Turns out making antibodies, even paper models of them, is complicated! Yesterday I taught a (virtual) summer camp lesson on protein structure and one of the types of proteins I talked about was the antibody. I had thought it could be fun to make paper models of antibodies, which you can print out templates for from the PDB. But I quickly realized that, as in real life, these antibodies are complicated. But antibodies are super useful so it’s important to know what they are and how they’re made (the real ways, not the tape way). They’ve been getting a lot of attention recently because of their role in immunity, but they’ve been used in biochemistry for a long time because they can help us molecule-see! But, monoclonal, polyclonal, what’s there to see, let’s begin with a look at a cell named B!
note: video added 3/30/22
When an animal is confronted with something unusual, like a protein from an invading virus or bacterium, it can mount an immune response to destroy the invader, but only if it knows that the thing is unusual, and the way it knows this thing is “foreign” is because it binds to proteins called antibodies that are made by a type of immune cell called B cells, which make antibodies and put them on display (sticking out of their membrane) to “keep watch”
The way they “keep watch” is through their antigen recognition site (paratope). An antigen is just the sciencey name for the thing that an antibody recognizes. If an antigen has a part (epitope) that matches the antibody’s antigen binding site, it can bind and trigger an immune response. B cells that make antibodies that recognize things that are supposed to be there (self things) don’t make it past quality control so you don’t just constantly set it off. That is, if all goes according to plan… an antigen can be *anything* perceived to be “foreign” that triggers an immune response (even something “harmless” like a peanut or a protein your own cells made), whereas a pathogen is a disease-causing microorganism like a virus or a bacterium.
So, for example, the coronavirus Spike protein (the protein that juts out from the viral membrane and binds to receptors on our cells) can act as an antigen. An infected person’s body makes antibodies that can bind it. But those antibodies can bind to Spike at different places, and these different binding places are called epitopes. So different anti-Spike antibodies can bind different epitopes (and with different strengths).
But how to you get all these different antibodies made? The B-cells don’t know what they’re going to have to face in the future – they can’t predict that someone with the coronavirus or a cold is about to sneeze on you. And they don’t even know what potential options are out there. So they don’t know what to make – so they “experiment” in a learning process called adaptive immunity. They recombine pieces of DNA to make antibodies and see what works. They can do this “easily” because antibodies have a kinda modular structure – they have a “generic” adapter region (constant region) and variable regions, which are unique to different antibodies and are where the experimentation occurs
Before I get too far into it, a couple key points to keep in mind. The adaptive immune system (aka acquired immunity) is like natural-selection based evolution of antibodies on a really tiny scale happening *inside* your body – and staying in your body (or at least in your cells & bodily fluid) – it occurs in somatic cells (cells that just “make you”) NOT germline cells (cells that can be used to make children from you), so the changes made don’t get passed on – although through an active transport mechanism in the placenta, some maternal antibodies get transferred from pregnant mother to fetus before birth, so for the first 6 months or so, newborns are protected by their mom’s antibodies while they start making their own.
And how do they make their own? the “experimenters” are a type of immune cell called progenitor B-cells. They’re like “blank slates” – gene rearrangement occurs (they mix & match different genetic “options” for the antibody’s final form) & they start making a unique antibody receptor that sticks out from its membrane. Now it’s a mature B-cell. But it’s still “naive” – it hasn’t encountered its matching antigen.
If it does, it’ll differentiate into 2 kinds of B cells – effector B-cells (aka plasma cells) & memory B-cells. Effector B cells make lots of that antibody (each can make millions of the antibody molecule) &, instead of displaying them on their surface, they secrete them into the bloodstream for a wider-reaching response
The memory B cells display the antibody on the surface like the original naive cell, but now in the “permanent collection.” This “permanent” memory isn’t stored in our genome, and it isn’t passed down in our genes. Only that subset of cells knows it. Those memory B cells live in the Bone marrow (which is why bone marrow transplants “swap” someone’s immune memory) & can circulate between lymph nodes – they secrete a low level of antibodies to keep watch. So if the body encounters that invader again, it doesn’t have to search through billions of potential antibodies to find a match – it has one in the “permanent” collection, it just has to make more of it.
In researching for this post I learned something new (to me) – I had always thought that these circulating antibodies where how babies got immunity from their moms – and wondered how formula-fed infants coped. But it turns out that, although breast milk does contain antibodies, and breast-fed infants do ingest those antibodies, they can’t absorb them into the bloodstream, so they’re restricted to playing defense in the gut – so they can protect against things like diarrheal diseases, but can’t provide longer-term protection like the several-months-worth of antibodies coming from the pre-birth placental transfer. Interestingly, in lots of other mammals, antibodies from breast milk can get absorbed during a short time window after birth. But anyways – I hope this provides a little reassurance for mothers who don’t breastfeed, and now lets get back to the antibody-making story.
The place to watch is the variable region – throughout the experimentation, the “generic” adapter region (constant region) is kept constant so that it can allow any variable region to be displayed. But note – that generic adapter part’s only generic for the particular animal that made it (i.e. the adapter part’s slightly different in mice & rats). You can think of it kind of like different versions of an iPhone that can all charge with the same charger (until Apple decides to create a new “species”) – Apple experiments with changes to the iPhone that don’t interfere with its “iPhone-ness” and compatibility with iPhone stuff. But you can’t use a charger that you’d use for your Android phone. I’m guessing that most of the things they try never make it to market, but their experimenting’s not “random” like the experimenting of our immune system, so there’s a lot less “failure” but also less potential for “creativity”
Antibodies offer a lot of room for “creativity” because, even though antibodies are pretty small proteins, they still have multiple peptide chains – antibodies have 2 copies of a heavy chain (~440 amino acids (protein letters) long) & 2 copies of a light chain (~ half that length) & there are different choices for how to make the chains – at least the variable regions. Both the heavy & the light chains have variable regions & they both contribute to the antigen-binding site that forms when the chains link up through disulfide bonds (those strong bonds that can form between cystine amino acids).
The “experimentation” occurs through a process called V(D)J recombination, which rearranges Variable, Joining, and (in the heavy chain but not the light chain) Diversity gene “pieces.” It’s like at a restaurant where you can choose 1 appetizer, 1 entree, and 1 dessert. But after you choose, the other options get cut out of the menu so that every time you go back to that restaurant you have to order the same thing.
The recombination process involves stitching together pieces of DNA and cutting out in-between parts. The recombination process changes the DNA, but since these cells are somatic cells, not germline cells, they won’t get passed down to your progeny, and they only affect the cells that come from them. So the other cells in your body still have all the possible parts to recombine (full menu), but cells that come from that particular mature B-cell can only make that one type of antibody.
This whole recombination thing might remind you of alternative RNA splicing – this is where the messenger RNA (mRNA) copies of DNA genes can be edited in different ways to make different products. more here: http://bit.ly/2AsVdjG
That kind of recombination is at the RNA level, not the DNA level, so it can’t even be passed down inside your body – it’s every cell – and every recipe copy – for itself – if they want to coordinate, they have to rely on external cues like splicing factors that only get expressed at certain times & in certain tissues.
To make antibodies even more complicated, there are different types of constant regions. When a cell undergoes “class switching” it basically swaps the variable region onto a different constant region and this allows it to perform new functions, while retaining the ability to bind the thing of interest. It occurs through a mechanism called “Class Switch Recombination” (CSR) that’s similar to V(D)J recombination in that you’re cutting out chunks of DNA.
The classical antibody you usually see in textbooks, etc. is IgG (Ig stands for Immunoglobulin and it’s another name for antibody). IgG antibodies are the main Ig in blood and they’re the type we usually use in the lab for detecting proteins, etc.. They’re what many antibody tests look for to see if someone’s been infected with some pathogen.
But IgG antibodies are not the first on the scene. Instead, a person’s body first makes mostly IgM antibodies. These typically bind more weakly than IgG antibodies because they haven’t had time to undergo “somatic hypermutation” to further evolve from the initial “hit” and get even better at binding to the antigen. To compensate for weaker binding, IgM antibodies work as pentamers (5 copies stuck together) which can form through disulfide bonds between their their M-style constant region (and a “J chain” that’s a separate little peptide that helps connect them). This way, if an antibody bumps into an antigen and one of the copies binds, one of the neighboring copies might bind before the antibody falls off. And that makes it bind more strongly. Allowing time for another copy to bind, etc. terminology note: this phenomenon whereby you have multiple interactions kinda reinforcing each other is called avidity, and the strength of the interactions is called affinity.
IgG is the most common antibody type in the blood, but it’s not the “most produced.” Instead, IgA wins that award. IgA hangs out in mucosal areas like your intestines, lungs, and urinary tract. It’s also secreted in saliva, tears, and breast milk, so babies can get these from breast milk. They can act alone or in teams (typically duos).
IgE antibodies are involved in allergic reactions – they bind to the peanut protein or pollen protein or whatever the allergen is and then cause histamine to be released from a type of cell called a mast cell. That sets off a sneezy chain of events. IgE are also involved in protection against parasitic worms.
There’s also IgD but seems like people aren’t quite sure what they do yet – something to do with B cell development. https://bit.ly/3aROxg6
But IgG are the type we usually use in the lab, so let’s talk about how we can use and choose them in a laboratory setting. Since antibodies can recognize specific things they’re great for detecting those things – if you have an antibody that recognizes a protein (Waldo in my “Where’s Waldo” analogy), you can do things like:
- a Western blot, where you take a mixture of proteins (maybe taken from inside a cell (a cell lysate)) and separate them by size using an SDS-PAGE gel, then transfer them to a membrane through electroblotting and use the antibody as a probe to see if Waldo’s there and where
- co-immunoprecipation (co-IP), where you coat little beads (resin) with the antibody and use it to “pull down” the protein with its binding partners from a cell lysate to see what it’s bound to
- stain tissues using fluoresently-labeled antibodies to see where different proteins are located in their natural setting
So they’re super handy – but getting them can be a challenge… Remember B-cells “randomly” choose what to make by piecing together different options for different parts of the variable regions of antibodies through V(D)J recombination. And once they choose, there’s no going back since they remove the other choices. So different B-cells randomly specialize to make different antibodies. Those antibodies all, by chance, happen to recognize different things. These things can be different molecules all together (e.g. Antibody 1 recognizes Protein A and Antibody 2 recognizes Protein B) OR the things can be different parts of the same molecule (e.g. Antibody 1 & 2 both recognize Protein A, but different parts of it)
I tend to anthropomorphize things, but when I say “recognize” it’s not that the antibody can go – “oh hey that’s Waldo!” – all it means is that the antibody binds it – and when it does, it still doesn’t even know the identity of the thing it’s bound to, just that it likes to bind it. And it binds because of the same reasons any molecules bind – intermolecular forces (shape, charge distribution, etc. complement each other)
It’s not like a fingerprint, where it’s recognizing something totally unique, just something that happens to have certain qualities that happen to complement its own certain qualities – it’s not looking for a soulmate, just something that meets its minimal standards.
As a result, you can get cross-reactivity -> antibodies that “recognize” more than one thing if they “look” similar -> a promiscuous antibody that’ll bind Waldo OR Joe (whose ID they still don’t know).
And, since the part they’re “recognizing” is just a little part of Waldo or Joe, and different antibodies “recognize” different parts, they can get “confused” by different proteins. (remember, they’re not really “looking” for anything – they’ll just bind what fits)
Say, for instance, you’re interested in detecting Waldo. You have 1 antibody that recognizes Waldo’s hat, 1 that recognizes his striped shirt, and 1 that recognizes his shoes. The 1st antibody might “confuse” someone with a similar hat as being Waldo whereas the other 2 don’t care about the hat. But the 2nd could get “confused” by someone with a similar shirt and the 3rd by someone with similar shoes.
A mix of antibodies that recognize the same antigen, but different parts (epitopes) of that antigen is called polyclonal – they’re derived from cell lineages (clonal) made from many (poly) different parent B cells, whereas monoclonal is where all the antibodies recognize the same epitope because they come from a single (mono) parent B cell
In the body, it can be good to have a mix of antibodies that recognize something because you can then get multiple antibodies binding to different parts of it, giving you a stronger response. And, in the lab, polyclonal antibodies can give you a stronger signal, which is good if you have a small amount of protein. But, the more different antibodies are in a mixture, the more chances for cross-reactivity.
Polyclonal antibodies are easier to produce and therefore cheaper to buy – but you may pay the cost in cross-reactivity. Monoclonal antibodies are harder to produce & thus more expensive – but there’s less risk of cross-reactivity and you know you’re always getting the same product when you buy it.
Typically, antibodies are produced for lab use by injecting an animal with some protein or part of protein or whatever you want to detect. That thing goes to the animal’s spleen, where B cells that happened to have made an antibody that can bind it bind it, leading to those B cells being selected for.
The B cells start making more copies of themselves, some of which (memory B cells) continue to display the antibody sticking out from their membrane, while others (effector or plasma B cells) secrete the antibody into the serum.
Scientists can collect the secreted antibodies from the serum (the cell-less part of blood you get when you spin blood in a centrifuge to separate out the cells). What they initially collect contains antibodies against your antigen, but also antibodies against other things – what you want is only ~2-5% of the antibodies. So they have to purify them, which they can do by selecting only things that really do bind the antigen.
If you want a polyclonal antibody, you can stop here. But if you want a monoclonal antibody, there’s still more work to do – after purification, you have antibodies that recognize your antigen – BUT you have a MIX of antibodies that recognize different parts of the antigen (different epitopes)
If you want a single one, you need to go straight to the source – the B cell that’s making that single antibody, not just the antibody it’s making. If you can isolate that B cell, you can be assured that all the antibody that comes from it and its progeny (clonal line) is the same antibody because they have the same edited menu.
Speaking of the cell’s progeny, you’re going to need a lot of them – a single cell won’t give you nearly enough. Problem is, these cells don’t like to grow outside of the body. So they need some help. So, the conventional method has been that scientists fuse them with cells that DO like to grow outside of the body – and inside of the body – cancer cells. They often use a type of cancer cell called myeloma cells
B cells from the spleen of the immunized animal (good at making antibody but bad at growing in a dish) + myeloma cells (can’t make antibody but good at growing in a dish) = hybridoma cells (good at making antibody AND growing)
Now you have an “endless” source of a specific antibody. The antibody can be produced in cell culture, or the hybridoma cells can be injected into the peritoneal cavity (abdomen outside of the organs) of a mouse or rat, where they’ll grow and secrete into the abdomen fluid (ascites) that can then be harvested & usually gives you more antibody than in a dish. So every time you order it you get the exact same thing -> better reproducibility. But you’ll likely get lower signal because there’s only 1 place on each protein that can get bound.
This is how antibodies have been made (at least historically) for use in the lab. But monoclonal antibodies can also be used to help treat or even prevent infections, like the coronavirus. As I talk in detail about in a past post, scientists have isolated “neutralizing antibodies” from the blood of recovered patients and immunized mice that were genetically engineered to produce humanized antibodies. Neutralizing antibodies are antibodies that bind to a virus and prevent it from infecting other cells (e.g. by blocking the cell receptor binding spot on the Spike protein). Instead of using animals or hybridomas to make lots of these antibodies, they’re taking the DNA for making the antibody and sticking them into expression cells to churn out large amounts. https://bit.ly/antibodytherapy
For treatment or prophylaxis, you typically want to use monoclonal antibodies, or a cocktail or a couple of them, because that allows you to choose the “best ones” and you know what you’re gonna get with each batch. Note: with convalescent plasma you are getting a polyclonal mix (and you’re also getting antibodies against multiple proteins including proteins from random past infections).
Let’s talk more about polyclonal antibodies in lab use… Polyclonal antibodies are usually produced from bigger animals like rabbits, donkeys, sheep, and goats, which have more serum that antibodies can be purified from. But you don’t have an “endless supply” and each batch will be different.
Be it monoclonal or polyclonal you still have the problem that you’re not always looking for the same Waldo & a downside of antibodies is that you have to find ones specific for whatever “Waldo” you’re looking for. Because antibody A won’t recognize protein B & vice versa. But, since antibodies only recognize a specific part (epitope), they’ll recognize & bind that epitope no matter where it is – it can be on Waldo, or Joe, or even a cat. So you can stick Waldo’s hat on the protein of your choice when you make it and then sending in antibody detectives to look for the hat.
The way to stick the hat on is to add it onto the end of the protein by putting the genetic instructions for it before or after the instructions for our gene when we express proteins RECOMBINANTLY -> we stick a gene for a protein into an easy-to-work-w/circular piece of DNA called a PLASMID VECTOR (this is the recombining part) then stick that into cells (often harmless bacteria or insect cells) to make the protein for us (this is the expression part). The gene in the plasmid has the instructions for making the protein, so we can change those instructions (alter the sequence) to change the protein.So we can insert a Waldo-hat-making recipe in front of our protein recipe to “tag” our protein.
We often do this to help with purification through affinity chromatography. For example if you stick a “His tag” on the end (just 6-8 histidine amino acids), you can get it to selectively bind a Ni-coated column, wash all the other stuff off, and then compete off the protein using a His mimic called imidazole. more here: http://bit.ly/2RRkUUE
And – since you’re already putting a hat on it, might as well use that hat to help you elsewhere! Antibodies can be produced against the tag itself so you can use an antibody that recognizes the tag to detect any protein for that tag. There are antibodies readily available for purchase against the common tags used, like His, GST, & streptavadin.
When I’m doing affinity chromatography, I’m doing it to purify the protein & remove all the stuff that’s stuck to it. But you can also use antibody-bound beads & milder conditions to “pull out” the protein WITH things it’s bound to to see *what* it’s bound to – this is called co-immunoprecipation (co-IP). And when you do that, instead of tons of lysate & do it in a column & lots of resin you use a little lysate, little resin, & do it in a test tube
if you want to make your own paper models: https://pdb101.rcsb.org/learn/paper-models/antibody
we ended up making GFP at camp, which is much easier! and if you want to learn about the GFP: https://bit.ly/gfpfunscience
Immunology’s not my field of expertise, but if you want more info, I found this great blog post: https://epomedicine.com/medical-students/vdj-somatic-recombination-made-easy/
more on topics mentioned (& others) #365DaysOfScience All (with topics listed) 👉 http://bit.ly/2OllAB0