Fig. 1. ¹H-¹⁵N HSQC spectra of Tlg2p fragments. (A) ¹H-¹⁵N HSQC spectrum of Tlg2p(65–192). (B and C) Superposition of the ¹H-¹⁵N HSQC spectra of Tlg2p(60–283) (black contours) (B), or Tlg2p(1–192) (cyan contours) (C), with the spectrum of Tlg2p(65–192) (red contours).

Fig. 3. ¹H-¹⁵N HSQC spectra of Pep12p fragments. (A) ¹H-¹⁵N HSQC spectrum of Pep12p(17–144). (B) Superposition of the ¹H-¹⁵N HSQC spectrum of Pep12p(17–253) (black contours) with the spectrum of Pep12p(17–144) (red contours). (C) Domain structure of Pep12p.

Fig. 5. Characterization of syntaxin 16. (A) Proteins encoded by various splice variants of human syntaxin 16. Three different full-length syntaxin 16 variants were reported in the literature: syntaxin 16H (synt16H) (Tang et al., 1998), syntaxin 16A (synt16A) and syntaxin 16B (synt16B) (Simonsen et al., 1998). Note that all three published syntaxin 16 sequences probably contain an internal frameshift error (corrected in our translation here, residues from Ala125 to Leu148 according to synt16H numbering) since the cDNAs we sequenced and the conceptual sequence deduced from the genome product (accession no. XM_030663) agree completely. The Synt16H, -A and -B sequences differ in an alternatively spliced region located between the N-terminal peptide that binds to mVps45 (see below) and the H_abc domain. The truncated group consists of a splice variant called syntaxin 16C (Simonsen et al., 1998) that contains a stop codon in the middle of the predicted H_b helix, and a new splice variant called syntaxin 16D (image clone 4413533; accession no. BG035083) that contains a stop codon after the H_abc domain but before the SNARE motif. The N-terminal variations that distinguish between Synt16H, -A and -B are probably also encoded by the mRNAs of the truncated syntaxin 16 variants. In the alignment, asterisks indicate positions of stop codons and dashes represent gaps. The double-underlined sequence corresponds to the N-terminal peptide sequence that binds to mVps45. Sequences in bold represent predicted α-helices of the H_abc domain (based on sequence comparison with Tlg2p), the italicized sequence underlined with a dotted line represents the SNARE motif, and the box demarks the transmembrane region. Numbers on the right correspond to syntaxin 16H splice variant. (B) Syntaxin 16 tissue distribution. Equal amounts of homogenates from the indicated mouse tissues were analyzed by immuno blotting with antibodies to syntaxin 16 (upper panel) and VCP (vasolin-containing protein, lower panel). Arrows indicate the position of syntaxin 16 (upper panel) and VCP (lower panel). (C) Expression of various splice variants of syntaxin 16 in a mammalian cell line. Mouse brain homogenate and extracts of COS-1 cells expressing the full-length human syntaxin 16C, syntaxin 16D and the entire cytoplasmic region of syntaxin 16H (residues 1–284) were analyzed by immunoblotting with antibodies to syntaxin 16. Numbers on the left indicate positions of molecular weight markers.

Fig. 7. Mammalian syntaxin 16C binds to mVps45p via an evolutionarily conserved N-terminal peptide motif. (A) Yeast two-hybrid assay. Yeast cells were co-transfected with pLexN-mVps45 and the indicated pVP16-syntaxin 16C prey clones (numbers indicate amino acid residues included in the constructs or mutated in the context of full-length syntaxin 16C). The relative β-galactosidase activity was calculated as described in the legend to Figure 4. Data shown are means ± SE from quadruple determinations. (B–D) GST-pulldown assays. In (B), glutathione–Sepharose beads containing GST alone or GST–syntaxin 16C fragments were incubated with extract of COS-1 cells expressing myc-tagged mVps45. After washing, the beads were analyzed by SDS–PAGE and visualized by immunoblotting with anti-myc·tag monoclonal antibody. In (C) and (D), GST alone, GST–syntaxin 1A (used as a negative control) and different GST–syntaxin 16 constructs were incubated with bacterial extract expressing T7-tagged mVps45. The beads were processed as in (B) and gels were visualized by Coomassie Blue staining (upper panels) or by immunoblotting with anti-T7·Tag monoclonal antibody (lower panels). Note that the recombinant mVps45 co-migrates with GST–syntaxin 16H^1–284 and cannot be seen on the Coomassie Blue-stained gel (D, upper panel). Numbers on the left indicate the positions of molecular mass markers. Asterisks indicate the positions of GST and GST-fusion proteins and the arrows indicate mVps45.

Fig. 9. Syntaxin–SM protein coupling in membrane traffic. (A) Two different modes of direct syntaxin/SM protein interactions currently known. In the upper scheme, the neuronal syntaxin 1 interacts with munc18-1 in a closed conformation (left), while core complex formation requires an open conformation of syntaxin 1 (right). Syntaxin 1 is colored in light blue (N-terminal H_abc domain), white (linker region) and blue (SNARE motif), while munc18-1 is colored in orange, and other SNARE motifs involved in the core complex are colored in yellow and red (for the vesicular SNARE). ‘N’ indicates the N-terminus of syntaxin 1, helices are represented by cylinders and the attachment points to the corresponding membranes are represented by thick gray lines. In the closed conformation, only part of the SNARE motif binds to the H_abc domain and has a defined structure (Dulubova et al., 1999; Misura et al., 2000). It is currently unknown if, in vivo, there is a true equilibrium between the two states of syntaxin 1 (hence the ‘?’). The lower scheme illustrates the N-terminal peptide-based interaction between the TGN/endosomal Tlg2p/syntaxin 16 and Vps45, which is analogous to the interaction of the ER Ufe1p/syntaxin 18 and the Golgi Sed5p/syntaxin 5 with Sly1p. The coloring scheme is the same, except that the N-terminal peptide motif of Tlg2p/syntaxin 16 is colored in green. The diagram emphasizes the fact that this mode of interaction is compatible with core complex formation and allows a simultaneous active role for the core complex and the SM protein in membrane fusion. (B) Summary of membrane traffic pathways in yeast. The syntaxins (blue labels) and SM proteins (orange labels) involved in each pathway are indicated. The red box encloses all the membrane compartments where syntaxin/SM protein coupling occurs by an N-terminal peptide-based mechanism. In the late endosome/prevacuolar compartment (PVC) and the vacuole (included in the blue box), syntaxins appear to interact indirectly with SM proteins incorporated into large protein complexes. A different mode of syntaxin/SM protein coupling may operate in the plasma membrane (green box). CV, constitutive secretory vesicles; EE, early endosome; ER, endoplasmic reticulum; LE, late endosome; TGN, trans-Golgi network. Adapted from Südhof and Scheller (2001).

Fig. 2. Domain structure of Tlg2p. (A) Definition of the three helical H_abc domain of Tlg2p. Differences (ΔδC_α) between the C_α chemical shifts observed for residues 65–192 of Tlg2p and those expected for a random coil are plotted against the residue number. Three regions with large positive ΔδC_α values indicate the three α-helices. (B) Schematic diagram of the Tlg2p sequence. Residue numbers are shown above the diagram. Hatched boxes correspond to different structural elements in Tlg2p. The approximate limits of the three helices of the H_abc domain were assigned based on chemical shift indices (Wishart and Sykes, 1994) deduced from the observed ΔδC_α values shown in (A).

Fig. 4. Tlg2p binds to Vps45p via a short evolutionarily conserved N-terminal sequence. (A) Yeast two-hybrid assay. Yeast cells were co-transfected with pLexN-Vps45p and the indicated pVP16-Tlg2p or pVP16-Pep12p prey clones (numbers indicate amino acid residues included in the constructs). The relative β-galactosidase activity was calculated as 1000 × OD₄₂₀/[min × vol. yeast (ml) × OD₆₀₀]. Data shown are means ± SE from quadruple determinations. (B–D) GST-pulldown assays. In (B), glutathione–Sepharose beads containing GST alone or the GST–Tlg2p(1–192) fragment were incubated with bacterial extracts prepared from cells expressing T7-tagged Vps45p. After extensive washes, the beads were separated on SDS–PAGE and visualized by Coomassie Blue-staining (upper panel) or by immunoblotting with T7·Tag antibody (lower panel). In (C), GST alone, GST– Pep12p(2–253) and different GST–Tlg2p fragments were similarly used in pulldown experiments and analyzed by immunoblotting with anti-T7·Tag monoclonal antibody. In (D), a similar procedure was applied to GST alone, or to GST–Tlg2p(1–192) fragments (either wild type or with the indicated mutations). The upper panel shows Coomassie Blue-stained PAGE and the lower panel shows the corresponding immunoblot with anti-T7·Tag monoclonal antibody. In (B–D), numbers on the left indicate the positions of molecular mass markers. Asterisks show the positions of GST and GST-fusion proteins and the arrows indicate Vps45p.

Fig. 6. ¹H-¹⁵N HSQC spectra of syntaxin 16 fragments. (A) ¹H-¹⁵N HSQC spectrum of the full-length syntaxin 16C. (B) ¹H-¹⁵N HSQC spectrum of the syntaxin 16H fragment corresponding to the predicted H_abc domain (residues 59–183).

Fig. 8. Specificity of syntaxin/SM protein interactions. (A) Sequence alignment of syntaxin N-terminal peptide motifs involved in binding to SM proteins. The aligned sequences have been divided in two classes: Tlg2/syntaxin 16 homologs, which bind to Vps45 (on the top), and Ufe1/syntaxin 18 and Sed5/syntaxin 5 homologs, which bind to Sly1 (on the bottom). Identical amino acid residues common to both classes and present in >50% of the sequences are shown on the yellow background. Class-specific residues present in >50% of the sequences from the given class are shown in white on either a blue (Vps45-binding syntaxins) or a pink (Sly1-binding syntaxins) background. Hs, Homo sapiens; Mm, Mus musculus; Ce, Caenorhabditis elegans; Dm, Drosophila melanogaster; At, Arabidopsis thaliana; Sc, Schizosaccharo myces cerevisiae; Sp, Schizosaccharomyces pombe. (B) Binding of Vps45p and Sly1p to different syntaxins. GST-pulldown assays with T7-tagged Vps45p (upper panel) or T7-tagged Sly1p (lower panel) were performed as described in Materials and methods, except that resins were washed with binding buffer containing 400 mM NaCl. Gels were analyzed by immunoblotting with anti-T7·Tag monoclonal antibody. Numbers on the left show the positions of molecular mass markers.