From: Aubrey de Grey (ag24@gen.cam.ac.uk)
Date: Wed Apr 23 2003 - 10:53:31 MDT
Rafal Smigrodzki wrote:
> Tarassov
This is indeed cool, but not obviously going anywhere therapeutically.
No one (even evolution) has succeeded in getting big RNAs into the
mitochondrion, and for this to work we would need lots of them (the
two ribosomal RNAs and also at least 11 messenger RNAs). For protein
import all we have to do is hijack a pathway that already works fine
for 1000 other proteins and use it for 13 more -- much easier.
> Guy
Right. This is the second one to be got working in mammalian cells;
ATP6 is the other (Nat Genet. 2002 Apr;30(4):394-9).
> only two mt proteins have *not* been found in the nucleus in one
> organism or the other.
Wrong (though often stated). Only two (cytochrome oxidase I and
cytochrome b) have always been found in the mt, but that's not the
same, because Complex I is completely absent in some yeasts and in
Plasmodium -- they get by with an enzyme that does the electron
transport but no proton pumping (for which they manage with the
contributions of Complexes III and IV). There are four subunits of
Complex I for which these are the only species that lack the genes in
the mtDNA, and of course they also lack them in the nDNA.
> Possibly the reason for stability of the mtDNA in
> evolution is not the protein import problem, but the difference in the
> nuclear and mito genetic code.
That works for animals, but not plants (see my mail of yesterday).
> ### Human mt-mRNA is polycistronic.
Sort of..... it's transcribed as one message for each strand, but they
are then chopped up and each bit polyadenylated before translation.
Two of the protein-coding products of the chopping-up are dicistronic,
the rest mono.
> my article on mt mutations in PD is written up, we are just working on
> the text. I can't tell you the details but let me say one thing - there
> is a lot more of them than most people thought.
I know of some other work in this area (but likewise I'm afraid I can't
divulge much at this point). I can say, though, that the question of
what causes what is made even more fraught by this data.
> If there is a different underlying process, dealing with the aggregates
> will not remove all of the problem, although it could ameliorate some of it.
> I would rather try to attack the ultimate cause, which most likely doesn't
> involve the proteins that actually aggregate
That depends. If there is a different underlying process that does not
exert its pathogenicity via the aggregates, then I agree, but that's a
big "if". If it does, then getting rid of the aggregates is a complete
solution -- and an ideal one, because we didn't have to mess about with
normal metabolic processes to try to make them "cleaner".
> in the absence of Elan, amyloid is extralysosomal, so it
> couldn't act by choking the lysosomes, and putting xenoenzymes in this
> compartment will not allow amyloid removal.
Right. It's possible that engineering transgenic microbial hydrolases
to be secreted to the extracellular medium would work, but I think that
would be more likely to have side-effects. I prefer the Elan approach
in conjunction with lysosomal enhancement.
> Is there some alpha-synuclein in lysosomes simply because stuff ends up
> in lysosomes, or is it's presence in the lysosomes the mechanism of its
> detrimental action? This has not been answered so far, AFAIK.
I agree, but who cares? -- the only thing that matters is whether NON-
lysosomal alpha-synuclein is bad for cells, and no one has postulated
any way that it could be.
> ### Atheromatous material may be difficult to digest, but is not
> non-digestible. With some treatments you can have regression of
> atherosclerotic changes implying digestion of macroscopic amounts of
> plaque. There is a dynamic balance between plaque accumulation and
> digestion.
That doesn't follow. At any given instant, there will certainly be a
proportion of the atheroma core that's digestible (and is indeed being
digested). That proportion will diminish as a result of treatments
that slow down the arrival of new material, etc., so the total size
of the plaque will also diminish. But that doesn't tell us whether
there is *also* stuff there that *can't* be degraded. In fact we can
say with confidence that there is, because the plaque appears out of
nothing. The originating macrophages that enter our artery wall in
youth, for example, can only become foam cells by hanging out ther
and taking in a steady stream of LDL etc and failing to break it all
down -- else they would remain as healthy macrophages forever. Same
goes for the core of the mature plaque -- there's a constant stream
on new macrophages arriving all the time to do battle with it, and
if they were able to break down everything there they would do so.
You're quite right that macrophages can eat incredibly tough stuff
-- good thing too, or we would have atherosclerosis a lot younger --
but oxidised cholesterol is *really* tough stuff.
Robert Bradbury wrote:
> Has there been any work to determine whether the lower metabolic
> rate directly correlates with a lower cancer rate for their mass
> (or # of cells)? I.e. do they have better cancer defenses than we
> do or is it indeed due precisely to lower free radical production?
Do you mean, are elephant/whale cells more resistant to transformation
in vitro (where the O2 tension and consumption can be controlled)? I
believe Jerry Shay is working on this, but it's a hard question because
you have to do such a lot to human cells to transform them that there's
no reason to suppose that an organism with equally good (but no better)
defences would have exactly the same defences.
> I was under the impression that
> there was a significant variance in mtDNA sizes
Not in animals -- 15 to 18 kb. Reviews of this sort of thing come out
regularly; also see <http://megasun.bch.umontreal.ca/ogmpproj.html>.
> I also disagree that evolution could not pull a rabbit out of the hat.
> It would depend entirely on how many UGAs there would be in a specific
> mtGene. I consider the prospect of adding a mt Signal Sequence to be
> much more difficult than fixing the UGA problem. After all one cannot
> know how frequently mutated mtDNA might substitute all of the UGA codes
> with UGG codons (allowing an alternate Trp coding) and subsequently
> transfer the genes to the nucleus.
It's certainly true that biologists are still very bad at guessing what
evolution will find easy and what hard. But we have plenty of data on
this; gene transfer to the nucleus when there is no code disparity and
no hydrophobicity problem has happened hundreds of times in evolution
(so apparently gaining a signal sequence is easy), whereas there is not
a single recorded case of gene transfer occurring across code disparity
(i.e. of the most economical phylogenetic tree implying that).
> http://users.rcn.com/jkimball.ma.ultranet/BiologyPages/C/Codons.html
Indeed. But the UGA switch probably matters most, because it causes a
subsequently transferred gene to encode a truncated protein.
> there isn't a lot of conclusive evidence
> that you can reduce the gene hydrophobicity enough to get all of the
> genes into the nucleus so one may always be left with some mtDNA.
The evidence is now pretty compelling, actually. We know that there
is a degree of hydrophobicity of each transmembrane segment that is
enough to make it stable in the membrane when it's assembled into a
complex but not enough to preclude import. The mt-coded proteins are
thus "unnecessarily" hydrophobic and I am now very confident that we
can tweak them pretty quickly into an importable but still fuctional
form. See, e.g., PNAS 2002 Aug 6;99(16):10510: just changing **two**
amino acids makes all the difference for one protein.
> Another question is whether the SO is being made by "perfect"
> proteins or damaged proteins? If the proteins are becoming
> damaged and increasing SO production then we are back to the
> protein recycling issue all over again.
We're back to the protein recycling issue either way. In my book I
suggested that complete allotopic expression might double lifespan,
but I definitely only said there was a 10% chance of this... these
days I'm more interested in a more comprehensive onslaught on aging,
and protein recycling (i.e. lipofuscin elimination) is a key plank
of my approach.
Rafal Smigrodzki wrote:
> it is indeed possible to replace full length mtDNA now, so it's
> just a matter of time before somebody manages to do something useful
> with it.
Probably not in respect of aging. Unfortunately it seems that the
mutant mtDNA molecule has a selective advantage in post-mitotic cells
and clonally expands at the expense of the wild-type, eventually
taking over the whole cell. This process would overcome any attempt
to replace mutant mtDNA by wild-type copies. No one has come up with
any feasible way to reverse this selective pressure (and not for want
of trying!). Putting in mtDNA that reduces superoxide production
may be a different matter though -- but (broken record time) that's
just retardation, not repair/reversal/obviation.
Damien Broderick wrote:
> I don't recall whether you've posted here until recently
No, I just signed on a few days ago. I'm already glad I chose the
digest version...
Aubrey de Grey
This archive was generated by hypermail 2.1.5 : Wed Apr 23 2003 - 11:04:36 MDT