Significance
Specialized sugar polymers (oligosaccharides) are necessary for life at the protein, organelle, cell, and organism levels. Processes for degrading oligosaccharides enhance their repertoire of functions, but a potential problem is short circuiting between degradation and synthesis. With the endoplasmic reticulum of mammalian cells, we show how lipid-linked oligosaccharides (LLOs) meant for attachment to proteins are segregated from an immunogenic process that involves their degradation to free oligosaccharides. Specifically, the STT3A isoform of oligosaccharyltransferase only transfers oligosaccharides, while the STT3B isoform can hydrolyze LLOs when not preoccupied with attachment of oligosaccharides to protein. This allows cells to perform both important processes in concert, without problems due to competition.
Keywords: dolichol, glycosylation, oligosaccharyltransferase, STT3A, STT3B
Abstract
Oligosaccharyltransferases (OSTs) N-glycosylate proteins by transferring oligosaccharides from lipid-linked oligosaccharides (LLOs) to asparaginyl residues of Asn-Xaa-Ser/Thr acceptor sequons. Mammals have OST isoforms with STT3A or STT3B catalytic subunits for cotranslational or posttranslational N-glycosylation, respectively. OSTs also hydrolyze LLOs, forming free oligosaccharides (fOSs). It has been unclear whether hydrolysis is due to one or both OSTs, segregated from N-glycosylation, and/or regulated. Transfer and hydrolysis were assayed in permeabilized HEK293 kidney and Huh7.5.1 liver cells lacking STT3A or STT3B. Transfer by both STT3A-OST and STT3B-OST with synthetic acceptors was robust. LLO hydrolysis by STT3B-OST was readily detected and surprisingly modulated: Without acceptors, STT3B-OST hydrolyzed Glc3Man9GlcNAc2-LLO but not Man9GlcNAc2-LLO, yet it hydrolyzed both LLOs with acceptors present. In contrast, LLO hydrolysis by STT3A-OST was negligible. STT3A-OST however may be regulatory, because it suppressed STT3B-OST–dependent fOSs. TREX1, a negative innate immunity factor that diminishes immunogenic fOSs derived from LLOs, acted through STT3B-OST as well. In summary, only STT3B-OST hydrolyzes LLOs, depending upon LLO quality and acceptor site occupancy. TREX1 and STT3A suppress STT3B-OST–dependent fOSs. Without strict kinetic limitations during posttranslational N-glycosylation, STT3B-OST can thus moonlight for LLO hydrolysis. In contrast, the STT3A-OST/translocon complex preserves LLOs for temporally fastidious cotranslational N-glycosylation.
Asparagine (N)-linked glycosylation occurs in all three domains of life—eukaryotes, bacteria, and archaea—with conservation of mechanism and topology. N-glycans serve multiple purposes ranging from the folding and function of individual glycoproteins, to cell–cell interactions and signaling (1). N-glycosylation begins with synthesis of an oligosaccharide on a polyisoprenyl phosphate carrier, i.e., a lipid-linked oligosaccharide (LLO), with the polyisoprene chain embedded in the membrane and the oligosaccharide unit displayed in an aqueous environment. In metazoans and many other eukaryotes (2) the LLO is glucose3mannose9N-acetylglucosamine2-P-P-dolicholC95 (G3M9-LLO) with the oligosaccharide facing the endoplasmic reticulum (ER) lumen. G3M9-LLO is the donor for N-glycosylation of asparaginyl residues in select Asn-Xaa-Ser/Thr/Cys sequons of polypeptides, also luminal, by membranous oligosaccharyltransferases (OSTs). These are single subunit enzymes in bacteria, archaea, and some single-cell eukaryotes, and multisubunit complexes in other single-cell eukaryotes such as Saccharomyces cerevisiae and all metazoans (3, 4).
Most sequons are N-glycosylated cotranslationally as they emerge from the translocon into the ER lumen, but a minority are N-glycosylated posttranslationally after synthesis of the polypeptide is complete. The latter include sequons within ∼50- to 55-aa residues of the carboxyl terminus (5), which are inside the translocon at the time of ribosomal release, and sequons unavailable for cotranslational modification due to nearby disulfide bonds (6). Mammalian OSTs evolved several features to manage the co/posttranslational workload (7): (i) Mammalian OST complexes have eight to nine subunits and exist as one of two major isoforms with either STT3A or STT3B catalytic subunits. (ii) STT3A-OST glycosylates cotranslationally, with DC2 and KCP2 subunits that direct STT3A-OST to the translocon (8–10). (iii) STT3B-OST glycosylates posttranslationally, is unassociated with the translocon, and has an oxidoreductase-type subunit—MagT1 or TUSC3 (6, 9, 10). (iv) These distinct roles are evident in two human Congenital Disorders of Glycosylation, STT3A-CDG and STT3B-CDG (11). (v) Both isoforms recognize G3M9-LLO, but STT3A-OST is less tolerant of the intermediate M9-LLO as a donor than is STT3B-OST (12). STT3A-OST and STT3B-OST are thus highly related, but function with important catalytic and temporal differences.
Unphosphorylated N-linked-type free oligosaccharides (fOSs) are released by hydrolysis of LLO pyrophosphate bonds under conditions that permit N-glycosylation by OSTs (13). In vitro LLO hydrolysis mirrors N-glycosylation regarding LLO structure and concentration dependence, and requirement for divalent metal ion. Thus, OST was suggested to cause the hydrolysis, but a separate LLO hydrolase was also possible (14). Our group reported “OST-like” LLO hydrolysis in the context of the phosphomannomutase 2 deficiency PMM2-CDG (15, 16), Herpes simplex virus-1 infection (17), and Three-Prime Repair Exonuclease 1 (TREX1) dysfunction (18). Importantly, the single OST isoform complex of S. cerevisiae hydrolyzes LLOs directly (19). LLO hydrolysis is not strictly eukaryotic: the single subunit PglB OST of Campylobacter jejuni hydrolyzes LLO in response to environmental stress (20, 21).
Eukaryotic cells generate N-linked-type fOSs by at least two processes (13). First, LLO hydrolysis by OST forms some fOSs, and is the major fOS source in mammals (22). After OST releases G3M9Gn2, the triglucosyl “cap” is removed by ER luminal glucosidases. M9Gn2 fOSs are then transported into the cytosol, where endoglycosidase (ENGase) (23, 24) removes one reducing terminal GlcNAc, and cytosolic α-mannosidase activity subsequently generates Man5GlcNAc1 with the same mannosyl configuration as the Man5-LLO biosynthetic intermediate. Finally, Man5GlcNAc1 is transported into lysosomes for further digestion (13, 25).
Besides LLO hydrolysis, fOSs are formed by ER-associated degradation (ERAD) after N-glycosylated misfolded proteins transit from the ER to the cytosol. This forms a minor portion of mammalian fOSs (22), but is the major source of S. cerevisiae fOSs (13, 19). ERAD of glycoproteins typically requires N-glycan trimming to expose an ER luminal α1,6-linked mannosyl residue (26), typically on an isomer of Man5Gn2, Man6Gn2, or Man7Gn2, which binds OS9/XTP-3-type lectin (27). After ER exit but preceding proteasomal degradation of the polypeptide, the N-glycan is released from the linking asparagine by cytosolic peptide-N-glycosidase (PNGase) (27, 28). Although the two fOS classes cross the ER membrane by different processes (13), cytosolic fOSs from either ERAD or LLO hydrolysis follow similar glycosidic routes, with ENGase and α-mannosidase digestion. Despite their similar terminal pathways and common LLO origins and OST requirements, the two fOS classes can be distinguished experimentally by their intermediate steps. ERAD fOSs are suppressed by inhibiting either trimming to Man5–7Gn2 isomers or PNGase (27), which should not affect the quantity of fOSs from LLO hydrolysis. Conversely, inhibiting ER luminal glycosidase digestion of LLO hydrolysis-derived fOSs should cause some to increase in size, and/or cause some formerly cytosolic fOSs to be retained in the ER lumen (13, 25).
We report that STT3B-OST, but not STT3A-OST, is a LLO hydrolase. LLO hydrolysis had an unexpected acceptor-dampened selectivity for G3M9-LLO over M9-LLO, supporting a gatekeeper role for external loop 5 (EL5) found in all OSTs (29, 30). STT3B-OST–dependent fOS pools were suppressed somewhat expectedly by TREX1, but surprisingly by STT3A. Mechanisms to protect co- and posttranslational N-glycosylation from competing LLO hydrolysis, and to regulate LLO-derived fOSs, are discussed.
Results
OST Transferase and LLO Hydrolysis Activities Are Balanced.
LLO hydrolysis was analyzed in monolayer cell cultures treated with the pore-forming toxin anthrolysin O (ALO) to gently and selectively permeabilize the plasma membrane, with minimal ER perturbation while allowing access to precursors (31). Such treatment preserves certain aspects of LLO biosynthesis that occur in live cells but are lost in microsomes: dependence upon MPDU1 for mannose-P-dolichol and glucose-P-dolichol–dependent glycosytransferases (32); restriction of the dolichol-P pool available for LLO synthesis (33); and limited access of UDP-GlcNAc:dolichol-P GlcNAc-1-P transferase to endogenous dolichol-P (34).
Permeabilized HEK293 kidney cells produced G3M9-LLO (SI Appendix, Fig. S1 A and B), and readily transferred G3M9Gn2 to acceptor peptides (APs) with functional sequons (acetyl-Asn-Tyr-Thr-CONH2), but not control peptides (CPs) with defective sequons (acetyl-Gln-Tyr-Thr-CONH2) (Fig. 1A). Aside from G3M9Gn2 peptide, three to four smaller glycopeptides were detected, likely from processing by glycosidase activity not fully blocked by added inhibitors deoxymannojirimycin (DMN) and castanospermine (CSN). Transferase products were diminished with an OST inhibitor (Fig. 1B), NGI-1 (35).
Free G3M9Gn2 was maximally generated with control peptide or no peptide, and suppressed by NGI-1 (Fig. 1C). Thus, OST activity hydrolyzed G3M9-LLO. Transfer and hydrolysis products plateaued by 30 min (SI Appendix, Fig. S1C). Acceptor peptide diminished free G3M9Gn2 revealing competition between transferase and hydrolysis reactions for LLO, and argued against fOS forming via ERAD-like PNGase with a peptidyl intermediate (36) since acceptor should have increased fOS, not decreased it. Transferase products (with acceptor peptide) exceeded hydrolysis products (control peptide) by ∼30-fold. STT3A-OST and STT3B-OST activities are therefore balanced. LLO hydrolysis is significant in the absence of acceptor, but transferase activity is strongly favored in the presence of acceptor and competitively suppresses hydrolysis.
STT3B-OST Hydrolyzes LLOs in HEK293 Cells.
Control HEK293 cells expressing both STT3A-OST and STT3B-OST, STT3A knockouts (STT3A-KO) expressing only STT3B-OST, and STT3B-knockouts (STT3B-KO) expressing only STT3A-OST (CRISPR/Cas9 technology, ref. 37) had similar LLO and total N-glycan pools (SI Appendix, Fig. S2). Each cell line also had significant transferase activity (Fig. 1 A and B). Surprisingly, LLO hydrolysis was measurable only in cells expressing STT3B-OST (Fig. 1 A and C). Such comparisons probably underestimate transferase activity because LLO concentrations declined as transferase reactions progressed (Fig. 2A). Most likely the LLO cycle and limited dolichol-P pool (33) kept pace with hydrolysis, but not with oligosaccharide transfer, even though nucleotide-sugar LLO precursors measured by fluorophore-assisted carbohydrate electrophoresis (FACE) (38, 39) were not significantly consumed. Suppression of hydrolysis by AP showed that the two STT3B-OST activities competed for the same LLO pool, and both were attenuated by NGI-1. Some inactive sequons can interact with OST (40), but this was inconsequential for hydrolysis because free G3M9Gn2 was formed equally with CP or no peptide.
LLO hydrolysis is thus a distinguishing feature of STT3B-OST, not a general OST side reaction. STT3A-OST, docked at the translocon, must cotranslationally N-glycosylate sequons during the brief moment they emerge from the translocation tunnel. LLO hydrolysis would sabotage this function. On the other hand, STT3B-OST acts posttranslationally without a fastidious kinetic requirement, and should tolerate “toggling” between transfer and hydrolysis.
Hydrolysis by STT3B-OST Is LLO Specific and Acceptor Dependent.
We anticipated that the LLO glycan specificity of STT3B-OST’s hydrolase activity would emulate its transferase activity, which used both G3M9-LLO and M9-LLO efficiently (12). This was tested by omitting UDP-glucose (UDP-Glc), the precursor of glucose-P-dolichol needed to convert M9-LLO to G3M9-LLO (Fig. 2 and SI Appendix, Fig. S1 A and B). LLOs synthesized with UDP-Glc were mostly G3M9-LLO, with small amounts of M9-LLO and other LLO intermediates. Without UDP-Glc, LLOs were primarily M9-LLO, but traces of G3M9-LLO were detected, perhaps from residual G3M9-LLO/glucose-P-dolichol present at the beginnings of reactions, or UDP-Glc impurities in the GDP-mannose or UDP-GlcNAc. STT3A-OST’s transferase activity (STT3B-KO cells) favored G3M9-LLO over M9-LLO at two donor ratios (depending on UDP-Glc addition). STT3B-OST, unlike STT3A-OST, used G3M9-LLO and M9-LLO as transferase donors equally. Thus, the STT3A-OST and STT3B-OST LLO specificities in HEK293 knockouts mirrored measurements with these OST complexes isolated biochemically (12).
With M9-LLO, like G3M9-LLO, hydrolysis by STT3A-OST was insignificant (SI Appendix, Fig. S1 A and B), but results for STT3B-OST were unanticipated. With CP or no peptide, STT3B-OST strongly preferred G3M9-LLO for hydrolysis over M9-LLO. Even when UDP-Glc was omitted, STT3B-OST generated mostly free G3M9Gn2, attributable to residual G3M9-LLO in the assay, and free M9Gn2 was essentially undetectable. In the presence of acceptor peptide, however, STT3B-OST became unselective for M9-LLO vs. G3M9-LLO as hydrolysis substrates, paralleling its transferase specificity. As expected addition of acceptor allowed transfer to outcompete hydrolysis. STT3B-OST’s unexpected acceptor-dependent LLO hydrolysis specificity implicates a role for EL5 (Discussion).
Hepatocyte STT3B-OST, Not STT3A-OST, Is an LLO Hydrolase.
We next tested the role of STT3A and STT3B in hepatocytes, which are major secretory cells producing most serum N-glycoproteins and are physiologically sensitive to OST dysfunction (11). OST subunit knockouts and rescue lines with hepatocyte-derived Huh7.5.1 cells (41) were thus used to examine LLO hydrolysis. Consistent with HEK293 cells, transferase activity was evident in all Huh7.5.1 STT3A/B knockouts and rescue lines, and LLO hydrolysis was insignificant in STT3B knockouts (Fig. 3). Importantly, LLO hydrolysis was restored by rescue of the STT3B knockout with normal (catalytically active) STT3B, but not with transferase-dead STT3B bearing point mutations for three evolutionarily conserved catalytic residues (41). STT3B-OST is therefore the sole LLO hydrolase in hepatocytes and kidney cells, with hydrolysis most likely at the catalytic site rather than a separate allosteric site (12, 42).
STT3A Diminishes STT3B-OST–Dependent fOSs.
Since STT3B-OST hydrolyzes LLOs, we expected fOSs to be depleted in STT3B-KO cells. Indeed, several fOSs were more abundant in STT3A-KO than in STT3B-KO cells (Fig. 4A and SI Appendix, Fig. S3A). However, control cell fOSs were similar to those in STT3B-KO cells, not STT3A-KO cells. Cellular fOSs therefore unexpectedly increased with the absence of STT3A, not the presence of STT3B. Acting cotranslationally, STT3A-OST should aid N-glycan–dependent folding. Glycoproteins normally bearing multiple N-glycans, but missing one if posttranslational glycosylation by STT3B-OST failed to fully compensate in STT3A-KO cells, could become candidates for ERAD. Subsequent cytosolic digestion by PNGase, ENGase, and mannosidases (27, 36) of additional ERAD substrates might elevate fOSs in STT3A-KO cells, supplementing the normally small contribution of ERAD to mammalian fOSs (22).
Elevation of STT3A-KO fOS levels was sensitive to NGI-1 and the LLO synthesis inhibitor tunicamycin (TN) (Fig. 4B and SI Appendix, Fig. S3B), demonstrating origination from LLOs. They varied in size corresponding to Glc3–8 standards. The migrations of the two most prominent fOSs (Fig. 4C and SI Appendix, Fig. S3; near Glc4 and Glc5) were consistent with M3Gn1–2–M5Gn1–2 structures which would form by partial trimming inside the ER lumen and further trimming by cytosolic/lysosomal glycosidases (13). Since STT3A-KO cells treated with ALO lost a subset of the Glc3–8 fOSs, they were a luminal/cytosolic mixture (Fig. 4C). Thus, the fOSs increased in STT3A-KO cells could be due to either LLO hydrolysis or PNGase/ENGase digestion of ERAD substrates.
PNGase is inhibited by z-VAD-FMK, a caspase inhibitor (43, 44), and in preliminary experiments it suppressed STT3A-KO fOSs. However, a control caspase inhibitor (Q-VD-Oph) which does not affect PNGase (43, 45) gave similar inhibition. This prompted an alternative approach for differentiating between LLOs and ERAD as the fOS source. Glycoprotein ERAD requires processing of N-linked oligosaccharides by kifunensine (KF)-sensitive ER mannosidase(s) (46). This forms structures with exposed α1,6-linked mannosyl residues in the ER lumen for OS9/XTP3-B binding (26). Retrotranslocation of such ERAD substrates and PNGase digestion releases ERAD-derived fOSs into the cytosol. KF should therefore attenuate cytosolic ERAD-associated fOSs (22). By comparison, cytosolic fOSs from LLO hydrolysis require removal of glucosyl residues by CSN-sensitive glucosidases I and II to exit the ER (13). As indicated previously (22) a mixture of KF and CSN should thus diminish LLO-derived cytosolic fOSs originating by either hydrolysis or N-glycosylation/ERAD. However, new and larger fOSs are expected only for LLO hydrolysis, with some being luminal. Inhibition of ERAD should simply eliminate PNGase-dependent fOSs with no new species (SI Appendix, Fig. S3C).
Fig. 4C displays fOSs after mixed KF/CSN treatment of both control and STT3A-KO cells, allowing STT3A-KO specific fOSs to be deduced by comparison. KF/CSN increased total fOSs, diminished two cytosolic fOSs (nos. 4 and 5), and caused the appearance of one cytosolic (no. 8) and one luminal (no. 6b) fOS. We conclude that the fOSs increased in STT3A-KO cells result from STT3B-OST–dependent LLO hydrolysis, not ERAD. STT3A thus unexpectedly suppresses this fOS pool (Discussion).
TREX1 Suppresses fOSs Generated by STT3B-OST.
TREX1 (DNase III) suppresses inappropriate activation of cGAS-STING innate immunity signaling (47). We reported a second TREX1 function requiring its carboxyl-terminal tail anchor, which localizes TREX1 to the cytosolic face of the ER and may explain TREX1 diseases with mutations that retain DNase activity (18). TREX1 knockout in mouse embryonic fibroblasts (MEFs) increases LLO hydrolase activity and LLO-derived fOSs, which interestingly activate expression of IFN-stimulated genes (18).
To test potential connections between TREX1 and STT3B-OST–specific LLO hydrolysis, we took advantage of fortuitously insignificant TREX1 levels in HEK293 cells (18). After stably expressing human TREX1 in control HEK293, STT3A-KO, and STT3B-KO cells, fOSs were analyzed. High cell density attenuated TREX1 effects on fOSs (SI Appendix, Fig. S4A), but at low cell density, control and STT3A-KO cells revealed suppression of STT3B-OST–dependent fOSs by TREX1 (Fig. 5). STT3B-KO fOSs were unaffected. The two STT3A-KO fOSs diminished most by TREX1 (Fig. 5A, arrows) may be part of the STT3A-suppressible fOS pool (Fig. 4C and SI Appendix, Fig. S3A), and the HEK293 cell density effect is possibly related to mitosis-dependent phosphorylation of TREX1’s C terminus (48). The fOS results with HEK293 cells were next expanded with STT3A and STT3B knockdowns in TREX1-KO MEFs. Target mRNAs were lowered by >90%, but only knockdown of STT3B suppressed fOSs in TREX1-KO MEFs, reaching fOS levels in WT MEFs (SI Appendix, Fig. S4B). In both HEK293 and MEF cells TREX1, a negative modulator of both innate immunity and LLO-derived fOSs, was thus linked to STT3B-OST.
Discussion
Spiro and Anumula suggested that mammalian OST was also a LLO hydrolase (14). Szymanski and coworkers, working with C. jejuni which has a single subunit OST (19), and then Suzuki’s group, working with S. cerevisiae which has a single OST complex (20), both demonstrated directly that those OSTs catalyzed LLO hydrolysis. This left little doubt that mammalian OST also hydrolyzed LLO, but important questions remained: whether the STT3A and/or STT3B-OST isoforms hydrolyzed LLOs; whether N-glycosylation was shielded from competing LLO hydrolysis; and whether LLO hydrolysis was biologically useful or regulatable.
Multiple Factors Shield N-Linked Glycosylation from LLO Hydrolysis.
In both kidney and liver cells LLOs were significantly hydrolyzed by STT3B-OST, not STT3A-OST, and thus hydrolysis is not a simple catalytic error generally affecting OSTs. Possibly STT3B-OST allows water into the catalytic site while STT3A-OST excludes it. Although the catalytic mechanism proposed for transfer of glycan by OST (30, 49) may not be easily reconciled with hydrolysis, the high concentration of water in the aqueous ER lumen could drive LLO hydrolysis by STT3B-OST. Alternatively, a latent LLO hydrolysis activity by STT3A-OST might be disabled upon docking to the translocon (9) perhaps via allosteric interaction or creation of a water-tight seal. Our data do not distinguish these models. Therefore, it will be interesting to see whether LLO hydrolysis is altered by losses of the STT3A-OST–specific subunits DC2 and KPC2, which are involved in translocon docking (8, 9), as well as by the STT3B-OST–specific oxidoreductase subunits MagT1 and TUSC3 needed for cellular STT3B-OST function with nascent polypeptides but not catalysis with simple acceptors by the isolated complex (6, 12).
STT3A-OST must cotranslationally glycosylate acceptor sequons as they emerge from the translocation tunnel near STT3A-OST’s catalytic site, and therefore STT3A-OST appears to reserve LLOs for N-glycosylation rather than hydrolyze them. In contrast, STT3B-OST glycosylates sequons posttranslationally, and should accommodate modest delays due to lack of bound, intact LLO. STT3B-OST thus appears amenable to LLO hydrolysis “moonlighting.” Transferase competition in the presence of acceptor also weakened hydrolysis by STT3B-OST. These properties of mammalian OSTs should shield both cotranslational and posttranslational N-glycosylation from LLO hydrolysis.
Regulatory Potential of STT3A and TREX1.
STT3A-OST does not hydrolyze LLOs, and in HEK293 cells STT3A knockout surprisingly elevated the TN- and NGI-1–sensitive fOSs linked to STT3B-OST and independent of ERAD. Such indirect modulation of STT3B-OST function by STT3A suggests that STT3A-OST may N-glycosylate a factor needed for fOS metabolism or transport. STT3A’s impact on HEK293 fOSs was also emulated by TREX1: like STT3A knockout, loss of TREX1 increased STT3B-OST–dependent fOSs. In MEFs, TREX1 knockout increases fOS levels and LLO hydrolysis (18). We reconfirmed here (SI Appendix, Fig. S5) the increased LLO hydrolysis activity in TREX1-KO MEFs (18), but we did not detect altered LLO hydrolysis in control or STT3A-KO HEK293 cells transfected with TREX1, even though fOSs were diminished (Fig. 5). Consequently, TREX1 may have two modes for modulating fOSs, only one regulating LLO hydrolysis, but both impacting TREX1’s suppression of IFN-stimulated genes (18).
External Loop 5 and LLO Selectivity.
LLO hydrolysis specificity for STT3B-OST was surprisingly acceptor dependent (Fig. 2). Campylobacter lari PglB protein (a single subunit OST) has a disordered loop, designated EL5, involved in binding of both LLO donors and peptide acceptors (29, 30, 50). EL5 is near a pocket proposed to accommodate the glycan of LLO and becomes ordered in its amino or carboxyl portions when LLO or acceptor peptides bind, respectively. The two conformational changes are modular and interactive. EL5 is most compact with both substrates bound, and suited to restrict access/egress of LLO to/from the pocket. In a proof-of-principle experiment with PglB, a lipid-linked monosaccharide maneuvered past/under EL5 more readily than a larger oligosaccharide (30). High-resolution structures of the S. cerevisiae and mammalian OST complexes have not probed EL5 function further (8, 49, 51). It thus remains unclear how the respective EL5s might explain STT3A-OST’s transferase preference for G3M9-LLO over M9-LLO, and/or the absence of a transferase preference by STT3B-OST (12).
In acceptor-free conditions, we found that STT3B-OST was surprisingly selective during LLO hydrolysis, resembling the transferase activity of STT3A-OST rather than STT3B-OST; G3M9-LLO was hydrolyzed but not M9-LLO. However, with acceptor peptide present, LLO hydrolysis by STT3B-OST mirrored its own transferase activity to utilize both G3M9-LLO and M9-LLO. In this way, STT3B-OST emulated STT3B-like PglB of C. jejuni, which hydrolyzes LLO in response to reduced osmotic stress, and seems more selective for full-length LLO as a hydrolase than as a transferase (20).
Extrapolating from the properties of C. lari PglB (29, 30, 50), we envision both the polyisoprenoid and oligosaccharide portions of LLO interacting with STT3B-OST while EL5 is still fully disordered. This would permit the oligosaccharide (independent of size) full access to its pocket. LLO-induced amino-terminal modular compaction of EL5 might then selectively hinder egress of oligosaccharide from the pocket. The larger dolichol-linked G3M9Gn2 would be trapped and subject to hydrolysis, while the smaller dolichol-linked M9Gn2 might pass through/under the partially compacted EL5 to dissociate from STT3B-OST without hydrolysis—a “catch-and-release” mechanism. Upon binding of peptide in addition to LLO, full ordering of EL5 in both its amino and carboxyl portions would form a more restrictive compact loop. M9-LLO as well as G3M9-LLO would then be trapped and available for both hydrolysis and transfer by STT3B-OST, as we observed (Fig. 2).
Summary.
STT3B-OST, but not STT3A-OST, hydrolyzes LLO to generate fOSs which can be suppressed by STT3A and TREX1. Additional features of STT3B-OST allow toggling between transferase and hydrolysis modes. N-glycosylation, especially at the translocon, is thus protected from competing LLO hydrolysis which generates regulatable fOSs (SI Appendix, Fig. S6).
Materials and Methods
Cell Culture.
Human HEK293-WT, STT3A-KO, and STT3B-KO kidney lines (37) and WT TREX1-KO mouse fibroblasts (18) were grown in DMEM (low glucose; Thermo Fisher no. 11885092) supplemented with 10% FBS (Atlanta Biological). Human Huh7.5.1 WT, STT3A-KO, STT3B-KO, and rescue hepatocyte lines (41) were grown in DMEM (high glucose; Thermo Fisher no. 11995065) with 10% heat-inactivated FBS, 1× l-glutamine and 1× nonessential amino acids. Cells were cultured in a humidified 5% CO2 atmosphere, then harvested directly with methanol for FACE analysis of saccharides (39) or permeabilized (32).
FACE.
Saccharide classes including LLOs, N-glycans, fOSs, and monosaccharides were isolated from methanolic sonicates of live or permeabilized cells and analyzed by FACE (38, 39). LLO glycans were released from dolichol with mild acid, and N-glycans were released with endo H (New England Biolabs no. P0703L) for glycopeptides or PNGase F (Prozyme no. GKE-5003) for glycoproteins. Glycans were labeled with 7-amino-1,3-naphthalenedisulfonic acid (ANDS) (AnaSpec no. 81529), monosaccharides were labeled with 2-aminoacridone (AMAC) (Sigma no. 06627), and samples were normalized to total cellular protein for FACE analysis. Gel images were acquired with a UVP Chemidoc-It II scanner, and saccharides were quantified with VisionWorks software.
Permeabilization of Plasma Membranes with Pore-Forming Toxin.
Cells in 10-cm dishes were refed within 24 h before ALO treatment to achieve 70–90% confluence, and permeabilized with ALO as described for streptolysin O (32). Dishes on ice were rinsed twice with ice-cold PBS, 20 nM ALO (4 mL, ice-cold in PBS) was added for 4 min with periodic swirling on ice, and dishes were rinsed again with ice-cold PBS. Dishes were then incubated for 4 min at 37 °C with prewarmed transport buffer (TB) (78 mM KCl, 4 mM MgCl2, and 50 mM K-Hepes, pH 7.2), then placed directly on ice for 10 min. If desired the conditioned TB was collected to obtain diffusible cytosolic contents, while the cell bodies were harvested for luminal contents or used for OST reactions.
OST Reactions.
ALO-permeabilized cells were incubated for 30–60 min in TB containing 0 or 400 μM UDP-Glc, 200 μM UDP-GlcNAc, 50 μM GDP-Man, 0.2 mM AMP, 10 mg/mL CSN, and 10 mg/mL DMN, with 50 μM CP/AP and 5 μM NGI-1 as indicated. After discarding the reaction buffer, cell bodies were scraped into methanol with sonication, followed by isolation of glycopeptides, fOSs, and LLOs all from the same dishes for direct comparison.
Statistics.
Graphpad Prism 7 was used to calculate mean ± SD and perform statistical analyses. Where two groups were compared, a two-tailed t test was used. For multiple groups, ordinary one-way ANOVA (no matching or pairing) was performed using Tukey’s multiple comparisons test. For both the ANOVA and the t test, P < 0.05 was considered significant.
Supplementary Material
Acknowledgments
We thank Dr. Jan Carette (Stanford University) for Huh7.5.1 WT/KO lines (41); Dr. Joseph Contessa (Yale University) for NGI-1 (35); Dr. Arun Radhakrishnan (UT Southwestern) for critical guidance and materials needed to prepare ALO; and the National Institutes of Health for Grants R01-GM038545 (to M.A.L.), R01-AR067135 (to N.Y.), and R01-GM043768 (to R.G.).
Footnotes
The authors declare no conflict of interest.
This article is a PNAS Direct Submission.
This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10.1073/pnas.1806034115/-/DCSupplemental.
References
- 1.Varki A. Biological roles of glycans. Glycobiology. 2017;27:3–49. doi: 10.1093/glycob/cww086. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Samuelson J, et al. The diversity of dolichol-linked precursors to Asn-linked glycans likely results from secondary loss of sets of glycosyltransferases. Proc Natl Acad Sci USA. 2005;102:1548–1553. doi: 10.1073/pnas.0409460102. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Helenius A, Aebi M. Roles of N-linked glycans in the endoplasmic reticulum. Annu Rev Biochem. 2004;73:1019–1049. doi: 10.1146/annurev.biochem.73.011303.073752. [DOI] [PubMed] [Google Scholar]
- 4.Kelleher DJ, Gilmore R. An evolving view of the eukaryotic oligosaccharyltransferase. Glycobiology. 2006;16:47R–62R. doi: 10.1093/glycob/cwj066. [DOI] [PubMed] [Google Scholar]
- 5.Shrimal S, Trueman SF, Gilmore R. Extreme C-terminal sites are posttranslocationally glycosylated by the STT3B isoform of the OST. J Cell Biol. 2013;201:81–95. doi: 10.1083/jcb.201301031. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Cherepanova NA, Shrimal S, Gilmore R. Oxidoreductase activity is necessary for N-glycosylation of cysteine-proximal acceptor sites in glycoproteins. J Cell Biol. 2014;206:525–539. doi: 10.1083/jcb.201404083. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Cherepanova N, Shrimal S, Gilmore R. N-linked glycosylation and homeostasis of the endoplasmic reticulum. Curr Opin Cell Biol. 2016;41:57–65. doi: 10.1016/j.ceb.2016.03.021. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Braunger K, et al. Structural basis for coupling protein transport and N-glycosylation at the mammalian endoplasmic reticulum. Science. 2018;360:215–219. doi: 10.1126/science.aar7899. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Shrimal S, Cherepanova NA, Gilmore R. DC2 and KCP2 mediate the interaction between the oligosaccharyltransferase and the ER translocon. J Cell Biol. 2017;216:3625–3638. doi: 10.1083/jcb.201702159. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Ruiz-Canada C, Kelleher DJ, Gilmore R. Cotranslational and posttranslational N-glycosylation of polypeptides by distinct mammalian OST isoforms. Cell. 2009;136:272–283. doi: 10.1016/j.cell.2008.11.047. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Shrimal S, Ng BG, Losfeld M-E, Gilmore R, Freeze HH. Mutations in STT3A and STT3B cause two congenital disorders of glycosylation. Hum Mol Genet. 2013;22:4638–4645. doi: 10.1093/hmg/ddt312. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Kelleher DJ, Karaoglu D, Mandon EC, Gilmore R. Oligosaccharyltransferase isoforms that contain different catalytic STT3 subunits have distinct enzymatic properties. Mol Cell. 2003;12:101–111. doi: 10.1016/s1097-2765(03)00243-0. [DOI] [PubMed] [Google Scholar]
- 13.Chantret I, Moore SEH. Free oligosaccharide regulation during mammalian protein N-glycosylation. Glycobiology. 2008;18:210–224. doi: 10.1093/glycob/cwn003. [DOI] [PubMed] [Google Scholar]
- 14.Anumula KR, Spiro RG. Release of glucose-containing polymannose oligosaccharides during glycoprotein biosynthesis. Studies with thyroid microsomal enzymes and slices. J Biol Chem. 1983;258:15274–15282. [PubMed] [Google Scholar]
- 15.Gao N, Shang J, Lehrman MA. Analysis of glycosylation in CDG-Ia fibroblasts by fluorophore-assisted carbohydrate electrophoresis: Implications for extracellular glucose and intracellular mannose 6-phosphate. J Biol Chem. 2005;280:17901–17909. doi: 10.1074/jbc.M500510200. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Cline A, et al. A zebrafish model of PMM2-CDG reveals altered neurogenesis and a substrate-accumulation mechanism for N-linked glycosylation deficiency. Mol Biol Cell. 2012;23:4175–4187. doi: 10.1091/mbc.E12-05-0411. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Gao N, et al. Mannose-6-phosphate regulates destruction of lipid-linked oligosaccharides. Mol Biol Cell. 2011;22:2994–3009. doi: 10.1091/mbc.E11-04-0286. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Hasan M, et al. Cytosolic nuclease TREX1 regulates oligosaccharyltransferase activity independent of nuclease activity to suppress immune activation. Immunity. 2015;43:463–474. doi: 10.1016/j.immuni.2015.07.022. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Harada Y, et al. Eukaryotic oligosaccharyltransferase generates free oligosaccharides during N-glycosylation. J Biol Chem. 2013;288:32673–32684. doi: 10.1074/jbc.M113.486985. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Nothaft H, Liu X, McNally DJ, Li J, Szymanski CM. Study of free oligosaccharides derived from the bacterial N-glycosylation pathway. Proc Natl Acad Sci USA. 2009;106:15019–15024. doi: 10.1073/pnas.0903078106. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Barre Y, et al. A conserved DGGK motif is essential for the function of the PglB oligosaccharyltransferase from Campylobacter jejuni. Glycobiology. 2017;27:978–989. doi: 10.1093/glycob/cwx067. [DOI] [PubMed] [Google Scholar]
- 22.Harada Y, Masahara-Negishi Y, Suzuki T. Cytosolic-free oligosaccharides are predominantly generated by the degradation of dolichol-linked oligosaccharides in mammalian cells. Glycobiology. 2015;25:1196–1205. doi: 10.1093/glycob/cwv055. [DOI] [PubMed] [Google Scholar]
- 23.Suzuki T, et al. Endo-β-N-acetylglucosaminidase, an enzyme involved in processing of free oligosaccharides in the cytosol. Proc Natl Acad Sci USA. 2002;99:9691–9696. doi: 10.1073/pnas.152333599. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Huang C, et al. Endo-β-N-acetylglucosaminidase forms N-GlcNAc protein aggregates during ER-associated degradation in Ngly1-defective cells. Proc Natl Acad Sci USA. 2015;112:1398–1403. doi: 10.1073/pnas.1414593112. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Suzuki T, Funakoshi Y. Free N-linked oligosaccharide chains: Formation and degradation. Glycoconj J. 2006;23:291–302. doi: 10.1007/s10719-006-6975-x. [DOI] [PubMed] [Google Scholar]
- 26.Quan EM, et al. Defining the glycan destruction signal for endoplasmic reticulum-associated degradation. Mol Cell. 2008;32:870–877. doi: 10.1016/j.molcel.2008.11.017. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Hebert DN, Molinari M. Flagging and docking: Dual roles for N-glycans in protein quality control and cellular proteostasis. Trends Biochem Sci. 2012;37:404–410. doi: 10.1016/j.tibs.2012.07.005. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Suzuki T, Kitajima K, Emori Y, Inoue Y, Inoue S. Site-specific de-N-glycosylation of diglycosylated ovalbumin in hen oviduct by endogenous peptide: N-glycanase as a quality control system for newly synthesized proteins. Proc Natl Acad Sci USA. 1997;94:6244–6249. doi: 10.1073/pnas.94.12.6244. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Lizak C, Gerber S, Numao S, Aebi M, Locher KP. X-ray structure of a bacterial oligosaccharyltransferase. Nature. 2011;474:350–355. doi: 10.1038/nature10151. [DOI] [PubMed] [Google Scholar]
- 30.Napiórkowska M, et al. Molecular basis of lipid-linked oligosaccharide recognition and processing by bacterial oligosaccharyltransferase. Nat Struct Mol Biol. 2017;24:1100–1106. doi: 10.1038/nsmb.3491. [DOI] [PubMed] [Google Scholar]
- 31.Martys JL, Shevell T, McGraw TE. Studies of transferrin recycling reconstituted in streptolysin O permeabilized Chinese hamster ovary cells. J Biol Chem. 1995;270:25976–25984. doi: 10.1074/jbc.270.43.25976. [DOI] [PubMed] [Google Scholar]
- 32.Anand M, et al. Requirement of the Lec35 gene for all known classes of monosaccharide-P-dolichol-dependent glycosyltransferase reactions in mammals. Mol Biol Cell. 2001;12:487–501. doi: 10.1091/mbc.12.2.487. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Gao N, Lehrman MA. Coupling of the dolichol-P-P-oligosaccharide pathway to translation by perturbation-sensitive regulation of the initiating enzyme, GlcNAc-1-P transferase. J Biol Chem. 2002;277:39425–39435. doi: 10.1074/jbc.M205195200. [DOI] [PubMed] [Google Scholar]
- 34.Gao N, Shang J, Lehrman MA. Unexpected basis for impaired Glc3Man9GlcNAc2-P-P-dolichol biosynthesis by elevated expression of GlcNAc-1-P transferase. Glycobiology. 2008;18:125–134. doi: 10.1093/glycob/cwm109. [DOI] [PubMed] [Google Scholar]
- 35.Lopez-Sambrooks C, et al. Oligosaccharyltransferase inhibition induces senescence in RTK-driven tumor cells. Nat Chem Biol. 2016;12:1023–1030. doi: 10.1038/nchembio.2194. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36.Harada Y, Hirayama H, Suzuki T. Generation and degradation of free asparagine-linked glycans. Cell Mol Life Sci. 2015;72:2509–2533. doi: 10.1007/s00018-015-1881-7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37.Cherepanova NA, Gilmore R. Mammalian cells lacking either the cotranslational or posttranslocational oligosaccharyltransferase complex display substrate-dependent defects in asparagine linked glycosylation. Sci Rep. 2016;6:20946. doi: 10.1038/srep20946. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38.Gao N, Lehrman MA. Non-radioactive analysis of lipid-linked oligosaccharide compositions by fluorophore-assisted carbohydrate electrophoresis. Methods Enzymol. 2006;415:3–20. doi: 10.1016/S0076-6879(06)15001-6. [DOI] [PubMed] [Google Scholar]
- 39.Gao N, Holmes J, Lehrman MA. Letter to the glycoforum: Improved protocols for preparing lipid-linked and related saccharides for fluorophore-assisted carbohydrate electrophoresis (FACE) Glycobiology. 2013;23:1111. doi: 10.1093/glycob/cwt067. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 40.Nilsson I, et al. Photocross-linking of nascent chains to the STT3 subunit of the oligosaccharyltransferase complex. J Cell Biol. 2003;161:715–725. doi: 10.1083/jcb.200301043. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 41.Marceau CD, et al. Genetic dissection of Flaviviridae host factors through genome-scale CRISPR screens. Nature. 2016;535:159–163. doi: 10.1038/nature18631. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 42.Karaoglu D, Kelleher DJ, Gilmore R. Allosteric regulation provides a molecular mechanism for preferential utilization of the fully assembled dolichol-linked oligosaccharide by the yeast oligosaccharyltransferase. Biochemistry. 2001;40:12193–12206. doi: 10.1021/bi0111911. [DOI] [PubMed] [Google Scholar]
- 43.Misaghi S, Pacold ME, Blom D, Ploegh HL, Korbel GA. Using a small molecule inhibitor of peptide: N-glycanase to probe its role in glycoprotein turnover. Chem Biol. 2004;11:1677–1687. doi: 10.1016/j.chembiol.2004.11.010. [DOI] [PubMed] [Google Scholar]
- 44.Misaghi S, Korbel GA, Kessler B, Spooner E, Ploegh HL. z-VAD-fmk inhibits peptide:N-glycanase and may result in ER stress. Cell Death Differ. 2006;13:163–165. doi: 10.1038/sj.cdd.4401716. [DOI] [PubMed] [Google Scholar]
- 45.Caserta TM, Smith AN, Gultice AD, Reedy MA, Brown TL. Q-VD-OPh, a broad spectrum caspase inhibitor with potent antiapoptotic properties. Apoptosis. 2003;8:345–352. doi: 10.1023/a:1024116916932. [DOI] [PubMed] [Google Scholar]
- 46.Liu Y, Choudhury P, Cabral CM, Sifers RN. Oligosaccharide modification in the early secretory pathway directs the selection of a misfolded glycoprotein for degradation by the proteasome. J Biol Chem. 1999;274:5861–5867. doi: 10.1074/jbc.274.9.5861. [DOI] [PubMed] [Google Scholar]
- 47.Hasan M, et al. Trex1 regulates lysosomal biogenesis and interferon-independent activation of antiviral genes. Nat Immunol. 2013;14:61–71. doi: 10.1038/ni.2475. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 48.Kucej M, Fermaintt CS, Yang K, Irizarry-Caro RA, Yan N. Mitotic phosphorylation of TREX1 C terminus disrupts TREX1 regulation of the oligosaccharyltransferase complex. Cell Rep. 2017;18:2600–2607. doi: 10.1016/j.celrep.2017.02.051. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 49.Wild R, et al. Structure of the yeast oligosaccharyltransferase complex gives insight into eukaryotic N-glycosylation. Science. 2018;359:545–550. doi: 10.1126/science.aar5140. [DOI] [PubMed] [Google Scholar]
- 50.Shrimal S, Cherepanova NA, Gilmore R. One flexible loop in OST lassos both substrates. Nat Struct Mol Biol. 2017;24:1009–1010. doi: 10.1038/nsmb.3508. [DOI] [PubMed] [Google Scholar]
- 51.Bai L, Wang T, Zhao G, Kovach A, Li H. The atomic structure of a eukaryotic oligosaccharyltransferase complex. Nature. 2018;555:328–333. doi: 10.1038/nature25755. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.