In this series of articles, NIZO Food Research Division Manager and FoodNavigator advisory panel member René Floris discusses some of the big issues in today’s food industry. This month, he continues his exploration of the protein transition by looking at the food safety for plant-based foods.

The market for meat- and dairy alternatives is growing at a massive rate. But while these plant-based products may deliver as close an eating experience as possible to their animal-based counterparts, the food safety challenge they present can be very different. To find out more, I spoke to Marjon Wells-Bennik, Principal Scientist Food Safety, who has over 20 years academic and industrial experience in food safety and microbiology.

Joined up thinking throughout the chain enables plant-based food safety

René Floris: How does the food safety picture for plant-based differ from animal-based food?

Marjon Wells-Bennik: Food safety is a much more varied challenge for plant-based foods and dairy alternatives. For example, in the dairy industry, there is basically one main ingredient – raw milk – which is fairly consistent in terms of its nutritional composition, physical characteristics and microbiological contaminants.

Plant-based proteins, on the other hand, come from many different sources. Each has its own mix of proteins, sugars etc. The variety of microbes present is also much greater – and the levels of microbes can vary greatly too. The soil where the source plant is grown, how the plant is harvested, how the plant materials are processed to obtain the protein, even how plant and resulting ingredients are stored and transported can play a role. So, you have very different microbial contamination starting points and growth conditions.

To make matters even more complex, plant-based proteins can have very different solubilities and reactions to heat, which affects how they can be processed. All this makes food safety a very challenging issue for plant-based products – one that requires a more holistic approach.

RF: It is often said that microbial contamination in plant-based foods is less well understood. Is that true?

MW-B: The dairy industry has been investigating microbial contamination for decades and, as I said, the conditions there are much more consistent. So it is not surprising that knowledge of microbial contamination in plant-based protein ingredients is less developed. But progress is being made. For example, at NIZO we have carried out our own research into microbial contamination of ingredients used to make dairy alternatives based on coconuts, oats, almond, peas and other legumes. We found that both the level and variety of microbes was higher than in raw milk. Some of the varieties found in plant-based ingredients are familiar from the dairy industry, such as Bacillus cereus and other Bacillus and Clostridia species. But some are new, and we do not know how the product composition influences their growth.

A large proportion of the total microbial count in plant-based ingredients consists of spores from bacteria such as Bacillus subtilis, B. licheniformis and B. amyloliquefaciens, but we alsofound B. cereus, and Geobacillus stearothermophilus. There can be as many as 1000 spores per gram in plant-based ingredients. These spores can survive high heat treatments and, upon germination, can go on to produce (heat-resistant) toxins that present a health risk. This contrasts with spore counts of around 1-10 per ml for raw milk, where the majority of bacteria are inactivated by pasteurisation.

Figure 1: Total aerobic and spore counts for samples of raw milk and various plant protein isolates.

RF: Are more extreme heat treatments than pasteurization required?

MW-B: That has two problems. Firstly, many plant proteins are denatured by high temperatures, affecting their taste, texture and nutritional value. And at the end of the day, the goal is to produce a food product that consumers want to eat or drink. Secondly, some of the spores identified in plant-based ingredients can survive even the most extreme heat treatments used in food processing. Novel techniques like Innovative Steam Injection (applying very high temperatures for very short times) are promising, particularly for inactivating spores in liquid end products, but they have not yet been widely applied to plant-based foods.

RF: Can we apply non-thermal techniques?

MW-B: There are techniques familiar from the dairy industry that we can apply to plant-based products, but we need to do so in a smart way and drawing on understanding of plant-based ingredients. For example, common heating processes such as pasteurization kill bacteria but not the spores. A technology like bactofugation has been successfully used to reduce bacterial spore levels in milk. This could also be attractive for plant-based products. However, it only works with liquids and the solubility of plant proteins varies greatly so you would need to consider that before developing a bactofugation-based process.

Fermentation is another possibility for improving product stability. Fermentation has been used for millennia in making cheese and yoghurt among other things, so is well understood. In most dairy fermentations, lactic acid is produced from lactose and helps preserve the food. Similar preservation can be achieved for plant-based products through fermentation or chemical acidification (for instance adding lactic acid). Spoilage by (heat-resistant) moulds is a major concern for yoghurt and cheese alternatives. Our own research has shown that rapid acidification through fermentation was more effective at preventing mould growth than chemical acidification – fermented samples of plant-based cream cheese were still mould free after six weeks at 6 °C.

However, plant-based ingredients have different sugars and fats as well as different proteins to milk. So you can’t just assume that you can use the same lactic acid bacteria for fermenting plant-based products as in dairy products. The presence of sugars other than lactose may allow spoilers to overgrow the fermentation culture, spoiling the product before it can be sufficiently acidified. You can avoid this by reformulating the product to adjust the sugar profile or using a different starter culture. The industry is starting to build up knowledge of which cultures work best with which plant-based ingredients, but there is still a long way to go. So today, it is often still a matter of trial and error.

RF: So how do you find the way through this complexity to a safe product?

MW-B: Food safety is always a matter of joined up thinking from ingredient to consumer. But this is even more true for plant-based products. The complexity of the microbiological situation, the different potentials for microbe growth mean you truly have to think about the whole chain holistically: where plants are grown, the extraction of proteins, transportation of ingredients, processing steps in manufacturing, the distribution and retail chains, desired shelf life, and how consumers will store and use products at home. Each of these steps could be a cause of contamination or pathogen and spoiler growth. But they can also be a potential solution.

For example, if you can’t use bactofugation in your final product because your protein isn’t soluble enough, perhaps you can prevent or remove contamination at an earlier stage in the chain. Maybe you need to consider using a protein concentrate obtained via a dry process instead of an isolate where the wet fractionation and subsequent evaporator process can encourage the growth of spore formers.

Predictive modelling is a powerful tool here. It allows you to carry out initial microbial risk assessments in silico, based on processing conditions, the intrinsic properties of the product, and the intended storage and consumption conditions. Such assessments can help you identify microbes of concern and make informed decisions on preventative measures and even product preservation strategies. For example, if you are planning a shelf-life stable product but the key microbes of concern are particular Bacillus bacteria which don’t grow at refrigeration temperatures, perhaps the best approach is to switch to a chilled product at least initially.

Figure 2: Oat drink spoiled by B. subtilis (left) and Almond drink spoiled by B. licheniformis (right).

Each new plant-based product can bring a new food safety challenge. But by taking this kind of end-to-end approach, it is possible to ensure ‘safety by design’ and deliver tasty, high-quality plant-based products.

Next month we will be discussing flavour, taste and mouthfeel of plant proteins.

In the previous blog, we looked at the range of alternative protein options available and how to choose the right one for your new product. Having chosen the right protein source (or sources), the challenge turns to ensuring you can maintain the desired functionality of that protein during processing and deliver the associated benefits to your customers.
In this blog, which was also shown recently in Food Navigator, the role of processing in protein functionality was put under the microscope by asking questions to Peter de Jong, Principal Scientist Processing at NIZO. In addition to his role at NIZO, Peter is professor of dairy process technology at Van Hall Larenstein University of Applied Sciences and director of New Technology Development for Food at the Institute for Sustainable Process Technology.

Why is protein functionality a key issue for process development?

The food industry is increasingly aware of the value of protein in food products. That goes a lot further than just the amount of protein, but also the functionality it brings. For example, as Fred van de Velde explained last month, protein functionality can influence the taste and texture of a food product, affecting how attractive the product is to consumers. It can also impact production efficiency. For example, processing can cause certain proteins to coagulate, leading to fouling and regular production shutdowns for cleaning. And of course, there is the nutritional impact of proteins, not just in terms of macronutrient properties but also more subtle effects such as binding vitamins and, particularly topical right now, their impact on immune response and anti-viral activity.
The interest in these effects is growing rapidly as the protein transition opens up new / alternative protein sources, many of which offer much greater protein functionality that meat does. For example, there is a lot of interest right now in raw milk because it seems effect our immune systems and possibly reduce allergies.

What is the impact of processing on protein functionality?

The goal of processing is to deliver a food product that is tasty, nutritious and safe. Traditionally, the food industry has taken the cautious approach – using heat treatments like pasteurisation and UHT that ensure all pathogens are killed or deactivated. However, temperatures above 80 C reduce or even destroy the functionality of many proteins. A good illustration is the Maillard reaction, where heat causes sugar molecules to bind with amino acids. You might want this when searing a steak or baking a biscuit, but when you are processing milk it reduces the bioavailability of vital amino acids like lysine which in turn can reduce benefits raw milk has for the immune system (Figure 1). Consequently, there is a big drive towards milder processing that still delivers maximum food safety while leaving more of the protein functionality intact.

So, can we just turn down the heat?

Unfortunately, it isn’t as simple as that. We at NIZO have analysed a lot of production processes and have found a great deal of variation in the protein functionality impact of seemingly similar process. Looking at the Maillard reaction I mentioned earlier, even processes as familiar as pasteurisation or UHT can vary in the amount of amino acid lost by a factor of two.
This shows two things. First, that there is plenty of room to optimize current processes. Second, food manufacturing processes are very complex, with multiple possible reactions between ingredients, each of which interacts differently with the process conditions. And that makes optimizing a process extremely challenging.

How do you start to optimize such a complex process?

One way is through data analytics: collecting as much data as you can from your factory and analysing it for any correlations. But this is a bit of a black box approach. It can help you identify which conditions or temperatures are linked to specific outcomes, but it doesn’t give you any insight into why or how to fix the issue. So, it isn’t really any help if you are trying to design a new process.
A better approach is through computer modelling of your factory set up. This does require deep understanding of the chemical reactions that can occur during food processing, but once you have built your model – or had it built for you – you can thoroughly explore the impact of variations in process conditions, either by manually tweaking process parameters in the model or by running simulations.
The results can be incredible. We have seen cases where manufacturers have been able to optimize process performance and improve bioavailability of nutrients by up to 30% without affecting the products physical properties or microbial quality specifications.

What other new technologies could help manufacturers retain protein functionality?

The industry is always innovating, finding new ways to make products better. One technology that I am excited by at the moment is called Innovative Steam Injection or ISI. This involves a very short blast of very high temperature – around 160 C for up to 1 second. This is enough to inactivate microorganisms in the food but, crucially, does not denature the proteins. Prototype ISI processes have been able to deliver “pasteurised” milk with shelf lives up to 60 days and just one-third the degradation of proteins such as β-lactoglobulin, immunoglobulin and lactoferrin (Figure 2). And even expert tasters couldn’t taste the difference.

And it is not just for dairy. ISI can be used with any pumpable fluid product. This could be very important for the protein transition as the microorganism contaminants in plant-based proteins are much more diverse and less well known. Currently, the plant-based food industry relies on extreme heat treatments which certainly kill off all pathogens but also destroy the desired functionality. As I said before, understanding how your protein interacts with your process allows you to find approaches that eliminate what needs to be eliminated (microorganisms, fouling, etc) but keeps what you want to keep in terms of nutrition, digestibility and flavour. The plant-based sector is starting to figure out what this means, but I think they could still learn a lot from the dairy sector.