Arabidopsis evolution

BIOCUKM at osucc.bitnet BIOCUKM at osucc.bitnet
Mon Apr 29 14:41:00 EST 1991

     Arabidopsis molecular biologists and geneticists should look at the
recent paper by Kemmerer, Lei and Wu (J. Mol. Evol. 32:227-237, 1991).
The paper reports the isolation and nucleotide sequencing of the
cytochrome c gene of Arabidopsis.  The unusual aspect of the paper is
that the predicted amino acid sequence is much more closely related to
those of the cytochromes c of Neurospora and yeasts than it is to the
cytochrome c sequences of another member of the crucifer family,
cauliflower, and, indeed, of all flowering plants whose cytochrome c
sequences are known.

     When I first saw the abstract, I thought that this would be another
example of grand conclusions from uncertain trees.  Calculated
phylogenetic trees have uncertainties associated with the placement of
branches.  Those uncertainties are rarely included in diagrams and are
often ignored when drawing conclusions.  On reading the paper, I found
the authors well aware of these difficulties.  They analyzed the
sequence data in many different ways.  Each way produced the same highly
significant conclusion.  A visual examination of the amino acid
sequences leads to the same conclusion, that the cytochrome c of the
sequenced gene is more similar to Neurosopora and yeast sequences than
it is to those of higher plants.

     What's going on?  Several explanations come to mind.
1.  The sequenced gene could encode a minor cytochrome c of plants, one
not detected by protein sequencing.  A S. cerevisiae probe was used to
screen an Arabidopsis library.  All 4 hybridizing library clones were
the same.  Thus, the gene is the only Arabidopsis cytochrome c gene
recognized by the yeast probe.  Perhaps the genes for the major
cytochromes c of plants are not recognized by the S. cerevisiae probe.
A difficulty with this and other technical explanations is that similar
conclusions about ancestry were reached with histone H3 comparisons
(though these are less certain).

2.  The sequenced gene was from a contaminant in the DNA preparation.
This is unlikely since a fragment of the cloned gene recognizes a
fragment of the same size in blots of Arabidopsis DNA.

3.  Lateral transfer of the cytochrome c gene (either from yeast to
Arabidopsis or, I suppose, vice versa).  This possibility, raised in the
paper, seems unlikely since the transferred gene would have to have
completely replaced an ancestral gene.

4.  My off-the-cuff favorite hypothesis is that in studying molecular
evolution over such large distances we are really not studying the
evolution of the molecules themselves.  We are really studying the
evolution of the fitness landscape.  Each of the cytochromes c has
evolved to be optimally suited for its particular ecological niche.
Perhaps the relevant features of the Arabidopsis cytochrome c ecological
niche are smallness and rapid growth.  Those properties seem more like
those of the yeasts than of cotton, mung bean, cauliflower, sesame,
sunflower, wheat, buckwheat and Gingko.  Against this view is the
presence of Chlamydomonas on the higher plant branch.

5.  A synthesis of explanations 3 and 4 would be that somewhere in
Arabidopsis' ancestry it was in close association (symbiosis?) with a
fungus.  The association was so close that heterokaryons formed.  With
time one of the copies of genes that were pesent in both the plant and
the fungal progenitor was lost.  Selection may have favored the fungal
cytochrome c.

     What do the rest of you think?  This is a question of economic, as
well as scientific, importance.  The paper makes the suggestion that
Arabidopsis may not be a good model for higher plants.  If this view
takes hold, there goes the funding!

!!!       Ulrich Melcher                     !!!
!!!       Department of Biochemistry         !!!
!!!       Oklahoma State University          !!!
!!!       Stillwater OK 74078  USA           !!!
!!!  BIOCUKM at OSUCC (Bitnet)    405-744-6210  !!!

More information about the Arab-gen mailing list