And I respond to Walter.
WR
>Steve Clark probably didn't intend it, but he made a fair argument for why
>the immune system must have been designed, rather than evolved:
In fact I didn't intend to make an argument one way or the other. I think
that you are trying to paint me into a corner to which I will not go.
Regarding the immune system, the evidence, as far as I am concerned, is
consistent with either design or evolutionary explanations. I generally do
not find the two explanations incompatible, because design could work via
evolutionary mechanisms. The randomness "required" for evolution poses no
problem for me since I believe that randomness is not the same as
purposelessness. While Darwinian evolution has a component of randomness, I
do not believe that there is an a priori requirement for purposelessness,
regardless of what Huxley, Dawkins, et al., beleive.
While mutation may be fairly random and non-directed, I fail to see how
anyone can view natural selection as anything but random because it
supposedly selects the "best" mutations that fit a particular environment.
By definition, selection implies a "choice" between those more and less fit.
Selection is not left to chance. Therefore, there seems to me to be plenty
of opportunity for a designer to work by, what appear to be evolutionary
mechanisms.
In sum, it makes sense that I find design an adequate explanation of the
immune system.
>My previous post speculated on a mechanism that would tune an *already
>existent* immune system. It did not attempt to explain the origin of the
>immune system or MHC genes. Rather, the mechanism modifies the sequence of
>*existing* MHC genes. As proposed, the mechanism would operate on
>populations (and species) over generations. (It would not fine-tune an
>individual within it's own lifetime, the immune system already has other
>mechanisms for that.)
>Steve described the established view in this field. Take that view exactly,
>and change one thing. In Steve's view, the gene conversion serves SOLELY to
>inject *random mutation* into MHC genes.
Oh? Methinks that you are a lumper rather than a splitter. That is, you
lump all who do not speak your language into a single metaphysical mold
without taking the time to understand their positions. Read on.
However, I speculate that the gene
>conversion also injects a component of useful information. I also speculate
>the information comes (in some way) from the result of viruses.
Ok on the first speculation, doubtful on the second. I also hope that you
recognize that this is only speculation and that you have offered no
evidence for your scenario.
Depending on one's definition of information in biological systems, I
contend that "random mutation" is not as inocuous as you imply and that it
involves information. So what is it that you mean here?
Further, you are wrong that gene conversion is random. Rather, it involves
a process similar to homologous recombination, in which similar but not
identical DNA sequences provide a template for changing the DNA sequence in
another related gene. The sequences that have been identified as candidates
for providing templates for gene conversion of MHC genes, are in fact
MHC-like pseudogenes that exist in numerous copies, at least in the mouse
genome, linked within the MHC gene complex. There is no reason to think
that viruses play any role in this process, whether by creating pseudogenes
via insertional mutagenesis, or by contributing to the template sequence.
In fact there is more reason to think this doesn't happen. A few viruses
can integrate into the mammalian genome and there exist transposable genetic
elements that are rudimentary viruses. All of these have a characteristic
fingerprint that allows us to follow them around in the genome. As far as I
know, this does not seem to be a major factor in creating MHC pseudogenes or
in contributing to the polymorphism of genes in the MHC. But this doesn't
mean it couldn't happen. Read on.
MHC genes belong to a huge family of cell surface proteins that have a
characteristic structure which consists of strings of individual protein
domains. These protein domains all maintain a conserved globular structure
known as a beta-barrel, and they represent the building blocks by which a
vast array of cell surface molecules are composed. Since each of these beta
barrel protein domains in different cell surface molecules are encoded by a
single genetic exon, evolutionary speculation is that they arose from a
primordial monocistronic gene that encoded a single beta-barrel-type protein
and that, over time, increased in complexity through amplification of these
singular domains into a variety of multi-cistronic genes, including the MHC
genes. Then there is the high degree of homology of MHC genes between
species. Not only are they homologous in DNA and protein sequence, they
exist as linked genetic complexes containing class I and class II MHC genes
(which have different function), MHC pseudogenes as well as genes for
unrelated serum proteins. There is no good evolutionary answer for
maintaining this complex so closely linked in divergent species. But it is
consistent for the evolution of the MHC complex prior to the divergence of
primates, rodents, and rabbits.
If you wish, you make invoke design at this stage, but as I said before, the
same data can have more than one equally valid conclusion. Design and
evolution are not necessarily mutually exclusive.
>Gene conversion is capable of what I describe, so the issue is whether there
>is information being injected, and (so far) Steve has not shown that
>implausible.
Nor do I have any desire to claim that information transfer here is
implausible. You put words into my mouth, I had no idea we were debating
whether this involved information transfer.
Certainly, sequence changes in the two outer domains of the MHC proteins
affect the way that they fold and present antigen to the immune cells.
Thus, gene conversion can contribute changes to the sequence of the MHC
molecules which, in turn, can dramatically change what type of antigens the
immune system can recognize. Thus, by changing the sequence information,
function is altered.
But doesn't anything that alters a protein sequence inject new
"information"? For instance, in ex vivo systems, cells treated with
mutagens that act fairly randomly with respect to the DNA sequence, but that
do not directly introduce new DNA sequences to a gene like gene conversion
does, can cause similar changes in MHC gene sequence and function. Does
this fit your criteria of injecting information?
>His post made one central argument. He argued that my mechanism is
>implausible because its RATE of adaptation would be too slow to do any good.
>He didn't supply any numbers, and he needn't bother -- his argument fails
>for a deeper reason.
What argument? The MHC genes and viruses just do not do what you speculated
them to do.
I never intended this mechanism to supply the
>super-rapid immunity that develops within one individual's lifetime.
>Compared with that, my mechanism provides a slower, less specific
>trans-generational course correction for the MHC genes. Let me emphasize
>again, my speculated mechanism is identical to Steve's view, with one key
>difference. In Steve's view, the gene conversion injects *random* mutation
Actually, this is a great extrapolation from anything I said in previous
messages. Gene conversion doesn't work by injecting random mutation.
Chemical mutation of cells, as I explained above can be considered random,
but gene conversion is sequence specific which is how it can be identified
as being different from "random mutation", and how the template pseudogenes
that contribute to the conversion can be identified.
>-- I speculate that it also injects information. Steve is ALREADY satisfied
>with his view, therefore he cannot argue that the addition of information
>makes the situation inadequate or too slow.
See my earlier comments regarding information. Why put words in my mouth
when the problem comes down to explaining what you mean by information?
>I think Steve misunderstood my idea. I will not fault him for that,
Thanks, I guess.
rather
>I take it as cause for more elaboration on my part.
Oh good.
My idea is about
>***information flow***,
Ah, I wish you had said so eariler. I don't think we disagree on the
information point, just on the way it could happen. I don't have any
problem with the concept of information transfer, but I think your model is
more complex than necessary.
>For example, the virus might directly turn an MHC gene into an MHC-like
>pseudogene.
That is possible, but would be detectable. There are many examples of
cellular genes being interrupted by "viral" sequences, but again, this
involves the general category of transposable elements that have very
characteristic fingerprints.
That pseudogene contains some information about the virus. And
>that information gets merged (by gene conversion) with the MHC genes, just
>as Steve described. Thus, I see a potential information pathway.
>The "information" I speak of is not necessarily the virus sequence itself.
>The mechanism could work even without any virus sequences ending up in the
>MHC genes. How? The virus preferentially attacks certain genes, and
>creates pseudogenes. The very EXISTENCE of that PARTICULAR pseudogene,
>signals the existence of the virus, and conveys information about its
>identity. For example, hemoglobin pseudogenes may signal that hemoglobin
>genes are under assault by viruses. In this way pseudogenes can convey real
>information about particular viruses.
This sort of "attack" has been documented and well studied. Such attack can
also cause inappropriate expression of cellular genes and lead to a
pathogenic state. As a matter of fact, I am currently writing paper for
submission to the Journal of Virology, in which we purposively tag such
viruses, let them do their thing, and when a malignancy arises in a mouse,
we identify which genes our tag is near. Thus, we identify new genes
involved in tumorigenesis.
To reiterate, I know of no example where an MHC pseudogene is a psuedogene
due to the action of such a virus. I think you make your information
transfer model needlessly complicated by all this.
>Information filter:
>
>I saw the potential of information flow. Gene conversion might serve to
>convey some information into the MHC genes and the immune system. But this
>BY ITSELF is woefully inadequate. The organism is full of DNA "information"
>and overwhelmed with it. There is too much DNA information, and it does no
>good to send it all to the MHC genes. You need a filter -- an information
>filter -- some way to preferentially filter INFORMATION ABOUT VIRUSES into
>the immune system. I saw potential for such an information filter, and it
>centered around pseudogenes.
>
>1) Viruses create pseudogenes
>2) MHC genes experience a gene conversion in which portions of other
> *** MHC-like pseudogenes *** are exchanged with homologous regions
> of expressed MHC genes.
>3) Thus, pseudogenes can act like an information filter. Not all DNA
>information is merged with the MHC genes -- only specific pseudogenes,
>... and viruses create pseudogenes.
Except for #1, and the last four words you are correct. The filter process
comes from the requirement for some sequence homology in order for gene
conversion to occur, much the same way that sequence similarity facilitates
homologous recombination in somatic cells. Forget the viruses, and you have
the essence of the gene conversion process.
>Above I discussed two examples where the mechanism might be confined solely
>to the MHC genes and MHC-like pseudogenes. On the other hand, it might
>conceivably work for many varieties of pseudogenes throughout the genome.
Perhaps, perhaps not. We probably can't answer this question at this stage
because it is very hard to show gene conversion in higher mammals. The
nature of the MHC genetic complex is very unique and made it easier to
identify gene conversion in this instance.
The scenario that you present provides an example of how one wishing to
identify whether information flow works in a biological system, could come
up with questions to test experimentally. However, the same experiments
have already been considered by naturalists who wish to learn if gene
conversion functions elswhere.
>If that is the case, then there are two additional information hurdles to
>overcome:
>
>1) The process would need an "information concentrator" or funnel, some
>mechanism that takes information from various pseudogenes around the genome,
>and funnels it into the MHC genes (or into the MHC-like pseudogenes).
Actually, the funnel is built in. The MHC-like pseudogenes provide the
homologous template for gene conversion. There also may be a proximity
requirement, which fits the pseudogenes that are found within the MHC
genetic complex. "Various pseudogenes around the geneome" may be irrelevant
here.
>2) I have not identified how the immune system or MHC genes would actually
>interpret and use information from various pseudogenes, such as from
>hemoglobin pseudogenes.
They may never get the chance.
>I thank Steve for his gracious comments on my admittedly speculative idea.
>He didn't give it the flailing I anticipated would come.
Yeah, well I do have a carefully cultivated reputation as a belligerent.
But if you think I'm bad, you ought to see my wife!
I emphasize again
>that my idea is about information flow, and Steve has not even argued, much
>less shown, that such information flow is implausible.
That is because I didn't know we were debating the topic of information
flow. At any rate, I don't think any biologist would deny that information
flow is central to biology, even to gene conversion. So I am sure what your
point is here. Your information flow to me seems to be filtered too much
(to put it into your words).
Steve
____________________________________________________________________________
Steven S. Clark, Ph.D. Phone: (608) 263-9137
Associate Professor FAX: (608) 263-4226
Dept. of Human Oncology and email: ssclark@facstaff.wisc.edu
UW Comprehensive Cancer Ctr
University of Wisconsin "To disdain philosophy is really to
Madison, WI 53792 be a philosopher." Blaise Pascal
____________________________________________________________________________