Molecular basis of nitrate uptake by the plant nitrate transporter NRT1.1

Article metrics

Abstract

The NRT1/PTR family of proton-coupled transporters are responsible for nitrogen assimilation in eukaryotes and bacteria through the uptake of peptides. However, in most plant species members of this family have evolved to transport nitrate as well as additional secondary metabolites and hormones. In response to falling nitrate levels, NRT1.1 is phosphorylated on an intracellular threonine that switches the transporter from a low-affinity to high-affinity state. Here we present both the apo and nitrate-bound crystal structures of Arabidopsis thaliana NRT1.1, which together with in vitro binding and transport data identify a key role for His 356 in nitrate binding. Our data support a model whereby phosphorylation increases structural flexibility and in turn the rate of transport. Comparison with peptide transporters further reveals how the NRT1/PTR family has evolved to recognize diverse nitrogenous ligands, while maintaining elements of a conserved coupling mechanism within this superfamily of nutrient transporters.

Access options

Rent or Buy article

Get time limited or full article access on ReadCube.

from$8.99

All prices are NET prices.

Figure 1: Structural and biophysical characterization of NRT1.1.
Figure 2: Crystal structure of nitrate bound NRT1.1.
Figure 3: Proposed mechanism for the effect of phosphorylation at Thr 101 on nitrate transport.
Figure 4: Alternating access model for nitrate-proton symport.

Accession codes

Accessions

Protein Data Bank

Data deposits

The atomic coordinates and structure factors of the apo and nitrate bound NRT1.1 are deposited in the Protein Data Bank with accession codes 4cl4 and 4cl5, respectively.

References

  1. 1

    Steiner, H.-Y., Naider, F. & Becker, J. M. The PTR family: a new group of peptide transporters. Mol. Microbiol. 16, 825–834 (1995)

  2. 2

    Daniel, H., Spanier, B., Kottra, G. & Weitz, D. From bacteria to man: archaic proton-dependent peptide transporters at work. Physiology (Bethesda) 21, 93–102 (2006)

  3. 3

    Crawford, N. M. Nitrate: nutrient and signal for plant growth. Plant Cell 7, 859–868 (1995)

  4. 4

    Tsay, Y. F. & Hsu, P. K. The role of plasma membrane nitrogen transporters in nitrogen acquisition and utilization. Plant Cell Monogr. 19, 223–236 (2011)

  5. 5

    Orsel, M. et al. Characterization of a two-component high-affinity nitrate uptake system in arabidopsis. physiology and protein-protein interaction. Plant Physiol. 142, 1304–1317 (2006)

  6. 6

    Dechorgnat, J. et al. From the soil to the seeds: the long journey of nitrate in plants. J. Exp. Bot. 62, 1349–1359 (2011)

  7. 7

    Tsay, Y. F., Schroeder, J. I., Feldmann, K. A. & Crawford, N. M. The herbicide sensitivity gene CHL1 of Arabidopsis encodes a nitrate-inducible nitrate transporter. Cell 72, 705–713 (1993)

  8. 8

    Léran, S. et al. A unified nomenclature of NITRATE TRANSPORTER 1/PEPTIDE TRANSPORTER family members in plants. Trends Plant Sci. 19, 5–9 (2013)

  9. 9

    Tsay, Y.-F., Chiu, C.-C., Tsai, C.-B., Ho, C.-H. & Hsu, P.-K. Nitrate transporters and peptide transporters. FEBS Lett. 581, 2290–2300 (2007)

  10. 10

    Boursiac, Y. et al. ABA transport and transporters. Trends Plant Sci. 18, 325–333 (2013)

  11. 11

    Nour-Eldin, H. H., Andersen, T. G., Burow, M. & Madsen, S. R. NRT/PTR transporters are essential for translocation of glucosinolate defence compounds to seeds. Nature 488, 531–534 (2012)

  12. 12

    Krouk, G. et al. Nitrate-regulated auxin transport by NRT1.1 defines a mechanism for nutrient sensing in plants. Dev. Cell 18, 927–937 (2010)

  13. 13

    Reddy, V. S., Shlykov, M. A., Castillo, R., Sun, E. I. & Saier, M. H. The major facilitator superfamily (MFS) revisited. FEBS J. 279, 2022–2035 (2012)

  14. 14

    Newstead, S. et al. Crystal structure of a prokaryotic homologue of the mammalian oligopeptide-proton symporters, PepT1 and PepT2. EMBO J. 30, 417–426 (2011)

  15. 15

    Fei, Y. J. et al. Expression cloning of a mammalian proton-coupled oligopeptide transporter. Nature 368, 563–566 (1994)

  16. 16

    Nakajima, H. et al. Cloning and functional expression in Escherichia coli of the gene encoding the di- and tripeptide transport protein of Lactobacillus helveticus. Appl. Environ. Microbiol. 63, 2213–2217 (1997)

  17. 17

    Chiang, C.-S., Stacey, G. & Tsay, Y.-F. Mechanisms and functional properties of two peptide transporters, AtPTR2 and fPTR2. J. Biol. Chem. 279, 30150–30157 (2004)

  18. 18

    Radestock, S. & Forrest, L. R. The alternating-access mechanism of MFS transporters arises from inverted-topology repeats. J. Mol. Biol. 407, 698–715 (2011)

  19. 19

    Yan, N. Structural advances for the major facilitator superfamily (MFS) transporters. Trends Biochem. Sci. 38, 151–159 (2013)

  20. 20

    Madej, M. G. & Kaback, H. R. Evolutionary mix-and-match with MFS transporters. Proc. Natl Acad. Sci. USA 110, E4831–E4838 (2013)

  21. 21

    Solcan, N. et al. Alternating access mechanism in the POT family of oligopeptide transporters. EMBO J. 31, 3411–3421 (2012)

  22. 22

    Doki, S. et al. Structural basis for dynamic mechanism of proton-coupled symport by the peptide transporter POT. Proc. Natl Acad. Sci. USA 110, 11343–11348 (2013)

  23. 23

    Liu, K. H., Huang, C. Y. & Tsay, Y. F. CHL1 is a dual-affinity nitrate transporter of Arabidopsis involved in multiple phases of nitrate uptake. Plant Cell 11, 865–874 (1999)

  24. 24

    Liu, K.-H. & Tsay, Y.-F. Switching between the two action modes of the dual-affinity nitrate transporter CHL1 by phosphorylation. EMBO J. 22, 1005–1013 (2003)

  25. 25

    Wang, Y.-Y., Hsu, P.-K. & Tsay, Y.-F. Uptake, allocation and signaling of nitrate. Trends Plant Sci. 17, 458–467 (2012)

  26. 26

    Baaske, P., Wienken, C., Willemsen, M. J. & Braun, D. Protein-binding assays in biological liquids using microscale thermophoresis. J. Biomol. Tech. 22, S55 (2011)

  27. 27

    Wienken, C. J., Baaske, P., Rothbauer, U. & Braun, D. Protein-binding assays in biological liquids using microscale thermophoresis. Nature Commun. 1, 100 (2010)

  28. 28

    Sun, L. et al. Crystal structure of a bacterial homologue of glucose transporters GLUT1–4. Nature 490, 361–366 (2012)

  29. 29

    Dang, S. et al. Structure of a fucose transporter in an outward-open conformation. Nature 467, 734–738 (2010)

  30. 30

    Guettou, F. et al. Structural insights into substrate recognition in proton-dependent oligopeptide transporters. EMBO Rep. 14, 804–810 (2013)

  31. 31

    Yan, H. et al. Structure and mechanism of a nitrate transporter. Cell Rep. 3, 716–723 (2013)

  32. 32

    Chen, X.-Z., Zhu, T., Smith, D. E. & Hediger, M. A. Stoichiometry and kinetics of the high-affinity H+-coupled peptide transporter PepT2. J. Biol. Chem. 274, 2773–2779 (1999)

  33. 33

    Chandrasekaran, A., Ojeda, A. M., Kolmakova, N. G. & Parsons, S. M. Mutational and bioinformatics analysis of proline- and glycine-rich motifs in vesicular acetylcholine transporter. J. Neurochem. 98, 1551–1559 (2006)

  34. 34

    Ugolev, Y., Segal, T., Yaffe, D., Gros, Y. & Schuldiner, S. Identification of conformationally sensitive residues essential for inhibition of vesicular monoamine transport by the noncompetitive inhibitor-tetrabenazine. J. Biol. Chem. 288, 32160–32171 (2013)

  35. 35

    Newstead, S. et al. High-throughput fluorescent-based optimization of eukaryotic membrane protein overexpression and purification in Saccharomyces cerevisiae. Proc. Natl Acad. Sci. USA 104, 13936–13941 (2007)

  36. 36

    Kazemi Seresht, A., Nørgaard, P., Palmqvist, E. A., Andersen, A. S. & Olsson, L. Modulating heterologous protein production in yeast: the applicability of truncated auxotrophic markers. Appl. Microbiol. Biotechnol. 97, 3939–3948 (2013)

  37. 37

    Drew, D. et al. GFP-based optimization scheme for the overexpression and purification of eukaryotic membrane proteins in Saccharomyces cerevisiae. Nature Protocols 3, 784–798 (2008)

  38. 38

    Newstead, S., Ferrandon, S. & Iwata, S. Rationalizing α-helical membrane protein crystallization. Protein Sci. 17, 466–472 (2008)

  39. 39

    Winter, G., Lobley, C. M. C. & Prince, S. M. Decision making in xia2. Acta Crystallogr. D 69, 1260–1273 (2013)

  40. 40

    Winter, G. xia2: an expert system for macromolecular crystallography data reduction. J. Appl. Cryst. 43, 186–190 (2009)

  41. 41

    Kabsch, W. XDS. Acta Crystallogr. D 66, 125–132 (2010)

  42. 42

    Evans, P. R. An introduction to data reduction: space-group determination, scaling and intensity statistics. Acta Crystallogr. D 67, 282–292 (2011)

  43. 43

    Sheldrick, G. M. Experimental phasing with SHELXC/D/E: combining chain tracing with density modification. Acta Crystallogr. D 66, 479–485 (2010)

  44. 44

    Abrahams, J. P. & Leslie, A. G. W. Methods used in the structure determination of bovine mitochondrial F1 ATPase. Acta Crystallogr. D 52, 30–42 (1996)

  45. 45

    Cowtan, K. DM: an automated procedure for phase improvement by density modification. Joint CCP4 and ESF-EACBM Newsletter on Protein Crystallography 31, 34–38 (1994)

  46. 46

    Pedersen, B. P., Morth, J. P. & Nissen, P. Structure determination using poorly diffracting membrane-protein crystals: the H+-ATPase and Na+,K+-ATPase case history. Acta Crystallogr. D 66, 309–313 (2010)

  47. 47

    Jones, T. A., Zou, J. Y., Cowan, S. W. & Kjeldgaard, M. Improved methods for building protein models in electron density maps and the location of errors in these models. Acta Crystallogr. A 47, 110–119 (1991)

  48. 48

    Yang, Z. et al. UCSF Chimera, MODELLER, and IMP: An integrated modeling system. J. Struct. Biol. 179, 269–278 (2012)

  49. 49

    Blanc, E. et al. Refinement of severely incomplete structures with maximum likelihood in BUSTER-TNT. Acta Crystallogr. D 60, 2210–2221 (2004)

  50. 50

    Davis, I. W., Murray, L. W., Richardson, J. S. & Richardson, D. C. MOLPROBITY: structure validation and all-atom contact analysis for nucleic acids and their complexes. Nucleic Acids Res. 32, W615–W619 (2004)

  51. 51

    Schrodinger, LLC. The PyMOL Molecular Graphics System, Version 1.3r1

  52. 52

    Foucaud, C. & Poolman, B. Lactose transport system of Streptococcus thermophilus. Functional reconstitution of the protein and characterization of the kinetic mechanism of transport. J. Biol. Chem. 267, 22087–22094 (1992)

  53. 53

    Knol, J. et al. Unidirectional reconstitution into detergent-destabilized liposomes of the purified lactose transport system of Streptococcus thermophilus. J. Biol. Chem. 271, 15358–15366 (1996)

  54. 54

    Smirnova, I. et al. Sugar binding induces an outward facing conformation of LacY. Proc. Natl Acad. Sci. USA 104, 16504–16509 (2007)

  55. 55

    Zheng, H., Wisedchaisri, G. & Gonen, T. Crystal structure of a nitrate/nitrite exchanger. Nature 497, 647–651 (2013)

  56. 56

    Katoh, K. & Toh, H. Recent developments in the MAFFT multiple sequence alignment program. Brief. Bioinform. 9, 286–298 (2008)

  57. 57

    Clamp, M., Cuff, J., Searle, S. M. & Barton, G. J. The Jalview Java alignment editor. Bioinformatics 20, 426–427 (2004)

Download references

Acknowledgements

We thank R. Flaig for help with additional access to beamline IO4 and we thank Diamond Light Source for access to beam lines I02, I03, I04 and I04-1 (MX7345) that contributed to the results presented here. This research was funded through the Medical Research Council (MRC) Career Development Award grant G0900399 and Royal Society grants (RG110211 and IE111401) to S.N.

Author information

J.L.P. and S.N. designed, performed and analysed all experiments and wrote the manuscript.

Correspondence to Joanne L. Parker or Simon Newstead.

Ethics declarations

Competing interests

The authors declare no competing financial interests.

Extended data figures and tables

Extended Data Figure 1 Alignment of NRT1.1 with PTR family members.

Amino acid alignment of A. thaliana NRT1.1 (Uniprot: Q05085) with PepTSo (Q8EKT7), PepTSt (Q5M4H8), human PepT1 (B2CQT6) and PepTGk (Q5KYD1) using MAFFT56 with manual adjustment in JalView57. AtNRT1.1 shares 23% identity to human PepT1 and 22% identity to PepTSo at the primary structure level. Identical residues shared across the NRT1/PTR family are highlighted in red with key functionally conserved residues highlighted by blue triangles. Nitrate binding site residues His 356, Tyr 388 are highlighted as magenta stars. Thr 101 is conserved in all mammalian PTR family members and a small subset of bacterial homologues including PepTSo. NRT1.1 topology: the central cavity is shown as an open triangle, representing the inward open state observed in the crystal structure. The 12 TM helices identified from the crystal structures are coloured from blue to red and the observed intracellular domain between TM6–TM7 is indicated in grey.

Extended Data Figure 2 Microscale thermophoresis binding assay as a technique to calculate KD for ligand binding to NRT1/PTR transporters.

a, The accuracy of the microscale thermophoresis binding assay was examined through comparison with data obtained from transport assays. PepTSo was functionally reconstituted into liposomes as described and an IC50 value calculated for Ala-Ala was determined and found to be approximately 50 μM (a, left). Using the NanoTemper monolith NT.115 instrument the KD for Ala-Ala was calculated to be approximately 35 μM (a, right) similar to the IC50 calculated from proton-driven transport assays. In this assay the orientation of the PepTSo molecules are highly likely to adopt multiple conformations, as observed for other members of the MFS54, and so the KD we have determined is a global parameter for this protein and should be the same for either conformation of the transporter. b, The binding of nitrate to NRT1.1 was determined at pH 7.0, 6.5 and 6.0; the KD calculated at each of these pH values did not markedly differ from one another. However, as the pH was reduced the fluorescent signal produced by GFP was partially quenched, therefore all subsequent analysis was performed at a pH of 6.5. Nitrite, alanine, di-alanine, phosphate and sulphate show no observable interaction with AtNRT1.1 using the MST binding assay under the same conditions as for nitrate. c, Chlorate was shown to bind NRT1.1 with a KD of approximately 5 mM. Data shown are representative of three independent experiments and mean and standard deviations are shown.

Extended Data Figure 3 Purification, secondary structure analysis and crystallization of NRT1.1.

a, Gel filtration trace of the purified NRT1.1 after tobacco etch virus cleavage of the GFPHis. The protein elutes as a single peak corresponding to a size of approximately 65 kDa following subtraction of the DDM micelle (50 kDa), which would correspond to a monomer in solution. b, The protein was purified to homogeneity as judged by Coomassie-stained SDS–PAGE. c, Analysis of secondary structure content of NRT1.1 using circular dichroism spectroscopy shows the protein is predominately alpha helical with 22% disordered regions and no significant beta strands. d, Typical crystal obtained with scale bar at 100 μm. e, AtNRT1.1 crystallizes as a dimer in the asymmetric unit. The buried surface area between the two monomers is 2,135 Å2 as calculated using PDBe PISA (http://www.ebi.ac.uk/msd-srv/prot_int/) and may be physiologically significant as both monomers are in the same orientation. Structural comparisons between bacterial members of the PTR family in different conformations of the transport cycle suggest the N-terminal bundle is less flexible compared to the C-terminal helices21, which may explain why we observe NRT1.1 packing through these helices in the crystal. The oligomeric state of the PTR family in the membrane is currently unclear, with a recent bacterial member conjectured to form a tetramer30. In this case however, the dimer interface was found to be via the C- rather than the N-terminal bundle.

Extended Data Figure 4 Extracellular gate of NRT1.1 is stabilized through hydrophobic interactions as opposed to the salt bridges observed in the bacterial and mammalian members of this family.

NRT1.1 is shown in the plane of the membrane with a slice through the protein volume highlighting the central nitrate binding site and the exit tunnel. The extracellular gate formed from TM1 and TM2 packing against TM7 and TM8 is highlighted and the interface between these helices is illustrated using electrostatic surface representations.

Extended Data Figure 5 Representative electron density maps used during model building and refinement.

Stereo view of final 2mFo − DFc electron density from refinement, contoured at 1σ around chain A (PDB: 4CL5). View shown is equivalent to Fig. 2b in main text (PDB: 4CL4). Nitrate binding site: the position of modelled nitrate was resolved much better in the electron density maps in chain A compared to chain B, possibly due to higher occupancy in this molecule and the analysis in the paper was therefore conducted on chain A. In chain B the nitrate was seen to occupy a slightly different position within the binding site, displaced down approximately 1.7 Å compared to the position in molecule A but still sitting close to and interacting with His 356. Given the resolution is 3.7 Å; it is difficult to ascertain the significance of this observation with respect to the coordination of nitrate. The low KD value determined from the binding experiments could be consistent with nitrate being able to adopt slightly different positions within the binding site with respect to His 356.

Extended Data Figure 6 Structural comparison between nitrate binding sites in NRT1.1 and the bacterial NNP family members NarK and NarU.

Although NRT1.1 shares no sequence identity to the nitrate-nitrite porter (NNP) family, they do both belong to the major facilitator superfamily, and share the same canonical architecture of 12 transmembrane helices4. a, Superposition of NarU and NarK (both from E. coli) with NRT1.1; NarU and NarK crystal structures were recently published in complex with nitrate and nitrite, respectively31,55. Consistent with previously determined crystal structures for ligand-bound MFS family members the ligand-binding site is located in the centre of the molecule, constructed from side chains contributed by the N- and C-terminal bundles. In NRT1.1 the nitrate in chain A is located much further into the cavity, approximately 9 Å, than observed in either NarU or NarK and makes far fewer interactions with side chain functional groups. However, as discussed in the main text the binding sites do exhibit similar physiochemical features. b, c, As illustrated in the close-up of the binding sites (b) and depicted in the cartoon schematic below (c), the nitrate in NarU is held tightly by two electrostatic interactions to Arg 87 and Arg 303 and by two further hydrogen bonds to Asn 173 and Tyr 261 (yellow sticks). These interactions are further sandwiched between two layers of hydrophobic residues, which will probably increase the relative strength of the electrostatic interactions by lowering the dielectric constant in the immediate vicinity of the nitrate. By comparison and as discussed in the main text, in NRT1.1 the nitrate is also held by an electrostatic interaction with a protonated His 356 with contributions in chain A from Thr 360 (purple sticks). The immediate vicinity of the nitrate in the present conformation of NRT1.1 is also hydrophobic and highlighted (purple sticks and black oval). However, as shown in the cartoon schematic, the number and strength of these interactions are substantially lower compared to the NNP family of transporters, potentially explaining the much lower KD we observe for NRT1.1 (1 mM) compared to NarU (33 μM).

Extended Data Figure 7 Superposition of NRT1.1 onto the outward facing crystal structure of the fucose permease, FucP.

The coordinates for the NRT1.1 structure was divided into the N-and C-terminal six-helix bundles, coloured blue and orange, respectively, before being superimposed onto the structure of FucP (PDB: 3O7Q) using the CCP4 Superpose program. The r.m.s.d. values are indicated. Helices 2, 7 and 11 were removed for clarity. Superposition of both N and C bundles of NRT1.1 onto the proton-coupled fucose transporter FucP places Lys 164 in the same position as a conserved Glu 135 in FucP, which upon protonation is predicted to trigger the switch to the inward-facing state29. Similarly, in superimposing the respective C-terminal domains, Glu 476 of NRT1.1, which this study suggests is involved in structural aspects of nitrate binding through an electrostatic interaction with His 356 (see Fig. 2b and Extended Data Fig. 5), is close to Tyr 365 in FucP, which is the proposed interaction partner of Glu 135. A speculative function of Lys 164 may therefore be to form a salt bridge with Glu 476 in the outward open conformation (Fig. 4a), facilitating the closer packing of helices H4 and H10 and sealing the binding site to the interior of the cell in the outward-facing state. Supporting this interaction in the POT family are mutational data from both PepTSt and GkPOT, which show these residues are important for proton-driven peptide transport in the POT family21,22.

Extended Data Figure 8 Proton transport is increased in the Thr101Asp variant of NRT1.1 when compared to wild-type protein.

A reconstituted proteoliposome system was used to study the effect of the Thr101Asp variant on nitrate uptake. a, In this assay the membrane impermeant pH indicator, pyranine (trisodium 8-hydroxypyrene-1,3,6-trisulfonate) was trapped within the NRT1.1 reconstituted liposomes. In the presence of a pH gradient, NRT1.1 couples the import of nitrate with protons into the lumen of the liposome. b, The trapped pyranine becomes protonated, whereupon the excitation maximum shifts from 460 to 415 nm, the emission maximum remaining constant at 510 nm. Therefore proton-coupled nitrate transport can be measured as a function of the proton-dependent quenching of pyranine. c, Protein was reconstituted into liposomes as detailed in the Methods and assays were performed on the same amount of protein for the wild-type, Thr101Asp and His356Ala variants. d, Representative data obtained from the nitrate transport assays showing three independent experiments for both protein free liposomes and those containing the variants were studied. Nitrate uptake was driven using a pH gradient (pH 6.8 inside, 6.0 outside) as the addition of valinomycin caused an anomalous proton leak in NRT1.1-containing liposomes. The difference between the addition of nitrate (blue) verses a no nitrate control (black) is shown, data were normalized to the values obtained before the addition of nitrate. Due to the use of a plate reader to measure the experiments, after the addition of nitrate there is a slight delay (10 s) before the first measurement is taken (time zero) which is why at time zero a drop in fluorescence is observed. The presence of wild-type NRT1.1 caused a larger change in fluorescence (indicative of acidification) in the presence of nitrate, which was not seen with the inactive His356Ala variant. The Thr101Asp variant, however, caused a substantial increase in acidification indicative of more protons and therefore nitrate being transported by this variant.

Extended Data Figure 9 Effects of mutating Thr 87 in PepTSo on protein stability, peptide binding and proton-coupled peptide uptake in a reconstituted proteoliposome system.

a, The N-terminal domains of NRT1.1 (also shown in Fig. 3a) and the bacterial peptide transporters are very similar. The location of Thr 101 and the equivalent threonine in PepTSo, Thr 87, are highlighted as magenta spheres, both threonines are located in a similar hydrophobic cavity formed at the intracellular ends of TM2 and TM4. b, Thr87Asp variant of PepTSo showed a fourfold increase transport rate, whereas the Thr87Arg variant was inactive. c, Mutation to alanine, serine or lysine had no significant impact on the rate of transport. Error bars represent s.d. from three independent experiments. d, Comparison of substrate specificity between wild type and The87Asp variant showed no significant change, indicating that the binding site has not been adversely affected. Error bars represent s.d. from three independent experiments. e, The stability as assessed using a high-performance circular dichroism spectrometer (Jasco J-1500 instrument) of the Thr87Asp (Tm = 52 °C) and Arg (Tm = 46 °C) mutants is reduced when compared to wild-type protein (56 °C). Loss of alpha helical secondary structure was monitored continuously at 220 nm while the temperature was increased from 20 °C to 80 °C over 1-degree increments. 10 μg of protein were monitored in a 100-μl cuvette. f, Binding isotherms showing there is no change in affinity for peptide upon mutating Thr 87 to Asp.

Extended Data Table 1 Data collection and refinement statistics

PowerPoint slides

Rights and permissions

Reprints and Permissions

About this article

Cite this article

Parker, J., Newstead, S. Molecular basis of nitrate uptake by the plant nitrate transporter NRT1.1. Nature 507, 68–72 (2014) doi:10.1038/nature13116

Download citation

Further reading

Comments

By submitting a comment you agree to abide by our Terms and Community Guidelines. If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.