Can a Single Nucleotide Mutation Significantly Change a Protein's Function?

[jaumzaum]

Hi. I'm in the first year of Medicine and I'm studying Genetics and Evolution. I have this doubt in the back of my mind and I'm not being able to move forward without someone explaining me what's wrong with my thinking.

So I've learned an allele is a variant form of a gene, all alleles of the same gene are slightly different from each other and code the same characteristic in different forms. We know proteins can have from 20 to 34350 amino acids. Let's take a protein with 4 thousand amino acids. The RNA that formed it had 4 thousand codons, which are 12 thousand nucleotides. I didn't study mutation yet but I assume a mutation can occur in a single nucleotide in the DNA that can or cannot code for a different protein (as some amino acids are coded by more than one codon).

Ok, so let's say a super energetic UV photon reaches one of my germinative cells and mutates a single nucleotide of my DNA that changes the provious amino acid. The protein has 4 thousand amino acids. Will a single one change its function or shape? I would say it is more likely not to change, and I don't know why this reasoning can be wrong. One amino acid in 4 thousand probably will have no effect in the overall shape and function and will form a slightly different/ almost identical protein with almost identical functions. If that is right, we should have an enormous number of alleles for the same gene (e.g. if we say less than 5 amino acids won't change the structure, that are 3,2 million different alleles), instead of 2, or 3, or 4, or some small number as I was taught.

I can think in 2 explanations: The first is that if, contrary to what I think, a small change in an unique amino acid could change the protein enough to make the individual die for example. Which I doubt. The second is if we in fact had millions of alleles, but all this alleles had similar or almost identical shapes, and we could segregate them in a small number of groups (2 for example) where the functions of the coded proteins are different from the other group but identical in the same group.

Am I going too far? Could anyone explain me why this is not possible?

 

[BillTre]

Your understanding of what an allele is seems OK to me.

The protein has 4 thousand amino acids. Will a single one change its function or shape? I would say it is more likely not to change, and I don't know why this reasoning can be wrong. One amino acid in 4 thousand probably will have no effect in the overall shape and function and will form a slightly different/ almost identical protein with almost identical functions.

You con not depend upon these statistics because there are a lot of different situations they are not taking into consideration.

• Each protein will have many different amino acids.
• Some will be more important to the protein's structure and function than others.
• Some changes in an amino acid at some locations in the sequence will have little or no noticeable effects.
• Other changes could have a big effect.
• Point mutations can also create stops in protein translation, where the RNA sequence does not say add the wrong amino acids, but instead just says stop. This results in a shorter protein, which can have big functional differences.
• There are other kinds of mutations, such as breakpoint mutations, where chunks of sequence can be removed, flipped around, or inserted, causing large changes in encoding sequence all at once.

In addition, when it comes to visible (larger scale) phenotypes (in people for example), they are often generated by the actions of many different proteins, possibly expressed in many different kinds of cells working together to generate the final features of the adult structure. Small changes in what amounts to a small component in this process (a single kind of protein) may or may not have an effect that can be observed at larger scales. It depepnds upon the specific changes to the component as well as its location and functional importance within the larger overall system. Small critical differences can have important effects.

we should have an enormous number of alleles for the same gene (e.g. if we say less than 5 amino acids won't change the structure, that are 3,2 million different alleles), instead of 2, or 3, or 4, or some small number as I was taught.

There is a numerical potential to have lots of different alleles for a particular gene.
However, there are some limitations to this approach:

• There is an upper limit to how much genetic diversity a diploid population could have, 2 X (the number of individuals in the population).
• Most of the versions of a gene you find, will be inherited from that individual's parents, in the form in which they are observed. Normal mutation rates are not so high that many new mutations, of a particular gene, in a particular individual would be expected to be found. There may be between a (1 in 1,000) to a (1 in 10,000) chance of getting a new mutation, in a particular gene, in a particular individual.
• Variation will be lost from a breeding population due to selection (removing negatively-adaptive alleles) or drift (random removing of alleles with no particular advantage), thus reducing the number of alleles.

 

[Ygggdrasil]

Ok, so let's say a super energetic UV photon reaches one of my germinative cells and mutates a single nucleotide of my DNA that changes the provious amino acid. The protein has 4 thousand amino acids. Will a single one change its function or shape? I would say it is more likely not to change, and I don't know why this reasoning can be wrong. One amino acid in 4 thousand probably will have no effect in the overall shape and function and will form a slightly different/ almost identical protein with almost identical functions.

This is generally correct. Most amino acid substitutions will have no effect or minimal effect on the protein's function. Experiments studying this question in the context of microbial evolution have found that maybe ~60% of amino acid substitutions have no significant effect on the protein's function (https://www.nature.comarticles/nrg2808). However, as @BillTre noted, there are single amino acid mutations or other types of mutation can completely change a protein's structure and eliminate its function. Indeed, in the study mentioned above, they found that ~10% of amino acid substitutions completely destroyed the protein's function.

If that is right, we should have an enormous number of alleles for the same gene (e.g. if we say less than 5 amino acids won't change the structure, that are 3,2 million different alleles), instead of 2, or 3, or 4, or some small number as I was taught.

Indeed, there can be an enormous amounts of variants of a gene. For example, in cancer biology, there have been efforts to sequence tumors in order to beter understand the mutations that might be present in cancer cells that drive the disease. One such study sequenced 91 brain cancers, and found 38 mutations in the the tumor suppressor gene TP53, that are spread all throughout the protein:

(source)

The x-axis is position along the protein sequence, y-axis shows the number of mutations found at that position. Note that this situation is slightly different from alleles in genetics because the mutations found in cancer cells are somatic (mutations that do not occur in germ cells and therefore cannot be inherited) and cancers often have elevated mutation rates compared to normal cells.

I can think in 2 explanations: The first is that if, contrary to what I think, a small change in an unique amino acid could change the protein enough to make the individual die for example. Which I doubt. The second is if we in fact had millions of alleles, but all this alleles had similar or almost identical shapes, and we could segregate them in a small number of groups (2 for example) where the functions of the coded proteins are different from the other group but identical in the same group.

The second situation is the correct explanation. Remember that the concept of alleles predates any of our knowledge of DNA, so alleles are generally classified by the phenotype they produce, not their specific genetic alteration. In many cases, different mutations can all have the same effect. For example, in the TP53 example posted above, most of the mutations are likely contributing to loss of function of the p53 protein encoded by the TP53 gene and essentially have the same effect. Similarly, genetic diseases can be associated with a number of different mutations. Here's a link to a database showing all of the different mutations found at the gene associated with Hemophilia A: https://www.ncbi.nlm.nih.gov/clinvar?term=300841[MIM]

Modern genetics tends to focus on changes at single nucleotides called SNPs (single nucleotide polymorphisms). In this case, there can be only four different alleles (corresponding to the four different nucleotides found in DNA).

Gene sequencing efforts in clinical settings find genetic variants that have not been observed previously. While scientists may be able to guess at whether the mutations can cause disease, many times these mutations get classified as variants of unknown significance (https://blogs.plos.org/dnascience/2018/05/03/whats-a-variant-of-uncertain-significance-a-vus/). Indeed, the clinvar database I linked above has a number of variants classified as VUSs. A major challenge in genetics research is to develop better tools to theoretically or experimentally determine the effets of VUSs.

 

[jaumzaum]

Thanks @Ygggdrasil and @BillTre, I never thought about alleles being classified by the phenotype rather than the genotype, that cleared my mind!

So we all agreed that the number possible genotypic alleles that will not kill a person or will make almost identical proteins is huge. That makes me think about the number or these alleles that has already occurred and are spread in the population.

I actually searched for some numbers, and what I found out is that mutations are far more common than what I thought. The problem with that is that I come with another apparent contradiction. Wikipedia says human mutation rates are 2,5 x 10^-8 mutations per base per generation. It is estimated 100 billion humans have already lived. One can calculate the expected value of the numberof mutations that has already happened for a specific base locus in human DNA history = 2500. As there are only 4 possible bases, almost all the bases should already have had a mutation gene, i.e. human DNA should be extremely different from person to person and it should be almost impossible to find 2 people with exactly the same gene for any given protein.

That clearly does not happen. I don’t have numbers, but if we compare the entire DNA sequence of 2 different individuals what is the percentage of similarity? I would guess more than 95%? So why is that possible?

 

[Ygggdrasil]

Given the human genome size of ~ 3e9 base pairs and a mutation rate of 2.5e-8, we expect ~ 75 new mutations each generation. According to Wikipedia, modern Homo sapiens emerged about 250,000 years ago. Assuming a generation time of ~20 years, this corresponds to ~12,500 generations. Thus, each modern human should have ~ 1 million differences from the ancestral human genome (assuming no selection). Thus two random individuals might expect ~ 2 million differences between their genomes.

Interestingly, studies have estimated that two individuals will typically have ~20 million differences between their genomes (https://en.wikipedia.org/wiki/Human_genetic_variation), so that back of the envelope calculation above is somehow too low by an order of magnitude.