kinetic curves of l-Trp at different pH values are shown in Figure a. As can be seen in Figure a, the adsorption of l-Trp on resin XDA-200 at pH 12 reaches equilibrium within 40 min. However, about 100 min is needed to reach adsorption equilibrium for l-Trp at pH 5.91 and 9. The phenomenon indicates that the adsorption of l-Trp at pH 12 is faster than that at pH 5.91 and 9. The adsorption kinetic curves of l-Trp were fitted by the intraparticle diffusion model established by Weber and Morris.[19] The model equation can be expressed as followswhere q is adsorption amount at time t (mmol/g) and kp is the diffusion rate parameter ((mmol/g)/min1/2).

Figure 7

(a) Adsorption kinetic curves of l-Trp; (b) fitting results of adsorption kinetic curves of l-Trp by the intraparticle diffusion model.

The fitting results are shown in Figure b. As can be seen in Figure b, there are two linear segments for each curve. In the first stage, the straight lines pass through the origin, which indicates that intraparticle diffusion is the sole rate-limiting step for l-Trp adsorption on resin XDA-200. The result is reasonable for the microporous resin XDA-200. The pore size distribution of resin XDA-200 is shown in Figure S5. Surface diffusion of adsorbates is the main way of mass transfer inside adsorbent particles for microporous adsorbents.[24] The contribution from pore diffusion of adsorbates is negligible relative to that from surface diffusion. Therefore, surface diffusion is the sole rate-limiting step for l-Trp adsorption on resin XDA-200. The surface diffusion rate of weakly retained adsorbates inside adsorbents is higher than that of strongly retained adsorbates.[25] Therefore, l-Trp– diffuses inside resin XDA-200 particles more fastly than l-Trp±. l-Trp exists mainly as l-Trp– in the aqueous solution at pH 12 (see Figure ). Therefore, the adsorption of l-Trp reaches equilibrium at pH 12 more fastly than that at pH 5.91 and 9. The fast mass transfer rate of l-Trp at pH 12 benefits the elution of l-Trp from resin XDA-200 using a NaOH aqueous solution at pH 12 as the eluant.

Separation of l-Trp in the Fixed Bed Packed with Resin XDA-200

The concentration profiles of l-Trp and l-Glu at the outlet of the fixed bed are shown in Figure a. As can be seen in Figure a, l-Glu flows out from the column before l-Trp. The phenomenon results from a weaker affinity between l-Glu and resin XDA-200 than that between l-Trp and resin XDA-200. The separation of l-Trp from l-Glu is realized satisfactorily with a resolution of 1.33. The concentration profile of Na+ and pH at the outlet of the fixed bed is shown in Figure b. When the feed solution is introduced into the fixed bed, the pH at the outlet of the fixed bed increases slowly from about 7.0 because the fixed bed is filled with deionized water initially. When the eluant (NaOH aqueous solution at pH 12) is introduced into the fixed bed (at about 30 mL), the pH at the outlet of the fixed bed increases quickly. However, with the outflow of l-Trp, the increase of pH slows down. The phenomenon can be explained by that l-Trp eluted from resins has a pH-buffering effect because l-Trp is an amphoteric compound. Na+ flows out from the fixed bed quickly because of no adsorption of Na+ on the weakly polar resin XDA-200. The satisfactory separation between l-Trp and Na+ with a resolution of 1.19 is obtained. The recovery rate and purity of l-Trp are all higher than 99%. In conclusion, adsorption separation of l-Trp and desalination based on the hyper-cross-linked resin XDA-200 are realized synchronously.

Figure 8

Concentration profiles of (a) l-Trp, l-Glu and (b) Na+ and pH at the outlet of the column.

Conclusions

The adsorption separation of l-Trp based on hyper-cross-linked resins was studied. The weakly polar resin XDA-200 is the best adsorbent for separating l-Trp because of its higher adsorption amount on l-Trp and easy elution. The resin XDA-200 has little adsorption on l-Glu. The adsorption amount of l-Trp on resin XDA-200 increases slightly when NaCl exists in the solution because of the salting-out effect. The adsorption of l-Trp on resin XDA-200 is a spontaneous process driven by adsorption enthalpy. The affinity between l-Trp± and resin XDA-200 is higher than that between resin XDA-200 and other l-Trp ions. The relatively strong affinity and low solubility of l-Trp± account for the high adsorption amount of l-Trp in the pH range of 4–8. The adsorption of l-Trp on resin XDA-200 is mainly based on π–π and hydrophobic interactions. l-Trp can be separated from l-Glu and soluble salts synchronously in the fixed bed packed with resin XDA-200. Separating l-Trp based on the hyper-cross-linked resin is a potential method to replace ion exchange. The study in this work can provide a significant reference for realizing the efficient separation of l-Trp and other ionizable organic compounds with hydrophobic groups.

Materials and Methods

Resins

The hyper-cross-linked resins used in this work were obtained from Sunresin New Materials Co. Ltd. (Xi’an, China). The physical and chemical properties of the resins are listed in Table . The pore structure of the resins was determined by N2 adsorption–desorption isotherms, which were measured at 77 K by a Micromeritics ASAP 2460 surface area and porosity analyzer. The pore size distribution of mesopores and micropores in resin XDA-200 were calculated by Barrett–Joyner–Halenda (BJH) and Horvath–Kawazoe methods, respectively. The functional groups on resins were determined by a Fourier transform infrared spectrometer (IR tracer-100, Shimadzu Corporation, Kyoto, Japan). The resins were pretreated by following steps before use. The resins were first soaked in absolute ethanol for more than 8 h to remove the residues from resin synthesis and then washed by deionized water until no ethanol was left in the resins.

Table 4

Physical and Chemical Properties of the Resins Used in This Work

resin	matrix	SBET (m2/g)	polarity	functional group
XDA-200	PS-DVBa	1018.1	weakly polar	carbonyl
XDA-1	PS-DVB	1336.9	nonpolar	none
XDA-6	PS-DVB	673.5	nonpolar	none
LX-T81	PS-DVB	877.1	weakly polar	carbonyl
LX-3020	PS-DVB	1139	strongly polar	acylamide, ester group
LX-68M	PS-DVB	1082.1	weakly polar	ester group
LX-60	PS-DVB	845.1	weakly polar	ester group
LSC-100	PS-DVB	b	strongly polar	acylamide
LSA-12	PS-DVB	647	weakly polar	aldehyde group, carbonyl
D101	PS-DVB	728.8	weakly polar	carbonyl

PS-DVB is the abbreviation of poly(styrene divinylbenzene).

The resin LSC-100 is a gel-type resin with a very low specific surface area.

Chemicals

l-Trp (purity ≥99%), l-Glu (purity ≥99%), NaHCO3 (purity ≥99.8%), sodium acetate (purity ≥99%), K2HPO4·3H2O (purity ≥99%), acetic acid (purity ≥99.5%), 2,4-dinitrofluorobenzene (purity ≥98%), and NaCl (purity ≥99.5%) were all purchased from Shanghai Macklin Biochemical Co., Ltd. (Shanghai, China). Ethanol (purity ≥99.7%) was supplied by Tianjin Tianli Chemical Reagent Co., Ltd. (Tianjin, China). Hydrochloric acid (36–38%, w/w) was purchased from China Pingmei Shenma Group Kaifeng Dongda Chemical Co., Ltd. (Kaifeng, China). NaOH (purity ≥96%) was obtained from Tianjin Jinbei Fine Chemical Co., Ltd. (Tianjin, China). Acetonitrile (chromatographic grade) was purchased from Tianjin Kemiou Chemical Reagent Co., Ltd. (Tianjin, China). The deionized water used in this work was produced by a water purifier (WP-UP-YJ-40, Sichuan Wortel Water Treatment Equipment Co., Ltd., Chengdu, China).

Screening Resin for Separation of l-Trp

The adsorption amounts of l-Trp on different types of resins were determined. Certain volumes (25 mL) of an l-Trp aqueous solution (3.43 × 10–2 mol/L) were put into several conical flasks (50 mL). A fixed amount (2 g) of different resins was transferred into different conical flasks. The flasks were sealed and shaken at 298 ± 1 K in a constant-temperature shaker (HNY-200B, Tianjin Honour Instrument Co., Ltd., Tianjin, China) for more than 12 h to attain adsorption equilibrium. The initial and adsorption equilibrium concentrations of l-Trp in the solutions were analyzed by a UV/VIS spectrophotometer. The adsorption amount of l-Trp on the resins was calculated by the following equationwhere qi,e is the equilibrium adsorption amount of solute i in the resins (mmol/g), ci,0 and ci,e are the initial and adsorption equilibrium concentrations in the bulk solutions, respectively, (mol/L), V is the volume of the bulk solutions (mL), and m is mass of the resins (g).

The elution performances of the resins (XDA-1, XDA-200, LX-3020, LX-68M, LX-T81) were compared using the above l-Trp-loaded resins. The solvent was removed from the resins by filtration. Then, the resins were contacted with 25 mL of an aqueous NaOH solution at pH 12 in different conical flasks. The other operating conditions were identical to those of the above adsorption experiments. The equilibrium concentrations of l-Trp in the aqueous NaOH solutions were determined by a UV/VIS spectrophotometer. The recovery rate f of l-Trp was calculated by the following equationwhere q is the initial adsorption amount of l-Trp on the resins. The values of q are equal to those of q in the above adsorption experiment.

Salt Resistance and Adsorption Selectivity of Resin XDA-200

The salt resistance and adsorption selectivity of resin XDA-200 were studied at different pH values. NaCl was selected as the representative of soluble salts because of the significantly higher concentration of NaCl than that of other salts in the fermentation broth of l-Trp. First, the adsorption equilibrium experiments of l-Trp including no NaCl in the feed solution were carried out at different pH values. An l-Trp aqueous solution (25 mL, 3.43 × 10–2 mol/L) was put into several Erlenmeyer flasks (50 mL). The solutions were adjusted to different pH values. Then, 2 g of the resin XDA-200 was transferred into the Erlenmeyer flasks and contacted with the l-Trp aqueous solutions. The other operating steps were identical to those of the above adsorption equilibrium experiments for selecting the resin. Then, the identical adsorption equilibrium experiments were carried out, except that about 0.63 mol/L NaCl existed in l-Trp aqueous solutions.

The adsorption equilibrium experiments of l-Glu at different pH values were carried out using similar operating steps to those of l-Trp. The initial and equilibrium concentrations of l-Glu in the solution were determined by high-performance liquid chromatography (HPLC).

Adsorption Isotherms of l-Trp at Different Temperatures

l-Trp aqueous solutions (25 mL) at different concentrations were contacted with 2 g of wet resin in several Erlenmeyer flasks (50 mL). The solution pH values were all adjusted to about 8.0 using an aqueous solution of NaOH. The flasks were kept on a shaker (HNY-200B, Tianjin Honour Instrument Co., Ltd., Tianjin, China) and shaken at 150 rpm and 298 ± 1 K for more than 12 h to attain adsorption equilibrium. The initial and equilibrium concentrations of the solutions were measured by a UV/VIS spectrophotometer. The adsorption amount of l-Trp on the resin was calculated by eq . The adsorption amount qi,e was plotted against the equilibrium concentration in bulk solutions ci,e to generate the adsorption isotherm at 298 K.

The adsorption isotherms of l-Trp at other temperatures (288 and 308 K) were measured using the same method as that for the above experiments.

Molecular Dynamics Simulations

The PDB coordinates of resin XDA-200 monomers, l-Trp, and l-Glu were sketched using the Discovery Studio Visualizer 4.5 (DSV). Molecular topologies of resin XDA-200 for Gromacs were generated using the PRODRG2 online server.[26] The molecular dynamics (MD) simulations were carried out using the Gromacs 2019 package.[27] The aggregates of XDA-200 monomers (see Figure S6) were used as the resin model in MD simulations of the adsorption process. The simulations for the systems of XDA-200, XDA-200-l-Glu, and XDA-200-l-Trp were calculated for 10, 5, and 5 ns, respectively, using the Gromos96 force field in combination with the extended simple point charge (SPCE) water model. The resin model was first immersed in a box for a presimulation, and then the specified number of amino acids were added to the system. System stability was verified by the root mean square deviation (RMSD) based on the system trajectory during the simulations. The particle mesh Ewald (PME) method was used to evaluate the electrostatic interactions. All other settings were kept at their defaults. The simulation trajectory was visualized using Visual Molecular Dynamics (VMD) and DSV. All simulations were performed on the Lenovo Think Station T910 (Lenovo, China).

Adsorption Kinetics of l-Trp at Different pH Values

An l-Trp aqueous solution (250 mL, 3.43 × 10–2 mol/L) at pH 12 was contacted with 10 g of wet resin XDA-200 in a round-bottom flask (500 mL). The flask was kept at 298 ± 1 K in a thermostat water bath. The solution in the flask was stirred by an agitator blade at 150 rpm. Several samples were taken using injectors at preset time intervals. The concentrations of the samples were determined by a UV/VIS spectrophotometer. The adsorption amounts of l-Trp at different times were calculated by eq .

The adsorption kinetics experiments of l-Trp at other pH values were carried out using the identical operating steps as those of the above experiments.

The mixed solution of l-Trp, l-Glu, and NaCl was introduced into the fixed bed in a downflow mode. The concentrations of l-Trp, l-Glu, and NaCl in the feed solution were about 5.88 × 10–2, 9.52 × 10–3, and 0.63 mol/L, respectively. The pH of the feed solution was about 8.0. The glass column was 1 cm in inner diameter. Also, the filling length of resin XDA-200 in the column was 15 cm. After 20 mL of the feed solution was introduced completely, an aqueous solution of NaOH at pH 12 was introduced into the column to elute l-Glu and l-Trp. The flow rate was kept at 0.4 mL/min by a constant-speed peristaltic pump (BT100-1L, Hebei, China). Several samples were collected at the outlet of the column at predetermined time intervals. The concentrations of l-Trp and l-Glu in the samples were determined by a UV/VIS spectrophotometer and HPLC, respectively.

Analytical Methods

The pH of solutions was measured by a laboratory pH meter (FE28, Mettler-Toledo, LLC, Zurich, Switzerland). The concentration of l-Trp in its aqueous solutions was determined by a UV/VIS spectrophotometer (BioSpectrometer, Eppendorf AG, Hamburg, Germany) at a wavelength of 218 nm. The concentration of Na+ was determined by a Na+ densimeter (DWS-295F, Shanghai Instrument Electric Science Instrument Limited by Share Ltd., Shanghai, China).

The concentration of l-Glu in its aqueous solutions was determined by an HPLC system (LC-20AT, Shimadzu Corporation, Kyoto, Japan) equipped with a Shim-pack GIST C18 column (250 × 4.6 mm2, 5 μm, Shimadzu Corporation, Kyoto, Japan) through a gradient elution. The column was kept at 306 K, and the total flow rate of the mobile phase was 1 mL/min throughout the measurement. The mobile phase consisted of solution A (aqueous acetonitrile solution at a concentration of 50%) and solution B (aqueous sodium acetate solution at a concentration of 4.1 g/L and pH 6.4 adjusted by acetic acid). The entire run time was 35.5 min. The gradient of the mobile phase is listed in Table S1. The detector wavelength (UV) was 360 nm. The sample was prepared by precolumn derivatization. Samples (10 μL) were placed into a 2 mL centrifuge tube. Then, 200 μL of a derivatization buffer solution (aqueous NaHCO3 solution at a concentration of 42 g/L) and 300 μL of a derivatization reagent (acetonitrile solution of 2,4-dinitrofluorobenzene at a concentration of 10 g/L) were added to the centrifuge tube successively. The mixed solution was fully mixed. The centrifuge tube was heated in dark for an hour in a 333 K water bath. Then, the solution was cooled down to room temperature. The volume of the solution was adjusted to 1.2 mL by adding a constant-volume buffer solution. The mixed solution after filtration by a microfiltration membrane was the sample for HPLC. The constant-volume buffer solution was prepared by dissolving 3.4 g of K2HPO4·3H2O using 0.1 mol/L NaOH and then adjusting the volume of the solution to 500 mL by adding 0.1 mol/L NaOH.

Theory

Adsorption Isotherm Models

Langmuir and Freundlich isotherm models[28,29] were widely used to describe the adsorption isotherms. The Langmuir model supposes that the adsorption is a monolayer process, while the Freundlich equation describes the adsorption on a heterogeneous surface. The Langmuir and Freundlich model equations are as followswhere qm is the maximum monolayer adsorption capacity of adsorbates (mmol/g), K is the Langmuir constant (L/mol), and KF (L/mol) and n are the Freundlich constants.

The Langmuirian equilibrium constant RL is an essential parameter. The constant can be expressed by the following equation[21]

The values of RL indicate that the adsorption isotherm of adsorbates is irreversible (RL = 0), favorable (0 < RL < 1), linear (RL = 1), or unfavorable (RL > 1).

Resolution

The resolution RS is calculated by the following equation[24]where tR1 and tR2 are the retention times of components 1 and 2, respectively (component 1 flows out first from the outlet of the chromatographic column) and w1 and w2 are the peak widths of components 1 and 2, respectively.

A resolution of 1.5 means the realization of baseline separation of two components. A resolution of 1.0 means the almost overlap of 3% of two adjacent chromatographic peaks.