Hi, I'm Rich Roberts ¦ and I'm Phil Sharp. In the '70s, while at Cold Spring Harbor Laboratory, we developed a number of important tools and techniques which led to our amazing discovery of interrupted genes. The first 'tool' we developed is a class of enzymes called restriction endonucleases - or simply restriction enzymes. In 1971, Hamilton Smith and Dan Nathans first used restriction enzymes to cut and analyze DNA. Restriction enzymes recognize and cut a specific sequence of DNA by breaking the sugar-phosphate backbone. Both Phil and I identified new restriction enzymes with different cutting specificities. The first 'tool' we developed is a class of enzymes called restriction endonucleases - or simply restriction enzymes. In 1971, Hamilton Smith and Dan Nathans first used restriction enzymes to cut and analyze DNA. Restriction enzymes recognize and cut a specific sequence of DNA by breaking the sugar-phosphate backbone. Both Phil and I identified new restriction enzymes with different cutting specificities. Using these restriction enzymes, a large piece of DNA can be cut into smaller gene-size pieces. This was important, because to study and characterize a gene, we needed a way to reproducibly generate gene-size DNA. Restriction sites are specific, cutting a piece of DNA with the same enzyme will always give the same DNA fragments. Depending on the frequency of the recognition sequence in a target DNA, a restriction enzyme can cut a DNA molecule many times or not at all. In the late '70s and early '80s, my lab at Cold Spring Harbor isolated and identified over three-quarters of the known restriction enzymes. Both Rich and I were also working on the DNA of simple viruses like adenovirus, which has 35,000 base pairs of DNA. We used restriction enzymes to cut adenovirus DNA into smaller pieces. For example, a restriction enzyme called EcoRI cuts DNA anywhere the sequence GAATTC is found, and breaks the backbone between the G and A. Having cut adenovirus DNA into small pieces, the DNA fragments needed to be separated so they could be analyzed. Ultracentrifugation, like that used by Meselson and Stahl, could work but was expensive and time-consuming. Gel electrophoresis - earlier used to separate proteins - was the ideal technique to separate DNA fragments. Electrophoresis uses an electric current to separate different-sized molecules in a porous, sponge-like matrix. Smaller molecules move more easily through the gel pores than larger molecules. For proteins, a polyacrylamide gel was used. However, polyacrylamide gels only separate small molecules and can't separate gene-sized fragments of DNA. While at Cold Spring Harbor, I worked with Joe Sambrook and Bill Sugden and developed an agarose gel, made from highly purified seaweed. This separated DNA molecules ranging from several hundred nucleotides in length to over 10,000 nucleotides. The gel is submersed in a tank filled with a salt solution that conducts electricity. Using a pipette, DNA samples are loaded into slots made in the agarose gel. The DNA samples are colorless, but we add a blue "tracking" dye. This makes it easier to load the samples and we can visually track the DNA migration through the gel. Remember that the phosphate groups in the DNA backbone carry negatively-charged oxygens - giving a DNA molecule an overall negative charge. In an electric current, negatively-charged DNA moves toward the positive pole of the electrophoresis chamber. The DNA molecules move through the gel by "reptation" - a reptile-like snaking through the pores of the agarose matrix. Smaller DNA fragments migrate faster and further over a given period of time than do larger fragments. This is how DNA fragments can be separated by size in an agarose gel. We also introduced the use of the fluorescent dye, ethidium bromide, to stain DNA. Ethidium bromide binds tightly to the double helix, and glows when illuminated with ultraviolet light. This lets us see where the separated DNA fragments ended up. A photo is taken of the gel for later analysis. The size of any DNA restriction fragment can be determined by comparing it to "markers" - DNA fragments of known sizes. A "map" of restriction enzyme sites can be generated by cutting a piece of DNA with different combinations of restriction enzymes. Let's go through an example. Suppose we had a 15,000 base-pair (15 kb) piece of DNA. When cut with EcoRI and run on a gel, we see two bands. We can size the bands by comparing them to 'marker' DNA - fragments of known sizes. As you can see, one is 8 kilobases (kb) and the other is 7kb. So, EcoRI cuts this 15-kb DNA only once. When we cut this 15-kb DNA with another enzyme, BamHI, two bands result: 10 and 5 kb. Like EcoRI, BamHI cuts the 15 kb of DNA once to give two pieces. But, since we don't know the relative order of the BamHI to the EcoRI sites, the BamHI sites can be arranged to give two maps. To figure out which BamHI map is correct, we can perform a "double digest" using both EcoRI and BamHI to cut the DNA. This lets us map the restriction sites relative to each other. Doing the double digest with our DNA gives bands whose lengths are 7, 5, and 3 kb. When we compare the double-digest data to the EcoRI and BamHI single digests, we see that the 8-kb EcoRI fragment is missing, and is "replaced" by fragments that are 5 and 3 kb long. This tells us that there is a single BamHI site within the 8-kb EcoRI fragment. The final EcoRI/BamHI map fits all the data from the double and the single enzyme digests. This strategy of restriction enzyme mapping was and is used to map DNA genomes. These maps are extremely useful because with them we can correlate the genetic map with a physical map of DNA segments. We can locate genes on pieces of DNA. By the 1970s, the details of transcription - making messenger RNA molecules from DNA coding regions - had been worked out for bacterial cells. In particular, we knew that bacterial mRNA molecules and their DNA counterparts were "colinear." That is, if aligned next to each other, the mRNA and DNA matched up along their entire length. However, when we did the same experiment with eukaryotic mRNA and DNA, we saw distinct "loops" of DNA between the matched RNA/DNA segments. We called these R-loops. These R-loops must be non-coding sequences. However, when we did the same experiment with eukaryotic mRNA and DNA, we saw distinct "loops" of DNA between the matched RNA/DNA segments. We called these R-loops. These R-loops must be non-coding sequences. Working with Rich Gelinas, a post-doc, we started looking at the differences between mRNA and DNA. Based on the information from the adenovirus restriction map that Phil Sharp and I made, I cut the adenovirus DNA and isolated specific fragments. I used a single-stranded BamHI DNA fragment ... ... and I mixed it with adenovirus messenger RNA. With help from Louise Chow and Tom Broker, we used the electron microscope to see what, if any, DNA/RNA hybrids and R-loops were made. We saw that one end of the mRNA hybridized to three points on the BamHI DNA fragment. This gave the two R-loops seen in the electron micrograph. The DNA loops out because the sequences are not present in the mRNA. It was obvious that RNA and DNA weren't colinear as we had assumed. The gene on the DNA was split compared to the mRNA sequence. By using the "right" restricted DNA fragment and comparing the results, I determined the start of the mRNA on the DNA, and which DNA segments coded for the rest of the mRNA. We used this electron micrograph in our 1977 Cell paper: An Amazing Sequence Arrangement at the 5' Ends of Adenovirus 2 Messenger RNA. While Rich was doing his work at Cold Spring Harbor, I was getting similar results from my experiments at MIT. The discovery of split genes revolutionized our thinking of how genes are organized.

dna restriction enzymes, sugar phosphate backbone, restriction endonucleases, target dna, restriction enzyme, restriction sites, hamilton smith, rich roberts, phil sharp