Work experience week at CPI

Ribbon structure of an Adeno-associated Virus (AAV) 

After a relaxing start to my summer holiday I arrived at CPI's Biologics Centre in Darlington on Monday 1st August 2016 for my work experience week on the train from Durham which was only a 15 minute train journey. I was given a card to swipe for all of the locked doors, and set up on the site's computer system with a CPI Email. Jonathan Gunnel, who I had emailed previously to set up the week, was my supervisor who arranged the things I would be doing. He graduated from Leeds University with a BSc in Biochemistry, so had a good idea of the sorts of practicals which would come in useful to have experience of for my university course. He showed me a quick tour so I knew where everything was. A short while after I was given the full safety briefing and hazard talk, and by 11am I had been invited to come to the fortnightly meeting where I was shown the details of the stages of projects and I than to see how the financial side of CPI worked. I was then set up with a desk and computer with an account to email and I got a full sense of just how integral teamwork is to CPI; the office space I was in was on a floor with around 50 other people and the seating was arranged so that all of the downstream scientists and upstream scientists were in the same area, but just across from them were the marketing team, and behind them were the financial team, admin, programme writers, everyone all able to communicate easily and face to face if need be. This sort of team work is needed for a place like CPI, to complete projects there needs to be ease of communication for the smoothness of progress. 

Centre for Process Innovation's National Biologics Manufacturing Centre 
Darlington 

On the scientist side of CPI there were three main departments; Upstream, Downstream and Analytical. Simply put, upstream cultures cells to produce specific molecules or proteins, downstream get given the "soup" of everything those cells have produced and are tasked with extraction and purification of a single type of protein and Analytical then test rigorously to make sure that the purity and correct protein or molecule has been collected. In my week I spent time with al three departments, but the main area of the work I was doing was downstream. 

Centre for Process Innovation's National Biologics 
Manufacturing Centre Darlington 
An industrial scale ion exchange chromatography column

For the first task, I was asked to research the basics of protein separation: Ion Exchange Chromatography, size exclusion, hydrophobic interaction and reverse phase chromatography. After a bit of searching around on the Internet and looking in some biochemistry books, I (hopefully) understood it: for ion exchange chromatography, a column is packed with beads made of, for example, cellulose with a negative carboxylate side chain and therefore negative charge overall. All proteins are made of a series of amino acids which each have a specific charge, hydrophobic affiliation and pH when in solution. These factors mean that every protein is slightly different overall, and by modifying the conditions of the beads and solution you can find a point at which the proteins which are highly positively charged all get attracted to the beads in adsobtion bands. The more positive the protein, the higher the band will be, as gravity or pressure pushes the solution down through the column. If you're trying to separate an overall negatively charged protein from the positively charged proteins, this method should allow the positively charged proteins to be repelled by the beads and so pure negatively charged proteins are collected at the bottom of the Ion Exchange Chromatography column. Each protein has an 'isoelectric point' (due to the addition of many amino acids with different charges) at which there is no overall charge on the protein. By controlling a buffer solution you have control of the rate at which the ions exchange. The beads can then be washed in a stage called Elution where a saline solution can be pushed through the beads, expelling the positively charged proteins to the bottom of the column. The Beads can then be used again, which is a good thing, because I was shown industrial scale columns for which which the beads are specific to the process and can be extremely expensive.

Centre for Process Innovation's National Biologics
 Manufacturing Centre Darlington
Gold Nanoparticle solution.. different to what I expected. 

Centre for Process Innovation's National
Biologics Manufacturing Centre Darlington 
Centrifuge make sure to balance it out!

Part of sitting in an open workspace means that everyone else is working near you on different projects, and I became friendly with a couple of other downstream Scientists who were working on a project which involved insulin attachment to Gold nanoparticles! This was the sort of thing that they give you in a Biology textbook for a 'did you know' box, but I was allowed to watch the process to understand what they were doing. I was surprised the Gold nanoparticle (GNP) solution was a dark dark brown-almost black- not dissimilar to Pepsi or Coke in colour. You just never think of metals as great absorbers of light, especially gold but at the nano scale it's not the same world as we are used to on this comparatively giant scale. Their task was to find the most efficient concentrations and ratios of GNPs to water to insulin for the drug's delivery in the human body, perhaps so that less insulin needed to be used because it could be more localised using this method or had more bioavailability in the human body. This was the first time that I saw the JANUS liquid Handler or 'the robot' as they all called it. Later in the week I programmed and used it, which was good to see in comparison to manual pipetteing.

Centre for Process Innovation's National
Biologics Manufacturing Centre Darlington
JANUS Liquid handler 

Their method was to use a fixed concentration of insulin and to dilute it with the water and GNP solution in different ratios in a plate with 96 wells so you can carry out 96 different versions of need be (or 32 and repeat each reaction 3 times) then Centifuged at 3220x Gravity for 5 minutes, so a pellet of
(hopefully) GNPs bound to insulin were collected at the bottom and then the robotic liquid handler removed the superb start without disturbing the Pellet at the bottom. They then conducted a BCA protein Assay (bicinchoninic acid assay) was used on the supernatant to assess how much of the insulin was left to then work out how much was bound to the GNPs on the pellet at the bottom of the well. The BCA assay turned green- apparently due to Cu²⁺ and ³⁺ ions- if not much protein was present and a deep purple if there was plenty of protein- you can quantitatively use this data if you enter the plate into a colourimeter and use 562nm green light. They explained to me the importance of balancing out the centrifuge, because if you don't it will apparently begin hopping around the room at high speeds! As shown by the image, some results with hardly any protein in the solution were obtained, so were successful. 

Centre for Process Innovation's National
 Biologics Manufacturing Centre Darlington 
Bicinchoninic Acid assay (BCA) 1 min after assay

Centre for Process Innovation's National
Biologics Manufacturing Centre Darlington 
End result of the BCA assay

Reverse osmosis seen through a scanning electron microscope

In the next couple of days I was asked to do some research around the different types of filtration that downstream scientists use to filter out impurities or to leave only larger molecules behind, either way separating them from other unwanted products of the cells. micro filtration, ultrafiltration, nano filtration and reverse osmosis are all different types of scientific filtrations which can be performed on solutions, with varying results. Micro filtration (0.1 micrometre pore size) is the usual type of filter paper which can be bought snd will filter out bacterial cells, whole cells and any suspended solids, such as the gold nanoparticles mentioned above, because they are around 300 nanometres across ( = 0.3 micrometers) however it does not filter out any ions in the water or virus cells. Nano filtration uses pore sizes of around 0.001 micrometers and so can take large ions out of the water such as calcium ions, meaning water can be de-hardened. The most extreme form of filtration is called reverse osmosis because it will only allow water to be let through, and salt ions will be flushed out, desalinating water. Sounds like a great solution to clean drinking water but unfortunately it requires a lot of energy to move the water around the filtration tube and isn't easy to do.

 I was introduced to another method of Protein assay, 280 nanometer Absorbance. To demonstrate this, one of the downstream scientists and I made up different concentrations of BSA  (Bovine serum Albumin- aka Fraction V) from crystalline powdered, pure protein with water.  This test is comprised of a small machine that looks like a miniature George Foreman grill, with tiny metal studs with pores of lights on the base of it, onto which a fraction of a microlitre of the Bovine Albumin serum is pipetted. In every protein there will be some amino acids with an aromatic side chain (such as proline, tyrosine or tryptophan) and these aromatic side chains happen to be very good at absorbing Ultraviolet light at a particular peak of 280nm, so the UV light is shone from the piece of equipment and can be observed in different intensities as it shines through the droplet and is detected on the other side when the lid is closed. Some computer programmes can map this out and tell you the exact concentration of proteins left the drop. The only drawback this method has is that the entire solution needs to be homogenised before pipetting onto the 280nm machine because otherwise you may get a particularly concentrated or weak part of the solution. There is also a coefficient we had to divide our results by which was specific to BSA, but we found that our results were spot on the the concentrations we theoretically had made with a percentage difference of around 5%. The limitations of 280nm absorbance are that it is very sensitive to non-protein matter, for example cell organelles will absorb almost all of the UV light, or even just a mixture of proteins since each has their own specific coefficients (which I assume depends on the quantity of aromatic rings present in the protein structure?).

Protein model of one of the serotypes of AAV

Every week or so at CPI they have a presentation  made by one of the members of staff on a topic or project that they have made progress with recently. I attended the week's lecture by Dr. Cristina Matos on the development of an industrial manufacturing platform for Adeno Associated Virus production (AAV). AAV is a very small virus- only about 20 nanometers in diameter- and lacks a protective lipid coating that most viruses have. There are 12 serotypes (types of virus with distinguishably different surface proteins) and although AAV is not harmful usually, particular serotypes are more effective than others at infecting certain tissues in the body. For example AAV8 is particularly suited to gaining access to liver, heart is suited and AAV1 6 and AAV7 is for skeletal muscle cells. I did not understand much of what Dr. Matos was saying because it was so in depth and she was presenting to a room full of scientists, but I enjoyed it nonetheless.

Next the Analytical department took me on a quick tour of their domain. I never knew there were so many different types of mass spectroscopy- gas chromatography (GCMS), Liquid chromatography (LCMS), and high performance- HPLCMS, even CEMS- capillary electrophoresis. I had not heard of electrophoresis before I saw this machine, so I have done a bit of research around it as I'm sure it will be something which I will be doing at some point in the future. Electrophoresis is defined as a way to separate macromolecules by putting an electric field through a gel. It can be used on proteins, RNA

and DNA, and usually involves a small gel filled chamber with wires attached at either end to create a difference in charge between each side of the chamber, pulling the molecules towards one end at different rates. the faster the rate the further it gets towards the positive terminal and so you can determine the approximate mass of the molecule band seen. Another curiosity I found when researching this type of analytical technique was the unit of mass used, the Dalton (Da). 1 Da is equal to 1 Au or 1/12th the mass of C-12, so the mass of Hydrogen is 1.0007 Da and carbon is 12 Da. Most proteins and DNA are measured in kDa (kiloDaltons). Titin, one of the biggest proteins ever found is 3.86 million Daltons, and around 30,000 amino acids long. so this means the rough mass of an average amino acid in this protein is about 128 Da- a useful conversion if you know the mass of any protein: Mass of protein in Da divided by 128 = approximate number of amino acids. (assuming that all amino acids are in similar ratios to how they are found in titin)

Centre for Process Innovation's National
Biologics Manufacturing Centre Darlington 
Me, all kitted up and pipetting into Eppendorf tubes

One of my favourite individual sections of practical lab work I was introduced to was the immunoassays. In Biology at school we learn how antibodies are Y shaped proteins which can sort of grab onto specific antigens of cells and because they are so specific they can be used by the human body to identify foreign objects or flag up harmful proteins or cells for destruction in the immune system. The Enzyme Linked immunosorbent assay or ELISA as it is commonly known is something which blew my mind when looking into the workings of it. You attach a selection of antibodies to an enzyme which takes colourless substrates and creates a coloured product, these antibodies are then introduced to a solution which you are trying to find a protein in, and they bind to this protein. You then wash away any other antibodies which are not attached to their corresponding antigens and add the two substrates to the solution, which should be catalysed by the enzyme into the coloured product. I was also introduced to the concepts of MIA and RIA- Magnetic immunoassay and Radio immunoassay. The basic idea with these is that you attach an antigen to a molecule of interest (if it does not already have one) and then you label a corresponding antibody with radioactive isotopes of carbon for RIA, or for MIA you attach a magnetic group onto the antibody which can then be detected extremely precisely.

Centre for Process Innovation's National Biologics 
Manufacturing Centre Darlington 
The JANUS pipetting machine uses syringe-like tubes to
create negative pressure and draw solutions up into the heads.

For the last day of my work experience I spent the morning logging new pieces of equipment into the warehouse's computer, and created a safety form for each of them. every piece of equipment from the tiniest 2ml eppendorf tube to a 100L bioreactor must be tested by a member of staff after one of the forms is made and only then can it be used for any new processes that need to be completed. There had been complaints from another member of staff that one of the liquid handling machines had not been accurate enough in pipetting when using a certain type of tip-(different tips can be programmed to be attached to the machine in order to minimise contamination or separate out two layers of a solution). I worked with another downstream scientist to create a protocol for the liquid handling machine to follow into order to test the exact accuracy of the machine using 2 different tips against manual pipetting. The first step was to prepare a batch each of BSA through manual pipetting. The only way to reliably create small solution concentrations by hand is through serial pipetting so I worked out the concentrations and ratios that would produce them, down to a 0.02 ug/ml and then produced them from the same crystalline BSA used before. After we had both made up 90 eppendorfs of various concentrations via serial pipetting, we set the liquid handler to work on its first task; creating the same numbers of Eppendorfs of the same solutions on a 96 well plate with just the normal head (i.e. no tip). The time difference from us creating ours to the robot doing it's was staggering. We took around 45 minutes, but the robot was finished in 5 due to having simultaneous capabilities of pipetting whole rows- (each head had to be washed in a water bath after each cycle to prevent contamination). The next set of 90 Eppendorfs was completed by the robot again, but this time was done using the conductive disposable tips, meaning that they could be sent an electrical signal and could stop pipetting as the machine reached a solid layer at the bottom. more importantly for our experiment was the disposable part of their name, so there would be no possibility of contamination because after every single solution was made, they were dropped off into a chute connected to a bin. This again ran for a very short time and the 90 Eppendorf tubes were collected.

Centre for Process
Innovation's National Biologics
 Manufacturing Centre Darlington
 JANUS robotic liquid handler

We used the 280nm absorbance machine mentioned earlier to quantify the concentrations of BSA we had made. I was then tasked with taking the data and preparing a spreadsheet with accompanying graph to work out the source, if any, of inaccuracy that the data reported and present my findings to the rest of the downstream department. I logged all of the 90 Eppendorf measurements in from the 280nm absorbance machine and averaged all of the readings. It was very pleasing to see that my own pipetting skills were a fraction higher or lower than the other Downstream scientist's who I had been working with, so we knew the were accurate baselines to use and compare the other two lines against. As I added more and more data to excel, the emerging graph showed that the single heads were at a slightly higher concentrations than our solutions, but only by a negligible amount at the lowest of concentrations- probably through small amounts not being entirely washed off. The real source of inaccuracy was found when I began to log in the disposable tips- each one at a far lower concentration than previously expected, and the graph showed a lack of linearity to the results obtained too; where the previous results were all almost a straight line, this disposable tip line had fluctuation all through the results. During my small presentation of my findings to the other scientists I showed this, and then suggested the cause of the error to be systematic, as I had noticed small bubbles of protein solution sometimes left on the tips after they were supposed to have expelled all of the solution. This would correspond to a random error in the concentrations, but would always make it lower than the original three lines, as was shown on the graph I produced. We had located the source of inaccuracy!

Centre for Process Innovation's National
Biologics Manufacturing Centre Darlington
Programming the liquid handler to carry
 out a set of instructions for dilution 

The suggestion after I showed my findings was that an air blow step was to be added into the instruction commands for the robot when using the disposable tips, in which a small amount of air would be blast through the end of the tip to pop any bubble left and add to the concentration.








Through my entire week at CPI I learned so many new skills and ideas beyond those taught at school which will no doubt be incredibly useful as I continue my scientific studies at university. I can only thank them for giving me the opportunity to do this. It was a great week and CPI is definitely somewhere I could see myself being employed in the future...