W John Lough and Mark Carlile
School of Pharmacy and Pharmaceutical Sciences, University of Sunderland, Chester Road, Sunderland, SR1 3SD, UK
Chromatography is a term that is used to describe a group of separation techniques that utilise the distribution of molecules between a stationary phase and a mobile phase [1]. Separation is achieved by exploitation of a molecule's physicochemical properties that promote it to interact with the stationary phase or stay within the mobile phase. Interactions with the stationary phase promote a slower passage through the separation system. It is this distribution and the modulation of the distribution between these phases that makes chromatography a very powerful and versatile separation technique. Different types of molecule have different distribution constants and hence come off the column at different times. Thus, they are separated from one another before being measured. Importantly, this means that quantitative determinations based on chromatography can be optimised to give good specificity. In an analytical method validation to demonstrate that an analytical method is fit for its intended purpose, accuracy, precision, linearity, limit of quantitation and limit of detection may be evaluated, but it is the test for specificity that is often the most challenging of the validation tests [2]. Therein lies perhaps the main reason for the very widespread use of chromatography in biology.
The field of ‘biology’ is vast and complex. Within this there are a large and very diverse number of applications for which the power of chromatography may be effectively exploited. Consequently, the entire range of manifestations of ‘chromatography’ can find some use in biology. However, especially in the purification and analysis of intact, large biological molecules, liquid column chromatography (Figure 7.1) is by far the most common form of chromatographic separation used [3,4]. However, the application of gas‐phase chromatographic separation is also becoming more common, especially for the analysis and profiling of cellular metabolism [5].
The most common forms of liquid chromatographic separation (Table 7.1) are referred to as modes of chromatography, which differ in terms of their stationary phase chemistry and physical properties and the associated mobile phase (one stationary phase can be used for more than one mode of chromatography). The stationary phase is commonly regarded as the solid matrix (usually particles/beads) and its associated surface chemistry. However, this is almost always porous and the ‘stationary phase’ might better be regarded as the porous matrix and the stagnant mobile phase therein (making constituting up to 90% of the total ‘stationary phase’). The solid matrix is typically composed of regular beads of <1–100 μm in diameter. The beads are either xerogels (e.g. Sephadex™: able to shrink and swell dependent on the mobile phase) and aerogels (e.g. silica, porous glass: volume is independent of the mobile phase). Silica finds greater use for modern quantitative analysis (cf. preparative/purification) where there has been adaptation and modification of technology common in high performance liquid chromatography (HPLC) of small molecules. The choice of composition of the mobile phase is dependent on the mode of the separation method and the nature of the samples applied.
Table 7.1 Different modes of liquid chromatographic separation commonly used in the analysis of biological samples.
Physicochemical property | Mode of chromatography | Pressure format | References |
Size and shape | Gel filtration/size exclusion separation for proteins and nucleic acids | HIGH LOW |
[6–8] |
Net charge | Anion and cation exchange chromatography for proteins. Water softening and demineralisation. pH quenching |
HIGH LOW |
[4, 9, 10] |
Isoelectric point | Chromatofocusing of proteins | MEDIUM | [11,12] |
Hydrophobicity | Hydrophobic interaction and reversed‐phase chromatography of proteins, nucleic acids and small molecule drugs | HIGH LOW |
[13,14] |
Biological function | Affinity chromatography for proteins and nucleic acids. Immobilised ligand chromatography for proteins and nucleic acids. Immunoadsorption chromatography for proteins |
HIGH LOW |
[15–17] |
Carbohydrate content | Lectin affinity chromatography | HIGH | [18] |
Free thiols | Chemisorption (covalent) chromatography applied to proteins | HIGH | [19,20] |
Metal binding | Immobilised metal ion affinity chromatography for proteins | HIGH LOW |
[21,22] |
Chiral | Typically separation of pharmaceutical drug enantiomers but also biomolecules (e.g. d, l‐ amino acids were amongst the first analytes to be enantio‐resolved and there is a revival of interest in the determination of trace d‐ enantiomer content in natural amino acids) | HIGH | [23–27] |
Ion pair | Used to separate charged substances using reversed‐phase HPLC. Works through the addition of mobile phase additives. Used to separate small molecule drug, metabolites, and nucleic acids | HIGH | [28–30] |
The mobile phase flow and analyte separation when applied to chromatography can be explained by several main mathematical relationships and physical parameters. Essentially a chromatogram, the signal response product of a chromatographic separation, is a plot of analyte concentration versus eluent volume or, less fundamentally, time (Figure 7.2). The retention time (t r ) or volume for a particular analyte is an important parameter for characterising the separation. Retention time is dependent on the volumetric flowrate. However, the retention time also depends on the dimensions of the column used. Therefore, the retention time is usually reported as a normalised retention quantity known as the capacity factor (or retention factor) (k). Selectivity (or separation factor) (α) is the ratio of the retention (capacity) factors (k) of two peaks. The efficiency (N) is reported as the number of theoretical plates for a column. Traditionally, a ‘plate’ (the name is taken from the ‘plates’ used in fractional distillation towers used in the petroleum industry) is defined as a hypothetical separation layer within the column in which the two phases establish equilibrium with each other. The higher the number of plates the more efficient the separation will be. The concept of a layer in which equilibrium is established is a rather abstract one in that chromatography is a dynamic process and equilibrium is never reached. For simplicity it may be useful to note that analytes elute from the chromatography column as a peak with a Gaussian distribution and the plate height is related to the variance of that peak. The Gaussian distribution arises from the band broadening processes that take place in the column, giving rise to random error in the time taken for molecules of a single analyte to elute from the column. Resolution (R) is defined as the separation of two analyte peaks, taking into account the individual peak widths. The main mathematical relationships used in chromatographic separation are shown below:
where
The above equation for resolution is used for the actual measurement of resolution from a chromatogram. This should not be confused with an alternative equation for resolution, which may be obtained by substituting the other three equations into the equation for resolution.
The Purnell equation [31] describes the resolution of two similar peaks under chromatographic separation taking into account the selectivity, capacity and plate number:
The Purnell equation is an extremely useful relationship that can be used to gain a deeper insight into what qualitatively is obvious. Better resolution may be had by having narrower peaks (better efficiency) and by having a larger ratio of capacity factors (selectivity). As the Purnell equation also indicates, better resolution may be had by increasing the capacity factor. However, not only is this at the expense of increasing analysis time but there is little benefit in terms of increasing resolution for k > 10.
The separation of two analytes (Figure 7.2) is only possible if their retention times differ by virtue of having different distribution coefficients between the mobile phase and stationary phase. While the principal goal in this case is to obtain reliable baseline resolution in as short a time as possible, an associated goal is to obtain this not just through good peak efficiency but also through good peak symmetry; i.e. there may be deviations from the anticipated symmetrical Gaussian peak shape in the form of fronting and/or tailing. There are several main causes for non‐Gaussian peak shapes, which include: column stationary phase deterioration, column under‐ or overpacking, mobile phase impurities or loss of liquid handling control and/or efficiency [32,33]. Peak width is usually measured at both the baseline inflection points and at half of the maximum signal height (Figure 7.2). By monitoring peak width over repeated separations, both column and system performance may be determined. For example, if a column is still being conditioned with the sample there will be a reduction in peak width over time. Also of note is the fact that as the concentration of an analyte in the sample increases there is usually a consequent broadening of the peak (mass overloading). Other factors leading to an increased peak width are dilute sample injections over long periods of time (usual in preparative protein chromatography and volume overloading), very shallow or elongated gradient elution profiles, mobile phase impurities or incorrect make‐up and/or problems with liquid handling.
Gradient elution is in contrast to isocratic elution. When eluting analytes in chromatographic separations, there are three distinct types of elution (Figure 7.3) available. Isocratic elution utilises one mobile phase consistently across the whole separation. Isocratic elution is used in size exclusion chromatography but is rarely used when other modes of liquid chromatography that are used for the separation of large biologicals. This applies in particular to preparative liquid chromatography. Gradient elution utilises two (or very occasionally more) mobile phases and over a period of time gradually mixes the phases so that the composition changes from the initial composition to a new composition. For the majority of liquid separations, the gradient is linear. Step elution, like gradient elution, utilises two (or more) mobile phases, but rather than gradually changing the composition it uses a series of ‘step‐up’ transitions to achieve a new mobile phase composition. In the analysis of small molecules a gradient is commonly used to shorten run times by bringing strongly retained sample constituents off the column more quickly. In preparative work on large biologicals the application of a gradient is more commonly used to facilitate separation and aid resolution of closely related analytes. Step elution is used to generate fast but crude ‘cuts’ of adsorbed species from the column. It is common to use step elution after the mobile phase composition needed for a specific analyte elution has been characterised via gradient elution.
In considering the theory of chromatography in the context of biological applications it must be remembered that it is not just about manipulating retention, selectivity and efficiency to obtain resolution of large molecules. Biological samples, particularly those containing small molecules, can be very complex, say several hundreds of thousands of components compared to maybe tens of components in a small molecule drug sample. A real problem can be that, even with very narrow peaks, there are simply too many sample constituents for them all to be resolved in a given run time. In other words, there is insufficient peak capacity. Temperature, flow rate, use of smaller particles, etc., can be used to increase peak capacity. However, for a step improvement it is necessary to use multidimensional chromatography where the peak capacity is the product of the peak capacity of each dimension. For this reason, comprehensive 2D‐LC (very small volumes of eluent from the first column are all chromatographed on a second column with different selectivity in real time) is attracting great interest [34]. In the meantime, in the quantitative determination of complex biological samples it is almost the norm to add another dimension when using a single column through the use of a mass spectrometer as the detector. Mass spectrometry in its own right is a powerful tool in the analysis of biologicals (Chapter ) but while the focus here is more on the actual chromatography of biological molecules, including for preparative purposes, it cannot be emphasised enough that liquid chromatography in combination with mass spectrometric detection is very powerful for the characterisation and quantitative determination of biologicals.
Biological samples are, by their very nature, complex, containing impurities that are analyte‐related (minority) and analyte‐unrelated (majority). Analytical and preparative separation methods need to be robust in terms of their reproducibility and, for preparative work, throughput (multiple injections/loadings), while maintaining a selectivity for a physicochemical property for the analyte. Chromatography as a separation technique can provide this selectivity requirement even when applied to heterogeneous sample mixtures from a variety of sources.
Chromatography can also be used in‐line with other preparative and analytical techniques since it can provide the analyte in an eluent that is suitable for further downstream separation and analysis (for example spectrophotometry, electrophoresis, circular dichroism, nuclear magnetic resonance [NMR] and mass spectrometry).
However, the dynamic nature of biological samples can be a technical challenge when applying any form of separation technique. Ensuring that the sample applied to the chromatographic technique is indeed representative of the true biological system from which the sample was taken needs to be considered at all stages of sample preparation through to data analysis. When applied correctly, chromatography can be non‐destructive, modulated in terms of physical properties (pH, temperature, conductivity), have efficient sample processing times, and take advantage of computation integration and processing methods to circumvent many of these challenges.
While there are chromatographic applications for isolating [35] and characterising intact cells [36], in just about all cases for chromatographic separation of biological samples some initial form of cellular disruption is required [37]. Disruption methods will release various compartmentalised hydrolases (phosphatases, glycosidases and proteases) that can alter the composition of the lysate [38]. In all modern experimental approaches, it is essential to preserve sample composition from the point of generation to analysis. Indeed, managing sample complexity can be an added challenge when working with biological material [39,40]. Even the most sophisticated modern instrument cannot produce meaningful data without some form of rational sample preparation. Biological samples, irrespective of their origin, will contain a heterogeneous mixture of small and macromolecular species (Table 7.2), all of which will behave differently when applied to chromatographic separation.
Table 7.2 General consideration and techniques for chromatographic separation of biological samples.
Sample type | Description | Technical considerations | |
Protein and peptides | Large bio‐macromolecule consisting of one or more chains of amino acids joined via peptide bonds. Proteins are found in all forms of life and have structural and catalytic functions. Peptides are short chains of 2 (di‐peptide) to 20–30 amino acids in length. The structure of a protein directs its function and so maintenance of the 3D conformation of a protein is important when preparing a sample. Proteins can be found in both soluble and insoluble forms depending on their function. |
|
[41–43] |
Nucleic acids and oligonucleotides | Large bio‐macromolecule consisting of one or two strands of nucleotides joined by phosphodiester bonds. Nucleic acids carry the genetic information for all known forms of life. DNA: Is a stable duplex of two strands of deoxyribonucleotides (Adenine, Thymine, Guanine and Cytosine). The stable structure is formed by the stacking of the nucleotide bases and the exclusion of water to the phosphate backbone on the outside of the duplex. RNA: Can be found in both single‐stranded and duplex conformations. RNA consists of strands of ribonucleotides and the base Thymine is replaced with Uracil. Oligonucleotides are short double‐ or single‐stranded nucleic acids that can be composed of ribo‐ or deoxyribonucleotides or unnatural nucleotide monomers. |
| |
|
|||
Lipids | A diverse family of biological macromolecules with mostly hydrophobic properties. Lipids can be long chained ‘fatty’ acids or shorter hydrocarbon chains with polar phosphate, saccharide groups or even proteins attached. Lipids are a major energy storage molecule in the cell and the main constituent of cell membranes. Steroids are also part of the lipid family. |
|
[44] |
Carbohydrates | Carbohydrates are a diverse family of biomolecules containing carbon, hydrogen and oxygen. The monomeric unit of carbohydrates is the saccharide. Carbohydrates are commonly referred to as sugars. Within biological systems, carbohydrates are used for their energy storage and structural properties. |
|
[45–47] |
Clinical samples | Clinical (patient‐derived) samples will contain a mixture of all biological macromolecules and bio‐inorganic analytes in different proportions dependent on the source. It should be noted that the generation of clinical samples usually indicates some form of perturbed physiological process associated with the patient and the sample. Common clinical samples are:
|
|
[39,48,49] |
At the start of the sample preparation phase, there are several important considerations that need to be taken into account since they impact the sample preparation strategy:
Table 7.3 Detection methods used with chromatographic separations.
Detection classification | Description and technical considerations | References |
General detectors |
UV/visual light absorbance: Visible light detection
(400–750 nm) and the more commonly used ultraviolet detection (100–400 nm; e.g. proteins absorb strongly at 280 nm by virtue of aromatic amino acids such as tryptophan) are a routine detection method used by biological sample analysis. The detector measures the absorbance of light by chromophores in the sample. For most applications UV/Vis detection obeys the Beer Lambert law but may show a non‐linear relationship at high analyte concentrations. Many detectors will allow multiple wavelengths to be monitored. Diode array detectors allow scanning of wavelength ranges and can generate 3D plotting of the detection signal. Picogram per ml detection levels are achievable. Refractive index: Involves measuring the change in refractive index of the column eluate. As an analyte passes through the detector the refractive index will change compared to the mobile phase. The greater the refractive index the greater the difference between the sample and mobile phase. Thus, the sensitivity will be higher for the higher difference in refractive index between the sample and mobile phase. The refractive index is said to be the only universal detection method for chromatography. Used in the analysis of carbohydrates and small molecules. Samples do not need a chromophore. Evaporative light scattering: Causes vaporisation of the eluate combined with quantification of the analyte by light scattering. The eluate emerging from the column is combined with a flow of air or nitrogen to form an aerosol – the eluent is evaporated from the aerosol by passage through an evaporator and the emerging dry particles of analyte are irradiated with a light source and the scattered light detected by a photodiode. The intensity of the scattered light is determined by the quantity of analyte present and its particle size. Samples do not need a chromophore. |
[51,52] [53] [54,55] |
Selective detectors |
Fluorescence: Can be applied to samples that fluoresce – a major limitation of its use (although many samples can be derivatised to induce fluorescence). Fluorescent detection is extremely sensitive down to the femtogram per ml level. Conductivity: Measures electrical impedance of the column eluate, which will change as analytes elute from the column. Its application is not common for biological samples but can be developed for protein separation. However, the sensitivity for protein is relatively low at the microgram per ml level. Electrochemical: Selective detection of electroactive analytes. Amperometric and coulometric forms of detection can be used. Analytes in the eluate flow through the flow cell, wherein molecules at the electrode surface undergo either an oxidation or a reduction, resulting in a current flow between the two electrodes. Sensitivity can be at the nanogram per ml level. Flame photometric: Applied in gas chromatographic separation for looking at residual metals such as tin, boron, arsenic and chromium as well as for sulphur and phosphorous containing compounds. Samples are carried in a hydrogen/air flame. Detection is via chemiluminescene at specific wavelengths, which when passed into a photomultiplier gives an electrical signal that can then be measured. Sensitivity can be at the picogram per ml level for some applications. |
[42,56,57] [58] |
Structure determining detectors (hyphenated analysis) | Combining sample separation with identification and/or structural deduction is a very powerful analytical method combination that is applied in the analysis of many biological samples. Mass spectrometry (MS): The analyte is introduced into a mass spectrometer after elution wherein its mass is determined. The mobile phase is mostly removed before the sample is introduced to the spectrometer. Detection can be via by total ion current (TIC) or selected ion monitoring (SIM). Mass spectrometry is able to distinguish analytes in overlapping peaks via ion monitoring as long as the analytes have a unique molecular ion or fragment ion. Mass spectrometry is very sensitive down to the femtogram per ml range for most analytes. Fourier transform inferred (FTIR) spectroscopy: Infrared spectroscopy arises because different analytes produce different spectral fingerprints when exposed to infrared radiation. The Fourier transform converts the detector output to an interpretable spectrum that provides structural insights. Sensitivity is in the nanogram per ml range. Nuclear magnetic resonance (NMR) spectroscopy: The sample is transferred directly into the NMR in the eluate solvent. The sample can be transferred during the chromatographic separation and the NMR spectra are then acquired either in on‐flow mode (continuously, while the chromatography is running), alternatively the NMR spectra are acquired under static conditions, requiring ‘parking’ of the analytes before analysis. Successful application has been seen with pharmaceutical analysis. Selectivity is in the femtogram per ml range. |
[59,60] [61,62] [63] |
Only after addressing these considerations can an appropriate strategy for sample pre‐treatment be determined.
Through sample extraction, it is possible to change the sample composition to suit the chromatographic separation method and/or reduce the sample complexity quite significantly. Common sample extraction methods used for biological samples prior to chromatographic separation (Table 7.4) and subsequent sample cleanup techniques that can be applied to an extracted sample (Table 7.5) can, in many cases, be applied in series to yield a sample that is ‘clean enough’ to be further processed via chromatographic separation.
Table 7.4 Sample extraction methods commonly used prior to chromatographic separation of biological samples.
Extraction method | Description and technical considerations | References |
Solid–liquid extraction | The sample is placed in a sealed container and solvent is added that dissolves/extracts/leaches the analyte of interest. The solution is then separated from the solid by filtration or centrifugation. Particularly useful for extracting small molecules from biological samples. | [64–66] |
Liquid–liquid extraction | Also known as solvent extraction or partitioning. Analytes in the sample are separated or portioned between two immiscible liquids (usually polar water and a non‐polar organic solvent) based on their relative solubility in each of the liquids. Particularly useful for extracting small molecules, pharmaceutical compounds and secondary metabolites. | [67,68] |
Homogenisation |
Biological tissue: Sample is placed in a blender or a mechanical homogeniser containing a solvent or aqueous buffer. The sample is homogenised to a homogeneous finely ‘chopped’ state. The liquid or suspended solids can be removed (filtration/centrifugation) for further workup. Microbial cultures: The culture is first harvested via centrifugation to yield a cell pellet. The pellet is then resuspended in a cold aqueous buffer and before mechanical homogenisation at high pressure (> 600 bar/8700 psi). Homogenisation is aided through the use of detergents. It is usual to carry out 3–5 passes through the homogeniser using optical density and/or microscopy to monitor cell lysis. |
[69–71] |
Sonication | The use of sound waves to vigorously agitate particles in a solution. The sound waves can be applied via the use of a probe immersed into the sample or through sound waves being applied to a sample container immersed in an ultrasonic bath. Particularly useful for resuspending solid samples (e.g. protein aggregates) and lysing small bacterial culture volumes, although care must be taken to not denature protein and nucleic acid analytes of interest. Also, can be used to generate emulsions and nano‐emulsions with more homogeneous characteristics. | [72–75] |
Table 7.5 Sample ‘clean‐up’ methods used prior to chromatographic separation.
Clean‐up method | Description and technical considerations | References |
Filtration | Sample is passed through a paper or membrane filter (or SPE cartridge) to remove suspended particulates; 0.2 μm filtration is recommended for all samples prior to chromatographic separation to ensure sample homogeneity and to protect the stationary phase. Various filter modalities are available, ranging from charged membranes to depth filters designed to remove specific particulate sizes. Note that filter flushing may be needed to recover valuable samples from filter media in some cases. Ultrafiltration can be used to concentrate samples and also remove specific impurities of a set MW by using filter membranes with pores of a specific molecular weight cut‐off (MWCO). This technique is available in centrifugation‐tube format. |
[76] |
Centrifugation and sedimentation |
Centrifugation: Sample is separated by centrifugal force to remove particulates to yield a clarified supernatant and a particulate pellet. Used to prepare cell‐free samples and to clarify protein and nucleic acid preparations. Often used as an alternative to filtration for smaller biological samples. Sedimentation: The sample is allowed to settle under gravity undisturbed. Sedimentation is a slow process and the rate of sedimentation is dependent on Stokes's radius. |
[77] |
Solid‐phase extraction (SPE) | A form of chromatography itself – in SPE the sample is passed through a small disposable cartridge or column filled with a stationary phase that is able to adsorb the analyte of interest while allowing contaminating analytes to pass through with little or no adsorption. Used frequently in an ion exchange (IEX) ‘charged’ format to remove contaminating proteins from test solutions. | [ 1, 28,78] |
Lyophilisation | Also known as freeze drying. An aqueous sample is frozen followed by controlled water removal via sublimation under vacuum. Particularly useful for thermally sensitive analytes (e.g. proteins) and samples containing volatile compounds (which will be removed during the drying). Can be used to concentrate an analyte from a large volume for subsequent reconstitution in a chromatography‐appropriate buffer or solution. | [ 10,79,80] |
Evaporation | Liquid is removed from the sample by gentle heating at atmospheric pressure, usually with flowing air (or under N2 gas). A vacuum can be applied to remove liquids with a low volatility. Generally used on small samples under centrifugal force to retain sample analytes. | [81,82] |
Dialysis | Analytes move between two aqueous liquid phases across a semi‐permeable membrane. Analytes transfer from one liquid to the other, based on differential concentration. Adaptations to this technique utilise membranes that have specific MWCO pores for more controlled analyte processing. Generally used to remove proteins from samples and for buffer exchange. | [83,84] |
Precipitation | Analytes can be ‘brought‐out’ of solution by modulation of their physicochemical properties. A common method for preparing proteins is to change the pH of the protein solution. As the pH approaches the isoelectric point of the protein it will precipitate out of solution and can be sedimented and harvested via centrifugation. An alternative method for protein precipitation is through the addition of ammonium sulphate, which will remove water from around the protein and cause it to precipitate out of solution. Both of these methods can be used to remove impurities while maintaining the analyte of interest in solution. | [85,86] |
As alluded to earlier, applications of chromatography to ‘biology’ are many and varied. However, the analysis of proteins has always been of paramount importance. Biochemists have been studying proteins since the early nineteenth century and the efficient chromatographic separation techniques in use today were developed in the early 1940s and 1950s and porous organic polymers were developed as part of the Manhattan Project (to develop the first atomic bomb!). Tiselius developed the basic chromatographic principles of chromatographic frontal analysis (continuous feed of the sample solution; the first constituent that breaks through may be obtained pure) in the early 1940s and applied a refined set of these techniques to the analysis of ovalbumin, serum albumin and immunoglobulins in the 1950s. In the subsequent six decades many different forms of chromatographic separation for protein preparation (purification) and analysis (concentration and structural determination) have been developed.
The first consideration for protein separation by chromatography is to be aware of problems specific to proteins that are not encountered in liquid chromatography of small molecules. The most obvious of these is size. In recent research, very small particles with no pores have been used. That notwithstanding, large pore particles (usually greater than 300 Å) must be used. The size is also responsible for slow diffusion in the mobile phase and slow mass transfer in and out of the stationary phase. The former can be an advantage at very low flowrates but the latter is more prevalent at normal flowrates and contributes to poor efficiency. Proteins may be denatured and so in preparative work it is prudent to work at close to natural conditions using aqueous buffers at pH not too far away from 7.4. It is best to avoid polar organic solvents such as methanol and acetonitrile. The use of very non‐polar stationary phases is also inadvisable given that proteins can denature the surface of these or even may adsorb irreversibly. Lastly, large proteins are heterogeneous and, in a small way, this contributes to band broadening.
Proteins have four main physicochemical properties that can be exploited by four different modes of chromatographic separation (Tables 7.1 and 7.2):
In size exclusion chromatography (SEC) (Figure 7.4a and b) the molecular sieving properties of porous materials are used to separate proteins based upon their size and shape. Traditionally in SEC, a column of microparticulate cross‐linked copolymers (styrene or divinylbenzene) and with a narrow range of pore sizes is in equilibrium with the mobile phase. More recently, it is common to use silica‐based particles typically with an immobilised diol surface for protein SEC. Large analytes are excluded from the pores and pass through the spaces between the particles and are the first analytes to elute from the column. Smaller analytes are distributed between the mobile phase inside and outside the particles and will therefore take longer to pass through the column. The smallest analytes are the last to elute from the column. The mobile phase within a particle is available to an analyte that is able to penetrate the particle. The distribution of an analyte in a SEC column is determined by the total volume of mobile phase inside and outside the particles. For a given type of particle, the distribution coefficient (K d ) of a particular analyte between the inner and outer mobile phase is a function of its molecular size. If the analyte is large and completely excluded from the mobile phase within the particle, K d = 0. If the analyte is sufficiently small to gain complete access to the inner mobile phase, K d = 1. However, a variation in pore size will mean that the K d values vary between 0 and 1. It is this complete variation of K d between these two limits that makes it possible to separate analytes within a narrow molecular size range on a given particle type. It should be noted that since the protein separations take place before t 0, SEC columns are usually longer and wider than those for other modes of chromatography.
The sizing properties of SEC make it a very powerful technique for characterising proteins in their native conformation, but SEC can also be applied for desalting protein preparations and concentration of particular analytes. The application of SEC has also been successful for separating nucleic acids [ 6,87], viral particles [88], large molecular weight formulation products [89] and multicomponent cellular assemblies [90]. For all of the applications, good knowledge of common SEC stationary phase polymers and the associated technical considerations (Table 7.6) is important.
Table 7.6 Chromatographic stationary phase description and technical considerations for the preparation and analysis of proteins.
Mode | Stationary phase description and technical considerations |
Size exclusion chromatography (SEC) |
Low pressure stationary phase polymers: Dextran (Sephadex, Sephacryl), Agarose (Sepharose) and Polyacrylamide (Bio‐gels) High pressure stationary phase polymers: Polyvinyl chloride (Fractogel) and Dextran (Superdex) SEC resins are prone to compression and sometimes difficult to pack. For some applications, fine resolution is only obtained with excessively long columns (30–100 cm), which can be prohibitive on many systems. SEC columns should be cleaned and stored in 200 mM sodium acetate, 20% v/v ethanol. 0.5 M NaOH can be used for brief robust cleaning if needed. |
Ion exchange chromatography (IEX) |
Anion exchangers:
Strong: Trimethylaminomethyl: ‐O‐CH2N+(CH3)3 (cellulose matrix) Triethylaminoethyl: ‐O‐CH2CH2N+(CH2CH3)3 (dextran or cellulose matrix) Weak: Aminoethyl: ‐O‐CH2CH2N+H3 on an agarose matrix Diethylaminoethyl: ‐O‐CH2CH2N+(CH2CH3)2H (cellulose matrix) Cation exchangers: Strong: Sulpho: −SO3 −, O‐CH2SO3 −, O‐CH2CH2CH2SO3 − (cellulose, dextran, or polystyrene matrix) Weak: Carboxymethyl: O‐CH2COO− (cellulose or dextran matrix) It is important that IEX columns are adequately equilibrated prior to sample application so as to bind and prepare all available functional groups with a counter ion (for proteins this would be Na+Cl‐) and to apply a set mobile phase pH across the entire column. IEX can be used in gradient or step elution modes and elution can be via pH changes or counterion concentration changes. A final high ionic strength wash after the elution step will allow the column to be cleaned and regenerated for subsequent reuse. 0.5 M NaOH can be used for brief robust cleaning. IEX columns should be stored in 20% v/v ethanol between uses. |
Hydrophobic interaction chromatography (HIC) |
Common HIC and RPC stationary phase chemistries:
The performance of HIC separation is mostly dependent on the properties of the protein and the stationary phase chemistry. The stronger the ligand hydrophobicity and/or the more surface hydrophobicity on the protein the stronger the interaction. However, these interactions are also affected by the presence of contaminating analytes, detergents, pH and the ligand density/degree of substitution. The dynamic binding of the analyte and the subsequent resolution is determined by the flowrate, salt concentration and mobile phase properties. Complete elution of bound species can be a problem if a very hydrophobic ligand is matched with a high salt concentration in the injection sample. Selecting a less hydrophobic ligand or a less substituted matrix can help solve this problem. Temperature can have an effect on some complex sample separations also. Columns must not be left with bound analytes and can be fully eluted and cleaned by 1 M NaOH and are usually stored in 20% v/v ethanol between uses. |
Affinity chromatography |
Established affinity chromatography applications:
Elution of the bound protein (or other analyte) needs to be complete to ensure performance upon reuse. Elution is usually through the application of high ionic strength buffers (1.5 M KCl or 2 M NaCl) or changes to the pH of the mobile phase that change the surface charge of the protein and allow it to dissociate from the column. Cleaning of the resin can be carried out using detergents, chaotropes in low or high pH buffers containing 1.5 M NaCl. Strong acids and bases should not be applied to affinity resins so as to maintain the ligand integrity. Storage should be in 20% v/v ethanol. |
This form of chromatographic separation relies on charge–charge interactions between the analyte and the stationary phase. Ion exchange chromatography (IEX) (Figure 7.5a and b) is a very versatile form of chromatography that can be applied over a wide range of biomolecules, synthetic polymers and small molecules. Both cation and anion exchanger stationary phases have a high binding capacity and their retention and resolution can be modulated through the ionic strength and pH of the mobile phase. The exchangers are either acidic (cation exchanger, negatively charged) or basic (anion exchanger, positively charged). Analytes are exchanged on the stationary phase through ionic interactions (a positively charged analyte interacting with a negatively charged stationary phase functional group, and vice versa). IEX can be carried out using either a strong or weak exchanger, wherein the strength of the exchanger refers to the extent to which the ionisation state of the functional groups varies with pH. It is not the strength of the interaction between the stationary phase and the analyte. The selection of either a weak or strong exchanger (Table 7.6) depends on the nature of the separation. If the separation is using a wide pH range (gradient pH‐based elution) then a strong exchanger is preferable as it ensures that retention via ion exchange will take place at both extremes of pH. Weak anion exchangers should be used when the binding capacity of the stationary phase is known to change as the pH of the mobile phase varies. In most IEX method development exercises it is advisable to select and screen strong exchangers first before moving to a weak exchanger to, if required, introduce an additional element of selectivity.
For protein separation via IEX the isoelectric point (pI, wherein the protein has no overall net charge) of the protein of interest should be obtained as this will allow an appropriate exchanger and mobile phase to be selected. This is particularly important if a pH gradient is being used instead of a buffer gradient. At a pH lower than the pI, the protein will have a net positive charge (suitable for cation exchange), while at a pH above the pI the protein will have a net negative charge (suitable for anion exchange). However, the stability of the protein as pH varies and the physicochemical properties of any putative impurities in the sample should also be considered so as to ensure that chromatograms obtained are representative of the initial sample. IEX may also be successfully employed for the separation of nucleic acids [91], drugs and impurities [10].
Hydrophobic interaction chromatography (HIC) (Figure 7.6a and b) has been applied for protein separation for over 30 years [92,93]. Separation is based on the adsorption of hydrophobic areas on the surface of an analyte on to the stationary phase (or, in some models, partition into a less polar area of mobile phase close to the stationary phase). For proteins the majority of hydrophobic residues are buried in the centre of the protein protected from the aqueous environment of the surface; indeed, it is the formation of this hydrophobic core that drives protein folding in globular proteins. Hydrophilic amino acids on the protein surface are covered with an ordered layer of water. Hydrophobic ‘patches’ on a protein surface can be exposed by treating the protein with salt ions (usually ammonium sulphate), which preferentially take up the ordered water molecules and allow the hydrophobic areas to be exploited for capture and separation activities. Interactions are via weak van der Waals forces between the protein and the stationary phase. However, protein–protein interactions are possible via the exposed hydrophobic areas and may cause aggregation of the protein and eventual precipitation. HIC can be applied in both low pressure protein capture (selective initial isolation step) and high pressure analysis formats. For low pressure chromatography, HIC makes an ideal intermediate or polishing step (further treatment of the ‘captured’ protein) in protein purification. The binding of proteins to an HIC stationary phase is normally in high salt concentrations and elution proceeds via the gradual decrease in salt concentration. This gradual change in salt concentration will cause a rehydration of the protein surface and a loss of interaction with the stationary phase. Changing the mobile phase pH can also be used to elute HIC separations, but the resulting separation can be quite variable for complex samples as this changes the separation to a more mixed mode (HIC:IEX) format and requires specialised stationary phase chemistries, which must have good batch‐to‐batch reproducibility. When applied in an HPLC format, HIC can be used to separate very closely related analytes using very shallow gradients and small resin beads or particle sizes (≤5 μm). The stationary phases commonly used for HIC are naturally hydrophobic, containing groups such as butyl, octyl and phenyl attached to a matrix, which promotes more protein–matrix interaction rather than protein–protein interactions. This quality of HIC makes it very versatile and able to accommodate high loading capacities. Common stationary phases include Phenyl Sepharose and Phenyl‐5PW for low pressure HIC and Bio‐Gel TSK Phenyl and Spherogel TSK Phenyl for HIC HPLC.
HIC is very similar to reversed‐phase chromatography (RPC, also commonly abbreviated as RPLC or RP‐HPLC). However, in RPC the mobile phase will contain a polar‐organic constituent, such as methanol or acetonitrile, that may promote sample denaturation, especially for proteins. RPC tends not to be suitable for protein purification, except in very special circumstances [79], but is a very versatile and commonly used application for protein analysis. RPC has a lower binding capacity than HIC. While HIC is commonly used for protein separation it has also been applied to lipids [94], carbohydrates [95] and nucleic acids [96]. Common HIC and RPC stationary phase chemistries (Table 7.6) can be used in either HIC or RPC but generally more hydrophobic chemistries are used in RPC with polar organic solvents in the mobile phase and less hydrophobic chemistries are used in HIC with highly or completely aqueous buffer containing mobile phases.
RPC is very much the most common mode of chromatography for small molecules [97]. Therefore, as far as proteins are concerned, RPC comes into its own in the elucidation of protein structure where it is used for peptide analysis of protein tryptic digests (Section 7.4).
Affinity chromatography (Figure a and b) exploits the biological function or specific structure of an analyte for separation making it extremely specific and capable of purification to high purity from very complex sample mixtures [17]. It is this specificity that makes affinity chromatography very expensive to use. Affinity chromatography is probably best characterised by the purification of enzymes by immobilising substrates/ligands (Table 7.6) on to chromatography stationary phase matrices. The enzymes would bind to the immobilised substrate once introduced to the column [98,99]. The interaction between the stationary phase ligand and the analyte needs to be reversible so that the column can be reused and so the analytes can be eluted. Elution is usually carried out through the application of free‐ligand changes in pH or salt gradients. Affinity chromatography is now used as the primary capture step for purifying immunoglobulin molecules used for therapeutic applications, but is also used for separating nucleic acids, membrane‐bound proteins and even whole cells. The use of affinity methods requires a detailed knowledge of the biological function and to some extent structure of the analyte of interest. The stationary phases need to be relatively inert in terms of their binding characteristics but possess suitable chemistry to allow attachment of the ligand. In this way retention is exclusively by the affinity interaction and resolution is not lost under the influence of non‐specific interactions. Ligands are usually added along with a spacer of up to 10 units (usually hydrocarbon chains but can contain carbonyl or amine groups) in length to allow correct orientation of the ligand and stop steric hindrance during binding to closely immobilised ligands. The matrix is usually an agarose or dextran base. Unlike other forms of chromatographic separation, affinity chromatography can be carried out in both batch‐resin and packed‐column formats.
For the isolation of native proteins from biological tissue or clinical samples, the chromatography may be limited by the physicochemical properties of the protein of interest. However, with the advent of recombinant DNA technologies, proteins can now be engineered to be more easily purified via chromatographic methods. There are several options available when trying to engineer a protein for a more simplified purification. There are three main approaches for engineering proteins for more efficient purification.
Immobilised metal affinity chromatography ( IMAC ) associated with the surface properties of a target protein. IMAC (also known as metal chelate chromatography) utilises immobilised divalent cations (Ni2+, Co2+, Zn2+, Cu2+) that interact via coordinate bonds with the imidazole groups of histidine, indole groups of tryptophan and to a lesser extent thiol groups of cysteines. Proteins harbouring these residues on their surface will be adsorbed to the stationary phase and can be selectively eluted by altering the pH of the mobile phase, the addition of a chelating agent (usually ethylenediaminetetraacetic acid [EDTA]) or a solution of a free functional group (imidazole, indole). The molecular biology techniques used to generate His‐tagged proteins are generated by molecular biology techniques (Section 7.2.3). Most IMAC resins have a cellulose base and incorporate a spacer arm between the metal ion and the resin bead. Proteins can be engineered via recombinant DNA technologies [100,101] to have these tags at either their N‐ or C‐terminal. The incorporation of 6× histidine residues is the most common form of tag used in IMAC separations, as cysteine incorporation tends to lead to protein aggregation and tryptophan incorporation can cause significant changes to the protein structure. Due to its specificity, IMAC can be used on very ‘dirty’ samples (crude cell extracts) and still achieve target protein purities of >95% from a single separation.
Introduction of a FLAG‐Tag or S‐Tag on to the N‐terminal of a protein allows immunoaffinity chromatography to be used for protein purification. The FLAG tag (Asp‐Tyr‐Lys‐Asp‐Asp‐Asp‐Asp‐Lys) is an 8‐reside hydrophilic amino acid sequence that is recognised by an immobilised monoclonal antibody on the stationary phase. The binding on the FLAG tag to the stationary phase is Ca2+‐dependent. Elution is carried out via the introduction of EDTA to the column and subsequent chelation of the Ca2+. The FLAG tag can be removed through treatment of the eluate with enterokinase [102]. Similarly, the 15‐amino acid residue S‐tag (Lys‐Glu‐Thr‐Ala‐Ala‐Ala‐Lys‐Phe‐Glu‐Arg‐Gln‐His‐Met‐Asp‐Ser) derived from pancreatic ribonuclease A, can be used for protein purification, usually in batch preparations [103]. Removal of the S‐tag is via subtilisin treatment.
Both of these small proprietary tag‐based purification systems are very useful for small preparations of recombinant proteins, but as the size of the preparation increases the financial costs may become prohibitive and the digestion product isolation can become more complex.
Polyarginine C‐terminal tags have been used in conjunction with cation exchange chromatography for protein purification. The arginine tag is sequentially removed through the application of carboxypeptidase B [104].
Both systems fuse the GST or MBP protein to the N‐ or C‐terminal of the protein of interest and then use immobilised ligands (Glutathione for GST fusions; amylose for MBP fusions) on agarose‐based stationary phases. Both systems can allow on‐column cleavage or elution via pH changes or the addition of free ligand and subsequent cleavage. Thrombin is used for GST cleavage and Factor Xa for MBP cleavage. If the digestion is carried out after the initial affinity capture step, then a follow‐on IEX separation is usually employed to separate the fusion protein from the protein of interest.
The generation of a fusion protein not only allows for efficient purification of target proteins but also confers stability to otherwise unstable proteins [106] and can enhance solubility for aggregation‐prone protein preparations [107].
Another application of recombinant DNA technologies, has allowed recombinant expressed proteins to be sorted and/or secreted to the periplasm (bacterial cells), cell membrane or growth media (bacterial, plant, insect and mammalian cells) for a more simplified purification strategy. As proteins are generated, specific sorting signals are built in to the amino acid sequence; these signals will ensure that a protein is sorted to the correct site of activity. These signals can be manipulated for engineering protein sorting [113,114]. Isolation of proteins from the periplasm of bacterial cells is carried out via an osmotic shock of the cells [115], while extracellular signals will ensure that expressed proteins are transported across the cell membrane into the growth media. As in the application of the strong promoter systems described above, sorting signals are much used in the generation of commercial‐scale protein therapeutics.
For an example on‐column refolding of recombinant lysozyme (Figure 7.8a), showing the associated size exclusion separation of key elution samples from the on‐column refolding (Figure 7.8a), the diethylaminoethanol (DEAE) elution of the refolded lysozyme is resolved into conjoined two peaks. The peak 1 refolded sample shows a shift in retention time (16 minutes) compared to the solubilised load material (main peak max. at 7 minutes). The solubilisation sample resolves into two conjoined peaks (peak max. at 7 and 10 minutes, respectively). The DEAE elution peak 2 contains higher molecular weight material (lower retention time) with a similar retention time peak max. of 10 minutes to some of the solubilised material. The DEAE column cleaning step contains high molecular weight material similar to the solubilisation sample (peak max. at 7 minutes). The conclusions from this on‐column refolding experiment are that lysozyme can be captured post‐solubilisation and refolded using on‐column methods. The post peak material seen on the DEAE column contains higher molecular weight material that could be aggregates or misfolded protein. The cleaning step removes material that is likely to be non‐refolded protein.
Associated analytical technologies: Assaying for protein structure can be a relatively complex undertaking. However, there are some techniques (circular dichroism, protein NMR or hydrogen‐deuterium exchange mass‐spectroscopy) that will provide detailed information on the folding dynamics and crucially the final 3D conformation of the protein. However, there are simpler analytical technologies that can be employed to monitor the changes in protein conformation, for example analytical ultracentrifugation and size exclusion chromatography.
Even after isolating an intact protein, it may still be difficult to characterise the purified product. However, such a ‘top‐down’ study of a large molecule can be complemented by a ‘bottom‐up’ approach to protein analysis. This could involve, for example, papain digestion of a monoclonal antibody, but this results in a few still relatively large fragments. On the other hand, deglycosylation of monoclonal antibodies allows glycan analysis by hydrophilic interaction liquid chromatography (HILIC) liquid chromatography–mass spectrometry (LC‐MS) to study the sugars responsible for molecular recognition, HILIC being a relatively recently popularised mode of LC involving a polar stationary phase and a mobile phase containing a high proportion of a polar organic solvent and a low proportion of aqueous buffer that is suitable for polar analytes and is characterised by good MS sensitivity [117]. This type of analysis is very valuable as it is post‐translational modifications such as glycosylation that endow proteins with many of their important properties. This is particularly the case in the study of monoclonal antibodies and ‘biosimilars’ used as biopharmaceutical drugs.
However, the most common form of the ‘bottom‐up’ approach is to carry out a tryptic digest of the protein and subsequently characterise the product peptides by LC (Figure 7.9) or, more frequently, LC‐MS or even 2D‐LC‐MS. This chromatography approach is now widely used and the chromatography used has been perfected to the point that the actual digestion is a significant variable. To this end, manufacturers are now introducing kits to carry out solid phase digestion to reduce the variability caused by time, temperature and autocatalytic digestion of the enzyme.
An even greater challenge is to determine low levels of biopharmaceutical drugs in biological fluids. This can be and is carried out on the intact protein but even here the bottom‐up approach can be applied. A signature product peptide may be found for a particular biopharmaceutical and levels of that signature peptide may be monitored as a surrogate for the biopharmaceutical.
While the content of this chapter has been focused mainly on the separation of proteins, advances in chromatographic hardware and column chemistry have allowed liquid and gas chromatography to be applied to many biologically relevant fields. This is not new. For many years catecholamines in biological fluids have been studied, e.g. by HPLC with electrochemical detection. Prostanoid levels have been monitored in the study of analgesic drugs and corticosteroids in the study of depression. Gas chromatography, after chemical derivatisation of analytes, can be used in the field of metabolomics and supercritical fluid chromatography has been used effectively with MS detection in the field of lipidomics. However, more recent illustrative examples include:
In summary, the application of chromatography to biological research has enabled biologists to gain a detailed insight into their sample composition and confidence in their data that was once only available to synthetic and analytical chemists.
That is not to say that chromatography in biology is a mature field. The pace of biomedical research shows no signs of abating. In the field of biopharmaceuticals the need for innovation is a constant and, going beyond large biological molecules as drugs, research into drug–antibody complexes and ‘new modalities’ such as oligonucleotides as drugs will throw up new challenges to the chromatographer.