In the mid-1950s Spiegelman turned his research focus to the enzymes involved in nucleic acid synthesis. Francis Crick and James Watson had published their landmark model of DNA in 1953, which indicated that the genetic code was contained in a linear sequence of chemical bases--thymine, adenine, guanine, and cytosine. The model also suggested that in duplication, each strand of the double helix served as a pattern--a template--for assembling a new double helix by making a complementary copy. (The bases always bond in a complementary way, adenine to thymine, and cytosine to guanine.) However, much was still unclear about exactly how replication occurred and how the genetic information (represented by the sequences of bases) was translated to direct protein synthesis. It was accepted that RNA had something to do with protein synthesis, probably as the intermediary between DNA and protein. The likely mechanism of DNA replication also implied that translation could occur via complementary RNA copies of the DNA (as indeed it does). But while the DNA base composition varied widely between different organisms, the composition of the RNA found in their cells was monotonously similar. It would later be found that most of the RNA in cells is ribosomal RNA which actually makes the proteins, but several other varieties, including messenger and transfer RNA, exist in a cell, in much smaller quantities, and only during certain phases of cell growth. In the mid-1950s, a simple relation between DNA and RNA appeared unlikely.

Nevertheless, Spiegelman's accumulated research with enzyme induction led him to believe that there was a large molecule in the cell cytoplasm directing the synthesis of each enzyme, and that it was probably RNA patterned on the cell's DNA. If so, there must be a special enzyme that could make complementary RNA copies of the DNA. Arthur Kornberg had recently isolated DNA polymerase--the enzyme that catalyzes the making of new DNA molecules--and Spiegelman was certain he would find a corresponding RNA polymerase. By 1958 he had found convincing evidence of such an enzyme--a DNA-dependent RNA polymerase--in E. coli bacteria. Rather than trying to purify the enzyme, he set about designing an experiment that would demonstrate that it actually operated in the cell. This work led directly to the technique of molecular hybridization, one of the most useful techniques of molecular biology.

For his experiment, Spiegelman drew on a curious study reported in 1956 by Eliot Volkin and Lazarus Astrachan at the Oak Ridge National Laboratory. It had long been observed that bacteria infected by viruses stopped making their own proteins. The viruses--known as bacteriophages or just phages--somehow took over the protein synthesis process in host cells. Volkin and Astrachan, tracing nucleic acid production with radioactive phosphorus, found that after infection by a DNA phage called T2, E. coli bacteria contained within them a strange RNA, whose chemical composition (specifically, the proportions of the four nitrogenous bases uracil, adenine, cytosine, and guanine) was quite similar to that of the phage DNA--and much different from the known composition of the bacterial DNA or RNA. Knowing that the RNA polymerase existed, Spiegelman saw that if the new RNA could be proved to be complementary to the phage DNA, and not that of the bacteria, it would demonstrate that the DNA transmitted its coding by making the RNA.

Using gel electrophoresis and sucrose-gradient centrifugation, he verified that the new RNA was molecularly distinct, and demonstrated that it could be physically separated from the rest of the normal RNA found in the uninfected bacteria. But how could its relationship to the phage DNA be proved? Just at the right time, in 1959, Paul Doty and Julius Marmur discovered independently that when double-stranded DNA was separated into single strands by heat (a process called denaturing), the single strands reconstituted spontaneously into double helices as the mixture slowly cooled; the sequences of complementary bases paired up again, adenine to thymine and guanine to cytosine. Further, Marmur found that such reconstitution would take place only between DNA strands that came from the same or closely related organisms. Spiegelman realized immediately that if the RNA were a copy of phage DNA, it should be complementary to one of the two phage DNA strands. Working with Benjamin Hall and Masayasu Nomura, he purified some of the T2 phage RNA, mixed it with denatured (thus single-stranded) T2 phage DNA, and then cooled the solution slowly. This yielded not only reconstituted DNA, but hybrid molecules made of one strand of DNA and one strand of RNA, with uracil-adenine pairs instead of thymine-adenine pairs. The methods they developed to detect the hybrids were ingenious and unusually intricate, requiring an arsenal of purification procedures, radioactive labeling, and five-day centrifuge runs with cesium-chloride gradients. Importantly, they also showed that the phage-related RNA would not form hybrids with DNA from other sources, not even that from phage T5, which is very similar in base composition. This proved that the new RNA was a specific complementary copy of the phage DNA, capable of directly carrying genetic instructions to the protein-making machinery in the cell cytoplasm.

During the next five years Spiegelman's team improved and simplified the hybridization procedure, making it accessible to a wider circle of researchers. They found a simpler way to detect the hybrids, using a special nitrocellulose filter paper--single strand DNA and DNA-RNA hybrids will adhere to nitrocellulose fibers, but RNA alone does not. The team also clarified some related problems of molecular biology showing, for example, that ribosomal and transfer RNA are specified by distinct portions of the cellular genome. They also demonstrated that hybridization could be used to locate physically the positions of particular sequences of DNA, i.e., genes. RNA-DNA hybridization rapidly became one of the major tools of genome research, including the development of recombinant DNA.