- freely available
Int. J. Mol. Sci. 2012, 13(6), 7354-7364; doi:10.3390/ijms13067354
Published: 14 June 2012
Abstract: CV_2116 is a small hypothetical protein of 82 amino acids from the Gram-negative coccobacillus Chromobacterium violaceum. A PSI-BLAST search using the CV_2116 sequence as a query identified only one hit (E = 2e−07) corresponding to a hypothetical protein OR16_04617 from Cupriavidus basilensis OR16, which failed to provide insight into the function of CV_2116. The CV_2116 gene was cloned into the p15TvLic expression plasmid, transformed into E. coli, and 13C- and 15N-labeled NMR samples of CV_2116 were overexpressed in E. coli and purified for structure determination using NMR spectroscopy. The resulting high-quality solution NMR structure of CV_2116 revealed a novel α + β fold containing two anti-parallel β-sheets in the N-terminal two-thirds of the protein and one α-helix in the C-terminal third of the protein. CV_2116 does not belong to any known protein sequence family and a Dali search indicated that no similar structures exist in the protein data bank. Although no function of CV_2116 could be derived from either sequence or structural similarity searches, the neighboring genes of CV_2116 encode various proteins annotated as similar to bacteriophage tail assembly proteins. Interestingly, C. violaceum exhibits an extensive network of bacteriophage tail-like structures that likely result from lateral gene transfer by incorporation of viral DNA into its genome (prophages) due to bacteriophage infection. Indeed, C. violaceum has been shown to contain four prophage elements and CV_2116 resides in the fourth of these elements. Analysis of the putative operon in which CV_2116 resides indicates that CV_2116 might be a component of the bacteriophage tail-like assembly that occurs in C. violaceum.
Chromobacterium violaceum is a Gram-negative non-sporulating coccobacillus. It produces a natural antibiotic called violacein, which may be useful for treatment of colon and other cancers . The complete genome of C. violaceum, which was published in 2003 , contains four identified integrated viral genomes (prophages) . Among these prophages, CvP4 is composed of 51 ORFs spanning CV2114 to CV2150, including CV_2116, which encodes a hypothetical protein of 82 residues (UniProtKB/TrEMBL ID Q7NW74_CHRVO). To date, CV_2116 does not belong to any known protein families (Pfam 26.0, November, 2011) and is simply annotated as a putative uncharacterized protein . This hypothetical protein was selected for structure determination as a target by the Northeast Structural Genomics (NESG) consortium (NESG Target ID CvT4). A goal of the United States National Institutes of Health’s Protein Structure Initiative (PSI) during the 10-year period spanning the PSI-1 and PSI-2 phases was to establish an understanding of all protein sequence/protein fold relationships that occur in nature with an emphasis on determining at least one experimental protein structure for every identified protein sequence family. However, not all proteins predicted by genome sequencing efforts belong to existing sequence families. Such proteins are referred to as “singletons”, which is currently the case for CV_2116. Therefore, despite not belonging to an existing protein family, knowledge of the three-dimensional structure of the hypothetical CV_2116 protein helps to fill in a missing part of the sequence-protein fold space that occurs in nature.
2. Results and Discussion
A sequence similarity search against CV_2116 using PSI-BLAST identified only one hit, a hypothetical protein OR16_04617 from Cupriavidus basilensis OR16 (E-value of 2e−07) with sequence identity of 43% over all 82 residues. Likewise, a PSI-BLAST search using OR16_04617 as the query identified CV_2116 as the only significant sequence match. Figure 1 shows the sequence alignment of CV_2116 to OR16_04617. The other two proteins identified by PSI-BLAST had E-values greater than 1.0, including GFO_0138 (E-value of 1.3) from Gramella forsetti (UniProtKB/TrEMBL ID A0LXN2_GRAFK)  with sequence identity of 33% over 60 residues, and Ribonuclease HII (E-value of 6.3) from Bacillus coagulans (UniProtKB/TrEMBL ID G2TJM8_BACCO)  with sequence identity of 36% over 42 resides. The low sequence similarity to other proteins indicated that CV_2116 is not a highly conserved protein domain, and the single PSI-BLAST hit to OR16_04617 failed to provide any insight into the function of CV_2116.
A high quality solution structure of CV_2116 was determined by NMR spectroscopy. CV_2116 exhibited an excellent 1H-15N-HSQC spectrum (Figure 2A) indicative of a well-behaved and structured protein in solution. The rotational correlation time of CV_2116 was determined from average 15N relaxation times extracted from 1D 15N-edited T1 and T2 experiments of the NC sample recorded on a 600 MHz spectrometer at 298 K. The overall isotropic rotational correlation time (τc) of 6.2 ± 0.5 ns derived from T1 and T2 measurements confirmed its monomeric state in solution based on a curve of molecular weight versus τc from other proteins characterized in our laboratory with molecular weights under 20 kDa as shown in Figure 2B .
Figure 3A shows the stereoview of the superimposition of the 20 lowest energy structures of CV_2116. The secondary structural elements are locally and globally well defined. CV_2116 contains two anti-parallel β sheets in the N-terminal two-thirds of the protein and one α helix in the C-terminal third of the protein (Figure 3B). The first β sheet is composed of four β-strands (β1, A4-Y6; β2, Y9-E12; β5, L29-K32; β6, I41-T44) in the order of β1-β2-β5-β6. The short β6 β-strand is present, as predicted by PROCHECK 3.5.4 [9,10], in 13 out of 20 structures in the final NMR ensemble.
Two β-strands (β3, H15-R18 and β4, K23-P26) compose the second anti-parallel β sheet. The 21-residue helix (E52-D72) in the C-terminal third of the protein flanks the two anti-parallel β sheets. Analysis of the NOEs and tertiary structure of CV_2116 revealed that Y6 from β1, Y9 from β2, I30 from β5, T44 from β6, and A63, I66, and I67 from α-helix constitute the hydrophobic core contacts to the first β sheet, whereas few NOEs between Y24 from β4 and A55 from the α-helix were observed to the second β sheet. Two longer loops, one between β5 and β6 (V33-P40) and one between β6 and α-helix (E52-D72), and the C-terminal tail (R73-G82) were flexible and disordered in the NMR structure. Structural statistics for CV_2116 are presented in Table 1.
CV_2116 adopted an apparently novel fold without any current matches in the CATH  and SCOP  structural classification databases. Comparison of the three-dimensional structure of CV_2116 using Dali (v.3)  revealed little similarity to any known protein structures in the PDB. Three proteins, peptide chain release factor (PDB ID: 3D5C-X), UDP-N-acetylmuramate-alanine ligase (PDB ID: 1P31-A), and hyponastic leave 1 (PDB ID: 2L2N-A), had the highest Z scores of 3.7, despite the fact that they had low sequence identities of 8%, 8% and 5%, respectively. No biochemical or biological functions could be derived from the structural comparisons with the known structures in the PDB.
The electrostatic surface diagram of CV_2116 is shown in Figure 3C. Positive and negative regions were observed on one side of the protein while one negative region was present on the opposite side. So far, no potential functions of these regions have been identified.
Although the function of the hypothetical CV_2116 protein could not be predicted by either sequence or structure similarity searches, the CV_2116 gene from C. violaceum is sandwiched between several neighboring genes that encode for apparent bacteriophage tail-like assembly proteins. In fact, C. violaceum has been observed to exhibit an unusual and extensive organization of bacteriophage tail-like particles . The locus two genes before CV_2116, namely CV_2114, encodes a probable pyocin R2_PP tail fiber protein. R-type pyocins are a component of the non-flexible tail structure of bacteriophage  The gene immediately preceding CV_2116, namely CV_2115, encodes a protein homologous to other bacteriophage J tail proteins belonging to the protein family PF04865, which are involved in baseplate assembly. Among the genes immediately following CV_2116, CV_2117 encodes a protein homologous to other bacteriophage V proteins (PF04717), which is also involved in baseplate assembly  and CV_2118 is homologous to several members of PF05954, which are phage late control gene D proteins. Hence, it is reasonable to speculate that CV_2116 is a component of a bacteriophage-like tail assembly specific to C. violaceum.
Previous analysis of the C. violaceum genome has shown that it contains four prophages denoted CvP1 through CvP4 . The latter, CvP4, is composed of 51 ORFs (CV_2114 to CV_2150) and is thought to be a chimeric prophage containing ORFs with similarity to different phages. Specifically, it has been suggested that the tail and tail fiber genes, which include CV_2114-CV_2118, are similar to P2-like phage genes , and in particular, that these genes exhibit similarity to those found in a P2-like prophage of Salmonella enterica subsp. enterica serovar Typhi, which are the so-called CT18 phages [14,17,18]. Collectively, it has been speculated that CvP4 was responsible for introducing a number of genes from S. enterica into the C. violaceum genome. A PSI-BLAST search of CV_2117 indicated ~100 hits to proteins with >30% identity, including NP_456044 from S. enterica subsp. enterica serovar Typhi str. CT18, which had 41% identity with 96% coverage and E = 9e−40. Examination of the genes surrounding NP_456044 indicated a putative operon containing 10 consecutive genes spanning NP_456044 to NP_456053, all encoding proteins involved in bacteriophage tail-like assembly. Examination of the strongest hit PSI-BLAST hit (E = 1e−105, 72% sequence identity) against CV_2117, namely YP_003745638 from Ralstonia solanacearum CFBP2957, which is an aerobic non-sporing, Gram-negative plant pathogenic bacterium, also revealed a putative operon with a collection of genes spanning YP_003745640 to YP_003745645 that are all annotated as bacteriophage tail-like assembly proteins. These comparisons further support the speculation that CV_2116 is likely a novel gene of prophage origin that plays a structural or functional role in bacteriophage tail-like assembly in C. violaceum.
3. Experimental Section
3.1. Preparation of Protein Samples
The full-length gene for CV_2116 from C. violaceum was cloned into the expression plasmid p15TvLic including an additional N-terminal hexa-histidine tag and TEV cleavage site (22 residues, MGSSHHHHHHSSGRENLYFQGH), and the plasmid was transformed into E. coli BL21 (DE3)-RIPL (Stratagene). E. coli cells were grown in 0.5 L of 2 × M9 minimal medium containing 15NH4Cl and 13C-glucose as the sole nitrogen and carbon sources, and supplemented with ZnSO4, thiamine, and biotin. Cells were grown at 37 °C to an OD600 of 1.0, followed by addition of 1 mM isopropyl β-d-thiogalactoside. The temperature was reduced to 15 °C, and cells grew overnight before harvesting. Cell pellets were lysed by sonication in a buffer containing 500 mM NaCl, 20 mM Tris and 5 mM imidazole (pH 8.0). The proteins were isolated from the lysate by nickel affinity chromatography (Qiagen) after washing the beads three times with five column volumes of 500 mM NaCl, 20 mM Tris (pH 8.0), 30 mM imidazole. Protein was eluted with five column volumes of 500 mM imidazole added to the washing buffer. For NMR structure determination, the hexahistidine tag of CV_2116 was cleaved with TEV protease and the mixture passed through a nickel affinity column. The purified 13C/15N labeled protein (referred to as the NC sample) was concentrated, and buffer was exchanged by ultrafiltration and dilution/reconcentration into the NMR buffer containing 10 mM Tris (pH 7.0), 300 mM NaCl, 10 mM DTT, 1 mM benzamidine, 0.01% NaN3, 1× inhibitor cocktail (Roche Applied Science), 95% H2O/5% D2O. Following the same procedure, a uniformly 100% 15N-labeled and 7% biosynthetically directed 13C (NC7) CV_2116 sample was also prepared for stereospecific assignments of isopropyl methyl groups of Val and Leu residues.
3.2. NMR Spectroscopy and Data Collection
NMR spectra were acquired on 250 μL NC and NC7 samples in 5 mm Shigemi NMR tubes at 293 K on Varian Inova 600 and 750 MHz NMR spectrometers. Backbone and sidechain resonance assignments were obtained from the analysis of 15N-HSQC (Figure 3A), 13C-HSQC, HNCO, HNCA, HN(CO)CA, HNCACB, CBCA(CO)NH, C(CO)NH, HC(CO)NH, HBHA(CO)NH, HCCH-COSY, HCCH-TOCSY and (H)CCH-TOCSY spectra. Stereospecific assignments of isopropyl methyl groups of Val and Leu residues were determined from characteristic 1H-13C coupling patterns in a high resolution 1H-13C HSQC of the NC7 sample . NOE distance restraints were derived from a 15N-edited NOESY-HSQC (mixing time τm = 70 ms) and two 13C-edited NOESY-HSQC (τm = 70 ms) optimized for either aliphatic or aromatic carbon detection. Additional NOEs were assigned from a 4D 13C-13C-HMQC-NOESY-HMQC (τm = 70 ms) recorded after lyophilization and dissolution into pure D2O. Spectra were processed by NMRPipe and analyzed with Sparky 3.110 . Nearly complete resonance assignments (99.3%) were determined. NMR assignments, raw NOESY FIDs, and NOESY peak lists have been deposited in the BioMagResBank (BMRB accession number 16,521).
3.3. NMR Structure Calculation
The solution NMR structure of CV_2116 was initially calculated using both AutoStructure  and CYANA 2.1 . Inputs for calculation were resonance assignments, NOESY peak lists from the four NOESY spectra that included peak intensities, and TALOS-derived dihedral restraints for ϕ and ψ dihedral angles. NOE assignments were initially made automatically in AutoStructure and CYANA and the assignments refined manually. NMR RPF quality assessment scores were used to assess “goodness of fit” between calculated structures and NOESY peaks lists and were used to guide manual and iterative refinement of NOESY peak picking. The NOE-derived distance constraints, dihedral constraints, and hydrogen-bond constraints derived from AutoStructure were converted to Xplor/CNS format using PDBStat and the upper bounds for the NOEs were increased by 10%. These constraints were used to calculate 150 structures using Xplor-NIH (version 2.25) with a standard simulated annealing protocol followed by refinement of the 20 lowest energy structure using the Xplor-NIH protocol [23,24]. For the final NMR ensemble, the 20 lowest energy structures were deposited in the Protein Data Bank (PDB ID, 2KON). NMR structure quality analysis and structural statistics were performed using the PSVS  and RPF  software packages. Structural statistics for CV_2116 are presented in Table 1.
In conclusion, protein sequence analysis indicated that the hypothetical protein CV_2116 from C. violaceum is not highly conserved among known protein families. The three-dimensional structure of CV_2116 was determined using solution state NMR spectroscopy and analysis of the structure indicated that CV_2116 adopted a novel fold containing two anti-parallel β-sheets and one α-helix. Despite determination of a high quality three-dimensional structure, no biochemical or biological function of CV_2116 could be predicted based on sequence similarity or structural alignment searches. However, analysis of the genes neighboring CV_2116 indicated that it resides in a putative operon that encodes for five proteins that are likely involved in bacteriophage tail-like assembly, and that the gene and its associated operon is of prophage origin, thereby giving C. violaceum distinctly bacteriophage-like characteristics that include the occurrence of a bacteriophage tail-like assembly. In closing, the analysis of CV_2116 from C. violaceum, which is both a hypothetical protein and a singleton, is typical of the challenge that is often encountered when researchers attempt to construct a hypothesis regarding the biochemical or cellular function of a hypothetical and/or singleton protein exclusively from sequence and structural information when little other knowledge about the protein is available.
This work was supported by the National Institute of General Medical Sciences; Protein Structure Initiative-Biology Program; Grant Number: U54-GM074958 and U54-GM094597. The majority of the data collection was conducted at the Ohio Center of Excellence in Biomedicine in Structural Biology and Metabonomics at Miami University. A portion of the NMR experiments was performed in the Environmental Molecular Sciences Laboratory, a national scientific user facility sponsored by the U.S. Department of Energy’s Office of Biological and Environmental Research and located at Pacific Northwest National Laboratory. The authors acknowledge Gaetano T. Montelione and Shuisong Ni for helpful discussions.
- Kodach, L.L.; Bos, C.L.; Durán, N.; Peppelenbosch, M.P.; Ferreira, C.V.; Hardwick, J.C.H. Violacein synergistically increases 5-fluorouracil cytotoxicity, induces apoptosis and inhibits Akt-mediated signal transduction in human colorectal cancer cells. Carcinogenesis 2006, 27, 508–516. [Google Scholar]
- De Almeida, D.F.; Hungria, M.; Guimarães, C.T.; Antônio, R.V.; Almeida, F.C.; de Almeida, L.G.P.; de Almeida, R.; Andrade, E.M.; Araripe, J.; de Araújo, M.F.F.; et al. The complete genome sequence of Chromobacterium violaceum reveals remarkable and exploitable bacterial adaptability. Proc. Nat. Acad. Sci. USA 2003, 100, 11660–11665. [Google Scholar]
- De Almeida, R.; Trevilato, P.B.; Bartoleti, L.A.; Proença-Módena, J.L.; Hanna, E.S.; Gregoracci, G.B.; Brocchi, M. Bacteriophages and insertion sequences of Chromobacterium violaceum ATCC 12472. Genet. Mol. Res 2004, 3, 76–84. [Google Scholar]
- Protein: Q7NW74_CHRVO (Q7NW74), Availbale online: http://pfam.sanger.ac.uk/protein/Q7NW74 accessed on 15 February 2012.
- Bauer, M.; Kube, M.; Teeling, H.; Richter, M.; Lombardot, T.; Allers, E.; Würdemann, C.A.; Quast, C.; Kuhl, H.; Knaust, F.; et al. Whole genome analysis of the marine Bacteroidetes “Gramella forsetii” reveals adaptations to degradation of polymeric organic matter. Environ. Microbiol 2006, 8, 2201–2213. [Google Scholar]
- Rhee, M.S.; Moritz, B.E.; Xie, G.; del Rio, T.G.; Dalin, E.; Tice, H.; Bruce, D.; Goodwin, L.; Chertkov, O.; Brettin, T.; et al. Complete Genome Sequence of a thermotolerant sporogenic lactic acid bacterium, Bacillus coagulans strain 36D1. Stand. Genomic Sci 2011, 5. [Google Scholar] [CrossRef]
- Larkin, M.A.; Blackshields, G.; Brown, N.P.; Chenna, R.; McGettigan, P.A.; McWilliam, H.; Valentin, F.; Wallace, I.M.; Wilm, A.; Lopez, R.; et al. Clustal W and Clustal X version 2.0. Bioinformatics 2007, 23, 2947–2948. [Google Scholar]
- Yang, Y.; Ramelot, T.A.; Cort, J.R.; Wang, H.; Ciccosanti, C.; Jiang, M.; Janjua, H.; Acton, T.B.; Xiao, R.; Everett, J.K.; et al. Solution NMR structure of Dsy0195 homodimer from Desulfitobacterium hafniense: First structure representative of the YabP domain family of proteins involved in spore coat assembly. J. Struct. Funct. Genomics 2011, 12, 175–179. [Google Scholar]
- Laskowski, R.A.; MacArthur, M.W.; Moss, D.S.; Thornton, J.M. PROCHECK—A program to check the stereochemical quality of protein structures. J. Appl. Cryst 1993, 26, 283–291. [Google Scholar]
- Laskowski, R.A.; Rullmannn, J.A.; MacArthur, M.W.; Kaptein, R.; Thornton, J.M. AQUA and PROCHECK-NMR: Programs for checking the quality of protein structures solved by NMR. J. Biomol. NMR 1996, 8, 477–486. [Google Scholar]
- CATH: Protein Structure Classification Database, Available online: http://www.cathdb.info/ accessed on 15 February 2012.
- SCOP: Structural Classification of Proteins, Available online: http://scop.mrc-lmb.cam.ac.uk/scop/ accessed on 15 February 2012.
- Holm, L.; Rosenstrom, P. Dali server: Conservation mapping in 3D. Nucleic. Acids. Res 2010, 38, W545–W549. [Google Scholar]
- Kumazaki, T.; Ishii, S. Comparative study on fibers isolated from four R-type pyocins, phage-tail-like bacteriocins of Pseudomonas aeruginosa. J. Biochem 1982, 92, 1559–1566. [Google Scholar]
- Haggard-Ljungquist, E.; Jacobsen, E.; Rishovd, S.; Six, E.W.; Sunshine, M.G.; Lindqvist, B.H.; Kim, K.-J.; Barreiro, V.; Koonin, E.V.; Calendar, R.; et al. Bacteriophage P2: Genes involved in baseplate assembly. Virology 1995, 213, 109–121. [Google Scholar]
- Rucinsky, T.E.; Gregory, J.P.; Cota-Robles, E.H. Organization of bacteriophage tail-like particles in cells of Chromobacterium violaceum. J. Bacteriol 1972, 110, 754–757. [Google Scholar]
- Canchaya, C.; Proux, C.; Fournous, G.; Bruttin, A.; Brüssow, H. Prophage genomics. Microbiol. Mol. Biol. Rev 2003, 67, 238–276. [Google Scholar]
- Parkhill, J.; Dougan, G.; James, K.D.; Thomson, N.R.; Pickard, D.; Wain, J.; Churcher, C.; Mungall, K.L.; Bentley, S.D.; Holden, M.T.G.; et al. Complete genome sequence of a multiple drug resistant Salmonella enterica serovar Typhi CT18. Nature 2001, 413, 848–852. [Google Scholar]
- Neri, D.; Szyperski, T.; Otting, G.; Senn, H.; Wuethrich, K. Stereospecific nuclear magnetic resonance assignments of the methyl groups of valine and leucine in the DNA-binding domain of the 434 repressor by biosynthetically directed fractional 13C labeling. Biochemistry 1989, 28, 7510–7516. [Google Scholar]
- Goddard, T.; Kneller, D. SPARKY, verison 3; University of California: San Fransisco, CA, USA, 2008. Available online: http://www.cgl.ucsf.edu/home/sparky/ accessed on 15 February 2012.
- Huang, Y.J.; Tejero, R.; Powers, R.; Montelione, G.T. A topology-constrained distance network algorithm for protein structure determination from NOESY data. Proteins 2006, 62, 587–603. [Google Scholar]
- Guntert, P. Automated NMR structure calculation with CYANA. Methods Mol. Biol 2004, 278, 353–378. [Google Scholar]
- Schwieters, C.D.; Kuszewski, J.J.; Clore, G.M. Using Xplor-NIH for NMR molecular structure determination. Progr. Nucl. Magn. Reson. Spectrosc 2006, 48, 47–62. [Google Scholar]
- Cai, M.; Huang, Y.; Suh, J.-Y.; Louis, J.M.; Ghirlando, R.; Craigie, R.; Clore, J.M. Solution NMR structure of the barrier-to-autointegration factor-Emerin complex. J. Biol. Chem 2007, 282, 14525–14535. [Google Scholar]
- Bhattacharya, A.; Tejero, R.; Montelione, G.T. Evaluating protein structures determined by structural genomics consortia. Proteins 2007, 66, 778–795. [Google Scholar]
- Huang, Y.J.; Powers, R.; Montelione, G.T. Protein NMR recall, precision, and F-measure scores (RPF scores): Structure quality assessment measures based on information retrieval statistics. J. Am. Chem. Soc 2005, 127, 1665–1674. [Google Scholar]
|Table 1. Summary of NMR and structure statistics for CV_2116 protein from Chromobacterium violaceum a.|
|Conformationally-Restricting Constraints b|
|Intra-residue (I = j)||306|
|Sequential (|i–j| = 1)||357|
|Medium-range (1 < |i–j| < 5)||193|
|Long-range (|i–j| ≥ 5)||419|
|Distance constraints per residue||15.5|
|Hydrogen bond constraints|
|Long-range (|i–j| ≥ 5)/total||20/56|
|Dihedral angle constraints||92|
|Total number of restricting constraints||1423|
|Number of constraints per residue||5.4|
|Residue Constraint Violations b|
|Average Number of Distance Violations Per Structure|
|Average RMS distance violation/constraint (Å)||0.01|
|Maximum distance violation (Å)||0.29|
|Average Number of Dihedral Angle Violations Per Structure|
|Average RMS dihedral angle violation/constraint (degree)||0.15|
|Maximum dihedral angle violation (degree)||2.60|
|RMSD from Average Coordinates (Å) c|
|Ramachandran Plot Statistics c|
|Most favored regions (%)||96.3|
|Additional allowed regions (%)||3.6|
|Generously allowed (%)||0.1|
|Disallowed regions (%)||0|
|Global Quality Scores (raw/Z-score) b|
|RPF Scores d|
aStructure statistics were computed for the ensemble of 20 deposited structures;bCalculated using PSVS 1.4 program. Average distance violations were calculated using the summation of r−6. Residues 1–82 were analyzed;cOrdered residues ranges (with sum of phi and psi > 1.8): 3–5, 8–13, 16–32, 40–45, 51–71, 79–81;dPF scores reflected the goodness-of-fit of the final ensemble of structures including disordered residues to the NMR data.
© 2012 by the authors; licensee Molecular Diversity Preservation International, Basel, Switzerland. This article is an open-access article distributed under the terms and conditions of the Creative Commons Attribution license (http://creativecommons.org/licenses/by/3.0/).