D.
La Mendola
*a,
F.
Arnesano
*b,
Ö.
Hansson
c,
C.
Giacomelli
a,
V.
Calò
b,
V.
Mangini
b,
A.
Magrì
d,
F.
Bellia
d,
M. L.
Trincavelli
a,
C.
Martini
a,
G.
Natile
b and
E.
Rizzarelli
d
aDepartment of Pharmacy, University of Pisa, via Bonanno Pisano 6, 56126, Pisa, Italy. E-mail: lamendola@farm.unipi.it
bDepartment of Chemistry, University of Bari “A. Moro”, via E. Orabona 4, 70125 Bari, Italy. E-mail: Fabio.arnesano@uniba.it
cDepartment of Chemistry and Molecular Biology, University of Gothenburg, Medicinaregatan 9C, PO Box 462, SE-40530 Göteborg, Sweden
dInstitute of Biostructure and Bioimaging, CNR, via P. Gaifami 18, 95126 Catania, Italy
First published on 12th November 2015
Angiogenin is a member of the ribonuclease family and a normal constituent of human plasma. It is one of the most potent angiogenic factors known and is overexpressed in different types of cancers. Copper is also an essential cofactor in angiogenesis and, during this process, it is mobilized from inside to outside of the cell. To date, contrasting results have been reported about copper(II) influencing angiogenin activity. However, in these studies, the recombinant form of the protein was used. Unlike recombinant angiogenin, that contains an extra methionine with a free terminal amino group, the naturally occurring protein present in human plasma starts with a glutamine residue that spontaneously cyclizes to pyroglutamate, a lactam derivative. Herein, we report spectroscopic evidence indicating that copper(II) experiences different coordination environments in the two protein isoforms, and affects their RNase and angiogenic activity differently. These results show how relatively small differences between recombinant and wild type proteins can result in markedly different behaviours.
The biological role of Ang is not limited to induction of angiogenesis, as suggested by its widespread expression in all human organs and tissues,5,6 but recent findings demonstrate that Ang is down-regulated both in the mouse model of Parkinson's7 and patients affected by Alzheimer's diseases.8 Moreover, Ang has emerged as one of the key agents in amyotrophic lateral sclerosis (ALS) where it acts as a motoneuron protective factor.9,10
Also copper is known to play a key role in neurodegenerative diseases and has been shown to be an essential angiogenesis cofactor in vivo.11 Serum copper levels are raised in a wide variety of human cancers and correlate with tumor malignancy.12 So far, the targets of Ang activity and the specific role of the metal remain unclear. It is known that during angiogenesis, there is an extracellular translocation of copper,13 thus metal binding to extracellular proteins involved in angiogenesis, such as Ang, is a possible pathway through which copper takes part in the signaling process. Different relationships between Ang and copper have been reported. In one case Cu2+ binding to the protein has been found to increase its interaction with endothelial cells;14,15 the binding details of this interaction have been addressed in a study concerning the formation of copper(II) complexes with a linear peptide encompassing the putative cellular binding site (residues 60–68) of angiogenin.16 In order to probe the copper(II) binding features of the recombinant protein, some peptide fragments of its N-terminal region have been reacted with copper(II) and the results showed similarities as well as differences.17 It has also been suggested that copper and Ang are both angiogenic factors, but through different and independent biological pathways.18 It should be noted, however, that most data reported so far have been obtained using the recombinant form of Ang (r-Ang), containing an extra methionine as the first residue.14,15 In contrast, the wild-type protein (wt-Ang) starts with a glutamine residue which spontaneously cyclizes to pyroglutamate, the γ-lactam form, so that wt-Ang normally present in human plasma has no free amino terminal group.19 Recently, some of us have shown that Cu2+ increases the expression of wt-Ang and modulates its intracellular localization in HUVEC.20 In the present work, Cu2+ binding to wt-Ang has been investigated by means of several techniques, including UV-vis, CD, ESI-MS, EPR, and NMR, and compared to that of r-Ang. The wild-type protein was obtained through the specific enzymatic cleavage of the first Met residue present in the recombinant form.21 Moreover, copper perturbation of the biological activities of the two proteins, such as RNase activity and angiogenesis induction, has been determined.
The wt-Ang was obtained incubating the recombinant protein (7–10 μM) with 1 nM Aeromonas aminopeptidase in 200 mM potassium phosphate buffer, pH 7.2, for 24 h at 37 °C under gentle shaking. The buffer was replaced with Tris-HCl 25 mM, EDTA 1 mM, and NaCl 0.1 M (pH 7.4) by dialysis (Spectra/por MWCO 6–8000) and the reaction mixture subjected to a cation-exchange purification step. The cyclization of the N-terminal glutamine residue was assessed measuring the molecular weight by means of Electrospray Mass spectrometry. The ribonucleolytic activity was determined as described in the literature to confirm the correct folding of the protein.22
Resonance assignment of the apoprotein was carried out with the aid of 2D TOCSY and NOESY along with 3D CBCANH and CBCA(CO)NH experiments, using the available 1H and 15N chemical shift data.23 The titration of the 1.0 mM samples of wt- and r-Ang with CuCl2 (0.1, 0.25, 0.5, 0.75, and 1.0 mol equivalents) at two different pH values (7.4 and 5.5, obtained by addition of HCl to the starting buffer solution) was followed by 1H,15N and 1H,13C heteronuclear single quantum coherence (HSQC) experiments. All spectra were collected on a Bruker Avance 600 with an Ultra Shield Plus magnet using a triple resonance probe equipped with z axis self-shielded gradient coils, and processed using the standard Bruker software (TOPSPIN) and analyzed using the program CARA (The Computer Aided Resonance Assignment Tutorial, R. Keller, 2004, CANTINA Verlag), developed at ETH-Zürich. Cross-peaks affected during Cu2+ titration were identified by comparing their intensities (I) with those of the same cross-peaks (I0) in the dataset of samples lacking Cu2+. The I/I0 ratios as a function of metal-to-protein ratio were fitted to a single-exponential decay function and the obtained decay constants k were plotted as a function of protein sequence.
The small effect on the secondary structure suggests that Cu2+ binding may involve little structural rearrangement in the protein. The Cu2+ titration curve of wt-Ang, shown in the inset of Fig. 1a, indicates that the intensity of the CD band decreases with Cu2+ addition up to one mole equivalent; no further change occurs up to the addition of 2–3 mol equivalents of Cu2+. These data are in accord with a 1:1 metal to ligand stoichiometry. Unlike wt-Ang, the analogous binding curve for r-Ang (inset of Fig. 1b) shows a decrease of the CD band intensity up to the addition of two mol equivalents of Cu2+, suggesting that r-Ang can bind up to two Cu2+ ions.
Electrospray Ionization Mass Spectrometry (ESI-MS) measurements, carried out at physiological pH, confirm the stoichiometry of the metal–protein complexes deduced above. After the addition of only one equivalent of metal ion, both wt-Ang and r-Ang show the signals corresponding to a 1:1 Cu2+–protein adduct. By increasing the amount of copper ion up to ten equivalents, no significant changes are observed for wt-Ang whereas a species with two copper ions per protein molecule is detected for r-Ang (Fig. S2 and S3, ESI†).
The EPR spectrum of wt-Ang in the presence of Cu2+ (1:1 molar ratio, pH 7.4) is typical of type 2 Cu2+–protein complexes (axial g matrix, Cu hyperfine coupling constant >130 × 10−4 cm−1). The calculated EPR parameters are: g∥ = 2.252 and A∥ = 185 × 10−4 cm−1 (Table 1 and Fig. 2).
Fig. 2 EPR spectra of Cu2+–r-Ang (black line) and Cu2+–wt-Ang (red line). [Ang] = [Cu2+] = 0.5 mM, 25 mM MOPS, pH 7.4. |
These Hamiltonian values are distinctly different from those previously reported for r-Ang,17 and indicate the formation of a copper complex in which the metal ion coordination shell may involve three (rather than four) nitrogen atoms with macro-chelate formation in a tetragonally distorted geometry.25,26
The UV-vis CD spectrum recorded at pH 7.4 for the Cu2+ complex formed by wt-Ang displays ligand-to-metal charge-transfer bands at 300 (Δε = −0.25), 326 (Δε = +0.19) and 364 nm (Δε = −0.05) (Fig. 3 and Table 1).
Fig. 3 UV-vis CD spectra at pH 7.4 of Cu2+–wt-Ang (red line) and Cu2+–r-Ang (black line). [Ang] = [Cu2+] = 0.5 mM, 10 mM MOPS, pH 7.4. |
These spectroscopic parameters are the hallmark of the simultaneous involvement of two different protein nitrogen atoms in Cu2+ binding. The first band is diagnostic of a deprotonated amide nitrogen while the second and third bands are originated by the charge transfer from imidazole nitrogen atoms to metal ions,27 clearly suggesting the anchoring role played by the histidine residues. For both proteins the d–d bands show a cross-over signal which is originated by the metal chelate formation. The different chirality around the metal ion (Fig. 3) is attributable to the different main anchoring site, the amino N-terminus for the r-Ang protein and the histidine imidazole for the wt-Ang protein, that determine a different disposition of the binding side chains with respect to the peptide backbone.
The increase of relaxation rates and NMR signal line widths of nuclei close to the paramagnetic center was exploited to obtain insights into the Cu2+ binding sites.28,29 In the case of wt-Ang, large exponential decay constants of signal intensity, as a function of Cu2+ concentration, were found for a large number of residues (Fig. S6a, Tables S1 and S2, ESI†). The most affected signals identify a region of the protein corresponding to the catalytic site, including His-13 and His-114 (Fig. 4a), while residues 60–68 and Asn-109 (the putative endothelial cell-binding site) are only marginally affected.
Fig. 4 Mapping the effects of Cu2+ titration, at pH 7.4, on the NMR signals of wt- (a) and r-Ang (b) (structure: PDB ID 1H52). Residues with an exponential decay constant of signal intensity larger than average plus one standard deviation are colored in red. Histidine residues (blue sticks) and disulfide bonds (yellow sticks) are shown. A red sphere of 10 Å radius is centered on the putative Cu2+ anchoring site. |
These data support the presence of a Cu(2NIm,N−,O) chromophore in the wild-type protein, in which His-114 is the anchoring residue while the neighbouring side chain of His-13 may provide the additional imidazole nitrogen. The involvement of His-8 and His-65 in Cu2+ coordination is less likely, because of the greater distance of their imidazole rings from the anchoring residue in the apoprotein structure and the limited structural rearrangement occurring upon copper binding, as deduced by CD spectra.
In contrast, the addition of CuCl2 to r-Ang mainly affects the signals of residues in the N-terminal region of the protein, including His-8 (Fig. 4b and Fig. S6b, Table S2, ESI†), thus supporting a 4N metal coordination mode (NH2,2N−,NIm). Therefore in r-Ang the Cu2+ anchoring group appears to be the free amino terminus of Met, while the deprotonated amide nitrogens of Glu-1 and Asp-2 and the imidazole nitrogen of His-8 most likely complete the coordination environment of the metal ion. The pH decrease (from 7.4 to 5.5) produces dramatic effects only on the NMR spectra of r-Ang. For this reason, new Cu2+ titration experiments were performed at lower pH (5.5).
While the NMR parameters indicated that the Cu2+-binding site of wt-Ang remained unchanged (Fig. S7a and b, ESI†), those pertinent to r-Ang clearly suggest a conversion to a Cu2+ coordination mode similar to that found for wt-Ang (Fig. S7c and d, ESI†). The terminal amino group of r-Ang is protonated at pH 5.5 and therefore unable to bind Cu2+; as a consequence the imidazole nitrogens of histidine residues present in the catalytic domain become the preferred coordination site for Cu2+, as in the case of wt-Ang. 1H,13C HSQC NMR experiments performed on wt-Ang at low pH and sub-stoichiometric molar ratios of Cu2+ confirm that His-114 serves as an anchoring residue located in the protein core (Fig. S8, ESI†).
Therefore, the catalytic activity of wt- and r-Ang was investigated in the presence of Cu2+ (Fig. 5).
The enzymatic assay was carried out by a modification of the Shapiro et al. procedure.4 In the absence of Cu2+, the two proteins show similar activity, which was comparable to that reported by Holloway et al.22 at pH 7.0. The increase of Cu2+ concentration (within the range used in the literature4) induces a decrease of RNase enzymatic activity in both proteins. This effect is more pronounced for wt-Ang, where His-114 of the catalytic site is the anchoring residue for Cu2+, than for r-Ang where the anchoring group is the free amino terminus. Therefore, a larger amount of Cu2+ has to be added to r-Ang to obtain the same decrease of RNase activity observed in wt-Ang, confirming that r-Ang can bind the metal ion also through sites different from the catalytic one.
The effect of Cu2+ on the RNase activity of the protein may influence its angiogenic activity and, to this end, the capillary-like tube formation test was performed on both r-Ang and wt-Ang in the presence of Cu2+ (Fig. 6).
In the unbound form, the two protein isoforms display the same ability to induce capillary-like tube formation. Conversely, Cu2+ alone, at the used concentrations (100–500 nM), was not able to induce significant tube formation. Noteworthily, the addition of copper ions decreases the activity of both proteins. However, also in this case the effect is much sharper in wt-Ang where it is already evident at an Ang/Cu2+ ratio of 1:1. This may be indirectly correlated with the involvement of His-114 in metal ion binding. To further corroborate this hypothesis, a protein mutant, in which His-114 has been substituted with a tyrosine, has been expressed and subjected to first-methionine removal so as to have a single point mutated form of wt-Ang (H114Y). Noteworthily, the activity of wt-Ang in the presence of Cu2+ was similar to that observed in Ang H114Y (Fig. 6). This confirms that Cu2+ binding to His-114 in wt-Ang inhibits the catalytic process so that its activity becomes similar to that observed for the mutated protein H114Y.
Our data highlight, for the first time, the relevant difference between recombinant and wild type Ang in binding copper and the influence of metal binding on Ang biological activity. The awareness of such a difference entails the need to use the wild-type form for the correct understanding of Ang–copper interaction, and, in turn, the ensuing effects on angiogenic processes.
Footnote |
† Electronic supplementary information (ESI) available: Peptide synthesis and purification; potentiometric and UV-vis characterization of peptide–copper(II) complex species; ESI-MS measurements; and NMR tables. See DOI: 10.1039/c5mt00216h |
This journal is © The Royal Society of Chemistry 2016 |