Selection of a suitable alkyl primary amine and a proper organic solvent
The designed method to quantify molar equivalents of NHS esters of derivatives of mPEGs in preparations requires a suitable alkyl primary amine, a proper organic solvent, an optimized concentration of the alkyl primary amine in the organic solvent and stoichiometry for the reaction between the primary amine and NHS esters of mPEGs. In appearance, any alkyl primary amine, any organic solvent compatible with NHS esters of mPEGs may be applicable. However, the consecutive reaction with TNBS to quantify the residual alkyl primary amine requires alkaline aqueous solutions, in which most alkyl primary amines and some organic solvents are incompatible. This complicated situation requires delicate selection of an organic solvent and an alkyl primary amine. Glycylglycine surely has unfavorable solubility in common organic solvents. Methylamine, ethylamine and propylamine are easily evaporated under room temperature. NHS esters have negligible reactivity with alcohol [15]. Hence, ethanolamine, glycine, ethylenediamine, n-butylamine with reasonable solubility in water were tested as candidate amines, while tetrahydrofurane (THF) and dimethylformamide (DMF) of unlimited solubility in water were compared as candidate organic solvents.
Those alkyl primary amines produced similar absorbance spectra after reaction with TNBS at 25°C for 120 min in alkaline aqueous solutions (Figure 1). To quantify an alkyl primary amine with TNBS, the absorbance at 420 nm was measured, but the conjugates of tested amines and TNBS had two absorption peaks, one around 420 nm while another around 350 nm. The conjugates of ethanolamine and n-butylamine with TNBS all had absorbance at 420 nm no less than that around 350 nm while that of glycine had slightly stronger absorbance at 350 nm. The situation with ethylenediamine is complicated. The conjugate of excessive ethylenediamine with TNBS had stronger absorbance at 420 nm than that at 350 nm, but ethylenediamine can not be quantified in this case. With TNBS in molar excess, both amino groups on ethylenediamine were conjugated with TNBS but the conjugate had stronger absorbance around 350 nm than that at 420 nm. These results made ethylenediamine unfavorable to quantify molar equivalents of NHS esters of mPEGs by the new method. On the other hand, in alkaline aqueous solution at 37°C, ethanolamine, glycine, ethylenediamine and n-butylamine displayed reaction rates with TNBS in a descent order (Figure 2). Solubility in water of ethanolamine or ethylenediamine ranks the first, that of glycine acts as the second while that of n-butylamine is the lowest. Solubility in THF and DMF of ethanolamine, ethylenediamine and n-butylamine is comparable, but that of glycine is too low for this new method to quantify molar equivalents of NHS esters of mPEGs. Hence, ethanolamine may be a suitable alkyl primary amine for the new method.
Both THF and DMF have no reactivity to NHS esters of mPEGs, and they are also good solvents for ethanolamine. When NHS carbonate ester of mPEG5k (NHS-CB-mPEG5k), or NHS ester of succinic monoester of mPEG5k (NHS-SC-mPEG5k), was used at a concentration over 1.0 mmol/L in THF plus 2.0 mmol/L ethanolamine, there were white sticky precipitates after reactions (Additional file 1: Figure S1, Supporting Material). But after the reaction mixture was diluted with water, such white precipitates disappeared. The mixture of ethanolamine and NHS at concentrations over 1.0 mmol/L in THF also produced such cloudy precipitates, but the mixture of either ethanolamine or NHS alone at the same concentration in THF produced no precipitates. However, the mixtures of NHS and ethanolamine at concentrations smaller than 60 mmol/L in DMF produced no such sticky precipitates. Hence, DMF is a suitable organic solvent for the new method. Ionic complexes might be formed between NHS (pKa, aq = 6.1) and ethanolamine (pKa, aq = 9.5) and precipitate in THF.
The reaction mixture of commercial DMF with TNBS produced negligible absorbance at 420 nm. The conjugate of TNBS with any of the tested primary amines has millimolar absorptivity over 10 (mmol/L·cm)-1 (Figure 2) [18–20]. Assuming the measurable absorbance below 1.500, the highest concentration of a suitable alkyl primary amine in DMF should be below 1.7 mmol/L if the reaction mixture is diluted by 21-fold for reaction with TNBS in an alkaline aqueous solution (the addition of just 50 μL DMF solution of ethanolamine and an NHS ester of mPEG as the sample to a total of 0.95 mL borate buffer plus 50 μL 0.4% aqueous solution of TNBS to quantify residual ethanolamine). For measuring absorbance below 1.200 at 420 nm with common spectrophotometers, final 0.83 mmol/L ethanolamine is employed in DMF to react with NHS esters of mPEGs for the new method.
To quantify molar equivalents of NHS esters of mPEGs by this new method, there should be ethanolamine in molar excess to NHS esters of mPEGs. However, the use of ethanolamine to react with NHS esters of mPEGs faces the potential production of esters to complicate stoichiometry between NHS esters of mPEGs and amino group. When the differences in reaction rates of NHS esters of mPEGs with amino and alcohol groups in ethanolamine is large enough, the reactions of NHS esters of mPEGs with amino group in excess should be completed before significant formation of esters with alcohol group and thus the potential interference should be negligible. Indeed, within 15 min under room temperature, the reaction between ethanolamine and either of two NHS esters of mPEG5K was completed while there was negligible reaction between either NHS ester with ethanol at the same concentration in DMF (Figure 3). The reaction rate of NHS-AC with amino group in ethanolamine was even faster than those of NHS-CB-mPEG and NHS-SC-mPEG (Data not given). Therefore, the reactions of ethanolamine with NHS esters of mPEGs may be completed in 15 min under room temperature and still follow 1:1 stoichiometry to simplify the quantification of molar equivalents of NHS esters of mPEGs in preparations.
Optimization of reaction conditions of ethanolamine with TNBS
The reaction of TNBS in excess with each of the tested alkyl primary amines at 25°C was not completed even after 90 min (Figure 2). The increase in reaction pH produced little improvement except increases in background absorbance. Ethanolamine has a high boiling point and its reaction with TNBS may be speeded up at higher temperatures. Indeed, the formation of the conjugates of TNBS and ethanolamine was greatly accelerated at 55°C so that the maximum absorbance was achieved after only 10-min reaction. Strangely, the absorbance at 420 nm of the conjugate of TNBS and ethanolamine began to decrease after 40-min reaction at 55°C or 90-min reaction at 45°C, but continued to increase for more than 120 min at 25°C (Figure 4). These decreases in absorbance at 420 nm may be owing to decomposition of conjugates at higher reaction temperature. As a result, after ethanolamine was reacted with TNBS at 55°C for 15 min, the reaction mixture was cooled to room temperature with tap water before the assay of absorbance at 420 nm. Since cooling, the absorbance at 420 nm remained constant within at least 120 min so that absorbance at 420 nm can be easily measured. This property makes the new method more advantageous over the classical spectrophotometric method based on the absorbance of NHS at 260 nm after reaction with ammonia or hydroxide ion in aqueous solutions [9, 11].
Potential interference from reaction products of ethanolamine and NHS esters of mPEG with the assay of ethanolamine was examined. In THF, NHS-CB-mPEG5k at 0.80 mmol/L and ethanolamine at 0.85 mmmol/L were mixed for 20-min reaction under room temperature. The amide was repetitively precipitated with 10-fold volume of diethyl ether and washed with THF till no ethanolamine in ether solution was detectable by reaction with TNBS. The amide of mPEG5k alone in DMF up to 1.0 mmol/L caused negligible absorbance at 420 nm after reaction with TNBS for 15 min at 55°C; they also caused no changes of absorbance at 420 nm for the conjugate of TNBS with ethanolamine. Similar results were observed with the amide from NHS-SC-mPEG5k. Moreover, NHS at concentrations smaller than 0.80 mmol/L in DMF did not alter the absorbance at 420 nm of the conjugate between TNBS and 0.83 mmol/L ethanolamine; it also caused no interference with the assay of other alkyl primary amines by reactions with TNBS (Additional file 1: Figure S2, Supporting Material). Therefore, the new method is resistant to pre-existed NHS and amide derivatives of mPEGs. Based on the reaction between ethanolamine and TNBS at 55°C for 15 min, the absorbance at 420 nm linearly responded to ethanolamine quantities with a slope about 92% of that after reaction at 37°C for 90 min (Figure 5). The difference in the response slopes may be attributed to increased spontaneous hydrolysis of TNBS and higher background absorbance at 55°C, but it resulted in no problems for the quantification of ethanolamine with TNBS.
Taken together, the optimized conditions for the new method are preset as follows. The reaction between 0.83 mmol/L ethanolamine in excess and a tested NHS ester of mPEG in DMF is allowed to continue for 15 min at 25°C for complete consumption of the NHS ester (Figure 3). The reaction between residual ethanolamine and TNBS in borate buffer at pH 9.2 continues for 15 min at 55°C followed by cooling with tap water to measure absorbance at 420 nm. These optimized conditions were used throughout, unless otherwise stated.
Preliminary applications of the new method
Reference substances of NHS esters of mPEGs for this new method are difficult to prepare, which is more pronounced with chromatographic methods. NHS acetate ester as synthesized has satisfactory purity and can serve as a reference compound for the new method. In addition to labor and instrumentation, the difference in availability of reference substances makes this new method advantageous over chromatographic methods. For the reaction of NHS-AC with ethanolamine, the absorbance at 420 nm of reaction mixtures with TNBS negatively correlated to NHS-AC quantities within a reasonable range in DMF (Figure 6). The absolute value of the slope for such a response displayed a difference just about 1% from that for the response of absorbance at 420 nm to ethanolamine quantities. Therefore, the reaction between ethanolamine and NHS-AC follows 1:1 stoichiometry.
By TLC analysis, no free NHS was detectable in the two NHS esters of mPEG5k. By the new method, there was a linear decrease in absorbance at 420 nm to quantities of either of the two NHS esters of mPEGs in DMF for reactions with ethanolamine. Based on the response slopes and 1:1 reaction stoichiometry, NHS-CB-mPEG5k and NHS-SC-mPEG5k in their preparations were about 90% and 60% of their theoretical values calculated from their average molecular weights, respectively. By the classical spectrophotometric method based on absorbance at 260 nm of NHS released upon the reaction with ammonia, consistent molar equivalents of these two NHS esters of mPEG5k in their preparations were obtained, correspondingly (data not given). NHS-CB-mPEG5k in its preparation had a molar equivalent close to its theoretical value, supporting its high purity and the reaction stoichiometry of 1:1 between NHS-CB-mPEG5k and ethanolamine. It is a putative that NHS-SC-mPEG has reactivity with ethanolamine comparable to NHS-CB-mPEG5k [21, 22]. The reaction between NHS-SC-mPEG and ethanolamine should be completed under the optimized reaction conditions; the lower molar equivalent of NHS-SC-mPEG in its preparation thus indicated its lower purity. Hence, both the new method and the classical spectrophotometric method are effective to check homogeneity of NHS esters of mPEGs in the absence of pre-existed NHS.
To test the potential advantages of the new method to quantify molar equivalents of NHS esters of mPEGs in preparations containing pre-existed NHS due to partial hydrolysis, the storage stability of powder samples of NHS-SC-mPEG5k and NHS-CB-mPEG5k was examined under different conditions. As expected, both the new method and the classical spectrophotometric method gave consistent molar equivalents of NHS esters of mPEG5k in such powders when they were isolated from water during storage. After exposure of such powders to humid air for 6 h, the new method already can detect a statistical significant decrease in molar equivalents of two NHS esters of mPEG5k, but neither TLC analysis nor the classical spectrophotometric method based on NHS absorbance upon reaction with ammonia can safely detect hydrolysis of the two NHS esters of mPEG5k (data not given). After exposure of such powders to humid air for 24 h at 37°C, the two NHS esters display nearly 30% hydrolyses by the new method (Figure 7), less than 20% hydrolyses by the classical spectrophotometric method, and detectable hydrolyses by TLC analyses (Additional file 1: Figure S3, Supporting Material). After exposure of such powders to humid air for 72 h at 37°C, both NHS esters exhibited more than 50% hydrolyses by the new method, but less than 30% hydrolyses by the classical spectrophotometric method. The differences in percentages of hydrolysis in powder samples of each NHS ester of mPEG5k by these two spectrophotometric methods grew larger after exposure of these samples to humid air for more time. It is a putative that hydrolysis of NHS esters of mPEG5k produces NHS. Thus, for quantifying molar equivalents of NHS esters of mPEGs in preparation containing pre-existed NHS, the new method is superior to the classical spectrophotometric method.
In conclusion, a facile method is developed to quantify molar equivalents of NHS esters of mPEGs in laboratory or commercial preparations via their reactions with ethanolamine in dimethylformamide and subsequent spectrophotometric assay of residual ethanolamine with 2,4,6-trinitrobenzenesulfonic acid. The new method consumes just about 35 min for each analysis, displays resistance to pre-existed NHS, is universally applicable to common active esters as long as they are stable in DMF, and requires just one easily-accessible reference compound for different NHS esters. The glycylglycine test of active esters requires over 60 min for each analysis and is unreliable to active esters susceptible to hydrolysis [1, 8, 13]. The classical spectrophotometric method based on NHS absorbance at 260 nm is susceptible to the interference of pre-existed NHS originated from partial decomposition/hydrolysis of NHS esters. Taken together, the new method for controlling quality of commercial or laboratory preparations of active esters of mPEG derivatives and optimizing PEGylation process of therapeutic proteins is advantageous over other conventional methods.