Tuesday, May 24, 2016

Minimalist food supply - synthetic amino acids

 Plants are not a particularly efficient source of protein. They tend to be better at producing carbohydrates. As a result, vegetarian diets often focus on a few high-nitrogen plants like beans, soy and peanuts.

 I tend to explore food systems that are familiar, but let's take a minimalist approach and see where it leads. Instead of deriving protein from plant sources directly, what if we use microorganisms to produce amino acids in bioreactors using plant starch as input? This is like that sci-fi staple 'vat meat', but with neither texture nor flavor. Still, amino acids can be stored for years (possibly decades) if powdered and sealed.

As with all my posts, this article is based largely on internet research. I am not a process biologist. I've included sources where possible, but these results should be considered preliminary at best.

Short results: vat-grown amino acids can be yours for 4-6m³ per person.
That includes vats, supporting equipment, hydroponic space and waste treatment.
You will still need to provide bulk calories and other nutrients.

More after the break.

 Let's start with demand. The indispensable amino acids  are tryptophan, threonine, isoleucine, leucine, lysine, methionine, phenylalanine, valine and histidine. Required amounts vary; women typically need the least at about 14g, men in the middle with about 18g and children requiring the most at about 22g. Contrast with protein requirements of 110g, 140g and 90g respectively. These figures are from my menu tracker which is based on US dietary recommendations, so do not take this as specific dietary advice.

 Dietary protein serves two purposes: first as a source of food energy and second as a source of amino acids. If we eat only amino acids then the 'missing' food energy needs to be provided by additional carbohydrates and fats. That also means we only need to eat enough amino acids to satisfy the body's need for building material. The science is not settled, so let's assume we need twice the minimum amount or an average of 36 grams per person per day. Some of this will be provided by other food items but let's ignore any other protein sources for now.

 Production is varied as there are several distinct chemical structures involved in the various amino acids. Each type or family requires a specific environment, feedstocks and process. A basic overview is available on Wikipedia, while a deeper look at processes for the aromatic amino acids is available here. This is just an example; the industry has been active for decades and there is a lot of material available on this technology.
 Plant starches can be readily decomposed into glucose using microbes or enzymes. (see for example the process of making sake; rice is decomposed by a fungus to produce sugar-rich material for yeast to ferment into alcohol.) Sweet potato mash with autolyzed yeast and/or spirulina should be a reasonable starting point for producing a viable nutrient solution.
 The chosen microorganism is grown in a starter culture (1-5l) for half a day and then transferred to a large vat (100-500l) for fermentation over an additional one to three days. For industrial and medical purposes the finished broth is lysed by freezing, vibration or centrifugation and then separated by centrifugation. Further purification steps are applied including filtration and fractional crystallization. For nutritional use this high-grade purification may not be necessary. An example might be alanine as a byproduct of valine production; for an industrial process this would be a contaminant but as a food source it's beneficial. In this case a single-step centrifugation/separation can be applied with the result tested for composition and then dried without further processing.

Yields listed below are lab results, specified in units of product per unit of glucose. Nonessential amino acids are in italics. Many of these values leave significant room for improvement.
There is also a note in the book "Corynebacterium glutamicum: Biology and Biotechnology"  (edited by Nami Tatsumi, Masayuki Inui) on page 111 giving a snapshot look at yields of industrial C. Glutamicum fermentation processes. These values are weight percent, given in {} brackets below; as you can see, most products scale up well but a few do not.

Glutamate - {50%}
Glutamine -  {40%}
Proline - 36% g/g
Arginine - 35% g/g

Aromatic amino acids
Phenylalanine - 25% g/g {50%}
Tyrosine - 30% g/g
Tryptophan - 14% mol/mol {22%}

Lysine - 42% mol/mol {50%}
Asparagine - unknown
Methionine - 20% {17%}
Threonine - 60% {45%}
Isoleucine - 22% {25%}

Ribose 5-phosphate
Histidine - 5.5%

Serine - 45% g/g {32%}
Glycine - unknown
Cysteine - unknown

Alanine - 86% g/g {50%}
Valine - 88% mol/mol {35%}
Leucine - 30% mol/mol

A ballpark value for the overall average yield is perhaps 40% by weight, requiring 90 grams of glucose per person per day.

 Final concentrations average around 20g per liter, with some exceptions (histidine in particular is perhaps 5g/l while some others are over 45g/l). That works out to 8g/l/day or 4.5l per person. Let's double that volume again (for redundancy) and call it about ten liters of bioreactor volume per person. Space is required for starter cultures, nutrient processing and storage; let's assume this supporting volume (and wasted space due to packing issues) is twice that of the vats and call it 30 liters of volume per person. At a thousand liters per cubic meter, 1m³ could serve a bit over 30 people.

Let's go over the drawbacks:

 - The bioreactors have to be controlled (pH, temperature, glucose, nitrogen, oxygen, agitation), which requires energy. Information on power requirements are difficult to find, so I can't offer even a bad guess. This also means some smart tech is required for automation.
 - Product yield is less than 1% of the final broth, so assume that the entire volume has to be treated as wastewater. About 5 liters per person per day.
 - As a biological process, cleanliness is essential. Quantities of soap, alcohol or bleach and washwater will be required.
 - Population control is important. Bacteria evolve rapidly and cross-contamination is a possibility. Vats must be monitored for invasive species and unexpected byproducts. Reserve supplies
 - At least nine separate product lines are required. More may be needed to completely replace food protein needs.
 - The nutrient solution has to provide everything needed for the microbes to grow and thrive. This is on par with intensive hydroponics for complexity.
 - Centrifuges are tricky in microgravity. Angular motion needs to be carefully balanced, so vats should always be counter-spun in mass-matched pairs.
 - The system requires carbohydrates as inputs. These in turn require production of enzymes to break the starch into glucose.
 - Amino acid powders are not particularly appetizing.

These are largely the same drawbacks that apply for alcohol production and can easily be integrated into that workflow. Nutrient supply can be integrated with the hydroponics workflow. Wastewater can be handled largely by evaporation or ultrafiltration (though centrifugation should yield concentrated sludge and fairly clean water as separate outputs), with concentrated wastes blasted in the SCWO reactor and recovered by spirulina. Carbohydrate supply can be provided by either sweet potato (24.0g/m²/day) or wheat (18.5g/m³/day), requiring 3.8 to 4.9 m³ per person. These inputs would co-produce edible protein extracts amounting to 8.6g or 16.5g respectively. Not much can be done for the taste other than combining with protein extracts or other food ingredients, unfortunately; perhaps that will change with applied research.

Here are the advantages:

 - No animals.
 - Simple waste streams.
 - Extremely compact.
 - The tools and techniques can be applied to other biosynthesis products such as ethanol and other industrial feedstocks as well as a broad variety of medicinal compounds.
 - The techniques for producing the necessary inputs are largely the same as for intensive hydroponics.
 - A variety of techniques are available for continual improvement, including selective breeding, directed mutation and outright genetic engineering. Microbial gene editing (such as with CRISPR) allows earth-based developments to be applied to in-space organisms by sending only data.
 - The process can be scaled easily for populations from one to a thousand or more.
 - All of the technology can be built, tested and optimized on Earth. A round of low-gravity performance testing and some engineering sweetness to handle periods of microgravity would be a good idea before deployment but not strictly necessary for use on a spinning habitat.

I still believe it would be beneficial to use all plant wastes as feed for insects, fish and possibly chickens. This represents recapture of energy that would otherwise have gone to waste into food products for variety. Even so, a vat process could be paired with suitable hydroponics for complete nutrition with high efficiency. Sweet potato greens are edible, so a menu could be devised that minimizes plant waste. A steady diet of sweet potatoes, salad and amino acid seasoning doesn't sound very appealing to me but it might be among the earliest self-sufficient space food systems.


  1. "Amino acid powders are not particularly appetizing."

    This is a considerable understatement.

    1. See for example:

      Citrus fruit juice is a tall order for self-contained hydroponics. Fortunately, citric, ascorbic and malic acids can all be produced by fermentation similar to the amino acid processes. An artificial citrus amino drink sweetened with glucose would be reasonably palatable, like faintly gritty lemonade with bitter undertones.
      An alternative might be to mix the hydrophobic amino acids with, say, peanut butter instead of mixing into a drink.

    2. In my experience mixing them with food goes a long way towards dealing with the taste. Histidine is still rather unpleasant that way, though.