Predicting Tankyrase Binders

Katie Pollock, Michael Ranes, Ian Collins, Sebastian Guettler

Published: 2021-09-03 DOI: 10.17504/protocols.io.bnxamfie

Abstract

This protocol is part of a collection: Identifying and Validating Tankyrase Binders and Substrates: A Candidate Approach

The poly(ADP-ribose)polymerase (PARP) enzyme tankyrase (TNKS/ARTD5, TNKS2/ARTD6) uses its ankyrin repeat clusters (ARCs) to recognize degenerate peptide motifs in a wide range of proteins, thereby recruiting such proteins and their complexes for scaffolding and/or poly(ADP-ribosyl)ation. Here, we provide guidance for predicting putative tankyrase-binding motifs, based on the previously delineated peptide sequence rules and existing structural information. We present a general method for the expression and purification of tankyrase ARCs from Escherichia coli and outline a fluorescence polarization assay to quantitatively assess direct ARC–TBM peptide interactions. We provide a basic protocol for evaluating binding and poly(ADP-ribosyl)ation of full-length candidate interacting proteins by full-length tankyrase in mammalian cells.

Before start

Note
Experimental approaches such as yeast-two-hybrid analysis [22] or co-immunoprecipitation coupled with tandem mass spectrometry [23] have been used to identify proteins present in complexes with tankyrase. It is desirable to pinpoint those proteins that directly interact with tankyrase. The existence of putative TBMs is the most reliable known indicator for direct interaction. Based on previously identified TBM sequence rules and structural information, we provide a guide to identifying strong TBM candidates for further validation. A previously generated list of ranked octapeptide motifs across the entire human proteome can be used [7]. Here, we recapitulate a simple step-by-step approach for identifying candidate TBMs without the need for specialized bioinformatics beyond the use of established programs and databases.

Steps

Predicting Tankyrase Binders

1.

Screen your protein of interest for potential TBMs. In an initial step, a search for Arg and Gly spaced 6 amino acids apart (i.e., with four amino acids in between, RxxxxG) will suffice. See step 4 below for atypical TBMs.

2.

Using UniProt (http://www.uniprot.org), assess the topology of the protein and whether the candidate motif would be located in the cytoplasm or nucleoplasm, accessible to tankyrase [24]. The relevant information can be found under  Subcellular location > Topology .

3.

Assess the occurrence of disordered stretches in your protein of interest. TBMs must be present in unstructured regions of the protein to be accessible to tankyrase and must not be part of a stable α-helix or β-strand. Secondary structure information, obtained from the Protein Data Bank [25], is available in UniProt (http://www.uniprot.org) under Structure > Secondary structure [24]. It remains possible that TBMs are present within loops of structured protein domains. If no structural information is available, a disorder predictor such as IUPred (http://iupred.enzim.hu) may be used to obtain an indication of protein disorder [26].

4.

Assess the short-listed sequences for their suitability as TBMs. REAGDGEE has been found to be the optimal 8-amino-acid TBM by positional scanning [7], but many TBMs only span six amino acids (positions 1–6). While the sequence rules from positional scanning prove valuable in predicting TBMs, it should be noted that some of their aspects may be specific to the context of the 3BP2 TBM peptide , which positional scanning was performed on [7]. Whereas Arg and Gly at positions 1 and 6, respectively, are essential, other residues are important as well, with the exception of positions 2 and 3, which can likely be ignored since they do not contribute to binding and mainly fulfil spacing roles (Fig.1b, d). Please refer to the motif description under Subheading 1 (Introduction) for the assessment. Of note, a large or strongly charged amino acid at position 4 very likely disqualifies the TBM candidate as a genuine tankyrase binder. Furthermore, Asp and Glu at position 8 can compensate for unfavorable amino acids at other positions [7], as seen in one of the TBMs of AXIN1/2, where a suboptimal Val at position 4 appears to be counterbalanced by a Glu at position 8 (Fig.1e) [7, 11]. Interestingly, a second TBM in AXIN1/2 shows that the Arg, normally at position 1, can be placed further N-terminal, resulting in looping-out of the “inserted” residues (Fig.1c, e). Likewise, the E3 ubiquitin ligase RNF146 has been proposed to contain numerous atypical TBMs with a nonconventional positioning of the Arg [27], and it is possible that a distal Arg is required for weak binding of the TBM of GRB14 to tankyrase (Fig.1e) [7, 8, 28]. Thus, the possibility of a distal Arg residue should be taken into consideration. It is likely that such extended TBMs display a weaker affinity for tankyrase. Numerous tankyrase binders contain multiple TBMs, such as AXIN1/2 [11] and MERIT40 [7] ( see below). Since avidity effects likely contribute to binding of these proteins to tankyrase, the affinity of some of these individual TBMs for the ARC may be lower, and this may be reflected in deviations from the “optimal” sequence.

Fig. 1Substrate binding by tankyrase. (a) Domain organization of human tankyrase and tankyrase 2 (modified from [31]). (b) and (c) Examples for ARC–TBM interactions studied by X-ray crystallography. (b) Human TNKS2 ARC4 is shown in surface representation with bound TBM peptides from 3BP2 and MCL1 shown in stick representation with the core TBM octapeptide colored purple and orange, respectively, and by heteroatom. TBM amino acid positions (1–8) and sequences shown. The figure was generated by superimposing the ARCs of the ARC4-3BP2 and ARC4-MCL1 crystal structures (PDB accession codes 3TWR and 3TWU, respectively) onto each other and showing ARC4 of the former [7]. The colored surface areas represent different contact areas, as indicated, that mediate binding of the TBM peptides (Modified from [7] with permission from Elsevier/Cell Press). (c) ARC2 (from ARC2–3) of murine Tnks bound by the N-terminus of murine Axin1 (PDB accession code 3UTM), which contains two TBMs [11]. Each TBM binds one copy each of ARC2 in a dimeric ARC2-3 assembly. The figure was generated by superimposing the two ARC2-3 copies onto each other; the surface of ARC2 bound by the first TBM is shown. TBMs are shown and labeled as in (b). The first TBM, shown in magenta, consists of a continuous stretch of eight amino acids. In the second TBM, shown in green, the Arg at position 1 is followed by a seven-amino-acid insertion (positions +1 to +7), as indicated in the sequences shown. The peptide insertion forms a loop. (d) TBM sequence rules represented by a sequence logo. (Reprinted from [7] with permission from Elsevier/Cell Press.) (e) Sequence alignment of known example TBMs ([7] and references therein, [11,22,23,37,38,39]), colored by identity with conservation graph, generated with ClustalX and Jalview [40,41]. UniProt IDs are indicated [24]. The asterisk indicates insertion sequences in AXIN1 and AXIN2. The TBMs of 3BP2, TRF1 (TERF1_HUMAN), and MERIT40 (BABA1_HUMAN), studied as model TBMs here, are highlighted
Fig. 1Substrate binding by tankyrase. (a) Domain organization of human tankyrase and tankyrase 2 (modified from [31]). (b) and (c) Examples for ARC–TBM interactions studied by X-ray crystallography. (b) Human TNKS2 ARC4 is shown in surface representation with bound TBM peptides from 3BP2 and MCL1 shown in stick representation with the core TBM octapeptide colored purple and orange, respectively, and by heteroatom. TBM amino acid positions (1–8) and sequences shown. The figure was generated by superimposing the ARCs of the ARC4-3BP2 and ARC4-MCL1 crystal structures (PDB accession codes 3TWR and 3TWU, respectively) onto each other and showing ARC4 of the former [7]. The colored surface areas represent different contact areas, as indicated, that mediate binding of the TBM peptides (Modified from [7] with permission from Elsevier/Cell Press). (c) ARC2 (from ARC2–3) of murine Tnks bound by the N-terminus of murine Axin1 (PDB accession code 3UTM), which contains two TBMs [11]. Each TBM binds one copy each of ARC2 in a dimeric ARC2-3 assembly. The figure was generated by superimposing the two ARC2-3 copies onto each other; the surface of ARC2 bound by the first TBM is shown. TBMs are shown and labeled as in (b). The first TBM, shown in magenta, consists of a continuous stretch of eight amino acids. In the second TBM, shown in green, the Arg at position 1 is followed by a seven-amino-acid insertion (positions +1 to +7), as indicated in the sequences shown. The peptide insertion forms a loop. (d) TBM sequence rules represented by a sequence logo. (Reprinted from [7] with permission from Elsevier/Cell Press.) (e) Sequence alignment of known example TBMs ([7] and references therein, [11,22,23,37,38,39]), colored by identity with conservation graph, generated with ClustalX and Jalview [40,41]. UniProt IDs are indicated [24]. The asterisk indicates insertion sequences in AXIN1 and AXIN2. The TBMs of 3BP2, TRF1 (TERF1_HUMAN), and MERIT40 (BABA1_HUMAN), studied as model TBMs here, are highlighted

推荐阅读

Nature Protocols
Protocols IO
Current Protocols
扫码咨询