Digitonin-high salt extracts prepared from cells 72 hr after siRNA transfection were processed for immunoblots using antisera for STT3A, STT3B, ribophorin I (R I), BiP, Sec61α, and the α-subunit of the F₁F₀-ATPase. Asterisks designate nonspecific bands. Similar results were obtained using a second set of each of the siRNAs. The negative control siRNA is abbreviated NC.

Cotranslational glycosylation of nascent glycoproteins in mammalian cells (A and B). Skipped sites can be cotranslationally glycosylated by STT3B (B). The signal sequence (wavy line) remains bound to the translocon until cleaved (C). Cotranslational glycosylation of an extreme N-terminal glycosylation site by the STT3B isoform (D). Reduced efficiency of cotranslational glycosylation in STT3A-depleted cells (E). Posttranslational glycosylation of a C-terminal sequon by the STT3B isoform (F) is limited by protein folding (G).

(A) Digitonin-high salt extracts analyzed in Figure 1 were assayed for OST activity. The values shown are averages of duplicate assays normalized to DNA content and expressed relative to negative control siRNA-treated cells.
(B) Native immunoprecipitation using anti-STT3B. Aliquots of the detergent extract (T), immunoprecipitate (P) and supernatant (S) fractions were analyzed on immunoblots. STT3B, R I, glyceraldehyde 3-phosphate dehydrogenase (GAPDH) and IgG heavy chain (HC) are labeled.
(C) The relative amount of OST complexes in siRNA-treated cells was estimated using protein immunoblot data (Figure 1) and the STT3A:STT3B ratio (60:40) determined in panel (B).

(A) Diagram of the FVII precursor showing the signal sequence (black), the gamma-carboxyglutamate domain (green), two EGF-like domains (cyan), the activation peptide (purple) and the catalytic domain (red). Glycosylation sites and disulfide bonds are shown. Cells cotransfected with wild-type (B–F), N360Q (G), or N183Q (H) FVII plasmids and negative control (B, D, G, and H), STT3A (C, E, G, and H), or STT3B (F–H) siRNAs were pulse labeled for 45 s (B and C) or 2 min (D–H) and chased as indicated. One pulse labeled sample was digested with Endo H (EH). Full-length FVII polypeptides that contain 0, 1, or 2 N-linked oligosaccharides were resolved from glycosylated nascent chains (g-nc) and nonglycosylated and deglycosylated nascent chains (nc). (D–F) The percentages of FVII molecules that have 0 (triangles), 1 (circles) or 2 (squares) N-linked oligosaccharides are shown.

(A) Diagram of the preprocathepsin C showing the signal sequence (black), retained propeptide (cyan), cleaved propeptide, (white), heavy chain (magenta) and light chain (yellow). Glycosylation sites, disulfide bonds (red lines) and free cysteines (red dots) are shown.
(B–D) Cells treated with the indicated siRNAs prior to pulse-chase labeling were used for pCatC immunoprecipitation. Results obtained with two siRNAs for ribophorin I are shown.
(E) Diagrams of pCatC-HA glycosylation site mutants. Cells cotransfected with pCatC-HA constructs and negative control (−) or STT3B (+) siRNAs were pulse labeled for 3 min, and immunoprecipitated using anti-HA. (B, D, and E) As indicated, samples from control cells were digested with Endo H (EH). Procathepsin C with 0-4 N-linked glycans were resolved by SDS-PAGE and the average number of glycans per pCatC chain was calculated.

(A–C) Control cells were pulsed for 5 min and chased as indicated. Cell extracts were treated with PEG-maleimide (PEG-Mal) as indicated prior to immnuoprecipitation of pCatC (A), pSAP (B), or FVII (C). Proteins are labeled as in Figure 3–5. Brackets labeled 1–4 designate the migration position of proteins that have 1–4 PEG-Mal adducts. Larger adducts (data not shown) migrated as a smear. The percentage of PEG-Mal resistant pCatC, FVII-1, FVII-2, and pSAP-65 relative to the untreated sample is shown below each panel. (D) The saposin domain is stabilized by three disulfide bonds (in red) and a hydrophobic core (A, I, L, V, and M residues are gray). The distance between the carboxamide oxygen of N80 and the side chain of T82 is 10.9 Å. N101 and T103 are on opposite faces of the third α-helix. The figure was created with MacPyMOL software using the saposin A structure (PDB 2DOB).

(A, C, and D) HeLa cells treated with the indicated siRNAs were pulse-labeled for 2 min (D) or 5 min (A and C) and chased as indicated (C and D).
(A) One plate of control cells was treated with tunicamycin (Tm) prior to labeling. Equal amounts of ³⁵S-labeled protein were incubated with Con A beads. Glycoproteins were eluted and resolved by SDS-PAGE. Circles designate glycoproteins that decrease in intensity in STT3A-depleted cells. The triangle designates hypoglycosylated variants of p65.
(B) Diagram of preprosaposin showing the signal sequence (black), the homologous N- and C-terminal domains (NT and CT; cyan), the saposin domains (yellow), and spacer segments (green). Glycosylation sites (e.g., 80 NAT) disulfide bonds (red lines) and lysosomal processing sites (arrowheads) are shown. The disulfide-bonding pattern of the invariant cysteine residues in the NT and CT domains is not known.
(C) Pulse-chase labeling of prosaposin in control cells. Intracellular (cells) and secreted (media) prosaposin was recovered by immunoprecipitation. The ER (pSAP-65) and Golgi (pSAP-75) forms of prosaposin are labeled. The arrowhead designates a nonspecific band. (D) Pulse-chase labeling of prosaposin. Glycoforms of pSAP-65 are labeled; fully glycosylated pSAP-75 is indicated by the bracket. The average number of glycans per pSAP-65 was calculated after quantification of the pSAP-65 glycoforms.