On the interaction between ARFs and AuxREs: from methods to models

The plant hormone auxin is a key regulator of many growth and developmental processes. The presence of auxin  changes the gene expression of the cell via a short pathway called the nuclear auxin pathway (NAP). The NAP contains three players: TIR1/AFB, Aux/IAA and the transcription factor ARF. Under conditions of low auxin,  Aux/IAA binds ARF repressing it. At higher concentrations of auxin, TIR1/AFB binds Aux/IAA and marks it for degradation, thereby freeing ARF. Members of the ARF family bind a DNA motif called AuxRE (TGTCTC) and their DNA binding domains (DBDs) are known to dimerize and bind cooperatively on response elements composed by two AuxREs present in an inverted orientation. Although the affinities at play inside the NAP and between ARFs and AuxREs had been studied extensively with a range of techniques,  a quantitative view on this system was still lacking. Throughout this thesis, we therefore developed and applied methods with the goal of obtaining a quantitative understanding of ARF-ARF and ARF-AuxRE interactions.

In chapter 2 we developed and utilized a method based on smFRET and smPIFE-FRET to study the interaction between ARF and composite DNA response elements. We tested several A. thaliana ARF-DBDs and showed that the differences in binding stability and kinetics are quantitatively captured by the method. The smFRET and smPIFE-FRET experiments in which  the affinity of a DNA-binding deficient mutant of AtARF5-DBD (R215A) was probed, proved that the shift in FRET efficiency seen when wt AtARF5-DBD is added in concentration of tens on nM is due to the specific protein interaction with the DNA motif. We then tested the binding preferences of Marchantia polymorpha ARF1 and ARF2, which are, respectively, the only members of the phylogenetically conserved class A-ARF and B-ARF in the species. Whereas A-ARFs are considered   activators, B-ARFs are considered repressors. In M.  polymorpha, a  mechanism in which B-ARFs repression is based on competition for the same AuxREs bound by A-ARFs had been proposed as part of a minimal auxin response system. We validated this model on a composite response element carrying two TGTCTC in an inverted repeat spaced by 7 basepair (IR7-TGTCTC) showing that the ratio between the affinity of MpARF2-DBD and MpARF1-DBD is ~5, allowing for competition; similarly, IR7-TGTCGG showed a ratio of ~3. Interestingly, IR7-TGTCAA showed a ratio of ~13 which hints to a  regulation for this element that does not involve competition between A and B-ARFs.

We then applied the method based on smPIFE-FRET to determine the effect of dimerization and oligomerization on the affinity and kinetics of ARF-AuxRE interaction. In chapter 3 we tested the binding affinity of different variants of AtARF2 towards an IR7 response element. We showed that the presence of the PB1 domain increases the affinity of this protein-DNA interaction and that this effect is caused by the increased stability of the protein dimer brought by the homotypic interaction of the PB1 domain; the effect of further oligomerization on the affinity towards this composite AuxRE was found to be negligible. The kinetics of AtARF2-IR7 interaction, showed an increase of the observed  kon and a  decrease in the observed koff in the full-length protein compared to the DBD, with full-length variants carrying mutations that reduce the interaction between the PB1 domains (K2S and OPCA) showing intermediate values. Remarkably, both the trends in the kinetics and the differences in affinity towards the DNA shown by the different AtARF2 variants were captured by a global fit based on a four-states association model. The model explained the differences between variants as being caused by changes in the dimerization equilibrium of the proteins and returned a dimerization Kd  for AtARF2-DBD of ~2 µM.

In chapter 4 we introduced two microfluidic devices aimed for single-molecule fluorescence detection on a TIRF microscope without requiring the immobilisation of DNA constructs. Both designs feature a (nano-)channel height of  200 nm which confines the fluorescent molecules into the evanescent field of the TIRF microscope allowing for single-molecule tracking to take place. The first design featuring parallel nanochannels was tested studying  the well characterized interaction between DNA polymerase I (KF) and a gapped DNA construct and the conformational dynamics of a DNA hairpin; these experiments returned results comparable with the ones obtained in smFRET experiments with immobilized DNA molecules. Using the second design featuring a T-shaped mixing (nano-)channel  we accessed non-equilibrium conditions by mixing primarily open DNA hairpins with a high-salt solution triggering the closing of DNA hairpins. Moreover, we observed polymerization of 25 bases on a DNA template by a DNA polymerase, illustrating that complex biological reactions can be followed in real time and in a continuous fashion. The parallel nanochannels proved useful in studying the interaction between ARF-DBDs and between ARF-DBD and its DNA response element.

In chapter 5 we demonstrated that quantitative data similar to the one coming from immobilized samples can be obtained from smFRET experiments using free-flowing doubly-labelled DNA constructs; moreover, a decrease in diffusion coefficient can be detected and used as an independent readout for the DNA binding. The values of the diffusion coefficients of fluorescently labelled AtARF5-DBD and AtARF1-DBD showed that they are present as monomers at the ~1 nm concentrations tested. Finally, flowing a mixture of AtARF5-DBDs labelled with a donor or an acceptor  showed that the addition of unlabelled DNA containing an IR7 response element is likely to stabilize the dimer as a  population displaying FRET and characterized by slow diffusion coefficient emerged. This stabilization  once again underlines the importance of cooperativity for the stability of biomolecular complexes.