MSc Chemistry
Analytical Chemistry
Master Thesis
Polymeric Monoliths for the Characterization of Intact Proteins
A study about developing a method compatible with mass spectrometry using
hydrophobic interaction chromatography or ion exchange chromatography for the
analysis of immunoglobulins
by
Laura de Wal
10756426
February 2020
48 EC
July 2019 – February 2020
Supervisor/Examiner:
Examiner:
Prof. dr. ir. P.J. (Peter) Schoenmakers
dr. R. (Rob) Haselberg
Prof. dr. E.F. (Emily) Hilder
Dr. R.D. (Dario) Arrua
Acknowledgements
The supervisors of this project Prof. Emily Hilder and Dr. Dario Arrua have provided excellent help and guidance during the eight months of this project. In every stage of this research project they provided inspiring new ideas to bring this project to the next level and always gave great encouragement to keep finding new answers, but also questions. Their supervision and involvement are much appreciated.
The supervision from Amsterdam, the Netherlands is thankfully acknowlegded. During this project at the other side of the world Prof. Peter Schoenmakers and Dr. Rob Haselberg were always open to help and support when needed. Also, the idea and opportunity by Prof. Peter Schoenmakers who suggested to do my research project at the Future Industries Institute of the University of South Australia is much appreciated.
Many others have contributed to make this a successful project, in particular Dr. Amin Khodabandeh and Dr. Ester Lubomirsky who were always open for help and created a great lab atmosphere. The E.H. research group deserves an acknowledgement as well, for making sure there always was a pleasant working environment. Outside this group, Ken Neubauer from Adelaide Microscopy and Dr. Nobuyki Kawashima gave crucial assistance during the SEM sessions, which contributed greatly to this project.
Table of Contents List of Abbreviations ... 5 Abstract ... 6 Samenvatting ... 7 Chapter 1: Introduction to the separation of intact proteins ... 8 1.1 Introduction ... 8 1.2 Proteomics ... 8 1.3 Immunoglobulins and monoclonal antibodies ... 10 1.4 Separation of large proteins ... 11 1.5 Monolith columns ... 12 1.6 Components in the polymerization mixture ... 15 1.7 Project aim... 17 Chapter 02: Experimental ...18 2.1 Chemicals ... 18 2.2 Equipment ... 19 2.3 Preparation of capillary columns ... 20 2.4 Capillary liquid chromatography experiments ... 22 Chapter 3: Polymeric monoliths for hydrophobic interaction chromatography for the separation of intact proteins ...23 3.1 Introduction ... 23 3.2 Hydrophobic interaction between the stationary phase and proteins ... 23 3.3 Hydrophobic interaction chromatography coupled mass spectrometry ... 24 3.4 PEGDA polymeric monoliths as stationary phase ... 25 3.5 Hydrophobic interaction chromatography on PEGDA column ... 27 3.6 Hydrophobic interaction with sulfate salts as mobile phase ... 28 3.7 HIC separation with ammonium acetate as mobile phase ... 29 Chapter 4: Polymeric monoliths for ion exchange chromatography for the separation of intact proteins ...35 4.1 Introduction ... 35 4.2 Ion exchange – mass spectrometry based on a salt gradient ... 36 4.3 Charge variant analysis - Protein separation based on a pH gradient ... 37 4.4 Results: protein separation with ion exchange chromatography ... 39 Chapter 5: Optimization of sulfoalkylbetaine polymeric monolith...47 5.1 Introduction ... 47 5.2 Repeatability ... 47 5.3 Permeability and porosity ... 49 5.3 Different ratios of monomer mixture ... 51 5.4 Thermal polymerization ... 52 5.5 Changes in the porogen mixture ... 54
Chapter 6: Conclusions and Future Work ...61
6.1 Conclusions ... 61
6.2 Future work ... 62
List of Abbreviations
mAbs Monoclonal antibodies
IgG Immunoglobulin
Fab Fragment antigen-binding
Fc Fragment crystallizable
(2D)LC-MS (two dimensional)Liquid chromatography
coupled mass spectrometry
PTMs Post Translational Modifications
ESI-MS Electrospray ionization mass spectrometry
RPLC Reversed phase liquid chromatography
UV Ultraviolet
IEC Ion exchange chromatography
EDMA Ethylene glycol dimethacrylate
PEG Poly (ethylene glycol)
Mw Molecular weight
HIC Hydrophobic interaction chromatography
PEGDA Poly (ethylene glycol) diacrylate
CAD collision-activated dissociation
CVA Charge variant analysis
SCX Strong cation exchange
HILIC Hydrophilic interaction liquid chromatography
CEC Capillary electro chromatography
SPE N,N-
dimethyl-N-methacryloyloxyethyl-N-(3-sulfopropyl) ammonium betaine
SEM Scanning electron microscopy
Abstract
Polymeric monolithic stationary phases are well suited for the separation and characterization of large structures using different chromatographic modes like hydrophobic interaction chromatography or ion exchange chromatography. Polymeric monolithic stationary phases show higher efficiencies and resolution for the separation of proteins.1
In order to preserve the higher order structure of proteins, chromatographic methods as hydrophobic interaction chromatography (HIC) or ion exchange chromatography (IEC) are of great value to obtain information about the intact structure of proteins. However, traditional approaches use high concentrations of non-volatile salts, which makes them not suitable to couple to mass spectrometry (MS). To overcome this challenge to characterize intact proteins with LC-MS, a successful HIC or IEC method needs to be developed with a low concentration of MS-compatible salts. Here, polymeric monolithic stationary phases are used to increase the affinity between the proteins and the stationary phase to a point that proteins are retained and eluted using a maximum salt concentration of 1M of a MS-compatible salt.2
Two different polymeric monolithic structures were prepared, for hydrophobic interaction chromatography it was based on an acrylate-based PEGDA Mw 258 g/mol (poly(ethylene glycol) diacrylate) monomer and for ion-exchange chromatography a monolithic structure based on a SPE (N,N- dimethyl-N-methacryloyloxyethyl-N-(3-sulfopropyl) ammonium betaine) zwitterionic monomer was prepared. 3,4 Both capillary columns showed good results when traditional salts were used. The separation of immunoglobulins is promising with the SPE zwitterionic stationary phase where proteins can be separated based on a pH gradient with only a low concentration of buffer in the mobile phase. The SPE zwitterionic stationary phase was further investigated and different parameters within the polymerization mixture were explored to optimize the conditions for protein separation.
Samenvatting
Een ruim deel van alle medicijnen ooit geproduceerd, bevatten eiwitten. Het succes van deze eiwitten zorgt voor een snelle omschakeling van medicijnen die bestaan uit kleine moleculen naar medicijnen gemaakt van grote biomoleculen. Naar verwachting blijft de wereldmarkt voor bio geneesmiddelen groeien met tenminste 10.9% in 2024. Op eiwit-gebaseerde geneesmiddelen hebben een wereld geopend voor de behandeling tegen levensbedreigende ziekten zoals kanker en AIDS. Daarnaast wordt er ook verwacht dat er 50% meer op eiwit-gebaseerde geneesmiddelen worden goedgekeurd en gebruikt. Om de veiligheid voor al deze nieuwe geneesmiddelen te waarborgen, groeit de vraag naar nieuwe analytische methodes om deze eiwitten te analyseren. Er zijn oneindig veel manieren voor eiwit-analyse, in dit project wordt er gebruikt gemaakt van capillaire vloeistofchromatografie. Wanneer de stationaire fase gemaakt is uit een polymeer-structuur, verleend dit zich uitstekend voor de scheiding en identificatie van grote biomoleculen zoals eiwitten en geeft een hogere resolutie voor de vloeistofchromatografie. Eiwitten zijn opgebouwd uit verschillende lagen en structuren, waar veel informatie in verborgen zit. Om deze hogere orde informatie te behouden, hebben zachte interactie vloeistof chromatografische methodes de voorkeur. Dit zijn Hydrophobic Interaction Chromatography (HIC) en Ion Exchange
Chromatography (IEC). Deze methodes bevatten echter een hoge concentratie aan zouten, die niet
samengaan met massa spectrometrie. In dit project wordt er een oplossing gezocht voor deze uitdaging. Op polymeer-gebaseerde stationaire fase wordt gemaakt om te onderzoeken of een nieuwe stationaire fase een verhoogde affiniteit heeft met deze eiwitten zodat traditionele zouten niet meer nodig zijn. Twee verschillende polymeren zijn onderzocht, een licht hydrofiele monomeer voor de HIC-metingen en een zwitterionische monomeer voor IEC-metingen. Beide kolommen gaven uitstekende resultaten voor analyse van eiwitten met behulp van traditionele zouten en veel belovende resultaten zijn behaald voor de analyse met eiwitten met behulp van vluchtige zouten. Verschillende parameters van de zwitterionische stationaire fase zijn verder onderzocht.
Chapter 1: Introduction to the separation of intact proteins 1.1 Introduction
Nowadays seven out of the ten best-selling therapeutics are protein based. This means there has been a rapid shift in the development of new biopharmaceuticals over small molecule-based therapies. Therapeutic proteins have enabled the treatment of a wide range of life-threatening diseases, which were thought to be incurable or untreatable before. Different kinds of new drugs for the treatment of cancer, AIDS and arthritis are on the market or are almost ready for approval. It is expected that within the current decade, more than 50% of the new regulatory drug approvals will be used. 5,6 This has led to a need for new analytical characterization that can support the development and characterization of new protein targets to ensure safety for new therapies. This chapter will introduce proteomics, which covers the study of proteins and specific proteins such as immunoglobulins. There are endless different ways to approach proteomics, one of the ways is with capillary liquid chromatography, as investigated in this research.
1.2 Proteomics
Proteins are large and complex macromolecules that play a crucial role in many different biological processes at the molecular level, considering they are products of transcriptional and translational processes.7 This contributes to the clinical relationship between proteins and diseases. This includes the role of biomarkers, where proteins can be found in tissue or plasma of patients suffering from a certain disease. Those proteins are expressed in a different way than in healthy patients. The study of proteomics monitors these differences. However, this might sound relatively easy, the reality shows there are approximately 20.000 protein-encoding genes in the human body.8
To get the right information about a particular protein, first the protein has to be isolated from their matrix, separated from other similar proteins and be analyzed. Within proteomics, mainly this is done within two ways, namely up and top-down proteomics. This started with bottom-up proteomics, where proteins are studied from the ‘bottom’ through enzymatic digestion with trypsin to cut the protein to smaller peptides. These peptides can be identified with tandem mass spectrometry and compared with theoretical tandem mass spectra from a protein database. Bottom-up proteomics relies on the analysis of peptides, which is suitable for routine proteomics due to the ease of fractionating, ionizing and fragmenting. 9 This process affords a high number of
identifications. However, with the analysis of only peptides a lot of information about the protein is lost. This arises when proteins undergo diverse modifications e.g. acetylation or methylation, but also with incomplete characterization of alternative splice forms or protein cleavages. This loss of information forms a challenge for protein analysis, which also occurs for protein-based therapeutics. 10
Any modification of either primary, secondary, tertiary or quaternary structure of protein therapeutics can in principle impact its function therefore safety profile.11 Therefore, it is important to be able to detect and characterize these post-translational or chemical modifications as this give information about variations of the amino acid sequence. Most changes in amino acid sequence involve a change in mass, which can be detected by mass spectrometry.12,13 It can be easily overlooked, but these standard proteomics methods lack the ability to identify all modifications sites or characterize the full complement of modification. Often it only gives partial sequence information. Also, not all expected peptides or fragments are found back in the MS/MS spectra.13 As full knowledge of the protein has become important, the interest rose in top-down proteomics, which overcomes this problem.
Top-down proteomics is a promise for the full description of the proteome by retaining the intact mass information. One of the key benefits in top-down proteomics is the prospect of comprehensively interrogating all post translational modifications (PTM) within the whole protein from a birds-eye view. This holds great promise for providing novel insights into cellular signal transduction and regulation, as well as disease mechanisms. This includes the identification, characterization and quantification of various proteoforms.14 Proteoforms can be described as all of the different molecular form in which the protein product of singe gene can be found. The different molecular forms arise from genetic variations, alternatively spliced RNA transcripts and post translational modifications. 15 Although top-down proteomics show great promise, still monoclonal antibodies (mAbs) and their derivatives are among the most complex biologic compounds. That is why analytical challenges remain for comprehensive characterization and quality control of mAbs.15 Numerically, the intact proteome appears to be much simpler mixture than its corresponding peptide digests. However, in practice protein-level fractionation and separation are daunting tasks due to the diverse physicochemical properties (e.g. size, charge, and hydrophobicity) and the wide dynamic range of the proteome. 16
1.3 Immunoglobulins and monoclonal antibodies
Monoclonal antibodies (mAbs) are proteins made by a unique parent immune cell. These monoclonal antibodies had the first approval in 1986 to be used as therapeutic drugs. Since then, therapeutically monoclonal antibodies show a continuing growth and are arguably the most important and promising class of therapeutics for a variety of different human diseases such as cancer, autoimmunity or infection diseases. 17,18 Monoclonal antibodies have the advantage of being target specific, selective, together with having a long half-life and reliable safety profile. Also, protein engineering with specific antibody-drug conjugates is improving, which stimulates the development of new mAb-based therapeutics, resulting in high demand for proteomics research.18 Therapeutic mAbs are large glycoproteins that belong to immunoglobulins with a molecular weight around 150 kDa. The immune system uses immunoglobulins to identify and neutralize foreign organisms or antigens. Immunoglobulins are divided in five groups, IgA, IgD, IgE, IgG and IgM. Currently, if the antibody is approved for clinical used, it is based on the immunoglobulin G. 19 Over sixty immunoglobulin and IgG derivatives have been approved for the use as indicators in cancer or inflammatory diseases. 20 IgG’s make up 75% of the total immunoglobulin present in the human serum and have the typical Y-shape (Figure 1.1).3 This Y-shape configuration contains two identical heavy chains and two light chains connected by two disulfide bonds (H and L chains). Both of the heavy chains contain one variable (VH) and three constant domains (CH1, CH2 and CH3), whereas both light chains contain one variable (VL) and one constant domain (CL). In the heavy chain, CH1 and CH2 are linked by sulfide bonds in what is called the hinge region (Hi). This hinge region can differ within the subclasses, e.g. IgG1 and IgG4 have two disulfide bonds in the hinge region, whereas IgG2 has four. 19
The heavy chain can be divided into a Fab and Fc section. The fragment antigen-binding (Fab) region (VH and CH1) is located at the amino terminal, which can be seen as the arm of the Y-shape. The base section of the antibody is known for the fragment crystallizable (Fc) region (CH2 and CH3). The variable domains contain the antigen-biding regions (CDR). These binding sites make immunoglobulins interact against foreign antigens, such as viruses and bacteria.19,21
Also, IgGs can be further divided into four different groups, IgG1,2,3, and 4 based on different patterns of the inter connection of disulfide bonds and heavy-chain sequences.19 IgG1,2, and 4 are
widely used in therapeutics. While IgG1 is most commonly developed, IgG3, is rarely used because of its shorter serum half-life. 17 Infections can change the proportions of each subclass significantly as a result of an immunological compromised situation. 3
Figure 1.1: The global structure of immunoglobulin G. This Y-shaped IgG structure has two identical heavy and light chains. Two constant regions from the heavy chain (CH2 and CH3) will form the Fc region. The variable regions with the close constant region will form the Fab region. The variable region from both the light and heavy chain contain antigen-binding regions.19
1.4 Separation of large proteins
The growing demands of characterizing even more complex protein-based therapeutics continue to stimulate development of new analytical methodologies. Use of hyphenated techniques such as LC-MS(MS) has proved particularly useful for characterization of complex protein samples both in the field of biopharmaceuticals and beyond. 16 The analysis with e.g. liquid chromatography coupled mass spectrometry is necessary to ensure the therapeutic proteins are of high quality and there are no unwanted PTMs that may be potentially immunogenic.13 Native ESI-MS is a promising detection tool for mass profiling and beyond for various species based on the analysis of charge state distributions in the mass spectra.22
Among all chromatographic methods reversed phase liquid chromatography (RPLC) can be seen as the golden standard and therefore is the most commonly used method because of its robustness, high efficiency and compatibility with mass spectrometry. 6,13 In RPLC, it is possible to analyze
therapeutic proteins in their intact forms. This approach is often considered for impurity profiling or heterogeneity evaluation. Typically, oxidized and reduced forms of the native protein can be separated and quantified in RPLC conditions. For example, Fekete et al.23 reported the separation of test proteins under a reversed phase mechanism with an gradient from 30-60% acetonitrile. However, stationary phases for RPLC generally consist a very hydrophobic material, often based on long C-chains (e.g. C18). The hydrophobic patches of the proteins will be absorbed on the stationary phase, which can give poor recovery of intact protein. 16 In order to elute the proteins a large percentage of organic solvent e.g. acetonitrile is required, which can cause denaturation of proteins. 13 As a result, any information on the higher order structure (polypeptide chain folding and or noncovalent assembly of multiunit protein complexes) is lost, and the MS analysis is focused exclusively on covalent structure. Due to the non-denaturation conditions, hydrophobic interaction chromatography (HIC) and ion exchange chromatography (IEC) are considered as promising alternatives for the analysis of PTMs of intact protein level. 18 HIC and IEC are further discussed in the next chapters.
The salt incompatibility could possibly be solved by online two-dimensional liquid chromatography coupled mass spectrometry (2D-LC/MS). Alvarez et al. first 24, and then followed by others 25,26, showed how online 2D-LC/MS could be used for the rapid characterization of monoclonal antibodies. Methods were developed where they use a IEC separation in the first dimension to measure the charge variants followed by reversed-phase chromatography in the second dimension coupled to mass spectrometry including a desalting step. However, even 2D-LC does not seem to be sufficient for low-abundance charge variants, especially those charge variants which are undetectable by UV detection at the first dimension can be easily overlooked. Besides, the 2D-LC methods are more time-consuming and require elaborate instrument set up and specialists. 27
1.5 Monolith columns
In the beginning of the 1990s monolith columns were introduced by Svec et al. 30 and Hjertén et al.
28,29 as an alternative to particle packed columns for low pressure chromatography. Originally the
monoliths were called a continuous polymer bed, because it can be seen as a continuous rod with large though-pores in a canal-like structure and smaller pores in a skeletal structure. Monoliths are usually prepared in one phase of the monomer mixture within a mold, such as a capillary.30
1.5.1 Characteristics of polymeric monoliths
Monoliths are of great interest because of their high permeability in combination with maintained efficiency, low resistance to mass transfer, and fast and simple in situ preparation within micro- or nano formats. This makes monoliths highly suitable for protein separation applications. Alternatively to packed columns, polymeric monoliths can offer an excellent solution for separation of large and complex proteins such as immunoglobulins.1 Branovic et al.31 show how to isolate immunoglobulin from a solution with a high abundance of albumin and transferrin using monolithic disks. An affinity chromatography type of disk was used to capture immunoglobulin whereas simultaneously albumin and transferrin were retained on an anion exchange chromatography type of disk.
As mentioned earlier, a significant advantage of monolithic columns is the permeability. To move mobile phase go through the column, a pressure (ΔP) is required. Assuming the velocity (u0) is the
velocity of unretained species, η is the viscosity of the mobile phase, 𝜀" is the total porosity, L is
the length of the column, and 𝐾$,& is the superficial velocity-based columns permeability Equation
1 follows:
∆𝑃 = 𝜂𝜀"𝑢-𝐿 𝐾$,&
Equation 2: Equation of the relation between e.g. the pressure, the permeability and velocity
For packed columns the permeability is defined by the specific permeability factor for packed columns (Kf) and the average diameter of the particles(𝑑$). The specific permeability of a monolith
column is undefined, however often Equation 1 and Equation 2 are used to compare the permeability of packed columns to monolithic columns and determine what particle size of the packed column is required to achieve the same permeability.32
𝐾$,& = 𝐾1𝑑$2
Equation 2: Equation of the relation between the permeability and the average diameter of the packed particles
Monoliths also offer good robustness, since the monolith is one continuous structure that can be covalently linked to the column wall and the possibility to tune the morphology such that the best compromise between efficiency and permeability can be obtained. 33 An additional advantage of polymeric monoliths as stationary phase compared to particle packed columns has to do with mobile phase mass transfer which reduces band broadening and the loss of efficiency and resolution.34 The through-pore size and skeletal dimensions can be varied independently. This is a
great opportunity to the engineer to develop monolith stationary phase with high porosities to reduce the resistance to mass transfer without decreasing column permeability. 35 This is beneficial compared to porous particle packed columns, the way to improve mass transfer would be to reduce the particle size and diffusional path lengths within the pores. However, this improvement would lead to a decrease in permeability and an increase of back pressure of the column. 33 High through put screening and analysis can also be achieved with monolithic materials by their low flow resistance at high flowrates.17
A drawback of polymeric monoliths as stationary phase can be the swelling and shrinking properties as the organic solvent in the mobile phase changes. 32 Also, polymeric monoliths can be limiting because there is a need for re-optimization of the polymerization conditions for each new system to achieve the desired macro porous structure. 33 However, one of the major concerns is the column-to-column repeatability. This is more difficult to achieve with monolith columns compared to packed columns because the packing of particle-based columns occurs all at the same time. The small difference in the polymerization mixture for monolith columns are of great influence in the end result of the monolithic structure. However, there are techniques to address these challenges as evidenced by the successful commercialization of polymeric monolithic columns by a number of companies. 36
1.5.2 Organic and inorganic monoliths
Monoliths used for intact protein separations can be organic-based, inorganic-based, or hybrid which is a combination of both. For inorganic monoliths compounds such as silica, zirconia, ceria, nickel and iron can be used as materials for the monolith, however silica is mostly used.32 Dependent on the materials used, the columns can have different pore morphologies and have been shown to be efficient for the separation of small molecules. 37
For protein separation, organic polymer monoliths are mainly used. These monoliths have large pores and a ‘cauliflower’ structure that are especially suitable for macromolecule separations. The preparation of organic monoliths starts with combining a mixture of monomers, porogenic solvents and an initiator as a polymerization mixture. Commonly used monomers are methacrylate, acrylamide and styrene based and can be chosen for their specific characteristics. 1-Propanol and 1,4-butanediol are commonly used porogens. Porogenic solvents are added to the polymerization mixture to help control the porosity of the column. This also allows the components to form one
homogenous solution. A cross linker is a monomeric unit with more than one double bond and is integrated to allow monoliths to grown in more than one direction and form an interconnected crosslinked polymer. 32
1.6 Components in the polymerization mixture
1.6.1 Monomer
Monomers and corresponding properties determine the choice of porogens, and therefore the structure of the monolith. The monomers determine the chemical functionality of the monolith. The final polymer properties can even be influenced by the monomer concentration. Hebb et al.38 showed the effect of monomer concentration for poly (trimethylolpropane trimethacrylate)(poly-TRIM) monoliths. At low monomer concentrations the polymeric monolith was not formed, whereas with a high monomer concentration the material was denser and friable.
1.6.2 Cross linker
A cross linker is a monomer which helps to make a rigid structure within the monolith. Also, the cross linker has influence on both the porous properties and chemical composition within the monolith. An increase of cross linker in the monomer mixture leads to a decrease in pore size due to early formation of highly crosslinked micro clusters. 39 One of the most popular crosslinking agents is ethyl dimethacrylate (EDMA) for the preparation of rigid macro porous methacrylate polymeric monoliths.39 The rigidity and homogeneity of the monolith may also be affected by the percentage of cross linker, in general the monolith structure will become more rigid when the cross linker ratio is increased. 40
1.6.3 Porogen
The most powerful ingredient to develop the desirable porous structure are porogenic solvents and their ratios in the polymerization mixture. What makes it powerful is that sometimes by changing the porogen only the porous structure is affected, whereas other parameters such as monomer/cross linker ratio or initiator can change the composition of the monolith structure. While in other situations it is also possible to change the exposed chemistry of the polymer and therefore the surface chemistry is changed.
The mechanism of pore formation using porogenic solvents can be explained as follows. The monomers are more thermodynamically favorable for the polymer than the porogens. As a result, the polymer is preferable swollen with monomers and since the monomer concentration within
the polymer is higher than in surrounding solution, further polymerization is kinetically preferred. Further polymerization will create micro globules that can form clusters. The final morphological structure is created by the continuous grow and crosslinking micro globules. The polymerization process ends with a two-phase system containing the solid monolith and inert liquid porogen filling the pores. 40
Porogens can either be good or poor solvents for the polymer. Initially the polymerization takes place in a homogenous polymerization mixture to grow crosslinked polymer chains. This continues until the polymer chains precipitate. When this moment of precipitation happens, is dependent on the porogen. If the porogen is a good solvent for the final polymer, this moment for phase separation will occur later and this will result in smaller pores. 40–42 A change in composition of the porogens affects the solubility of the growing polymer chain at the early stages of the polymerization reaction. If a poor porogenic solvent is used, earlier phase separation will occur which results generally in larger pores. Larger pores are also obtained by a higher percentage of porogens relative to the monomer ratio. 40
Another approach is to use polymeric porogens such as poly(ethylene glycol) (PEG). Courtois et al.43 demonstrated how the mean pore diameter was increased when longer PEG chains were used. The mean pore diameter increased from 0.2 μm when PEG Mw 4000 g/mol was used to 1 μm for PEG Mw 20 000 g/mol. In this way the pore size and permeability were increased, but the functional morphology stayed the same.
1.6.4 Initiator
To start the formation of the polymeric monolith, an initiator is added to the polymerization mixture to initiate the polymerization reaction. For example, this happens when the initiator decomposes, and a free radical is generated (free radical polymerization). Decomposition can happen mainly thermally or UV-initiated. These two are the most commonly used, so therefore only discussed here. As the free radicals initiate the polymerization, the polymer chain starts to grow. In the reaction mixture the solubility decreases, and the polymer chains start to precipitate.40 The concentration of free radicals that are generated influence the polymerization kinetics and thus the final monolithic structure obtained.39
When the polymerization is initiated thermally, commonly done with azobisisobutyronitrile (AIBN), the temperature of the polymerization process has to be controlled precisely in order to obtain
monoliths with reproducible and uniform structure. The temperature has a significant effect on the growth of the polymer. At higher temperatures, the initiator will decompose faster, also the rate of distribution and the number of growing nuclei will become larger, which results in smaller final pore size. 39 Another risk at high temperature is that fast polymerization can cause less uniform porous structure. Although it seems that a low temperature is preferred, the decomposition temperature of the initiator should always be taken into account. 40,44
In contrast to thermally initiated polymerization, UV-initiated polymerization is usually performed at room temperature. This makes UV-initiated polymerization suitable for volatile solvents with a low boiling point. Solvents such as ethanol, tetrahydrofuran, acetonitrile, chloroform, ethyl acetate and hexane can in this way also be used as porogens for the preparation of polymeric monoliths. Also compared to thermally initiated polymerization, UV-initiated polymerization is very fast where polymerization can be achieved in a few minutes. The main factor affecting the photo polymerization is the light intensity, which influences the free radical concentration.40
1.7 Project aim
The aim of this project is to develop a MS-compatible method to characterize intact proteins with the use of polymeric monoliths. In order to maintain the proteins in their native forms, the possibilities within the soft interaction chromatographic methods such as hydrophobic interaction chromatography (HIC) and ion exchange chromatography (IEC) are explored. This holds a great challenge because of the salt incompatibility of the chromatographic system to the mass spectrometry. Therefore, the possibilities of using volatile salts for HIC and IEC is explored.
To support the protein retention with volatile salts a variety of monolith capillary columns will be synthesized by free radical polymerization. The morphology and difference in protein retention will be observed, and difference will be noted and where possible explained. The polymeric monolith stationary phase should be able to give a reproducible method and give the optimum conditions between permeability and efficiency. Besides the chromatographic results, scanning electron microscopy (SEM) will be used to obtain information about the materials morphology.
These new polymeric monolith stationary phases will be used for the separation of immunoglobulin G subclass 1 and its variants λ and κ. The IgG1 variants λ and κ are very similar so that demands for a specific separation method.
Chapter 02: Experimental
This chapter briefly describes the materials, instrumentation and general procedures that are used throughout this research.
2.1 Chemicals
All chemicals used in this work are listed in Table 2.1
Table 2.1: List of the used chemical in this project
Chemical name Grade Supplier
Acetic acid ≥99% Sigma Aldrich
Acetone >98% Chem-Supply
Aluminiumoxide - Sigma Aldrich
Ammonium acetate ≥98% Sigma Aldrich
Ammonium bicarbonate >99% Sigma Aldrich
Ammonium hydroxide 1M ACR Chemical Reactants
Ammonium phosphate dibasic dihydrate
>98% Sigma Aldrich
Ammonium sulfate >99% Sigma Aldrich
2,2'-Azobis(2-methylpropionitrile) 12 wt.% in acetone Sigma Aldrich α-Chymotripsinogen A, from bovine
pancreas
>95% Sigma Aldrich
1,4-Butanediol >99% Sigma Aldrich
Bovine Serum Albumin ≥98% Sigma Aldrich
Butyl acrylate 98% Sigma Aldrich
Cytochrome C, from equine heart ≥95% Sigma Aldrich
1-Decanol 99% Sigma Aldrich
Dichloromethane 99% Sigma Aldrich
2,2-Dimethoxy-2-phenylacetophenone
99% Sigma Aldrich
Ethanol 100% Chem-Supply
Ethylene glycol dimethacrylate 98% Sigma Aldrich
Hydrochloric acid 37% Sigma Aldrich
IgG1 lambda from Human Plasma >95% Sigma Aldrich
IgG2 lambda from Human Plasma >95% Sigma Aldrich
IgG from Human serum ≥95% Sigma Aldrich
Lysozyme, from chicken egg white ≥90% Sigma Aldrich
Methanol 99,7% Sigma Aldrich
Myoglobin, from horse heart ≥90% Sigma Aldrich
N,N-dimethyl-N- methacryloyloxyethyl-N-(3-sulfopropyl)ammonium betaine
95% Sigma Aldrich
Poly ethylene glycol Mw 10.000 g/mol
- Sigma Aldrich
Poly(ethylene glycol) diacrylate Mw 258 g/mol
- Sigma Aldrich
1-Propanol >98% Sigma Aldrich
Sodium carbonate ≥98% Sigma Aldrich
Sodium hydroxide ≥98% Sigma Aldrich
Sodium phosphate dibasic
dihydrate
≥98.5% Sigma Aldrich
Sodium sulfate (anhydrous) ≥99% Sigma Aldrich
3-(Trimethoxylsily) propyl methacrylate
≥98% Sigma Aldrich
2.2 Equipment
2.2.1 Capillary liquid chromatography
Chromatographic separations were performed using an Agilent 1260 Infinity Series capillary micro-LC system equipped with 1260 Cap Pump (G1376A) including the 1260 HiP Degasser (G4225A), 1260 DAD VL+ Detector (G1315C), a flow cell (G1376-60001) and a 1260 HiP ALS Auto sampler with 1 μL loop 9 (G1377A). Chemstation software (Agilent LC, online) was used for system control and Agilent LC (offline) was used for data processing (data collection rate 2.5 Hz). Chromatograms were converted into CSV files and imported in Microsoft Excel.
Scanning electron micrographs were recorded on Zeiss Gemini SEM as part of Adelaide Microscopy, operated under high vacuum mode with an acceleration voltage of 2 kV. Secondary electrons were detected. Specific surface area and pore size distribution was determined by nitrogen
adsorption/desorption at 77K by Brunauer-Emmet-Teller (BET) Tristar II (Micrometrics) in the dry state. The pore size was assessed by using the non-localized density functional theory (NLDFT). UV polymerization was performed using 260nm UV lamp (OAI 500W DUV) with lamp calibrations to 20.0 mW/cm2, which were performed with OAI Model 308 Intensity meter with 260nm probe.
2.3 Preparation of capillary columns
2.3.1 Surface modification of fused-silica capillaries
The polyimide-coated and Teflon capillaries were both surface modified based on a procedure by Rohr et al. 45 The capillaries were first rinsed with acetone, followed by MiliQ water. Then a solution of 0.2M sodium hydroxide was pumped through the capillary with a syringe pump at a flowrate of 30 μl/h for 30 minutes. The capillaries were washed with MiliQ water before a 0.2M hydrochloric acid solution was pumped through the capillaries with the same flowrate of 30 μl/h for 30 minutes. The pH was measured at every step. The capillaries were rinsed with water and ethanol at pH 5. A solution of 20 wt.% 3-(trimethoxysily) propyl methacrylate in ethanol pH 5 was pumped through the capillaries at 30 μl/h for 1 hour. At last the capillaries were rinsed with acetone, purged with nitrogen gas and left at room temperature for at least 24h.
2.3.2 General procedure of the purification of monomers
All the monomers that are used in chapters 3,4 and 5, are purified regarding the following method. The PEGDA monomer is one of the exceptions, as described in this chapter as well. SPE is the other exception, because it is a solid, it does not have any inhibitors and therefore does not need to be purified. A column was made of a glass pipette with glass wool with aluminum oxide on top. The monomer was purified by letting it flow though the column and collecting it at the bottom. To speed up the process nitrogen gas can be used.
2.3.3 Purification of the poly(ethylene glycol) diacrylate monomer
The PEGDA monomer was purified to remove impurities and inhibitors based on the procedure described by Lui et al. 46 Here 50 mL of PEGDA (Mw 258 g/mol) and 30 mL of saturated solution of sodium carbonate were added to a 250 mL separation funnel and shaken vigorously. The funnel was placed on a clip and phase separation occurred. The lower layer (carbonate solution) was removed, this washing step was repeated twice more, followed by a washing step of three times 25mL of MiliQ water to remove residual carbonate solution. In this step the upper layer was removed. Then the PEGDA monomer was extracted from the remaining water layer using twice 50
mL of dichloromethane. After the extraction with dichloromethane no real water layer was visible anymore, so sodium sulfate was added to remove the remaining water that was left in the monomer. This was filtered through Whatman cellulose-based filter paper. The dichloromethane was removed by rotary evaporation.
2.3.4 Preparation of the Hydrophobic Interaction Stationary Phase
For the hydrophobic interaction chromatography, a stationary phase is prepared based on the PEGDA monomer (poly(ethylene glycol) diacrylate Mw 258 g/mol). 3 The PEGDA-monoliths are all prepared in modified polyimide-coated capillaries with an internal diameter of 200 μm.
The polymerization mixture was a combination of the monomer, porogen and initiator. All polymerization mixtures contained 1.00 g of PEGDA 258, 1.00 g of PEG 10.000 (20 wt.% in methanol), 0.5g of decanol and 0.01 g of AIBN (respectively 1 wt.% of the monomer). The mixture sonicated for at least 30 minutes and degassed with nitrogen gas before the capillaries were filled. The capillaries were then together with the remaining polymerization mixture placed in a water bath of 60°C for at least 20 hours.
After polymerization the capillaries were rinsed with methanol with a syringe pump at 2 μL/min for 1 h. The remaining mixture was washed by a Soxhlet apparatus using methanol for 48h to remove the residual porogen. Finally, the mixture was left in a vacuum oven for at least a week.
2.3.5 Preparation of the Ion Exchange Chromatography Stationary Phase
For ion exchange chromatography, a stationary phase is prepared based on the zwitterionic sulfoalkylbetaine monomer (N,N-dimethy –N– methacryloyloxyethyl –N - (3-sulfopropyl) ammonium betaine)4. The sulfoalkylbetaine monoliths polymerized with photo initiator are all prepared in modified Teflon capillaries with an internal diameter of 100 μm.
The polymerization mixture was a combination of the monomers, porogen and initiator. The polymerization mixture was prepared as described here, unless stated different elsewhere. The mixture contained 3 units of monomer mixture which consisted of 47 wt.% sulfoalkylbetaine monomer and 53 wt.% of EDMA and 7 units of methanol. DMPA is added with an amount of 1 wt.% with respect to the monomer mixture. The mixture was sonicated for at least 30 minutes and degassed with N2 for 5 minutes. The remaining polymerization mixture was poured in between two
glass slides to polymerize as well for material characterization. The photo polymerization took place at 260 nm for 15 minutes. After polymerization the columns were flushed overnight with water by the LC-system and syringe pump.
When instead of photo polymerization, thermal polymerization was used, the capillaries were changed for modified polyimide-coated capillaries with an internal diameter of 200 μm, initiator was changed to AIBN and the capillaries were placed in a 60°C water bath for at least 20h.
2.4 Capillary liquid chromatography experiments
The LC experiments were performed under gradient conditions and used 100 nL injections with the aid of an auto sampler. UV detection was employed at 214 nm and 280 nm. Buffers were prepared by appropriate dilution of salts in MiliQ water (18MΩ cm, Millipore) and filtered with 45 μm cellulose membranes (Millipore) before being degassed, on the degas cycle of an ultrasonic cleaner (Branson 1860) to remove air bubbles. The pH of the mobile phases was measured with Oaklon PC2700. All proteins were diluted with MilliQ water to the appropriate concentrations for storage and further diluted in the appropriate mobile phase for analysis.
Permeability measurements were performed by recording the column back pressure at various flowrates between 0.5 μL/min and 5.0 μL/min in methanol and water. The pressure was allowed to stabiles for 5 minutes before pressure was recorded.
Chapter 3: Polymeric monoliths for hydrophobic interaction chromatography for the separation of intact proteins
3.1 Introduction
Recently, hydrophobic interaction chromatography has increased in importance for the characterization of bio therapeutics, also allowing to obtain complementary information to reversed phase chromatography. 13 Hydrophobic interaction chromatography is commonly used for separation of monoclonal antibodies and other intact proteins because of the soft interactions which allow the functionality of the protein to be preserved.47 To promote the soft interactions for hydrophobic interaction chromatography the stationary phase should contain slightly hydrophobic ligands, these ligands are often short alkyl-based or poly (ethylene glycol) (PEG)-based. This is combined with a mobile phase, which often consists of a salt solution.
This chapter will discuss the use of polymeric monoliths for hydrophobic interaction chromatography with volatile salts as mobile phase for protein separation.
3.2 Hydrophobic interaction between the stationary phase and proteins
The ideal stationary phase for hydrophobic interaction chromatography for the separation of proteins should only interact with the hydrophobic ligands present.48 Back in 1948, the first method based on hydrophobic interaction was discussed by Tiselius about the absorption of proteins on a silica gel under the presence of salts. They called this principle salting-out chromatography. 49 The mechanism of HIC is based on the interaction between hydrophobic patches of proteins and weakly hydrophobic ligands in the stationary phase. 50 Proteins will create a shield of ordered water molecules when they are in an aqueous environment, which prevents hydrophobic interaction with the stationary phase. When salts are introduced in the mobile phase, they will compete for water molecules with the proteins, as a result the ordered water layer will break down. This promotes the hydrophobic interaction between the hydrophobic patches of the protein and the stationary phase. Protein elution is based on the difference in available hydrophobic surface of the protein which is achieved by a decreasing salt gradient in the mobile phase over time. 13 HIC separations typically have aqueous buffers as mobile phase in combination with a linear decreasing salt gradient starting at a high salt concentration.50
The hydrophobic ligands are often present in a lower concentration compared to reversed phase stationary phases. The hydrophobic patches of the protein will bind strongly to hydrophobic ligands, if the hydrophobic ligand of the stationary phase is large, as in reversed phase chromatography, a stronger mobile phase is required to elute the proteins. These kinds of strong mobile phases often contain organic solvents, which causes denaturation of the proteins. 13
Baca et al.50 compared different commercially available HIC columns for protein separation based on effects of eluent type and concentration on protein retention. As expected, the chemistry of the ligand of the stationary phase and the type of salt and its concentration were the two biggest effect.
3.3 Hydrophobic interaction chromatography coupled mass spectrometry
The non-denaturation method of HIC could be used for the direct coupling with the mass spectrometer as HIC is complementary to reversed phase chromatography and has high resolution.51 Usually in HIC, a decreasing salt gradient of a salt which stands high in the Hofmeister series is used. The Hofmeister series scales the salts in kosmotropic salts on one hand and chaotropic salts, on the other hand. Kosmotropic salts have a higher polarity than chaotropic salts and interact strongly with water. As a result, a large aqueous layer is formed around the kosmotropic salt, leaving the hydrophobic patches of the stationary phase exposed, to promote hydrophobic interaction between the analytes and stationary phase. Chaotropic salts disrupt hydrogen bonding and reduce the hydrophobic effect and therefore weaken the hydrophobic interaction between the analytes and stationary phase. 50
Commonly used salts in HIC, including sulfate and phosphate salts, are kosmotropic salts and stand high in the Hofmeister series. Unfortunately, these salts are non-volatile and therefore not compatible with mass spectrometry. Ammonium acetate is often used as buffer with mass spectrometry because it is a volatile salt. However, it provides inadequate protein retention when used with conventional HIC materials.14
Already in the early days of HIC, in 1990, Nakamura et al.52 found that in order to use volatile salts for hydrophobic interaction chromatography, the chemistry of the column should be more hydrophobic to have sufficient protein retention. Their work showed protein separation with ammonium acetate as volatile salt when a less hydrophilic polymethacrylate-based column is used instead of a hydrophilic silica column.
Chen et al.2 show a potential solution to use ammonium acetate in a HIC separation. Instead of using just conventional HIC materials such as POLYPROPYL A columns, different alkyl chain lengths were investigated to observe if the change in chemistry causes more retention between the stationary phase and proteins. By increasing the hydrophobic character of the column, a point was reached where they were able to retain a series of proteins and elute them using a concentration of 1M or less of ammonium acetate.
Follow up research by the same group focused not only on a mixture of test proteins, but also on intact monoclonal antibodies such as immunoglobulins.53 As hydrophobic interaction chromatography remains a powerful approach to analyze monoclonal antibodies 54. Chen et al.53 demonstrated an online HIC-MS analysis of intact monoclonal antibodies using quadrupole-time-of-flight (Q-TOF) and Fourier transform ion cyclotron resonance (FTICR) mass spectrometers, showing the difference in hydrophobicity and glycosylation of proteoforms of the monoclonal antibodies.
3.4 PEGDA polymeric monoliths as stationary phase
Poly (ethylene glycol) -based monomers are highly suitable for hydrophobic interaction-based protein separation. This has to do with the character of the PEG polymeric structure. PEG is a hydrophilic polymer that can undergo phase separation in the presence of kosmotropic salts. Under these salt conditions, the PEG polymer can have a hydrophobic character.55 This reversible change between hydrophilic and hydrophobic can be extremely useful for hydrophobic interaction-based separation of proteins. Under a high salt concentration, the PEG-based polymer has a hydrophobic character and will bind to the proteins. However, under low salt concentrations, the PEG-based polymer has a hydrophilic character and will completely release the proteins. This could be beneficial over traditional hydrophobic stationary phases as phenyl, butyl and octyl which can give poor release of the proteins.55 Li et al.48 demonstrated that monoliths prepared poly(ethylene glycol) diacrylate (PEGDA) or poly(ethylene glycol) methyl ether acrylate (PEGMEA) as monomer and cross linker showed negligible nonspecific adsorption of large proteins e.g. bovine serum albumin.
Figure 3.1: Structure of poly (ethylene glycol) diacrylate (PEGDA) monomer
H2C O O O CH2 O n
Yu et al.55 show that PEG-based monomers are also suitable in paper membrane. When the PEG monomer is grafted on paper-based membrane, it can be an affordable alternative for the high-resolution purification processes of monoclonal antibodies. In other work of Xin et al.56 they used polymeric monolithic capillaries for solid phase extraction based on acrylate monomers.
Aggarwal et al.35 show that poly (ethylene glycol) diacrylate (PEGDA) monoliths are well suitable for the separation of small molecules. Also, PEGDA monoliths have been shown excellent results for the separation of proteins. 36,30
Another characteristic of PEGDA as capillary monolith shows to be excellent for hydrophobic interaction chromatography, this has to do with its hydrophilic character, proteins do not adsorb on the ethylene glycol structure, which makes PEGDA highly compatible for the separation of proteins. 30 Also, this slightly hydrophilic matrix makes it an excellent choice for hydrophobic interaction chromatography. Biomolecules tend to maintain their native conformation when there is interaction with the PEG-stationary phase and it typically do not affect their biological activities if appropriate conditions are used.3
Li et al.36 compare different PEGDA molecular weights ranging from PEGDA Mw 258 to 700 g/mol and they conclude that the PEGDA Mw 258 g/mol monolith column represented the best result with regard to peak capacity, resolution and peak shape. Desire et al.3 reported a method to prepare successfully PEGDA-based monolithic column for hydrophobic interaction chromatography. This method is used for the separation of the different subclasses and variants of Immunoglobulin G. However, the use of a sulfate salt in the mobile phase does not allow the coupling with mass spectrometry.
To the best of my knowledge no work has been done using monolithic columns for coupling HIC to mass spectrometry. As discussed before, PEGDA shows promising results as stationary phase in a capillary column. Monolithic columns are flexible, and the monolith structure can be adjusted by small changes in the polymerization mixture. Changing the polymeric structure can be of great value, because by changing the chemistry, it is possible to change the hydrophobicity. A higher hydrophobicity could be of importance when a different salt is used. Traditionally, sulfate salts are used for hydrophobic interaction chromatography, which are salts with a high ionic strength. When there is a switch towards ammonium acetate, which has a lower ionic strength, the interactions
might not be strong enough to retain the proteins on a column which was suitable for sulfate buffers as mobile phase. The aim is to find the right chemistry within the stationary phase with an increase in hydrophobicity, so that proteins are retained, and eluted with a relatively low concentration (~1M) of ammonium acetate.
3.5 Hydrophobic interaction chromatography on PEGDA column
Here, the first step was to prepare a PEGDA column as described in Chapter 2. The polymerization of the PEGDA monolith is based on free radical polymerization. The polymerization mixture consisted of the four different components. Poly (ethylene glycol) diacrylate Mw 258 g/mol as monomer, PEG10.000 in methanol and 1-decanol as porogens and AIBN as initiator. To start the polymerization, energy is added thermally by putting the capillaries in a 60° C water bath. With an increase in temperature AIBN will introduce free radicals which will promote polymerization. As PEGDA has the double acrylate groups on either side it works as a cross linker as well. A somewhat rigid structure was formed inside the capillary, when this monolith column was washed with methanol for the porogens to be flushed out and the crosslinked structure of PEGDA remains in the capillary.
3.5.1 Permeability measurements of PEGDA columns
The permeability of the capillaries was measured by observing the pressure at different flowrates. The comparison of the different permeability for the different columns is shown in Figure 3.2. The PEGDA and PEGDA-co-butyl acrylate columns were used for permeability testing. PEGDA-co-butyl acrylate columns were prepared to investigate if a more hydrophobic character could be obtained. The steeper the slope of the graph, the higher the pressure at a certain flowrate so the less permeable the column is. Figure 3.2 shows that the original PEGDA column is less permeable than the PEGDA-butyl acrylate column. The PEGDA monomer can cross link with its own structure to form a rigid crosslinked polymeric monolith, if the amount of PEGDA in the monomer mixture is decreased, less crosslinking can occur, and this could give a possible explanation why the monolithic structure with PEGDA-co-butyl acrylate (10 wt.%) is more permeable.
A column should have a certain permeability such as the mobile phase and analytes can easily flow through the column and higher flowrates and pressures can be achieved. However, a high permeability might look preferable for fast separations, it also has its limits. Often a higher
permeability comes from less functional stationary phase, with less interaction sites. Another factor is the surface area, the higher the surface area, the more rigid the structure. So the monolith structure ideally should have the right balance between permeability, surface area and interaction sites. 33,57
Figure 3.2: Permeability of both PEGDA (black) and PEGDA-co-butyl acrylate (10wt%) (grey) stationary phase. The permeability is shown by the pressure (y-axis) at different flowrates (x-axis).
3.6 Hydrophobic interaction with sulfate salts as mobile phase
First, similar conditions were applied as shown in the work of Desire et al.3 These conditions included a HIC separation with a PEGDA column described as above and based on a salt gradient starting from 3M ammonium sulfate in buffer solution to only phosphate buffer. This resulted in a successful separation of a mixture of five different proteins. The set of test proteins consisted of cytochrome C, myoglobin, ribonuclease A, lysozyme and α-chymotripsinogen and are identified in Figure 3.3. 0 10 20 30 40 50 60 70 80 90 0 1 2 3 4 5 6 Pr es su re (b ar ) Flowrate (μL/min)
Figure 3.3: Chromatograms showing the separation of a protein mixture with the identification of a) Cytochrome C b) Myoglobin c) Ribonuclease A d) Lysozyme e) + f) α-Chymotripsinogen. Conditions: 16 cm x 200 μm i.d. column, PEGDA monolith as stationary phase; Mobile phase A was mobile phase B with 3M (NH4)2SO4, mobile phase B was 0.1M Na2HPO4•H2O in MiliQ water; 15-minute
linear gradient from 0%B to 100%B, then isocratic elution at 100%B for 5 minutes before returning to 0%B in 5 minutes. Allow the system to re-equilibrate for at least 15 minutes; Flowrate 2 μL/min; injection volume 100 nL; protein concentration 0,1 mg/mL; detection at 214 nm.
From this result can be stated that a PEGDA polymeric monolith is sufficient for protein retention under high salt concentrations. However, these high concentrations of sulfate and phosphate salts are non-volatile and therefore, this method cannot be coupled to the mass spectrometer.
3.7 HIC separation with ammonium acetate as mobile phase
The following step was to investigate the possibilities of protein retention on the PEGDA column with a volatile salt. A commonly used volatile salt is ammonium acetate. 58 Individual proteins were used to observe the retention with the PEGDA column when ammonium acetate is used as mobile phase. Figure 3.4 shows the retention pattern of different proteins when ammonium acetate is used as salt in the mobile phase. It seems that the ionic strength of ammonium acetate is not strong enough to retrieve protein retention when a PEGDA column is used. As a result, none of the test proteins gave any retention and all proteins eluted at the void volume.
0 50 100 150 200 250 300 350 0 5 10 15 20 25 30 35 40 45 Si gn al (m AU ) Time (minutes) c) d) b) a) e) f)
Figure 3.4: Chromatograms showing the retention under HIC conditions of Immunoglobulin (a), Ribonuclease A (b) and Bovine Serum Albumin (c), with the blank (grey) and the protein (black) chromatograms on a PEGDA-based stationary phase. Conditions: 16 cm x 200 μm i.d. column, PEGDA monolith as stationary phase; Mobile phase A was 2M NH4OAc and mobile phase B was 20mM NH4OAC,
both in MiliQ water; 15-minute linear gradient from 0%B to 100%B, then isocratic elution at 100%B for 5 minutes before returning to 0%B in 5 minutes. Allow the system to re-equilibrate for at least 15 minutes; Flowrate 2 μL/min; injection volume 100 nL; protein concentration 0,1 mg/mL; detection at 280 nm.
There is a possibility to use ammonium acetate as mobile phase. Chen et al. 2 show when a POLYHEPTYL column is used instead of a traditional POLYPROPYL and POLYBUTYL columns the three test proteins of lysozyme, ribonuclease A and α-chymotripsinogen show slightly more retention. If PEGDA is compared to the traditional POLYPROPYL or POLYHEPTYL it is already more hydrophilic because of the oxygen present. The hydrophobicity can be measured by the partitioning coefficient logP. LogP is a partitioning factor which indicates whether the molecule has a higher affinity for an aqueous or organic phase. A positive logP value means that the analyte has a higher affinity for the organic phase. A logP value of 0 means that the analyte is equally partitioned between the two phases. A negative value means that the analyte rather stays in the aqueous
0 10 20 30 40 50 Si gn al (m AU ) Time (minutes)
IgG
0 10 20 30 40 50 Time (minutes)Rionuclease A
0 10 20 30 40 50 Time (minutes)phase. The partitioning factor can also be referred as Kow or Pow. 59 The XlogP3 is a more evolved method for determine the partitioning factor by a computational algorithm. 60 For example a poly(propyl) structure has a XlogP3 of 3.84, while poly(ethylene glycol) diacrylate has only a XlogP3 of 1.2, note here that trimethyl (propyl) silane is used as reference.
3.7.1 Modification of the chemistry within the monolith
In order to achieve protein retention, the monolith stationary phase should have a more hydrophobic character. The hypothesis is that if the chemistry causes a higher hydrophobicity, a salt with a lower ionic strength can still be used for protein retention. To achieve this the polymerization mixture was changed by added different percentages of 1wt%, 5wt% and 10%wt buytl acrylate with respect to the amount of PEGDA as monomer. All the other conditions were kept the same. The monolith structure was kept as similar as possible by using an acrylate-based alkyl chain which is expected to polymerize the same way as the PEGDA monomer. Also, a relatively low weight percentage of butyl acrylate is added to the polymerization mixture.
Figure 3.5: Structure of butyl acrylate
First separation was performed under the same conditions as with the standard PEGDA-column separation, as is shown in Figure 3.6. Although the same protein mixture was injected, slightly different peak shapes were observed. Mainly the peak identified as Cytochrome C did not show a consistent appearance. This could possibly be explained by the changing chromatographic behavior of Cytochrome C. The protein cytochrome C can show different retention even though there is a high degree of homology. 61
H2C
O O
If Figure 3.3 and Figure 3.6 are compared, similar retention times are observed. However, the only condition that is different compared to the separation described in Figure 3.3, is the flowrate. The flowrate in this condition is increased to 4 μL/min to improve peak shape by reducing diffusion. A higher flowrate also introduces shorter retention times. Because of this, it can be assumed that the proteins have more hydrophobic interaction with the stationary phase if butyl acrylate is included in the monolith structure.
Figure 3.6: Chromatogram showing the separation of a protein mixture that consists of Cytochrome C a) Myoglobin b) Ribonuclease A c) Lysozyme d) + e) α-Chymotripsinogen on Conditions: 15 cm x 200 μm i.d. PEGDA with 10% butyl acrylate monolith as stationary phase; ; Mobile phase A was mobile phase B with 3M (NH4)2SO4, mobile phase B was 0.1M Na2HPO4•H2O in MiliQ water; 15 minute
linear gradient from 0%B to 100%B, then isocratic elution at 100%B for 5 minutes before returning to 0%B in 5 minutes. Allow the system to re-equilibrate for at least 15 minutes; Flowrate 4 μL/min; injection volume 100 nL; protein concentration 0,1 mg/mL; detection at 214 nm
The possibility to use this PEGDA-co-butyl acrylate stationary phase for hydrophobic interaction chromatography with ammonium acetate as salt in the mobile phase is also investigated. However, it did not seem to be hydrophobic enough to retain the different proteins on the column with an ammonium acetate solution as mobile phase. The proteins all eluted around the void volume similar to the original PEGDA column (Figure 3.7).
0 50 100 150 200 250 300 350 400 0 5 10 15 20 25 30 35 40 45 Si gn al (m AU ) Time (minutes) a) b) c) d) e)
Figure 3.7: Chromatograms showing the retention under HIC conditions of Immunoglobulin(IgG), Ribonuclease A (RiA) and Bovine Serum Albumin (BSA), with a PEGDA-based stationary phase. Conditions: 16 cm x 200 μm i.d. column, PEGDA-co-butyl acrylate (10%) monolith as stationary phase; Mobile phase A was 2M NH4OAc and mobile phase B was 20mM NH4OAc, both in MiliQ water;
15-minute linear gradient from 0%B to 100%B, then isocratic elution at 100%B for 5 15-minutes before returning to 0%B in 5 15-minutes. Allow the system to re-equilibrate for at least 15 minutes; Flowrate 2 μL/min; injection volume 100 nL; protein concentration 0,1 mg/mL; detection at 280 nm.
To make sure that the protein was eluting around the void volume one of the test proteins was used in different concentrations, assuming that the surface area of the peak increases with the concentration. This can be confirmed by Figure 3.8, where an increasing peak area is observed with an increasing concentration of the protein bovine serum albumin (BSA).
Figure 3.8: Chromatogram showing the retention under HIC conditions of BSA with a PEGDA-butyl acrylate stationary phase. Conditions: 15 cm x 200 μm i.d. column, PEGDA-butyl acrylate (10wt%) monolith as stationary phase; Mobile phase A was 2M NH4OAc and mobile phase B was 20mM NH4OAC, both in MiliQ water; 15-minute linear gradient from 0%B to 100%B, then isocratic
elution at 100%B for 5 minutes before returning to 0%B in 5 minutes. Allow the system to re-equilibrate for at least 15 minutes; Only the first 10 minutes are shown; Flowrate 2 μL/min; injection volume 100 nL; protein concentration 0,2 mg/mL and 0,4 mg/mL; detection at 280 nm. 0 5 10 15 20 25 30 35 Si gn al (m AU ) Time (minutes) IgG BSA RiA 0 2 4 6 8 10 12 Si gn al (m AU ) Time (minutes) Blank BSA 0.2 mg/mL BSA 0.4 mg/mL
3.7.2. Path towards a hydrophobic PEGDA column
As shown above, butyl acrylate does not provide sufficient hydrophobicity to obtain protein retention. Besides butyl acrylate, there are also other acrylate-based monomers with a longer alkyl chain available. Hexyl acrylate could be a promising alternative. This longer chain length could give the monolith structure a more hydrophobic character. Also, the tert-butyl monomer could be an interesting choice, because the tert-butyl structure gives a more hydrophobic character than its linear equivalent.
Another option could include changing the molecular weight of the PEGDA monomer. In this work only PEGDA Mw 258 g/mol is used, but different variants of the PEG chain are also available. Li et al.36 showed no big differences in retention between different PEGDA columns ranging from PEGDA 258 until PEGDA 700. The PEG chain length seemed to only have influence on the peak capacity, resolution and peak shape. However, this did not seem to have a great impact on the hydrophobicity and its correlation retention time.
The last option is to switch to a different volatile salt. Although the choice of volatile salts is not that elaborate, there are two more possible options. For separations which require a low pH, ammonium formate is worth mentioning. 52 Ammonium tartrate can be another possibility. Ammonium tartrate dissolved in an ammonium acetate buffer provided similar elution strength compared to ammonium sulfate. This mixture was only demonstrated to be MS compatible after a quick desalting step.14 Until this point, direct LC-MS has not been realized for HIC with ammonium tartrate as buffer.
Chapter 4: Polymeric monoliths for ion exchange chromatography for the separation of intact proteins
4.1 Introduction
Cohn et al. 62 showed the importance of ion exchange chromatography for biochemistry, already in 1949 by performing cation and anion-exchange for a trace-analysis study on formation and degradation of enzymes. Also in the recent years, ion exchange chromatography (IEC) is a widely applied method to control product quality of bio therapeutics. The fact that ion exchange chromatography has a broad range of separation conditions e.g. salt concentrations and pH values, makes it very versatile and attractive for protein analysis. 13
Ion exchange chromatography is a nondenaturing separation technique to separate protein charge variants based on increasing salt or pH gradient. If the separation is performed under a salt gradient, the mobile phase pH should be kept constant. An increasing salt concentration in the mobile phase will give a higher ionic strength. This promotes protein elution as the salt ions compete with the protein molecules for the charged sites on the stationary phase. Also, proteins can be separated based on a pH gradient through the column. 13 When pH gradients are generated with IEC, it is called chromatofocusing. Chromatofocusing is similar to iso-electrofocusing (IEF) elution, only in chromatofocusing the pH is externally generated and pressure-driven.63 Two modes of chromatofocusing can be applied i.e. cation chromatofocusing where the stationary phase has cation exchange properties with a pH ranging from a low pH to a high pH. Oppositely, anion chromatofocusing has anion exchange properties and will cause protein elution by a pH gradient going from a high pH to a low pH. The pH gradient allows proteins to focus and releasing them once the pH gradient reaches the pI of the protein.13,63
This chapter will discuss the use of ion exchange chromatography for the separation of immunoglobulins based on both salt and pH gradient with zwitterionic polymeric monoliths as stationary phase.
