Stannin (SNN)

SNN is an 88 amino acid protein expressed only in vertebrates and exhibits a high level of conservation; the rat, mouse, and human SNN differ by only two amino acids in the carboxyl terminus. Lower vertebrates, such as frog (Xenopus laevis) and Zebrafish (Danio rerio), share 89 and 91% similarity with the human protein sequence, respectively. Structural analysis using.

Unlike other environmental neurotoxins (i.e., methylmercury, lead, m-dinitrobenzene, or polychlorinated biphenyls, etc.), organotins possess a high specificity of action. While trimethyltin chloride (TMT) causes lesions in specific regions of the hippocampus and neocortex, triethyltin chloride (TET) damage is localized within the spinal cord. Interestingly, it has been found that in mammalian organs such as brain, liver, and kidneys, organotins are progressively dealkylated to inorganic Sn(IV). It has been shown that the extent of this dealkylation correlates inversely with the length and stability of alkyl chains. Furthermore, delayed toxic action of polysubstituted organotins has been observed and associated with the rate of in vivo conversion of highly substituted organotins into their metabolites. Organotin dealkylation has also been detected in the environment, where alkyl group removal has been attributed to the action of UV light, chemical cleavage, and biological degradation by bacteria. Although the chemistry of organotin degradation has never been elucidated, organotin reactivity has been attributed to the nature of the C-Sn bond that can be attacked by both nucleophilic and electrophilic reagents.

To study the chemistry of organotin binding to biological dithiols, we synthesized a nine-residue peptide (SNN-PEP) corresponding to amino acids 29-37 of stannin (ILGCWCYLR) and examined its binding with different organotin compounds. Using circular dichroism (CD) and electrospray ionization mass spectrometry (ESI-MS) as probes, we determined the affinity and the stoichiometry of the SNN-PEP/organotin complexes formed (Buck et al., 2003).

CD spectra of SNN-PEP show that it undergoes a distinct structural change upon addition of different organotins.

CD spectra of SNN-PEP free (black) and titrated with DMT (red), TMT (orange), DET (green), TET (blue), and TPrT (purple) chlorides.

The peptide titrations with both TMT and DMT were carried out at pH 4.0 and monitored using 1D-¹H NMR spectroscopy (Buck, et al). Upon the addition of increasing amounts of TMT, two different populations of peaks are observed for the peptide in the proton spectrum. This is apparent, where the amide and the aromatic portions of the proton spectra are shown. In particular, the appearance of a second species in solution is evident for the Hε of W5 around 10.2 ppm, as indicated by arrows in below.

As confirmed by exchange spectroscopy, peak doubling corresponds to the free and alkyltin bound species of the peptide in slow exchange under the NMR time scale. E, above shows the cross correlation between the free and bound species of the peptide carried out at saturating concentrations of TMT. At lower concentrations of TMT, the peptide was predominantly in the free form, but at a ratio of 6:1 [TMT]:[SNN-PEP], the reaction equilibrium was approximately 50:50 between the free and bound peptide (A-C). These results are in agreement with the K_d values determined previously using CD spectroscopy. Only a large excess of TMT (greater than 50:1, [TMT]:[SNN-PEP]) drove the reaction equilibrium to the predominantly bound form of the peptide.

Since both TMT and DMT form identical complexes with SNN-PEP in solution, we proceeded by determining the structure of the peptide bound to DMT (Buck, et al). This allowed us to use equimolar concentrations of peptide and alkyltin, while titration with TMT would require a large excess of ligand complicating the NMR spectra due to the intense methyl resonance at ~1 ppm. Complete resonance assignments for both free and DMT bound SNN-PEP were determined from 2D-¹H-TOCSY spectra at pH 6.5. In both the free and DMT bound forms, the resonances were well dispersed allowing for straightforward spectral assignment. In the free form, the chemical shifts were representative of a random coil structure.

To determine the full backbone assignments of SNN, TROSY-based experiments as well as specifically ¹⁵N-labeled SNN samples were necessary (Buck, et al). While TROSY experiments improve the sensitivity of the slowly tumbling structural domains of SNN, they also ameliorate the resolution of resonances from the unstructured loop in ¹H/¹⁵N correlated experiments by suppressing the line broadening due to chemical exchange. The 3D walk for complete resonance assignment was determined using the HNCA/HN(CO)CA pathway. Three additional TROSY-based triple resonance experiments including HNCACB, CBCA(CO)NH and HNCO, were used to assign the C^β and C′ chemical shifts, respectively, contributing further to the SNN backbone assignment.

Superposition of the N^H, C^α, and C′ atoms in the transmembrane (residues 10–33) and the cytoplasmic domains (residues 61–79) for the 25 lowest energy structures of SNN. (a) Alignment of the backbone atoms with side-chains (backbone RMSD values of 0.26(±0.07)Å and 0.41(±0.17)Å), and heavy-atom side-chain values of 2.00(±0.48)Å and 2.07(±0.33)Å for the transmembrane and cytoplasmic domains, respectively). (b) Average minimized structure of SNN including the linker region. RMSD values were determined from comparison to the average minimized structure.

 

While the structural analysis of SNN in detergent micelles allows us to identify its secondary structural elements, the architecture of SNN in the membrane can be elucidated only by using solid-state NMR with SNN reconstituted in mechanically aligned lipid bilayers. For the solid-state NMR experiments, we used ³¹P spectroscopy to determine the formation of a uniformly orientated lipid bilayer, and ¹⁵N chemical shift anisotropy to trace the organization of SNN with respect to the lipid bilayer. The membrane orientation of SNN was obtained in a 4:1 mixture of 1,2-dioleoyl-sn-glycero-3-phosphocholine (DOPC):1,2-dioleoyl-sn-glycero-3-phosphoethanolamine (DOPE) and as determined by 1D-³¹P analysis on the aligned sample, revealing an oriented lipid bilayer with a characteristic resonance at 25ppm (a, below). In addition to the oriented component, a peak representing another residual lipid phase is present at ~15ppm. In analogy with studies carried out with other mixed membrane systems, the peaks at 25ppm and 15ppm have been assigned to protein-induced, DOPE-rich and DOPE-poor domains.

(a) Representative ³¹P spectrum showing the alignment of the DOPC/DOPE lipid bilayer after protein incorporation. (b) Powder pattern spectrum of full-length SNN in 4:1 DOPC:DOPE lipid bilayers.

The spectrum of mechanically aligned, uniformly ¹⁵N labeled SNN reported in (below, c) has three major components: a cluster of peaks from ~50–100ppm, a broad peak from ~100 to 150ppm, and a cluster of peaks from ~160 to 230ppm. To help de-convolute the spectrum of the full protein, we dissected SNN into two major fragments (SNN_1-48 and SNN_30-88) and studied them independently. Dissection of membrane proteins to facilitate the assignment of solid-state NMR spectra was applied previously to the monotopic membrane protein Vpu.

¹⁵N cross-polarization spectra of oriented (a) SNN_30-88, (b)SNN_1-48, and (c) SNN aligned in 4:1 DOPC:DOPE lipid bilayers. SIMSPEC calculations of the chemical shift at various tilt angles indicates that the transmembrane domain and cytoplasmic domains are aligned at 20° (blue) and 80° (red) with respect to the magnetic field, respectively.

The ¹⁵N spectra of oriented uniformly ¹⁵N labeled SNN, SNN_1-48 and SNN_30-88 are shown above. The spectrum for uniformly ¹⁵N labeled SNN_30-88 is characterized by an envelope of resonances located between 50ppm and 100ppm, indicating that the fragment that contains the cytoplasmic helix of SNN is oriented approximately parallel with the membrane bilayer plane. This spectrum aligns well with that of the full-length SNN, confirming that the cytoplasmic helix is localized preferentially on the surface of the membrane bilayer. These results are consistent with the solution NMR paramagnetic quenching and H/D exchange data. On the other hand, the spectrum from SNN_1-48 gives two major peaks, one spanning from 100 to 150ppm and the second from 160 to 230ppm. The intense signal centered at ~117ppm lies close to the isotropic limit of the ¹⁵N chemical shift tensor and can be assigned to the unstructured regions of the protein, while the signals ranging between 160ppm and 230pm indicate that this fragment is oriented approximately perpendicular to the plane of the membrane bilayer. Again, this spectrum aligns well with that of the oriented full-length SNN, and can be attributed to the transmembrane domain.

Publications

Buck-Koehntop, B., Mascioni, A., Buffy, J.J, and Veglia, G., Structure, Dynamics, and Membrane Topology of Stannin: A Mediator of Neuronal Cell Apoptosis Induced by Trimethyltin Chloride J. Mol. Biol. 354, 3, 2005, 652-665.

Buck, B.A., Mascioni, A., Cramer, C.J., Veglia, G., Interactions of alkyltin salts with biological dithiols: dealkylation and induction of a regular beta-turn structure in peptides J. Am. Chem. Soc. 2004 Nov 10;126(44):14400-10.

Buck, B., Mascioni, A., Que, L. Jr., and Veglia G., De-alkylation of Organotin Compounds by Biological Di-Thiols: Toward the Chemistry of Organotin Toxicity, J. Am. Chem. Soc. 125(44); 13316-13317 (2003).