Gmds^Y187*/Y187* (Gmds^m1Anu/m1Anu ) mice were generated by N-ethyl-N-nitrosourea mutagenesis, and the causal mutation was mapped using a previously published strategy (29). For the mapping process, an unaffected carrier was bred with a CBA mouse, and the resulting offspring were randomly intercrossed. The F2 mice were phenotyped, and all the mice from breeder pairs that produced affected offspring were genotyped for a panel of markers evenly distributed across all chromosomes. This defined an area on Chromosome 13, between rs13481734 and rs3701164, that was associated with the phenotype.

In parallel, we performed exome sequencing on an affected mouse to identify homozygous variants, as previously described (30). The only protein changing variant identified in the interval was the GMDS^Y187* mutation. To eliminate any unrelated mutations, the line was maintained for more than ten generations by crossing a Gmds^Y187*/+ carrier with a wild-type C57BL/6J mouse. Experimental mice were generated by crossing heterozygous Gmds^Y187*/+ mice from the backcross. PCR amplification was performed from DNA isolated from mouse-ear punches for genotyping. All experimental mice were housed under specific pathogen-free conditions at the Australian National University, and all animal procedures were approved by the Australian National University Animal Ethics and Experimentation Committee under protocols A2014/62 and A2017/54.

Flow cytometric analysis of cell suspensions from bone marrow, thymus and spleen was performed as previously described (30, 31). Briefly, single cell suspensions were prepared and red blood cells were lysed for spleen samples. For surface staining, cells were counted using the ViCELL cell counter (Beckham Coulter, Brea, CA, USA) and equal numbers of cells were first treated with purified rat anti-mouse CD16/CD32 (BD, San Jose, CA, USA; clone 2.4G2) and anti-biotin antibodies in ice-cold staining buffer (phosphate-buffered saline containing 2% bovine serum and 0.1% NaN₃) and incubated for 30 min at 4°C in the dark. Cells were then washed and stained with a primary antibody cocktail containing an appropriate combination of the antibodies and streptavidin in ice-cold staining buffer and incubated for 30 min at 4°C in the dark. Samples were washed with staining buffer and resuspended in the same buffer, followed by analysis on an LSR II (BD, San Jose, CA, USA) or LSR Fortessa flow cytometer (BD, San Jose, CA, USA) and FlowJo 887 software (FlowJo, LLC, Ashland, OR, USA) was used for data analysis.

The following antibodies were used for flow cytometry analysis: anti-CD3 Alexa Fluor 700 (eBioscience, San Diego, CA, USA; clone 145-2C11), anti-CD4 APC (BD, San Jose, CA, USA; clone RM4-5), anti-CD4 APC Cy7 (BioLegend, San Diego, CA, USA; clone GK1.5), anti-CD8a PE (BD; clone 53-6.7), anti-CD8a PE Cy7 (Biolegend; clone 53-6.7), anti-CD11b APC Alexa Fluor 750 (eBioscience; clone M1/70), anti-CD19 Brilliant Violet 510 (Biolegend; clone 6D5), anti-CD19 APC Vio770 (Miltenyi Biotec, Bergisch Gladbach, Germany; clone 6D5), anti-CD21/35 FITC (BD; clone 7G6), anti-CD23 Pacific Blue (Biolegend; clone B3B4), anti-CD23 PE (BD; clone B3B4), anti-CD25 Biotin (BD; clone 7D4), anti-CD25 FITC (BD; clone 7D4), anti-CD44 Pacific Blue (Biolegend; clone IM7), anti-CD45R/B220 APC (BD; clone RA3-6B2), anti-CD45R/B220 AF700 (BD; clone RA3-6B2), anti-CD45R/B220 Brilliant Violet 605 (Biolegend; clone RA3-6B2), anti-CD45.1 Alexa Fluor 700 (Biolegend; clone A20), anti-CD45.1 FITC (BD; clone A20), anti-CD45.2 APC (eBioscience; clone B104), anti-CD45.2 Pacific Blue (Biolegend; clone B104), CD62L Brilliant Violet 605 (Biolegend; clone MEL-14), anti-CD93 Biotin (eBioscience; AA4.1), anti-Gr-1 Biotin (BD; clone RB6-8C5), anti-IgD FITC (eBioscience; clone 11-26c), anti-IgM PE Cy7 (BioLegend; clone 53-6.7), anti-NK1.1 PE (BD; clone PK136), Streptavidin Brilliant Violet 605 (Biolegend), Streptavidin PE Cy7 (eBioscience), anti-TCRβ APC Cy7 (Biolegend; clone H57-597), anti-TCRγδ FITC (BD; clone GL3) and 7-aminoactinomycin D (Thermo Fisher Scientific, Waltham, MA, USA).

100% chimeras were generated using bone marrow cells extracted and counted from Gmds^+/+ and Gmds^Y187*/Y187* donor mice and injected intravenously into Rag1 ^-/- mice (32) that were irradiated with a single dose of 500 rads. Mixed chimeras were generated using bone marrow cells extracted from either Gmds^+/+ CD45.2 or Gmds^Y187*/Y187* CD45.2, counted and mixed 1:1 with Gmds^+/+ CD45.1/2 animals. Gmds^+/+ CD45.1 recipient mice were irradiated with two doses of 450 rads prior to intravenous injection of bone marrow donor cells for each recipient mouse. The second cohort of mixed chimeras was generated by extracting, counting, and mixing 1:1 bone marrow cells from either Gmds^+/+ CD45.2 or Gmds^Y187*/Y187* CD45.2 with Gmds^+/+ CD45.1 animals. Rag1 ^-/- recipient mice were irradiated with a single dose of 500 rads prior to intravenous injection of with bone marrow donor cells for each recipient mouse.

Data were statistically analyzed using GraphPad Prism 9 for Mac OS X (GraphPad Software, San Diego, CA, USA). Student’s t-test was used to compare only two groups. When multiple experimental groups were compared, one-way ANOVA with Tukey’s multiple comparison test was used. For all statistical analysis, differences were considered to be significant when p < 0.05.

To identify novel genes with critical roles in the development and/or function of immune cells (29), we screened mice carrying N-ethyl-N-nitrosourea-induced point mutations. Among the offspring, several mice exhibited severe growth retardation after weaning and a reduction in body weight of approximately 42% ( Figure 1A ). The causative mutation was mapped to the gene encoding GMDS, which catalyzes the conversion of GDP-mannose to GDP-4-keto-6-deoxymannose, the first step in the synthesis of GDP-fucose from GDP-mannose, which can be used as a substrate for fucosylation (5). The mutation is a T-to-A substitution at position 818 on cDNA that results in the conversion of tyrosine to a STOP codon at amino acid 187 (p.Y187*). The mutation is predicted to remove the last 185 amino acids of GMDS, potentially leading to loss of cellular fucosylation. The median survival of mice with the growth defect was 5 weeks and all Gmds^Y187*/Y187* mice died by 95 days after birth ( Figure 1B ). Genotyping of offspring from intercrosses of Gmds^Y187*/+ mice showed that Gmds^+/+ and Gmds^Y187*/+ mice were born at the expected 1:2 ratio but less than 7% of all weaned mice were homozygous ( Figure 1C ), suggesting the requirement for proper fucosylation for embryonic development or survival in the early postnatal phase before weaning (14–16). However, Gmds^Y187*/+ mice were virtually indistinguishable from their Gmds^+/+ littermates in terms of size and survival ( Figures 1A, B ).

While the reduced survival could be attributed to the requirement for fucosylation during embryonic and/or postnatal development (14–16), we focused our investigation on the role of GMDS in lymphocyte development. As we observed the death of Gmds^Y187*/Y187* mice at approximately 5 weeks of age, we immunophenotyped the mice at 3 weeks of age by analyzing the thymus and spleen. Gmds^Y187*/Y187* mice had lower thymic cellularity compared to their Gmds^Y187*/+ and Gmds^+/+ littermates ( Figure 2A ). Analysis of different T cell subsets in the thymus of Gmds^Y187*/Y187* mice at 3 weeks of age revealed a small but statistically significant decrease in the frequency of double positive (DP) cells with a concomitant increase in the percentage of DN cells, CD4 single positive (SP) and CD8SP cells ( Figure 2B ). In absolute terms, this phenotype resulted in normal numbers of DN thymocytes, but significantly reduced numbers of DP, CD4SP and CD8SP cells ( Figure 2B ). This observation is consistent with the up-regulation of core fucosylation at the transition from DN to DP cells in the thymus (26).

Depending on the expression of the adhesion molecule CD44 and the interleukin (IL)-2 receptor α chain CD25, DN cells can be divided into four subpopulations, namely CD44⁺CD25^– DN1 cells, CD44⁺CD25⁺ DN2 cells, CD44^–CD25⁺ DN3 cells, and CD44^–CD25^– DN4 cells (33). Detailed analysis of the DN subsets based on their CD44 and CD25 expression revealed a lower CD25 expression on Gmds^Y187*/Y187* thymocytes ( Figure 2C ). This resulted in a trend towards a reduced percentage of DN3 cells and increased DN4 cells in the thymus of Gmds^Y187*/Y187* mice ( Figure 2C ). However, the absolute numbers of different DN subsets in the thymus of Gmds^Y187*/Y187* mice appeared normal compared to their Gmds^Y187*/+ and Gmds^+/+ littermates ( Figure 2C ). Interestingly, reduced expression of CD25 was specific to DN cells, with CD4⁺FoxP3⁺ cells in the thymus and spleen of Gmds^Y187*/Y187* mice exhibiting comparable levels of CD25 expression to those of Gmds^Y187*/+ and Gmds^+/+ ( Supplementary Figure 1 ). These data suggest that the mutation in the gene encoding GMDS affects the development of T cells in the thymus.

We then examined the different immune cell subsets in the spleen of mice. At 3 weeks of age, the absolute number of total splenocytes in Gmds^Y187*/Y187* mice was less than 50% of their Gmds^Y187*/+ and Gmds^+/+ littermates ( Figure 3A ). Surprisingly, despite the fact that Gmds^Y187*/Y187* mice had a partial developmental block at the DN stage of T cell development in the thymus ( Figure 2 ) and that the total number of splenocytes was significantly reduced, Gmds^Y187*/Y187* mice had comparable number of T cells in the spleen ( Figure 3B ). Analysis of CD62L and CD44 expression on CD4⁺ and CD8⁺ T cells also revealed a largely normal profile in the spleen of Gmds^Y187*/Y187* mice compared to their Gmds^Y187*/+ and Gmds^+/+ littermates ( Figure 3C ; Supplementary Figure 2 ).

The total numbers of B cells and CD11b⁺Gr1⁺ myeloid cells were significantly reduced in the spleen of Gmds^Y187*/Y187* mice ( Figure 3B ). Detailed analysis revealed the normal presence of B cell subsets in the spleen of Gmds^Y187*/Y187* mice ( Figure 3D ; Supplementary Figure 3 ). As Gmds^Y187*/Y187* mice had essentially normal splenic B cell subsets, the reduction in total B cells and CD11b⁺Gr1⁺ numbers could simply be attributed to either the lower total number of splenocytes in Gmds^Y187*/Y187* mice ( Figure 3A ) or an unknown effect of the GMDS^Y187* mutation on the development and/or survival of other immune cell subsets, including innate lymphoid cells (ILCs).

Since Gmds^Y187*/Y187* mice showed poor postnatal survival and a partial block in T cell development ( Figures 1 , 2 ), we next wanted to determine whether the T cell phenotype observed in intact mice resulted from lack of GMDS-mediated fucosylation in hematopoietic cells. To this end, we performed bone marrow transplantation, where Gmds^+/+ or Gmds^Y187*/Y187* bone marrow cells were transferred into irradiated lymphoid-deficient recipients (Rag1^-/- ) followed by analysis after reconstitution. Consistent with the analysis of intact Gmds^Y187*/Y187* mice, these experiments showed that the total number of thymocytes was significantly reduced in Rag1^-/- mice transplanted with Gmds^Y187*/Y187* bone marrow cells compared to Rag1^-/- mice transplanted with Gmds^+/+ bone marrow cells ( Figure 4A ). While the number of DN cell subsets in mice transplanted with Gmds^Y187*/Y187* marrows was similar to that in mice transplanted with Gmds^+/+ cells, there was a significant reduction in the number of Gmds^Y187*/Y187* -derived DP cells in the recipients ( Figure 4A ). There was a trend towards reduced numbers of CD4SP and CD8SP cells in the thymus of recipient mice, but the difference was not statistically significant ( Figure 4A ). Analysis of the spleen of recipient mice demonstrated that the absolute numbers of Gmds^Y187*/Y187* -derived total T cells, CD4⁺ and CD8⁺ T cell subsets were comparable to Gmds^+/+ -derived cells ( Figure 4B ). The ratio of CD4 and CD8 T cells, and the CD44^lo naïve and CD44^hi memory T cell subsets within CD4⁺ and CD8⁺ cells in the spleen of Rag1^-/- recipients engrafted with Gmds^Y187*/Y187* bone marrow cells were also comparable to those of recipients transplanted with Gmds^+/+ bone marrow cells ( Figure 4C ).

We also analyzed B cell development in the bone marrow and spleen of Rag1^-/- recipient mice transplanted with Gmds^Y187*/Y187* bone marrow and found that B cell subsets were present in normal numbers compared to animals engrafted with Gmds^+/+ cells ( Figure 4D, E ). Interestingly, there was an increase in the percentage of SSC^hiFSC^hi myeloid cells in the spleen of recipient mice engrafted with Gmds^Y187*/Y187* bone marrow cells ( Figure 4F ). Apart from the differences in immunological phenotypes, there was no apparent morbidity in Rag1^-/- mice that were transplanted with Gmds^Y187*/Y187* bone marrow cells alone.

In order to further assess whether GMDS controls T cell development in a cell-autonomous manner, we also analyzed the development and persistence of different immune cell types in a cohort of mixed bone marrow chimeras. The chimeras were established with a 1:1 ratio of bone marrow cells from Gmds^+/+ (CD45.2) or Gmds^Y187*/Y187* (CD45.2) and Gmds^+/+ (CD45.1/2), which were injected into sublethally irradiated Gmds^+/+ (CD45.1) recipients ( Figure 5A ). Consistent with our findings presented above, we found a profound developmental block at the DN stage of T cell development, as evidenced by a reduction in the proportion of Gmds^Y187*/Y187* -derived DP, CD4SP and CD8SP T cells in the thymus of the chimeric recipients ( Figure 5B ). As a result of a defect in early development, Gmds^Y187*/Y187* -derived T cells failed to accumulate in the spleen of irradiated recipients ( Figure 5B ). The percentage of Gmds^Y187*/Y187* -derived TCRγδ cells was also reduced in the thymus and spleen of recipient mice, but the reduction was less compared to TCRαβ T cell subsets ( Figure 5B ). Analysis of different DN subsets revealed that there were relatively fewer Gmds^Y187*/Y187* -derived DN3 and DN4 cells compared to DN1 and DN2 cells ( Figures 5C, D ). Interestingly, we again observed a reduced expression of CD25 on DN thymocytes derived from Gmds^Y187*/Y187* bone marrow but not on Gmds^+/+ T cells developing in the same animals, demonstrating that this is a cell-intrinsic phenotype and not secondary to any cell extrinsic alteration ( Figure 5C ).

We next tested the percentages of other immune cell types and found that B cells in the bone marrow and B cell subsets in the spleen of recipient mice transplanted with Gmds^Y187*/Y187* cells were broadly comparable to those transplanted with Gmds^+/+ bone marrow cells ( Supplementary Figure 4A ). Similarly, the percentage of Gmds^Y187*/Y187* -derived CD11b⁺Gr1^-, CD11b⁺Gr1^int and CD11b⁺Gr1^hi cells in the spleen of recipients was comparable to those transplanted with Gmds^+/+ cells ( Supplementary Figure 4B ). Interestingly, there was a decrease in the percentage of Gmds^Y187*/Y187* -derived NK1.1⁺ Natural Killer (NK) cells compared to other cell types in the spleen of recipient mice ( Supplementary Figure 4B ).

These results were confirmed in a second cohort of mixed bone marrow chimeras but using irradiated immunocompromised (Rag1^-/- ) mice as recipients ( Supplementary Figures 5A–C ). Collectively, these findings demonstrate that the GMDS^Y187* mutation acts cell intrinsically to regulate T cell development in vivo and is dispensable for B cell development.