The science of fermentation

The current state of target selection 

Biology provides food developers with an almost boundless palette of molecules from which to assemble flavors, textures, and aromas. However, not all these ingredients are easily sourced at large volumes and low prices. By using microbial cells as the production host, precision fermentation allows for highly scalable manufacture of virtually any ingredient.

Target selection and design is the starting point for the process of precision fermentation. The molecule or molecules of interest are referred to as the target. The target can be a protein, a lipid, a flavor compound, a fragrance, an enzyme, a growth factor, a pigment, or another class of molecule. 

Fermentation-derived ingredients are already widely used across the food industry.

The majority of vitamins in nutritional supplements and fortified processed foods, such as B12 and riboflavin, are produced through fermentation, as are many flavoring components. The food industry was among the first to leverage fermentation to displace animal products in everyday use. 

Commercialization of fermentation-produced chymosin (the major enzyme in calf rennet, taken from the lining of calves’ stomachs) in the 1980s rendered calf rennet’s previously vital use as a coagulant in cheesemaking obsolete for most global cheese production. 

A notable recent example of precision fermentation’s use in alternative proteins is Impossible Foods’ use of purified soy leghemoglobin. “Heme,” produced using Pichia pastoris, is used as a flavoring ingredient in their burger to produce a suite of organoleptic properties in the cooked product. 

Other recombinant proteins, such as casein and whey, are key targets because of their unique functionality in dairy products. These proteins can be combined with plant-derived ingredients to create a final product. For example, sugar, coconut oil, and sunflower oil are combined with fermentation-produced recombinant whey to make Perfect Day’s ice cream base. 

Precision fermentation targets specific molecules.

Target molecules such as animal-origin-free growth factors are used in the production of cultivated meat. Several companies, including ORF Genetics, Richcore, and Peprotech, already work in this space. Furthermore, proteins such as collagen or fibronectin produced through fermentation may serve as key animal-free components of scaffolding for more complex, highly-structured cultivated meat products. 

In the case of a protein target, the instruction manual for synthesizing the protein is encoded in the host organism’s DNA, either as a naturally occurring gene or as a gene introduced through engineering. Depending on the target, both engineered and non-engineered approaches may be possible. 

For example, the soy leghemoglobin protein produced by Impossible Foods is engineered into a yeast host strain for efficient, scalable production. On the other hand, microalgae company Triton Algae Innovations is commercializing heme proteins that are native to their algal strains, so no engineering is involved. 

Synthesis of non-protein targets cannot be encoded directly in the host’s DNA. Instead, the genome encodes a series of enzymes that compose the biosynthetic pathway for producing the target molecules. 

For example, the target molecules for algal omega-3 production are the fatty acids DHA and EPA, but the instruction manual for manufacturing these fatty acids consists of several gene-encoded enzymes that convert precursor fatty acids into these desirable fatty acids within the cell. As with protein targets, molecules like fats or flavoring molecules can be produced in microbial hosts either with or without the use of engineering techniques, depending on the specific target and the choice of host organism.

Where target selection innovation for precision fermentation is headed 

Fermentation allows for a decoupling of the original source of a target molecule and its production method. This decoupling vastly expands the search landscape for biomolecules with unique and valuable functions.

First, ideal targets may originate in species that are extraordinarily rare, difficult to harvest, expensive, or otherwise inaccessible or impractical. Fermentation provides a mechanism for manufacturing these molecules at scales and prices suitable for commercial viability.

Second, targets are not limited to those found in nature: Novel variants of target molecules can be engineered through random alteration and screening (directed evolution) or through rational design, leading to targets that exceed the performance of any naturally occurring version.

Mastodon collagen anyone?

Geltor’s collagen production platform illustrates both aspects of this expanded search landscape. Gelatin (a form of collagen) from conventional animal sources is limited to a few species (predominantly pig and cow, although fish gelatin is also commercially available) that are processed in large quantities. But collagen is ubiquitous in the animal kingdom, and Geltor can manufacture collagen proteins from any species, including extinct species.

In 2018, the company showcased the versatility of their platform with an animal-free leather binding made from jellyfish collagen and gummy snacks made with mastodon collagen. Geltor also makes bespoke versions of collagen that are precisely tuned to the characteristics desired for a particular application — for example, gelatin that exhibits a specific gelling viscosity, elasticity, or melt temperature.

Similarly, fermentation allows for enzymes to be adapted or engineered to exhibit higher activity, novel substrate specificity, greater stability, or robustness under specific processing conditions, with dramatic implications for cost reduction. Such enzymes serve many purposes across the alternative protein industry.

These examples show that fermentation holds immense potential to screen for natural variants of targets and to design new variants for augmented sensory, functional, or nutritional properties or for attributes that reduce costs and streamline manufacturing processes.

Current state of microbial strain development 

In theory, microbial fermentation encompasses an enormous variety of species with vast biological diversity, ranging from fungi to bacteria to microalgae. However, exceedingly few microbial species have ever been commercialized for use in food.

New technologies allow us to overhaul the search for new lines.

The emergence of high-throughput screening and characterization tools, in addition to in silico capabilities, merit a recanvassing of all known microbial species for their potential suitability as protein sources.

For example, the strain Fusarium venenatum, the filamentous fungus first commercialized by Quorn, emerged from precisely such a screening effort. This particular search was initiated in the 1960s over concerns about an impending protein shortage, but the constraints of that era’s research tools meant that discovery efforts were limited in scope.

Likewise, the catalog of host strains used as microbial factories for producing high-value targets is also overdue for an overhaul. For decades, the fermentation sector has relied predominantly on a small number of well-established staple species.

While there is room to further design these species for higher yields, more robust cultivation, faster growth, and other ideal production traits, the sector has been limited to these species because of their longstanding use and familiarity, as well as regulatory barriers to commercializing new host species.

The future of strain development for fermentation

A program of systematic strain discovery and development conducted with today’s tools — and bolstered by computational insights based on the wealth of genomic data now available for tens of thousands of microbial species — could yield entirely new candidates for microbial protein production, in addition to providing a roadmap for continued improvement of existing strains.

High-throughput methods of strain selection, adaptation, screening, and engineering enable innovators to iterate new strains with greater speed and precision. They can select for more nuanced attributes, such as precise flavor-enhancing metabolite profiles, rather than simple traits like growth rates or temperature tolerance.

While some of the strain development work in this sector is likely to involve biotechnological tools, such as gene editing and genetic engineering, vast progress remains to be made through simple adaptation and breeding strategies powered by advanced genomic insights.

Clearing the way for the next generation

Strain development research can pave the way for new workhorse strains that can significantly outperform the incumbents. We also need visionary regulatory leadership to streamline commercial adoption of new candidates. Fermentation should not be stymied by historical allegiance to legacy host strains or outdated regulatory procedures. 

In addition, comprehensive efforts should be undertaken to assimilate systems biology insights regarding metabolic pathways across species to aid in identification and design of novel hosts with ideal attributes. These attributes may include prolonged generational stability to support long-term continuous culture, metabolic pathways that use a wide range of feedstocks, a desirable flavor profile, and low levels of nucleic acid or other undesirable metabolites.

The current state of feedstocks for fermentation

Feedstocks provide the nutrients — the basic building blocks of life — to support microorganisms’ growth during fermentation. Much of the resiliency and adaptability of fermentation derives from its innate malleability with regard to these feedstock raw material inputs.

At the same time, feedstocks are a major cost driver for most fermentation processes. Thus, a great deal of optimization is possible in engineering industrial-scale production schemes to use unconventional feedstocks, including potential sidestreams from other industries. This presents potential gains for both economic viability and sustainability. 

At present, the majority of fermentation relies on fairly standardized, refined, sugar-based feedstocks. These have a long history of validated use in both food and industrial biotechnology fermentation processes. To reach mass commercialization, cheaper and more sustainable substrates must become widely available. Additional research is needed to move beyond this paradigm and empower fermentation companies to leverage more diverse inputs.

The future of better feedstocks

An increasing number of companies and researchers are capitalizing on the potential to convert waste products or agro-industrial byproducts into high-quality protein biomass. Nature’s Fynd produces protein from extremophile fungi isolated from a thermal spring in Yellowstone National Park.

These fungi exhibit wide metabolic flexibility and therefore can use diverse feedstocks. 3F Bio and Mycorena in Sweden also position themselves as leaders in sustainable feedstock use. Other startups, including Air Protein, leverage gaseous feedstocks, deriving energy from chemical reactions involving hydrogen, methane, or carbon dioxide gas.

Building a global bioeconomy

Feedstock optimization should be considered in the context of global shifts in demand across many biological raw materials. The rise in demand for fermentation feedstocks is driven by a wholesale shift toward a bioeconomy model of production. This bioeconomy could potentially leverage microbial platforms for manufacturing not just food and pharma products but also green chemical products, biopolymers, and fuels that have historically been dominated by petrochemical-based production.

With this perspective, it is possible to engage in more strategic decision-making regarding the location of new fermentation facilities, placing them near abundant low-cost feedstock sources.

Feedstocks should also be examined across all alternative protein production platforms, including plant-based and cultivated. All these production modalities currently require slightly different feedstocks as primary inputs, and strategic forecasting of raw material demands across all sectors informs better decision-making regarding processing, sourcing, and formulation. 

The industry’s ability to nimbly tap into diverse, unconventional feedstocks will also be bolstered by the adoption of globally recognized standards and the development of novel characterization technologies. These will give purchasers confidence in the quality and performance of the feedstock material they buy. These standards will also equip them with the predictive capacity to adapt their process as needed to suit a given lot, even if it is from a source or of a composition they have not routinely used in the past.

The current state of bioprocess design

For decades, industrial microbial fermentation has operated at massive scales, with individual cultivation tanks as large as hundreds of thousands of liters. However, these production facilities represent relatively limited options for process design to accommodate novel organisms or to suit the manufacturing requirements of alternative protein applications of fermentation.

For example, the vast majority of fermentation facilities currently operational for industrial biotechnology and bioethanol production use submerged fermentation, meaning that the microbial cells are suspended in a liquid nutrient medium. However, some of the structured and intact uses of whole-cell biomass (e.g., whole meat cuts) rely on solid-state fermentation, where the microbes are inoculated onto a moistened solid feedstock that may be enclosed or even grown in open air.

Even within submerged fermentation bioprocesses, the scale, cost sensitivity, and sustainability considerations associated with alternative protein applications may warrant approaches distinct from classical stirred-tank bioreactors. Some fermentation companies are successfully using novel bioprocess and bioreactor designs, but relatively little attention has been paid to further optimization or iteration of these designs simply because few companies have used them to date.

Clearly, there is still ample room for innovations in bioprocess design to meet the unique needs of the alternative protein industry. 

What could be next for bioprocess design?

Alternative protein players using fermentation have already demonstrated some bold thinking in bioprocess innovation. For example, Quorn pioneered a bioreactor design called air-lift fermentation, which requires substantially less energy than conventional bioreactors while accommodating large volumes. This bioreactor design is well suited for filamentous fungi, which increase the viscosity of the solution more than non-filamentous fungi or bacteria.

Animal feed companies using gaseous feedstocks — such as Unibio and Calysta, who work with methanotrophic bacteria — often apply a similar concept, whereby feedstock gases and other gases circulate the cells and media in an enclosed loop, eliminating the need for an impeller.

Several companies using fermentation to create whole muscle cuts or to convert plant-based proteins into more functional ingredients are exploring the use of solid-state fermentation platforms. These systems may offer cost savings and lower barriers to entry because they do not require the same capital-intensive stainless-steel bioreactors needed for submerged fermentation.

Solid-state fermentation platforms also open the door to scale-out approaches rather than scale-up approaches — increasing capacity through parallel small-scale units rather than larger-volume single units. This approach mitigates some of the technical risks and capital costs associated with scaling.

Are existing facilities an opportunity?

Another key opportunity area is research into retrofitting existing manufacturing facilities and equipment to suit the needs of alternative protein applications. For example, most existing fermentation infrastructure was built for anaerobic bioethanol production.

As the world moves increasingly toward renewable energy and electrification, these facilities may be decommissioned in the coming decades. However, deeper analysis is needed to understand under what circumstances converting these facilities for alternative protein applications would be technically feasible or fiscally merited. 

White Dog Labs recently purchased an ethanol plant with plans to convert it to aquaculture-feed protein production. They can do this because their microorganism tolerates anaerobic growth. This example illustrates the importance of coordinated R&D efforts across all the key technology development areas to capitalize on insights at the interfaces of these disciplines.

Most microorganisms currently used for food ingredient fermentation require aerobic growth, but if technoeconomic and engineering analyses indicate that conversion from bioethanol facilities to anaerobic food production facilities is highly feasible, then anaerobic growth should be a key screening condition within comprehensive strain assessment efforts.

Explore research opportunities in bioprocess design

Fermentation-derived ingredients cost less and create better texture.

There is ample room to develop other novel, relatively low-tech and low-cost structuring solutions. These can further improve texture without the capital costs associated with the high-moisture extrusion used to produce many current plant-based proteins.

Individual companies will iteratively improve on their recipes, potentially incorporating flavorings, fats, binders, functional enzymes, and nutritional fortification to achieve sensory profiles that more closely mimic those of their animal-derived counterparts.

In turn, many of these ingredients may themselves be produced by B2B providers of fermentation-derived ingredients. Thus, many opportunities for innovation at the end product level are identical to those of plant-based products.

Explore research opportunities in fermentation

Meet the author

Liz Specht, Ph.D.

DIRECTOR OF SCIENCE AND TECHNOLOGY

Liz Specht oversees GFI’s Science and Technology department to build a roadmap for accelerating alternative protein research while empowering scientists to execute on this vision.

More about Liz