Lessons from the pandemic: A new look at an new virus: patterns of mutation accumulation in SARS-CoV-2 since 2019

[quote=“Mercer, post:59, topic:15113”]

It does. The specificity of an antibody comes entirely from the variable region of the light chain (VL) and the variable region of the heavy chain (VH), both of which resulting from the process you mentioned.

It doesn’t.

Question for you, since you claim to understand this so deeply: What is the percent similarity between the rearranged and mutated V region of an IgG gene from a MATURE B-cell clone and the germline?

There’s no speculation required. This is about what we know happens in real time.

It does. Note that I am not saying that the specificity of an antibody comes entirely from the V, D, J segments, but from the variables regions of the final immunoglobulin. This is basic knowledge in immunology.

No, you weren’t.

You wrote,

It is, as is the importance of sloppy recombination and somatic mutation in creating new information critical to the evolution of antibodies. But you were using that as your opener to denying that basic knowledge. Have you changed your mind?

Pleese consider the paper by Keefe and Szostak published in Nature titled « Functional proteins from a random-sequence library » where the authors estimate that roughly 1 in 10^11 of all random-sequence proteins have ATP-binding activity comparable to the proteins isolated in the study. According to the Gregory Winter experiments @RonSewell referred to at 57, this frequency, 1 in 10^11, is about 1000 times lower than the frequency of antibodies exhibiting a given specificity in a naive repertoire. But things are much worse than this for Keefe and Szostak found that for all the rare polypeptide sequences able to bind ATP, only a small proportion (~10%) of the corresponding population of polypeptides adopt a bona fide folded structure allowing binding, whereas for antibodies, it is nearly 100%! So no, it is definitely not the case that « there is nothing to indicate that any individual VDJ segment is closer in sequence space to matching a never before encountered antigen epitope than just any sequence pulled at random from sequence space ». Functional proteins from a random-sequence library - PMC

You’re right. I had not considered the structural aspects of the VDJ segments and was thinking only of the potential mutual affinity exhibited by a CDR to some arbitrary antigent epitope. And even here I think it is unavoidable these small CDRs will also have some effect on structure that restricts their valid sequence space and are thus under selection to stay within certain regions that don’t too negatively affect IG structure. I concede the point.

1 Like

Please contrast this with Doug Axe’s absurd extrapolation.

There are multiple problems with your attempt to compare these, because binding is not digital and specificity is not required to start B-cell selection.

For example, would the frequency Keefe and Szostak found go up or down if they had eluted with 0.5 mM ATP instead of 5 mM? With 15 mM instead of 5?

It is suspect to me that an intrinsic attribute of an object is contingent on the existence or potential existence of another remote object, but I think we agree that the aggregate antibody shape space is immense. If the information defining that space actually was contained in the immunoglobulin genes, what would that look like?

The most certain definition would to be to hard code as in the sense of smell; one gene, one molecular shape. The information required to enable olfactory responses to particular odorants repose in specific genes. For antibodies, that would require hundreds of millions of genes, which would have to be maintained to anticipate each specific threat. That rules out genetic encoding of the information.

Apart from hard coding, the two mutually exclusive ways to saturate the needed shape space would be systematically or by maximizing randomness. The characterization of antibody production as controlled or systematic must be qualified, as that might imply some sort of serialization or state memory which is not present. The recombination enacted by one B cell is not determined by any result of the previous cell.

True, but that misses the point. By nucleotide, the possible combinations would be about 4^110, much much larger than 2.1 million H and L chain combinations available before any mutations. But a particular chain combination is still close to the odds of drawing a royal flush of spades. While a card shuffling machine or lottery machine may be controlled, the result of combinatorial shuffling is not contained in the contributing elements. Quite the opposite. Systematic bias is a flaw in randomization, and there are people on the prowl to capitalize on any they find. The object of shuffling is to eliminate any transfer of ordered information and send it to a chaotic abyss. Given the range of antibodies produced, it seems that VDJ recombination is reasonably effective in approaching this.

Combinatorial diversity, however, is just to get started - the early stages of B cell maturation.

Antibody Engineering - Generation of Antibody Diversity

Because of the huge diversity, the immunoglobulin repertoire cannot be encoded by static genes, which would explode the genomic capacity comprising about 20,000–25,000 human genes.

Molecular Biology of the Cell - Generation of Antibody Diversity

During V(D)J recombination the diversity of immunoglobulins is further increased by incorporation of additional nucleotides between the junctions of the V, D, and J gene segment of the heavy and V and J gene segment of the light chain.

Note that junctional diversity essentially is drawing from a bucket of nucleotides, and this takes place prior to hypermutation.

Even in the absence of antigen stimulation, a human can probably make more than 10^12 different antibody molecules—its preimmune antibody repertoire. Moreover, the antigen-binding sites of many antibodies can cross-react with a variety of related but different antigenic determinants, making the antibody defense force even more formidable. The preimmune repertoire is apparently large enough to ensure that there will be an antigen-binding site to fit almost any potential antigenic determinant, albeit with low affinity.

This is what the whole ID narrative of protein binding attempts to downplay. Whether dealing with virus - receptor, protein - protein, or antibody - antitope, affinity and disruptive forces oppose, so there will be a dynamic equilibrium between particles at large and attached. This equilibrium is far from binary. If there were no degree of binding affinity between antigen and the base of naïve antibodies, we would all succumb to infection. Low affinity binding can be biologically useful, for one by activating somatic hypermutation. Any trace information which might have been in the original VDJ segments is inconsequential in defining the repertoire shape space found in antibodies resulting from diversification and hypermutation.

2 Likes

I should add with respect to the whole “is the information already there?” question that I can see how the sequence information does actually matter with respect to evolution, and this is actually one of the reasons why novel functional proteins can evolve by combining fragments of other already existing proteins.

2 Likes

In a naive repertoire, the variable regions of antibodies are encoded almost entirely by the V, D and J segments which are brought together by a somatic recombination process to produce a complete V-region exon. These variable regions adopt the immunoglobulin fold, which is composed of nine beta sheets. The antibody-binding site is formed by three loops of amino acids known as hyper variable regions HV1, HV2 and HV3, or also CDR1, CDR2 and CDR3. These loop are located between the pairs of beta sheets B and C,C’ and C’’, and F and G. In addition to these HV regions, the variable regions are composed of four framework regions (FR1, FR2, FR3 and FR4). The variable regions of antibodies are about 120 aa long polypeptides that can be represented as follow:
FR1/HV1/FR2/HV2/FR3/HV3/FR4. In a naive repertoire, the four FR regions as well as HV1, HV2 and part of HV3 are entirely encoded by the VDJ segments and only a part of HV3 (representing less than 10% of the length of the variable regions) is not encoded by them.
Moreover, whereas the HV regions determine most of the specificity and affinity of the interaction of an antibody with its epitopes, the FR regions provide a stable structural framework for the HVs and help enormously to determine their position and conformation.
The bottom line is as follow:

  • most of the information for a functional antibody resides in the VDJ segments.
  • my safe lock analogy at 45 works pretty well

He’s referring to matured

So why are you going on about the naive repertoire as though you are refuting Ron, Gil?

1 Like

It doesn’t matter whether Ron is referring to the naive or mature repertoire, for in both case, his claim is wrong. I’ve presented at 70 arguments showing he was wrong for the naive repertoire, but the same type of arguments can be used to show that he is wrong also for the mature repertoire.

This is probably not the case for the folds. Is there any evidence that a given fold such as, say, the immunoglobulin fold, can evolve by combining functional sub fragments?
To return to the paper by Keefe and Szostak referenced at 65, another interesting point to emphasize is that the results it reports are highly concilient with the view of ID proponents that the protein folds found in nature are rare and special.

You should watch this lecture:

There is a considerable literature now on fold-evolution and the many different connections between known protein folds.
Two recent articles:

While I don’t know whether the IG itself owes it’s origin to such a process(haven’t looked into it), the IG fold has itself lend fragments to the emergence of other protein folds:

2 Likes

No, Gil, we’re both pointing out that your claim:

is false.

The converse is true; most of the information is generated de novo by sloppy recombination and somatic mutation, in only two weeks.

The vast majority of functional proteins will not bind ATP, and yet they are still functional. The vast majority of functional antibodies will not bind ATP, and yet they are still functional. You may want to think about that.

The variable region of the antibody is within the larger sequence which allows for functional folding. The same could be said about the evolution of new function from already existing proteins that can effectively fold into stable structures. That’s not a problem.

1 Like

Take any molecular entity exhibiting about the same size/complexity than ATP. What can be inferred from the paper by Keefe and Szostak is that roughly 1 in 10^11 of all random 80 aa long peptides will have the ability to bind this particular molecular entity, not only ATP. This frequency (1 in 10^11) must be contrasted with the frequency of antibodies exhibiting a given specificity in a naive repertoire, which, at roughly 1 in 10^8, is about 1000 times more frequent. So it is clearly not the case that « there is nothing to indicate that any individual VDJ segment is closer in sequence space to matching a never before encountered antigen epitope than just any sequence pulled at random from sequence space ».

No, the variable regions of antibodies have the ability to adopt a functional folded structure by themselves, independently of the larger structure. scFv are a case in point.

I don’t understand what you’re saying here, can you rephrase it?

Yeah I agree I made an unwarranted extrapolation when I said that. I was thinking about the affinity of the parts of the VDJ segments known as the complementary determining regions (CDRs), which are those that directly contact the antigen epitope and bind to it, while not thinking about the effect of the CDR itself on the structure of the variable region.

But again I think we need to be specific when we say either the information isn’t there—or it is—what kind of information we are talking about. And I think it is a mistake to conflate sequence information with functional information, even though it’s clear that they are in some sense related and contribute to the evolution of both proteins in general and antibodies more specifically.

Thinking about it some more I think it’s because some higher-order functions are actually the product of other simpler/less complex functions. The whole is in some sense more than the sum of the parts.
Some parts of proteins contribute to the overall function of the protein (say enzymatic activity or ligand binding) by providing stability, and/or flexibility, and/or folding information. In this sense the protein’s overall function can be considered the product of the combination of a host of sub-functions.

I think the way the immune system functions basically provides a demonstration that a process of sampling of different combinations of sub-functions(or things being near valid sub-functions) massively speeds up the search for higher order functions that a priori can be extremely rare in protein sequence space.

So when thinking of protein evolution more broadly, I think it teaches us that the discovery of very rare protein functions in sequence space most probably proceeds by first discovering very simple functions, and then these providing stepping stones and building blocks towards more complex and rare functions.

1 Like

No, binding is not binary. You left out the affinity that they specified. Why?

Not at the same affinity, Gil. Binding affinity is not binary.

Yes there is: affinity. What’s clear is that you are ignoring the requirement to specify affinity and/or activity.

That fudges the definitions of “fold” and “structure” beyond any recognition. All such regions are constrained to not disrupt immunoglobulin structure.

Still, a naïve inbred mouse repertoire (about 10^8) contains more than one antibody that has measurable beta-lactamase function.

That demolishes Axe’s and your conclusions, which ignore both functional measurements and the fact that binding is not binary.

Did I mention that binding is not binary?

The point here is that the frequency of 1 in 10^11 does not only concern ATP but can be generalized to any molecular entity exhibiting about the same size/complexity than ATP. Note that this generalization is made by Keefe and Szostak in their conclusion.