How does one isolate a specific antibody from the blood? Presumably there are thousands (millions?) of different ones in there for all kinds of pathogens.
Typically, you first test the blood to see if there are any antibodies in that blood of interest. E.g., collect blood sample, harvest serum from blood by coagulation and centrifugation, apply serum to surface coated with COVID protein(s), wash away unbound material, detect whether antibodies to serum have bound to surface coated with COVID proteins. If yes, then this is a blood sample of interest. You could also test for neutralizing antibodies in parallel/before/after/instead in a viral replication assay. Once you have identified the blood as containing what you want, you then harvest more and collect mature B cells. Mature B cells each produce one antibody. You isolate the B cells by limiting dilution and/or immortalize them by fusion with a special type of cell to make a hybridoma to isolate clonal populations of single cells that are maintainable. Then you test each clonal population for whether it produces the antibody of interest, isolate the nucleic acid that encodes that antibody and put it into a cell type suitable for manufacturing. There are other ways to do it, but the above is fairly standard. It is laborious and easy to mess up and takes time (1 to 2 months) to do correctly. Any step can go wrong and require starting over.
> You isolate the B cells by limiting dilution and/or immortalize them by fusion with a special type of cell to make a hybridoma to isolate clonal populations of single cells that are maintainable. Then you test each clonal population for whether it produces the antibody of interest
Do people actually still do it this way? I would have thought that you would use more targeted approaches where you use an antigen to fish for cells of interest before culturing. There are a few such techniques described here:
Naively, i would have thought you could do something with affinity columns or magnetic beads, too - coat beads with antigen, then use those to extract B cells expressing a matching surface immunoglobulin.
Magnetic bead filteration is commonly used to filter immune cells, but they are expensive. I used to work for a lab that did this on a daily basis. Common problems with it were getting enough cells to run the assays, and the process of extraction is damaging to the cells. So you need a lot of starting material and then after that they don't stay alive for very long. You can treat and freeze after isolation but thawing only decreases the amount of healthy cells.
Here’s a noob question. The coronavirus is made out of a whole bunch of proteins. Have we fully mapped out all the protein structures and corresponding DNA code?
The antigen is also a protein, I assume the DNA sequence for it is well known. Right?
How far are we in terms of tech to print custom proteins from arbitrary DNA sequences?
Is understanding protein folding and protein to protein interaction the holy grail of making massive improvements in molecular biology? What are the big unsolved problems?
Like if we know the virus’s DNA and it’s 3D protein architecture, we can solve for antigen proteins in a computer that outputs possible DNA sequences and we can manufacture them the next day in a protein printer. How far away are we to that future?
Protein folding is part of it. The other is finding which parts are antigenic to the immune cells. Epitope mapping is a common method to screen small bits of the proteins to see if any are hits for immune cells to recognize and kill. There are algorithms that can take the RNA/DNA, predict proteins, and then guess a percentage of those that may be important. But you still need to synthesize those in mass quantities and start testing each. Once you have candidates you then test them in mouse models (typically) to see if they actually provide an immune response. If interested check out the iedb. https://www.iedb.org/
"Have we fully mapped out all the protein structures and corresponding DNA code [of the COVID-19 virus]?"
Yes
"The antigen is also a protein, I assume the DNA sequence for it is well known. Right?"
Yes
"How far are we in terms of tech to print custom proteins from arbitrary DNA sequences?"
Generally that is something a first year graduate student can accomplish.
"Is understanding protein folding and protein to protein interaction the holy grail of making massive improvements in molecular biology? What are the big unsolved problems?"
There are too many unsolved problems to count. There have been great advances lately in de novo prediction of protein folding and to a lesser extent protein:protein interactions. But even if you had perfect knowledge of all that, you still can't just like design the perfect vaccine.
"Like if we know the virus’s DNA and it’s 3D protein architecture, we can solve for antigen proteins in a computer that outputs possible DNA sequences and we can manufacture them the next day in a protein printer. How far away are we to that future?"
We (the world) accomplished that within a couple of weeks of identifying the COVID virus.
> It is laborious and easy to mess up and takes time (1 to 2 months) to do correctly.
Sounds like a manual software testing. What are the chances of automating entire process? Whenever I see bio/chemists working it seems very manual job. I assume someone already tried it, but perhaps only for specific area rather than making universal robot?
Complete automation would be tricky. We are dealing with a lot of starting and intermediate materials that need to be precisely incubated under demanding and entirely sterile conditions. You need incubations at liquid nitrogen temperatures (for storing cell lines), -80 C, -20 C, 4 C, room temp, 30 C, 37 C, 37 C 5% CO2, etc. THen you need a way to go to and from each of these environments and a way of sterilizing in between steps. I'm not saying it's impossible, but it seems difficult.
IIRC the virus was isolated last year, so wouldn't at least some antibodies have been identified by now? Or, it's just very common to have to start over?
Is it unusually difficult for this coronavirus? eg I've heard it's unusually large.
Antibodies have been identified by now. They are being tested as we speak. Coronavirus is not more difficult than any other target really. It just takes time.
They're far too large and complex to be synthesized from scratch, that is simply not possible.
They are produced in various biological systems, with nature doing the synthesis. As far as I understand you can scale that up reasonably well with some effort.
To add to this, you don't even need to know what the antibody looks like, or even if it exists. Very simplified example: It's possible to grow a dish of cancer cells (because they divide quickly and are immortal, e.g HeLa cells), "purify" away everything that isn't protein (cell lysis), then run the proteins through a gel that separates them by size (Western blot). Comparing these to a control will show you which proteins enabled survival, which gives you your candidates for sequencing and further tests.
However, right now we have humans synthesizing large amounts of antibodies that are proven to work (because the humans creating them survived and cleared the virus). It may ultimately be faster to isolate antibodies from the serum and engineer cell cultures (through adding DNA to them, "recombinant DNA") to create more of those antibodies, resulting in much stronger synthesis of a single antibody (so-called "monoclonal antibody drugs").
It's nearly certain that both of these are happening many times over around the world right now. All of the science here was already in lab use the first time I worked in a bio lab in 2003, and nowadays we have methods that didn't exist then (such as CRISPR for DNA manipulation and fast sequencing of both nucleic acids and proteins).
The reason it's not researched much is because giving someone an antibody will not confer immunity - it will only treat the immediate virus.
It has value, but there are usually better treatments available that more broadly fight a variety of viruses and don't need to be so specifically customized as an antibody.
I suppose testing would be easier for a cure (compared to a vaccine)? Given that only the "adapter" needs to be specialized, I was wondering why there don't seem to be any approaches based on antibody mass-production....