Using Precision Fermentation to Make Rennets and Challenges to Producing Complex Components in Milk-Like Casein Micelles

Precision engineering is one of the new terms used nowadays when attempting to harness the power of fermentation for food and biopharmaceutical production. Back in the 1980s, the dairy industry had already recognised and pioneered this type of technology. This was one of the first examples using precision fermentation to produce recombinant chymosin  or rennet for cheesemaking. We will also discuss the production of casein micelles for cheese manufacture.

Subsequently, the use of precision fermentation to produce simple whey proteins was followed up but the challenges remain with recreating complex biological systems like casein micelles. To produce a complex protein system such as a casein micelle is challenging to the say the least. As we discuss towards the end of the article, the key is recreating numerous post-translational modifications that are required in order for these micelles to be bio-assembled. Producing a rennet is straightforward but generating a casein is trickier because of these modifications. What appears critical in forming a casein micelle is that they are able to bind any milk salts as well as stabilize the correspondingly large supramolecular aggregates we know as casein micelles. The modifications needed to an initial and in context, simple protein sequence are not commercially available because industrial bacteria and yeasts don’t have the necessary enzymes to do the job. We will see that  physiological conditions needed to control the aggregation of caseins into micelles are a careful balance between electrostatic and hydrophobic interactions otherwise we might experience unwanted protein precipitation. Finally, we discuss the current viability of producing high-value dairy ingredients as a more secure and commercial industry in producing high-value dairy ingredients.

Introduction to Production of Rennet using Precision Fermentation 

The precision fermentation approach means extracting a gene from a calf, making a copy of it and inserting that into a plasmid. The gene of interest was that of chymosin which is what rennet is. Rennet, or chymosin as it is more accurately known, is a protease that causes the coagulation of caseins which then makes cheese. The plasmid was first put into an industrial strain of bacteria and then yeast cells to produce chymosin. The yeast cells are themselves modified. Fermentation of the microorganism led to production of the rennet protein which was then isolated and purified.

Recombinant rennet was the first food ingredient approved by the Food and Drug Administration (FDA) in 1991 that used this type of genetic recombinant technology and it was a certain cause celebre for the societal and political ramifications that followed. It was so novel in fact the FDA turned it into a case study and published it. It was a Nature article discussing the rigor of the new technology and the process it had to go through to be approved for use in food.

To summarise the technology, rennet which is in fact  chymosin, is a protease found in the guts of cattle. The gene for chymosin is taken from a calf and inserted into an industrial strain of bacteria but was later also inserted into yeasts.

In a more recent situation, camel rennet has been made by this precision fermentation approach. We now find because of the success of this method that simple functional proteins or enzymes with sufficient value are made commercially made by this approach and have done so for decades.

Harboe & Budtz in their research paper describe the process for the fermentation of the fungus Aspergillus to produce chymosin. It is a typical process for producing fermentation derived rennets!

1. An original medium is prepared
2. It typically uses submerged fermentation
3. After fermentation is complete, an acid treatment destroys the fungus including residual DNA & RNA
4. The mycelium and fermentation liquid are separated where the liquid contains the enzyme.
5. Purification by chromatography is required because so many other enzymes and other impurities are present in the bulk liquid
6. The enzyme is formulated and stabilized at the correct pH using salt and another stabiliser.
7. Sterile membrane  filtration removes unwanted bacteria
8. There is final quality control on the amount and concentration of the enzyme in solution.

Recently, precision fermentation was used to make proteins and in particular milks without the use of cows as in animal-free types of milk. The marketing strap-line is ‘claiming to replace cow’s milk’. An example is ‘Perfect Day’ which is an animal-free milk. The company focused on a single and simple whey protein (beta-lactoglobulin) from a protein chemistry perspective which is not present in human milk.

A number of start-ups are now exploiting precision fermentation. In this process they use dairy flora and sugar from plants as a fermentation media to produce a real dairy but animal-free protein that can be used by ice cream and cheese makers (Check out web-site: Perfect Day, Inc.). In more detail, it involves producing larger batches from a small fermenter to a large one and then transfer to the largest vessels for the purposes of industrial fermentation. The protein is harvested followed by separation and concentration before drying. These processes are proprietary and covered by trade secret in many instances. In these recent methods, steps such as chromatography may have been replaced because they are costly and complex  but they do enhance safety and purity. Bear in mind, rennets are a trace ingredient in cheese. If you wanted to replace the animal caseins in cheese where caseins would be around 25% of the cheese, that would be another prospect and thus a major ingredient is using this technology for that.

Change Foods for example is on that mission to replace casein. They are stating ‘cow cheese without the cow’. The challenges they face are extremely high.

Can precision fermentation make complex protein structures like the casein micelles in milk?

We know we have the means to make simple proteins like whey proteins and rennets. We need to use the milk secretion process to understand the challenges.

The Dairy Processing Handbook (1995) shows an udder which has a cistern connected to a teat cistern and then to a teat channel from which the milk comes out of. Milk is secreted from the udder of a cow. The key cells are the secretory cells or alveolus that secrete milk into the udder’s cistern. The raw materials for milk comes from blood which include water, amino acids (Swaisgood, Food Chemistry 1996), key nutrients and components, to build up the complicated structures present in the milk.

The secretory cell has two parts. Caseins are synthesized in the rough ER from genes in the cow. The caseins undergo post-translational modifications in the golgi apparatus and then assembled into micelles or aggregates before pushed out into the lumen. These are the cells that produce all the raw materials.

Milk Secretion and Formation of Casein Micelles

This occurs within the mammary gland (secretory cells). It requires:-

• protein synthesis
• post-translational modifications – a critical step such as phosphorylation (needed for calcium binding) by adding phosphates to serine groups and glycosylation needed for micelle stability.

Milk is important for transport of minerals such as calcium and phosphate for nutrition. The purpose of the micelles  is to allow safe transport of those minerals.

The association of caseins is via hydrophobic and electrostatic interactions including calcium phosphate nanoclusters which are crystals of calcium phosphate within these casein micelles makes for a very complicated structure.

Casein Micelle Biosynthesis is complex!

Casein micelle production requires not just basic protein synthesis but specific translational modification necessary for the caseins to be in the right shape and conformation that they can start to aggregate. They then bind significant amounts of calcium and phosphate to trigger association.

Glycosylation is another key post-translational modification that controls the role of one of the caseins: kappa-casein on stabilising the micelle. It controls its role on the outside of these micelles and in turn stabilises the rest of the caseins from precipitation which other wise easily happens.

It’s not exactly clear which enzymes and environmental conditions dictate biosynthesis functions. We probably know which enzymes are now involved but we’re not 100% sure. Neither are we sure of all the environmental conditions dictating the biosynthetic processes. From a scale-up or start-up perspective, there are no GRAS enzymes available and not at commercial scale to replicate what is done in the mammary gland.

The ionic environment is complex, with the different types of salts and concentrations used.  These need to be produced at the right level and right sequence as was found during artificial micelle formation. If it’s not in the right sequence, the whole micelle just precipitates. 

The casein micelle structure was examined in its entirety back in 1998 (McMahon & McManus, 1998) alongside a binding model developed by Horne (1998).

Studies show a dual binding casein micelle model of Horne (1998).  A model for casein micelle aggregates. It uses alpha S2, alpha S1, beta- and kappa-caseins with CCP. It involves a polymerization or gelation reaction where the caseins are aggregating together to make a large, extended open structure. Two features of the model are: in the triangles formed are CCP or calcium phosphate nanoclusters or crystals that actually precipitate or are formed within these micelles. About 9% of the dry weight of casein micelles is calcium phosphate and citrate, so its very rich in these nutrients. The kappa-caseins are glycosylated on the outside of the micelle. This has a stabilising feature.

Many models available

The key is the formation of the nanoclusters. These calcium phosphates are formed at centers within these micelles and they’re stabilized also by these caseins so they don’t precipitate out. (Think of kidney stones with uncontrolled precipitation of calcium-based salts).  The phosphorylation is critical. Just look at the amino-acid structure of the Bos alpha1-casein B-8P. (McI.Whitney (1988) Fundamentals of Dairy Chemistry). All the phosphate groups are formed in clusters on the alpha S1 casein molecule. This clustering of phosphate groups allows it to become a super centre to bind calcium and form these nanoclusters. They’re not randomly spaced throughout the casein micelle. We know the phosphoserines are responsible (see fig 2.20 Biochem. J 127 (1971) pp. 237 for most of the binding of calcium. Because we do isotherms where we can add more calcium to see how much is bound. Its proportional to the amount  of phosphoserines on a casein molecule. If you remove the phosphate groups using enzymes, the calcium binding dramatically decreases. It loses its calcium binding ability.

Glycosylation is only on the kappa-casein part of the protein. It’s at the 133 bond of kappa-casein between a serine and threonine amino-acid. the sugar is a complicated sialic acid-type group which is n-acetyl-neuraminic acid, which is a complicated sugar.  Its an important part because rennin acts at bond 105 and 106 between a phenylanine and methionine in the amino-acid chain. The protein chain after bond 106 is removed during the renneting process. That starts a very rapid cascade of aggregation and gelation in the cheese-making process with the sugar group on the discarded part. The cleavage of the protein and the displacement of the sugar moiety from the maintenance of the kappa-casein three-D structure plays a loss of stability in the micelles and hence the basis of cheese making.

What is the physical/chemical basis for that stability? It is steric stabilisation. It’s the outside peptides that have sugar groups on the micelles on the kappa-casein on the outside based on the model of Horne. These prevent other micelles coming too close together.

The steric stabilization by glycosylated kappa-casein hairs prevents aggregation of casein micelles. These hairs are removed by rennet and they start aggregating. The stabilizing layer is lost. The micelles aggregate closer to start gelation.

Summary of process

There are 4 steps involved. These are:

1. Make the Protein sequence. Could be done with lab grown proteins only accomplish the 1st step. We have 4 major types of caseins, so 4 different fermentation processes to generate all the caseins.  It is too complex to grow and separate all 4 caseins in one fermentation medium.
2. Add phosphate and sugar groups as part of the post-translational modification -specific enzymes add phosphate and sugar groups on the caseins on specific sites. We want to cluster the phosphates to have the maximum ability to bind phosphates.
3. Assembly phase – the CaP nanoclusters form along with association of caseins. These nanoclusters are entrapped safely in casein micelle aggregates to get the final micelle form.
4. Micelle forms – sugar groups on outside provides stability to the casein micelle. These have the phosphoserines in the middle and on the outside these sugar groups providing the stability.

It is generally an extremely complicated process to form a micelle and thus difficult to reproduce in the lab and synthetically.

We can study the structures formed in the mammary cells themselves. The kappa-casein acts to terminate the aggregation process because it lacks the phosphoserine cluster to continue aggregation process. Read about this in the chapter by Farrell (1988) in ‘Fundamentals of Dairy Chemistry’.

Its important because insoluble calcium phosphate is a key crosslinking material in the cheese matrix.

• a cheese with a high insoluble calcium content has a very interconnected protein matrix and a poor melt.
• the same cheese but with a low insoluble calcium content: less interconnected protein matrix, excellent melt.

The amount of calcium phosphate within or cross-linking these casein micelles is critical for functionality of cheese. If you make cheese without any acid development, you have a very highly interconnected or cross-linked cheese that doesn’t want to melt. To have melting cheeses requires acid development to remove some of these Ca phosphate nanoclusters. These are readily dissolved by acid. It makes them much more pliable, softer, more meltable and producing a stretchable cheese. So we have the basis of mozzarella.

The properties of artificial casein micelles were being studied in the 1970s (Schmidt, 1979) alongside their composition with respect to αs1-, β and κ-casein, colloidal phosphate and citrate. In the laboratory, you can create artificial casein micelles from individual caseins that are already post-translationally modified by careful addition of milk salts like calcium, phosphate and citrate. These artificially created casein micelles look similar with many similar properties of a natural casein micelle. The cow had already produced the caseins for the laboratory study! The study also hinted at how the post-translational modification works to a certain extent.

Impacts On Climate Change

The climate footprint (GWP) is calculated as kg of CO₂ per kg product. The skim milk and oat beverage is very similar at 0.5 carbon equivalents per 1 kg of product. An almond beverage is 3.5 (highest), the Perfect Day whey protein is 2.75, milk powder manufacture is 1.5 and dry whey processing product is 0.75. These results are taken from the web-site https://denstoreklimatedatabase.dk/en. We also recommend checking the article by Finnegan et al., (2017) and on life-cyle assessment of the cheese and whey industry by Aguirre-Villegas et al., (2012).

There will be a carbon footprint for making new milk products using precision technology as in Perfect Day’s whey protein product. We know of no peer-reviewed studies yet for what the carbon footprint is in precision fermentation – we are relying on web-site data.

Opportunities for Precision Fermentation in The Dairy Industry?

The focus quite rightly is on high value components such as rennet because it is still the most expensive ingredient in cheese making. But perhaps we should now pay attention to those components, especially proteins that are present in low concentrations in milk. Exploiting precision fermentation here would make it all more cost effective.

It’s best to focus on the simple proteins – the ones where post-translation and further bioassembly is not required. These include lactoferrin, lactoperoxidase, immunoglobulins, etc.

The use of lactose-rich dairy feedstocks e.g. permeate, acid whey for the fermentation process improves the sustainability of diary processing.

Summary

Simple proteins or enzymes can be successfully made using precision a fermentation approach such as rennet and beta-lactoglobulin.

Complex proteins like casein require complex post-translation modification, but its unclear how to do these outside the mammary cells. It’s unclear how to do a bio-assembly of complex protein structures like casein micelles done without post-translational modification and without mammary cells. What then is the functionality of the non-cow derived casein proteins? We should really focus on high-value proteins milk at low concentrations such as lactoferrin because a cost-benefit analysis is critical.

References

Aguirre-Villegas, H. A., Milani, F. X., Kraatz, S., & Reinemann, D. J. (2012). Life cycle impact assessment and allocation methods development for cheese and whey processing. Transactions of the ASABE, 55(2), pp. 613-627.

Finnegan, W., Goggins, J., Clifford, E., & Zhan, X. (2017). Environmental impacts of milk powder and butter manufactured in the Republic of Ireland. Science of the Total Environment, 579, pp. 159-168

Horne, D. S. (1998). Casein interactions: casting light on the black boxes, the structure in dairy products. International Dairy Journal, 8(3), pp. 171-177

McMahon, D. J., & McManus, W. R. (1998). Rethinking casein micelle structure using electron microscopy. Journal of Dairy Science, 81(11), pp. 2985-2993 (Article).

Schmidt, D. G. (1979). Properties of artificial casein micelles. Journal of Dairy Research, 46(2), pp. 351-355 (Article).