Besides this isothermal amplification a second test system was chosen to evaluate the microfluidic integration. The reaction cycle for an autocatalytic amplification was recently published by Zhang et al. (Science 318, 1121 (2007)). Here an autocatalyst (34mer) reacts with a substrate consisting of a terplex (50mer template, 34mer autocatalyst and 36mer signal) leading to free signal (36mer). This signal then reacts with the reporter (duplex, one 20mer labelled with Alexa 488 and one 30mer labelled with Dabcyl) leading to an increase of fluorescence as the signal displaces the Alexa labelled oligonucleotide. In the next step the new-formed terplex (template, 2 x autocatalyst) releases both autocatalysts as the fuel (44mer) reacts with the template. So in each reaction cycle the amount of autocatalyst is doubled and the consequence is an exponential increase in fluorescence intensity.
We have shown that it is possible to perform this chemical reaction within our microfluidic environment. With a stopped flow procedure, which is comparable to a reaction in a cuvette, we observed a curve showing exponential saturation kinetics as expected. Starting and stopping the flow in a 10-minute-interval we could reproduce this curve progression. Furthermore we did the reaction using different flow rates to show different points of the reaction corresponding to the distance that is covered by the reaction mixture. For example after mixing of the oligonucleotides (fuel, reporter, autocatalyst) with the substrate that is needed for the reaction to start the reaction mixture covers a distance of 10 mm in 8 minutes using a flow rate of 0.5 µl/h. So the corresponding increase at this position represents the progress of the reaction after 8 minutes. The red coloured curve in the upper left image represents the autocatalytic reaction under stopped-flow condition in comparison to curves representing flow rates at 0.2, 0.4 and 0.6 µl/h. As expected with increasing flow rate the progress of reaction at the selected point on the chip decreases.
Furthermore we tried to use the electrodes implemented in our microfluidic devices to control the reaction within the "H"-shaped reaction-chamber. Using the electroosmotic flow, we concentrated the chemical components in the reaction chamber (see lower figure) between the two channels 1 and 2 (upper figure) that carry the substrate and the other oligonucleotides respectively. Using the electrodes with a defined cycle being positive, negative or uncharged we got a concentration of the oligonucleotides in the reaction chamber. What was not desired is that several parts of the oligonucleotides precipitate due to the high concentration of MgCl2 (12.5 mM) in the used buffer. Hence it is not clear if the increase in fluorescence intensity can be traced back on the autocatalytic reaction as we also got an increase of intensity doing the reaction without the substrate that activates the reaction. It seems to be a mixture of both processes: the reaction and the precipitation with the MgCl2. Tests to do the reaction with a smaller amount of MgCl2 or NaCl as an alternative salt led to a breakdown of the reaction so that we could not get any further results with this system.