Information Technology Reference
In-Depth Information
telomere binding protein in
Oxytricha nova
is
present in the micronucleus in the permuted order 1-3-5-7-9-11-2-4-6-8-10-12-
13-14 [12]. The gene encoding DNA polymerase
The gene encoding the
α
in
S. lemnae
is apparently
broken into more than 48 MDSs scrambled in an odd/even order [10], with
14 MDSs inverted on the opposite strand of another 24 MDSs, 9 additional
MDSs on a separate locus, and 1 MDS not present on either locus. MDSs have
to be reordered and intervening IESs eliminated to construct the functional
macronuclear gene.
The process of unscrambling the micronuclear DNA into the macronuclear
genes is ostensibly aided by the presence of short pointer sequences present at
the junctions between MDSs and IESs. More precisely, at the junction between
the
n
th MDS and the adjacent IES following, it there is a sequence (of 9-
13 bp in actin I and 6-19 bp in
α
telomere binding protein) that is repeated
somewhere else in the gene—namely, at the junction between the
(n
+
α
1
)
st
MDS
and the adjacent IES preceding it. After aligning two pointers, homologous
recombination between them would then join MDSs
n
and
(n
+
1
)
in the correct
order and eliminate one copy of the pointer.
This process can be viewed as a computation solved during gene unscram-
bling by homologous DNA recombinations. If we assume that the cell's bio-
chemistry can identify those DNA segments that represent pointers, and if the
pointer pairs were unique in the micronuclear DNA, then one could argue that
the ciliate is solving the computational problem of sorting a permuted sequence
in the correct order. However, the computational problem facing the cell is
much more complex given that some pointer sequences occur more than 13
times in a single gene (e.g., DNA polymerase
in
S. lemnae
). Taking into
account the multiplicities of each pointer (the raw number of occurrences of
the sequence representing the pointer in the micronuclear sequence), the num-
ber of combinations the cell would need to explore in order to arrive at the
correct solution could be greater than 14 trillion for DNA polymerase
α
in
S.
lemnae
[2]. Clearly, even assuming a priori knowledge of the pointer sequences,
blind searching of matching pointer pairs is not a realistic explanation of gene
unscrambling.
Other factors such as knowledge of which DNA sequences represent MDSs,
IESs, and pointers, together with geometric folding of the micronuclear DNA
that brings corresponding pointers together, have been suggested as mechanisms
of gene unscrambling [17]. Alternative or complementary information guid-
ing unscrambling may be the presence of contexts that flank correct matching
pointer pairs and that might be responsible for solving this seemingly difficult
computational problem [17, 18]. The details of the gene rearrangement pro-
cess are still elusive. In the following sections we explore formal models for
the homologous recombinations that lead to gene unscrambling in ciliates and
investigate their computational power.
α
