The plasmids used in this study and their important features are described in Table 1. Standard molecular biological techniques were used for plasmid construction, and mutations that altered various amino acids (see Table 2) were introduced using overlapping PCR with mutagenic primers (BGI, Shenzhen, China) and verified by DNA sequence analysis.

Human embryonic kidney cells, HEK293, were cultured in DMEM supplemented with 10% FBS (FCS, Biological Industries, Beit Haemek, Israel), penicillin and streptomycin (1 μg/ml) in a humidified 5% CO₂ atmosphere at 37°C.

E. coli Rosetta cells (DE3, Merck, NJ, United States) harboring both pET26b-pelB-FLAG-ALG13-iso2 and pCDFDuet-6His-ALG14 plasmids were cultured in 200 ml of Terrific-Broth medium (TB, 1.2% tryptone, 2.4% yeast extract and 0.5% glycerol) at 37°C to an OD600 of 1.0 and then cooled to 16°C. After adding 0.1 mM isopropyl-β-d-thiogalactopyranoside (IPTG, Sangon Biotech, Shanghai, China), the cells were incubated for additional 18 h–24 h at 16°C to induce protein expression. The harvested cells were resuspended in 15 ml lysis buffer (150 mM NaCl, 50 mM Tris/HCl, pH 8.0) and homogenized by sonication. The cell lysate was centrifuged at 4,000×g to remove debris, and the supernatant was collected (this supernatant was used as the “crude extract”). To obtain the membrane fraction containing ALG13/14 complex, the crude extract was further centrifuged at 12,000×g to collect precipitated membranes. Any remaining cytosolic ALG13 was removed by washing the membranes with 15 ml lysis buffer three times. This washed membrane fraction was then solubilized by resuspension in 5 ml lysis buffer containing 1% Triton X-100 and incubated on ice for 15 min to obtain the detergent extract. This detergent extract was used for purification of the recombinant ALG13/14 complex following the procedure described below. This detergent extract was also used for the co-immunoprecipitation experiment presented in Figure 2A.

The detergent extract prepared from the E. coli cells co-expressing FLAG-ALG13-iso2 and His-ALG14 was used to purify the recombinant ALG13/14 complex. First, His-tagged ALG14 proteins, including the recombinant FLAG-ALG13-iso2/His-ALG14 complex in detergent extract, were trapped by a HisTrap HP affinity column (GE Healthcare Life Sciences). Their eluted solution was then incubated with anti-FLAG Affinity Matrix (Sigma–Aldrich, MO, United States) to enrich the ALG13/14 complex and remove the free His-ALG14. After elution with FLAG peptide, the purified ALG13/14 complex was analyzed by native PAGE and SDS–PAGE followed by silver staining to confirm the purity. Native PAGE was performed using a Blue/Clear Native PAGE Electrophoresis Kit (Real-time, Beijing, China) according to the manufacturer’s instructions.

HEK293 cells were seeded in 6-well plate and transfection was performed when the cell density was 80%–90%. Transfection was performed using Lipofectamine 2,000 (Thermo Fisher Scientific, MA, United States) according to the manufacturer’s instructions. 4 μg plasmid (when co-transfect two plasmids, 2 μg per plasmid) and 5 μL lipofectamine 2000 were diluted using 250 μl OPTI-MEM respectively, after incubated at room temperature for 5 min, then mixed and incubated for another 20 min. The mixed reagent was added in HEK293 cells which grown in 2 ml DMEM medium with 10% FBS. Changed the mixture to DMEM medium with 10% FBS after 8 h and harvested the cell after 72 h.

To prepare the protein extract, 1 × 10⁶ cells were solubilized in HEPES buffer with 1% NP-40 (25 mM HEPES, pH 7.4, 150 mM NaCl, 1%, 1 mM PMSF and 1 × Protein inhibitor cocktail) and incubated on ice for 30 min. Cell lysates were centrifuged at 15,000×g for 10 min at 4°C. The resulting supernatants were collected as the protein extracts. To prepare the membrane fraction which contain endogenous ALG13/14, HEK293 cells (1 × 10⁶ cells) were resuspended in 1 ml of hypotonic buffer (10 mM Tris/HCl, pH 7.4, 1.5 mM MgCl₂, 10 mM KCl, protease inhibitor), followed by incubation on ice for 30 min. Cells were then homogenized by passing them 36 times through a homogenizer. Nuclei and cell debris were pelleted by centrifugation at 2,500×g for 5 min at 4°C. The supernatant was collected and centrifuged again at 12,000×g for 30 min at 4°C. The obtained membrane fraction (equivalent to 30 μg membrane protein) was used as the enzyme fraction of ALG13/14 GnTase in studies that characterized the suitability of synthetic Gn-PDol (C95) as substrate, as presented in Figure 1C.

E. coli co-expressing FLAG-ALG13-iso2 and 6His-ALG14 were cultured, and crude extracts were prepared as described above. Ten micrograms of this crude extract was separated by 12% SDS–PAGE analysis and transferred to a PVDF membrane (Bio-Rad, CA, United States). Immunoblotting was performed with an anti-His mouse antibody (TransGen Biotech, Beijing, China) and anti-FLAG mouse monoclonal antibody (TransGen Biotech, Beijing, China) as the primary antibody and the anti-mouse IgG-horseradish peroxidase (HRP) conjugate (TransGen Biotech, Beijing, China) as the secondary antibody followed by chemiluminescence (ECL) detection (Bio-Rad, CA, United States). For co-immunoprecipitation, 6His-ALG14 or FLAG-ALG13-iso2 was immunoprecipitated with Ni-NTA His binding resin (TransGen Biotech, Beijing, China) or anti-FLAG Affinity Matrix (Sigma–Aldrich, St. Louis, MO, United States) as described elsewhere (Gao et al., 2005). After separation by 12% SDS–PAGE, the precipitated proteins were transferred to an Immobilon-PVDF membrane. The membrane was incubated with an anti-FLAG mouse monoclonal antibody or anti-His mouse monoclonal antibody, washed, and incubated with anti-mouse IgG–HRP conjugate. Immunoreactive bands were visualized by ECL.

Similarly, HEK293 cells co-expressing HA-ALG13-iso1/FLAG-ALG14 or HA-ALG13-iso2/FLAG-ALG14 were cultured, and their protein extracts were prepared as described above. Ten μg of each protein extract was analyzed using western blot. Immunoblotting was performed with anti-FLAG mouse monoclonal antibody (Sigma-Aldrich, MO, United States) and anti-HA rabbit polyclonal antibody (Cell Signaling Technology, MA, United States) as the primary antibody and the anti-mouse IgG-HRP conjugate or the anti-rabbit IgG-HRP conjugate (TransGen Biotech, Beijing, China) as the secondary antibody. Immunoreactive bands were visualized by ECL. For co-immunoprecipitation assays, protein extracts were immunoprecipitated with anti-HA rabbit polyclonal antibody followed by incubation with 40 μl Protein A + G Agarose (Beyotime Shanghai, China) for 3 h to pull down the HA-tagged ALG13-iso1 or ALG13-iso2 protein. After washing 3 times with HEPES buffer, the immunoprecipitate was solubilized, separated by SDS-PAGE and analyzed by western blotting with anti-FLAG mouse monoclonal antibody to detect the co-immunoprecipitated FLAG-ALG14.

For the purification of ALG13-iso1 and ALG13-iso2, HEK293 cells (3 × 10⁶ cells) expressing HA-ALG13-iso1 or HA-ALG13-iso2 were harvested and the protein extract was prepared as described above. The protein extract was incubated with anti-HA rabbit polyclonal antibody and 40 μl Protein A + G Agarose for 4 h to enrich the HA-tagged ALG13-iso1 or ALG13-iso2 protein. After washing the agarose by PBS for 3 times, the agarose-bound immunoprecipitated proteins were divided into two parts. One part was immunoblotted with anti-ALG13 rabbit polyclonal antibody (Proteintech, Wuhan, China) to detect HA-ALG13 while the other part was employed to test the GnTase activity.

The standard reaction mixture for ALG13/14 GnTase assay contained in a volume of 50 μl: 28 mM Tris/HCl, pH 7.5, 0.3% NP-40, 23% glycerol, 40 μM Gn-PDol, 2 mM UDP-GlcNAc (Sigma-Aldrich, MO, United States) and protein. After incubation at 37°C, the reaction was stopped by heating to 100°C. The specific amount of protein included and reaction times are described in detail in each section. The donor concentration using 2 mM, which was 50 folds than the substrate Gn-PDol concentration. ALG13/14 can only transfer one GlcNAc from UDP-GlcNAc to Gn-PDol, so the donor concentration was excess which could facilitate the reaction.

After terminating the reaction, glycans were released by adding 50 μl of HCl (40 mM) to the reaction and shifting the temperature to 100°C for 1 h. The hydrolyzed glycans were purified as described (Li et al., 2017a; Li et al., 2017b; Xu et al., 2018; Li et al., 2019; Xiang et al., 2022). The desalted sample was injected into a Dionex Ultimate 3,000 UPLC (Thermo Scientific, MA, United States) and separated using a Waters Acquity UPLC BEH Amide Column (1.7 μm 2.1 × 150 mm) at a flow rate of 0.2 ml/min. The UPLC conditions were same as previous report while the ESI-MS range was adjusted to 100–1,500 m/z (positive mode). The oligosaccharide transfer rate was calculated using the peak intensity in LC-ESI-MS through Xcalibur (Version 2.0, Thermo Scientific, MA, United States).

Enzymatic assays were performed using purified protein with the final concentration as 0.3 ng/μl (Figures 3B–E) or 0.2 ng/μl (Figure 3F) under the standard conditions for 15 min as described above. The optimized reaction conditions were as follows: temperature from 10°C to 60°C, pH from 4.0 mM to 10.0 mM, 10 mM ions (Co²⁺, Mn²⁺ and Mg²⁺, respectively) or 10 mM EDTA for depletion conditions. The apparent K _m and k _cat values for Gn-PDol (20 μM–100 μM) were calculated by nonlinear regression curve fitting (GraphPad Prism 8.0, GraphPad Software Inc.). The donor specificity of ALG13/14 was confirmed using different donor (UDP-Glc, UDP-GlcNAc and GDP-Man) with same concentration (2 mM).

The stereochemistry of the linkage between two GlcNAc residues in Gn2-PDol (C95), the product of ALG13/14 GnTase, was confirmed by using yeast Alg1 (yAlg1) MTase. Purified recombinant His-yAlg1ΔTM and GDP-Man were added to the recombinant ALG13/14 reaction system, followed by 1 h of incubation at 30°C. The products were analyzed by LC–MS as described above.

The yeast strain XGY186 (as in W303a; ∆alg13::His ⁺⁵ ; ∆alg14::Kan ⁺ and GAL1/10pr-3HA-yALG14/yALG13-FLAG::trp1), contains a deletion of ALG13 and ALG14 in the chromosome and a GAL1/10 promoter-driven yALG13 and yALG14. This strain was used to test for the ability of human ALG13 different isoforms and ALG14 to complement growth in the presence of glucose. The URA3-marked YEp352GAPII-3HA-ALG13-iso1 or YEp352GAPII-3HA-ALG13-iso2 plasmid was co-transformed with the LEU2-marked YEp351GAPII-ALG14-FLAG into XGY186 and selected on SG (-Ura-Leu) plates. Transformants were cultured in medium containing glucose, followed by growth on solid YPA medium (1% yeast extract, 2% peptone, 50 mg/L adenine) supplemented with 2% glucose (YPAD) or 2% galactose (YPAG) at 30°C for 2 days.

The enzymatic properties, kinetic analysis and CDG variants relative activity experiments were independently repeated three times. Sample size, mean, and kinetic standard deviation (SD) value are provided in the figure, figure legend and table footnote. Kinetic parameters were determined by nonlinear regression fitted to Michaelis-Menten equation in GraphPad Prism 8.0.

The human ALG13/14 (ALG13/14) complex catalyzes the addition of a β1,4-linked GlcNAc to Gn-PDol to produce Gn2-PDol in the DLO synthetic pathway. We aimed to establish an in vitro assay system for quantitatively detecting UDP-GlcNAc transferase (GnTase) activity. We previously described an LC–MS-based quantitative assay for LLO mannosyltransferases (Li et al., 2019) and this assay was modified for quantitative measurements of ALG13/14 GnTase activity with a suitable Gn-PDol acceptor substrate. Unlike other ALG GTases, which can use phytanyl-linked oligosaccharides as their acceptor substrates for enzymatic studies (Li et al., 2017a; Li et al., 2017b; Xu et al., 2018; Li et al., 2019; Xiang et al., 2022), ALG13/14 GnTase does not recognize GlcNAc-pyrophosphate-phytanyl as the acceptor (data not shown). Therefore, we tried to prepare its natural Gn-PDol acceptor using semisynthetic methods. Since the lipid tails of the DLO in mammalian cells usually contain 17–21 isoprene units (C85–C105), we synthesized a Gn-PDol (C95) substrate that contains 19 isoprene units of lipid tails (Figure 1A). A polyprenol (C95) obtained from ginkgo leaves was applied to the asymmetric hydrogenation to generate dolichol, which was coupled to phosphorylated GlcNAc to obtain the target substrate Gn-PDol (C95, Figure 1A). The product structure was characterized by mass and nuclear magnetic resonance (NMR) analyses. The detailed synthetic procedures and data will be published elsewhere.

To test if this Gn-PDol (C95) is a suitable substrate for ALG13/14 GnTase, the HEK293 cells membrane fraction (equivalent to 30 μg membrane protein, which was confirmed contain endogenous ALG13 and ALG14 using anti-ALG13 antibody and anti-ALG14 antibody, data not shown) was prepared and applied to a standard reaction mixture with a final volume of 50 μl, in which Gn-PDol (C95) and UDP-GlcNAc were used as the acceptor and donor substrates (see Material and methods for details). Hydrochloric acid was added to hydrolyze the saccharide moieties, followed by purification and detection using LC–MS for quantitative analysis (Figure 1B). Two group peaks appeared on UPLC (Figure 1C). The corresponding ESI-MS analysis showed that the first peak eluted at 4.90 min was [Gn + H]⁺, and the mass peak at m/z 222.12 (Supplementary Figure S4) was hydrolyzed from the substrate Gn-PDol. The second group peak eluted at 7.39 min and 7.79 min was [Gn2+Na]⁺, with the mass peak m/z: 447.09, which was hydrolyzed from the product Gn2-PDol (Figures 1C,D). The two peaks of Gn2 have been proven to be glycan anomeric isomers, which were designated alpha (α) and beta (β) (Li et al., 2017a). The conversion rate was quantified by calculating the peaks intensity of Gn2 (two peaks)/peak intensity of (Gn + Gn2), which was 79.8% in Figure 1C. The boiled membrane fraction was applied to the standard reaction mixture as negative control, only one peak representing Gn was detected (data not shown). These results demonstrated the production of Gn2-PDol, indicating that Gn-PDol (C95) is a suitable substrate for ALG13/14 GnTase.

Although neither ALG13 nor ALG14 alone can rescue yeast alg13 or alg14 mutants, when co-expressed, human ALG13 (ALG13-iso2) and ALG14 can rescue the growth defect caused by either the loss of yeast ALG13 or ALG14 (Gao et al., 2005). These results demonstrated the formation of an active human ALG13/14 GnTase complex in yeast cells. Thus, to prepare active recombinant ALG13/14 complex, we chose to co-express human ALG13-iso2 and ALG14 in E. coli. Using N-FLAG- or His-tags, ALG13-iso2 and ALG14 were cloned into the pET26b and pCDFDuet expression vectors, respectively (Table 1). The resulting plasmids pET26b-pelB-FLAG-ALG13-iso2 and pCDFDuet-6His-ALG14 were transformed into E. coli Rosetta (DE3) cells. After induction with IPTG, the expression of both FLAG-ALG13-iso2 and His-ALG14 proteins was confirmed using anti-FLAG or anti-His antibodies (Figure 2A, left panel). Co-immunoprecipitation assays confirmed the formation of the ALG13/14 complex (Figure 2A, right panel). Immunoprecipitation of His-ALG14 with Ni-NTA His binding resin brought down the FLAG-ALG13-iso2 protein, which was detected by anti-FLAG antibody, and vice versa. These results suggested that ALG13/14 form a heterodimer in E. coli.

To determine if these recombinant proteins are active, GnTase activity of crude extracts was measured. Total protein (5 μg) prepared from E. coli that co-expressed ALG13-iso2 and ALG14 was incubated with the standard reaction mixture containing Gn-PDol (C95) and UDP-GlcNAc for 1 h, followed by detection of Gn2 product using LC–MS. The Gn2 product can be detected as two peaks at retention times of 7.39 min and 7.78 min (Figure 2B, left panel) with a mass of [Gn2+Na]⁺ (data not shown). When extracts prepared from two different E. coli strains (5 μg total protein from each strain), which individually expressed ALG13-iso2 and ALG14 were mixed and assayed for activity under the same conditions, no product peaks were detected (Figure 2B, right panel). These results indicate that the GnTase activity of the ALG13/14 complex cannot be obtained by mixing the individual recombinant ALG13 and ALG14 proteins. Instead, they imply that complex formation occurs in vivo and once formed, is stable.

To further confirm the stereochemistry of the GlcNAc-GlcNAc linkage in the newly formed product, we asked if it could be recognized as a substrate by yAlg1, the mannosyltransferase that catalyzes the reaction following ALG13/14 GnTase. yAlg1, adds β1,4-Man to GlcNAc(β1,4)GlcNAc-PDol (Gn2-PDol) with strict substrate specificity (Li et al., 2017a). To test this, recombinant yAlg1 was added to ALG13/14 GnTase reaction buffer with 2 mM GDP-mannose (GDP-Man) and incubated for another 1 h. After incubation, peaks corresponding to Man-Gn2 were observed at 10.41 min and 10.71 min, with a mass peak m/z of 587.27 (Figures 2C,D). The Man-Gn2 also was proved as glycan anomeric isomers. The ability of yAlg1 to produce this product implies the presence a β1,4 linkage in the GlcNAc-GlcNAc (Gn2) product of ALG13/14 GnTase. Taken together, our results demonstrated that co-expression of human ALG13-iso2 and ALG14 genes in E. coli can form an enzymatically active GnTase complex.

To further study its enzymatic properties, recombinant human ALG13/14 complex from E. coli harboring both pET26b-pelB-FLAG-ALG13-iso2 and pCDFduet-6His-ALG14 plasmids was purified as described in “Material and methods”. The purified protein was calculated as 2.6 mg/ml (0.1 ml), the concentration exhibiting a 500-fold compared with E. coli membrane extract. Native-PAGE analysis, in which the complex remains associated, revealed a single band with a molecular weight of approximately 45 kDa–66 kDa, while denaturing SDS–PAGE analysis revealed two bands, the individual ALG13-iso2 (25 kDa) and ALG14 (22 kDa) proteins (Figure 3A). Together, these results suggested that the purified recombinant ALG13/14 complex exists as a heterodimer.

Purified ALG13/14 was used to optimize reaction conditions (see Material and methods for details). As shown in Figures 3B–E, using 0.3 ng/μl recombinant proteins, our experiments demonstrated that the optimum pH and reaction temperature were 7.0°C and 37°C, respectively (Figures 3B,C). Moreover, of the three nucleotide sugars tested as donor substrates (UDP-GlcNAc, GDP-Man, and UDP-Glc), only UDP-GlcNAc was specifically recognized by the ALG13/14 complex (Figure 3D). Since additional divalent metal cations were required for most glycosyltransferase activities, the effects of various divalent metal ions on GnTase activity of ALG13/14 were tested. There were no improved activities confirmed in the presence of Mn²⁺, Mg²⁺ and EDTA, whereas Co²⁺ even inhibited the activity to some extent (Figure 3E). These results suggested that ALG13/14 GnTase does not require divalent metal ions for its enzyme activity. Furthermore, to investigate the kinetics of the recombinant ALG13/14, the rate of conversion to Gn2-PDol by purified recombinant ALG13/14 (0.2 ng/μl) was measured for the Gn-PDol (C95) acceptor from 20 mM to 100 mM with a fixed concentration of 2 mM UDP-GlcNAc as the donor substrate. The K _m value was calculated as 94.1 μM, and the k _cat value was 919.5 pmol/min by nonlinear regression fitted to Michaelis–Menten (Figure 3F).

As the canonical human isoform, the longer ALG13-iso1 contains not only an N-terminal glycosyltransferase 28 domain, but also several other domains including an ovarian tumor deubiquitinase domain (Ng et al., 2020). Compared to isoform 1, the short ALG13-iso2 lacks a large portion of the C-terminal region and uses an alternate 3′ terminal exon. The resulting iso2 protein shares N-terminal 127 amino acids with isoform 1, which contains the catalytic domain, and possesses a distinct C-terminal region consisting of 38 amino acids (Figure 4A). Among the ALG13 isoforms, only isoforms 1 and 2 contain the catalytic domain of ALG13/14 GnTase. It has been reported that ALG13-iso2 forms an active GnTase complex with ALG14 (Gao et al., 2005; Gao et al., 2008) (Figures 2, 3), but to date, no functional studies have ruled out that ALG13-iso1 also has catalytic activity.

To compare their GnTase activity, HA-tagged ALG13-iso1 protein was over-expressed alone in HEK293 and analyzed for enzymatic activity in vitro (Table 2). Cultured HEK293 cells (3 × 10⁶ cells) expressing HA-ALG13-iso1 were harvested and resuspended in 300 μl of HEPES buffer containing 1% Triton X-100 (see Material and methods for detailed composition). After incubation for 30 min on ice to solubilize membrane proteins, HA-ALG13-iso1 was immunoprecipitated from the cell lysate with anti-HA antibody and Protein A + G Agarose beads. The collected agarose beads were resuspended in 40 μL of reaction buffer. One half of this mixture (20 µl) was incubated with our standard reaction mixture to detect GnTase activity while the other half was subjected to SDS–PAGE analysis and immunoblotted with anti-ALG13 rabbit antibodies to confirm the expression and enrichment of HA-ALG13-iso1 proteins (data not shown). In parallel, HA-ALG13-iso2 was also expressed in HEK 293 cells and immunoprecipitated with anti-HA agarose (Table 2) (Gao et al., 2005). This pulled down ALG13-iso2 was selected as the positive control for GnTase activity because it can form an active GnTase complex with ALG14 in the ER membrane (Gao et al., 2005). Unexpectedly, by this assay HA-ALG13-iso1 did not show any GnTase activity, while HA-ALG13-iso2 could catalyze the formation of Gn2-PDol under the same reaction conditions (Figure 4B).

Because heterodimer formation is essential for ALG13/14 GnTase activity (Gao et al., 2008), we sought to determine the efficiency of complex formation between human ALG13-iso1 and ALG14 in HEK293 cells using co-immunoprecipitation assays (see Material and methods). HA-tagged ALG13-iso1 was co-expressed with FLAG-tagged ALG14 in HEK293 cells (Table 2). After extraction of the membrane proteins with 1% Triton X-100, HA-ALG13-iso1 was pulled down by anti-HA agarose. FLAG-ALG14 that co-precipitated with the HA-ALG13 was measured by immunoblotted with anti-FLAG antibodies. This experiment demonstrated that FLAG-tagged ALG14 completely failed to coprecipitate with HA-ALG13-iso1, although it efficiently interacted with HA-tagged ALG13-iso2 (Figure 4C). These experiments demonstrated that, unlike the shorter isoform 2, ALG13-iso1 does not display GnTase activity nor form a heterometric complex with ALG14.

To confirm the above results, a yeast complementation assay was performed to test whether ALG13-iso1 has any detectable in vivo GnTase activity. HA-ALG13-iso1 and ALG14-FLAG were co-expressed in yeast XGY186 strain (Table 2), in which both ALG13 and ALG14 chromosomal loci are deleted. The survival of this strain depends on the galactose-driven promoter ALG13/14 (GAL1pr-ALG13/14), which can be shut off by shifting to growth in medium containing glucose. To determine if ALG13-iso1 and ALG14 have in vivo activity, we measured their ability to rescue growth of this yeast strain in the presence of galactose versus glucose. As shown in Figure 4D, HA-ALG13-iso1 and ALG14-FLAG co-expression did not rescue the growth defect caused by the inhibition of yeast Alg13/14; in contrast, HA-ALG13-iso2 and ALG14-FLAG co-expression rescued the growth defect. Consistent with our in vitro results (Figure 4B), this complementation assay further confirmed that ALG13-iso1 does not possess GnTase activity in vivo or in vitro.

To date, a total of sixteen ALG13-CDG and nine ALG14-CDG missense mutations have been identified from patients (de Ligt et al., 2012; Timal et al., 2012; Esposito et al., 2013; Judith et al., 2013; Smith-Packard et al., 2015; Dimassi et al., 2016; Kobayashi et al., 2016; Bastaki et al., 2017; Hamici et al., 2017; Schorling et al., 2017; Kvarnung et al., 2018; Ng et al., 2020; Alsharhan et al., 2021; Datta et al., 2021; Palombo et al., 2021; Gang et al., 2022; Katata et al., 2022). Among these ALG13-CDG mutations, eight are located in the N-terminal region of ALG13 that is shared by isoform 1 and 2 (Figure 5A), while the remaining seven mutations are located in regions unique to ALG13-iso1, and only one mutation is located in short C-terminal of isoform 2 (data not shown). Based on the results from our study on ALG13 isoforms (Figures 1, 4), we predicted that the eight ALG13-CDG mutations located in the common N-terminal region may affect GnTase activity. To test this idea, we developed a semiquantitative assay to measure the effect of ALG13-CDG and ALG14-CDG mutations on GnTase activity.

We first chose ALG13 N107S and ALG14 P65L as candidates for this analysis as these two mutations have been identified in multiple CDG patients (see Figures 5A,B). These mutations were introduced into plasmids and co-expressed as FLAG (ALG13) or 6His (ALG14) tagged alleles in E. coli to produce strains co-expressing wild type (WT) ALG13 + WT ALG14, WT ALG13 + mutant ALG14 P65L or mutant ALG13 N107S and WT ALG14 (see Materials and Methods). Membrane fractions containing the ALG13 N107S/ALG14 or ALG13/ALG14 P65L complex were collected and solubilized in 5 ml lysis buffer with 1% Triton X-100 (Figure 5B, see Material and methods for details). These detergent extracts were analyzed with anti-FLAG or anti-His antibodies to detect the expression of the ALG13/ALG14 complex (Supplementary Figure S5A). To quickly quantify the concentration of the ALG13/ALG14 complex in the extract, a standard curve method was used, in which purified recombinant yeast Alg13 was used as the standard (Supplementary Figure S5B).

To compare the GnTase activity of the WT and CDG mutants, detergent extracts containing 0.1 ng/μl of ALG13 protein were incubated with the standard reaction mixture at 37°C for 15 min, followed by LC-MS analysis. The use of E. coli membrane which is absent from endogenous ALG13/14 would retain the enzyme activity and avoid the laborious protein purification steps. As shown in Figure 5C, WT ALG13/ALG14 complex GnTase catalyzed the production of Gn2-PDol (C95) with a conversion rate of 74%, while the conversion rate catalyzed by ALG13 N107S/ALG14 or ALG13/ALG14 P65L complex was 6% or 59%, respectively. These results demonstrated that ALG13 N107S and ALG14 P65L CDG mutations diminished the GnTase activity of ALG13/ALG14.

Similar to ALG13 N107S and ALG14 P65L, GnTase activities in several other ALG13- and ALG14-CDG mutants (Figure 5A, marked with red) were tested. Plasmids containing FLAG-tagged ALG13-iso2 I17N, E69del, A81T, K94E or T141L were transformed into E. coli cells and co-expressed with WT 6His-tagged ALG14. Similarly, plasmids containing 6His-tagged ALG14 R104X, D74N, R109Q or V141G were co-expressed with WT FLAG-tagged ALG13. After confirming protein expression of each ALG13-iso2 and ALG14 variant (Supplementary Figure S5C,E), detergent extracts prepared from each of these E. coli strains were assayed for GnTase activity. For a more rigorous comparison with WT ALG13/ALG14 GnTase, the relative activity of each mutation was introduced to show the remaining activity of the ALG13/ALG14 complex mutants. For example, as shown in Table 3, the relative activities were defined using the conversion rate of the mutants/the conversion rate of the WT, and ALG13 N107S/ALG14 and ALG13/ALG14 P65L were calculated as 8% and 80%, respectively. Notably, all of the ALG13/ALG14-CDG mutants that were tested displayed reduced GnTase activity despite the fact that most patients carrying these mutants had a normal N-linked glycosylation phenotype as judged by CDT analysis (Table 2) (Judith et al., 2013; Gadomski et al., 2017; Alsharhan et al., 2021). In conclusion, our results confirmed the GnTase deficiency of these ALG13/ALG14-CDG variants.