Purification of Peptides Using Surrogate Stationary Phases on Reversed-Phase Columns

Submitted: Sept. 24, 2015
Accepted: Dec. 2, 2015

Prior to this report, there were just two ways to increase the amount of sample that could be purified in a single run by preparative reversed-phase high-performance liquid chromatography (Prep-RP-HPLC). The traditional approach was to use a bigger column, which provided a greater amount of stationary phase; and the second alternative was by displacement chromatography, which uses the stationary phase more effectively. This paper describes a unique Prep-RP-HPLC technique that uses a C-18/C-8 derivatized silica coated with a hydrophobic quaternary ammonium salt or quaternary phosphonium salt that acts as an additional/surrogate stationary phase (ASP/SSP). The SSP is bound to the C-18/C-8 chains and silanols of the stationary phase via van der Waals and ionic interactions. The SSP resides in the interstitial volume available in the C-18/C-8 silica stationary phase. The result is a seven to 12-fold increase in the sample loading in contrast to the conventional Prep-RP-HPLC technique.

Reversed-phase high-performance liquid chromatography (RP-HPLC) is used ubiquitously in the pharmaceutical industry. Analytical RP-HPLC is used for the release and characterization of raw materials, intermediates, and APIs. Preparative RP-HPLC (Prep-RP-HPLC) is used for the commercial production of peptide APIs and most other complex APIs that are not amenable to crystallization.

Prep-RP-HPLC in the elution mode is the most widely practiced and preferred mode for purifying crude peptide mixtures and other small complex organic molecules due to the ease of elution-mode operation. The first-generation Prep-RP-HPLC stationary supports were irregular silica particles that were derivatized with C-18 or C-8 chains, and they suffered from high back pressure (1-3). The high back pressure limited their use with respect to the quantity that could be purified in a single run and to relatively smaller diameter columns. Recent advances in Prep-RP-HPLC have focused on producing spherical silica and the development of new bonding chemistries to furnish stationary supports that have improved stability and selectivity. The commercial manufacture of spherical silica that has been derivatized by C-18, C-8, and other ligands has overcome these challenges and has extended the utility of preparative HPLC vastly. The technological advances in process HPLC instrumentation and the bonded silica supports have made possible commercial production of complex peptides such as Fuzeon, a 36-amino acid peptide, in hundreds of kilos quantities. Unfortunately, these large-scale HPLC instruments and the associated column hardware are costly and restrict the affordability of such methods. Also, none of these improvements have addressed the loading capacity of a given column nor have they resulted in significant increase in the amount of purified product (output/mL of the packed column).

The elution Prep-RP-HPLC mode is limited in terms of the quantity of sample that can be purified in a single run by several factors including the resolution between the desired product and its closest eluting related substance, the capacity factor, and the number of theoretical plates of the preparative column (1-3). The typical loading capacity of synthetic peptides is in the range of 1 to 2 mg per mL of packed column volume (namely, 0.1% to 0.2% with respect to total column volume). Wu et al. have disclosed preparative HPLC of a glucagon-like peptide-1 (GLP-1) analog (4). The loading was 0.225% with respect to total column volume (approximately 45 mg on a 20-mL C-18 substituted (octadecyldimethylsilyl) silica column with particle size of 15 microns). Carl et al have reported Prep-RP-HPLC of (Aib8,35) GLP-1(7-36)-NH2 at loadings up to 20 g/L (2% with respect to total column volume) (5). Table I lists the loading capacity of various chromatographic techniques. Entries 1 to 4 were taken from reference (1). Entry 1 describes the Prep-RP-HPLC of a commercial peptide and the loading here was 0.91% with respect to the total column volume. In contrast, loadings of approximately 8% with respect to the column volume are achievable in normal-phase silica chromatography.

Table I: Loading capacity of various chromatographic techniques. RP-HPLC is reversed-phase high-performance liquid chromatography; TBAHS-SSP-RP-HPLC is tetra-n-butylammonium hydrogen sulphate, surrogate stationary phase, reversed-phase high-performance liquid chromatography

A more efficient alternative to elution Prep-RP-HPLC would be to use it in the displacement mode. Table I, entry 2 describes the displacement Prep-RP-HPLC of a 25 amino acid peptide. Although significantly higher loadings are possible, the method’s limitations are that it is not generally applicable and that it is often arduous to develop. Displacement chromatography uses as a mobile phase, a displacer molecule that has higher affinity for the stationary phase material than the sample components. Haymore describes the displacement chromatographic purification of angiotensin (6). The loading capacity with respect to total column volume was approximately 3.7%, and the relative loading capacity with respect to traditional HPLC was about 4%. Displacement chromatography is best suited for ion-exchange mode, and it has found numerous recent applications (6-7).

The surrogate stationary phase (SSP) Prep-RP-HPLC method is unique and distinct from displacement chromatography. In displacement chromatography, the sample components are applied to the stationary phase, and the subsequent resolution of the various components of the sample mixture is achieved by the use of a displacer molecule. In SSP-Prep-RP-HPLC, the SSP is bound to the stationary phase before the sample is applied. The reversed-phase displacement chromatography for separating organic compounds from a mixture involves the step of applying to the hydrophobic stationary phase a mixture comprising organic compounds to be separated before the addition of the displacer (6-9).

Table I, entry 3 lists the loading capacity during the enantiomers separation (normal phase) using the box-car injection technique (1). It was about 1.0%, which is close to the sample loading observed in standard Prep-RP-HPLC (Table I, entry 1).

Table II. Purification of leuprolide: Comparison of the surrogate stationary phase aided Prep-RP-HPLC with the standard Prep-RP-HPLC. Prep-RP-HPLC is preparative reversed-phase high-performance liquid chromatography; TBAHS-SSP is tetra-n-butylammonium hydrogen sulphate, surrogate stationary phase.

The results of SSP-Prep-RP-HPLC of leuprolide (Pyr-His-Trp-Ser-Tyr-D-Leu-Leu-Arg-Pro-NHEt) are summarized in Table II. The purified product (leuprolide) output of the standard Prep-RP-HPLC is 2.4 mg/mL of column volume (Table II, entry 1). In contrast, the purified product output of the SSP-Prep-RP-HPLC is 29.6 mg/mL of column volume (Table II, entry 2) and 16.3 mg/mL of column volume (Table II, entry 3). These results suggest that loadings of 7 to 12 times capacity of conventional prep-RP-HPLC are achievable with the processes described in the present paper (10-11).

This study also provides a process for the removal of the SSP reagents such as hydrophobic quaternary ammonium salt or quaternary phosphonium from the C-18 or C-8 column by using sodium tetrafluoroborate or potassium hexafluorophosphate with organic modifier.

The SSP-Prep-RP-HPLC technique described in this study has the following advantages--increased sample loading; economic use (lesser quantity) of solvents; reduced waste disposal; simple operation; and reduced scale of the equipment used to chromatograph, elute, concentrate, and recover the desired components.

Materials and methods

Prep-RP-HPLC using YMC-ODS-AQ [50 mm internal diameter (ID) X 250 mm length (L), C-18, 10 μ particles, 120 Å pore diameter; 14% carbon content]; Waters Symmetry [19 mm (ID) X 50 mm (L), C-8, 5 μ particles, 100 Å pore diameter; 11.7% carbon content]; and Discovery Bio Wide Pore [10 mm (ID) X 250 mm (L), C-18, 5 μ particles, 300 Å pore diameter, 9.2% carbon content] columns was performed on a Waters LC2000 equipped with a variable wavelength ultraviolet (UV) detector. Prep-RP-HPLC using Grace Vydac C-18 cartridges (12 g C-18, 40 μ particles, 60 Å pore diameter) was performed on a Reveleris flash chromatography system equipped with a variable wavelength UV and evaporative light scattering detector (ELSD). Analytical RP-HPLC analyses were performed on a Waters Alliance HPLC system equipped with a photodiode array (PDA) UV detector, and YMC-ODS-A [4.6 mm (ID) X 250 mm (L), 5 μ particles, 120 Å pore diameter]. Purified leuprolide samples obtained with the YMC, Discovery, and Waters Symmetry columns were assayed for purity using the method described in the United States Pharmacopeia (USP) monograph (12) on a C-18 YMC or equivalent column [4.6 mm (ID) X 100 mm (L), 3 μ]. The leuprolide obtained in this study was identical in all aspects including the infrared (IR), nuclear magnetic resonance (NMR), and electrospray mass spectrometry (ES-MS) spectra, amino acid analyses, and co-elution on RP-HPLC with an authetic leuprolide USP standard. Partially purified leuprolide (approximately 95% pure) was analyzed by an in-house gradient modification of the USP method.

Crude leuprolide used in these Prep-RP-HPLC studies was obtained using an innovative, proprietary solution-phase peptide synthesis route. The crude product contained 81.7% of leuprolide by quantitative HPLC assay.

General procedure

The C-18/C-8 reversed-phase column was equilibrated with 5-10 column volumes (Vcs) of 5% to 10% aqueous acetonitrile containing 10 mM tetra-n-butylammonium hydrogen sulphate (TBAHS, Buffer A). The pH of the starting buffer was not adjusted and was approximately 1.95. It is important to keep the concentration of acetonitrile at 5% to 10% lower than the concentration needed to elute the product on an analytical HPLC column. The crude compound to be purified was dissolved in starting buffer A or aqueous trifluoroacetic acid (TFA) or aqueous acetic acid (HOAc) and loaded onto the column. After the loading was complete, the column was equilibrated with 2 Vcs of buffer A. Next, the gradient elution process was started. The buffer B is usually 300 mM to 500 mM TBAHS in 5% to 10% aqueous acetonitrile or 5-25 mM quaternary ammonium/phosphonium salt in 50% aqueous acetonitrile. A linear gradient of 0% to 100% buffer B over 10 Vcs was applied. When the product of interest was about to elute, a gradient hold was applied until all the desired product had eluted from the column. Alternately, if the aim was to elute the product in a concentrated form, the gradient was allowed to run its course. The fractions containing the pure product were combined after confirming that the pooled fraction met the purification criteria. The approximate quantity of the associated quaternary ammonium/phosphonium salt was calculated, and treated with 1.5 to 2 equivalents of sodium tetrafluoroborate (NaBF4) and extracted three times with chloroform to remove the quaternary ammonium/phosphonium cation as its tetrafluoroborate salt. The aqueous residue was then loaded on to a C-18/C-8 column from which all the SSP (quaternary ammonium/phosphonium salt) had been removed. Removal of SSP from the C-18/C-8 column was accomplished by the following steps: The column was first washed with at least three Vcs of 80% acetonitrile, 20% water. Next, the column was washed with three Vcs of 100 mM NaBF4 in 80% acetonitrile, 20% water. The column was equilibrated with 1M acetic acid in 1% aqueous acetonitrile (10 Vcs). The aqueous phase containing “pure product” and excess NaBF4 was diluted with water (five times its volume) and loaded on to the C-18/C-8 column. The column was washed with five to 10 Vcs of 1% phosphoric acid, 1% acetonitrile, 98% water to exchange the BF4 anions for phosphate anions. The column was then washed with five to 10 Vcs of 100 mM aqueous guanidine hydrochloride to remove the phosphate anions and to exchange the phosphate anions to chloride anions. Finally, the chloride anions were exchanged for acetate anions. The fractions containing the “pure product acetate salt” were combined, and the organic volatiles were removed under reduced pressure. The aqueous residue was lyophilized or precipitated after removal of water. The final product was analyzed according to the USP/EP Methods of Analysis (12). Table II summarizes the results of purifying crude leuprolide using the standard Prep-RP-HPLC and the SSP-Prep-RP-HPLC technique.

Examples

Example 1: TBAHS-SSP-Prep-RP-HPLC of leuprolide acetate (Pyr-His-Trp-Ser-Tyr-D-Leu-Leu-Arg-Pro-NHEt). Two different columns were evaluated for the purification of leuprolide: A Discovery Bio Wide Pore column [column parameters: 10 mm (ID) X 250 mm (L), C-18, 5 u particles, 300Ǻ pore diameter, 9.2% carbon content; amount loaded was 1.2 g of crude leuprolide (prepared by solution phase peptide synthesis)] and a Waters Symmetry Column [column parameters: 19 mm (ID) X 50 mm (L), C-8, 5 μ particles, 120Ǻ pore diameter, 11.7% carbon content; amount loaded was 1.4 g of crude leuprolide (prepared by solution phase synthesis)] were used. Buffer A was 10 mM TBAHS in 10% aqueous acetonitrile. Buffer B was 300 mM TBAHS in 10% aqueous acetonitrile. The purification process was done in triplicates to confirm reproducibility, and followed the general procedure described in the previous section. Lyophilized leuprolide acetate API product was assayed according to the USP method (see Figures 1 and 2). The purification output was 0.42 g (30% yield for Waters Symmetry Column) and 0.32 g (26.7% yield for Discovery Bio Wide Pore Column) (see Table II).

Example 2: TBAHS-SSP-Prep-RP-HPLC of triptorelin acetate (Pyr-His-Trp-Ser-Tyr-D-Trp-Leu-Arg-Pro-Gly-NH2). A Discovery Bio Wide Pore column [column parameters: 10 mm (ID) X 250 mm (L), C-18, 5 μ particles, 300Ǻ pore diameter; amount loaded was 1.0 g of crude triptorelin] was used.

Buffer A was 10 mM TBAHS in 10% aqueous acetonitrile. Buffer B was 300 mM TBAHS in 10% aqueous acetonitrile. The purification process was done in triplicates to confirm reproducibility, and followed the general procedure described in the previous section. The yield of lyophilized triptorelin acetate was approximately 25%. 

Example 3: Tetra-n-butylammonium bromide (TBA-Br)-SSP-Prep-RP-HPLC of leuprolide. The C-18 reversed-phase column [Grace Vydac column parameters: 12 g C-18, 40 μ particles, 60 Å pore diameter] was saturated with 36 g of TBA-Br in 360 mL of water at a flow rate of 8.0 mL/min. The column was then equilibrated with 10 Vcs of buffer A (25mM TBA-Br in water) at a flow rate of 8.0 mL/min. The crude leuprolide trifluoroacetate salt was dissolved in buffer A and loaded on to the column. After loading was complete, the gradient elution process was started. The buffer B was 25 mM TBA-Br in 50% aqueous acetonitrile. A linear gradient of 0% of buffer B to 100% of buffer B over 10 Vcs was applied. When leuprolide was about to elute, a gradient hold was applied until all the API had eluted from the column. The fractions containing >95% purity of leuprolide were combined after confirming the purity on an analytical HPLC column. Yield was 66.4% (see Table III, entry 1).

Table III. Performance of quaternary ammonium and phosphonium as an additional/surrogate stationary phase (ASP/ SSP). Experimental parameters: Reveleris C-18, 12 g, CV 18 mL, 40 μ, 60 Å; flow rate was 8 mL/min. The analyte was crude leuprolide, 800 mg, net weight by Edelhoch method; 84.6% pure by modified United States Pharmacopeia method.

Example 4: TBAHS-SSP-Prep-RP-HPLC of leuprolide. The procedure described in example 3 was followed except that the buffer A was 25 mM TBAHS in water, and buffer B was 25mM of TBAHS in 50% aqueous acetonitrile. Yield was 64.4% (see Table III, entry 2).

Example 5: Cetyltrimethylammonium bromide (CTA-Br)-Prep-RP-HPLC of leuprolide. The C-18 reversed phase column [Grace Vydac column parameters: 12 g of C-18, 40 μ particles, 60Å pore diameter] was saturated with an aqueous solution of 1mM CTA-Br as described in example 3. The procedure described in example 3 was followed except that the buffer A was 5 mM CTA-Br in water, and buffer B was 5 mM of CTA-Br in 50% aqueous acetonitrile. Yield was 61.4% (see Table III, entry 3).

Example 6: Tetra-n-butylphosphonium chloride (TBP-Cl)-SSP-Prep-RP-HPLC of leuprolide. The procedure described in example 3 was followed except that the buffer A was 25 mM TBP-Cl in water, and buffer B was 25 mM of TBP-Cl in 50% aqueous acetonitrile. Yield was 60.3% (see Table III, entry 4).

Example 7: Tetra-n-butylammonium chloride (TBA-Cl)-SSP-Prep-RP-HPLC of leuprolide. The procedure described in example 3 was followed except that the buffer A was 25 mM TBA-Cl in water, and buffer B was 25 mM of TBA-Cl in 50% aqueous acetonitrile. Yield was 53.5% (see Table III, entry 5).

Discussion

The loading capacities of various chromatographic techniques are summarized in Table I. The lowest loading capacity amongst these techniques was seen with the traditional Prep-RP-HPLC technique (Table I, entry 1 and Table II, entry 1). The low loading capacity of traditional elution mode Prep-RP-HPLC may be due to many factors that include resolution between the product of interest and its closest eluting related substance, the number of theoretical plates, selectivity, molecular size of the analyte, size and pore diameter of the stationary phase, and carbon content, among others. These factors were studied using octreotide as an analyte and are described later in this section. The coating of C-18/C-8 stationary phase with SSP has beneficial effects on the efficiency, selectivity, and retention characteristics of the column.

Table IV: Loading capacity in classical preparative reversed-phase high-performance liquid chromatography (Prep- RP HPLC): control experiment. Control experiments parameters: Reveleris C-18, 12 g, 40 μ, 60 Å. The analyte was crude leuprolide, 84.6% purity determined by modified United States Pharmacopeia (USP) method. Flow rate was 8 mL/min. Buffer A = 0.1% trifluoroacetic acid (TFA) in water; Buffer B = 0.1% TFA in 50% aqueous acetonitrile.

Control experiments with various loadings of leuprolide on an uncoated Grace Vydac C-18 cartridge were done to evaluate the deleterious contribution of non-specific binding of the analyte with the ionic sites on the reversed-phase column. The results are summarized in Table IV. The C-18 column was loaded with various finite amounts of crude leuprolide [84.6% purity by HPLC; peptide assay was done by the Edelhoch method (13)]. Four parameters were studied to evaluate the chromatography performance:

• Flow through: The amount of leuprolide in the flow-through during loading was measured to ascertain whether the capacity of column during loading was exceeded.

• Pool of fractions containing at least 95.0% leuprolide: Several pools of fractions were made and the amount of leuprolide was quantified using the Edelhoch method and by quantitative HPLC assay.

• Purest leuprolide fraction (appraisal of resolution): The fraction containing the highest purity of leuprolide was determined. This fraction was helpful in assessing the resolution of leuprolide from its closest eluting impurities.

• Mass balance of the entire eluent from the chromatographic run: This parameter was measured using the Edelhoch method. This parameter was helpful in determining the loss of leuprolide and similar analogs due to non-specific ionic binding to residual silanol groups present on the column.

Examination of Table IV reveals that:

• The output (% purification yield, Table IV, entry 6) of >95% leuprolide ranged from 11.9% to 19.1%. Stated simply, 80.9% to 88.1% of crude leuprolide was lost to ionic interactions with the stationary phase.

• The mass balance of the individual complete chromatography run was in the range of 88.4% to 96.5% (Table IV, entry 8). This finding suggests the high contribution of non-reversed type of interactions between the analyte and the stationary phase to be the cause of poor purification performance with respect to the output of >95% leuprolide.

• The purity of the individual fractions ranged from 97.8% (when 100 mg of crude leuprolide was loaded) to 95.2% (when 800 mg of crude leuprolide was loaded). The purity was 95.5% when 1200 mg of crude leuprolide was loaded. This result may be due to the “self-displacement” contribution.

The beneficial effects of SSP-Prep-RP-HPLC can be seen from an examination of Table III:

• The output (% purification yield) of >95% leuprolide ranged from 53.5% to 66.4%. Absence of leuprolide in the 100% B wash confirms that ionic interactions with stationary phase are negligible in the SSP-Prep-RP-HPLC.

• The mass balance of the individual complete chromatography run was in the range of 91.0% to 96.0%.

• With a constant loading of 800 mg of leuprolide, the highest purity of individual fraction ranged from 96.4% to 98.3% and the yields of >95% purity leuprolide ranged from 53.5% to 66.4%.

The resolution is given by Equation 1:

It is well known to chromatographers that the greatest gain in resolution is obtained by increasing the selectivity parameter. To understand the reasons for the superior performance of SSP-Prep-RP-HPLC, several C-18 analytical columns [0.46 cm (ID) X 25 cm (L), 5 μ particles] were chosen and their resolution parameters were studied before and after coating with tetra-alkylammonium salts. The resolution of octreotide [H-D-Phe-Cyclo (Cys-Phe-D-Trp-Lys-Thr-Cys)-L-Threoninol] from an early eluting isoacyl analog was studied. A reduction in “k” (retention factor) and “USP tailing factor” was observed with all the C-18 columns. The plate count (N) increased or remained similar to uncoated columns. In all cases, the selectivity parameter (α) increased significantly.

Figure 1: Analytical RP HPLC profile (USP method) of leuprolide acetate obtained using TBAHS-SSP-Prep-RP HPLC. A Discovery Bio Wide Pore (10 mm (ID) X 250 mm (L), C-18, 5  μ particles, and 300Å pore diameter, 9.2% carbon content) column was used for the TBAHS-SSP- Prep-RP HPLC process.

An examination of Table II reveals that loading capacities of the ASP/SSP technique are in the range of 7.5X to 12.1X. The C-8 column (11.7% carbon content) has a higher sample loading capacity than the C-18 column (9.2% carbon content) and consequently, a higher amount of the adsorbed SSP. The higher sample loading observed with SSP-aided Prep-RP-HPLC is ascribed to the increased hydrophobic surface area that is available as a consequence of the SSP/ASP assembling into a three-dimensional matrix.

Figure 2: Analytical RP HPLC profile (USP method) of leuprolide acetate obtained using TBAHS-SSP-Prep- RP HPLC. A Waters Symmetry (19 mm (ID) X 50 mm (L), C-8, 5 μ  particles, 120 Å pore diameter, 11.7% carbon content) column was used for the TBAHS-SSP-Prep-RP-HPLC process.

Conclusion

A novel and generally applicable Prep-RP-HPLC method has been developed. In this method, a quaternary ammonium/phosphonium salt is used as an additional stationary phase (ASP)/surrogate stationary phase (SSP). The increase in sample loading and purification output is due to the ASP/SSP occupying part of the interstial volume of the porous silica C-18/C-8 particles, and is directly proportional to % carbon loading. The 7- to 12-fold beneficial increase in the loading and the increase in purification output is the synergistic result of the ASP/SSP increasing the plate count (N, efficiency), increasing the selectivity parameter (α), and decreasing the tailing factor of the RP-HPLC column.

References

1. D. A. Wellings, A Practical Handbook Preparative HPLC (Elsevier, Amsterdam, UK, 2006).
2. M.-I. Aguilar (Editor), “HPLC of Peptides and Proteins,” in Methods in Molecular Biology (Humana Press, New Jersey, USA, 2004).
3. S. Kromidas (Editor), HPLC Made to Measure--A practical handbook for
optimization (Wiley-VCH, Germany, 2006).
4. X. Wu, and A. Bogsnes, “Preparative RP-HPLC Method for Purifying Peptides,” US Patent 0322976, December 2012.
5. C. Carl, M. Rothe, C. Saladin, D. Strub, and F. Vix, “Reversed Phase HPLC Purification of a glp-1 Analogue,” US Patent 0313131, December 2011.
6. B. L. Haymore, “Cationic Displacer Molecules for Hydrophobic Displacement Chromatography,” WO Patent 052539A3, June 2013.
7. K. M. Sunasara, R. G. Rupp, and S. M. Cramer, Biotechnology Progress 17 (5) 897-906 (2001).
8. S. M. Cramer, A. A. Shukla, and K. M. Sunasara, “Low Molecular Weight Displacers for Protein Purification in Hydrophobic interaction and Reversed Phase Chromatographic Systems,” US Patent 6239262, May 2001.
9. B. L. Haymore, “Neutral Zwitterionic Displacer Molecules for Hydrophobic Displacement Chromatography,” WO 052087, April 2013.
10. M. K. Anwer, “Purification of Organic Compounds Using Surrogate Stationary Phases on Reversed Phase Columns,” WO patent 118797, August 2014.
11. M. K. Anwer, “Purification of Organic Compounds by Surfactant Mediated Preparative HPLC,” WO patent 040635A2, March 2015.
12. USP-NF, Monograph 34, Methods of Analysis, page 3279-3281.
13. H. Edelhoch, Biochemistry 6(7) 1948-1954 (1967).

Article Details

Pharmaceutical Technology
Vol. 40, No. 8
Pages: 37–44

Citation: 

When referring to this article, please cite as M. Anwer et al., "Purification of Peptides Using Surrogate Stationary Phases on Reversed-Phase Columns," Pharmaceutical Technology 40 (8) 2016.