POR_genetic_analysis

Summary:

Turns out that I am also a carrier of the R457H variant that causes Antly-Bixler syndrome (ABS) in people that get two defective copies.

Common low-cost genotyping technologies:

So it looks like the genetic test you took is registered with the government here (https://www.ncbi.nlm.nih.gov/gtr/tests/528277/overview/). There isn’t anything particularly interesting there; I just think it’s notable that they have (voluntarily) become listed on this website, which keeps track of clinical genetic tests. On the other hand, 23andme is not listed on there, likely because they are direct-to-consumer and don’t try to go through the same processes (I’m assuming regulations) to really be considered a clinical genetic test.

The specific genetic test you took (CarrierMapGEN test from “Cooper Genomics” / “Recombine”) is very similar to the technology used by 23andme. Looking at the website, it says “CarrierMapGEN (genotyping) is built on Illumina’s proven genotyping platform”. Illumina is a company that specialized in offering a couple types of DNA sequencing / genotyping technologies; we actually use a different one of their products a lot in the our department / lab). The genotyping platform used by Cooper Genomics is apparently this chip (https://www.illumina.com/products/by-type/microarray-kits/infinium-iselect-custom-genotyping.html). 23andme also uses Illumina microarrays chips for genotyping. This youtube video here is pretty informative in describing how these chips work: (https://www.youtube.com/watch?v=lVG04dAAyvY). In short, there are certain positions in the genome where some small-to-moderate subset of the population has a different nucleotide than most of the rest of the population. Knowing this, these chips were printed to look at these nucleotide differences at tens of thousands of these positions. It’s not true DNA sequencing since it’s only looking at tens of thousands of possible nucleotide differences, rather than the 6 billion nucleotides that exist in each human cell, but it’s a more practical (fast and cost-effective) way to look at some major sites of genetic differences. Thus, the technology used by both services is likely very similar, though each company may customize / change exactly which DNA changes they’re looking for on their own versions of the chip.

The reason I bring this up, is that the single nucleotide polymorphism (change) you tested positive for is actually captured by 23andme; it’s just that they don’t report it, likely due to federal regulations / hoops they haven’t fully jumped through. The report from Cooper Genetics tells you the ABS carrier variation is in the POR gene, with the nucleotide at positions 1370, which is normally a G (guanine) instead being an A (adenine). This nucleotide change ends up changing the sequence of the protein encoded by the POR gene, where positions 457, which is normally an arginine, is now a histidine. In short, this variation is referred to as “R457H” at the protein level. These single nucleotide polymorphisms that exist at small-to-moderate subsets of the population (and thus are built into the genotyping chips supplied by Illumina) are listed in a government database, called dbSNP, and given a unique identifier name / number. In the case of this POR R457H variation, it’s listed here: (https://www.ncbi.nlm.nih.gov/SNP/snp_ref.cgi?rs=rs28931608), and the associated unique identifier is “rs28931608”.

Importantly, if I (or you, for that matter) go to the 23andme webpage, hover over “tools” in the top menubar, and click on “Browse raw data” on the subsequent drop-down menu, it brings you to a page where you can enter the specific SNP id (so “rs28931608” in this case). It gave me a result that looked like this:

My results upon entering rs28931608 into 23andme

As the above image shows, it lists my genotype as being “A / G”, showing the nucleotide I have at this position, on each of my different copies of chromosome 7. One is the “G” found in the general population, encoding arginine at position 457 in that copy of the POR gene. On the other hand, the other version chromosome 7 I have encodes an “A”, resulting in histidine at position 457 in that second copy of POR gene. Thus, I seem to be a carrier as well. Based on my reseearch, it looks like this specific variant (R457H) is very rare, but still sporadically observed in some people of Japanese ancestry, suggesting that we likely inhereted it from mom. Dad’s 23andme result can likely shed some light on this.

So what is encoded by the POR gene, and what does this genetic change do?

While there are numerous useful resources online, I found this review article to be rather informative, so I uploaded here for you to see it without having to hit a paywall: (http://visualizedlife.com/wp-content/uploads/2017/10/Miller-et-al-2011-PMID-21070833.pdf). The POR gene encodes a moderately sized (not too big, but certainly not small) protein, which resides in the endoplasmic reticulum within just about every one of the cells of the human body. There are a certain group of important enzymes called cytochrome p450 enzymes, which are important for the synthesis and breakdown of a bunch of important small molecules, such as hormones (including estrogen and testosterone synthesis and metabolism), cholesterol synthesis, and vitamin D. Notably, these enzymes have to be powered to do their jobs, and the protein product of the POR gene, called cytochrome P450 Reductase or P450 oxidoreductase, is the protein responsible for powering these cytochrome P450 enzymes. Apparently, it uses NADPH, which is a molecule used to store energy for certain chemical reactions. POR helps to take some electrons from NADPH and pass them to an FAD (flavin adenine dinucleotide) molecule, and then onto an FMN (flavin mononucleotide) molecule, all the while changing shape to interact with the nearby cytochrome P450 enzymes. It then presses up against the nearby P450 enzymes and passes the electrons to the iron atom at the center of a heme molecule attached to the P450 enzymes, which is then used to perform its actual enzymatic function.

From Miller et al; A cartoon depiction of the electron relay between POR and P450 enzymes

So where does the R457H mutation fit into all of this? Since it’s a pretty important protein, and it behaves pretty well when expressed and manipulated in the lab, some scientists have been able to determine how the protein folds, and a model of this protein can be found at this public database of molecular structures (https://www.rcsb.org/pdb/explore.do?structureId=3QE2). Below, I’m using software called Pymol to look at this structural model of POR, zooming in around R457. I’ve colored the POR protein green, while the FAD molecule is colored magenta, blue, orange, and red. Notably, in this image, I’m showing the side chain of R457 (colored green, with blue atoms of nitrogen). The yellow dotted lines are denoting that R457 physically contacts with the phosphate groups on the FAD molecules, and helps it bind to the protein. When mutated to a histidine, FAD now can’t make these same contacts, preventing FAD from binding to the protein in the way it needs to for doing it’s job transporting electrons, leaving the nearby P450 proteins unpowered and non-functional.

Image of R457 contacting FAD

People have run various biochemical experiments on cells with the R457H version of POR, and found that many of the nearby P450 atoms were now nonfunctional. Thus, because R457 is such an important amino acid for POR protein function, mutating it into a histidine makes the enzyme no longer work. And since POR is so important for how a whole class of proteins responsible for many drug / metabolite / hormone metabolisms, this variant shuts down the creation or breakdown for many hormones / small molecules.

Table of variant activities

So what’s actually going on in my own cells? Amazingly, all of my cells are going to have this dead enzyme just floating around the ER, doing absolutely nothing. But, the beauty of having two copies of POR – one on chromosome 7 from mom and one on chromosome 7 from dad – means that there are just as many good / functional versions of the POR protein floating around the ER doing its job, as there are dead versions doing nothing. So even though roughly only half of the POR protein in each cell is doing its job, that’s still enough to power the nearby cytochrome P450 enzymes and allow me to live a normal life. Thus, this an autosomal recessive disease, allowing me to merely be a carrier of his nonfunctional genetic copy of POR, without giving me any observable symptoms.

Only individuals with two bad copies will have too little POR activity to shut the cytochrome P450 enzymes down. Since they aren’t able to make a bunch of different hormones, such as testosterone, the cells within their bodies can’t communicate correctly with each other as a baby (with these two bad copies) is growing during pregnancy. This causes a bunch of different problems, which altogether contribute to the affected baby dying early in its life. Thus, while it’s fine to have some of the POR protein for your cells and body to work correctly, catastrophic problems occur when none of the POR protein in your cells and body work correctly.

Estimating the liklihood of a carrier having a child with ABS

So since I have one normal copy and one R457H copy, any child I have will have a 50% chance of getting the R457H loss-of-function gene from me. So if we had no information about the mother of this possible child, what’s an estimate of the child getting a maternal copy of a loss-of-function POR allele. Looking at this database of POR variations observed in genomes and exomes from largely healthy people (without ABS at least; http://gnomad.broadinstitute.org/gene/ENSG00000127948), almost all of the missense or loss-of-function alleles are rare variants. A503V is relatively common, but the same review article above shows that this variant is partially active; we can assume it functions normally enough for the purpose of this estimation.

This website (http://www.orpha.net/consor/cgi-bin/OC_Exp.php?Expert=83) states that the prevalence of ABS worldwide is < 1 in 1,000,000. We’ll consider this a rough estimate, but if that’s the case and with ABS being autosomal recessive, the estimated frequency of a pathogenic POR variant allele in the population is:

## [1] "Pathogenic POR variant allele frequency: 0.001 or 1 in 1000"

## [1] "Since there's a 50% chance of a child getting the R457H inactive POR variant from me, the chance of a child being affected can be estimates as: 5e-04 or 1 in 2000"

Thus, while it’s thought to be a less than a 1 in a million chance for two random people in the population to have a child with ABS, there’s a small but much more realistic chance for me to have a child with ABS, due to the fact that I already carry an inactive copy of the POR gene. Since it’s such an impactful disease, I think it will be worth it for us to have Anna tested, to help figure how just how cautious we’ll have to be about having a child with ABS.