WO2010125360A1

WO2010125360A1 - Protein crystallisation

Info

Publication number: WO2010125360A1
Application number: PCT/GB2010/000880
Authority: WO
Inventors: Jerry Heng; Michael Mark Roberts; Daryl Williams
Original assignee: Imperial Innovations Limited
Priority date: 2009-04-30
Filing date: 2010-04-30
Publication date: 2010-11-04
Also published as: GB0907504D0; EP2424876A1

Abstract

A method and apparatus for producing crystalline protein particles. A protein solution is re-circulated from a reservoir container (1) using a pump (2), through one or more capillaries (3) and back to the reservoir (1) through a re-circulation loop (5) and an outlet (6). In the capillary zone the solution is treated such that the protein solubility decreases, causing the protein to precipitate.

Description

PROTEIN CRYSTALLISATION

The present invention is concerned with protein purification and in particular an apparatus and a method for the crystallisation of proteins.

Proteins are very important in a number of different areas of medicine and chemistry. In many cases it is desirable or necessary to have a pure protein but this can be a difficult and expensive process. Most methods of protein purification involve a series of separation techniques, including chromatography. This can be both time-consuming and costly. Chromatography does not always provide protein in the stable native state which is required for biological activity. Many proteins are used as biopharmaceuticals, such as insulin, interferons and vaccines. These are required to be administered in the biologically active state. The rate and frequency of dosage is also crucial. In many cases, these medicines have to be targeted to specific sites in the body in a more concentrated form to be effective. Purification and formulation are therefore very important aspects to these medicines.

A further problem with protein purification is shown in the area of structure determination. The biggest hindrance to solving the molecular structures of proteins is the ability to produce sufficiently good crystals for X-ray crystallography. This is the primary technique used to solve protein structures to the greatest precision. This is also a vital route to drug discovery based on the structural characterisation of the active sites of proteins and potentially new chemical compounds that would bind to them as chemotherapeutic agents. The production of crystals > 1 mm makes it possible to perform other spectroscopic and physical measurements on proteins due to their high concentration in the crystalline state, such as neutron diffraction. The production of proteins in a regular crystalline form would offer numerous benefits in tackling these issues and would provide protein in its purest and most stable form that could be isolated from a mixture by removal of supernatant as a one-stage process. A narrow size distribution of large crystals is the most desirable for purification where the removal of impurities in the supernatant is required. A carefully engineered crystal size distribution is also necessary for drug or vaccine foπnulations for a controlled dosage release over a specified period.

Crystals contain the protein in its most concentrated form, enabling a targeted therapeutic delivery to be made. The production of stable crystalline forms also provides a number of additional drug delivery routes including aerosols for pulmonary delivery, for example. The long shelf life of crystalline dosage forms would also offer major benefits over the relatively short shelf life of freeze dried or liquid based solutions which are used in parenteral delivery.

It is therefore an object of the present invention to provide a method of producing crystalline protein and apparatus for doing the same.

Protein crystallisation for structure determinations is traditionally done in closed containers from a quiescent saturated protein solution that is allowed to stand until crystals form. A substantial disadvantage of this approach is that it creates erratic crystal growth due to uneven diffusion effects in the supernatant caused by convection. Furthermore, the extent of crystal growth in a quiescent system is limited by the rate of diffusion of the molecules into the crystal within the immediate vicinity. This produces a concentration gradient around each crystal, leaving uncrystallised supersaturated protein outside this zone. These factors combine and result in crystals of reduced and inconsistent size. It is therefore a further object of the present invention to provide a method and apparatus for producing crystalline protein of a more uniform and larger size.

According to the present invention, there is provided a method of producing crystalline protein, in which a protein solution flows from a reservoir through one or more glass capillaries and is treated in the capillary zone such that the protein solubility decreases and protein crystallises out.

Through the application of a controlled flow of a larger volume of protein solution into a smaller crystallisation space (glass capillaries), it is possible to increase the crystal size to a more consistent value. This is beneficial for the reasons set out above.

In one embodiment, there may be many capillary tubes arranged in series or in parallel. The protein crystallises out throughout the length of the capillaries and may be removed by any suitable means.

The treatment in the capillaries is normally temperature variation. This may be cooling of the solution, which may be achieved by any suitable means, for example by using iced water.

The temperature variation could alternatively be heating, which may be achieved by using hot recirculated water.

Alternatively, the treatment in the capillaries may be chemical treatment by way of a buffer exchange. In this case, the capillaries would take the form of hollow fibre dialyzers made from transparent regenerated cellulose, supplied by Spectrum. These are manufactured with a 0.2 mm internal diameter and molecular weight cutoffs of 13 and 18kDa. The former would retain most proteins. This technology has been applied to remove salt from protein solution buffers by online flow dialysis to enable the mass spectral analysis of proteins (S. Canarelli, L Fisch and R. Freitag. "On-line microdialysis of proteins with high- salt buffers for direct coupling of electrospray ionization mass spectrometry and liquid chromatography", J. Chromatogr. A, 948 (2002) 139-149). In the present application, the protein solution would pass through the hollow fibre dialyzers running through the flow cell containing the precipitating solution recirculated in reverse flow mode to the direction of protein flow, or simply stirred with a magnetic stirrer with the inlet/outlet ports closed. This would produce a supersaturated protein solution that leads into a glass capillary outside the flow cell where crystal growth would occur as a continuous process.

The advantage of this process is that crystal nucleation centres of a consistent size should be produced, corresponding to the transportation time across the hollow fibre dialyzer. By necessity, this solution would flow through the capillary in a single passage without recirculation. Through a careful optimisation of the protein solution flow rate, it should be possible to complete the buffer exchange by the time the solution enters the glass capillary. There is a well-documented range of suitable precipitation buffers for the crystallisation of many proteins by inducing supersaturation. In this method the concentration of the precipitating agent should be chosen with the aim that this is a buffer exchange rather than buffer mixing. Consequently, this method requires half the concentration of precipitating agent to that used in a 1: 1 batch crystallisation premix of protein and precipitant. As the crystallisation experiment proceeds, the concentration of precipitating agent in the flow cell will gradually decrease, reducing the frequency of nucleation and promoting crystal growth. This would reduce the appearance of late, smaller crystals and promote larger crystals of a more consistent size from nucleation centres that appeared in the early stages of the experiment. For all crystallisation methods that employ this flow procedure, including dialysis, batch and temperature change, glass capillaries of internal diameters in the range 0.5-6mm, for example 0.8-5mm or 1-4 mm may be used. These may be used to produce crystals of a size in the range 0.05- 1.0mm, for example 0.07- 0.7mm or 0.1 -0.4mm. For example, glass capillaries in the range l-3mm or 1- 2mm may be used to produce crystals in the range 0.1 -0.4mm.

The flow through the system may be controlled by means of a peristaltic pump that draws protein solution into the capillary rather than pumping it in, to avoid protein denaturation. For smaller volumes of protein solution, an oscillating dual piston mechanism may be used to induce alternating solution flow through the capillary without passage through the peristaltic pump that might denature the protein.

The invention also extends to an apparatus for producing crystalline protein, comprising a reservoir for protein solution, connected to one or more glass capillaries in a treatment zone, and further connected to a flow generating means, in which the protein solution is pulled through the glass capillary.

The flow generating means may comprise a peristaltic pump or an oscillating dual piston mechanism as discussed above. The treatment zone may comprise a heat exchange zone between the glass capillaries and a heat exchange medium. The heat exchange zone may be a tube and shell heat exchanger such as a flow cell in which the cooling or heating fluid circulates around the outside of the glass capillaries which contain the protein solution. The treatment zone may alternatively be a buffer exchange zone as discussed above. This invention may be put into effect in a number of ways and an embodiment is described below with reference to the following Ωgures, in which: Figure 1 shows schematically an example of a flow crystallisation circuit according to the present invention; Figure 2 shows schematically a specially constructed flow cell cooled by ice-cold water for use in the circuit of figure 1 according to the present invention; and Figures 3 and 4 show crystal formation in capillaries for quiescent and flow systems.

The inventors have demonstrated the invention with the crystallisation of hen egg white lysozyme (HEWL) using the cooled liquid flow system described in the next section.

Referring to figure 1, a flow system is shown where a protein solution is drawn from the bottom of a reservoir 1 by a peristaltic pump 2 at different flow velocities into a series of glass capillaries 3. These capillaries are maintained at a reduced temperature by heat exchange means 4 as is better seen in figure 2.

Crystallisation in the glass capillaries is induced by reducing the temperature of the protein solution to a sufficiently high degree of supersaturation. The different flow rates are achieved simultaneously by using a selection of different Tygon® tube sizes at 0.38, 0.57, 0.89, 1.09, 1.3, 1.65, 2.06 and 2.79 mm internal diameter on IPC N-8 peristaltic pump rollers operating at a specified rotation speed to give flow rates of 8.6, 17.7, 39.5, 57.4, 79.9, 120.8, 178.8 and 299.4 μl/min, for example.

The protein solution is pulled rather than pushed into the capillaries to avoid pump pulsation and denaturation effects. The different flow rates are generated in these experiments to find the optimum flow velocity for crystallisation of a particular protein. In practice, once optimum conditions have been found for a specified crystal preparation, a single diameter capillary and the optimum flow rate would be used.

In our example, 30mg/ml of HEWL in 0.1 M sodium acetate at pH 4.8 was mixed with an equal volume of precipitation buffer (6.5 % sodium chloride in 0.1 M sodium acetate at pH 4.8) to increase the solution supersaturation prior to being drawn into the glass capillaries for cooling.

The protein solution outflow from the peristaltic pump can be recirculated through a recirculation loop 5 by dispensation into the original reservoir container

1 through an outlet 6 that allows solution flow down the side of the reservoir container 1. This avoids denaturation and ensures that the entire starting volume of protein solution is used before the recirculated solution which remains overlaid. In this way, protein can be repeatedly passed through the capillaries until it is completely crystallised or the supernatant in the capillaries has approached the solubility limit. Such a lysozyme flow experiment conducted over

40 hours in 16 cm x 2 mm i.d. capillaries produced a yield of 80 % of total protein, compared to 81 % for a sealed batch control. This suggests that denaturation by the rollers of the peristaltic pump is minimal. Recirculation therefore greatly increases the efficiency of protein recovery by crystallisation from solution. Recirculation would not be used where the protein concentration is to be kept constant, to maximise the mass of crystal growth, for example.

Connections between the protein solution reservoirs, capillaries and pump tubing are made by combinations of silicone and Teflon tubing. The capillary tube cooling is achieved by means of an ice/water bath or a specially constructed flow cell cooled by ice-cold water such as that shown in figure 2. The flow cell in figure 2 is constructed from Perspex and accommodates up to ten capillary tubes 10 in parallel through a series of 3 mm holes on both sides. Note that for glass tubes of external diameter over 3 mm, alternative temperature regulation methods need to be used, such as an ice/water bath. In figure 2 the protein solution flows from left to right in the direction of the arrows from the protein reservoir through the capillaries and to the pump. Water is recirculated through the cell externally from a water bath to control the temperature of the capillaries to a length of 12 cm, including sections of the housing. In figure 2, water enters the cell through water port 20, passes over the capillaries 10 in counter-current mode and exits the cell through water port 22 passing back to the water bath (not shown). The cell contains an additional port 24 in between the water ports 20, 22 to monitor the water temperature by insertion of a mercury thermometer. The capillary positions are placed 5 mm underneath an observation window to be visible within the focal length of the microscope set up to observe the crystal growth in the capillaries. The bottom capillary 15 in figure 2 is a sealed batch system with no flow to operate as a crystallisation control.

As soon as the protein solution flow enters the capillaries, it is rapidly cooled to the water temperature, in this case 0⁰C. Crystal growth in the tubes with protein flow is much greater and gives crystals of a more consistent size than in a sealed tube without flow because protein is supplied at a constant rate that overrides the diffusion effects. The flow velocity can be increased up to the rate-limiting step defined by the incorporation of protein molecules into the crystalline state at the crystal face. Any further increase in flow velocity may result in reduced crystal growth. Using the recirculation system, the optimum flow velocity for HEWL giving maximum crystal growth is closest to the protein solution flow setting of 63 mm/min. This is not a universal value and it will vary depending on the volume of solution in the reservoir, the capillary diameter and the time the experiment is running. For this optimum flow setting the capillary diameter was 1 mm and the solution volume in the reservoir was 4 ml for an experimental run of 29 hours. Note also that the crystal growth at some point may restrict the solution flow. For unrecirculated protein solution, crystal growth is greater and continues to increase with higher flow velocities.

In a comparative experiment, protein solution is held for 32 hours on ice without flow recirculation, in 5cm x lmm glass capillaries. HEWL crystals grow to 0.3-0.37 mm under flow conditions but only 0.05-0.1 mm under quiescent conditions, as measured across the { 1 10} faces according to the Miller Indices. The crystal mass in the capillary of the quiescent system was 0.438mg (see figure 3) and that in the flow recirculation system was 32.87mg (see figure 4). This represents a 75-fold increase of the concentration of crystal mass relative to a sealed batch crystallisation, attained within an equivalent capillary space using the flow technique of the present invention with a greater volume of protein solution. This is shown by the comparison between figures 3 and 4.

No crystal growth was observed in the flow lines outside the cooled zone where the ambient temperature was above the crystallisation limit under these conditions.

The present inventors have demonstrated how this flow cell can be used to induce protein crystallisation .by cooling. However, it can also be applied in cases where protein solubility decreases with temperature. Here, heating can be applied to induce crystallisation.

Other applications beyond temperature control can be used with this flow cell, such as the use of protein flow through tubes of dialysis membranes with different buffer conditions circulated through the cell to favour crystallisation conditions as described earlier. In any case, this flow cell allows up to 10 flow crystallisation experiments to be conducted simultaneously under controlled temperature or potential buffer exchange conditions.

The key step described here is a method for generating optimum protein crystallisation using a liquid flow system through a capillary tube rather than a traditional quiescent liquid environment, using either plates or capillaries. A number of additional innovations follow including closed loop liquid flow systems, with either re-circulatory or reciprocating liquid flows, as well as the use of capillary arrays for either scaling up the level of protein crystallisation or for allowing for rapid optimisation of experimental conditions for protein crystallisation. The approach described here can be used in conjunction with crystallisation by temperature or buffer solution change.

This solution flow technique for protein crystallisation has a number of advantages over traditional crystallisation methods using quiescent protein solutions in closed containers. It is able to crystallise protein from a much larger solution volume into a more confined space. This is far more practical than setting up multiple sealed batch crystallisation experiments for the same volume to achieve the same crystal yield. Furthermore, the crystals are of a much larger size and have a better defined crystal size distribution. Also, by manipulating the flow rate it is possible to control the crystallisation process. This makes crystal preparation both easier and more reliable. One of the most important advantages of the flow system is the time-saving factor. It takes a week to grow HEWL crystals to 0.24mm under quiescent conditions compared to 0.37mm in a day by flow methods in our hands using the same acetate-sodium chloride buffer system described above. This would be a great cost saving for the growth of other protein crystals in the biotechnology and pharmaceutical industries. In one prior art system, a solution flow method was tested on HEWL growth. HEWL solution was pumped in a similar acetate buffer to that set out above through tubes of porous polypropylene membranes (E. Curcio, S. Simone, G. Di Profio, E. Drioli, A. Cassefta and D. Lambda. Membrane crystallisation of lysozyme under forced flow, J. Membr. ScL, 257 (2005) 134-143). In this case, vapour diffusion occurs through the membranes into a stripping solution flowing as a counter-current outside the tubes. This increases the supersaturation of the HEWL solution, enabling crystals to form in the tubes as the solution flows through.

However, the crystals growing on the tubes were of a substantially smaller size (< 0.1 mm), even though the initial protein concentration used in that study (40 mg/ml) was higher than that used in the present example. The membrane flow system is complicated by the fact that there are two factors influencing crystallisation: the vapour diffusion flux and fluid flow. The flow causes an anisotropic growth of the tetragonal crystal faces, producing crystals of a rod- shaped habit quite different from the block-shaped crystals produced in our hands at similar flow velocities. This suggests that the liquid flow mechanics in each case differ or the hydrophobic nature of the membrane has an influence.

In glass capillaries, there is a very good heat transfer surface for temperature control and HEWL readily undergoes heterogeneous nucleation onto glass surfaces. The growing crystals remain adhered to the glass as they are fed more protein from the solution flow. This in turn creates more turbulent flow within the capillary that captures additional protein to crystallise from solution that would otherwise pass out of the capillary in a laminar flow regime. These factors contribute to the increased crystal growth in glass capillaries under flow conditions. The flow system of the present invention has a great potential for development in terms of its applications to proteins other than HEWL and to mixtures of proteins to determine the effectiveness of crystallisation as a means of purification. To aid this process, the internal surfaces of the capillaries could, for example, be derivatised with chemical groups or with optimised surface topography to specifically attract a protein from such a mixture to crystallise.

The flow mechanism could be modified to an oscillatory motion back and forth along the capillary instead of a one-way flow using a dual piston motion. In this way, identical volumes of protein solution in capillaries could be compared in static and dynamic states. This would also enable less protein solution to be used and avoid the need for temperature change to induce localised supersaturation, since the liquid would be localised in the capillary throughout the experiment.

The flow cell accommodates capillaries with internal diameters up to 2 mm and applies temperature control over a length of 12 cm. However, we have successfully demonstrated HEWL flow crystallisation in larger glass tubes up to 30 cm long and 4 mm internal diameter in an ice bath.

The crystals formed in the capillaries may be removed by any suitable means. Possible mechanisms for this include the use of a plunger, sonication, a steel ring drawn through the tube with a magnet, or even a laser beam.

Claims

1. A method of producing crystalline protein, in which a protein solution flows from a reservoir through one or more glass capillaries and is treated in the capillary zone such that the protein solubility decreases and protein crystallises out.

2. A method as claimed in claim 1, in which there may be many capillary tubes arranged in series or in parallel.

3. A method as claimed in claim 1 or claim 2, in which the treatment in the capillaries is a temperature variation.

4. A method as claimed in claim 3, in which the temperature variation is cooling of the protein solution.

5. A method as claimed in claim 4, in which the cooling means is iced water.

6. A method as claimed in claim 3, in which the temperature variation is heating of the protein solution.

7. A method as claimed in claim 1 or claim 2, in which the treatment in the capillaries is chemical treatment by way of a buffer exchange.

8. A method as claimed in any previous claim, in which the glass capillaries have a diameter in the range 0.5-6mm, or 0.8-5mm, or 1-4 mm, or 1-3 mm or 1-2 mm.

9. A method as claimed in any preceding claim, in which the flow through the system is generated and controlled by means of a pump

10. A method as claimed in claim 9, in which the pump is a peristaltic pump.

1 1. Apparatus for producing crystalline protein, comprising a reservoir for protein solution, connected to one or more glass capillaries in a treatment zone, and further connected to a flow generating means, in which the protein solution is pulled through the glass capillary.

12. Apparatus as claimed in claim 11, in which the flow generating means comprises a pump.

13. Apparatus as claimed in claim 12, in which the pump is a peristaltic pump.

14. Apparatus as claimed in claim 12, in which the pump is a dual piston pump.

15. Apparatus as claimed in any one of claims 11 to 14, in which the treatment zone comprises a heat exchange zone between the glass capillaries and a heat exchange medium.

16. Apparatus as claimed in claim 15, in which the heat exchange zone is a tube and shell heat exchanger in which the cooling or heating fluid circulates around the outside of the glass capillaries which contain the protein solution.