The story of a first-hand discovery

DNA polymerase as a regulator of immunity

Peter Starokadomsky, "Biomolecule"

Recently, I was lucky enough to participate in an international study, the results of which were published on the pages of Nature Immunology. We were able to discover a new mutation leading to a very rare immunological syndrome; we managed to understand the molecular basis of this syndrome, and also discovered a new type of biological molecules regulating intracellular immunity. It was an unforgettable experience, which I want to share with a wide audience from fresh memory.

Background

Back in 2006, in Kiev, I solemnly framed the diploma of Candidate of Biological Sciences and began to think about how to live on. There were few options at that time – I really wanted to do immunology, and this field of science requires expensive equipment and qualified colleagues. Therefore, at that time I could only find a good laboratory abroad. By the way, at the same time I joined a newly created young site – Biomolecule.ru (if I'm not mistaken, my account goes under number 35).

So, I received an invitation to work as a post-doc at the Weizmann Institute (Israel), and everything started spinning. I was lucky: in Israel I got into a renowned laboratory and was able to master the most modern methods, gradually gaining a good specialization in the field of molecular cell immunology. And after three years in the promised land, I got to the USA, where I received a second post-doc in one of the departments of the University of Texas (UT Southwestern Medical Center). I was lucky again: the university is engaged in medical research, at the same time being an operating hospital. This allowed me to gradually combine my immunological studies with the problems of specific patients. And although I am not a doctor myself and do not have access to patients, I work as part of a large team of specialists, including doctors, geneticists, bioinformatics, etc. My duties are to investigate materials obtained from patients using molecular methods: as a rule, these are blood cells or skin fibroblast cultures. One of these projects is described below, which unexpectedly turned into something much bigger for all of us.

XLPDR – a syndrome about which nothing is known

Once geneticists came to us for advice. They have been struggling for a long time to identify the gene that causes a very rare syndrome, which does not even have a name in Russian - X–linked reticulate pigmentary disorder (XLPDR). Today, only about a dozen families around the world are known to have cases of this disease. Only boys are ill (the disease is classically linked to the X chromosome), but girls are carriers without any particularly noticeable manifestations of the disease. The syndrome is characterized by spotty (similar to bright freckles) and dry skin all over the body (most patients have chronic anhidrosis – there are no sweat glands from birth) and a very specific face with hair directed backwards (Fig. 1). However, the most unpleasant thing is that in addition to phenotypic signs, these patients have constant inflammatory processes in the lungs, kidneys and the gastrointestinal tract. Nothing about the nature and mechanisms of this rare disease was known at the time of the start of work.

Figure 1. Patients with XLPDR syndrome. Photos from websites edition.cnn.com and xlpdr.com .

...So, geneticists have asked us to study the molecular mechanism of XLPDR syndrome. The traditional rapid genome sequencing (sequence sequencing of exons only, exome sequencing) did not bring results, and geneticists could not find a mutation using relatively cheap methods. Therefore, we were asked to look at the problem from the point of view of molecular immunology – we had just published a similar type of article describing molecular changes in intracellular signaling pathways in patients with another rare syndrome - XLID (hereditary dementia, also linked to the X chromosome and also manifested only in boys) [1]. That is, by that time we already had experience in immunological studies of real patients "from scratch". In addition, just two patients with XLPDR were observed in our hospital. In general, we willingly got involved in a new project.

Hope for quick success

At first, the matter was disputed: all cell lines obtained from patients (skin fibroblasts and blood cells) showed orders of magnitude higher expression of the entire palette of immunological markers-genes. Both NF-kB and interferon-dependent genes were expressed over the edge for no apparent reason. It seemed that some important regulatory gene had lost its functions in patients, and the cells constantly felt the invasion of insidious viruses, even when they grew up in sterile incubator conditions. Having documented all the artifacts, we began to wait for geneticists to find out the name of the gene that would contain the mutation in all patients.

We didn't have to wait long – since the syndrome is rare, and the genomic DNA of all known patients was on our hands, geneticists conducted a complete genome sequencing of all sick children (as already mentioned, the syndrome affects only boys). And suddenly it turned out that all patients have the same point substitution (SNP) in the thirteenth intron of the POLA1 gene – the catalytic subunit of DNA polymerase alpha*.

* – The article "Traces of polymerase α" tells very clearly about the normal – that is, rather negligent – work of this enzyme and the mechanisms of fixing its errors in the genome [2]. – Ed

This polymerase is known to all biologists from the university as an enzyme that synthesizes Okazaki fragments during genome replication. The mutation found in the intron of the POLA1 gene leads to pre-mRNA misplacing and provokes a shift in the reading frame (Fig. 2). As a result, during translation, ribosomes fly into a sudden stop codon, translation stops, and the whole complex degrades. As a result, patients' cells contain very little correct mRNA (only 10-20% of the norm) and, accordingly, very little polymerase itself. This is strange, because the lack of polymerase should have been fatal (as it turned out later, people survive only with this mutation, and other mutations in the same gene are most likely lethal). However, surprisingly, the patients did not have any special growth or development defects, and therefore we simply took this fact into account and began further searches – there were many more surprises ahead of us.

Figure 2. Diagram of the POLA1 gene and mRNA in healthy people (a) and in XLPDR patients (b). As a result of point replacement in intron 13 (asterisk), a false alternative splicing site is formed, and most of the pre-mRNA forms non-viable mRNA, which degrades rapidly. A smaller part of the mRNA is still assembled correctly, which ensures cell survival (replication), but this amount of POLA1 is not enough to perform alternative functions (see below).

The first riddle

The first hypothesis – that patients get sick due to the inability of lymphocytes to multiply rapidly and suppress infection during infection - was refuted by a simple blood test. We found that the mutation does not affect the number of immune cells in the blood of patients, which means they divide normally. Therefore, a metabolic disorder occurs inside the cells themselves, and this in some unknown way provokes an immune response.

And it's true – after examining the patients' cells, we found that polymerase deficiency for some reason leads to chronic activation of many signaling pathways. In addition, if the amount of POLA1 in normal cells is reduced with the help of a specific siRNA [3], inflammation also begins in them. All the data indicated that POLA1 polymerase somehow keeps cells from a constant inflammatory response, that is, it serves as a negative regulator of immunity. It remained only to understand how this could be.

Molecular routine

Then there were endless experiments using all the equipment that we could reach. Using microscopy, we found that a significant part of the POLA1 protein is located in the cytoplasm, which, again, is quite unusual for DNA-dependent DNA polymerase. Then we (quite justifiably) began to look for the points of contact of POLA1 with other known cytoplasmic regulators of the immune response. There are a number of immune sensors and signaling pathways in cells, and most of these proteins interact directly with each other when activated. We have even developed a special method that simplifies the search for such interactions [4]. However, in the case of POLA1, nothing worked – the polymerase was not directly part of any known signaling pathway.

The next maneuver was an attempt to determine which signaling pathway is regulated by POLA1 polymerase. With the help of kiRNA, we reduced the concentration of POLA1 in cultures of various normal cells to the level of patients (this is called gene knockdown) and then checked how sensitive the cells became to a particular immune stimulus. I must say, human cells so masterfully recognize any infection with the help of more than a dozen different receptors, which significantly confuses thoughts and complicates work. Therefore, we consistently treated the cells with all imaginable ligands for toll-like receptors* (TLR; infection receptors, of which a person has a dozen, and everyone "feels" the presence of a certain type of bacteria or fungi).

* – A selection of articles about these molecules: "Toll-like receptors: from Charles Janeway's Revolutionary Idea to the 2011 Nobel Prize" [5], "Immunological Nobel Prize (2011)" [6], "Finding the CpG motif, or the Fine Work of toll-like receptor 9" [7]. – Ed

We also stimulated cells with interleukin-1 and TNF (universal inflammatory factors that cause a person to have a fever and painful fatigue), and also simulated viral infection (which activates intracellular DNA and RNA sensors, for example, RIG-I, MAVS, STING). It was logical to assume that if a certain signaling pathway is associated with POLA1, a cell with an artificial POLA1 knockdown should become hypersensitive only to a certain ligand, and the effect of other stimuli should not differ much from the control. However, a surprise awaited us again: when POLA1 was knocked down, most cells acquired hypersensitivity to all stimuli at once. It turned out like hell: none of us had heard of the existence of such global regulators of the entire immunity, and most importantly, we still did not understand the mechanism.

Here we had serious suspicions – suddenly we missed some obvious factor that causes such an artifact? In order to exclude possible influences on the experiment of our biofields, barabashkas and other otherworldly forces, we turned to several laboratories with a request to repeat key experiments. But all the results were confirmed regardless of the phase of the moon and the location of the laboratories. It became clear that we had caught something more than just a mutation.

A Path in Darkness

An accident helped us to get closer to the disclosure of the mechanism. In the experiments with kiRNA described above, we noticed that the knockdown of POLA1 in one of the most popular laboratory cell lines – kidney carcinoma, NEC293 – did not cause the expected inflammation. At that point, we just noted it and switched to other lines. However, now we have remembered why NEK293 cells are one of the most popular laboratory cell lines: they are very easily transfected by any DNA. And this happens in particular due to the fact that in NEK293, due to a number of mutations, there are almost completely no internal antiviral sensors that recognize viral RNA and DNA molecules that have penetrated into the cell. The fact that this particular cell line turned out to be the only immune system to changes in the amount of POLA1 gave us a decisive clue: the observed effect must somehow be related to the concentration of nucleic acids in the cytoplasm and the ability of the cell to distinguish nucleic acids in the cytoplasm on the principle of "friend-foe".

It is known that cells very carefully control the amount of nucleic acids in the cytoplasm. During infection, viruses inject their DNA (or RNA) into cells, and cell sensors recognize foreign nucleic acids by a number of features. When foreign DNA/RNA is detected, interferon protection is activated – the basis of the antiviral response. This process involves the arrest of metabolism (so that viral particles do not have time to gather), the secretion of interferons (notifying neighboring cells and the immune system of infection) and the degradation inside cells of anything that looks suspicious (general cleaning, let's call it that). And what we saw in the cell lines of patients with our syndrome was very similar to the antiviral response – with only one difference: the cells were in sterile conditions and were not infected with viruses.

Now, if we put together everything we knew about the syndrome, the picture was filled with a new meaning:

1. POLA1 synthesizes DNA (in the form of Okazaki fragments), while using a non-standard RNA primer for synthesis.
2. POLA1 is found in large quantities in the cytoplasm of the cell.
3. With POLA1 deficiency, cells behave as if they are infected with viruses.
4. Cells that are not sensitive to POLA1 deficiency (this is the NEC293 line) do not have active registration systems for viral RNA and DNA, as well as RNA:DNA chimeras (which are formed during reverse transcription of many viruses).

In general, everything indicated that POLA1 somehow reduces the amount of viral or proprietary (for example, retrotransposon) RNA or DNA in the cytoplasm. And if POLA1 is absent, then "junk" nucleic acids accumulate in the cytoplasm and chronically activate antiviral signaling pathways, which leads to constant inflammation, which we see in patients.

This is all logical, but I still wanted to understand the mechanism of POLA1. It cannot directly destroy junk DNA (as a number of enzymes do, for example, TREX1 [8]), because POLA1 is one of the few polymerases with an inactive exonuclease domain. In addition, we had to learn how to determine the amount of nucleic acids in the cytoplasm, since existing methods – microscopy or flow cytometry using specific antibodies to DNA – worked extremely poorly on our models, because we were only interested in the cytoplasm, and the nucleus full of nucleic acids invariably gave a very high background signal.

A tale of lost time

In parallel, another problem arose: due to the "hot" information (after all, at that time the most important thing we discovered was a mutation in the POLA1 gene) and a large number of co–authors, we were constantly being rushed to publish what we have - otherwise any other geneticist could accidentally stumble upon an interesting mutation and publish an article, thereby by leveling the value of our research.

We began to feel a lot of pressure – they say, let's not chase the top magazines, we will publish what is in an average genetic publication, and later we will calmly find the mechanism and publish it again, also in a good molecular journal. We (my supervisor and I) had to show remarkable political qualities, convincing the collaborators that we were almost done and that there was no need to worry, we also have a parabellum... However, this did not raise the mood – we really did not know which side to approach the mechanism from, and time was running out. Each new version of the test required two or three months of hard work, and we had a maximum of two or three attempts left until the collaborators lost their last patience.

Green light

Solitaire developed somehow suddenly. We already had at hand a complete set of antibodies to nucleic acids: to double- and single-stranded DNA, as well as to RNA:DNA-hybrid molecules. The antibodies worked well, but we couldn't come up with a normal registration method. And suddenly it dawned on us: if the cytoplasmic fraction of cells is separated from the nuclei (this is easy), then it is cytoplasmic DNA or DNA that can be purified with the help of antibodies:RNA. And if antibodies are attached to sepharose beads (Protein-G agarose beads is a reagent that is traditionally used in laboratories for affinity chromatography), then these beads can simply be colored with a specific dye for nucleic acids (with the invention of qPCR, a lot of such dyes are sold, and we chose PicoGreen because it is green and beautiful), and then compare the glow levels under a conventional microscope with an ultraviolet lamp.

So, if there are a lot of DNA or RNA molecules in the cytoplasm:DNA, the balls will glow bright green, and if not enough, then pale green. Having quickly tested the technique, we began to compare the cytoplasm in various healthy cells and in the cells of our patients with XLRPD syndrome.

The first experiment amazed us: healthy cells had a lot of cytoplasmic RNA in their cytoplasm:DNA. In the cytoplasm of the cells of our patients, RNA:There were practically no DNA molecules at all (Fig. 3)! Exactly the opposite result to what we expected, but there was no stopping us.

Figure 3. Typical result of cytoplasmic RNA analysis:DNA in various cell models. a – Individual sorbent grains with immobilized RNAs:DNA antibodies (RNA:DNA) or non-specific control antibodies (IgG). Three types of cells are compared – HEK293 and HeLa with POLA1 knockdown, as well as fibroblasts of healthy people and patients with XLPDR. The numbers under the pictures are relative fluorescence. b – Sorbent grains with a smaller increase show a comparative amount of RNA:DNA in the cytoplasm of XLPDR patients and in the same cells after adding a healthy copy of the POLA1 gene to the genome. The numbers are relative fluorescence: as you can see, the difference is more than tenfold.

Making sure that the main difference between a healthy cell and a cell with a POLA1 deficiency is a mysterious RNA:DNA in the cytoplasm, we once again belatedly slapped ourselves on the forehead. After all, what the university course in molecular biology says: the Okazaki fragment – the main product of POLA1 – is a small RNA seed to which the polymerase adds DNA nucleotides. It turns out that POLA1 synthesizes similar RNAs in its free time from replication:DNA complexes in the cytoplasm.

We tried to purify these cytoplasmic chimeras and transfect them into the cells of patients with XLPDR syndrome. Indeed, after that, the inflammation disappeared for a while. This means that POLA1 constantly synthesizes some immunologically inert RNAs:DNA complexes (about 100 nucleotide pairs in size) that serve as a kind of "white noise", preventing cytoplasmic sensors from being overexcited due to minor leaks of nucleic acids, for example, from the nucleus or when mitochondria are damaged. And only when the concentration of suspicious DNA exceeds the noise level, the entire antivirus system starts. Apparently, it turned out that the task of POLA1 was to constantly produce "nucleic noise" in order to prevent the false triggering of the interferon response.

Publication, in kind

We did it! Usually in fairy tales and in teen films, the credits begin at this place, and the characters go into the distance, into the sunset. Alas, in reality it does not work so elegantly.

Firstly, most were skeptical about publishing in a top magazine – the competition there is huge, and any even very good article can get stuck at the review stage for many months. Secondly, it takes about a month to write and issue an article for submission to each journal. And, of course, we had to write ourselves – despite the abundance of co-authors, only a few provided real help.

We started by seniority: August 2015 – Science. The answer in two weeks is no. September 2015 – the rewritten article went to Nature. The answer in a week and a half is no. October 2015 – Nature Immunology, the article is rewritten again and the drawings are redone. And suddenly we were lucky – the editor's answer came almost immediately: "I like it, but let's see what the reviewers say." A week later, all three reviewers answered the same way – we will take it. And then events developed fantastically quickly: it took us only a month to respond to the main comments – almost all the comments were on the case, and most of the material we already had on hand (we did not include it in the article due to space savings, because the volume of text and drawings is strictly regulated). The revised version was sent to the reviewers, and we received a positive response the very next day. After three years of work on the project, it was truly a royal award [9]!

In your opinion, is this the final?

Of course, our idea is about RNA:DNA duplexes in the role of "nucleic noise" are still only a working hypothesis. We have not yet been able to determine the sequence of nucleotides in these duplexes, so we do not represent their source and targets inside the cells. The idea of "nucleic noise", although elegant, has also not yet been proven. And in general, there are more questions than answers at the moment. But one of the results of this work is already possible to be proud of: we can now genotype the sisters of sick children to determine whether they are carriers of the mutation or may not be afraid to have children in the future. And we were finally able to tell ten patients why they are sick and what is wrong with them, and now we can tell their attending doctors in which direction their metabolism should be adjusted. Because it's terrible to know that you have a rare disease, about which only one name is known. And even that – still not translated into Russian.

Literature

1. Starokadomskyy P., Gluck N., Li H., Chen B., Wallis M., Maine G.N. et al. (2013). CCDC22 deficiency in humans blunts activation of pro-inflammatory NF-κB signaling. J. Clin. Invest. 123, 2244–2256;
2. Biomolecule: "Traces of polymerase α";
3. biomolecule: "About all RNAs in the world, large and small";
4. Starokadomskyy P. and Burstein E. (2014). Bimolecular affinity purification – a variation of TAP with multiple applications. Methods Mol. Biol. 1177, 193–209;
5. Biomolecule: "Toll-like receptors: from Charles Janeway's Revolutionary Idea to the 2011 Nobel Prize";
6. Biomolecule: "Immunological Nobel Prize (2011)";
7. Biomolecule: "Finding a CpG motif, or the fine work of a toll-like receptor 9";
8. Stetson D., Ko J., Heidmann T., Medzhitov R. (2008). Trex1 prevents cell-intrinsic initiation of autoimmunity. Cell. 134, 587–598;
9. Starokadomskyy P., Gemelli T., Rios J., Xing C., Wang R.C., Li H. et al. (2016). DNA polymerase-α regulates type I interferon activation through cytosolic RNA:DNA synthesis. Nat. Immunol. doi: 10.1038/ni.3409.

Portal "Eternal youth" http://vechnayamolodost.ru  29.03.2016