Light-Activatable Photocaged UNC2025 for Triggering TAM Kinase Inhibition in Bladder Cancer.

Photopharmacology is an emerging field that utilizes photo-responsive molecules to enable control over the activity of a drug using light. The aim is to limit the therapeutic action of a drug at the level of diseased tissues and organs. Considering the well-known implications of protein kinases in cancer and the therapeutic issues associated with protein kinase inhibitors, the photopharmacology is seen as an innovative and alternative solution with great potential in oncology. In this context, we developed the first photocaged TAM kinase inhibitors based on UNC2025, a first-in-class small molecule kinase inhibitor. These prodrugs showed good stability in biologically relevant buffer and rapid photorelease of the photoremovable protecting group upon UV-light irradiation (<10 min.). These light-activatable prodrugs led to a 16-fold decrease to a complete loss of kinase inhibition, depending on the protein and the position at which the coumarin-type phototrigger was introduced. The most promising candidate was the N,O-dicaged compound, showing the superiority of having two photolabile protecting groups on UNC2025 for being entirely inactive on TAM kinases. Under UV-light irradiation, the N,O-dicaged compound recovered its inhibitory potency in enzymatic assays and displayed excellent antiproliferative activity in RT112 cell lines.

Type I arginine methyltransferases are intervention points to unveil the oncogenic Epstein-Barr virus to the immune system.

The oncogenic Epstein-Barr virus (EBV) evades the immune system but has an Achilles heel: its genome maintenance protein EBNA1. Indeed, EBNA1 is essential for viral genome maintenance but is also highly antigenic. Hence, EBV seemingly evolved a system in which the glycine-alanine repeat (GAr) of EBNA1 limits the translation of its own mRNA to the minimal level to ensure its essential function, thereby, at the same time, minimizing immune recognition. Therefore, defining intervention points at which to interfere with GAr-based inhibition of translation is an important step to trigger an immune response against EBV-carrying cancers. The host protein nucleolin (NCL) plays a critical role in this process via a direct interaction with G-quadruplexes (G4) formed in the GAr-encoding sequence of the viral EBNA1 mRNA. Here we show that the C-terminal arginine-glycine-rich (RGG) motif of NCL is crucial for its role in GAr-based inhibition of translation by mediating interaction of NCL with G4 of EBNA1 mRNA. We also show that this interaction depends on the type I arginine methyltransferase family, notably PRMT1 and PRMT3: drugs or small interfering RNA that target these enzymes prevent efficient binding of NCL on G4 of EBNA1 mRNA and relieve GAr-based inhibition of translation and of antigen presentation. Hence, this work defines type I arginine methyltransferases as therapeutic targets to interfere with EBNA1 and EBV immune evasion.

Protein Dimerization via Tyr Residues: Highlight of a Slow Process with Co-Existence of Numerous Intermediates and Final Products.

Protein dimerization via tyrosine residues is a crucial process in response to an oxidative attack, which has been identified in many ageing-related pathologies. Recently, it has been found that for isolated tyrosine amino acid, dimerization occurs through three types of tyrosine-tyrosine crosslinks and leads to at least four final products. Herein, considering two protected tyrosine residues, tyrosine-containing peptides and finally proteins, we investigate the dimerization behavior of tyrosine when embedded in a peptidic sequence. After azide radical oxidation and by combining UPLC-MS and H/D exchange analyzes, we were able to evidence: (i) the slow kinetics of Michael Addition Dimers (MAD) formation, i.e., more than 48 h; (ii) the co-existence of intermediates and final cyclized dimer products; and (iii) the probable involvement of amide functions to achieve Michael additions even in proteins. This raises the question of the possible in vivo existence of both intermediates and final entities as well as their toxicity and the potential consequences on protein structure and/or function.

Acridine-O6-benzylguanine hybrids: Synthesis, DNA binding, MGMT inhibition and antiproliferative activity.

O6-Methylguanine-DNA-methyltransferase (MGMT) is a key DNA repair enzyme involved in chemoresistance to DNA-alkylating anti-cancer drugs such as Temozolomide (TMZ) through direct repair of drug-induced O6-methylguanine residues in DNA. MGMT substrate analogues, such as O6-benzylguanine (BG), efficiently inactivate MGMT in vitro and in cells; however, these drugs failed to reach the clinic due to adverse side effects. Here, we designed hybrid drugs combining a BG residue covalently linked to a DNA-interacting moiety (6-chloro-2-methoxy-9-aminoacridine). Specifically, two series of hybrids, encompassing three compounds each, were obtained by varying the position of the attachment point of BG (N9 of guanine vs. the benzyl group) and the length and nature of the linker. UV/vis absorption and fluorescence data indicate that all six hybrids adopt an intramolecularly stacked conformation in aqueous solutions in a wide range of temperatures. All hybrids interact with double-stranded DNA, as clearly evidenced by spectrophotometric titrations, without intercalation of the acridine ring and do not induce thermal stabilization of the duplex. All hybrids, as well as the reference DNA intercalator (6-chloro-2-methoxy-9-aminoacridine 8), irreversibly inhibit MGMT in vitro with variable efficiency, comparable to that of BG. In a multidrug-resistant glioblastoma cell line T98G, benzyl-linked hybrids 7a-c and the N9-linked hybrid 19b are moderately cytotoxic (GI50 ≥ 15 µM after 96 h), while N9-linked hybrids 19a and 19c are strongly cytotoxic (GI50 = 1-2 µM), similarly to acridine 8 (GI50 = 0.6 µM). Among all compounds, hybrids 19a and 19c, similarly to BG, display synergic cytotoxic effect upon co-treatment with subtoxic doses of TMZ, with combination index (CI) values as low as 0.2-0.3. In agreement with in vitro results, compound 19a inactivates cellular MGMT but, unlike BG, does not induce significant levels of DNA damage, either alone or in combination with TMZ, as indicated by the results of γH2AX immunostaining experiments. Instead, and unlike BG, compound 19a alone induces significant apoptosis of T98G cells, which is not further increased in a combination with TMZ. These results indicate that molecular mechanisms underlying the cytotoxicity of 19a and its combination with TMZ are distinct from that of BG. The strongly synergic properties of this combination represent an interesting therapeutic opportunity in treating TMZ-resistant cancers.

BrdU immuno-tagged G-quadruplex ligands: a new ligand-guided immunofluorescence approach for tracking G-quadruplexes in cells.

G-quadruplexes (G4s) are secondary structures forming in G-rich nucleic acids. G4s are assumed to play critical roles in biology, nonetheless their detection in cells is still challenging. For tracking G4s, synthetic molecules (G4 ligands) can be used as reporters and have found wide application for this purpose through chemical functionalization with a fluorescent tag. However, this approach is limited by a low-labeling degree impeding precise visualization in specific subcellular regions. Herein, we present a new visualization strategy based on the immuno-recognition of 5-bromo-2'-deoxyuridine (5-BrdU) modified G4 ligands, functionalized prior- or post-G4-target binding by CuAAC. Remarkably, recognition of the tag by antibodies leads to the detection of the modified ligands exclusively when bound to a G4 target both in vitro, as shown by ELISA, and in cells, thereby providing a highly efficient G4-ligand Guided Immunofluorescence Staining (G4-GIS) approach. The obtained signal amplification revealed well-defined fluorescent foci located in the perinuclear space and RNase treatment revealed the preferential binding to G4-RNA. Furthermore, ligand treatment affected significantly BG4 foci formation in cells. Our work headed to the development of a new imaging approach combining the advantages of immunostaining and G4-recognition by G4 ligands leading to visualization of G4/ligands species in cells with unrivaled precision and sensitivity.

Oxidative radicals (HO• or N3•) induce several di-tyrosine bridge isomers at the protein scale.

Among protein oxidative damages, di-tyrosine bridges formation has been evidenced in many neuropathological diseases. Combining oxidative radical production by gamma radiolysis with very performant chromatographic separation coupled to mass spectrometry detection, we brought into light new insights of tyrosine dimerization. Hydroxyl and azide radical tyrosine oxidation leading to di-tyrosine bridges formation was studied for different biological compounds: a full-length protein (Δ25-centrin 2), a five amino acid peptide (KTSLY) and free tyrosine. We highlighted that both radicals generate high proportion of dimers even for low doses. Surprisingly, no less than five different di-tyrosine isomers were evidenced for the protein and the peptide. For tyrosine alone, at least four distinct dimers were evidenced. These results raise some questions about their respective role in vivo and hence their relative toxicity. Also, as di-tyrosine is often used as a biomarker, a better knowledge of the type of dimer detected in vivo is now required.

A decrease in NAMPT activity impairs basal PARP-1 activity in cytidine deaminase deficient-cells, independently of NAD.

Cytidine deaminase (CDA) deficiency causes pyrimidine pool disequilibrium. We previously reported that the excess cellular dC and dCTP resulting from CDA deficiency jeopardizes genome stability, decreasing basal poly(ADP-ribose) polymerase 1 (PARP-1) activity and increasing ultrafine anaphase bridge (UFB) formation. Here, we investigated the mechanism underlying the decrease in PARP-1 activity in CDA-deficient cells. PARP-1 activity is dependent on intracellular NAD+ concentration. We therefore hypothesized that defects of the NAD+ salvage pathway might result in decreases in PARP-1 activity. We found that the inhibition or depletion of nicotinamide phosphoribosyltransferase (NAMPT), the rate-limiting enzyme in the NAD+ salvage biosynthesis pathway, mimicked CDA deficiency, resulting in a decrease in basal PARP-1 activity, regardless of NAD+ levels. Furthermore, the expression of exogenous wild-type NAMPT fully restored basal PARP-1 activity and prevented the increase in UFB frequency in CDA-deficient cells. No such effect was observed with the catalytic mutant. Our findings demonstrate that (1) the inhibition of NAMPT activity in CDA-proficient cells lowers basal PARP-1 activity, and (2) the expression of exogenous wild-type NAMPT, but not of the catalytic mutant, fully restores basal PARP-1 activity in CDA-deficient cells; these results strongly suggest that basal PARP-1 activity in CDA-deficient cells decreases due to a reduction of NAMPT activity.

Proper chromosome alignment depends on BRCA2 phosphorylation by PLK1.

The BRCA2 tumor suppressor protein is involved in the maintenance of genome integrity through its role in homologous recombination. In mitosis, BRCA2 is phosphorylated by Polo-like kinase 1 (PLK1). Here we describe how this phosphorylation contributes to the control of mitosis. We identify a conserved phosphorylation site at T207 of BRCA2 that constitutes a bona fide docking site for PLK1 and is phosphorylated in mitotic cells. We show that BRCA2 bound to PLK1 forms a complex with the phosphatase PP2A and phosphorylated-BUBR1. Reducing BRCA2 binding to PLK1, as observed in BRCA2 breast cancer variants S206C and T207A, alters the tetrameric complex resulting in unstable kinetochore-microtubule interactions, misaligned chromosomes, faulty chromosome segregation and aneuploidy. We thus reveal a role of BRCA2 in the alignment of chromosomes, distinct from its DNA repair function, with important consequences on chromosome stability. These findings may explain in part the aneuploidy observed in BRCA2-mutated tumors.

Nucleolin Discriminates Drastically between Long-Loop and Short-Loop Quadruplexes.

We investigate herein the interaction between nucleolin (NCL) and a set of G4 sequences derived from the CEB25 human minisatellite that adopt a parallel topology while differing in the length of the central loop (from nine nucleotides to one nucleotide). It is revealed that NCL strongly binds to long-loop (five to nine nucleotides) G4 while interacting weakly with the shorter variants (loop with fewer than three nucleotides). Photo-cross-linking experiments using 5-bromo-2'-deoxyuridine (BrU)-modified sequences further confirmed the loop-length dependency, thereby indicating that the WT-CEB25-L191 (nine-nucleotide loop) is the best G4 substrate. Quantitative proteomic analysis (LC-MS/MS) of the product(s) obtained by photo-cross-linking NCL to this sequence enabled the identification of one contact site corresponding to a 15-amino acid fragment located in helix α2 of RNA binding domain 2 (RBD2), which sheds light on the role of this structural element in G4-loop recognition. Then, the ability of a panel of benchmark G4 ligands to prevent the NCL-G4 interaction was explored. It was found that only the most potent ligand PhenDC3 can inhibit NCL binding, thereby suggesting that the terminal guanine quartet is also a strong determinant of G4 recognition, putatively through interaction with the RGG domain. This study describes the molecular mechanism by which NCL recognizes G4-containing long loops and leads to the proposal of a model implying a concerted action of RBD2 and RGG domains to achieve specific G4 recognition via a dual loop-quartet interaction.

How protein structure affects redox reactivity: example of Human centrin 2.

Electron transfer inside proteins plays a central role in their reactivity and biological functions. Herein, we developed a combined approach by gamma radiolysis and electrochemistry which allowed a deep insight into the reactivity of Human centrin 2, a protein very sensitive to oxidative stress and involved in several key biological processes. This protein bears a single terminal tyrosine and was observed to be extremely sensitive to ionizing radiation sources, leading to a tyrosine dimer. By cyclic voltammetry in the 100-1000 V s(-1) range, its redox potential and dimerization rate could be evaluated. Accordingly, reaction in solution with a redox mediator revealed an efficient catalysis. Finally, protein denaturation by a progressive increase in temperature was proportional to a decrease of dimerization radiolytic yield. Our results thus demonstrated that the protein structure plays a major role in oxidation sensitivity. This leads to meaningful results to understand protein redox reactivity.

"On-the-fly" kinetics of enzymatic racemization using deuterium NMR in DNA-based chiral oriented media.

We report the in situ and real-time monitoring of the interconversion of L- and D-alanine-d3 by alanine racemase from Bacillus stearothermophilus directly observed by (2)H NMR spectroscopy in anisotropic phase. The enantiomers are distinguished by the difference of their (2)H quadrupolar splittings in a chiral liquid crystal containing short DNA fragments. The proof-of-principle, the reliability, and the robustness of this new method is demonstrated by the determination of the turnover rates of the enzyme using the Michaelis-Menten model.

Oxidative stress induces mainly human centrin 2 polymerisation.

PURPOSE: To determine the human centrin 2 (Hscen 2) protein response to oxidising radicals in vitro and to evaluate the consequences on its biological functions. MATERIALS AND METHODS: Hscen 2 was submitted to hydroxyl and azide radicals produced by radiolysis in the absence of oxygen. The resulting products were characterised by biochemical, spectroscopic and mass spectrometry techniques. Their thermodynamics parameters of complexation with C-terminal fragment of Xeroderma pigmentosum C protein (C-XPC), one of the Hscen 2 cellular partners, were quantified by isothermal titration calorimetry (ITC). RESULTS: Both hydroxyl and azide radicals induce centrin 2 polymerisation as we characterised several intermolecular cross-links generating dimers, trimers, tetramers and higher molecular mass species. These cross-links result from the formation of a covalent bond between the only tyrosine residue (Tyr 172) located in the C-terminal region of each monomer. Remarkably, dimerisation occurs for doses as low as a few grays. Moreover, this Hscen2 dimer has a lower affinity and stoechiometry binding to C-XPC. CONCLUSIONS: These results show that as oxidative radicals induce high proportions of irreversible damages (polymerisation) centrin 2 is highly sensitive to ionising radiation. This could have important consequences on its biological functions.

The carboxy-terminal domain of xeroderma pigmentosum complementation group C protein, involved in TFIIH and centrin binding, is highly disordered.

Xeroderma pigmentousum group C protein (XPC) is involved in the first step of nucleotide excision repair, with multiple functional roles including DNA damage recognition and recruitment of the repair machinery. This human protein of 940 residues forms a strong heterotrimeric complex with Rad23B and centrin 2. The structure of XPC is actually not known, and lack of significant sequence homology with proteins from structural data bases precludes any relevant prediction. Here, we present the molecular and structural characterization of a C-terminal fragment of XPC (C-XPC: 126 residues, 815-940), which was shown to be involved in centrin 2 and TFIIH binding. C-XPC may be highly expressed in E. coli, but because of its limited solubility it was purified under 6 M urea. Using bioinformatics tools, and a combination of several experimental methods (circular dichroism, fluorescence, nuclear magnetic resonance, and small-angle X-ray scattering), we show that C-XPC has a highly flexible structure under native physiological conditions, with a propensity to form helical secondary structures. Isothermal titration calorimetry experiments show that the C-XPC fragment binds human centrin 2 with high affinity and a 1:1 stoichiometry. NMR analysis indicates that the physical interaction between C-XPC and centrin 2 induces only minor conformational changes into XPC, localized around the 17-mer segment (847-863), showed to be critically involved in the centrin binding.

Structural, thermodynamic, and cellular characterization of human centrin 2 interaction with xeroderma pigmentosum group C protein.

Human centrin 2 (HsCen2), an EF-hand calcium binding protein, plays a regulatory role in the DNA damage recognition during the first steps of the nucleotide excision repair. This biological action is mediated by the binding to a short fragment (N847-R863) from the C-terminal region of xeroderma pigmentosum group C (XPC) protein. This work presents a detailed structural and energetic characterization of the HsCen2/XPC interaction. Using a truncated form of HsCen2 we obtained a high resolution (1.8 A) X-ray structure of the complex with the peptide N847-R863 from XPC. Structural and thermodynamic analysis of the interface revealed the existence of both electrostatic and apolar inter-molecular interactions, but the binding energy is mainly determined by the burial of apolar bulky side-chains into the hydrophobic pocket of the HsCen2 C-terminal domain. Binding studies with various peptide variants showed that XPC residues W848 and L851 constitute the critical anchoring side-chains. This enabled us to define a minimal centrin binding peptide variant of five residues, which accounts for about 75% of the total free energy of interaction between the two proteins. Immunofluorescence imaging in HeLa cells demonstrated that HsCen2 binding to the integral XPC protein may be observed in living cells, and is determined by the same interface residues identified in the X-ray structure of the complex. Overexpression of XPC perturbs the cellular distribution of HsCen2, by inducing a translocation of centrin molecules from the cytoplasm to the nucleus. The present data confirm that the in vitro structural features of the centrin/XPC peptide complex are highly relevant to the cellular context.

High sensitivity of human centrin 2 toward radiolytical oxidation: C-terminal tyrosinyl residue as the main target.

Centrins are calcium-binding proteins that play a significant role in the maintenance of the centrosomal organization, mainly in the continuity between centrosome and microtubular network. Recent data showed that centrosome duplication abnormalities, like overduplication for example, could be due to hydrogen peroxide, suggesting an important impact of oxidative stress. To challenge this hypothesis, we performed one-electron oxidation experiments with human centrin 2, starting from azide radicals. Our results first revealed several intermolecular cross-links generating dimers, tetramers, hexamers, and higher molecular mass species. Dimers result from covalent bond linking the C-terminal tyrosines of each monomer. Second, the methionyl residue at position 19 was oxidized on the monomeric centrin. Further, electron microscopy experiments on centrin 2 showed a preexisting hexameric organization that was stabilized by covalent bonds as a result of irradiation. Overall, these results show that centrin 2 is highly sensitive to ionizing radiation, which could have important consequences on its biological functions.

Slow backbone dynamics of the C-terminal fragment of human centrin 2 in complex with a target peptide probed by cross-correlated relaxation in multiple-quantum NMR spectroscopy.

The C-terminal domain of human centrin 2 (C-HsCen2) strongly binds to P1-XPC, a peptide comprising 17 amino acids with a NWKLLAKGLLIRERLKR sequence. This peptide corresponds to residues N847-R863 of XPC, a protein involved in the recognition of damaged DNA during the initial step of the nucleotide excision repair pathway. The slow internal dynamics of the protein backbone in the C-HsCen-P1-XPC complex was studied by measuring the relaxation rates of zero- and double-quantum coherences involving neighboring pairs of carbonyl 13C and amide 15N nuclei. These relaxation rates, which reflect dynamics on time scales in the range of micro- to milliseconds, vary significantly along the protein backbone. Analysis of the relaxation rates at different CaCl2 concentrations and ionic strengths shows that these slow motions are mainly affected by the binding of a Ca2+ ion to the lower-affinity EF-hand III. Moreover, we discuss the possible functional role of residues that undergo differential exchange in the formation of HsCen homodimers.

Binding of human centrin 2 to the centrosomal protein hSfi1.

hSfi1, a human centrosomal protein with homologs in other eukaryotic organisms, includes 23 repeats, each of 23 amino acids, separated by 10 residue linkers. The main molecular partner in the centrosome is a small, calcium-binding EF-hand protein, the human centrin 2. Using isothermal titration calorimetry experiments, we characterized the centrin-binding capacity of three isolated hSfi1 repeats, two exhibiting the general consensus motif and the third being the unique Pro-containing human repeat. The two standard peptides bind human centrin 2 and its isolated C-terminal domain with high affinity (approximately 10(7) M(-1)) by an enthalpy-driven mechanism, with a moderate Ca2+ dependence. The Pro-containing repeat shows a binding affinity that is two orders of magnitude lower. The target binding site is localized within the C-terminal domain of human centrin 2. Fluorescence titration and NMR spectroscopy show that the well-conserved Trp residue situated in the C-terminus of each repeat is deeply embedded in a protein hydrophobic cavity, indicating that the peptide direction is reversed relative to previously studied centrin targets. The present results suggest that almost all of the repeats of the Sfi1 protein may independently bind centrin molecules. On the basis of this hypothesis and previous studies on centrin self-assembly, we propose a working model for the role of centrin-Sfi1 interactions in the dynamic structure of centrosome-associated contractile fibers.

Crystallization and preliminary X-ray diffraction data of the complex between human centrin 2 and a peptide from the protein XPC.

Centrins are highly conserved calcium-binding proteins involved in the nucleotide-excision repair pathway as a subunit of the heterotrimer including the XPC and hHR23B proteins. A complex formed by a Ca2+-bound human centrin 2 construct (the wild type lacking the first 25 amino acids) with a 17-mer peptide derived from the XPC sequence (residues Asn847-Arg863) was crystallized. Data were collected to 1.65 angstroms resolution from crystals grown in 30% monomethyl polyethylene glycol (MPEG) 500, 100 mM NaCl and 100 mM Bicine pH 9.0. Crystals are monoclinic and belong to space group C2, with two molecules in the asymmetric unit. The unit-cell parameters are a = 60.28, b = 59.42, c = 105.14 angstroms, alpha = gamma = 90, beta = 94.67 degrees. A heavy-atom derivative was obtained by co-crystallization with Sr2+. The substitution was rationalized by calorimetry experiments, which indicate a binding constant for Sr2+ of 4.0 x 10(4) M(-1).

Flexibility and plasticity of human centrin 2 binding to the xeroderma pigmentosum group C protein (XPC) from nuclear excision repair.

Human centrin 2 is a component of the nucleotide excision repair system, as a subunit of the heterotrimer including xeroderma pigmentosum group C protein (XPC) and hHR23B. The C-terminal domain of centrin (C-HsCen2) binds strongly a peptide from the XPC protein (P1-XPC: N(847)-R(863)). Here, we characterize the solution Ca(2+)-dependent structural and molecular features of the C-HsCen2 in complex with P1-XPC, mainly using NMR spectroscopy and molecular modeling. The N-terminal half of the peptide, organized as an alpha helix is anchored into a deep hydrophobic cavity of the protein, because of three bulky hydrophobic residues in position 1-4-8 and electrostatic contacts with the centrin helix E. Investigation of the whole centrin interactions shows that the N-terminal domain of the protein is not involved in the complex formation and is structurally independent from the peptide-bound C-terminal domain. The complex may exist in three different binding conformations corresponding to zero, one, and two Ca(2+)-bound states, which may exchange with various rates and have distinct structural stability. The various features of the intermolecular interaction presented here constitute a centrin-specific mode for the target binding.

The N-terminal domain of human centrin 2 has a closed structure, binds calcium with a very low affinity, and plays a role in the protein self-assembly.

Centrins are well-conserved calcium binding proteins from the EF-hand superfamily implicated in various cellular functions, such as centrosome duplication, DNA repair, and nuclear mRNA export. The intrinsic molecular flexibility and the self-association tendency make difficult the structural characterization of the integral protein. In this paper we report the solution structure, the Ca2+ binding properties, and the intermolecular interactions of the N-terminal domain of two human centrin isoforms, HsCen1 and HsCen2. In the absence of Ca2+, the N-terminal construct of HsCen2 revealed a compact core conformation including four almost antiparallel alpha-helices and a short antiparallel beta-sheet, very similar to the apo state structure of other calcium regulatory EF-hand domains. The first 25 residues show a highly irregular and dynamic structure. The three-dimensional model for the N-terminal domain of HsCen1, based on the high sequence conservation and NMR spectroscopic data, shows very close structural properties. Ca2+ titration of the apo-N-terminal domain of HsCen1 and HsCen2, monitored by NMR spectroscopy, revealed a very weak affinity (10(2)-10(3) M(-1)), suggesting that the cellular role of this domain is not calcium dependent. Isothermal calorimetric titrations showed that an 18-residue peptide, derived from the N-terminal unstructured fragment, has a significant affinity (approximately 10(5) M(-1)) for the isolated C-terminal domain, suggesting an active role in the self-assembly of centrin molecules.