Using Tobacco plants to grow antibodies for better breastmilk replication

Every single doctor and article out there will always say “Breast is best”. But I want change that, by changing the way we approach infant formula.

Elizabeth Trykin
14 min readMar 2, 2021

The stigma attached to mothers who do not breastfeed has gotten so bad, that there are entire groups of “breastfeeding bullies”.

Unfortunately, if we take away the lack of respect, empathy, and inclusiveness of these people, for the most part they are correct. And so are the articles and doctors that recommend breastfeeding. Breast is much better than formula.

Mothers breastfeeding their children is always the ideal situation. But in reality not many women are in the situation where they have both the time, the comfort and the physical ability to breastfeed.

There are many many factors that play into women’s decisions to stop, or to never even start breastfeeding:

  1. Sore nipples
  2. Baby isn’t latching on properly
  3. Breast engorgement → too much breast milk
  4. Infections (for ex. Mastitis)
  5. Not enough breastmilk

In terms of non-health related issues:

  1. Society’s standards → Women can’t just breastfeed anywhere they’d like to, so living life outside of their home is quite difficult.
  2. Breastfeeding takes about 10 hours a day, its a full time job that no one other than the mother is responsible for.
  3. Having to go back to work, which entails leaving children at home with babysitters, another guardian or in daycare.

So, about 60% of women who begin breastfeeding will stop after 3–4 weeks.

And not to mention, 5% of women physically cannot produce breastmilk. So even if they really wanted to and ignored the issues, they still couldn’t. Similar goes for another 23%, who at some point, cannot produce enough milk.

This leaves only about 1/4 mothers that breastfeed exclusively.

Infant formula — The status quo

The solution for the other 75% of mothers is infant formula, which usually consists of purified cow’s milk whey and casein as a protein source, a blend of vegetable oils as a fat source, lactose as a carb source and a vitamin-mineral mix.

In comparison, human breast milk contains live cells (stem cells), more than 1,000 different proteins, enzymes (that speed up chemical reactions in the body), growth factors, hormones, vitamins and minerals, antibodies (also called immunoglobulins), fatty acids and bacteria.


The problem: Breastmilk and current formula are quite different.

The solution: Instead of replacing the components of breastmilk with completely new ingredients, I’m going to replicate the components 1 for 1.

In this article I’m going to walk you through only one of the components, and in the future I will begin prototyping for more.

Secretory Immunoglobulin A (SIgA)

Antibodies (also called immunoglobulins), are a type of protective protein produced by all animal and human immune systems.

Antibodies work by attaching themselves to antigens, which are small parts of a bacteria, viruses, alergens, toxins or chemicals. After its attached, the goal of the antibody is to remove the antigen from the body.

Different antibodies are responsible for different antigens, which also means that there are many different types of antibodies.

Antibodies are quite complicated, with lots of different components, but all you need to understand is the basics.

There are many different antigens out there, so the antigen binding site of almost every antibody is different. All other parts are more or less constant across all antibodies. There are 5 different heavy chain types, so all antibodies are classified into either IgA, IgD, IgE, IgG or IgM. This part determines how the antibody fights the antigen.

But other than that, there is very little differentiation. This also means that the process of replication of each of these antibodies is very very similar.

Right now, I’m going to talk specifically about Immunoglobulin A (IgA), which is the most important antibody contained in breastmilk.

The purpose of SIgA

The gut is sterile and bacteria-free at birth. Then, all of a sudden an infant is exposed to a load of new bacteria, viruses, microbes etc., which is where sIgA (the secretory form of IgA) comes in.

The secretory form of IgA consists of now 4 heavy chains, 4 light chains, as well as a J chain and a Secretory component (binding the two proteins)

SIgA binds to antigens in the gut and prevents them from attaching to the inside of your stomach.

Think of it like a protective coating on the inside of a person’s stomach — the antibodies basically creates a boundary between gut microbes and the inner lining of the stomach. These antibodies prevent viral infection and the spread of bacteria across the stomach.

Memory antibodies

Every time a new pathogen (any bacteria, virus or anything else that can harm the body) enters the body, white blood cells produce new antibodies. These antibodies then go on to fight the “invader”.

But as all of this is done, the cells also make sure to write down the instructions for fighting this specific pathogen. This is a protective mechanism of our body to make sure that if the same organism ever reenters the body, it already knows how to fight it.

Now, in relation to the mother-child

A mother’s cells that already have the instructions for antibodies begin creating more antibodies and sending them to a newborn through breastmilk.

This happens through the entero-mammaric link, which is a pathway connecting the gut lymphoid tissues (where the antibodies are made in the body) to the lactating mammary glands (glands that lead to the nipple).

In a more practical example:

  1. The mother already has instructions for several SIgA antibodies against the her previous gut microbes
  2. This protection is transferred to the child through the entero-mammaric link.
  3. The breastfed infant receives the antibodies, which then coat the gut and moderate the exposure of the newborn’s gut to microbes.

Using plants as bioreactors to manufacture sIgA Antibodies

When babies are fed with formula, they lack the antibodies transferred to them by their mothers.

The child mortality as a cause of diarrhoea (2nd leading cause of child mortality) is 25 times higher in non-breastfed compared to those who were exclusively breastfed. And the leading cause of that is the absence of lining of a SIgA antibodies, which means that all sorts of pathogens enter and stay in the newborn’s gut.

So if we can’t have access to the mother’s antibodies, the next best thing is to replicate them.

When I first thought about this problem, my mind went straight to animals. If we can mix cow milk into infant formula, why not also mix in their antibodies?

Well in theory, that’s a great solution. But in practice, its extremely expensive, unsafe and therefore, not at all scalable.

So my next thought: plants. Unfortunately, plants do not naturally produce antibodies, but since they have the abilities to create proteins (keep in mind that antibodies are also a type of protein), I can engineer them to do so.

Using plants as bioreactors (like tiny factories) to produce of antibodies makes sense because its:

Renewable — there are no greenhouse gases produced

Easy — we’re pretty good at growing plants

Cheap — Plants are super cheap to grow

Global — Plants can be grown anywhere, which means that this can be produced directly in the country that needs the product

Feasible — This technology already exists

Rapid — The entire process only take 2–3 weeks

Intrinsically safe — Animal and human pathogens are not involved

Low entry barrier — Few requirements for infrastructure, skills, approvals and sterile conditions

And all of these factors make it extremely scalable.

Tobacco Plants

And going a little deeper, I specifically decided to use Tobacco plants as antibody factories because they have…

  1. Much greater ease of transformation, compared to other plants
  2. High biomass per unit area (means that there is high output with low input)
  3. The ability to production of large number of seeds
  4. Extremely high protein expression levels compared to other plants.

The process that I’m going to outline in the next part of this article is showing how we can basically engineer plants to make a 1–1 copy of any antibody that humans produce.

The production process

SIgA is the secretory form of Immunoglobulin A, which as I previously mentioned means that it has double the components. Most other antibodies, as well as the regular form of IgA have 2 light chains and 2 heavy chains. 1 gene can code for 2 chains, which means that these regular antibodies need 2 genes.

But, since sIgA has 4 heavy and 4 light chains, in order to be expressed in its complete form, there need to be 4 different genes.

Agrobacterium Tumefaciens — used to insert genes into plant cells

Agrobacterium Tumefaciens is a bacteria that can also act as a type of natural genetic engineer.

Just like Eukaryotic (animal) cells, bacteria have nucleases which contain chromosomes with DNA.

But what makes bacteria so different is that they also have extra DNA , stored in plasmids. Plasmids are small circular pieces of DNA that replicate independently from the host’s chromosomal DNA, and are not inside of a nucleus. Agrobacterium has a specific type of plasmid, called a tumor inducing (TI) plasmid.

The reason these plasmids exist is that they help bacteria much more easily transfer DNA between cells, and in result also infect other cells with less effort. So in nature, these ti plasmids are meant to infect other cells, but we can leverage them for other purposes.

Ti plasmids have an average of 200 to 800 kbp (kilobase pair, an equivilent to 1000 base pairs). A small portion of that (10–30 kbp) is called the T-region, which contains the T-DNA (transfer DNA). This portion of the plasmid is the storage place for the DNA (containing the gene) that needs to be transferred from the bacteria cell to another cell is store.

Although plasmids have many other components, the two that are required for this process are the t-DNA and the virulence (vir) region.

The virulence (vir) region is composed of virA, virB, virC, virD, virE, virF, virG and virH genes, which help control and mediate the T-DNA transfer process.

So… here’s the actual process of growing the antibodies!

The process

  1. Recognition and Attachment

Acetosyrinzon is a natural chemical compound that is released from plants when they are wounded. The virA protein (in the virulence region) senses this compound, which acts like a turn-on button for the rest of the process.

When doing this in vitro (in a lab), wounding the plant is not ideal, so instead we simply add the Acetosyrinzon chemical into the media.

VirA is a transmembrane (spans the entire membrane of the cell) sensor protein with the C-terminal portion (one end of the protein) in the cytoplasm and the N-terminal (the other end of the protein) portion in the periplasm. The reason for the different ends in different sections of the cells is so that the protein can perform more than one action.

The C-terminal region, which uses its autokinase activity to Autophosphorylate (the addition of a phosphate to the protein). Those are big words, but the process is quite simple.

Autkinase is what allows enzymes to attach phosphate groups to molecules. Autophosphorylation is the actual process of the addition of phosphate groups to molecules.

A visual of phosphorylation

As soon as it is phosphorylated, this acts like an on switch for the protein. This part is the responsibility of the other end of VirA, the N-terminal region. The phosphate is transferred from the C to the N-terminal, which then switches “on” the protein.

2. Activation of the vir genes

The VirG protein is responsible for activating the rest of the virulence proteins. The protein binds to a DNA sequence that is present in the promotor region (on and off switch of a gene) of all vir genes.

Notice how I keep using the on and off switch, since this is all a big chain reaction.

Once the protein is bound to the DNA sequence, the rest of the vir genes are activated.

3. Recognition of T-DNA

The activated VirD1/VirD2 proteins recognize the 25bp borders around the T-DNA. They then make a small cut to the bottom strand of the T-DNA at the borders (using endonuclease activity, which is the process of enzymes cutting DNA).

Whenever DNA is transferred, always only either the top or the bottom strand is used. The strand used then acts as a template for the creation of the other strand. Doing this is simply much easier than transferring both strands of DNA.

So, the DNA in between the two cuts is now the DNA that will be transferred to the other cell.

The VirD2 protein remains covalently bound (chemical bond where atoms share electrons) to the 5' end of the single stranded T-DNA. One reason for this is to prevent exonucleotic attacks, where exonucleases attack the ends of nucleic acid molecules and try remove pieces of the single stranded DNA. The 5' end is now the leading end for the T-DNA transfer.

4. Transfer into the plant cell

So now comes the actual transportation part, the complex is in a UPS delivery truck🚛. The virD2 protein is the driver of the truck, the ssT-DNA are the packages that need to be delivered, and the virE2 protein (this ones new) is the metal that makes up the truck.

The VirE2 protein is a single strand DNA binding protein that coats and protects the single stranded T-DNA-VirD2 complex. The VirE2 protein also elongates the DNA strand to reduce the diameter to about 2nm. Because of the way DNA is built, if you stretch it out, it reduces in width, which makes it easier for the complex to fit through the membrane channels (the tunnel from cell to cell).

5. Creation of the membrane-spanning protein channel

This is where the VirB proteins come in and assemble the channel between the two cells.

The VirB7 and VirB9 proteins travel to the membrane and form covalently cross-linked protein structures called heterodimmers with one another. This protein structure is then transported to the outer membrane, where it assembles into the membrane channel.

6. Integration into the plant genome

The VirE2 protein contains two plant signals and VirD2 contains one, which means that once the T-DNA structure enters the plant cell, the nucleus can sense that the complex is there.

The 3' end of the T’DNA senses homologies (similarities) with plant DNA, which creates the first contact called synapses.

Restriction enzymes then make cuts in the plant DNA and inserts the single stranded DNA into the bottom strand. Since Adenine pairs with thymine, and cytosine pairs with guanine, the bottom strand is used as a template for the upper strand.

Cool visual diagram of the process, if this helps you understand:)

Then comes the construction of the protein, which is now completely the job of the plant. The first step is that ribosomes help create a long chain of amino acids (the building blocks of proteins). The order of these amino acids is determined by mRNA.

The growth of the protein takes only about a week, which then leads us into the final steps of the process.

Extraction and Purification

SIgA consists of 4 monomeric structures, a complex of 2 heavy and 2 light chains, joined by a J chain and a secretory component. I talked about this earlier, but this is very important!!

And since this process can only support 1 gene at a time, a series of sexual crosses (cross-pollination) have to be performed between plants in order to generate plants that simultaneously have all four protein chains expressed.

So, there are now 4 different versions of plants, each expressing independently either the light chain, heavy chain, J chain, or the secretory component. These 4 versions of plants are cross-bread, which results in a new plant that expresses all 4 genes and is ready for extraction + purification.

Downstream Processing

Everything I’ve talked about now is called the upstream processing — when the biomolecules are actually grown. However, the most complex + expensive (representing more than 3/4 of the total production cost) part is the downstream processing — recovery and the purification of biosynthetic products.

There are different ways we can extract proteins from plants, but I’ll walk you through only one of them right now.

  1. Tissue Disinegration

The first step is basically breaking down the plant tissue and disrupting the cell walls, in order to maximize the amount of protein that is released during extraction. So now we’re just breaking apart the plant, as much as possible using hammer mills for wet and dry grinding, high-shear rotor-stator mixers, and high-pressure homogenizers. Basically blenders on 10x mode.

This is a hammer mill🔨

2. Solid-liquid separation

This part can get complicated, but I’m going to simple it down to a 1 minute video.

Tl;dr of what happens: A centrifugation process is used to separate particles based on density. This takes out all of the solid material and leaves us only with leaf juice.🧃

3. Using the Aqueous two-phase extraction method to clarify and pretreat the extracts

The previously performed tissue + cell disinegration also releases several other water-soluble cell components, such as DNA, chlorophyll, pigments, alkaloids, phenolics, soluble cell wall polysaccharides and proteases. These all have to removed to make sure that the purification yields are as high as possible.

If you’re interested, here’s another 1 minute video that shows the process.

4. Using Protein A to capture and purify the antibodies

The antibodies in a solution of all sorts of other components is placed into protein A resin. The solution passes through a column, where the protein binds to the antibodies and the rest of the components just pass through the column.

The antibody is then separated from the ligand (protein a) by lowering the pH levels, which changes the buffer conditions allowing the antibody to set free from the protein.

In theory and in practice, this process works perfectly, but the price of protein A purification is still extremely high.

So as of right now, my goal is to figure out the best and cheapest ways to purify these proteins.


  • There are many components in breastmilk that are not present in current baby formula. Since only 25% of women exclusively breastfeed, this creates significant problems for the 75% of children who are given formula or regular food.
  • SIgA is the most important antibody contained in breastmilk, and the process of replication is fairly simple
  • Because plants can produce proteins, I turned to plants to produce these antibodies. Using animals is not very safe, cheap, scalable or renewable.
  • I’m using a type of bacteria called Agrobacterium Tumefaciens, which is a natural genetic engineer. This is used to insert antibody producing genes into a plant, which will then kickstart the plant’s protein production process.
  • A current setback is the price of purification of antibodies, which is more than 3/4 of the price it takes to do this. That’s what I’ll be working on for next little bit, and will soon write about it.

The female body is so well adapted to carrying a child, giving birth, breastfeeding etc., but there are still so many problems associated with the entire process.

There are still mothers dying post-partum, miscarriages happen all the time, breastfeeding can be an extremely draining process, women regularly cannot get pregnant etc.

Replicating breastmilk to combat the problems every mothers experiences with breastfeeding is a step closer to helping women all around the world live a more comfortable life. And I’m so excited for it.

In my next article I’m going to talk about the implementation, costs, and mass production of this process, so subscribe to stay updated:)

Thank you for reading this!! I’d love to connect, and if you have any, feel free to send me a question:)