8.3: DNA Structure

By 1878 a substance in the pus of wounded soldiers derived from cell nuclei (called nuclein) was shown to be composed of five bases (the familiar ones of DNA and RNA). The four bases known to make up DNA (as part of nucleotides) were thought to be connected through the phosphate groups, in short repeating chains of four nucleotides. By the 1940s, we knew that DNA was a long polymer. Nevertheless, it was still considered too simple to account for genes. After the Hershey and Chase experiments, only a few holdouts would not accept DNA as the genetic material. When the scientific community finally accepted DNA as the stuff of genes, the next questions were the following:

• What did DNA look like?
• How did its structure account for its ability to encode and reproduce life? While the four phosphate-linked nucleotide composition of DNA was known for some time, it became mandatory to explain how such a “simple molecule” could inform the thousands of proteins necessary for life. The answer to this question lay (at least in part) in an understanding of the physical structure of DNA, made possible by the advent of X-ray crystallography.

8.3.1. X-Ray Crystallography and the Beginnings of Molecular Biology

If a substance can be crystallized, the crystal will diffract X-rays at angles revealing regular (repeating) structures of the crystal. William Astbury demonstrated that high molecular-weight DNA had just such a regular structure. His crystallographs suggested DNA to be a linear polymer of stacked bases (nucleotides), each separated from the next by 0.34 nm. Astbury is also remembered for coining the term “molecular biology” to describe his studies. The term now covers all aspects of biomolecular structure, as well as molecular functions (e.g., replication, transcription, translation, and gene regulation).

In an irony of history, the Russian biologist Nikolai Konstantinovich Koltsov had already intuited in 1927 that the basis for the genetic transfer of traits would be a “giant hereditary molecule” made up of “two mirror strands that would replicate in a semiconservative fashion using each strand as a template” (The Remarkable Nikolai Koltsov - scroll down to read the biography). This was a pretty fantastic inference if you think about it, since it was proposed long before Watson, Crick, and their colleagues worked out the structure of the DNA double-helix!

8.3.2. Wilkins, Franklin, Watson, and Crick—DNA Structure Revealed

Maurice Wilkins, an English biochemist, was the first to isolate highly pure DNA with a high molecular weight. Working in Wilkins’s laboratory, Rosalind Franklin was able to crystalize this DNA and to produce very high-resolution X-ray diffraction images of the DNA crystals. Franklin’s most famous (and definitive) crystallograph was “Photo 51” (Figure 8.5).

Figure 8.5: Photo 51, the X-ray crystallograph of DNA taken by Rosalind Franklin, revealed three crucial molecular dimensions that ultimately led to the double-helical DNA model.

This image confirmed Astbury’s 0.34 nm repeated dimension and revealed two more numbers, 3.4 nm and 2 nm, reflecting additional repeat structures in the DNA crystal. When James Watson and Francis Crick got hold of these numbers, they used them along with other data to build DNA models out of nuts, bolts, and plumbing.

Their models eventually revealed DNA to be a pair of antiparallel, complementary strands of nucleic acid polymers—shades of Koltsov’s mirror-image macromolecules! Each strand is a string of nucleotides linked by phosphodiester linkages, the two strands held together in a double helix by complementary H-bond interactions. Let’s look at the evidence for these conclusions. As we do, refer to the two illustrations of the double helix in Figure 8.6.

Figure 8.6: The three molecular dimensions from Franklin’s Photo 51 could be best explained by the proposition of a DNA double helix (left). Model-building by Watson and Crick revealed the antiparallel structure of the double helix (right).

Recalling that Astbury’s 0.34 nm dimension was the distance between successive nucleotides in a DNA strand, Watson and Crick surmised that the 3.4 nm repeat was a structurally meaningful tenfold multiple of Astbury’s number. When they began building their DNA models, they realized from the bond angles connecting the nucleotides that the strand was forming a helix, from which they concluded that the 3.4 nm repeat was the pitch of the helix (i.e., the length of one complete turn of the helix). This meant that there were ten bases per turn of the helix. They further reasoned that the 2.0 nm number might reflect the diameter of the helix. When their scale model of a single-stranded DNA helix predicted a helical diameter much less than 2.0 nm, they were able to model a double helix that more nearly met the requirement for a 2.0 nm diameter. In building their double helix, Watson and Crick realized that bases in opposing strands would come together to form H-bonds, holding the helices together. However, for their double helix to have a constant diameter of 2.0 nm, they also realized that the smaller pyrimidine bases, thymine (T) and cytosine (C), would have to H-bond to the larger purine bases, adenine (A) and guanine (G), and that neither A-G nor C-T would form pairs in the double helix.

Now to the question of how a “simple” DNA molecule could have the structural diversity needed to encode thousands of different polypeptides and proteins. In early studies, purified E. coli DNA was chemically hydrolyzed down to nucleotide monomers. The hydrolytic products contained nearly equal amounts of each base, reinforcing the notion that DNA was that simple molecule that could not encode genes. But Watson and Crick had private access to data from Erwin Chargaff, who had determined the base composition of DNA isolated from different species, including E. coli. He found that the base composition of DNA from different species was not always equimolar, meaning that for some species, the DNA was not composed of equal amounts of each of the four bases (see some of this data in Table 8.1).

The mere fact that DNA from some species could have base compositions that deviated from equimolarity put to rest the argument that DNA had to be a very simple sequence. Finally, it was safe to accept the obvious, namely that DNA sequences could vary almost infinitely and could indeed be the stuff of genes.

Chargaff’s data also showed a unique pattern of base ratios. Although base compositions could vary between species, the A/T and G/C ratio was always 1, for every species; likewise, the ratio of (A+C)/(G+T) and (A+G)/(C+T). From this information, Watson and Crick inferred that A (a purine) would H-bond with T (a pyrimidine), and that G (a purine) would H-bond with C (a pyrimidine) in the double helix. When building their model with this new information, they also found H-bonding between the complementary bases would be maximal only if the two DNA strands were antiparallel, leading to the most stable structure of the double helix. Watson and Crick published their paper A Structure for Deoxyribose Nucleic Acid in 1953 in Nature (read an annotated version of this seminal article at The Classic 1953 Watson-Crick.

Their article is also famous for predicting a semiconservative mechanism of replication, something that had also been predicted by Koltsov twenty-six years earlier, albeit based on intuition—and much less evidence! Watson, Crick, and Wilkins shared a Nobel Prize in 1962 for their work on DNA structure. Unfortunately, Franklin died in 1958, and Nobel prizes were not awarded posthumously. She did eventually get a measure of delayed and well-deserved recognition, including a university in Chicago named in her honor! Nevertheless, there is still controversy about why Franklin did not get appropriate credit at the time for her role in the work.

169-2 Unraveling the Structure of DNA

8.3.3. Meselson and Stahl’s Experiment—Replication is Semiconservative

Confirmation of Watson and Crick’s suggestion of semiconservative replication came from Matthew Meselson and Franklin Stahl’s very elegant experiment, which tested the three possible models of replication (Figure 8.7)

Figure 8.7: Three possible modes of DNA replication could be imagined.

In their experiment, E. coli cells were grown in a medium containing \({}^{15} \rm N\), a “heavy” nitrogen isotope. After many generations, the DNA in all of the cells had become labeled with the heavy isotope. At that point, the \({}^{15} \rm N\)-tagged cells were placed back into a medium containing the more common, “light” \({}^{14} \rm N\) isotope and were allowed to grow for exactly one generation. Figure 8.8 shows Meselson and Stahl’s predictions for their experiment. Meselson and Stahl knew that \({}^{14} \rm N\)-labeled and \({}^{15} \rm N\)-labeled DNA would form separate bands after centrifugation on CsCl (cesium chloride) density gradients.

Figure 8.8: Meselson and Stahl predicted experimental outcomes based on possible modes of replication.

They tested their predictions by purifying and centrifuging the DNA from the \({}^{15} \rm N\)-labeled cells grown in the \({}^{14} \rm N\) medium for one generation. They found that this DNA formed a single band with a density between that of \({}^{15} \rm N\)-labeled DNA and \({}^{14} \rm N\)-labeled DNA, thereby eliminating the conservative model of DNA replication (possibility #1), as Watson and Crick had also predicted.

That left two possibilities: replication was either semiconservative (possibility #2) or dispersive (possibility #3). The dispersive model was eliminated when DNA isolated from cells that had been grown for a second generation on \({}^{14} \rm N\) were shown to contain two bands of DNA on the CsCl density gradients.

170 Replication is Semiconservative