Back to:    Aids    Home

MUTATIONS: DAMAGE AND REPAIR OF DNA
     

              Endogenous (or spontaneous) DNA damage
Instability of the DNA molecule 
     Premutagenic base changes 
          Tautomeric shifts 
     Loss of bases 
Errors in replication
     Strand slippage 
          Trinucleotide repeat diseases   

             Exogenous (or environmental) DNA damage 
Ionizing radiation
Ultraviolet radiation
Chemical agents  
     Alkylating agents 
     Cross-linking agents and bulky adducts 
     Inactive chemicals that are metabolized to reactive mutagens 

                        DNA repair systems
Reversal of damage
Nucleotide excision repair
Mismatch repair
Other repair processes

 
Mutations, hereditary alterations of the DNA sequence, develop in a two-step process.  In the first
step, sometimes called DNA damage, DNA lesion or premutation, the composition or structure of
the DNA is modified.  The second step, fixation of the mutation, is the re-establishment of the
normal DNA structure but with a new DNA sequence that all progeny cells inherit (Note that the
expression "a mutation is fixed", used in this sense, means that a mutation will be transmitted,
and not that one has been corrected).  Once a premutation occurs, whether it results in a mutation
or not depends on which of two cellular processes gets to the damage site first: the next round of
replication or a DNA repair system   although some types of damage act as replication blocks,
such that DNA synthesis does not proceed through them until the damage is repaired.  Any one of
several repair processes may be involved, and the usual outcome is that the normal chemical
structure of the DNA is restored and the original nucleotide sequence is recovered   although in
the repair of some forms of damage, such as multiple double strand breaks, the chemical integrity
of the DNA may be restored, but the original sequence is not. 
     Mutations may occur in the germ cells and be transmitted to the next generation or they
may occur in a somatic cell and affect only a patch of tissue.  DNA damage may have exogenous
causes   mutagenic chemicals, ultraviolet light, X-rays   or endogenous causes, i.e., they occur
"spontaneously" in the absence of an external insult.  It is useful to consider these two sources
separately. 

              Endogenous (or spontaneous) DNA damage
There are three main sources of DNA damage in the normal course of biological processes: 1)
instability of the DNA molecule, 2) errors in replication (DNA synthesis), 3) errors in
recombination.  

Instability of the DNA molecule.
In some respects, DNA is a remarkably stable compund; it is very resistant to hydrolysis of the
phosphodiester backbone   thanks to the absence of a 2' hydroxyl   and it is also resistant to
denaturation.  Several bonds, specially in the nitrogen bases, however, are quite reactive; chemical
processes that are particularly relevant to human genetics are: tautomeric shifts, deamination and
oxidation of bases, and hydrolysis of glycosidic bonds with the concommitant loss of bases and
creation of apurinic/apyrimidinic (AP) sites.  We will consider these types of damage in turn.
     Premutagenic base changes.  These are changes in the chemical structure of the bases
that alter their capacity for hydrogen bond formation.  We will discuss briefly tautomeric shifts. 
Other changes include deamination, oxydation and methylation of bases.
          Tautomeric shifts.  The bases T (at position C4) and G (at C6) have oxygen
substitutions on their rings.  The favored state of these oxygen atoms is in the form of carbonyl
groups (C=O), which act as acceptors of protons to form H bonds.  A small fraction of molecules
can be found in the unstable hydroxyl form (OH)   called a tautomeric shift   for which the
proton is supplied by the neighboring N.  If this transition occurs during the replication process,
the shifted base will tend to mispair (*T G and *G T, where * marks the shifted base and the
symbol "C G" is used to indicate nucleotides on complementary strands attached to each other by
H-bonds.)  If left uncorrected these mispairs would give rise to transitions.  A transition is a
mutation in which a purine is replaced by another purine, while in a transvertion a purine is
replaced by a pyrimidine, or viceversa.
          Loss of bases.  Hydrolysis of the glycosyl bond between carbon 1' and the nitrogen
base occurs in native DNA much more frequently in purine than pyrimidine nucleotides, and for
this reason  these lesions are sometimes called apurinic sites; the abbreviation AP, however,
stands for apurinic/apyrimidinic site.  The consequence of unrepaired AP sites is a block in DNA
replication.  

Errors in DNA replication
DNA replication is involved in mutagenesis in two ways.  One is the fixation of premutation
lesions (of the type described in the previous section, and others) into base substitutions, when the
replication fork reaches a damaged base before repair occurs.  Replication is also involved in
creating premutation DNA damage: errors of replication include mis-incorporation of nucleotides
leading to base substitutions, and strand slippage leading to small deletions and duplications.  If
the accuracy of base pairing were entirely dependent on the H-bonding abilities of the four bases,
in the absence of the specificity provided by DNA polymerase and postreplication repair
mechanisms, misincorporation of nucleotides might be as high as 1-10 %.
     Strand slippage.  This type of misalignment occurs in clusters of repeating sequences
with several types of mutations as possible consequences.  Fig. 7.9 illustrates the production of a
one base deletion as a result of slippage in the template strand.  Similar slippage of the priming
strand results in base duplications.  If the repeating unit is longer than a single base the deletion or
duplication of longer segments is possible.  The best documentation of this type of mutations in
humans is provided by the trinucleotide repeat expansion diseases. 
          Trinucleotide repeat diseases (TRD).  More than ten loci have been identified, in
which expansion of trinucleotides repeats interfered with normal gene function.  The location of
the repeat within the gene is not constant in the various loci.  In most of these genes the number
of repeats in the normal population is highly polymorphic, which suggests that mutations changing
the number of repeats   both increases and decreases caused by slippage during replication 
occur rather frequently.  Nevertheless, the mutation rate is not so high as to prevent the stable
transmission of alleles within a family.  Although increases and decreases in the number of repeats
have been detected, there seems to be a slight overall tendency toward expansion.  When repeat
expansion reaches a certain level, however, the element becomes unstable, expansion accelerates,
and it reaches the threshold at which it causes a phenotype usually in one generation.  In some
cases, if the number of repeats is only slightly over the threshold, symptoms may be less severe, or
appear later in life.  Because of the inexorable expansion of such unstable repeats, however, any
offspring of such a patient would carry significantly larger repeats, accompanied by more severe
symptoms, or earlier appearance of the disease.  This peculiar aggravation of a genetic disease
from one generation to the next was quite puzzling until the molecular basis of these mutations
was discovered. 
     The largest group among the trinucleotide repeat diseases is that of the neuropathies
associated with polyglutamine stretches.  Two-to-three fold expansion of these polyglutamine
tracts seems to produce a gain-of-function, therefore dominant, allele. 

Huntington's disease (HD).  HD is an autosomal dominant, fully penetrant, neurodegenerative
disease.  This is one of the rare genetic conditions in which the heterozygous and homozygous
mutant phenotypes are indistinguishable (Fig. 1-3).  HD symptoms usually start in the thirties and
forties and culminate with death 10 to 20 years later; they include progressively severe involuntary
movements and mental disorders caused by neuronal attrition.
     HDH, the gene responsible, has 67 exons that extend through 180 kb; huntingtin, the
corresponding protein is 3144 amino-acids long and has no obvious sequence similarity to other
proteins of known function.  Seventeen codons downstream from the AUG site there is a short
stretch of 10-30 CAG repeats encoding polyglutamine, with most alleles carrying less than 24
repeats.  Expansion beyond 40 repeats is responsible for the majority of HD cases; no HD cases
have fewer than 30 repeats.  There is an inverse relationship between the age of onset of the
disease and the number of repeats.  Individuals with 30 to 40 repeats are usually asymptomatic
but very likely to transmit further expansion to the progeny.  In alleles of this intermediate range,
changes in repeat lengths from one generation to the next are much greater in males than females;
cell divisions in the male germ line can lead to a doubling of number of repeats, while in the
maternal line changes are usually limited to increases or decreases of less than four repeats.  HDH
has a high level of expression in neurons and several other tissues.  Heterozygous for null
mutations do not present any of HD symptoms, but no homozygotes for such nulls are known. 
Expansion of the polyglutamine tract is thought to be a gain-of-function mutation; perhaps by
increasing the interaction ("stickiness") of huntingtin with other cellular components.  The end
result, at the cellular level, is the stimulation of apoptosis,or programmed cell death in affected
neurons. 

             Exogenous (or environmental) DNA damage
Mutagenic agents are usually divided into physical and chemical.  Some environmental agents
increase the incidence of damage that occurs spontaneously, while others have unique and specific
effects.  
Ionizing radiation
X-rays and  -rays represent the very short end of the wavelength spectrum of electromagnetic
radiation; they interact with water and biological molecules to create reactive ions and free
radicals. 

Ultraviolet radiation
UV light is electromagnetic radiation with wavelength between 100 and 400 nm, it is non-ionizing
and is significant as a mutagen mainly in the wavlength around 254 nm, the absorption peak of
DNA nitrogen bases.  Sunlight is the major source of UV radiation, and although absorption of
wavelengths below 320 nm by stratospheric ozone obviates the vast majority of the problems that
sunlight might pose for humans, its pre-mutagenic action is quite significant.  Damage is probably
due to a combination of the small amount of 254 nm light that reaches the surface of the earth and
attenuated but non-zero efficiency of longer wavelengths to cause DNA damage.  
     As opposed to the damage caused by ionizing radiation, UV light's effects result from
direct interaction with DNA.  UV light approaching 254 nm causes adjacent pyrimidines in DNA
to form covalent bonds (Figs. 7-11a).  DNA synthesis seems to be able to proceed across
unrepaired pyrimidine dimers, but in an error prone process that has a preference for inserting A
on the complementary strand   thus resulting in frequent G C   A T transitions.  Although
pyrimidine dimers are the main premutation lesions in UV irradiated DNA, other photoproducts
are found, including the cross-linking of DNA to proteins, and strand breakage.
Chemical agents
Numerous chemicals are capable of reacting with DNA, the main goups include:
     Alkylating agents.  These are reagents that can modify most N and O atoms in the four
bases by addition of alkyl groups.  Alkylated bases induce a weakening of the glycosyl bond and
often result in AP sites. 
     Cross-linking agents and bulky adducts.  These are chemicals that cause significant
distortion of the secondary structure of DNA.  Cross-linking can be intrastrand or interstrand;
interstrand links effectively block DNA replication and agents that promote them, such a nitrogen
mustard, mitomycin, platinum derivatives and psoralen derivatives have been used in cancer
chemotherapy. 
     Inactive chemicals that are metabolized to reactive mutagens.  These have been
recognized for their carcinogenic effect in the tissues where the conversion occurs (extensive
studies have shown that carcinogenic compunds are also mutagenic).  Polycyclic aromatic
hydrocarbons, produced during coal combustion (and present in cigarette smoke), and aflatoxins,
produced by some fungi, are examples of such metabolically activated agents.

                        DNA repair systems
There are many enzymes whose function it is to restore the molecular integrity of the DNA
molecule and to preserve its sequence.  Some of them are highly specific and correct only one
type of damage   such as removal of methyl groups from O6-mG   while others are more
general and can repair an entire category of damages   such as a deformed double helix due to
cross-linking or bulky modification of bases.  Many repair enzyme are remarkably conserved in
function, and to a certain extent in amino acid sequence, in all life forms. 

Reversal of damage
These are reactions that neatly reverse the premutation chemical change.  One example is the very
efficient removal of methyl groups from O6-mG by the enzyme O6-methylguanine-DNA
methyltransferase (O6-MGT) also known as the Ada protein. 
     The classic example of damage reversal   and the first form of DNA repair found, in the
1940s   is photoreactivation (PR): the light-dependent monomerization of pyrimidine dimers by
photolyase or PR enzyme.  PR is a widespread process among prokaryotes and eukaryotes,
which, however, does not seem to occur in placental mammals.

Nucleotide excision repair (NER)
This process corresponds to the dark repair (as opposed to light repair or PR) of pyrimidine
dimers originally described in E. coli.  NER exists in all organisms; it is not as specific as the
glycosylase-initiated reactions, but rather, it is a more generic repair system.  NER is the primary
repair system involved in the correction of damage that distorts the double helix, such as
pyrimidine dimers or bulky adducts   for example benzo[a]pyrene-guanine from smoke, or
thymine-psoralen and guanine-cisplatin, from chemotherapeutic drugs   but it can also act on
non-distorting damage such as methylated, oxidized, and even mismatched bases.  NER is carried
out by a complex multienzyme aggregate.  It includes proteins that recognize and bind to
damaged sites, specific excision nucleases that introduce single strand breaks on either sides of the
damaged nucleotide(s), and helicases that unwind the DNA.  A 29-np oligonucleotide, including
the damaged nucleotide is thus removed, and DNA synthesis and ligation then repair the gap (Fig.
7-22 and lecture figure).  That, in the absence of photoreactivation, the brunt of thymine-dimer
repair falls on NER is illustrated by the genetic disease xeroderma pigmentosum (XP), a condition
caused by deficiencies in NER.  XP is a very rare disease characterized by the occurrence of
dermatitis, excessive freckling and numerous skin tumors on parts exposed to sunlight (Fig. 7-27
and 7-28).  Seven genes have been identified, which can cause the XP phenotype, NER is,
however, a very complex process kown from biochemical studies to involve at least twelve
proteins. 

Xeroderma pigmentosum (XP) and nucleotide excision repair.  XP behaves as an autosomal
recessive condition, but it is heterogeneous in origin; that is to say, it may be caused by mutations
in any one of several genes.  When it was discovered that cultured fibroblasts from XP patients
showed increased sensitivity to UV radiation, the suspicion that these patients were deficient in
some form of DNA repair was reinforced; that observation also led to a convenient in vitro assay
for genetic studies.  One group of studies consisted of fusing cells from many patients in various
pairwise combinations.  The surprising result was that many of the cell hybrids had normal
tolerance to UV light.  This genetic complementation indicated that the same clinical condition
arose from mutations in different genes.  Cell hybrids that were as sensitive as the parental cells,
showing failure to complement, pointed to mutations that were in the same gene.  In this fashion
seven complementation groups were defined XPA through XPG.  

Mismatch repair
Mismatch repair is a relatively specific DNA repair system that acts on endogenous damage.  It
detects and corrects the incorporation of mismatched bases and the single-stranded loops created
by slippage along stretches of repeated DNA sequence DNA looped-out single strands, specially
those that occur in connection with DNA replication. 
     The type of damage repaired by mismatch repair is unique in that all the elements present
  such as T G mispairs in newly synthesized DNA   are normal components of the DNA
molecule and there is, therefore, ambiguity concerning the original information: should the T be
replaced with a C or the G replaced with an A?  Both replacements would correct the damage,
but while the proper correction would recover the original sequence, the improper one would fix
a mutation.  Establishing which is the right replacement requires that the repair system be able to
identify which is the original, template, strand and which is the priming or newly synthesized one. 
In E.  coli, DNA is methylated at GATC sites some time after replication.  Thus, newly
synthesized DNA is hemimethylated, the priming strand being as yet unmethylated, and this
asymmetry allows the repair system to recognize the strand with the error and replace it.  It
appears, however, that methylation is not the signal used in mammalian genomes to distinguish
between newly synthesized and template strands (7-25).

Other repair processes
There are other, less well-defined, repair processes that may be functionally analogous to bacterial
error-prone or SOS repair and post-replication repair.  In general, these are systems that restore
the structural integrity of the DNA even when the informational content is compromised.  Post
replication repair is defined by some mutant rodent cell lines that display slightly increased
sensitivity to UV light, but reduced UV mutability.  That is to say, normal cells seem able to 
complete repair of UV damage by a process that is prone to cause mutations; a deficiency in some
step in this process renders the cells more sensitive to UV radiation   because, unrepaired, the
lesions are cytotoxic   but less mutable.  Similar characteristics are found in cells of Fanconi's
anemia patients, an autosomal recessive disease characterized by increased predisposition to
cancer.

                 Somatic mutations and mosaicism
In the traditional view of mutations, these can be either somatic or germinal.  The former give rise
to patches of mutant tissue in individuals, which are called mosaics, and, unless the mutation
affects skin morphology or pigmentation, somatic mutations usually go undetected.  The size and
shape of a mutant patch depends on the stage in development when the mutation occurred (Fig. 7-29 and 7-30). 
     Germline mutations, on the other hand are hereditary and transmitted to the progeny in
Mendelian fashion.  This dichotomy was a very formalistic view by necessity, since not much
evidence was available to shed light on the origin of mutations, or the cells and individuals in
which they occurred.  Advances in cellular and molecular techniques have made possible studies
in this very area.  
     Those studies have blurred the distinction between somatic and germinal mutations.  The
notion of a single, ill-starred, mutant gamete in a pool of normal congeneres is probably largely
incorrect   except for non-disjunction and other meiotic errors that are only received by the
gamete(s) that result from one particular meiocyte.  Many hereditary mutations seem to start out
in mosaic individuals who carry the mutation in a significant fraction of germline cells or
sometimes in mixed patches of germline and somatic cells. 
     Examples of mosaicism are available in de novo cases of several genetic diseases; as in the
case of an unaffected man who transmitted the same partial deletion of the Duchene muscular
dystrophy  gene   X-linked recessive mutation   to two daughters, but not to three others. 
There are also cases in which the mutation can be detected in the normal parent of an affected
child.  For example, in a family in which two normal parents had two children with
neurofibromatosis type 1   autosomal dominant mutation   and a 12 kb deletion of NF1 was
detected in 10% of paternal sperm.