Skip Navigation

This Article
Right arrow Abstract Freely available
Right arrow Print PDF (234K) Freely available
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Add to My Personal Archive
Right arrow Download to citation manager
Right arrow Search for citing articles in:
ISI Web of Science (18)
Right arrowRequest Permissions
Right arrow Commercial Re-use Guidelines
for Open Access NAR Content
Google Scholar
Right arrow Articles by Guerri, A.
Right arrow Articles by Neidle, S.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Guerri, A.
Right arrow Articles by Neidle, S.
Social Bookmarking
Add to CiteULike   Add to Connotea   Add to Del.icio.us  
What's this?

Nucleic Acids Research Pages 2873-2878  

Visualisation of extensive water ribbons and networks in a DNA minor-groove drug complex
Introduction
Materials And Methods
Results
Discussion
Acknowledgements
References

Visualisation of extensive water ribbons and networks in a DNA minor-groove drug complex

Visualisation of extensive water ribbons and networks in a DNA minor-groove drug complex

A. Guerri, I. J. Simpson, S. Neidle*

The CRC Biomolecular Structure Unit, The Institute of Cancer Research, Sutton, Surrey SM2 5NG, UK

Received March 18, 1998; Revised and Accepted May 5, 1998

NDB no. GDL 056

ABSTRACT

The crystal structure is reported of a complex between an ethyl derivative of the minor-groove drug furamidine and the dodecanucleotide duplex d(CGCGAATTCGCG)2, which has been refined to 1.85 Å resolution and an R factor of 16.6% for data collected at -173°C. An exceptionally large number (220) of water molecules have been located. The majority of these occur in the first coordination shell of solvation. There are extensive networks of connected waters, both in the major and minor grooves. In particular, there are 21 water molecules associated with the minor-groove drug, via hydrogen bonds from the four charged nitrogen atoms. One cluster of four waters is situated in the groove itself; the majority are on the outer edge of the groove, and serve to bridge between the outward-directed drug nitrogen atoms and backbone phosphate oxygen atoms. These bridges are both intra- and inter-strand, with the net effect that the outer edge of the drug molecule is covered by ribbons of water molecules.

INTRODUCTION

Nucleic acids require an environment of water molecules in order to maintain their stability (1). X-ray crystallographic studies have shown, in particular, that the functional groups on the edges of bases are able to act as effective hydrogen bond donors/acceptors with water molecules, and the resulting distribution of available sites has been extensively documented (2). These sites can also be used by amino acid side-chains (and occasionally main-chains) of regulatory proteins for direct DNA sequence recognition and readout (3). Hydration of phosphate groups is believed to be of especial importance in overall stabilisation and in the transitions of nucleic acids from one polymorph to another (4). Water molecules can also form networks around nucleic acids, as seen some years ago in the crystal structure of an intercalation complex (5). The well-characterised spine of hydration (effectively an extended ribbon of water molecules) in the narrow A/T minor groove of the dodecanucleotide duplex d(CGCGAATTCGCG)2, initially observed by crystallography (6), has been independently observed by NMR techniques (7,8), and in molecular dynamics simulation studies (9).

Narrow A/T minor groove sequences are the preferred sites of binding for many drugs, typified by netropsin, distamycin, Hoechst 33258, and the bis-amidine compounds berenil and pentamidine (10,11). The basis for their biological action remains to be fully elucidated; there is evidence, for example, that they can compete with regulatory proteins such as TBP when it binds and bends the minor groove at the TATA-box site for the origin of transcription (12). In doing so, it is reasonable to assume that these drugs compete with, and displace bound water molecules, especially the ordered spine of waters in runs of A[bull]T base pairs.

Crystallographic and molecular modelling studies have been reported on oligonucleotide complexes of several derivatives of pentamidine and berenil (13-16), notably those with a central furan ring linking the two phenylamidinium groups (17). Several of these compounds (`furamidines') show activity against the Pneumocystis carinii pathogen (18,19), and have cytotoxic activity against tumour cells in vitro, probably due to their ability to effectively compete with protein-DNA contacts (20). The structural studies have shown that the drugs bind in the minor groove with the amidinium groups usually forming hydrogen bonds with O2 substituent atoms on thymine base edges.

We report here the crystal structure of a further derivative of furamidine complexed to a dodecanucleotide duplex, with ethyl groups attached to one nitrogen atom of each terminal amidinium group (Fig. 1). The present structure is at significantly higher resolution than earlier determinations of minor-groove drug complexes. This improvement in resolution, together with the fact that data collection took place at -173°C, has enabled a greater number of water molecules to be located than hitherto. It was to be expected that these would populate available first-shell hydration sites deep in the major and minor grooves, and such arrangements have indeed been observed, albeit rather more complete than we have previously found. Unexpectedly, we also observe extended networks and chains involving higher-shell water molecules, which connect to both drug and DNA molecules in ways which have not previously been described.ab


Figure 1. (a) The structure of bis-(ethylamidiniumphenyl) furan. (b) Sequence and numbering scheme for the dodecanucleotide duplex.

MATERIALS AND METHODS

Crystals were obtained at 15°C with the hanging-drop method. Droplets, buffered at pH 6.3 with sodium cacodylate, contained 2 µl 3 mM d(CGCGAATTCGCG) (from Oswell), 1 µl 10 mM spermine, 2 µl 3 mM drug, 2 µl 75 mM MgCl2 and 2 µl 25% (v/w) methyl 1,2-pentane-diol (MPD) at pH 7.0; the reservoir contained 30% (v/w) MPD. Crystals of an acceptable size grew in 1-2 weeks.

Table 1. Crystallographic data and statistics
Cell dimensions: a = 24.240, b = 39.940, c = 65.880 Å
Space group: P212121
Total number of reflections collected: 36613
Number of unique reflections: 5634
Completeness: 96.2%
Rsym: 3.4%
Resolution range of refinement: 7.0-1.85 Å
Reflections with Fo > 4[sigma](Fo): 5386
Completeness: 92.0%
Final R: 16.6%
Final R for reflections with Fo > 4[sigma](Fo): 16.9%
Final Rfree: 22.7%
Number of DNA atoms: 486
Number of drug atoms: 27
Number of water molecules: 220
R.m.s.d. bond lengths: 0.009 Å
R.m.s.d. bond angles: 0.023°

Diffraction data were collected with a Rigaku RAXIS IV image plate detector using mirror-focussed CuK[alpha] radiation from a Rigaku RU200 rotating-anode generator operating at 100 mA and 50 kV. The crystals were flash-frozen and maintained at -173°C with the aid of a Cryostream instrument (Oxford Cryosystems). Data were collected to a resolution of 1.80 Å, and were processed with the DENZO and SCALEPACK packages (21); there were <50% observed reflections in the range 1.80-1.85 Å, so this shell of data was not used in subsequent calculations. Table 1 details relevant crystallographic data.

The structure is isomorphous with other dodecanucleotide-drug crystal structures solved in this laboratory and elsewhere. Starting coordinates for the DNA were taken from those of the complex with the closely-related drug having an isopropyl group at the termini instead of the present ethyl groups (Nucleic Acid Database entry No. GDL044). Initial refinement was conducted with the X-PLOR v3.1 program (22), using the DNA parameter set recently described (23). Initial low-resolution refinement was confined to nucleotide groups, which was subsequently extended to restrained individual atomic coordinate refinement. Convergence was achieved at an R factor of ~30%, when an Fo-Fc map was calculated. This clearly showed the drug molecule in the minor groove of the dodecanucleotide. Refinement was continued, initially with X-PLOR and a dictionary for the drug geometry based on that previously determined for the isopropyl and cyclopropyl analogues (16). Charges were obtained from a MOPAC calculation (24). The SHELX-97 package (25) was then used for all subsequent refinement, continuing with the same set of standard DNA and drug geometries as used in the X-PLOR refinement. The refinement protocol involved initial low-resolution refinement, followed by gradual extension to successively higher resolution, and frequent checks on structural integrity with 2Fo-Fc maps. The orientation for the terminal carbon atom of each ethyl group on the drug side-chain was determined from these maps, as was the relative orientation of the phenyl and furan rings. A hydrated magnesium ion was found in these maps, in the same position as reported in other minor groove complexes (26). Water molecules were located by two independent procedures (by A.G. and S.N., respectively), in order to minimise `investigator bias'. One employed systematic hand selection of difference electron density peaks and their verification on the basis of acceptable hydrogen-bonding distance and angle geometry. Water assignments were then verified by successive rounds of refinement, with those having B values >60 Å2 being rejected. The other used the automated water location procedure in SHELX-97. This was found to give the same set of water molecules, albeit rather more speedily. The value of Rfree on adding successive water molecules was monitored with both procedures, and was found to be a sensitive indicator of the correctness of water assignments. The full set of water molecules, together with the dodecanucleotide and drug molecules and hydrated magnesium ion, was subjected to a final round of cycles of refinement, until convergence was judged to have been achieved. Final atomic coordinates and structure factors have been deposited with the Nucleic Acid Database as entry No. GDL056.

RESULTS

The structure shows the drug lying in the A/T region of the dodecanucleotide minor groove in a position closely similar to that observed for the same sequence with furamidine itself (16) (which has unsubstituted terminal amidinium groups), and with cyclopropyl and isopropyl substituents (16). The water molecules found are distributed in both major and minor grooves, and also cluster around phosphate groups and the hydrated magnesium ion. Most hydrogen bond donor-acceptor distances involving the waters are in the range 2.7-3.2 Å. A few distances are significantly shorter than this lower limit, possibly due to the involvement of disordered or mobile water molecules. A significant number of the water molecules located are solely hydrogen-bonded to other waters, and can be considered as second or higher-shell solvent. The O2 oxygen atoms of the phosphate groups are directed towards the major groove, and they form the starting-point for numerous water networks in the major groove, which tend to contact base edges (3), especially of the purine bases. These networks are not discussed further here.

At both ends of the drug molecule, an amidinium group is oriented such that one of its nitrogen atoms is directed towards the floor of the groove, where it makes hydrogen-bonded contact with a thymine O2 atom (O2 atoms of thymines 8 and 20). At the 5[prime]-end of the drug, this nitrogen, N1, also contacts a water molecule, O119, which in turn hydrogen-bonds to O2 of the next base, cytosine9 on the same strand, as well as to N3 of adenine17 (Fig. 2). At the other end of the drug molecule there is a more extensive array of water molecules in the form of an irregular ribbon of four waters, which start from the amidinium N1[prime] atom (Fig. 3). O162 is hydrogen-bonded to N1[prime], and also contacts N3 of adenine5 (which is complementary to thymine20, whose O2 atom hydrogen-bonds to N1[prime]). This water molecule O162 in turn hydrogen bonds to O197, which makes two DNA contacts, to the O4[prime] ring sugar atom of guanine22 and to O2 of cytosine21, as well as to another water molecule, O198. This second-shell water bridges to O157, which then hydrogen bonds to N3 of guanine22, and rather more weakly, to O4[prime] of the next nucleotide, cytosine 23.


Figure 2. The structure of the complex, at the 5[prime]-end of the drug binding site. Water molecules are shown as small yellow spheres, and the drug molecule is coloured brown. Hydrogen bonds are indicated by white lines.


Figure 3. The structure of the complex, at the 3[prime]-end of the drug binding site. The ribbon of four water molecules starting from N1[prime], is in the centre of the figure.

The outer nitrogen atoms (N2, N2[prime]) of each amidinium group, which have an ethyl group attached, are oriented towards the mouth of the groove. In both instances, the nitrogen atom is hydrogen-bonded to a water molecule, which then leads off into clusters and ribbons of further waters (Figs 2 and 4).


Figure 4. A view of the complete bound drug molecule, showing all three ribbons of inter-strand water molecules, passing over the centre and both ends of the bound drug.

At the 5[prime]-end of the drug, N2 contacts water O146 (Fig. 2). This water is hydrogen-bonded to two second-shell waters, O142 and O186, which lead off in three separate directions. The first, from O142, leads directly to the phosphate oxygen atom of thymine20, which is oriented in towards the minor groove. O142 is also hydrogen-bonded to water O292, situated in the centre of the groove mouth, which then contacts O247 on the side of the opposite strand. There is a final hydrogen bonded interaction from this water to the phosphate oxygen atom of cytosine9 on the opposite strand. Water O186, close to the first strand, bridges two further waters, O195 and O282, which finally contact the phosphate oxygen atoms of thymine19 and thymine20 via O132 and O225 respectively. Overall, this network emanating from N2, involves nine water molecules and three phosphate groups.

The network of hydrogen-bonded waters arising from the N2[prime] atom at the other end of the drug molecule is similar (Fig. 4), with three phosphate groups also being the end-points of water ribbons. In this instance, two ribbon of waters bridge the two DNA strands. The pattern of waters differs in detail: the phosphate group on thymine8 is the terminal of an irregular zig-zag of three waters, starting at N2[prime], viz O275, O185 and O211. O185 also makes a weak hydrogen bond to the O3[prime] atom of this phosphate group. A chain of four further waters (O237, O271, O310 and O168) connects this phosphate with that of cytosine21 on the opposite strand. Water O275 is the starting-point for a further ribbon of three waters (O260, O314 and O129), which lead back to the phosphate group of guanine 22. Thus, these three waters together with O260, O275, O165 and O211, form the third continuous inter-strand network, which as with the other two, runs over the bound drug molecule (Fig. 4).

Values of the temperature factors for the water molecules are generally in accord with their pattern of hydrogen-bonding and role in first versus second-shell hydration. Thus, the water molecules O119 and O162 directly hydrogen-bonded to the inward-facing nitrogen atoms N1 and N1[prime], have B factors of 32 and 34 Å2 respectively. Water molecules forming the three inter-strand bridges, mostly have solely water-water contacts, and have a higher average B factor, of 49 Å2.

DISCUSSION

Small numbers of groove-bound water molecules have been previously observed in drug-oligonucleotide minor-groove complexes, although none in as extensive an arrangement as found here at the 3[prime] end of the drug. This short ribbon of water molecules resembles the spine of hydration found in the native d(CGCGAATTCGCG) structure (6), although the set of direct and continuous water-water contacts seen here is quite distinct from those in the spine. The single water molecule observed at the 5[prime] end, is deeply embedded in the groove. It fulfils an important stabilising role, resulting in additional indirect interactions between drug and several DNA acceptor sites in addition to the direct N1 ... O2 hydrogen bond. In this respect it is analogous to the bridging water molecule found in the berenil-d(CGCGAATTCGCG)2 complex (28).

The networks of water molecules linking cationic charged nitrogen atoms in the drug (N2, N2[prime]) with the anionic phosphates, have not been previously observed. The three strings of inter-strand linking waters, which effectively cover the mouth of the groove (Fig. 5), form in effect a protective cover over the hydrophobic outer edge of the drug. We have previously noted, in the crystal structures of two pentamidine complexes (14,27), a small number of unexplained water molecules approximately at the positions of waters O260, O271 and O292 in the present structure. These may well have been the least mobile, and therefore, the only observable part of equivalent networks in these lower resolution, room-temperature structures.


Figure 5. A view of the complex, now looking down the minor groove, along the long axis of the drug molecule. A semi-transparent representation of the solvent-accessible surface of the DNA is shown. The three inter-strand ribbons of water molecules can be seen to be covering the mouth of the minor groove.

Electrostatic, hydrophobic, van der Waals and hydrogen bonding interactions have been generally considered to be the major factors contributing to the stability of molecules binding in the minor groove (10). The contribution to overall binding of the terminal charged nitrogen atoms of the drug by means of hydrogen-bonding to atoms on base edges is well-documented (10,13-17). The present structure shows that these charged atoms take part in extensive hydrogen bonding to water molecules, and so may be relaying electrostatic effects from these cationic nitrogen atoms to anionic phosphates.

Although no thermodynamic data is as yet available on the present drug, several studies have been made of netropsin binding to A/T-containing poly- and oligonucleotides. This is accompanied by a positive entropy change (29). Analysis of volume changes on binding (30) have indicated that major changes in hydration of both drug and DNA occur on binding, contributing to the overall entropic change. Analogous effects have been noted for the binding of two distamycin molecules (with a large van der Waals component) to an A/T sequence (31), and it would be therefore reasonable, by analogy, to suppose that the DNA binding of the present substituted furamidine is also accompanied by a significant entropic change. This is in accord with the present finding of a large number of water molecules associated with the drug-oligonucleotide complex.

It has been proposed (32) that the enthalpic gain of binding a ligand to a macromolecule may outweigh the entropic loss due to the immobilisation of water molecules. This loss will be much less than might initially be suggested, based purely on the large number of bound waters reported in this study: an estimate (33) of the entropic cost of transferring a water molecule from liquid to bound state gives a range of 0-7 cal/mol/C, but only for firmly-bound waters with low B values such as O119 and O162, which bridge between drug and DNA. The majority of water molecules located in this study are much more mobile, and will therefore make only a small contribution to the overall change in entropy on binding. A recent molecular dynamics simulation (34) of a DNA-homeodomain complex with explicit solvent suggested that highly mobile water molecules play an important role in the protein-DNA recognition process. We suggest that the ensemble of bound water molecules observed in the present structure play an analogous role in drug-DNA recognition, and will need to be taken into account in order to more fully understand drug-DNA binding processes.

ACKNOWLEDGEMENTS

This work was supported by Cancer Research Campaign Programme Grant SP1384 to S.N., and by a Research Studentship from the Institute of Cancer Research to I.J.S. We are grateful to Professors D.W.Boykin, W.D.Wilson and Dr A.Kumar (Georgia State University) for the provision of drug samples, and much useful discussion. We are also grateful to Professor H.M.Berman (Rutgers) for discussions on hydration issues. A.G. thanks the University of Florence for support of her stay at The Institute of Cancer Research.

REFERENCES

1. Westhof, E. (ed.) (1993) Water and Biological Macromolecules. Macmillan Press, London.

2. Schneider, B., Cohen, D. and Berman, (1992) Biopolymers, 32, 725-750. MEDLINE Abstract

3. Seeman, N.D., Rosenberg, J.M. and Rich, A. (1976) Proc. Natl. Acad. Sci. USA, 73, 804-808. MEDLINE Abstract

4. Saenger, W., Hunter, W.N. and Kennard, O. (1986) Nature, 324, 385-388. MEDLINE Abstract

5. Neidle, S., Berman, H.M. and Shieh, H.-S. (1980) Nature, 288, 129-133. MEDLINE Abstract

6. Drew, H.R. and Dickerson, R.E. (1981) J. Mol. Biol., 151, 535-556. MEDLINE Abstract

7. Kubinec, M.G. and Wemmer, D.E. (1992) J. Am. Chem. Soc., 114, 8739-8740.

8. Liepinsh, E., Otting, G. and Wüthrich, K. (1992) Nucleic Acids Res., 20, 6549-6553. MEDLINE Abstract

9. Duan, Y., Eilkosz, P., Crowley, M. and Rosenberg, J.M. (1997)J. Mol. Biol., 272, 553-572. MEDLINE Abstract

10. Neidle, S. (1997) Biopolymers, 44, 105-121.

11. Zimmer, C. and Wähnert, U. (1986) Prog. Biophys. Mol. Biol., 47, 31-112. MEDLINE Abstract

12. Chiang, S.-Y., Welch, J., Rauscher, F.J. and Beerman, T.A. (1994) Biochemistry, 33, 7033-7040. MEDLINE Abstract

13. Brown, D.G., Sanderson, M.R., Garman, E. and Neidle, S. (1992)J. Mol. Biol., 226, 481-491.

14. Nunn, C.M., Jenkins, T.C. and Neidle, S. (1994) Eur. J. Biochem., 226, 953-961. MEDLINE Abstract

15. Nunn, C.M. and Neidle, S. (1995) J. Med. Chem., 38, 2317-2325. MEDLINE Abstract

16. Laughton, C.A., Tanious, F., Nunn, C.M., Boykin, D.W., Wilson, W.D. and Neidle, S. (1996) Biochemistry, 35, 5655-5661. MEDLINE Abstract

17. Trent, J.O., Clark, G.R., Kumar, A., Wilson, W.D., Boykin, D.W., Hall, J.E., Tidwell, R.R., Blagburn, B.L. and Neidle, S. (1996) J. Med. Chem., 39, 4554-4562. MEDLINE Abstract

18. Boykin, D.W., Kumar, A., Spychala, J., Zhou, M., Lombardy, R.J., Wilson, W.D., Dykstra, C.C., Jones, S.K., Hall, J.E., Tidwell, R.R., Laughton, C., Nunn, C.M. and Neidle, S. (1995) J. Med. Chem., 38, 912-916. MEDLINE Abstract

19. Boykin, D.W., Kumar, A., Xiao, G., Wilson, W.D., McCurdy, D.R., Hall, J.E. and Tidwell, R.R. (1998) J. Med. Chem., 41, 124-129. MEDLINE Abstract

20. Neidle, S., Kelland, L.R., Trent, J.O., Simpson, I.J., Boykin, D.W., Kumar, A. and Wilson, W.D. (1997) Bioorg. Med. Chem. Lett., 7, 1403-1408.

21. Otwinowski, Z. and Minor, W. (1993) In Sawyer, L., Isaacs, N.W. and Bailey, S. (eds), Data Collection and Processing. SERC Daresbury Laboratory, Warrington, UK.

22. Brünger, A.T., Kuriyan, J. and Karplus, M. (1987) Science, 235, 458-460.

23. Parkinson, G., Vojtechovsky, J., Clowney, L., Brünger, A.T. and Berman, H.M. (1996) Acta Crystallogr., D52, 57-64.

24. Stewart, J.P.P. Mopac 6.0 (QCPE), available from the Quantum Chemistry Program Exchange, Indiana University, Bloomington, IN 47405.

25. SHELX-97, a crystallographic refinement programme, Sheldrick, G.M., University of Göttingen, Germany.

26. Wood, A.A., Nunn, C.M., Czarny, A., Boykin, D.W and Neidle, S. (1995) Nucleic Acids Res., 23, 3678-3684. MEDLINE Abstract

27. Nunn, C.M., Jenkins, T.C. and Neidle, S. (1993) Biochemistry, 32, 13838-13843. MEDLINE Abstract

28. Brown, D.G., Sanderson, M.R., Skelly, J.V., Jenkins, T.C., Brown, T., Garman, E., Stuart, D.I. and Neidle, S. (1990) EMBO J., 9, 1329-1334. MEDLINE Abstract

29. Marky, L.A. and Breslauer, K.J. (1987) Proc. Natl. Acad. Sci. USA, 84, 4359-4363. MEDLINE Abstract

30. Chalikian, T.V., Plum, G.E., Sarvazyan, A.P. and Breslauer, K.J. (1994) Biochemistry, 33, 8629-8640. MEDLINE Abstract

31. Rentzeperis, D., Marky, L.A., Dwyer, T.J., Geierstanger, B.H., Pelton, J.G. and Wemmer, D.E. (1995) Biochemistry, 34, 2937-2945. MEDLINE Abstract

32. Ladbury, J.E. (1996) Chem. Biol., 3, 973-980. MEDLINE Abstract

33. Dunitz, J.D. (1994). Science, 264, 670.

34. Billeter, M., Güntert, P., Luginbühl, P. and Wüthrich, K. (1996) Cell, 85, 1057-1065. MEDLINE Abstract

*To whom correspondence should be addressed. Tel/Fax: +44 181 643 1675; Email: steve@iris5.icr.ac.uk

This page is run by Oxford University Press, Great Clarendon Street, Oxford OX2 6DP, as part of the OUP Journals
Comments and feedback: www-admin{at}oup.co.uk
Last modification: 4 Jun 1998
Copyright©Oxford University Press, 1998.
Add to CiteULike CiteULike   Add to Connotea Connotea   Add to Del.icio.us Del.icio.us    What's this?



This Article
Right arrow Abstract Freely available
Right arrow Print PDF (234K) Freely available
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Add to My Personal Archive
Right arrow Download to citation manager
Right arrow Search for citing articles in:
ISI Web of Science (18)
Right arrowRequest Permissions
Right arrow Commercial Re-use Guidelines
for Open Access NAR Content
Google Scholar
Right arrow Articles by Guerri, A.
Right arrow Articles by Neidle, S.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Guerri, A.
Right arrow Articles by Neidle, S.
Social Bookmarking
Add to CiteULike   Add to Connotea   Add to Del.icio.us  
What's this?