Visualizing Introns
Pictures Considered #58
by Christoph
When we talk about genes and genomes today we almost reflexively think of long strings of the letters G, A, T, C. By reading sequences we easily detect stretches that code for proteins. We detect promoters, transcription factor binding motifs, start/stop signals for transcription, signals where introns are removed by splicing from primary transcripts (pre‑mRNA) to yield mature messenger RNA (mRNA), and much more. It wasn't always so easy.
Figure 1. Diagrammatic representation of two nonallelic β‑globin genes. The first and third lines of the figure represent the EcoRI·HindIII fragment of MβG3 and the EcoRI fragment of MβG2, respectively, which were used to perform the R‑loops. The fragments are drawn with the 5' globin mRNA sequences to the left. The positions of the globin mRNA sequences (G) and the intervening sequences (IVS) are indicated. The middle line represents an enlargement of the equivalent, sequences of both fragments that hybridized to 15S RNA, with the BamHI site at the codons for amino acids 98–100 indicated. The numbers below indicate the sizes of the globin and intervening sequences in bp. The smaller, putative intervening sequence located close to the 5' end of the globin structural sequence in MβG2 is not shown in the diagram. Source
In fact, it was rather the other way around when molecular biology was still in its infancy in the 1960s and in its adolescence in the 1970s. Historically, genetics came first. Then the physical analysis of DNA and RNA, their interactions with each other and with proteins. And lastly, sequencing by both Maxam-Gilbert and Sanger sequencing methods in 1977. Introns were literally seen first, before their features as exact lengths and splice-site motifs were determined by sequencing. That's what today's post is about.
Tilghman et al. (1977) had cloned a complete mouse β‑globin gene from chromosomal DNA and found it considerably longer than the full-length cDNA clone they used as hybridization probe for its identification. They set out to clarify this discrepancy by analyzing the β‑globin gene transcripts since they had learned from very recent publications by the labs of Phil Sharp and Rich Roberts that eukaryotic genes can occur in a fragmented configuration, resulting in the translation of "complete" proteins only when portions are removed from the primary transcripts and the remaining ones are stitched together again, that is, "spliced."
Figure 2. Electron micrographs of R‑Ioop structures between (A) 15S RNA and MβG3, and (B) 10S RNA and MβG3. The dotted lines in the drawings indicate the positions of the RNA molecules. (×140,000 magnification). the ends of the 6.25 kb EcoRI·HindIII MβG3 fragments (see Figure 1) are marked with pink arrowheads. Source
The method of choice was visualization, using electron microscopy, of the hybrids formed between the cloned β‑globin gene and the RNA transcripts. This approach had been successfully used much earlier. For example, for mapping deletions in phage DNA by hybridizing mutant and wild-type phage DNA as featured in our Pictures Considered #6, and also by the Sharp and Roberts labs in their work on Adenovirus 2 mentioned above.
In their follow-up paper, Tilghman et al. (1978) present their analyses of RNA·DNA hybrids of the mouse β‑globin gene with its transcripts. It's not easy to adequately summarize their work and careful optimization of methods so I recommend you read the full article (Open Access PDF available here). But briefly, from 8 L of hematopoietic Friend mouse cells (clone 707/M2) cultured over several days they were able to isolate and purify 200 mg of RNA. After several enrichment steps they finally obtained two β‑globin-specific RNA fractions, a (shorter) 10S RNA and a (longer) 15S RNA.
For the electron microscopic analysis, they hybridized both β‑globin-specific RNA fractions separately to the MβG2 β‑globin gene and the non‑allelic MβG3 β‑globin gene. You see in Figure 2 example DNA·RNA hybrid molecules for MβG3 with 15S RNA (A) and 10S RNA (B). Note that in the electron micrographs, the nucleic acids are not seen "naked" but stained with uranyl acetate and, to increase their electron density, shaded with heavy metals platinum/palladium/carbon, the common method for visualizing nucleic acids in EM. Also note that single-stranded nucleic acids appear a bit thinner than double‑stranded nucleic acids, here the DNA·RNA hybrid molecules.
The longer 15S transcript hybridizes over its full length with the β‑globin gene while hybrids of the β‑globin gene with the shorter 10S transcript have a ~550 nucleotide‑long "displaced" single‑stranded region where the RNA cannot hybridize with DNA (R‑loop). They summarized their measurements in Figure 1: both the MβG2 and MβG3 β‑globin gene have a length of ~1300 bp and the coding region is interrupted by a ~550 bp long intron.
Figure 3. Map and opisometer. Source
In 1978, excellent electron microscopes were available but not digital cameras for documentation. So Tilghman et al. captured their electron microscopy images onto film, and produced positive prints from their negatives at a size that allowed measuring the length of the DNA·RNA hybrid molecules. The actual length measurements were then done with a digitizer. Did you know that the lengths of DNA molecules in electron micrographs were measured with a so-called opisometer before there were digitizers? No, opsiometers are not steampunk! Similar devices were independently invented in the industrialized countries – and called curvimètre in France, or Kurvimeter in Germany – in the 2nd half of the 19th century, and were used by cartographers (and the military!) to measure curved paths and roads on maps (Figure 3). Using opisometers was quite easy if you knew the magnification – that is, the scale – of your electron micrograph. You traced the length of the DNA molecule with the small wheel of the opisometer on the magnified print of your electron micrograph, and then you could read the "distance traveled" on the dial. If you had a decent standard deviation in a number of repeat measurements for a given DNA molecule, you could determine its length with an accuracy of ±10 bp (for comparison: the determination of the length of restriction fragments by agarose gel electrophoresis is rarely better). Time consuming, but not bad, right?
Do you want to comment on this post? We would be happy about it! Please comment on Mastodon or on Bluesky.
