All of the essential mechanisms in our system were veriﬁed in solution using polyacrylamide gel electrophoresis. These mechanisms include: walker-track binding, triggering the walker, walking from one track to another, picking up cargo, walking while carrying cargo, triggering the cargo goal, dropping oﬀ cargo, and irreversibly walking from tracks to the walker goal. Each gel has several control lanes (marked in green), where control lanes are either single strands, or previously veriﬁed combinations of strands. Positive controls are marked with (+) below the lane number, and negative controls are marked with (-) below the lane number. We deﬁne a negative control as one that should not appear in any reaction in that gel, even though some may in practice appear due to stoichiometric errors etc. Lanes that test the binding of certain strands are labeled in blue, whereas lanes that test strand displacement reactions are labeled in red. Null experiments (ones where we expect no reaction to occur, but provide comparisons for reactions where we expect the end result to be the same or similar) are labeled in black. All inputs in a lane were intended to be equimolar unless otherwise noted. Probes were used when it seemed necessary to distinguish strands or complexes of similar length (for example PTR2 was used with track 2 to distinguish it from track 1, although we found that TR1 always appears slightly higher in the gel than TR2, and likewise when they are bound to the walker, perhaps due to some minor secondary structure).
While the mechanisms all appear to behave properly in solution, there were a few mysteries, which may or may not matter as the mechanisms are translated onto origami. One such mystery is the apparent absence of TR2-PTR2 in lanes 6 and 7 of gel 4. There a couple possibilities for this. One is that it is just very low intensity, and since it should be close to the bands for W-TR1-C1, it blends in. Another is that there were stoichiometric issues, so almost all of the TR2 is in W-TR2-PTR2-C1. In any case, it does not seem possible for it to not exist if stoichiometry was correct, given the results in those two lanes. Another mystery is that in lanes 7 and 8 of gel 6 the bands that correspond to the walker appear somewhat higher than expected. No logical explanation can be found for this, and it seems ignorable, except for the fact that in a similar gel using old cargo and cargo goal strands, the same eﬀect appeared! Finally, in lane 4 of gel 7, the walker goal appears much higher than a 47 nucleotide strand should appear (compare to the tracks). This was also seen in other gels, which suggests there is probably some polymerization of the strand (or otherwise some secondary structure), but since the rest of the gel suggests the walker goal behaves as expected, it will be ignored.