Jim
*****
rticle: 357213 of talk.origins
From: iz028@cleveland.Freenet.Edu (Julie Thomas)
Newsgroups: talk.origins
Subject: A Critique of Behe's Critic [Part 1]
Date: 4 May 1997 07:30:24 GMT
Organization: Case Western Reserve University, Cleveland, OH (USA)
Keith Robison has a review of Michael Behe's book among the FAQs for
talk.origins. After reading it, I couldn't resist offering my own
review
of Keith's review. Here is Part 1 of my reply (I hope to finish the
other
parts later this week if time permits).
>Pseudogenes
>One argument against an intelligent designer is the amazing amount of
>flotsam and jetsam in genomes. The human genome is 90-95%
>apparent junk, useless sequences, many of which resemble functional
>genes, but are clearly beaten up beyond working order (pseudogenes).
When Keith asserts that "one argument against an intelligent designer is
the amazing amount of flotsam and jetsam in genomes" he essentially
makes a point that relies on two value judgments: much of the genome
is "junk" and a designer wouldn't include "junk."
First, is it really true that most of the genome is "flotsam and
jetsam?"
In my opinion, this is a misguided assertion. Keith's mistake is one
that
is common among molecular biologists, namely, the only "function" of
DNA is to encode something. But this is a view that is becoming harder
and harder to hold. For example, Keith might consider highly repetitive
DNA "flotsam and jetsam," but he would be wrong. Such DNA is part of
the sequences that make up centromeres and telomeres. And while
these chromosomal regions do not code for proteins or RNA, one could
hardly dismiss them as "jetsam and flotsam."
What the telomeric and centromeric sequences show is that DNA
sequences can be essential/important for cellular life although they
don't code for any protein or RNA molecule. In fact, there are other
examples of important, non-coding DNA. While it is still often claimed
that introns are useless "junk," this view again reflects the view that
DNA must code for something or be useless. But there is a growing
body of evidence that suggests introns may indeed be useful. While they
may not code for anything, their splicing may be coupled with transport
>from the nucleus. In fact, many researchers who construct transgenic
animals have found that by putting introns into cDNA constructs, the
level of expression is significantly raised.
Furthermore, I would bet that as our techniques improve, we will find
chromosomes are not randomly arranged within the nucleus and that
the skeletal network of the nucleoplasm itself is "informational" (ie,
it
imposes specificity on the arrangement of DNA). We already know that
the ends of chromosomes interect with the proteins on the nuclear side
of the nuclear envelope. This would suggest that certain sequences may
actually serve the purpose of making semi-specific contacts with
nuclear-skeletal proteins. Thus, these sequences might be involved in
the specific arrangement and orientation of chromosomes within the
nucleoplasm.
Furthermore, one gets the feeling that Keith is viewing the eucaryotic
genome through procaryotic filters. Unlike procaryotes, eucaryotes
often employ rather large upstream sequences to regulate gene
expression. And what appears to be "junk" between regulatory
elements may in fact be crucial spacing. Thus, the so-called junk
actually allows for a far more complicated method of regulating gene
expression.
In other words, if your "no designer" hypothesis predicts most of the
genome is "flotsam and jetsam," the prediction may ultimately be
erroneous. The sequence of nucleotides in DNA need not only function
to encode protein and RNA. They can also be involved in DNA
replication (telomeres), chromosomal disjunction (centromeres),
gene expression, chromosmal arrangement and orientation, chromosome
packaging, the expression levels of genes through rates of nuclear
export and other dynamics we have yet to uncover and understand.
Of course, even if it is true that many DNA sequences serve no purpose
for the individual organism, they may/do serve a purpose in evolution.
Introns make exon shuffling possible. They also make alternative
splicing possible, where the same gene can give rise to different gene
products in different cell lines (a very handy and economical feature
for
a multicelluar critter). Also, much of the non-coding DNA is made up of
"jumping genes" which in turn increase the genetic diversity in a
population and over time. Furthermore, a genome that is loaded with
useless DNA is also buffered against mutations. If Keith's figures
about
the human genome are correct, then 90-95% of the mutations occur in
"flotsam and jetsam." That is, the "flotsam" may be a mutation-sink.
Thus, if the process of evolution itself was designed/intended, these
type of DNA sequences are not inconsistent with a intelligent designer.
And this takes us to pseudogenes. Are they really an argument against
an intelligent designer? That depends. Let's consider some relevant
points.
1. Pseudogenes are really better evidence of common decent. But
common descent is not evidence of "no designer." The concepts of
"common descent" and "design" are not mutually exclusive (although
this often appears to be a default assumption on this board) The true
opposite of "common descent" is "no common descent". In other words,
it's a debate between continuity and discontinuity. And an intelligent
designer may have employed either common descent or no common
descent. Thus, while pseudogenes may pose a serious problem for
creationists (those who champion discontinuity), it poses little or no
problem for those who adhere to some notion of design that
incorporates continuity.
2. Ah, but would an intelligent designer have included such a thing as
a
"useless" pseudogene? There are two basic responses to this.
a. First, we really don't know that pseudogenes are "useless." If the
standard of "usefulness" is a coding-gene, then yes, the pseudogene is
useless. But as I have already explained, that standard is not always
true (ie, centromeric sequence). Pseudogenes may serve functions that
have nothing to do with encodement and labelling them useless may
merely be symptomatic of our ignorance. What might pseudogenes do?
Some may contain regulatory elements that may titrate factors in the
regulation of their coding "sister" and thus be involved in the
sensitive
regulation of gene expression. They may play subtle, yet functional
roles in "homology search" during recombination. They may serve as
contact points with nuclear proteins that are part of the nuclear-
cytoskeleton and thus provide fine-tuned positioning of chromosomal
domains on a very small level. They may even serve as gene
conversion templates that respond to certain cues not normally present
(it's interesting to note that this is exactly how chickens generate the
diversity of variant sites in their antibodies!). The point is that we
don't
know. And until we know, we can't simply dismiss them as "useless"
because they don't code for proteins.
b. If the Designer employed evolution, there is not much of a problem.
Keith already acknowledges that the generation of pseudogenes is a
by-product of crucial events such as recombination. Thus, while
pseudogenes may not be directly designed, they may still be the
ultimate by-product of design. Of course, one could at this point
attempt to score *theological* points about "sloppy design" or something
like that, but this would not be a convincing reply. For the Designer
would not just be designing cells and organisms, but evolution itself
would be designed. And while some may still dismiss this as "sloppy,"
that is probably our *cultural* expression whereby we equate design
with engineering. That's why we think inefficiency is evidence against
design. But if there is a Designer, I expect He/She/It designs more
like
an Artist than an Engineer. And artists are not nearly as obsessed with
neatness and efficiency as are engineers.
All in all, I would agree that pseudogenes are good evidence of common
descent, but I don't see them as evidence against an intelligent
designer.
>In attempting to rebut a passage by a Dr. Ken Miller on pseudogenes, >Behe
claims (p.226):
> The second reason why Miller's argument fails to persuade is that
>The fact of the matter is, the answer can be found in almost any
>genetics textbook. There are two major mechanisms for producing such
>duplications in biology, and both have been demonstrated >experimentally.
The two major mechanisms that Keith refers to are unequal crossing-
over and reverse transcription/integration. After briefly explaining
some of the evidence that supports the existence of these mechanisms,
Keith concludes:
"Hence we see that the available body of biological knowledge predicts
that pseudogenes are an inevitable phenomenon -- given enough time.
The complex machinery that Behe claims is necessary for pseudogene
formation not only exists, but it exists for completely different
purposes, in all living systems. "
Now, while Keith claims that "Behe is apparently completely ignorant of
the enormous amount of literature on tandem duplication, in which one
copy of a gene spawns multiple copies," I think Keith totally missed
Behe's point. Keith answer's Behe as if he needs merely to demonstrate
the existence of mechanisms that could generate pseudogenes. But Behe
didn't claim such mechanisms don't exist! He asserted that the "complex
machinery" Keith speaks of, the machinery that generates the
pseudogenes, has not be accounted for by a "Darwinian step-by-step
process." And it hasn't. Neither has Keith offered such an
explanation.
So let us consider the process of generating gene duplications. This is
an
especially important consideration since Keith relies heavily on this
process to generate his Darwinian speculations on the origin of
specified
and complex systems.
The first thing to note is that an explanation that attempts to explain
how the complex machinery needed for gene duplication arose cannot
itself invoke gene duplications. You can't assume the systematic
existence of gene duplications to explain the origin of processes that
generate gene duplications!
With that in mind, it appears safe to conclude that most, if not all,
gene
duplications are going to involve some form of recombination activity.
And both mechanisms that Keith appeals to do indeed involve
recombination. So what is involved in recombination? Well, as Behe
notes, "a dozen sophisticated proteins are required: to pry apart the
two DNA strands, to align the copying machinery at the right place, to
stitch the nucleotides together into a string, to insert the pseudocopy
back into the DNA, and much more." Let's consider the best-studied
system from E. coli.
In E. coli, recombination proceeds by at least five steps:
1. A single stranded loop is generated in the DNA.
2. A nick is introduced into one of the strands of the ss loop.
3. The nicked single-stranded invades another region of the DNA
4. A Holliday-type structure is thereby created followed by migration
of the cross-over point.
5. Endonuclease and ligation activity cleave and rejoin the Holliday
intermediate to generate recombinant products.
Let's consider what is involved in these steps:
Steps 1 and 2 are generated by the RecBCD enzyme complex. Right from
the beginning we have another candidate for irreducible complexity as
this enzyme is the product of three seperate genes - RecB, C, and D.
This
enzyme complex carries out various sequential reactions:
a. First, it has DNA-binding activity. Then, once bound it
demonstrates
helicase activity and begins unwinding the DNA. This is a process that
depletes ATP from the cell as it requires energy. The unwinding
activity generates single-stranded loops but the cut occurs near
specific
sites on the DNA - the so-called CHI sites (GCTGGTGG). When the RecBCD
complex passes by a CHI site, a specific nuclease activity generates a
single-stranded cleavage. At this point, we need to ask what good is
DNA-binding activity without the helicase and nuclease activity? And
what good is the helicase activity without the nuclease activity? In
fact,
since helicase consumes ATP, organisms that simply unwound their DNA
for no reason would be at an energetic disadvantage compared to those
that didn't engage in this line of Darwinian-experimentation.
Furthermore, that nuclease activity better be controlled. If you nick
the
same strand of DNA not far upstream of the intial cut, you've just
cleaved a hunk out of your DNA. Finally, what good is a single-stranded
region if the remaining steps (3-5) are not in place?
b. In step 3, single-strand invasion is mediated by RecA activity.
This
too is a reaction that consumes more ATP. Thus, unless this activity is
already part of a functioning system of recombination, it is just one
more way to flush ATP down the toilet and would thus put the
organism at an energetic disadvantage.
[Of course, some might note that RecA is also involved in the SOS
response of E. coli, thus the strand-invasion activity may have evolved
secondarily. But this would simply mean we would have to look at the
SOS response and we'd soon find that it too is a complex and intricate
system.]
Step 4 proceeds because of RuvA and RuvB activity. RuvA is the
protein that recognizes the Holliday structure and RuvB also has
helicase activity. So here is a system that needs RecA, RecB, RecC, and
RecD to generate the Holliday structure. But what good is such a
structure without RuvA and B? And what good is RuvA and B without
RecA,B,C, and D?
Step 5 proceeds because of RuvC which specifically recognizes the
Holliday recombination intermediates and makes the appropriate cuts
with an endonuclease activity to generate the recombinant molecule.
How did this system arise by Darwinian evolution? Instead of accusing
Behe of ignorance, Keith should have tackled this system and outlined
its Darwinian evolution (of course, without begging the question by
appealing to gene duplications). Here is a system where RuvC activity
is
of no obvious function without RuvA and B. In fact, it would be
interesting to know how many amino acid residues are important for
the site-specific recognition of RuvC - could *it* have been built up
one
amino acid substitution at a time? And Ruv A and B are of no
obvious function without RecA,B,C, and D.
So let's proceed from the other end. First, we need that DNA-binding
activity. This by itself may seem like no obstacle given the
commonality of such proteins, but if we were to take a closer look at at
generalized protein-nucleic acid interactions, we would find that
several
amino acids are usually involved in the interaction in an irreducibly
complex manner (for example, in zinc finger proteins, 3/4 or 2/4 of the
zinc ligands are no more functional than 1/4 or 0/4 zinc ligands).
But let's say we have our DNA-binding protein. What is it doing aside
>from binding? If it is doing nothing, then it would acquire mutations at
the rate of a pseudogene as it would have no function. Thus, selection
would not weed out deleterious mutations. This in turn would mean
that an assigned function would have to develop very rapidly.
Okay, so some helicase activity becomes associated with it. But what
selective advantage does this impart? In fact, since helicase activity
consumes ATP, a clear disadvanatage emerges in the random,
nonpurposeful, unwinding of DNA. Of course, helicase activity is
involved with DNA replication (and that's another candidate for
irreducible complexity), but that's a different gene (DnaB). Finally,
we
have the endonuclease activity, but what selective advantage is
imparted by putting random nicks in the DNA?
Put simple, as a *whole* system, all the players work together to impart
a partcular function. But I don't see how you can tease it apart in a
darwinian fashion. In *that* was Behe's point.
Keith concludes this section by saying:
>Evolution predicts that pseudogenes will be born, will decay or be
>deleted, and that common ancestors will often share common
>pseudogenes. Evolution also provides a means for distinguishing
>pseudogenes from real genes (by counting synonymous vs non->synonymous
>mutations).
It's pretty easy to say that evolution would predict pseudogenes will be
born after we know they exist. But can Keith cite any evolutionist who
actually made this prediction before pseudogenes were known to exist?
I would agree, however, that pseudogenes are indeed good evidence for
evolution. But I don't agree that they damage Behe's case.