Regular article
Screening of a library of phage-displayed peptides identifies human Bcl-2 as a taxol-binding protein1

https://doi.org/10.1006/jmbi.1998.2303Get rights and content

Abstract

A random library of phage displayed peptides was screened for binding to a biotinylated derivative of paclitaxel (Taxol). Affinity-selected peptides were analyzed for similarity to human proteins. There was no significant similarity between the paclitaxel-selected peptides and tubulin. However, a subset of the peptides was identified that exhibits significant similarity to a non-conserved region of the anti-apoptotic human protein Bcl-2: ELISA assays confirmed binding of paclitaxel to Bcl-2, and circular dichroism spectroscopy demonstrated that a substantial conformational change accompanies this binding. In vivo, treatment with paclitaxel has been shown to lead to Bcl-2 inactivation with concomitant phosphorylation of residues in a disordered, regulatory loop region of the protein. Similarity between paclitaxel-selected peptides and this loop region implicate these residues in drug binding, and suggest that the apoptotic action of paclitaxel may involve the binding of paclitaxel to Bcl-2. These results demonstrate that peptides displayed on the surface of bacteriophage particles can mimic the ligand-binding properties of disordered regions of proteins.

Introduction

Paclitaxel (trade name Taxol) is a highly effective anti-neoplastic agent known to halt mitosis and induce apoptosis. Its anti-mitotic activity is a direct effect of its induction of tubulin polymerization and suppression of microtubule dynamics (Schiff et al., 1979). Its apoptotic activity is presumed to be a secondary effect brought on by the mitotic block. A number of other microtubule-active drugs are now known to act by mechanisms that appear similar (Haldar et al., 1997). A key step in the induction of apoptosis by these drugs is the phosphorylation and concomitant inactivation of the anti-apoptotic protein Bcl-2 Haldar et al 1995, Haldar et al 1996. Numerous recent papers have suggested the involvement of several cellular pathways in paclitaxel action, and characterization of these pathways is critical to a complete understanding of the biological activity of paclitaxel.

To characterize better the types of interactions paclitaxel makes with proteins, a phage displayed peptide library was screened for members with relatively high affinity for a derivative of paclitaxel biotinylated at C7 of the taxane ring. The dodecapeptides in this library are attached to the N terminus of pIII, the host-binding protein of bacteriophage M13. Their conformation and environment on pIII is unknown. Some of the displayed peptides may interfere with infection of host by the phage, resulting in censorship of the library. Other metabolic and structural effects also lead to censorship of phage libraries (Rodi & Makowski, 1997). For instance, cysteine is present at very low levels in the library because of the action of the Escherichia coli dsb system which catalyzes the formation of disulfide bonds in the periplasm, leading to the covalent dimerization of pIII which precludes their incorporation into the assembling phage particles (Makowski & Russel, 1997). In spite of these factors, the library has very substantial sequence diversity, containing a significant fraction of the theoretically possible sequences (A. Soares & L.M. unpublished results). Nevertheless, the resulting censorship patterns must be taken into account in any statistical analysis of an affinity-selection process.

Screening of the library (Scott & Smith, 1990) by biopanning (Kay et al., 1996) involves incubation of the library above an immobilized ligand, in this case a biotinylated derivative of paclitaxel bound to a streptavidin coated plate. Peptides selected by biopanning are a mixture of those exhibiting preferential affinity for paclitaxel, those exhibiting favorable growth properties, and those randomly binding non-specifically to the substrate. Identification of those peptides with preferential affinity for paclitaxel requires the observation of patterns of sequence that are not observed in unselected populations. If the affinity is relatively weak, identification of these sequence patterns may require sequencing of significant numbers of selected peptides.

It is unclear to what extent the sequences of phage-displayed peptides with affinity for paclitaxel may reflect the sequences of proteins that bind to paclitaxel. The binding properties of peptides depend heavily on their environment, and there is no reason to believe that the environment at the N terminus of pIII of a filamentous bacteriophage would mimic in any way that of a paclitaxel binding site on a naturally occurring protein. Furthermore, the binding sites for small molecule ligands usually involve several short stretches of peptide separated by widely varying lengths of peptide, and arranged in a protein scaffold to form a well-defined ligand binding site. These facts notwithstanding, many ligand-binding sites are disordered loops on the surface of a protein prior to interaction with the ligand Dunker et al 1998, Romero et al 1998, and the importance of the surrounding scaffold and other aspects of the environment are less obvious in these cases. Therefore, a comparison of the sequences of phage-displayed peptides exhibiting affinity for paclitaxel with the sequences of proteins, such as β-tubulin, that are known to bind paclitaxel may be informative as to the ways in which this drug interacts with proteins.

Section snippets

Library screening

Filamentous bacteriophage particles whose recombinant pIII proteins might bind to paclitaxel were isolated by biopanning (Kay et al., 1996) as detailed in Materials and Methods using a derivatized paclitaxel with a biotin group covalently attached to C7 (see Figure 3) and immobilized on a streptavidin-coated plate. The five pIII structural proteins present at the tip of the virion each possess a random 12 amino acid extension of their amino terminus, coded for by a random synthetic

Discussion

The results presented here demonstrate that similarity between the amino acid sequences of ligand-selected peptides and that of an intact protein can be predictive for the binding of the ligand to that protein. The disordered loop of Bcl-2 appears to have ligand-binding properties adequately similar to those of peptides displayed on the surface of a bacteriophage that similar sequences in the two environments result in similar binding properties. Lower similarity between the selected peptide

Library screening

Thirty micrograms of C7-biotinylated paclitaxel (molecular mass 1118 Da, unpublished synthesis), solubilized in HPLC-grade dimethylsulfoxide and diluted into Tris-buffered saline plus 0.1 % (v/v) Tween 20 (TBST.1), was attached to a streptavidin-coated petri dish according to standard procedures (Kay et al., 1996). A random 12 amino acid pIII display library was obtained from New England Biolabs (NEB PHD-12) and screened according to recommended instructions with the following modifications.

Acknowledgements

We thank C. B. Thompson at the University of Chicago for purified human Bcl-XL, M. Glucksman at Mount Sinai School of Medicine for purified S. japonicum GST, Alex Soares at Florida State University for help with statistical analyses, and G. Myers at CBI for helpful discussions. This work was funded by a grant from the Lucille P. Markey Foundation and a grant from the National Science Foundation (L.M.).

References (30)

  • S. Haldar et al.

    Inactivatin of Bcl-2 by phosphorylation

    Proc. Natl Acad. Sci. USA

    (1995)
  • S. Haldar et al.

    Paclitaxel induces Bcl-2 phosphorylation and death of prostate cancer cells

    Cancer Res.

    (1996)
  • S. Haldar et al.

    Bcl2 is the guardian of microtubule integrity

    Cancer Res.

    (1997)
  • S. Haldar et al.

    Serine-70 is one of the critical sites for drug-induced Bc12 phosphorylation in cancer cells

    Cancer Res.

    (1998)
  • A.M. Ibrado et al.

    Bcl-XL overexpression inhibits paclitaxel-inducing Yama protease activity and apoptosis

    Cell. Growth Differ.

    (1996)
  • Cited by (226)

    • NCS-1 is a regulator of calcium signaling in health and disease

      2018, Biochimica et Biophysica Acta - Molecular Cell Research
      Citation Excerpt :

      The main mechanism of action of paclitaxel is the induction of mitotic arrest through stabilization and polymerization of microtubules [105]. Interaction with several proteins other than tubulin, including heat shock proteins and the antiapoptotic protein Bcl-2 [106,107], has been reported but the functional consequences of these interactions are unclear [108]. In contrast, NCS-1 was found to be a binding partner of paclitaxel [90] and this binding was shown to initiate an undesired side effect of paclitaxel therapy, CIPN.

    • Neurotoxic mechanisms of paclitaxel are local to the distal axon and independent of transport defects

      2017, Experimental Neurology
      Citation Excerpt :

      As a first step in defining mechanistic pathways that lead to paclitaxel neurotoxicity, we tested the hypothesis that increased microtubule stability, rather than off-target effects, is the primary cause of paclitaxel neurotoxicity. Although a change in microtubule dynamics after paclitaxel exposure has been a dominant hypothesis to explain paclitaxel neurotoxicity, alternative paclitaxel binding targets have been described: neuronal calcium sensor-1 and the apoptosis regulator Bcl-2 (Boehmerle et al., 2006; Ferlini et al., 2009; Rodi et al., 1999). To test this hypothesis, we examined epothilone B, a microtubule stabilizing compound that binds to a site on tubulin that overlaps the taxane site but is structurally different from paclitaxel; epothilone B would therefore be unlikely to have the same off-target effects (Goodin et al., 2004).

    • Graph-Based Motif Discovery in Mimotope Profiles of Serum Antibody Repertoire

      2023, Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
    View all citing articles on Scopus
    1

    Edited by I. A. Wilson

    View full text