Techniques that have been used to evaluate Pin1 binding include fluorescence anisotropy, isothermal titration calorimetry (ITC), and nuclear magnetic resonance (NMR) titrations. As Pin1 is a two domain protein with two distinct binding sites, NMR is the most sensitive method for determining binding since the two binding sites can be evaluated simultaneously and residue specifically (microscopic binding affinities). The main drawback for NMR is that it is ideal only for weaker binding events in the μM to mM regime. Fluorescence anisotropy and ITC have the potential for showing two separate binding events, but the K_D of the two events must be different by orders of magnitude, as if they are too similar they are indistinguishable. Even if two binding events were measured using either of these two methods, these methods alone would not be enough to disentangle which binding affinity corresponds to each binding site. In all the studies that have reported binding affinities of Pin1 using fluorescence anisotropy or ITC, only one binding affinity was reported for the whole protein, likely because the binding affinities of each domain are in the same order of magnitude, resulting in an effective K_D. The smaller the K_D value, the larger the binding affinity between the two molecules.

To quantify Pin1 isomerase activity, most mutants were studied using NMR exchange spectroscopy (EXSY) by fitting the cis-trans and trans-cis cross and the respective diagonal peaks in a NOESY experiment using many different mixing times [34]. Before performing this EXSY experiment, the ¹H chemical shifts of ligand are assigned without enzyme present. The ¹H chemical shifts are often distinct for trans and cis species due to slow exchange, typically with peak intensity ratios around 10:1 trans to cis. During a NOESY experiment, usually used to measure through-space interactions, each hydrogen is frequency labeled by its respective chemical shift and after a certain mixing time (time given for isomerization to occur), the hydrogens are frequency labeled again. If the proton isomerized to a different chemical shift during that time, there will be a cross peak between the two distinct frequencies. For example, if a ligand is in the trans conformation during the first frequency detection and then undergoes isomerization during the mixing time, the cis conformation is detected at the second frequency labeling event. Isomerization of a ligand can occur without enzyme present, but this is often a slow process (seconds-minutes) that is unlikely to occur during a NOESY mixing time (milliseconds-seconds). For pCDC25c and FFpSPR ligands, no cross peaks are detected without enzyme present. Once Pin1 is added, cross-peak intensities can be measured at various mixing times to extract rates of exchange. A perk of this experiment is that both c→t and t→c rates can be measured, and the sum of which is the exchange rate (k_CT + k_TC = k_EXSY). While many studies report both k_CT and k_TC, some studies only report k_EXSY. Note that the larger the k_EXSY value, the higher is the activity of the enzyme.

Another method to measure activity by determining k_cat/K_M has been performed through a chymotrypsin-coupled chromophoric assay of p-nitroaniline (pNA) [35]. Chymotrypsin is highly selective towards X-Pro-Phe-pNA and will only hydrolyze the C-terminal pNA if the X-Pro bond is in the trans conformer. Therefore, the substrate used to study Pin1 activity is suc-AEPF-pNA, with glutamate acting as a phosphomimetic. A similar assay can be done with trypsin when peptides contain Pro-Arg-pNA motifs, like with peptide WFYpSPR-pNA [3].

Figure 2 and Figure 3 summarize the affinity and activity, respectively, of many Pin1 mutants. Single measurements are reported only with a bar, while variants with multiple measurements show an error bar with the standard deviations. Note that some values have a high standard deviation do to various measurements not agreeing well across different studies.

Both domains have been studied in isolation, where they have different binding affinities and activities than the full-length (FL) protein (Figure 2A). It should be noted that constructs that are tested in the full-length protein always have “FL” written, while the isolated domains do not have any distinction. For example, “WW” refers to the isolated WW domain, while “FL WW” is the WW domain in full-length Pin1. This nomenclature is carried out throughout this review. While the isolated WW domain has similar affinity to the WW domain in full-length Pin1 (albeit marginally weaker), the isolated PPIase has typically 100× weaker affinity than the PPIase in the full-length protein [10]. It should be noted the high variability in PPIase binding: values range from 8 μM [5] to 10 mM [36]. Using GST pull-downs, the isolated WW domain was able to pull-down mitotic phospho-proteins (containing MPM-2 and Plk1 antibody recognition motifs) with similar efficiency as full-length Pin1, while the isolated PPIase domain had no ability to pull-down these targets [37]. Overall, we can conclude that the PPIase has lower binding affinity than the WW domain to the tested ligands and that the presence of the partnering domain has an effect on the affinity of each domain. All mutants evaluated in this study are located on the structure of Pin1 in Figure 4.

All reported values for binding affinities and activities refer to phosphorylated ligands. It has been shown, however, that Pin1 can also bind unphosphorylated motifs. For example, the activity of Pin1 towards unphosphorylated WFYSPR-pNA is 170 mM⁻¹s⁻¹, it increases to 20,160 mM⁻¹s⁻¹ for WFYpSPR-pNA [3]. The initial crystal structure, 1pin, contained a bound phosphate ion in the catalytic site [4]. Interestingly, other PPIases such as the cyclophilins and FKBPs are unable to isomerize phosphorylated motifs.

3. Conclusions

The changes in affinity and activity of Pin1 upon mutations are mapped on the Pin1 structure in Figure 4. There is no direct correlation between the effects on affinity and activity. Unsurprisingly with Pin1, like with most proteins, if the binding/active sites are mutated, in most cases a reduction in affinity and/or PPIase activity results (Figure 4). On the other hand, mutations at distal, strategic sites, like I28A and S138A in the interdomain interface, actually increase the isomerase activity through an allosteric mechanism (Figure 4B). In addition, changing the linker length between the two domains also changes the catalytic activity of Pin1. For ligand pCDC25c, the activity is higher for the isolated PPIase domain than in the presence of the WW domain. This questions the hypothesis that the purpose of the WW domain is the increase of local ligand concentration in order to increase the activity of Pin1. Instead, the WW domain appears to play a more subtle role in activity regulation, possibly in context with ligand-specific allostery.

Conclusively, this entry summarized mutations previously studied in Pin1 with their impacts on affinity and activity, and we hope other researchers can use this to evaluate and develop models to describe allostery. A suitable model for allostery would have to be able to explain the extensively mapped out effects of the mutations. This overview could also serve as an aid to make further mutations. Some future projects could look at the double mutants I28A and A140I, as these two residues have been predicted as a co-evolving pair and could give further insights into interdomain allostery [5]. Previous work suggests a conserved, hydrophobic conduit linking the interdomain interface to the PPIase catalytic site [53]. Residues I28, T29, A140, V150, T152, L60, L61, and V62 could be mutant targets for understanding this putative allosteric conduit. These mutations must trigger a coupled response of intra- and interdomain structural dynamics of Pin1. We are currently investigating such effects by a combination of nuclear magnetic resonance (NMR) and electron paramagnetic resonance (EPR) spectroscopy to characterize intra- and interdomain spatial sampling of Pin1, respectively.