Archaebacteria and the Three Kingdoms

peter at carrot.mcb.uconn.edu peter at carrot.mcb.uconn.edu
Mon Apr 11 11:57:28 EST 1994


In article <Co2HvB.LDC at gpu.utcc.utoronto.ca> lamoran at gpu.utcc.utoronto.ca(L.A. Moran) writes:

>Benachenbou-Lahfa et al. (1993) have been critical of the EF-Tu and
ATPase
>data. They point out that their glutamate dehydrogenase dendrograms do
>not support the Three Domain hypothesis. How do you respond to their
>criticism of your results and to Sidow and Bowman (1991) who claim that
>your methodology was flawed? This is an ideal forum to continue this
debate

As far as I remember the paper from Sidow and Bowman their criticism of
or initial paper in PNAS was that we had only one archaebacterial species
included in our analysis.   This has changed since (Sulfolobus,
Methanococcus, Methanosarcina, Halobacterium, Thermoplasma).  They
criticized Iwabe et al. for using only one algorithm (we had used
Parsimony, distance matrix, evolutionary parsimony and maximum likelihood
methods).

As I pointed out before, their are several molecular markers that group
the archaebacteria with the low GC gram negatives, some of them appear to
be well resolved.  To me these data (as well as the eubacterial type
ATPase encoding gene in Methanosarcina, and taken together with the data
that show archaebacteria and eubacteria as distinct "domains") suggest
horizontal gene transfer.  The finding that different molecular markers
reflect the same association suggests that it was not a transfer of a
single gene but the "transfer" of a substantial portion of the genome.
The alternative explanation of paralogous genes already present in the
last common ancestor requires a lot of coincidences (the archaebacteria
and the low GC gram negatives would have had to loose the same copies of
glutamin synthetases, HSP70 homologues, nif genes, ATPases), furthermore,
a lot of duplicated genes with identical function would have had to been
present in the last common ancestor (V and FATPases, 4 glutamine
synthetases, 2 HSP70 homologues...)
The glutamate dehydrogenase data from Forterre et al. when analyzed after
bootstrapping do not appear very conclusive; therefore, I did not
include them in my last posting.



> (You may have done this in Hilario and Gogarten
>(1993) - I can't find this paper; but I haven't given up yet.)

If you let me know your mail address, I'll send you a reprint.

Peter Gogarten



More information about the Mol-evol mailing list