Begin 02/25/16 (#14)

[Note: some material overlaps ENVR 430 and was condensed in lecture.] We
had concluded with the flow diagram of eukaryotic BER, which is a workhorse in repair in eukaryotic systems much more important than in
prokaryotes. A number of glycosylases that specifically target commonly
occurring lesions have been identified and several of these have been
crystallized as complexes with damaged DNA. By determining the residues
contacting the damaged bases it is possible to gain insight into the
mechanisms of deglycosylation. Different glycosylases break the glycosidic
bond by different mechanisms; however, a unifying aspect of all mechanisms
illustrated in the following series slides is rotation of the target base along
with its sugar completely out of the helix. The extrahelical rotation is
postulated to be the way in which the glycosylase interrogates DNA and
identifies mismatches or other damage.
Given the frequency of GT mismatches because of guanine keto-enol
tautomerism, an important eukaryotic repair targets GT mismatches in
mammalian cells and operates to remove the T in favor of restoring the GC
pair. (Statistically, this correction would be the appropriate repair.) There are
two glycosylases that can accomplish this repair, one is MBD4 (methyl-CpG
binding domain protein 4) and TDG (thymine DNA glycosylase). In addition to
the C-terminal glycosylase domain, MBD4 contains methyl-CpG binding
domain, and can excise 5-hydroxymethyluracil in a G5-HMU mismatch as
well as the T in a GT mismatch or a U in a GU mismatch. The presence of
the

Me

CpG-binding domain and the glycosylase activity of MBD4 with respect

to 5-hydroxymethyluracil points to a dual role for this glycosylase in

rearranging the epigenetic landscape, which we shall discuss later. TDG
catalyzes the removal of the T in the mismatched GT pair by activating a
water molecule to serve as a nucleophile. The next overhead shows a
structure of human TDG, at a GT mismatch, after the glycosidic bond has
been broken and the mismatched T has been removed. This has created an

abasic site (abasic sugar in blue), with the sugar still rotated out of the helix.
This slide along with series of slides following should help to clarify what is
going on. In the slide captions, remember the nomenclature convention that
a lower case h prefix refers to a human protein.
Slide 1. On this slide, you can see the abasic sugar in blue, the partner G of
the excised T in purple, and contacts with the undamaged strand made by A
(green) and R (black) in insertion loop of the glycosylase.
These two contacts provide the selectivity for the GT in the CpG context.
Series 3 5 to illustrate in stages
The next two slides show the same structure in layers, to make the
identification of key interactions easier.
Slide 3: This slide shows the DNA in stick form to making it easy to see the
rotated abasic sugar The phosphodiester bonds connecting the abasic sugar
in the backbone of the DNA strand were not represented by the software, but
I have filled them in here.
4: The next slide shows the insertion loop of the glycosylase in blue
containing the R and A bases that force the rotation of the dT out of the
helix.
5: The next slide shows the color-coded components all added back into the
picture.
A structure of the human 8-oxoguanine glycosylase or hOGG1 is shown on
the next slide. As the name indicates, hOGG1 is responsible for excising the
oxidized base 8-OG, a frequent promutagenic lesion resulting from oxidative
stress. I showed how Hoogsteen pairing of 8-OG caused a mutation, and I
think that 8-OG was mentioned in ENVR 430 as well. As a reminder, the
structure of 8-oxoG is at the bottom left.

Slide 6. hOGG1 glycosylase,.

Since hOGG1 recognizes a major product of oxidative damage, it has a very
important repair function. In the crystal structure, the oxidized base is still
attached to the sugar and the two views show the rotation of the 8-oxo-dGuo
out of the helix. Left is the structure with DNA in wire frame representation
and right is the same structure with DNA in ribbon form.
The next slide shows the mechanism by which the oxidized base is removed.
Slide 7. Structure of hOGG1 K249Q.
The bright blue residue is a basic amino acid which displaces the 8-OG from
the sugar C1 by nucleophilic attack to generate an apurinic site. This
bimolecular type of displacement is called an SN2 reaction. The aa in the
wild-type glycosylase is the strongly basic amino acid lysine, which we
discussed in the context of DNA binding, which has a 5-carbon chain with a
terminal NH2 group (lover left). In this situation, however, the amino group is
acting as nucleophile to displace the 8-OG from the sugar. To obtain the
crystal structure, though, lysine has been replaced with glutamine, which
also has a terminal amino group, buit cannot participate in the nucleophilic
displacement for two reasons: first, the residue chain is 1-carbon shorter
than Lys, having a 4-carbon chain instead of a 5-carbon chain, and NH2
cant approach C1 closely enough to act as a nucleophile and secondly, the
terminal NH2 belongs to an amide rather than an amine, the amido group
(C=O)-NH2 is less basic and the greatly diminished nucleophilicity allows a
stable intermediate for structural determination.
The next overhead shows DNA uracil glycosylase complexed to a sequence
of DNA containing dU, which is not a DNA base, but generated with high
frequency from the spontaneous deamination of dCyd. The resulting GU
lesion causes a GC AT transition mutation if not repaired. Since
deamination of dCyd is another frequently-occurring lesion its repair is

important. The same pattern of rotation out of target nucleoside out of the
helix is evident: illustrated by the right hand panel U and the sugar are
highlighted in in stick form.
Slide 8. Uracil glycosylase.
In order to have a stable structure, the complex was actually crystallized with
a nucleotide analogue in which sugar has been substituted by 2-deoxy-4thioribose (O in the sugar ring replaced by S). The right panel shows clearly
the sugar analog is rotated out of the helix. In this structure, the glycosidic
bond has been cleaved, but uracil is still present at the site. The hydrolysis
product 1-hydroxy-substituted sugar is evident. In contrast to the
mechanism of hOGG1, this uracil glycosylase does not remove U by an SN2
attack by a basic residue but appears to induce an SN1-like hydrolysis that
is an unassisted removal of the base - by distorting the deoxyribose ring and
stabilizing the transition state glycosidic bond cleavage by formation of a
network of H-bonds between the glycosylase and both the sugar and the
departing base. The left panel shows the same structure, but I have
processed it to show only the disconnected uracil base.
Finally, here is the structure of alkyladenine glycosylase, for which the
acronym is AAG, that recognizes and removes a variety of alkylated purine
bases, including: 3-MeAde, 7-MeGua, hypoxanthine, which is the product of
deamination of Ade, and the tricyclic base adduct 1,N6-Ade. 1,N6-Ade is a
lesion generated by the activated metabolite of vinyl chloride, and also by
oxidative stress, either endogenous or induced by xenobiotic chemicals. The
1,N6-Ade lesion is strongly mutagenic and is directly connected to a
carcinogenic endpoint.
Slide 9. Alkyladenine glycosylase.
In this overhead, the glycosylase is shown as a complex with the tricyclic
base 1,N6-Ade, the structure of the etheno base at right. The pattern of

rotation of the damaged deoxynucleotide out of the helix is obvious here as

well.
AAG removes the modified adenine by yet another mechanism it breaks
the glycosidic linkage by glutamic acid bridging through an H2O molecule
acting as a nucleophile to displace the base, rather than direct attack by a
protein residue.
Slide 10. Hydrolysis mechanism for AAG.
This structure appears in Genes XI on p. 405, Fig 16.15, except that the
damaged base is incorrectly identified as 3-Me-A and survived two editions.
The tricyclic base is really obvious in the figure.
We have essentially already described the broad picture of repair of abasic
sites in the BER pathway, for the special case in which the abasic site is
generated by glycosylases, and the next slide is a repeat of the scheme from
the slide at beginning of lecture with a few additional details. It is important
to note that the same pathway applies to any abasic site in eukaryotes,
whether it is generated during BER by a glycosylase or by depurinating
lesions. As I said previously, the repair usually follows left hand branch,
which is very short patch repair.
Slide 11. From EMBO J, 2001 on pol .
A refinement in detail in this slide is that after an incision is made to the 5
side of the abasic site by APE1 the nick is followed by excision of the abasic
site through the action of pol by a base-catalyzed -elimination
(mechanism on next slide).
Slide 12. -elimination (fix slide to conform to BER)
The double-headed arrow on slide indicates the nick by APE-1. The elimination formally involves the base-catalyzed abstraction of a proton from
C2' followed by oxy anion elimination from C3'. Note especially generation of

a single-strand break during the repair. The elimination of the abasic site is
followed by replacement of the single nucleotide which is also done by pol ,
and then the nick is sealed by a ligase.
If the abasic lesion is resistant to removal by -elimination, for example if the
sugar is modified, then the right hand branch, which is long patch repair, is
followed.
Slide 12.
The displacement and strand synthesis is accomplished by pol or pol /
and the displaced strand is removed by FEN1, the Flap Elimination
Endonuclease (we have encountered as the endonuclease involved in the
removal of RNA primers during eukaryotic replication).
If damage is not repaired by one of the excision systems prior to replication,
gaps may be left opposite lesions following replication. Repair systems
dealing with this situation are post replication repair systems and were
detected initially in prokaryotes when it was discovered that E. coli deficient
in excision repair could still repair UV damage. Because post replication
repair involves retrieval of the correct genetic information corresponding to
the gaps opposite lesions, it is called retrieval repair. This repair pathway
was not covered in ENVR 430. The retrieval is accomplished by transfer of
the missing DNA from the intact homologous duplex generated during
replication. The general strategy of the retrieval repair pathway for single
strand gaps is outlined in the following overhead:
Slide 13. Retrieval repair strategy.
A blocking lesion in a parent strand (in panel 1) results in a gap following
replication (in panel 2). The cell fills the gap by transfer of the appropriate
sequence from the intact homologous duplex, in panel 3 followed by filling of
the gap created by the transfer using the undamaged template in panel 4.

In prokaryotes, retrieval repair is mediated by a protein belonging to a family

of genes called rec (from recombination), and is initiated by the RecA
protein, which we shall see is a highly versatile protein. The next slide is a
model of how Rec A functions in retrieval repair.
III, Slide 14. Nature, model for retrieval repair by RecA.
A complex of RecA molecules represented by the cylinder, bind at the gap
and promote homologous pairing of the damaged and intact replicated
duplexes (homologous pairing means that the same sequences are lined up
adjacent to each other). Here the gap has been created by a TT dimer. The
undamaged duplex is nicked and the 3 terminus of the nick is transferred by
RecA onto the gap. At the same time, the 3 terminus of the gap is
transferred onto the undamaged duplex to create a cross-over feature called
a Holliday junction. The junction can be driven along the duplexes in a
process called branch migration and at some point the branch is cut (arrow
heads) and the nicks sealed to complete the repair. (The dotted line is where
the gap created by strand transfer has been filled in on the undamaged
template. Also note that the damage has not been removed, but simply
bypassed for repair at a later time.)
The next slide shows that once exchange has been initiated by RecA, the
products of the E. coli family of ruv genes are involved in completing the
strand exchange driving the branch migration. A multimeric complex of
proteins comprised of a tetramer of RuvA which contacts all four strands and
hexamers of RuvB, which are the same proteins that are we encountered in
excision repair, can displace RecA and drive the migration of the crossover as
illustrated in the next overhead.
III, Slide 15. Holliday branch migration
The ruv gene product, RuvC, which is associated here with the Holliday
junction, recognizes the consensus tetranucleotide sequence 5-(A/T)TT(G/C)-

3 makes symmetric nicks in the Holliday junction between the third and
fourth positions of the tetranucleotide sequence and releases the duplexes.
The resolution of the helices can be achieved in two ways, depending on how
the nicks are made by RuvC. The bottom panel illustrates nicks that result in
recombinant DNA (except for the region where the cross-over and migration
occurred = heteroduplex region), the duplexes on either side of the
heteroduplex region have been exchanged). Cuts at 90

to those illustrated

(arrows at right) result in regeneration of the original helices with a

heteroduplex region where strands were exchanged. At this point, all the
elements of repair are in place and the complete pathway is summarized on
the next slide.
III, Slide 16. Summary of retrieval repair
This mechanism operates in eukaryotes as well. RAD51 and DMC1, which
function in a manner homologous to RecA have been characterized in yeast
and other eukaryotes. RAD51 is involved in the search for the homologous
regions and strand pairing stages of the retrieval process during mitosis (cell
duplication). DMC1appears to have overlapping functions with RAD51, but its
activity is restricted to meiosis.