I invented a technique that was almost as sensitive as PCR, without using exponential amplification, a feat Helen Lee of Abbott called “impossible” – even more impossible when one considers that my technique was also more quantitative, used less extensive sample preparation, and had a much faster time to first result.
I did not know at the time, and neither (I dare say) did Kary Mullis, the inventor of PCR, that a future need would determine the ultimate fate of these two techniques.
In the early 1980’s, that future need was multiplex analysis – the ability to analyze dozens, hundreds, even thousands of sequences at once – something PCR was adept at doing. After all, in human cells, thousands of genes are replicated during the S phase of the cell cycle.
My biggest mistake was not inventing PCR when I had the chance in 1979.
My second biggest mistake: not inverting the format of my own sensitive hybridization technique, called RTC, reversible target capture. With that inversion, my technique could have been as sensitive as PCR, and more sensitive than RT-PCR (which has the inefficient reverse transcriptase step). The RTC technique might still be in use today for applications requiring superior quantitation, and especially with RNA targets.
A multiplex RTC with more than a dozen targets would seem to be a stretch without alternative, more efficient capture methods, requiring fewer beads. (A bead binning mechanism that would separate the different capture beads containing the different targets after hybridization would be required. Compare the technique of Illumina).
By “inversion” I mean using the homopolymers for quantitative linear signal amplification, not capture. Today I would design two or three non-interacting capture extender probe sequences to be used in the RTC. I would do a liquid phase hybridization with the three capture extender sequences and one to ten label extender sequences. The label extender probes that hybridize to the target would be tailed with poly(dC), with about 3000 or more residues. I would add six dC residues to the 3′ end of each oligonucleotide probe to make the tailing more uniform.
After washing away excess capture and label extender probes at high stringency, I would use first amplifier probes, composed of oligo(dG)-poly(dA) to hybridize to the label extenders that are hybridized to the captured target. Oligo(dG) would be short, 10-12 nucleotides, maybe 15, because of aggregation (solubility) problems, and an oligo (dA), six nucleotides in length, would be used during oligonucleotide DNA synthesis of these generic first amplifier probes, to get better, more uniform poly(dA) tailing [and longer dA sequences if adding them to the oligo(dG) would in fact reduce aggregation problems with the sequences]. The poly(dA) tail would be about 3,000 or more nucleotides. After washing, I would hybridize oligo(dT)-enzyme complexes to the poly(dA) tails. The oligo(dT) would be optimally between 20 and 30 nucleotides for decent stability.
After stringent washing, I would use displacement hybridization of the capture extender-capture probe, to do the release steps at low stringency (so that most of the non-specifically bound background would be left on the solid support). The displacement hybridization would be facilitated by having the capture sequence and its complementary displacing sequence longer than the capture extender sequence.
Then I would bind the probe-target complex to a second solid support containing a second capture sequence. Wash stringently, release by low stringency displacement hybridization. This second capture step would reduce both major types of background noise: non-specific binding of labeling sequences and non-specific hybridization between labeling sequences and capturing sequences.
At this point, the enzyme could be detected in solution.
Finally, if necessary, capture a third time to reduce the background (non-specific binding and non-specific hybridization), and detect the enzyme label. Alternatively, detect the enzyme label after the third release.
In the case of micron scale beads as a solid support, and as a general rule, I would filter the solution after each release to trap beads and bead fragments that contaminate the eluant.
For even greater sensitivity, the target could be concentrated into a “dot” prior to detection. A microdot, a nanodot, a picodot, etc. Whatever is required. Electrofocusing would be one method of concentrating nucleic acid targets. For even greater sensitivity, the target could be cleaved into many targetlets prior to the first hybridizations.
For even greater sensitivity, “smart” systems, such as two types of label extender sequences [prepared by a long series of ligations of unique sequences] could be used along with a signal generating scheme that required both sequences to be present. For example, the second type of label extender could bind a preamplifier and amplifier that brings the chemiluminescent enhancer directly onto the target. May save a round of capture.
Sensitivity, speed, quantification, with generic linear, but powerful amplification sequences – there is a need to validate so-called quantitative PCR assays. This i-RTC method could be used to great advantage. i = inverted. There is no need to label i-RTC as quantitative; it is that in spades. The use of multiple oligonucleotides – at least 3 or 4 total is needed for improving hybridization efficiency to near 100% in strongly structured targets like certain RNAs. Every hybridization reaction is driven to completion to ensure the accuracy and precision of the quantitation. The linearity of the signal amplification ensures the most accurate quantification in the dose-response curve.
1. Improving the RTC method: for the first multiplex capture, use a single capture sequence for convenience, or use multiple sequences for greater specificity.
2. More signal: Hybridize an oligo(dT)- branched DNA to the poly(dA) tail, and wash well. Hybridize a complementary enzyme-labeled sequence, wash well, release, recapture, wash, release and detect.