Dissociation mechanics and stability of type A botulinum neurotoxin complex by means of biophysical evaluation

Biophysical characterisation of type A botulinum neurotoxin (BoNT/A) could be challenging since it exists in association with neurotoxin associated proteins (NAPs) as large protein complexes. The objective of this study was to elucidate the dissociation mechanics of BoNT/A complex along with its thermodynamic stability through a combination of analytical tools. Size exclusion chromatography (SEC) was mainly used to study the behavior of BoNT/A complex at various pH. In addition, multi-angled light scattering (MALS), enzyme linked immunosorbent assay (ELISA), and dynamic light scattering (DLS) were utilized to validate its chromatographic results. The dissociation of BoNT/A from its complex was found to be strongly dependent on its pH with higher dissociation towards alkaline pH which was further accelerated with time and temperature. In addition, dissociated BoNT/A at pH 7.4 showed lower thermal stability compared to the associated state even in the presence of polysorbate, as revealed by the SEC chromatogram and aggregation onset temperature. Moreover, the partial reversibility of the dissociated BoNT/A complex after titration of pH to 6.0, suggested vulnerability of BoNT/A towards formation of irreversible aggregates as it dissociates, signifying lower stability profile of the neurotoxin after dissociation. Overall, BoNT/A was more stable when associated with NAPs at pH 6.0 compared to its dissociated state at pH 7.4. Moreover, the conventional analytical used could be utilized to relatively quantify its amount in different formulations.


Introduction
Type A botulinum neurotoxin (BoNT/A), a 150 kDa protein, is comprised of a 50 kDa light chain and a 100 kDa heavy chain, covalently linked by an inter-chain disulfide bond (Montecucco C & Schiavo 1995;Tighe and Schiavo 2013). It is generally produced in association with a group of neurotoxin-associated proteins (NAPs), comprising a non-toxic non-hemagglutinating subunit (NTNHA) and hemagglutinin subunits (HAs) HA-17, HA-33, and HA-70, yielding the final molecular weight of ~ 900 kDa (Bryant et al. 2013). The majority of botulinum neurotoxin products available on the market are in the form of purified neurotoxin complexes, except for Xeomin®, which includes a 150 kDa neurotoxin without the complexing proteins (Frevert 2009). Meanwhile, as reported earlier, NAPs are known to enhance the structural stability (Brin 2009;Wortzman and Pickett 2009) and activity of the neurotoxin (e.g., BoNT/A) as well (Sharma and Singh 2004;Ghosal et al. 2018). Moreover, its stabilizing effect against digestive proteases and acidic environment in the gastrointestinal tract supports in maintaining its oral toxicity (Sakaguchi 1982;Sharma and Singh 1998;Lee et al. 2013;Lam and Jin 2015). Nevertheless, the pharmacological effect is dependent on the free 150 kDa neurotoxin dissociated from the complex as it approaches the pH of around 7.4 in the circulation (Eisele et al. 2011;Tighe and Schiavo 2013).
Marketed products of botulinum neurotoxin formulated as vacuum-or freeze-dried powder such as Botox®, Dys-port® and Xeomin® exists with the reconstituted pH values of approximately 7, except for a ready to use solution Myobloc® (approximately pH 5.6) (Pirazzini et al. 2017;Fonfria et al. 2018). Existing research states that neurotoxins readily dissociate from the complex within a minute at the physiological pH (around pH 7.4), indicating the existence of free form of 150 kDa botulinum neurotoxin in most products after reconstitution (Frevert and Dressler 2010;Nestor et al. 2020). Yet several investigations claim the protective activity of NAPs on the neurotoxin in associated state against extreme pH conditions (Sharma and Singh 2004;Kukreja and Singh 2007;Chellappan et al. 2014). Moreover, studies suggest pH and ionic strength as major driving forces for dissociation of neurotoxin from its complex (Eisele et al. 2011;Matsui et al. 2014). Since, pH is vital for both stability and activity of protein therapeutics, it is critical to understand its characteristics based on different pH conditions, especially in the formulation development field (Maity et al. 2009;Chaudhuri et al. 2014).
On the other hand, all products mentioned contain human serum albumin (HSA) in the formulation to protect BoNT/A from nonspecific binding (Kutschenko et al. 2019). It stabilizes the neurotoxic proteins during and after manufacturing, suppressing the aggregation of neurotoxic proteins (Malhotra et al. 2003). However, the inclusion of HSA introduces the risk of viral and other pathogens transmission since it is derived from humans along with the difficulty in maintaining a uniform quality (Kim et al. 2021). To counter this risk, recombinant HSA could be an alternative. However, challenges still exist in the analytical method development, as the selective analysis and quantification of the active proteins and their degradation products or aggregates are often difficult in presence of the secondary protein in the form of HSA. For the reasons, the market demands HSA-free formulations of botulinum toxins. Consequently, the stability of the toxins and the mechanism of its dissociation should be investigated thoroughly.
In fact, the existence of the neurotoxin as a complex with NAPs poses challenges, as each of the proteins in the complex portrays individual behavior in different solution conditions. Given their complexity, different methods intended to investigate their similar attributes are often necessary to provide independent confirmation of the protein properties. Enzyme-linked immunoassay (ELISA) has been widely used as a highly sensitive quantification tool for neurotoxins (Keller 2008;Kim et al. 2008;Stanker et al. 2008Stanker et al. , 2013. However, the process is limited to detection and quantification of neurotoxin regardless of its associated or dissociated state as the neurotoxins are exposed to reagents at pH 7.2-7.4 during the analysis. For size exclusion chromatography (SEC), there is no relevant report of its use in quantifying and discriminating associated and dissociated neurotoxin. Generally, this analysis has been conventionally used for the quantification of proteins in various states as monomers, oligomers, aggregates, and even fragments. The major benefit is the use of elution conditions that allow the characterization of proteins with the least impact in their local environment and conformational structure (Fekete et al. 2014). In this study, the state of neurotoxin was investigated mainly using SEC to evaluate the dissociated BoNT/A from NAPs at different pHs, incubation times, and storage temperatures. Moreover, the impact of polysorbate 20 is investigated since it has been used as a stabilizer in marketed formulations and can be an excipient as a substitute for HSA (Kwak et al. 2020;Lee et al. 2020;Kim et al. 2021). Additionally, multiangle light scattering (MALS) and ELISA were utilized further to interpret the SEC chromatograms. Lastly, dynamic light scattering (DLS) was used to access the size distribution and zeta potential of the complexes at different pHs.

Materials and sample preparation
Type A botulinum neurotoxin complex (with an approximate molecular weight of 900 kDa) was supplied by JETEMA Co., Ltd. (Seoul, South Korea) as a 1.14 mg/mL solution in 50 mM sodium phosphate buffer with 150 mM NaCl at pH 6.0 stored at − 80 °C. HSA, sodium phosphate dibasic dihydrate, sodium phosphate monobasic monohydrate, sodium chloride, and polysorbate 20 were purchased from Sigma-Aldrich (St. Louis, MO, USA). All of the other reagents used were of analytical grade. Prior to analysis, all prepared samples were filtered through a sterile Spin-X 0.22 μm cellulose acetate centrifuge tube filter (Costar, Corning Incorp., Salt Lake, UT, USA) centrifuged at 8000 rpm for two minutes.

SEC analysis
SEC analysis was performed using the Shimadzu LC 20 HPLC system (Shimadzu Corporation, Kyoto, Japan). PRO-TEIN KW-804 (7 μm, 8.0 mm × 300 mm) (Showa Denko, Tokyo, Japan) was used as a column with the temperature set at 25 °C, while a mobile phase comprised of 50 mM sodium phosphate buffer pH 6.0 with 150 mM NaCl was used for the elution at a flow rate 1 mL/min. The injection volume was 60 µL, and the UV absorbance was measured at 278 nm. For the dissociation and stability study of the BoNT/A, the samples were incubated at the temperatures of 25 and 37 °C using the Eppendorf ThermoStat C (Eppendorf AG, Hamburg, Germany) while 0.01% polysorbate 20 was added to the samples to prevent the loss due to interfacial adsorption. The baseline obtained from the buffer chromatogram was subtracted from the sample's chromatogram to remove the interference of polysorbate 20 peak elution. The relative area percentage of the respective peaks were then calculated using the following formula: where, 'A t ' is the area of the respective peak at time 't' and 'A 0 pH6.0 ' is the initial area of peak 1 at pH 6.0 referred to as the undissociated state. The area percentage of aggregates was calculated with respect to the total peak areas of the same sample at the respective time.

Molecular weight determination by MALS
The molecular weight of the separated peaks from SEC was determined using a miniDawn TREOS light scattering system (Wyatt Technology Corp., CA, USA) equipped with a three-angled (43.6°, 90° and 136.4°) detector and a 685.0 nm laser beam. The molecular weight was processed through the software ASTRA 7 (version 7.1.3.15) and was calculated based on the dn/dc (0.185) value of the BoNT/A complex.

Neurotoxin quantification by ELISA
The collected fraction from the SEC was assayed for the investigation of neurotoxin content using a Botulinum Neurotoxin Type A DuoSet® ELISA kit (R&D Systems, Inc., Minneapolis, USA, DY008) after 500-fold dilution.
BoNT-A standard supplied within the kit (DY4489-05) was used to obtain the calibration curve. Botulinum antitoxin equine type A specific for botulinum neurotoxin was used as a capture antibody, and Streptavidin-HRP (streptavidin conjugated to horseradish-peroxidase) was incorporated as a detection reagent with the minimum detectable concentration at 0.39 ng/mL. The plate was incubated at 25 °C at each step. The 96-well microplate was first coated with the plate coating buffer containing the capture antibody and incubated for 16-18 h. Once coated, 100 µL of 1× reagent diluent was dispensed into each well and incubated for 2 h. Then, 100 µL of standard and samples were dispensed into the designated wells. After 2 h of incubation, 100 µL of streptavidin-HRP was loaded into the microplate well followed by incubation at room temperature for 20 min. The plates were washed three times after each step with the diluted wash buffer provided in the kit. Successively, 100 µL of color reagent was introduced into the wells. Finally, after 20 min, 50 µL of stop solution was added into all wells to end the enzymesubstrate reaction. The absorbance was measured using a SpectraMaxM2 spectrophotometer (Molecular Devices, Sunnyvale, USA) at 450 nm. The absorbance of the blank was subtracted from the resulting absorbance of the respective samples. The standard curve was fitted using a fourparameter logistic curve fitting method and the concentrations were determined using the equation derived from the curve fitting. The final BoNT/A concentration was calculated considering the dilution factor and total volume of the collected fraction.

Zeta potential and size distribution measurement by DLS
Zeta potential measurements were carried out using a Zetasizer Nano ZS90 (Malvern Instruments, Worcestershire, UK). To avoid the corrosion of the electrodes in the cell, the sample was desalted by dialyzing 0.25 mg/mL BoNT/A complex sample against 50 mM sodium phosphate buffer pH 4.8 using a 0.1 mL volume 10,000 MWCO Slide-A-Lyzer™ mini dialysis device (Thermo Fisher Scientific Solutions Co., Ltd., Waltham, MA, USA). The dialyzed samples were then loaded into the disposable folded capillary cells (DTS1070, Malvern Instruments Ltd., Worcestershire, UK) and the zeta potential was scanned from acidic to alkaline pH through manual titration. NaOH of 1 N and 0.1 N were used to adjust the pH. Additionally, the hydrodynamic size of 0.5 mg/mL BoNT/A complex was monitored with increasing temperature starting from 20 to 80 °C at 2 °C/step. First, 100 µL of sample was loaded into the QS High precision cell 3 × 3 mm (Hellma Analytics, Müllheim, Germany) and the samples were equilibrated at the respective temperatures for 60 s. All the measurements were repeated three times. Zetasizer software version 7.11 was utilized to achieve the parameters from the auto-correlated function. Figure 1a represents SEC chromatograms of the BoNT/A complex at pH 6.0, pH 7.4, and pH 8.0 analyzed after an hour incubation at 25 °C. A dominant peak (peak 1) was observed at all three pH levels, followed by an additional peak at retention time around 10 min (peak 2) at pH 7.4 and 8.0. The relative area of peak 1 was reduced approximately by 5% and 8% at pH 7.4 and 8.0, respectively. Additional analysis through MALS identified the molecular weights of peak 1 as 834.7, 743.6 and 578 kDa at pH 6.0, 7.4 and 8.0, respectively (Fig. 1a). The decrease in molecular weight along with a shift in retention times and appearance of the second peak with increasing pH, could indicate the dissociation of protein components from its complex at higher pH values.

pH-based dissociation of BoNT/A complex in SEC
The pH dependent assembly/disassembly of neurotoxin complexes has been confirmed previously in several research (Chen et al. 1998;Eisele et al. 2011;Matsui et al. 2014). Furthermore, high level of stability (i.e. stability of association) between the neurotoxin and NTNHA at pH 6.0 with a dissociation constant of ~ 30.8 nM and no detectable interaction at pH 7.5 observed formerly , suggests peak 2 to be the dissociated BoNT/A owing to the basic solution pH. However, the molecular weight of peak 2 was not detectable by MALS due to the low signal intensity.
To verify the peak 2 as dissociated BoNT/A, a more specific analytical tool, ELISA was applied on the fractions of SEC, since it allows for the selective analysis of BoNT/A and is more sensitive compared to SEC. The Duo-Set® ELISA development system used in this experiment is known to selectively detect and quantify 150 kDa BoNT/A existing either in free or complex form. As shown in Fig. 1b, a relatively larger portion of neurotoxin was detected at peak 1 at pH 6.0 indicating the greater presence of BoNT/A in an associated state. On the contrary, the ELISA result at pH 7.4 showed the dominant presence of BoNT/A in the fraction of peak 2, signifying the dissociation of a high amount of toxin from the NAPs as it is exposed to the basic pH. Hence, it can be established that peak 1 represents the overlay of 'BoNT/A complex' and 'neurotoxin free complex' peaks, whereas peak 2 represents 'free 150 kDa BoNT/A'.
To further investigate the concentration dependent behavior of the dissociated BoNT/A from its complex, different concentrations of the BoNT/A complex were analyzed from 0.01 to 0.5 mg/mL at pH 7.4 (Fig. 1c). To minimize the loss of dissociated BoNT/A through interfacial adsorption, polysorbate 20 was added to each solution. The area under the curve (AUC) of peak 1 was found to be linear with an R 2 value of 0.9973. In comparison, peak 2 showed less linearity (R 2 = 0.9676). The dissociation of the BoNT/A from the complex was previously found to be dependent on the complex's concentration with a high amount of dissociation at a lower concentration (Eisele et al. 2011;Benefield et al. 2013), which could result in the deviation of its linearity in respect to the complex's concentration. Consequently, the behavior of higher dissociation at lower concentration would be related to degree of its activity coefficient supporting that the dissociated neurotoxin would be ionic. Besides, limit of detection (LOD) and limit of quantitation (LOQ) for peak 2 were found to be 22 ng/mL and 65 ng/mL of the BoNT/A complex, respectively.

Reversibility of the BoNT/A association
While the current study along with earlier studies confirms the dissociation of neurotoxin at basic pH, limited information is reported on its reversibility to the original complex once reexposed to acidic pH. Matsui et al. introduced the favorable formation of a complex of BoNT/A and NTNHA at acidic pH using small-angle X-ray scattering (Matsui et al. 2014). Sagane et al. demonstrated the reversibility of a dissociated HA subcomplex into a neurotoxin complex using SDS-PAGE (Sagane et al. 2017). Herein, SEC was utilized to assess the reversibility of BoNT/A with NAPs as shown in Fig. 2. To evaluate its reversibility, sample solution incubated for an hour at pH 7.4 was titrated back to pH 6.0 using 0.6 M hydrochloric acid. Interestingly, the SEC chromatogram of the pH-reversed sample showed the disappearance of peak 2 compared to pH 7.4 (Fig. 2a). Moreover, the area of peak 1 was increased, suggesting the re-association of dissociated BoNT/A into its complex form at acidic pH. The calculated AUC-based [i.e., peak 1; [(AUC pH7.4 )/(AUC pH6.0 )] × 100] reversibility was 81.01%. To further confirm the reversibility of the complex, the collected SEC fraction was analyzed by ELISA. Supportively, the neurotoxin at peak 1 was recovered after the pH was adjusted back into 6.0 (Fig. 2b). The calculated BoNT/A-based reversibility 1 3 Fig. 1 Overlaid SEC chromatograms of 0.25 mg/mL BoNT/A complex at pH 6.0, 7.4 and 8.0 with the representation of molar mass (kDa) obtained through MALS (a). Concentration of BoNT/A in the collected SEC fractions from peak 1 and peak 2 at pH 6.0 and 7.4 investigated through ELISA (b).Overlaid SEC chromatograms of BoNT/A complex at pH 7.4 in the concentration range 0.1-0.5 mg/mL (c) was around 84.45%. In summary, the propensity of the two calculated reversibility was similar and the value itself suggests high reversibility to form the initial complex. The relative loss around 16-19% could be due to the formation of insoluble/ irreversible submicron or even larger particles however it was not considered further in this study.

Effect of time and temperature on the dissociation of BoNT/A
The time-dependent dissociation of the BoNT/A from its complex at pH 7.4 was investigated by incubating the sample at 25 °C for 96 h ( Fig. 3a and b). Compared to its undissociated state at pH 6.0, 96% of the relative AUC of peak 1 was retained until 1.5 h. It decreased to around 87% and 80% after 24 and 96 h respectively with an increase in the relative AUC of peak 2. Consequently, the dissociation occurred in a time dependent manner.
To further investigate the behavior of its dissociation and to accelerate the reaction, the incubation temperature was increased to 37 °C in both pH 6.0 and 7.4 for 14 days ( Fig. 3c and d). As a result, relatively rapid dissociation was observed at pH 7.4 compared to the sample incubated at 25 °C. However, a new peak arose at a shorter time as the Fig. 2 Overlaid SEC chromatograms of 0.25 mg/mL BoNT/A complex at pH 6.0, 7.4 and pH 6.0 titrated back from pH 7.4 after an hour incubation at 25 °C (a). Concentration of BoNT/A in the collected SEC fractions from peak 1 and peak 2 of the sample titrated back from pH 7.4 to pH 6.0 investigated through ELISA (b) area of peak 1 and 2 decreased, indicating the formation of soluble aggregates (< 100 nm) at both pH 6.0 and 7.4. This demonstrates that increasing the incubation temperature from 25 to 37 °C not only accelerates the dissociation but also induces protein unfolding or partial unfolding of any subunits in the complex and even aggregation of free BoNT/A. In comparison, the aggregates increased from 14% on the 3rd day to 31% on the 14th day at pH 7.4, while it was only around 6% on the 14th day at pH 6.0, suggesting higher thermodynamic (i.e., conformational) stability at acidic pH. On the other hand, the limited dissociation of BoNT/A at pH 6.0 along with the lower level of aggregation, suggests that a stronger association with NAPs could increase its stability against thermal stress. Nevertheless, the result additionally demonstrates the importance of controlling the temperature for in vitro tests since it causes separate aggregation kinetics rather than just dissociation reactions.

Isoelectric point and aggregation onset of BoNT/A
Zeta potential and hydrodynamic properties of proteins are used as an important physical parameter to determine their colloidal stability (Strickler et al. 2006;Gribenko and Makhatadze 2007). To better comprehend the relationship of the surface charge properties and aggregation behavior of the BoNT/A complex with the pH and temperature, DLS was employed to measure its zeta potential and hydrodynamic size distribution (i.e., Z-average size).
The isoelectric point (pI) of the BoNT/A complex in solution was found to be at pH 5.69 (Fig. 4a), where the net charge is close to zero. It has been suggested that the pI of the individual subunits within the complex should be considered to elucidate its behavior (Leney 2019). However, the information about the overall surface charge could be beneficial in interpreting the net stability of the complex. BoNT/A and NTNHA possess a large solvent accessible area through multivalent interfaces, which might make the interlocked interface sensitive to pH changes . Considering the whole complex as a single unit, as the pH moves further away from the pI toward a basic pH, more deprotonations occur along with an increasing number of negatively charged residues. The apparent zeta potential of the complex shifted from − 3 mV at pH 6.0 to approximately − 12 mV at pH 7.4. As previously elucidated, the positively charged residues at pH 6.0, particularly aspartate and glutamate present on the interface of the complex, deprotonate BoNT/A at different time periods. SEC chromatograms of BoNT/A complex incubated at 37 °C in pH 6.0 (c) and pH 7.4 (d) at pH 7.4, generating repulsive interactions with negatively charged NTNHA and finally inducing the dissociation of the complex . On the other hand, a high aggregation tendency was observed at pH 7.4 through SEC. Ideally, the protein-protein repulsion should be higher at pH 7.4 since it is far from the pI, reducing the probability of aggregation. However, in case of BoNT/A complex, the dissociation resulted in individualization of behavior of each dissociated component. Exposed free binding sites of the proteins after dissociation might have led to hydrophobic interactions forming high molecular weight species.
To confirm the aggregation behavior of BoNT/A complex at different pHs, Z-average size was analyzed with increasing temperature (Fig. 4b). Prior to heating, the Z-average size of the BoNT/A complex was found to be 27.19 nm, 26.12 nm, and 26.20 nm at pH 6.0, 7.4 and 8.0, respectively. Since the Z-average size provides a measure of the average particle size distribution, the initial reduction in the size as the pH shifts from 6.0 to 7.4 and 8.0 might specify dissociation of the neurotoxin leading to a decrease in its average size distribution. This result additionally supports the shift in the retention time detected in the SEC chromatogram toward an alkaline pH. Upon heating, the complex at pH 7.4 and 8.0 exhibited a similar sharp transition in the Z-average size with the aggregation onset (T onset ) at around 52 °C. As observed for several large proteins, the temperature induced unfolding of the BoNT/A complex resulted in the formation of larger aggregates (> 1000 nm). On the contrary, pH 6.0 exhibited a slower transition curve with the T onset of 54 °C. Comparable to the earlier SEC result, a slower aggregation tendency was observed at pH 6.0. The similar thermodynamic behavior of the BoNT/A complex was also noted in earlier reports with T m s of 58.5 °C, 56.7 °C and 55.9 °C at pH 6.0, 7.0, and 8.0, respectively, using circular dichroism (CD) (Brandau et al. 2007). Evidently, the overlaid size distribution at 50 and 54 °C clearly depicts the state of the complex with temperatures ( Fig. 4c and  d). These result support that pH 6.0 is thermodynamically more stable than pH 7.4 and that the increased stability is the result of the strong interaction between BoNT/A and NAPs. As BoNT/A dissociates from the complex, its vulnerability toward aggregation increases, which might in turn decrease the activity required for a proper therapeutic application.

BoNT/A complex vs. free BoNT/A: role of NAP in stability and future perspective
The present work supports the earlier studies that BoNT/A within its complex at acidic pH is structurally more stable against heat stress than in a free state (Brandau et al. 2007;Chellappan et al. 2014;Kukreja and Singh 2007). Most of the marketed lyophilized formulations exist in a complex form with a molecular weight ranging from 500 kDa in Dys-port® to 900 kDa in Botox® (Fonfria et al. 2018). Yet, their reconstitution pH value approaching the physiological pH (around 7.4) increases the probability of dissociation into the free 150 kDa neurotoxin. It has been demonstrated that the domains of botulinum neurotoxins are highly sensitive to even mild agitation (Toth et al. 2009). Therefore, the dissociation followed by aggregation of BoNT/A at neutral or alkaline pH could be the main reason for the necessity of HSA as a stabilizer. Especially in the product Xeomin® (BoNT/A without NAPs) where twice the amount of HSA in Botox® has been used. However, the risk of viral and pathogens transmission on using HSA, and challenges in the analysis of the BoNT/A in their presence may require albumin-free formulations in the pharmaceutical market. Moreover, SEC results suggest that the inclusion of polysorbate 20 might still be insufficient in preventing the aggregation tendency due to strong hydrophobic interactions especially at elevated temperature suggesting the requirement of more efficient stabilizer. In such a scenario, formulating the BoNT/A complex products at acidic pH can minimize the extent of hydrophobic interactions and decrease the aggregation tendency as represented in Fig. 5. In summary, the analytical tool kit (i.e., SEC, ELISA, and DLS) utilized in this study was very useful in elucidating the biophysical properties of the botulinum complex in different solutions as well as for providing a key to developing an in vitro release test method.

Conclusion
The biophysical characterization of BoNT/A complex along with its thermodynamic stability was assessed through a combination of conventional analytical tools for biopharmaceuticals. The key findings of this study are: (i) the time, temperature, and pH-dependent dissociation of the neurotoxin from its complex, (ii) the vulnerability of the dissociated BoNT/A toward aggregation compared to its associated state within the complex, and (iii) the discovery of the reversible association of the BoNT/A with NAPs when the pH is titrated to acidic pH. Overall, the data suggested that the neurotoxin is in a more stable state when it is associated with NAPs at pH 6.0 compared to its dissociated state at pH 7.4. However, it should be noted that NAPs alone cannot solely replace the use of stabilizers.

Conflict of interest
The authors declare no conflicts of interest.

Statement of human and animal rights
This article does not contain any studies with human and animal subjects performed by any of the authors.