What is SLC35A2?

SLC35A2 is the name of a gene that encodes a UDP-galactose transporter protein. SLC35A2 transporters call Golgi membranes home. Its function is to concentrate inside Golgi compartments an oligosaccharide building block and nucleotide sugar fusion called UDP-galactose. SLC35A2 is a close relative of SLC35A1 but closer still is SLC35A3, which transports a chemically related nucleotide sugar called UDP-N-acetylglucosamine. Mutations in any of those genes cause a Type 2 congenital disorder of glycosylation. SLC35A2-CDG was referred to as CDG-2m. (See the SLC35A1-CDG Disease page for more on Type 2 CDGs).

Type 2 CDGs involve the later stages of N-linked glycosylation when young proteins have left the endoplasmic reticulum for the Golgi to mature. Like an automated robotic assembly line, a procession of Golgi resident enzymes adds and subtracts various flavors of sugar to and from maturing proteins, lengthening and pruning polysaccharide plumage as proteins pass from one Golgi lobe to the next. SLC35A2 is evolutionarily conserved in mice, zebrafish, flies, worms and fission yeast.

How does SLC35A2 work?

Unlike SLC35A1-CDG, SLC35A2-CDG is X-linked, meaning the SLC35A2 gene lives on the X chromosome. Mutations causing SLC35A2-CDG are either maternally inherited or spontaneous mutations that arise during fetal development or rare cases of late-onset epilepsy in the brain (Winawer et al., 2018). There’s a complication called somatic mosaicism in both males and females, with some cells expressing mutant SLC35A2 and other expressing the normal copy. X-inactivation drives clinical variability in females. In girls the clinical presentation affects fewer organ systems because cells that express the good copy of SLC35A2 outcompete sister cells disadvantaged by a mutant copy of SLC35A2. Only three boys with SLC35A2-CDG have been described to date. Because boys have one X chromosome and therefore one copy of SLC35A2, only mild mutations are compatible with life. The reason why the preponderance of patients are female is because males are not surviving to term and the affected males that do survive are protected by mosaicism and disease modifiers.

Mosaicism is the product of a process called X inactivation, which in a clinical setting appears to spare “specific skin features, coagulopathies, immune dysfunction, cardiomyopathy, and renal dysfunction, which are often observed in glycosylation disorders, were not found in these three patients, although they did present epileptic seizures (EOEE),” as noted by Kimizu et al., 2013

A conclusion that can be drawn from which organ systems are spared and which organ systems are vulnerable is that the brain is particularly sensitive to loss of SLC35A2 transport activity that leaves proteins with glycan appendages shortchanged of galactose. In the absence of SLC35A2, UDP-galactose can’t get inside the Golgi where resident enzymes incorporate it into not just N-linked glycans but also O-linked glycans, glycosaminoglycans and glycosphingolipids. SLC35A2-CDG patient fibroblasts manifest glycosylation defects that can be robustly measured in the lab.

Clinical descriptions of SLC35A2-CDG

One of the first clinical reports features a small cohort of three unrelated CDG-1x children, two boys and one girl, whose diagnostic odysseys led them to a causal gene named SLC35A2 (Ng et al., 2013).

One of the affected children had a somatic mutation and fewer organ defects. A somatic mutation in SLC35A2 means it was not present in either parents and so arose spontaneously post fertilization.

Figure 1

Figure 1 from Ng et al., 2013 showing the glycosylation status of serum transferrin in a control fibroblast (A) compared to a male SLC35A2-CDG patient with an early premature stop mutation at 5 months old (B) and then at 38 months old (C). A minor fraction of the patient’s transferrin lacks terminal galactose sialic acid disaccharides but this defect resolved by the time he turned 3.

The surface glycoproteins of SLC35A2-CDG patient fibroblasts are indeed truncated. Further studies of patient fibroblasts revealed that mutant SLC35A2 transporters have reduced but not eliminated transport function. This result raises the possibility that one of SLC35A2’s close relations is picking up the transport slack, which has implications for therapeutic strategies.

Figure 3

Figure 3 from Ng et al., 2013 showing a 50% reduction in the transport of a tritiated (as in tritium, a radioactive isotope of hydrogen) form of UDP-galactose in all three SLC35A2-CDG patient fibroblast lines (CDG-341, CDG-348, CDG-352) compared to control fibroblasts. 

Fast forward six years to Ng et al., 2019, which presents the largest clinical report to date: 29 newly diagnosed SLC35A2-CDG patients and 26 never before seen pathogenic variants. Most SLC35A2-CDG patients are not diagnosed until the age of 3. 97%-100% of SLC35A2-CDG patients in this cohort has intellectual disability, developmental delay and neurological symptoms. 77% of the cohort had failure to thrive. However the mortality rate appears to be low.

SLC35A2 and infection

Loss of SLC35A2, as is the case for most genes, is like dropping a stone in a pond. The ripple effects can be unexpected and far-reaching. For example, SLC35A2 is required for infection by influenza virus (Moskovskich et al., 2019).

Figure 1A-C

Figure 1A-C from Moskovskich et al., 2019 showing the results of a genetic screen to identify genes whose loss leads to survival of cells after infection by influenza virus. SLC35A2 and SLC35A1 are the only statistically significant hits.  

SLC35A2 is also required for cholera toxin to work in cells (Guimaraes et al., 2011). The mechanism doesn’t appear to involve N-linked glycoproteins but rather N-linked glycolipids. Complex N-glycan linked lipids call gangliosides are receptors for cholera toxin and their formation requires galactose building blocks.

Steps toward the clinic?

Standard-issue advise for any CDG community is to find more patients like you so a natural history study can be conducted. In the specific case of SLC35A2-CDG, grow the community by improving the diagnostic rate of unsolved intellectual disability cases using clinical whole exome sequencing. Keep in mind that the shelf life of serum transferrin as a diagnostic criterion is limited and probably only useful very early in life.
Once again, there is a dearth of SLC35A2-CDG disease models. SLC35A2 loss-of-function mutations can theoretically be modeled in mice, zebrafish, flies, worms and fission yeast.

The orthologs, or ancestral versions, of SLC35A2 across species.

Can we compensate for sluggish mutant SLC35A2 transporters by feeding SLC35A2-CDG patients more dietary galactose? Results from Ng et al., 2013 suggest that might not work but they only tested galactose supplementation on fibroblasts. They noted: “We tried to improve galactosylation in each fibroblast line by providing 150–300 µM galactose for between 2–4 days, but we saw no significant improvement in GSII binding.”

Even though the SLC35A2 gene encodes the only UDP-galactose transporter in the human genome, SLC35A2 has close cousins that could in a pinch step in for it. SLC35A2-CDG patient fibroblasts have residual UDP-galactose transport activity, which suggests SLC35A3 or SLC35A4 or SLC35A5 or even SLC35A1 are filling the breach.

SLC35A2 and SLC35A3 are known to form heterodimers, i.e., one SLC35A2 protein links arms with a SLC35A3 protein (Maszczak-Seneczko et al., 2015). So maybe pharmacologically super-charged SLC35A1/3/4/5 could partner up with a faulty SLC35A2 and snap it back into shape? First things first. Cellular and animal disease models need to be generated and SLC35A2 functional assays, e.g., UDP-galactose transport, need to be optimized before unbiased or targeted drug screening can take place.


Dörre K, Olczak M, Wada Y, Sosicka P, Grüneberg M, Reunert J, Kurlemann G, Fiedler B, Biskup S, Hörtnagel K, Rust S, Marquardt T. (2015). A new case of UDP-galactose transporter deficiency (SLC35A2-CDG): molecular basis, clinical phenotype, and therapeutic approach. Journal of Inherited Metabolic Diseases. 38: 931-940.

Guimaraes CP, Carette JE, Varadarajan M, Antos J, Popp MW, Spooner E, Brummelkamp TR, Ploegh HL. (2011). Identification of host cell factors required for intoxication through use of modified cholera toxin. Journal of Cell Biology. 195: 751-64.

Kimizu T, Takahashi Y, Oboshi T, Horino A, Koike T, Yoshitomi S, Mori T, Yamaguchi T, Ikeda H, Okamoto N, Nakashima M, Saitsu H, Kato M, Matsumoto N, Imai K. (2017). A case of early onset epileptic encephalopathy with de novo mutation in SLC35A2: Clinical features and treatment for epilepsy. Brain Development. 39: 256-260.

Kodera H, Nakamura K, Osaka H, Maegaki Y, Haginoya K, Mizumoto S, Kato M, Okamoto N, Iai M, Kondo Y, Nishiyama K, Tsurusaki Y, Nakashima M, Miyake N, Hayasaka K, Sugahara K, Yuasa I, Wada Y, Matsumoto N, Saitsu H. (2013). De novo mutations in SLC35A2 encoding a UDP-galactose transporter cause early-onset epileptic encephalopathy. Human Mutations. 34: 1708-1714.

Maszczak-Seneczko D, Sosicka P, Kaczmarek B, Majkowski M, Luzarowski M, Olczak T, Olczak M. (2015). UDP-galactose (SLC35A2) and UDP-N-acetylglucosamine (SLC35A3) Transporters Form Glycosylation-related Complexes with Mannoside Acetylglucosaminyltransferases (Mgats). Journal of Biological Chemistry. 290: 15475-15486.

Moskovskich A, Goldmann U, Kartnig F, Lindinger S, Konecka J, Fiume G, Girardi E, Superti-Furga G. (2019). The transporters SLC35A1 and SLC30A1 play opposite roles in cell survival upon VSV virus infection. Scientific Reports. 9: 10471. 

Ng BG, Sosicka P, Agadi S, Almannai M, Bacino CA, Barone R, Botto LD, Burton JE, Carlston C, Chung BH, Cohen JS, Coman D, Dipple KM, Dorrani N, Dobyns WB, Elias AF, Epstein L, Gahl WA, Garozzo D, Hammer TB, Haven J, Héron D, Herzog M, Hoganson GE, Hunter JM, Jain M, Juusola J, Lakhani S, Lee H, Lee J, Lewis K, Longo N, Lourenço CM, Mak CCY, McKnight D, Mendelsohn BA, Mignot C, Mirzaa G, Mitchell W, Muhle H, Nelson SF, Olczak M, Palmer CGS, Partikian A, Patterson MC, Pierson TM, Quinonez SC, Regan BM, Ross ME, Guillen Sacoto MJ, Scaglia F, Scheffer IE, Segal D, Singhal NS, Striano P, Sturiale L, Symonds JD, Tang S, Vilain E, Willis M, Wolfe LA, Yang H, Yano S, Powis Z, Suchy SF, Rosenfeld JA, Edmondson AC, Grunewald S, Freeze HH. (2019). SLC35A2-CDG: Functional characterization, expanded molecular, clinical, and biochemical phenotypes of 30 unreported Individuals. Human Mutations. Feb 28.

Ng BG, Buckingham KJ, Raymond K, Kircher M, Turner EH, He M, Smith JD, Eroshkin A, Szybowska M, Losfeld ME, Chong JX, Kozenko M, Li C, Patterson MC, Gilbert RD, Nickerson DA, Shendure J, Bamshad MJ; University of Washington Center for Mendelian Genomics, Freeze HH. 2013). Mosaicism of the UDP-galactose transporter SLC35A2 causes a congenital disorder of glycosylation. American Journal Human Genetics. 92: 632-636.
Winawer MR, Griffin NG, Samanamud J, Baugh EH, Rathakrishnan D, Ramalingam S, Zagzag D, Schevon CA, Dugan P, Hegde M, Sheth SA, McKhann GM, Doyle WK, Grant GA, Porter BE, Mikati MA, Muh CR, Malone CD, Bergin AMR, Peters JM, McBrian DK, Pack AM, Akman CI, LaCoursiere CM, Keever KM, Madsen JR, Yang E, Lidov HGW, Shain C, Allen AS, Canoll PD, Crino PB, Poduri AH, Heinzen EL. (2018). Somatic SLC35A2 variants in the brain are associated with intractable neocortical epilepsy.Annals of Neurology. 83: 1133-1146.

Sign up for the CDG Hub mailing list

Would you like to be contacted by (pick all that apply)