Genomics- The Science and Technology Behind the Human Genome Project. Charles R. Cantor, Cassandra L / genomics1-10 / 1

Figure 1.3

Symmetry and pseudosymmetry in DNA double helices.

(a) The vertical lines indicate

hydrogen bonds between base pairs. The base pairs and their attached backbone residues can be

flipped by a 180° rotation around an axis through the plane of the bases and perpendicular to the he-

lix axis (shown as a filled lens-shaped

object).

(b) Conventional way of representing a double-

stranded DNA.

(c) Example of a

DNA hairpin.

(d) Example of a DNA pseudoknot.

In a typical DNA duplex, the phosphodiester backbones are on the outside of the struc-

ture; the base pairs are internal (Fig. 1.4). The structure of the double helix appears to be reg-

ular because the A–T base pair fills a three-dimensional space in a manner similar to a G–C

base pair. The spaces between the two backbones are called

grooves.

Usually one groove is

much broader (the major groove) than the other (the minor groove). The structure appears as

a pair of wires held fairly close together and wrapped loosely around a cylinder. Three major

classes of DNA helical structures (secondary structures) have been found thus far. DNA B,

the structure first analyzed by Watson and Crick, has 10 base pairs per turn. DNA A, which is

very similar to the structure almost always present in RNA,

has 11 base pairs per turn. Z

DNA has 12 base pairs per turn; unlike DNA A and B, it is a left-handed helix. Only a very

restricted set of DNA sequences appears able to adopt the Z helical structure. The biological

significance of the range of structures accessible to particular DNA sequences is still not well

understood. Elementary texts often focus on hydrogen bonding between the bases as a major

force behind the stability of the double helix. The pattern of hydrogen bonds in fact is re-

sponsible for the specificity of base–base interactions but

not their stability. Stability is

largely determined by electrostatic and hydrophobic interactions

between parallel overlap-

ping base planes which generate an attractive force called

base stacking.

METHYLATED BASES

5

Figure 1.4

Schematic view of the three-dimensional structure of the DNA double helix. Ten base

pairs per turn fill a central core; the two backbone chains are wrapped around this in a right-handed

screw sense. Note that the A–T, T–A, G–C, and C–G base pairs all fit into

exactly

the same space

between the backbones.

METHYLATED

BASES

DNA from most higher organisms and from many lower organisms have additional ana-

logues of the normal bases. In bacteria these are principally

N6 -methyl A and 5-methyl C

(or 4-methyl C). Higher organisms contain 5-methyl C. The presence of methylated bases

has

a

strong

biological effect. Once in place, the methylated bases are apt

to

maintain

their

methylated status after DNA replication. This is because the hemi-methylated du-

plex

produced

by one round of DNA synthesis is a far

better substrate for the enzymes

that insert the methyl groups (methylases) than the unmethylated sequence. The result is a

form

of inheritance (epigenetic) that goes beyond the

ordinary DNA sequence: DNA

has

a memory of where it has been methylated.

The role

of these modified bases is understood in

some detail. In bacteria these

bases

DNA,

which is methylated,

and the DNA from an invading bacterial virus (bacterio-

phage)

or plasmid, which is

unmethylated and can be selectively destroyed. Unless the

from

destruction. In

bacteria

particular methylated sequences function to control initia-

tion of DNA replication, DNA repair, gene expression, and movement of transposable el-

ements.

In

bacteria,

methylases

and

their

cognate nucleases

recognize specific

sequences

that

range

in

size from

4 to 8

base

pairs

in length. Each

site is independently

methylated.

In

6

DNA

CHEMISTRY AND

BIOLOGY

which is

converted to

m CpG. Although the

eukaryotic methylases recognize only

a dinu-

cleotide sequence, methylation (or unmethylated) at CpGs appears to be regionally specific,

suggesting that nearby

m CpG sequences interact. This plays a role in allowing cells with the

same exact DNA sequence to maintain stable, different patterns of gene expression (cell dif-

ferentiation), and it also allows the contributions of the genomes of two parents to be distin-

guished in offspring, since they often have different patterns of DNA methylation.

The fifth base in the DNA of humans and other vertebrates is 5-methyl C. Its presence

has profound consequences for the properties of these DNA. The influence of

m

C on DNA

three-dimensional structure is not yet fully explored. We know, however, that it favors the

formation of some unusual DNA structures, like the left-handed Z helix. Its biological

importance remains to be elucidated. However, it is on the level of the DNA sequence, the

primary structure, where the effect

of

m C is most profoundly felt. To

understand

the

im-

pact, it is useful to consider why DNA contains the

base T (5-methyl U) instead of U,

which predominates, overwhelmingly, in RNA. While the base T conveys a bit of extra

stability in DNA duplexes because of interactions between the methyl group and nearby

bases, the most decisive influence of the methyl group of T is felt in the repair of certain

potential mutagenic DNA lesions. By

far the most common mutagenic DNA damage

event in nature appears to be deamination of C. As shown in Figure 1.5, this yields, in du-

plex DNA, a U mispaired with a G. Random repair of this lesion would result, 50% of the

time, in replacement of the G by an A, a mutation, instead of replacement of the U by a

C, restoring the original sequence.

Apparently the intrinsic rate of the C to U mutagenic process is far too great for opti-

mum evolution. Some rate of mutation is always needed; otherwise, a species could not

adapt or evolve. Too high a rate can lead to deleterious mutations that interfere with re-

production. Thus the mutation rate must be carefully tuned. Nature accomplishes this for

deamination of C by a special repair system that recognizes the G–U mismatch and selec-

tively excises the U and replaces

it with a C. This system, centered about an enzyme

called

Uracil DNA glycosylase,

biases the repair process in a way that effectively avoids

most mutations. However, a problem

arises when the base 5-

m C is

present in the

DNA.

Then, as shown in Figure 1.5,

deamination produces T which is a normally occurring

base. Although the G–T mismatch is still repaired with a bias toward restoring the pre-

sumptive original

m C, the process is not nearly as efficient (nor should it be, since some of

the time the G–T mismatch will have come by misincorporation of a G for an A). The re-

sult is that

m C represents a mutation hotspot within DNA sequences that contain it.

In the DNA of vertebrates, all known

m C occur in the sequence

m

CpG. About 80% of

this sequence occurs in the methylated form (on both strands). Strikingly the total occur-

rence of CpG (methylated or not) is only 20% of that expected from simple binomial sta-

tistics based on the frequency of occurrence of the four bases in DNA:

X CpG

0.2

X C X G

where

X indicates the mole fraction. The remainder of the expected CpG has apparently been

lost through mutation. The occurrence of the product of the mutation, TpG, is elevated, as ex-

pected. This is a remarkable example of how a small bias, over the evolutionary time scale,

can lead to dramatic alteration in properties. Presumably the rate of mutation of

m

CpG’s con-

tinues to slow as the target size decreases. There must also be considerable functional con-

straints on the remaining

m CpG’s that prevent their further loss.

PLASTICITY IN DNA STRUCTURE

7

Figure 1.5

Mutagenesis and repair processes that alter DNA sequences.

(a) Deamination of C and

m C

produce

U and T, respectively.

(b) Repair of uracil-containing

DNA can occur without errors,

while repair of mismatched T–G pairs incurs some risk of mutagenesis.

(c) Consequences of exten-

sive

m CpG mutagenesis in mammalian DNA.

The vertebrate immune system appears to have learned about the striking statistical ab-

normality of

vertebrates. Injections of DNA from other sources with a high G

C con-

tent, presumably with a normal ratio to CpG, act as an adjuvant; that is, these injections

stimulate a generally heightened immune response.

PLASTICITY

IN DNA STRUCTURE

8

DNA CHEMISTRY AND BIOLOGY

average forms. DNA is revealed to be fairly irregular by methods that do not have to aver-

age over long expanses of structure. DNA is a very plastic molecule with a backbone eas-

ily distorted and with optimal geometry very much influenced by its local sequences. For

example, certain DNA sequences, like properly spaced sets of ApA’s promote extensive

curvature of the backbone. Thus, while base pairs

predominate, the angle between the

base pairs, the extent of their stacking (which holds DNA together) above and below

neighbors, their planarity, and their disposition relative to helix axis can vary substan-

tially. Almost all known DNA structures can be viewed in detail by accessing the Nucleic

Acid Database, NDB

http://ndbserver.rutgers.edu/

.

We do not really know

enough about the properties of proteins that

recognize DNA.

One extreme view is that these proteins look at the bases directly and, if necessary, distort

the helix into a form that fits well with the structure of protein residues in contact with the

DNA. The other extreme view has a key role played by the DNA structure with proteins

able to recognize structural variants, without explicit consideration of the sequence that

generated them. These views have very different implications for proteins that might rec-

ognize classes of DNA sequences rather than just

distinct single sequences. We are not

yet able to decide among these alternative views or to adopt some sort of compromise po-

sition. The structures of the few protein-nucleic acid complexes known can be viewed in

the NDB.

DNA

SYNTHESIS

Our ability to manufacture specific DNA sequences in almost any desired amounts is well

developed. Nucleic acid chemists have long learned and practiced the powerful approach

of combining chemical and enzymatic syntheses to accomplish their aims. Automated in-

struments exist that perform stepwise chemical synthesis of short DNA strands (oligonu-

cleotides) principally by

the phosphoramidite method. Synthesis proceeds

from the 3

-

end of the desired sequence using an immobilized nucleotide as the starting material (Fig.

1.6a ). To this are added, successively, the

desired nucleotides in a blocked, activated

form. After each condenses with the end of the growing chain, it is deblocked to allow the

next step to proceed. It is a routine procedure to synthesize several compounds 20 nu-

cleotides long in a day. Recently instruments have been developed that allow almost a

hundred compounds to be made simultaneously. Typical instruments produce about a

thousand times the amount of material needed for most

biological experiments. The cost

is about $0.50 to $1.00 per nucleotide in relatively efficient settings. This arises primarily

from the costs of the chemicals needed for the synthesis. Scaling down the process will

reduce the cost accordingly, and efforts to do this are a subject of intense interest. For cer-

tain strategies of large-scale DNA analysis, large

numbers of different oligonucleotides

are required. The resulting cost will be a significant factor in evaluating the merits of the

overall scheme. The currently used synthetic schemes make it very easy to incorporate

unusual or modified nucleotides at desired places in the sequence, if appropriate deriva-

tives

are available. They

also make it very easy to

add, at the ends

of the DNA strand,