In contrast, Francis and I were far from being on our own. One flight up was the clever Bill Cochran, who put the Bessel functions of helical diffraction theory into Francis’s working vocabulary, whence they entered mine. Even more important, Jerry Donohue’s spartan desk was no more than 12 feet from mine and Francis’s when his quantum chemistry expertise squelched my initial desire to build a double helix based on like-with-like base pairing (e.g., A-A and T-T). The Cavendish then was a magnet for minds that wanted to be challenged by others of equal power. In contrast, Linus Pauling’s Caltech was a chemistry garden of mortals hovered over by a god who saw no need to assimilate the ideas and facts of others. If Linus had only spent a few days in Caltech’s libraries perusing the literature on DNA that fall, he would most likely have hit upon the idea of base pairing and would now be celebrated for both the alpha helix and the double helix.
Virtually everyone who came to our now even more cramped Cavendish office to see the large 3-D model made in early April was thrilled by its implications. Any doubt as to whether DNA, and not protein, was the genetic information-bearing molecule suddenly vanished. The complementary nature of the base sequences on the opposing chains of the double helix had to be the physical counterpart of the Pauling-Delbrück theoretical postulation of gene copying through the creation of complementary intermediates. DNA double helices as they exist in nature must reflect single-stranded template chains hydrogen-bonded to their single-stranded products of complementary sequence. Two of the three big questions in molecular genetics, the DNA structure by which genetic information is carried and how it is copied, were thus suddenly resolved through the discovery of base-pair hydrogen bonding.
Still to be ascertained was how the information conveyed by the sequence of DNA’s four bases (adenine, guanine, thymine, and cytosine) determines the order of the amino acids in the polypeptide products–the stuff of the proteins forming all living things–of individual genes. Since there were known to be 20 amino acids and only four DNA bases, groups of several bases must be used to specify, or code for, a single amino acid. I initially thought the language of DNA would be best approached not through further work on the DNA structure but by work on the 3-D structure of its close chemical relative ribonucleic acid (RNA). My decision to move on from DNA to RNA reflected the already several-years-old observation that polypeptide (protein) chains are not assembled on DNA-containing chromosomes. Instead, they are made in the cytoplasm on small RNA-containing particles called ribosomes. Even before we found the double helix, I postulated that the genetic information of DNA must be passed on to RNA chains of complementary sequences that in turn function as the direct templates for polypeptide synthesis. Naively, I then believed that amino acids bonded to specific cavities linearly located on the surfaces of the ribosome RNA components.
After three subsequent years of x-ray studies–the first two at Caltech and the last back with the “unit” in Cambridge, England, in which I was joined by the Pauling- and Harvard Medical School-trained Alex Rich–I failed to generate a plausible 3-D structure for RNA. Though RNA from many different sources produced the same general x-ray diffraction pattern, the pattern’s diffuse nature gave no solid clues as to whether the underlying RNA structure contained one or two chains. By early 1956, I decided to change my focus from x-ray studies on RNA to biochemical investigations on ribosomes when I returned to the States to begin teaching in the fall at Harvard. Also then seeking a more tractable challenge was the Swiss-born biochemist Alfred Tissières, then studying oxidative metabolism at the Molteno Institute in Cambridge. He had already briefly dabbled with ribosomes from bacteria and liked the idea of our seeking out how they work across the Atlantic in the other Cambridge.