REPORT

PGM3 Mutations Cause a Congenital Disorderof Glycosylation with Severe Immunodeficiencyand Skeletal Dysplasia

Asbjørg Stray-Pedersen,1,2,3,4,* Paul H. Backe,5,6,7 Hanne S. Sorte,4 Lars Mørkrid,6,7 Niti Y. Chokshi,3,8

Hans Christian Erichsen,9 Tomasz Gambin,1 Katja B.P. Elgstøen,6 Magnar Bjøras,5,7

Marcin W. Wlodarski,10 Marcus Kruger,10 Shalini N. Jhangiani,1,11 Donna M. Muzny,1,11 Ankita Patel,12

Kimiyo M. Raymond,13 Ghadir S. Sasa,8,14 Robert A. Krance,8,14 Caridad A. Martinez,8,14

Shirley M. Abraham,15 Carsten Speckmann,10 Stephan Ehl,10 Patricia Hall,16 Lisa R. Forbes,2,3,8

Else Merckoll,17 Jostein Westvik,17 Gen Nishimura,18 Cecilie F. Rustad,4 Tore G. Abrahamsen,7,9

Arild Rønnestad,9 Liv T. Osnes,19 Torstein Egeland,7,19 Olaug K. Rødningen,4 Christine R. Beck,1

Baylor-Johns Hopkins Center for Mendelian Genomics, Eric A. Boerwinkle,1,11,20 Richard A. Gibbs,1,11

James R. Lupski,1,8,11,12,21,* Jordan S. Orange,2,3,8,21 Ekkehart Lausch,10,21 and I. Celine Hanson3,8,21

Human phosphoglucomutase 3 (PGM3) catalyzes the conversion of N-acetyl-glucosamine (GlcNAc)-6-phosphate into GlcNAc-1-phos-

phate during the synthesis of uridine diphosphate (UDP)-GlcNAc, a sugar nucleotide critical to multiple glycosylation pathways. We

identified three unrelated children with recurrent infections, congenital leukopenia including neutropenia, B and T cell lymphopenia,

and progression to bone marrow failure. Whole-exome sequencing demonstrated deleterious mutations in PGM3 in all three subjects,

delineating their disease to be due to an unsuspected congenital disorder of glycosylation (CDG). Functional studies of the disease-asso-

ciated PGM3 variants in E. coli cells demonstrated reduced PGM3 activity for all mutants tested. Two of the three children had skeletal

anomalies resembling Desbuquois dysplasia: short stature, brachydactyly, dysmorphic facial features, and intellectual disability. How-

ever, these additional features were absent in the third child, showing the clinical variability of the disease. Two children received

hematopoietic stem cell transplantation of cord blood and bonemarrow frommatched related donors; both had successful engraftment

and correction of neutropenia and lymphopenia. We define PGM3-CDG as a treatable immunodeficiency, document the power of

whole-exome sequencing in gene discoveries for rare disorders, and illustrate the utility of genomic analyses in studying combined

and variable phenotypes.

Glycosylation is a ubiquitous posttranslational modifica-

tion essential for the proper functioning of a broad spec-

trum of proteins and lipids. In this process, glycans are

constructed from a cellular pool of activated monosaccha-

rides, the sugar nucleotides. The structural diversity of the

glycans ensures specific and selective molecular interac-

tions. Mammals utilize nine sugar-nucleotide donors for

glycosyltransferases: uridine diphosphate (UDP)-glucose,

UDP-galactose, guanosine diphosphate (GDP)-mannose,

GDP-fucose, UDP-xylose, UDP-glucuronic acid, cytidine

monophosphate (CMP)-sialic acid, UDP-N-acetyl-galactos-

amine (UDP-GalNAc), and UDP-N-acetyl-glucosamine

(UDP-GlcNAc) (Figure S1, available online). Glycans are

1Department of Molecular and Human Genetics, Baylor College of Medicine,

dren’s Hospital, Houston, TX 77030, USA; 3Section of Immunology, Allergy, an

Texas Children’s Hospital, Houston, TX 77030, USA; 4Department of Medica

Microbiology, Oslo University Hospital, 0424 Oslo, Norway; 6Department of M

of Clinical Medicine, University of Oslo, 0318 Oslo, Norway; 8Department of Pe

of Pediatrics, Oslo University Hospital, 0424 Oslo, Norway; 10Department of P

Genome Sequencing Center, Baylor College of Medicine, Houston, TX 77030,

lor College of Medicine, Houston, TX 77030, USA; 13Department of Laboratory

USA; 14Center for Cell and Gene Therapy and Texas Children’s Cancer and Hem

Houston, TX 77030, USA; 15Pediatric Hematology Oncology, University of N

Department of Human Genetics, Emory University, Decatur, GA 30033, USA;18Department of Pediatric Imaging, Tokyo Metropolitan Children’s Medical C

Immunology and Transfusion Medicine, Oslo University Hospital, 0424 Oslo, N

ter, Houston, TX 77030, USA21These authors contributed equally to this work

*Correspondence: strayped@bcm.edu (A.S.-P.), jlupski@bcm.tmc.edu (J.R.L.)

http://dx.doi.org/10.1016/j.ajhg.2014.05.007. �2014 by The American Societ

96 The American Journal of Human Genetics 95, 96–107, July 3, 2014

attached to proteins via a nitrogen atom of an asparagine

(N-linked glycan) or an oxygen atom of a serine or threo-

nine (O-linked glycan). These two major glycosylation

mechanisms in eukaryotic cells differ in the protein targets

and cellular localization. Defects in genes encoding the

formation of sugar nucleotides or different steps of the

glycosylation processes result in the disruption of distinct

glycosylation pathways and might lead to congenital dis-

orders of glycosylation (CDGs).

UDP-GlcNAc, the end product of the hexosamine

biosynthetic pathway (Figure S1), is an activated precursor

for both N-linked and O-linked glycosylation of proteins1,2

and is needed for the generation of glycosaminoglycans,

Houston, TX 77030, USA; 2Center for Human Immunobiology, Texas Chil-

d Rheumatology, Department of Pediatrics, Baylor College of Medicine and

l Genetics, Oslo University Hospital, 0424 Oslo, Norway; 5Department of

edical Biochemistry, Oslo University Hospital, 0424 Oslo, Norway; 7Institute

diatrics, Baylor College ofMedicine, Houston, TX 77030, USA; 9Department

ediatrics, Freiburg University Hospital, 79106 Freiburg, Germany; 11Human

USA; 12Medical Genetics Laboratories, Molecular and Human Genetics, Bay-

Medicine and Pathology, Mayo College of Medicine, Rochester, MN 55905,

atology Centers, Baylor College of Medicine and Texas Children’s Hospital,

ew Mexico, Albuquerque, NM 87106, USA; 16Emory Genetics Laboratory,17Department of Radiology, Oslo University Hospital, 0424 Oslo, Norway;

enter, 2-8-29 Musashidai, Fuchu, Tokyo 183-8561, Japan; 19Department of

orway; 20Human Genetics Center, University of Texas Health Science Cen-

y of Human Genetics. All rights reserved.

proteoglycans, and glycolipids. Specifically, UDP-GlcNAc

is incorporated into N-glycans, O-glycans, and glycosyl-

phosphatidylinositol (GPI)-anchored proteins and is also

a donor for the reversible addition of O-GlcNAc to

proteins, i.e., in proteoglycan synthesis. In the yeast

Saccharomyces cerevisiae, as well as in higher eukaryotes,

phosphoglucomutase 3 (Pgm3) catalyzes an important

step in the synthesis of UDP-GlcNAc: the conversion of

GlcNAc-6-phosphate (GlcNAc-6-P) into GlcNAc-1-phos-

phate (GlcNAc-1-P) (Figure S1).1,3 Homozygous knockout

of Pgm3 in mice is embryonically lethal, whereas homozy-

gous hypomorphic alleles cause trilineage cytopenias

(anemia, thrombocytopenia, and leukopenia), ascribed to

decreased UDP-GlcNAc.3 The human homolog, PGM3

(MIM 172100), synonymously designated phosphoacetyl-

glucosamine mutase 1 (AGM1), is most abundantly ex-

pressed in the pancreas, prostate, and testis and is also

expressed in the bone marrow, placenta, salivary glands,

digestive tract, and liver, but not in lung tissue.4–6 In the

relevant tissue, the protein has mainly been found to

localize to the nucleus and cytoplasm and be associated

with the cytoskeleton.6 Mutations in another phosphoglu-

comutase-encoding gene, PGM1 (MIM 171900), cause hu-

man PGM1-CDG (CDG type It [MIM 614921]; Figure S1),

clinically characterized by growth retardation, hepatop-

athy, myopathy, dilated cardiomyopathy, hypoglycemia,

and the bifid uvula.7–10 Disease-causing hypomorphic

PGM3 mutations have been reported in two young-adult

cohorts with clinical presentation of eczema, recurrent in-

fections, immunoglobulin E (IgE)-mediated disease, bron-

chiectasis, variable degrees of neurocognitive impairment,

kyphoscoliosis, and CD4 or CD8 T cell lymphopenia.11,12

None of the described individuals had severe immune defi-

ciency, bone marrow failure, or skeletal dysplasia.

We describe PGM3-CDG, a CDG detected by whole-

exome sequencing (WES) in three children with a similar

hematological phenotype. Compared to the phenotype

of recently published affected individuals, their unique

clinical and immunological PGM3-CDG phenotype in-

cluded recurrent infections, combined immunodeficiency,

neutropenia with progression to bone marrow failure, and

variable dysmorphic features. Importantly, two individuals

presented with a recognizable skeletal dysplasia phenotype

resembling Desbuquois dysplasia (DBQD [MIM 251450]);

Table 1). DBQD is an autosomal-recessive osteochondro-

dysplasia characterized by growth retardation, short ex-

tremities (rhizomelic and mesomelic shortening), joint

laxity, and progressive kyphoscoliosis. Affected individuals

have facial dysmorphisms, a short neck, shortened tubular

bones with metaphyseal flaring, an exaggerated trochanter

minor of the proximal femur (monkey-wrench malforma-

tion), and advanced bone age. DBQD type 1 includes

hand anomalies such as an extra ossification center distal

to the second metacarpal bone, bifid distal-thumb pha-

lanx, and dislocation of the interphalangeal joints.

Mutations in the gene encoding calcium-activated nucleo-

tidase 1 (CANT1 [MIM 613165]), located at 17q25.3, have

The

been reported in DBQD type 1, but affected individuals

without detected CANT1 mutations suggest genetic het-

erogeneity.13–15 CANT1 functions in proteoglycan meta-

bolism.16,17 Proteoglycan synthesis is also disrupted in

DBQD type 2 as a result of a deficiency in xylosyltransfer-

ase 1 (XYLT1 [MIM 608124]).18 Severe combined immuno-

deficiency (SCID) or other types of congenital immunode-

ficiencies have not been reported in DBQD.

Three unrelated children from distinct world popula-

tions—a female with Afghani parents (P1), a male with

Mexican-American parentage (P2), and another male

from Germany (P3)—were studied (Figure 2). Clinically,

P1 and P3 constitute a distinct third Desbuquois variant

we here propose to classify as DBQD type 3.

Informed consent for research studies was obtained from

the probands, siblings, and parents through protocols

approved by the institutional review boards at Baylor

College of Medicine and Universitatsklinikum Freiburg

and through institutional research protocols approved by

regional ethics committees; all followed the principles

stated in the Declaration of Helsinki. In P3 and his family,

molecular analyses were also performed in accordance

with the German Genetic Diagnosis Act (GenDG), and in

P1, analyses were performed in accordance with the

National Biotechnology Act. Specific parental releases

were obtained from parents for the use of clinical data

(from P1–P3) in this manuscript.

After birth, the female child (P1; subject A.II-2 in

Figure 2) presented with respiratory distress with radio-

graphically verified pneumonias despite antimicrobial

therapy. She had leukopenia with neutropenia and

SCID with low numbers of T lymphocytes and

B cells but normal numbers of natural killer (NK) cells

(T�B�NKþSCID phenotype) and no anemia, thrombocy-

topenia, or splenic anomalies. T cell receptor excision

circles were not low in peripheral blood at birth.19

Lymphocyte subsets as measured by flow cytometry at

older age points revealed decreasing numbers of B cells

(Table S1).19,20 Striking skeletal abnormalities were noted

clinically and by radiography at birth: rhizomelic short-

ening of tubular bones with brachydactyly, short meta-

carpal and metatarsal bones and phalanges, and pectus

carinatum (Figure 1A). Dysmorphic facial features in-

cluded downturned corners of the mouth, midface

hypoplasia, and micrognathia (Figure 1B). Other skeletal

anomalies included bilateral exaggerated trochanter

minor, coronal clefts of the caudal lumbar vertebrae,

and cranial Wormian bones (Figures 1E–1G). She did not

have microcephaly or hydrocephalus, and MRI showed

normal cerebral myelination patterns at 5 months of

age. She had eczematous skin lesions from 2 months of

age. At 4 months of age, she received a hematopoietic

stem cell transplant (HSCT) from a 6/6 matched cord

blood donor, and at 1 year of age (6 months after

HSCT), she had leukocytes within normal ranges and

no infections. She was globally developmental delayed

(4-month stage at 1 year of age).

American Journal of Human Genetics 95, 96–107, July 3, 2014 97

Table 1. Clinical Characteristics of the Three Children Presenting with PGM3-CDG

Individual

P1 P2 P3

Gender female male male

Ethnicity Afghani Mexican German

Parental consanguinity no but same clan no no

Family history one healthy older sibling,no other cases

two deceased older siblings (5 monthsand 7 months old) with a similardisease, one additional older healthysibling

two healthy younger siblings,no other cases

Birth weight (percentilea) 4,135 g (95th ) 4,080 g (90th) 3,015 g (25th)

Birth length (percentilea) 46 cm (5th) 48 cm (10th–25th) 42 cm (4 cm < 2nd)

Birth OFC (percentilea) 38 cm (85th) NA 34 cm (35th)

Follow-up weight 8.7 kg (3rd) at 18 months 23.4 kg (50th–75th) at 6 years NA

Follow-up length 73 cm (2 cm < 2nd) at 18 m 120 cm (50th) at 6 years NA

Follow-up OFC 44 cm (5th) at 12 months NA NA

Skeletal dysplasia with short-limbeddwarfism, brachydactyly, and ‘‘monkey-wrench’’ femora

þ � þ

Pectus carinatum þ � þ

Dysmorphic facial features, downturnedcorners of mouth, midface hypoplasia, andmicrognathia

þ � þ

Developmental delay and/or intellectualdisabilities

þ � þ

Other one evaluation forhydrocephalus, no shuntneeded

� seizures since 4 months,hydrocephalus verified, shuntimplantation at 5 months

Age at onset of infections and/orimmunodeficiency

birth birth birth

T�B�NKþ SCID þ þ þ

Neutropenia þ þ þ

Anemia (þ) � þ

Thrombocytopenia � � �

Splenomegaly � � �

Recurrent respiratory infections, otitismedia, and pneumonia

þ þ þ

Skin infections þ þ þ

Eczema since 2 months since 2–3 months since 2 months

Gastrointestinal problems GERD GERD, persistent diarrhea GERD

Serum immunoglobulins low IgM and IgA, normalIgG and IgE

low IgM, normal IgG and IgA, highIgE (1,233–1,768 kU/l)

normal IgM, IgA, and IgE,low IgG from 3 months

Start age for antibiotics and antifungaltherapy

birth intermittent since birth, prophylacticsince 2.5 years

birth

Start age for immunoglobulin substitution 4 weeks 2.5 years 3 months

Start age for G-CSF injections 3 weeks 1 year 2 months

RBC transfusions (SAG-M) three SAG-M in total � first SAG-M at 6 weeks,repeated every 2–3 weeks

HSCT þ þ �

HSCT recipient age 4 months 6 years no HSCT

(Continued on next page)

98 The American Journal of Human Genetics 95, 96–107, July 3, 2014

Table 1. Continued

Individual

P1 P2 P3

HSCT donor type cord blood, unrelated,6/6 HLA match

HLA-identical sibling no HSCT

HSCT outcome successfully cured successfully cured no HSCT

Age at latest evaluation 1.5 years (living) 6.5 years (living) deceased at 7 months

PGM3 mutations (RefSeq NM_015599.2) c.[737A>G];[737A>G] c.715G>C and chr6.hg19:g.(83,013,454_83,145,962)_(84,389,166_84,395,825)del

c.[737dupA];[1352A>G]

Predicted PGM3 changes p.[Asn246Ser];[Asn246Ser] p.[Asp239His];[0] p.[Asn246Lysfs*7];[Gln451Arg]

Genes tested (with normal results) bySanger sequencing prior to WES

RAG1, RAG2, JAK3, ELANE,and HAX1

NEMO, CD40L, ELANE, HAX1, SBDS,SH2D1A, WAS, FOXP3, MAGT1, STK4,IFNGR1, IFNGR2, CXCR4, and GFI1

CANT1, CHST3, IMPAD1, SBDS,and RMRP

CMA (with normal results) prior to WES CMA Agilent 180K no CMA CMA Agilent 244K

Abbreviations are as follows: CMA, chromosomal microarray; G-CSF, granulocyte colony-stimulating factor; GERD, gastresophageal reflux disease; HLA, humanleukocyte antigen; HSCT, hematopoietic stem cell transplantation; IgA, immunoglobulin A; IgE, immunoglobulin E; IgG, immunoglobulin G; IgM, immunoglob-ulin M; NA, not available; OFC, occipitofrontal circumference; RBC, red blood cell; SAG-M, saline-adenine-glucose-mannitol stored RBC unit, T�B�NKþSCID,severe combined immunodeficiency with a lack of T and B lymphocytes but presence of NK cells; and WES, whole-exome sequencing.aAge percentiles according to WHO Child Growth Standards.

The Mexican-American male (P2; subject B.II-4 in

Figure 2) presented with recurrent infections (upper-respi-

ratory-tract infections, skin abscesses, and chronic otitis

media) soon after birth. Family history was significant for

two older full male siblings (B.II-2 and B.II-3 in Figure 2),

who died early from infection at 7 and 5 months of age,

respectively. These children exhibited persistent vomiting,

failure to thrive, pneumonia, and eczema. Neutropenia

was documented in only a single sibling. Neither P2 nor

his siblings had evidence of skeletal abnormalities, e.g.,

they had a normally configured thorax and proportionate

stature without facial dysmorphic features. This boy had

infancy-onset eczema and IgE-mediated food allergy with

associated anaphylaxis. Neutropenia diagnosed at 1 year

of age was responsive to granulocyte colony-stimulating

factor (G-CSF) therapy. NK cells, platelets, and red blood

cell (RBC) counts were normal, but his disease progressed

with loss of peripheral blood B and T cells (Table S2).

When he was 5 years old, his bone marrow aspirate was

hypocellular, had a 40% reduction of cellularity in compar-

ison to previous bone marrow samples, and showed evi-

dence of bonemarrow failure. At 6 years of age, he received

a matched-related HSCT from his human-leukocyte-

antigen-identical healthy brother. This boy had successful

engraftment and resolution of his neutropenia and return

of lymphocyte function. He currently attends school and

has mild speech delay but normal cognition.

Like the female child (P1), the most severely affected

child (P3; subject C.II-1 in Figure 2) had clinically apparent

DBQD-like disease. Short limbs and a small thoracic diam-

eter were noted on fetal ultrasound, and short-limbed

dwarfism and brachydactyly, along with pectus carinatum

and facial dysmorphism, were diagnosed after birth (Fig-

ures 1C and 1D). Skeletal radiographs demonstrated short

The

tubular bones, several phalangeal and tarsal dislocations

(Figure 1J), short femoral necks with metaphyseal beaking,

and exaggerated lesser trochanters (Figure 1K). P3 also had

leukopenia from birth, neutropenia with reduced response

to G-CSF, and a T�B�NKþSCID phenotype (Table S3) with

recurrent and severe infections. From 2 months of age, he

had a eczematous scalp and intertriginous skin lesions.

At 3 months, he required ventilatory and circulatory

support after influenza and coincident generalized bacte-

rial infection. He developed intermittent tonic seizures

with hypsarrhythmia on electroencephalography, and

MRI demonstrated internal and external hydrocephalus,

delayed myelination, and periventricular white-matter

lesions. He had complex neurological deterioration and

died at 7 months of age from overwhelming infection.

In summary, all three children had recurrent infections

since birth, congenital neutropenia, and a combined

immunodeficiency characterized by low numbers of

T cells, an increased CD4/CD8 ratio, progressive loss of

B cells with age, and persistently normal NK cells (Tables

S1–S3). None of them had thrombocytopenia or signifi-

cant anemia. Two children (P1 and P3) had skeletal

anomalies consistent with DBQD (Figure 1). The

Mexican-American boy (P2) had two older male siblings

(one with neutropenia and most likely the same disorder

and neither with skeletal dysplasia) who died in infancy

from infection. Two children (P1 and P2) were successfully

treated with HSCTwith neutrophil and T and B cell correc-

tion. The most severely affected child (P3) died prior to

receiving a transplant.

WES was performed on all three individuals with the use

of genomic DNA extracted from whole blood prior to

HSCT.21,22WES for two subjects (P1 and P2) was performed

at the Baylor College of Medicine Human Genome

American Journal of Human Genetics 95, 96–107, July 3, 2014 99

Figure 1. Dysmorphic Features andRadiological Manifestations in P1 and P3(A) P1 at 1 month of age. Note the shortfingers and short long bones (rhizomelic)and the bell-shaped thorax with pectuscarinatum.(B) P1 at 1 year of age. Note the dysmor-phic facial features, including a down-turned mouth, midface hypoplasia, andmicrognathia. The pectus carinatum is stillprominent, but body stature appears lessdisproportionate. The child is unable tosit by herself.(C and D) Neonatal photographs of P3show strikingly similar dysmorphic fea-tures and physical findings.(E and F) Skull X-rays of P1 at 3 months ofage demonstrate cranial Wormian bones(arrows).(G) A hand X-ray of P1 at 3 months of ageshows the extra ossification center, thepseudoepiphysis (arrow), proximal to thesecond metacarpal bone.(H) A hip X-ray of the pelvic bones andfemora of P1 at 3 months of age demon-strates the exaggerated trochanter minor(monkey-wrench appearance) of the prox-imal femur on both sides (arrows).(I–K) Neonatal radiographs of P3 demon-strate (I) no Wormian bones in the skull,(J) severe brachydactyly and phalangealdislocations, but no advanced skeletalmaturation or extra ossicles in the carpo-gram (also not present at 3 months), and(K) typical monkey-wrench morphology(arrows) of the proximal femora; there ismetaphyseal flaring at the distal femoralends.Specific parental releases were obtainedfrom parents for the use of photographs(of P1 and P3) in this manuscript.

Sequencing Center (BCM-HGSC) as part of the Baylor-

Johns Hopkins Center for Mendelian Genomics. WES

testing in P3 was performed by the Department of Pediat-

rics at Freiburg University Hospital with a patient-parent

trio design on a SOLiD5500xl platform as previously

described.23 The statistical summary of variants detected

by WES is summarized in Table S4. The candidate dis-

ease-associated variants identified by WES were indepen-

dently confirmed by Sanger sequencing and analyzed for

familial segregation (primers in Table S5). The WES

method used at BCM-HGSC, including the HGSC CORE

design, has been described.21,22,24–29 Annotation data

were added to the variant-call-format file with a suite of

annotation tools, designated ‘‘Cassandra.’’30 Rare variants

were selected on the basis of the NHLBI Exome Sequencing

Project (ESP) Exome Variant Server, 1000 Genomes (as

of October 2013), and two in-house-generated data-

bases that include results of whole-exome-sequenced sam-

ples from ~4,000 (Arterosclerosis Risk in Communities

[ARIC]) and >200 (Baylor College of Medicine Center for

Mendelian Genomics [BCM-CMG]) different individuals.

Variants of interest were selected on the basis of both

100 The American Journal of Human Genetics 95, 96–107, July 3, 201

rarity and evaluation by the following prediction tools:

PhyloP, SIFT, PolyPhen-2, likelihood-ratio test (LRT), and

MutationTaster. In addition, knowledge of gene function,

pathways, and expression patterns and results from other

model systems were considered. Prior to evaluation of

genes absent from publically available databases, all

variants in exonic and the captured intronic regions

of HGMD and disease-related OMIM genes were evaluated.

Bioinformatic prediction of copy-number variants (CNVs)

from WES data were based on BAM files analyzed by the

Integrative Genomics Viewer (IGV) and CoNIFER.31 The

Baylor College of Medicine chromosomal microarray

(BCM CMA) used in P2 was a custom-designed genome-

wide Agilent oligoarray (BCM CMA version 10) with

exon coverage of 4,200 genes, including PGM3 and 300

genes known to bemutated in various primary immunode-

ficiency diseases.32,33 This CMA readily identifies intra-

genic CNV alleles as recessive carrier states.34–36 For P1

and P3, standardized Agilent oligoarrays 180K and 244K

were performed as part of the diagnostic workup.

WES studies of P1 and P2 identified PGM3 as a candidate

gene, and adding the WES results from P3 allowed us to

4

Figure 2. PedigreesThe families of P1 (A), P2 (B), and P3 (C) are shown with segregation of mutant alleles. Partial chromatograms of Sanger confirma-tion analysis of PGM3 mutations are shown only for the probands; arrows indicate respective nucleotide changes. For homozygousand hemizygous mutations, normal control sequences are given below the mutated allele. Also below the pedigree in (B) isthe result from the chromosomal microarray; it shows the probes in the deleted region (red). The 6q14.1–q14.2 deletion(chr6.hg19:g.(83,013,454_83,145,962)_(84,389,166_84,395,825)del) encompassed PGM3 and three neighboring OMIM genes(UBE3D, ME1, and SNAP91) on P2’s paternal allele. Samples from his deceased affected brothers were not available for genetic testing.Abbreviations are as follows: del, 1.2 Mb deletion CNV; and WT, wild-type.

conclude that the described clinical phenotype in-

cluding the DBQD-like skeletal dysplasia was most likely

due to mutations in PGM3 (MIM 172100). None of the

single-nucleotide variants (SNVs) detected in PGM3 in

the three children had been reported in the NHLBI ESP

Exome Variant Server, 1000 Genomes, dbSNP, ARIC, or

BCM-CMG.

Exome sequencing of the female child (P1) identified

a homozygous nonsynonymous SNV in exon 6 of

PGM3. Asparagine was replaced by serine at residue 246

(c.737A>G [p.Asn246Ser]; RefSeq accession number

NM_015599.2]). Sanger sequencing confirmed the mis-

sense variant in the proband, heterozygous carrier status

in both parents, and homozygous wild-type status in the

healthy older sister (Figure 2A). The variant-prediction

programs (SIFT, PolyPhen-2, LRT, MutationTaster, and

PhyloP) evaluated the variant as most likely disease

causing and the affected location as conserved. The

amino acid Asn246 is highly conserved across species

(Figure S2). Although there was no known parental

consanguinity, the parents were from the same clan,

and two large genomic intervals with absence of heterozy-

gosity surrounding PGM3were observed on chromosome 6

only (Figure S3). Neither CANT1 mutations nor CHST3

(MIM 603799) disease-causing mutations were detected

in the WES results of P1. The WES coverage of these

two genes was adequate, except for exons 1 and 2 of

CANT1 (RefSeq NM_00159772.1) and exon 1 of CHST3

(RefSeq NM_004273.4), which were subsequently Sanger

sequenced.

In P2, WES identified a nonsynonymous SNV also

located in exon 6 in PGM3. This rare variant was confirmed

by Sanger sequencing (Figure 2B) and is predicted to cause

aspartic acid replacement by histidine at position 239

(c.715G>C [p.Asp239His); RefSeqNM_015599.2; Figure 3).

The A

Prediction programs (SIFT, PolyPhen-2, LRT, PhyloP, and

MutationTaster) evaluated the variant as most likely

disease causing, and the affected amino acid is con-

served across species (Figure S2). Asp239 is located seven

amino acids from both the active site and Asn246,

altered in P1. In segregation analyses, only the mother

was heterozygous for the SNV; however, bioinformatic

interpretation of WES (IGV and CoNIFER) provided sug-

gestive evidence of a deletion CNV involving PGM3

(Figure S3), suggesting that the proband was hemizygous

for the SNV. Chromosomal microarray (BCM CMA ver-

sion 10) confirmed a 1.2 Mb deletion of 6q14.1–q14.2,

chr6.hg19:g.(83,013,454_83,145,962)_(84,389,166_84,395,

825)del, involving the entire PGM3 and three neighboring

OMIM genes (UBE3D [MIM 612495], ME1 [MIM 154250],

and SNAP91 [MIM 607923]) on the boy’s paternal allele

(Figure 2B). The deletion was inherited from his healthy

father (individual B.I-1 in Figure 2). Samples from the

deceased affected brothers were not available for genetic

testing.

In P3, who had the most severe phenotype, com-

pound-heterozygous variants were detected in PGM3.

In exon 6, we detected a 1 bp duplication (c.737dupA

[p.Asn246Lysfs*7]; RefSeq NM_015599.2). This duplica-

tion is predicted to cause nonsense-mediated degradation

of the mutant mRNA. Accordingly, quantitative real-time

PCR analysis (primers in Table S5) of PGM3 transcripts in

P3’s blood cells revealed a reduction to approximately

50% (Figure S4), which is in keeping with degradation of

the mRNA with a premature stop codon.37 On the other

allele, a missense mutation was detected in exon 11

(c.1352A>G [p.Gln451Arg]; RefSeq NM_015599.2); this is

predicted to be disease causing (deleterious according to

MutationTaster and PolyPhen but tolerated according to

SIFT). The affected amino acid is moderately conserved

merican Journal of Human Genetics 95, 96–107, July 3, 2014 101

Figure 3. Homology Model of Human PGM3(A) Structural overview of human PGM3. The three amino acidsAsp239 (D239), Asn246 (N246), and Gln451 (Q451) are coloredred and depicted in ball-and-stick representation. The greensphere represents the magnesium (Mg) ion.(B–D) Close-up view of the interaction between the altered aminoacids and their surroundings. (B) Interaction between Asp239(D239) and Arg222 (R222). (C) Interactions between Asn246(N246) and the active serine loop and between Asn246 and themetal-binding loop. (D) The environment around amino acidGln451 (Q451). The substrate GlcNAc-6-P is shown in magenta.Other amino acids are as follows: S64, Ser64; H65, His65; N66,Asn66; and P67, Pro67.

across species (Figure S2). Family segregation by Sanger

sequencing confirmed that the variants were trans-alleles,

fulfilling Mendelian expectations (Figure 2C).

The crystalline structure of human PGM3 has not been

determined; therefore, we performed structural analysis

of the three amino acid substitutions (p.Asp239His,

p.Asn246Ser, and p.Gln451Arg) in human PGM3 (UniProt

ID O95394) on the basis of a homology model made by

SWISS-MODEL38 and provided by the Protein Model

Portal.39 The homology model of human PGM3 (Figure 3)

is based on the experimental X-ray structure of Aspergillus

fumigatus Pgm340 (Protein Data Bank [PDB] ID 4BJU),

which has ~50% sequence identity with the human pro-

tein. A model of PGM3 in complex with the substrate

GlcNAc-6-P was obtained by superposition with the

Candida albicans (C. albicans) Pgm3-GlcNAc-6-P com-

plex41 (PDB ID 2DKC). As for the other members of this

superfamily, PGM3 consists of four domains (Figure 3A),

and the active site of the protein is made up of one loop

from each of the four: the serine loop in domain 1, the

metal-binding loop in domain 2, the sugar-binding loop

102 The American Journal of Human Genetics 95, 96–107, July 3, 201

in domain 3, and the phosphate-binding loop in domain 4.

Asn246 and Asp239 are both located on the same loop in

domain 2 and are positioned close to the active site (Fig-

ures 3A–3C). Asn246 is directed toward the active site

and most likely stabilizes both the serine loop and the

metal-binding loop by two and one hydrogen bond,

respectively (Figure 3C). The substitution of this amino

acid with serine is predicted to abolish these intramolecu-

lar interactions and result in an overly flexible active site.

Furthermore, the model suggests that Asp239 participates

in a hydrogen bond with Arg222, which also is likely to

have a stabilizing effect (Figure 3B). Finally, Gln451 is

located in domain 4 and is directed toward the substrate

GlcNAc-6-P (Figure 3D). If residue Gln451 is changed to

an arginine, both the alteration in electrostatic charge,

from neutral to positive, and the longer side chain will

most likely influence the interaction between the protein

and its substrate.

Clinical screening for disorders of glycosylation did not

show abnormalities. Capillary-zone electrophoresis (CZE)

of serum, taken both before and after HSCT, demonstrated

normal serum transferrin in the reported female child (P1)

in a sample collected at 1 month of age (before immuno-

globulin substitution and RBC transfusion), at 2 months

of age, and after HSCT. Likewise, mass spectrometry anal-

ysis showed normal serum transferrin and apolipopro-

tein-CIII (apoCIII) glycoforms from all serum samples

collected in this child. For CZE analysis, the serum trans-

ferrin sialoform pattern was examined by capillary electro-

phoresis (P/ACE-SYSTEM MDQ). Serum transferrin and

apoCIII glycoforms were analyzed by simultaneous online

immunoaffinity chromatography electrospray ionization

mass spectrometry (SCIEX API4000 tandem mass spec-

trometer with Turbo V spray source).31 Total serum glycan

was analyzed after the samples were denatured and di-

gested to release the N-glycans prior to clean up. After clean

up, and permethylation with iodomethane, N-glycans

were analyzed with MALDI-TOF mass spectrometry.42,43

Serum T antigen (T) and sialylated T antigen (ST) were

quantified with MALDI-TOF. The T/ST ratios were normal

in all P1’s serum samples (Table S6). In the male Mexican-

American boy (P2), N-glycan transferrin analysis was qual-

itatively and quantitatively normal, and the apoCIII profile

and T/ST ratio were normal in serum collected after HSCT.

Serum was not available for glycosylation testing prior

to HSCT. Finally, for the most severely affected male (P3),

CZE of serum sample from 1 and 4 months of age demon-

strated normal N-glycosylation of transferrin.

In order to examine the effect of the p.Asn246Ser,

p.Asp239His, and p.Gln451Arg substitutions on PGM3

activity, we generated recombinant clones containing

the mutant alleles, expressed them in Escherichia coli

(E. coli), and subsequently purified the recombinant

PGM3 proteins. Indeed, all three proteins demonstrated

reduced phosphate-group transfer from position GlcNAc-

6-P to GlcNAc-1-P, reflected by reduced substrate con-

sumption (Table 2). The p.Asn246Ser substitution was

4

Table 2. PGM3 Activities

PGM3No. ofExperiments Mean (%) SEM (%)

Wild-type 5 100 0.0

p.Asn246Ser 5 1 8.1

p.Asp239His 5 59 11.8

p.Gln451Arg 5 50 10.0

PGM3 activities were assayed in a 200 ml standard reaction mixture containing50 mM Tris-HCl (pH 8.0), 5 mM MgCl2, 10% (v/v) glycerol, 200 mM GlcNAc-6-P substrate, and 50 mg of the indicated PGM3 at 30�C for 10 min. Reactionswere then inactivated by incubation at 80�C for 5 min. The effect of the aminoacid substitutions on PGM3 was tested by mass spectrometry in ‘‘multiple re-action monitoring’’ mode; the transition from the molecular ion (m/z 300)to a fragment specific to the substrate (GlcNAc-6-P) (m/z 138) was used formeasuring substrate consumption in relation to that of the wild-type. Datawere calculated from five independent experiments.

completely inactive even with 4-fold higher enzyme

concentrations.

Our experiments in conjunction with recent reports

document that that PGM3 mutations in humans cause a

phenotypically variable CDG.11,12 From extensive clinical

studies in three families reported here, the phenotype of

PGM3-CDG might include dysmorphic facial features,

cognitive impairment, leukopenia, T�B�NKþSCID and

neutropenia, and skeletal dysplasia. Our subjects had

more pronounced immunodeficiency, including severe

neutropenia (Tables S1–S3). Only P2 had elevated serum

IgE levels, whereas the affected individuals reported by

the two other groups all had hyper IgE syndrome, and

none of them progressed to bone marrow failure.11,12

Other disorders of glycosylation are also known to cause

a wide phenotypic spectrum, from mild to severe pheno-

types.44 Both genotype-phenotype correlation and other

modifying factors, including immunodeficiency, intellec-

tual disabilities, and skeletal dysplasia, might contribute

to the variety of phenotypes observed in CDGs. The differ-

ences observed between our three subjects might be related

to specific genotypes, i.e., p.Asn246Ser causes a more

pronounced block of enzyme activity in comparison to

hypomorphic variants such as p.Asp239His, as demon-

strated by our mutant model testing. The data from enzy-

matic activity of p.Asn246Ser and p.Asp239His correspond

with the findings in the Pgm3 mouse models, where loss-

of-function mutants confer more severe outcomes.3 How

and why p.Gln451Arg conferred substrate consumption

similar to that of p.Asp239His in model testing, whereas

P3 had the same severe skeletal phenotype as P1, will

require further investigation. Our protein homology

model predicts that p.Gln451Arg, the variant detected in

P3, might alter the substrate affinity, and high-affinity

binding of substrate might completely block the conver-

sion to product GlcNAc-1-P, even if our mutant model

testing showed half enzyme activity with regard to sub-

strate consumption. Of the six variants reported by the

other two research groups, five (p.Leu83Ser, p.Asp325Glu,

p.Asp502Tyr, p.Glu529Gln, and p.Glu340del) were dem-

The A

onstrated to be hypomorphic with sustained enzyme

activity.11,12

Other in vivo modifying factors, such as infections and

autoinflammatory changes, might also have contributed

to the severity and phenotypic spectrum observed in the

PGM3-CDG individuals. Compared to the other disorders

of carbohydrate metabolism associated with immunodefi-

ciency, such as SLC37A4-CDG (severe congenital neutro-

penia type 4 [MIM 612541], caused by mutations in

G6PC3 [MIM 611045]) and SLC35C1-CDG (CDG type IIc

[MIM 266265], caused by mutations in SLC35C1 [MIM

605881]), the involvement of lymphocytes in addition

to neutrophils demonstrates expanded immunological

impact. Unlike ALG12-CDG (CDG type Ig [MIM

607143])-affected individuals, who have B cell deficiency

and a defect in N-glycosylation, our PGM3-CDG probands

did not show any alterations in serum transferrin or

apoCIII profiles (data not shown) or T/ST ratios (Table

S6), leading us to conclude that N- and O-glycosylation

in the liver for transferrin and apoCIII is not affected by

the PGM3 defect. Thus, this might reflect the tissue- and

organ-specific functions of PGM3, as well as a critical role

for certain hematopoietic lineages. Another explanation

for the normal apoCIII profile is that it only tests core 1

O-glycosylation, which is not dependent on UDP-GlcNAc.

The first monosaccharide attached in the synthesis of

O-linked glycans is GalNAc. A core 1 structure is generated

by the addition of galactose. A core 2 structure is generated

by the addition of GlcNAc to the GalNAc of the core 1

structure. Core 3 and core 4 structures are generated by

the addition of a single GlcNAc to the original GalNAc

and by the addition of a second GlcNAc to the core 3 struc-

ture, respectively. Hence, formation of cores 2–4 is depen-

dent on UDP-GlcNAc. Reduced levels of UDP-GlcNAc in

C. albicans do not block N-glycosylation but might cause

reduced and shorter glycosylation branching,45,46 which

wewere unable to demonstrate in the sera from affected in-

dividuals. Both Sassi et al. and Zhang et al. reported a

normal transferrin pattern in their subjects with PGM3

mutations. Zhang et al. reported high serum T antigen

levels and elevated T/ST ratios, but we could not demon-

strate the same.12 Interestingly, Sassi et al. detected reduced

bi-, tri-, and tetra-antennary N-glycan branching in

leukocytes from affected individuals and suggested a geno-

type-phenotype correlation on the basis of their study sub-

jects with three different mutations.11 Homozygotes for

p.Glu340del had the most altered glycosylation branching

pattern. Our study subjects had either received bone

marrow transplantation or died before the CGD diagnosis

was made, and no leukocytes were available for further

functional studies.

The skeletal abnormalities, including X-ray findings, in

two children (P1 and P3) resembled DBQD; however, the

linear growth was less restricted in the subject with trans-

planted PGM3 deficiency than has previously been re-

ported in classical DBQD.13 Unlike in classical DBQD,

bone age was not advanced in P3. Another difference of

merican Journal of Human Genetics 95, 96–107, July 3, 2014 103

note in these children with PGM3-CDG is the extra ossifi-

cation center placed proximal rather than distal to the sec-

ond metacarpal bone, as is typical in DBQD type 1. The

similarities between the skeletal dysplasia in CANT1 defi-

ciency and PGM3 deficiency might be due to their com-

mon effect on the proteoglycan synthesis. Others have

shown that CANT1 deficiency causes reduced availability

of UDP-GlcNAc and thereby reduced glycosyltransferase

activities.13 Proteoglycan synthesis is also disrupted in

DBQD2 as a result of xylosyltransferase 1 (XYLT1 [MIM

608124]) deficiency.18 Occurrences of skeletal dysplasia

have been reported in other CDGs, such as PMM2-CDG

(CDG type Ia [MIM 212065], caused by mutations in

PMM2 [MIM 601785]),9,10,47 ALG6-CDG (CDG type Ic

[MIM 603147], caused by mutations in ALG6 [MIM

604566]),48 ALG12-CDG (caused by mutations in ALG12

[MIM 607144]),49 COG1-CDG (CDG type IIg [MIM

611209], caused by mutations in COG1 [MIM 606973]),50

COG7-CDG (CDG type IIe [MIM 608779], caused bymuta-

tions in COG7 [MIM 606978]),51,52 and TMEM165-CDG

(CDG type IIk [MIM 614727], caused by mutations in

TMEM165 [MIM 614726]).53 The kyphoscoliosis noted in

one-quarter of the individuals with PGM3 mutations re-

ported by Sassi et al. and Zhang et al. might represent

the milder spectrum of the PGM3-related skeletal

dysplasia.11,12 Zhang et al. reported decreased PGM3 cata-

lytic activity and reduced intracellular UDP-GlcNAc levels

in fibroblasts from three of their PGM3-CDG-affected indi-

viduals. The severity and variability of the phenotypes

observed between the persons with different PGM3 muta-

tions might be directly correlated with UDP-GlcNAc levels.

A genotype-phenotype correlation has been shown in

mice: mice compound heterozygous for mutant Pgm3

alleles with a radical effect on the protein or a loss-of-func-

tion allele are more severely affected than mice homozy-

gous for a mild mutant allele.3 Pgm3�/� mice die in early

embryogenesis and show a dramatic reduction in UDP-

GlcNAc. With defects restricted to the salivary glands,

pancreas, testis, kidney, and hematopoietic cells, mice

with partial PGM3 activity are viable. Mice with partial

PGM3 deficiency have profound B cell defects (a normal

number of naive B cells but a loss of mature B cells), an

increased CD4/CD8 ratio, and mild anemia and thrombo-

cytopenia but normal numbers of neutrophils, eosino-

phils, and monocytes.3 Whether PGM3-CDG individuals

presenting in infancy with immunodeficiency will develop

the same symptoms as mice (such as male infertility,

exocrine pancreatic insufficiency, and glomerulonephritis)

is questionable. For instance, neutropenia was a profound

hallmark in these PGM3-CDG-affected children, but not in

the mouse model or the described cohorts of older individ-

uals. The potential neurological abnormalities in PGM3-

CDG remain to be defined, and whether it is present at

birth and/or progressive with age is unclear. Interestingly,

brain microglia are derived from hematopoietic precursors

of mesoderm origin and can be replaced by blood-derived

monocytes. It is currently not clear which of the other dis-

104 The American Journal of Human Genetics 95, 96–107, July 3, 201

ease manifestations, in addition to the hematological

defects, HSCT rescues. Skeletal dysplasia has not been re-

ported in mouse models but is perhaps reflected by the

growth restriction observed in mutant mice.3 Genetic

studies, including maternal and zygotic loss-of-function

screens in Drosophila, have revealed that mutations in

nesthocker (nst), the fruit fly’s PGM3 ortholog, block meso-

dermal and tracheal development. This is interesting

because the mesoderm gives rise to bone, cartilage, and

hematopoietic precursors, including microglia. Embryos

lacking maternal and zygotic nst products show low

amounts of intracellular UDP-GlcNAc and defective

O-GlcNAcylation of fibroblast growth factor receptor

(FGFR)-specific adaptor protein, which impairs FGFR-

dependent migration of mesodermal and tracheal cells.54

The identification of a role for nst in FGFR signaling is

compelling in the light of the skeletal dysplasia observed

in our affected subjects.

Some CDGs are treatable with supplements of substrates

for the defective glycosylation pathway. For instance, indi-

viduals deficient in mannose-6-phosphate isomerase (Fru-

6-P to Man-6-P conversion) lack sufficient Man-6-P for

complete physiologic N-glycosylation, and daily supple-

ments of mannose can correct this glycosylation defi-

ciency.55 Given that PGM3 catalyzes an important step

in the synthesis of glycans, it is possible that substitution

of a compound that enhances the enzymatic reaction per-

formed by PGM3 (GlcNAc-6-P to GlcNAc-1-P conversion)

or bypasses the block, such as N-acetyl-galactosamine

(GalNAc; Figure S1), might ameliorate the pathologic

phenotype. For example, supplemental therapy with

galactose was recently found to be effective for the homol-

ogous disease PGM1-CDG (Figure S1).44 However, treat-

ment with HSCT is lifesaving because it corrects the

immunodeficiency, and two of our described children

were successfully cured, whereas the others (P3 and the

two affected brothers of P2) died of infectious complica-

tions (presumably resulting from immunodeficiency)

before transplantation was initiated.

We provide evidence that PGM3 mutations in children

can cause a CDG with leukopenia, skeletal dysplasia, dys-

morphic facial features, and cognitive impairment. The

immunological abnormality is further defined as severe

neutropenia, T and B cell lymphopenia, and progression

to complete bone marrow failure. Genotype together with

other modifying factors might contribute to the pheno-

typic variation and disease severity observed in this CDG,

as is the case in other glycosylation disorders. Our study

demonstrates PGM3-CDG as a severe infancy-onset immu-

nodeficiency in which HSCT is lifesaving and defines the

power of WES in gene discoveries for rare disorders.

Supplemental Data

Supplemental Data include three figures and six tables and can be

found with this article online at http://dx.doi.org/10.1016/j.ajhg.

2014.05.007.

4

Acknowledgments

The authors are grateful to the families for their participation

in this study. Special thanks go to Eric A. Smith for his technical

contributions. The study was performed at the Baylor-Johns

Hopkins Center for Mendelian Genomics, funded by the NIH

National Human Genome Research Institute (U54HG006542

and U54HG003273). The Centers for Mendelian Genomics repre-

sents a cooperative, international research effort to determine the

genetic cause(s) of Mendelian disorders. The German branch of

this study was supported by a grant from the German Federal Min-

istry of Education and Research to E.L. (FACE consortium TP1,

01GM1109A). E.L. was also supported by the European Commis-

sion Seventh Framework Programme (the SYBIL consortium, grant

agreement 602300). P.H.B. was supported by the South-Eastern

Norway Regional Health Authority’s Technology Platform for

Structural Biology and Bioinformatics (grant 2012085). J.R.L.

holds stock ownership in 23andMe Inc. and is a coinventor on

multiple United States and European patents related to molecular

diagnostics. The Department of Molecular and HumanGenetics at

Baylor College of Medicine derives revenue from molecular ge-

netic testing offered in the Medical Genetics Laboratories.

Received: January 14, 2014

Accepted: May 16, 2014

Published: June 12, 2014

Web Resources

The URLs for data presented herein are as follows:

1000 Genomes Browser, http://browser.1000genomes.org/index.

html/

Arteriosclerosis Risk in Communities (ARIC) Study, http://www2.

cscc.unc.edu/aric/

Baylor-Hopkins Center for Mendelian Genomics, https://

mendeliangenomics.org/

Centers for Mendelian Genomics, http://www.mendelian.org/

dbGaP, http://www.ncbi.nlm.nih.gov/gap/

dbSNP, http://www.ncbi.nlm.nih.gov/projects/SNP/

Enzyme Nomenclature, http://www.chem.qmul.ac.uk/iubmb/

enzyme/

HUGO Gene Nomenclature Committee (HGNC), http://www.

genenames.org/

The Human Protein Atlas, PGM3, http://www.proteinatlas.org/

ENSG00000013375/tissue/

Integrative Genomics Viewer (IGV), http://www.broadinstitute.

org/igv/

Likelihood-ratio test, http://www.genetics.wustl.edu/jflab/lrt_

query.html

Medical Genetics Laboratories at Baylor College of Medicine,

http://www.bcm.edu/geneticlabs/

MutationTaster, http://www.mutationtaster.org

NHLBI Exome Sequencing Project (ESP) Exome Variant Server,

http://evs.gs.washington.edu/EVS/

Online Mendelian Inheritance in Man (OMIM), http://www.

omim.org/

PhenoDB, https://mendeliangenomics.org/

PhyloP, http://compgen.bscb.cornell.edu/phast/

PolyPhen-2, http://genetics.bwh.harvard.edu/pph2/

Protein Data Bank (PDB), http://www.rcsb.org/pdb/home/

home.do

RefSeq, http://www.ncbi.nlm.nih.gov/RefSeq

The A

SIFT, http://sift.jcvi.org

UCSC Genome Browser, http://genome.ucsc.edu/

UniProt, http://www.uniprot.org/

WHOChild Growth Standards, http://www.who.int/childgrowth/

en/

Accession Numbers

The PhenoDB accession numbers for the phenotype data reported

in this paper are BH3596 in P1 and BH2704 in P2.

References

1. Freeze, H.H. (2013). Understanding human glycosylation dis-

orders: biochemistry leads the charge. J. Biol. Chem. 288,

6936–6945.

2. Freeze, H.H., and Elbein, A.D. (2009). Glycosylation Precur-

sors. In Essentials of Glycobiology, A. Varki, R.D. Cummings,

J.D. Esko, H.H. Freeze, P. Stanley, C.R. Bertozzi, G.W. Hart,

and M.E. Etzler, eds. (Cold Spring Harbor: Cold Spring Harbor

Laboratory Press), pp. 47–61.

3. Greig, K.T., Antonchuk, J., Metcalf, D., Morgan, P.O., Krebs,

D.L., Zhang, J.G., Hacking, D.F., Bode, L., Robb, L., Kranz,

C., et al. (2007). Agm1/Pgm3-mediated sugar nucleotide syn-

thesis is essential for hematopoiesis and development. Mol.

Cell. Biol. 27, 5849–5859.

4. Pang, H., Koda, Y., Soejima, M., and Kimura, H. (2002).

Identification of human phosphoglucomutase 3 (PGM3) as

N-acetylglucosamine-phosphate mutase (AGM1). Ann. Hum.

Genet. 66, 139–144.

5. Li, C., Rodriguez, M., and Banerjee, D. (2000). Cloning and

characterization of complementary DNA encoding human

N-acetylglucosamine-phosphate mutase protein. Gene 242,

97–103.

6. Uhlen, M., Oksvold, P., Fagerberg, L., Lundberg, E., Jonasson,

K., Forsberg, M., Zwahlen, M., Kampf, C., Wester, K., Hober,

S., et al. (2010). Towards a knowledge-based Human Protein

Atlas. Nat. Biotechnol. 28, 1248–1250.

7. Perez, B., Medrano, C., Ecay, M.J., Ruiz-Sala, P., Martınez-

Pardo, M., Ugarte, M., and Perez-Cerda, C. (2013). A novel

congenital disorder of glycosylation type without central ner-

vous system involvement caused by mutations in the phos-

phoglucomutase 1 gene. J. Inherit. Metab. Dis. 36, 535–542.

8. Timal, S., Hoischen, A., Lehle, L., Adamowicz, M., Huijben, K.,

Sykut-Cegielska, J., Paprocka, J., Jamroz, E., van Spronsen, F.J.,

Korner, C., et al. (2012). Gene identification in the congenital

disorders of glycosylation type I by whole-exome sequencing.

Hum. Mol. Genet. 21, 4151–4161.

9. Jaeken, J., Hennet, T., Freeze, H.H., and Matthijs, G. (2008).

On the nomenclature of congenital disorders of glycosylation

(CDG). J. Inherit. Metab. Dis. 31, 669–672.

10. Jaeken, J., Hennet, T., Matthijs, G., and Freeze, H.H. (2009).

CDG nomenclature: time for a change!. Biochim. Biophys.

Acta 1792, 825–826.

11. Sassi, A., Lazaroski, S., Wu, G., Haslam, S.M., Fliegauf, M.,

Mellouli, F., Patiroglu, T., Unal, E., Ozdemir, M.A., Jouhadi,

Z., et al. (2014). Hypomorphic homozygous mutations in

phosphoglucomutase 3 (PGM3) impair immunity and in-

crease serum IgE levels. J. Allergy Clin. Immunol. 133, 1410–

1419, e13.

12. Zhang, Y., Yu, X., Ichikawa, M., Lyons, J.J., Datta, S., Lamborn,

I.T., Jing, H., Kim, E.S., Biancalana, M., Wolfe, L.A., et al.

merican Journal of Human Genetics 95, 96–107, July 3, 2014 105

(2014). Autosomal recessive phosphoglucomutase 3 (PGM3)

mutations link glycosylation defects to atopy, immune defi-

ciency, autoimmunity, and neurocognitive impairment.

J. Allergy Clin. Immunol. 133, 1400–1409, e5.

13. Faivre, L., Le Merrer, M., Zerres, K., Ben Hariz, M., Scheffer, D.,

Young, I.D., Maroteaux, P., Munnich, A., and Cormier-Daire,

V. (2004). Clinical and genetic heterogeneity in Desbuquois

dysplasia. Am. J. Med. Genet. A. 128A, 29–32.

14. Huber, C., Oules, B., Bertoli, M., Chami, M., Fradin, M., Ala-

nay, Y., Al-Gazali, L.I., Ausems, M.G., Bitoun, P., Cavalcanti,

D.P., et al. (2009). Identification of CANT1 mutations in Des-

buquois dysplasia. Am. J. Hum. Genet. 85, 706–710.

15. Baratela, W.A., Bober, M.B., Tiller, G.E., Okenfuss, E., Ditro, C.,

Duker, A., Krakow, D., Stabley, D.L., Sol-Church, K., Macken-

zie, W., et al. (2012). A newly recognized syndrome with char-

acteristic facial features, skeletal dysplasia, and developmental

delay. Am. J. Med. Genet. A. 158A, 1815–1822.

16. Nizon, M., Huber, C., De Leonardis, F., Merrina, R., Forlino, A.,

Fradin, M., Tuysuz, B., Abu-Libdeh, B.Y., Alanay, Y., Albrecht,

B., et al. (2012). Further delineation of CANT1 phenotypic

spectrum and demonstration of its role in proteoglycan syn-

thesis. Hum. Mutat. 33, 1261–1266.

17. Calı, T., Fedrizzi, L., Ottolini, D., Gomez-Villafuertes, R., Mell-

strom, B., Naranjo, J.R., Carafoli, E., and Brini, M. (2012).

Ca2þ-activated nucleotidase 1, a novel target gene for the

transcriptional repressor DREAM (downstream regulatory

element antagonist modulator), is involved in protein folding

and degradation. J. Biol. Chem. 287, 18478–18491.

18. Bui, C., Huber, C., Tuysuz, B., Alanay, Y., Bole-Feysot, C.,

Leroy, J.G., Mortier, G., Nitschke, P., Munnich, A., and Corm-

ier-Daire, V. (2014). XYLT1mutations in Desbuquois dysplasia

type 2. Am. J. Hum. Genet. 94, 405–414.

19. Shearer, W.T., Rosenblatt, H.M., Gelman, R.S., Oyomopito, R.,

Plaeger, S., Stiehm, E.R., Wara, D.W., Douglas, S.D., Luzuriaga,

K., McFarland, E.J., et al.; Pediatric AIDS Clinical Trials Group

(2003). Lymphocyte subsets in healthy children from birth

through 18 years of age: the Pediatric AIDS Clinical Trials

Group P1009 study. J. Allergy Clin. Immunol. 112, 973–980.

20. Chan, K., and Puck, J.M. (2005). Development of population-

based newborn screening for severe combined immunodefi-

ciency. J. Allergy Clin. Immunol. 115, 391–398.

21. Lupski, J.R., Gonzaga-Jauregui, C., Yang, Y., Bainbridge, M.N.,

Jhangiani, S., Buhay, C.J., Kovar, C.L., Wang, M., Hawes,

A.C., Reid, J.G., et al. (2013). Exome sequencing resolves

apparent incidental findings and reveals further complexity

of SH3TC2 variant alleles causing Charcot-Marie-Tooth neu-

ropathy. Genome Med. 5, 57.

22. Yang, Y., Muzny, D.M., Reid, J.G., Bainbridge, M.N., Willis, A.,

Ward, P.A., Braxton, A., Beuten, J., Xia, F., Niu, Z., et al. (2013).

Clinical whole-exome sequencing for the diagnosis of mende-

lian disorders. N. Engl. J. Med. 369, 1502–1511.

23. Vissers, L.E., Lausch, E., Unger, S., Campos-Xavier, A.B.,

Gilissen, C., Rossi, A., Del Rosario, M., Venselaar, H., Knoll,

U., Nampoothiri, S., et al. (2011). Chondrodysplasia and

abnormal joint development associated with mutations in

IMPAD1, encoding the Golgi-resident nucleotide phospha-

tase, gPAPP. Am. J. Hum. Genet. 88, 608–615.

24. Bainbridge, M.N., Wang, M., Wu, Y., Newsham, I., Muzny,

D.M., Jefferies, J.L., Albert, T.J., Burgess, D.L., and Gibbs,

R.A. (2011). Targeted enrichment beyond the consensus cod-

ing DNA sequence exome reveals exons with higher variant

densities. Genome Biol. 12, R68.

106 The American Journal of Human Genetics 95, 96–107, July 3, 201

25. Li, H., and Durbin, R. (2009). Fast and accurate short read

alignment with Burrows-Wheeler transform. Bioinformatics

25, 1754–1760.

26. Li, H., Handsaker, B., Wysoker, A., Fennell, T., Ruan, J., Homer,

N., Marth, G., Abecasis, G., and Durbin, R.; 1000 Genome

Project Data Processing Subgroup (2009). The Sequence

Alignment/Map format and SAMtools. Bioinformatics 25,

2078–2079.

27. DePristo, M.A., Banks, E., Poplin, R., Garimella, K.V., Maguire,

J.R., Hartl, C., Philippakis, A.A., del Angel, G., Rivas, M.A.,

Hanna, M., et al. (2011). A framework for variation discovery

and genotyping using next-generation DNA sequencing data.

Nat. Genet. 43, 491–498.

28. Challis, D., Yu, J., Evani, U.S., Jackson, A.R., Paithankar, S.,

Coarfa, C., Milosavljevic, A., Gibbs, R.A., and Yu, F. (2012).

An integrative variant analysis suite for whole exome next-

generation sequencing data. BMC Bioinformatics 13, 8.

29. Danecek, P., Auton, A., Abecasis, G., Albers, C.A., Banks, E.,

DePristo, M.A., Handsaker, R.E., Lunter, G., Marth, G.T.,

Sherry, S.T., et al.; 1000 Genomes Project Analysis Group

(2011). The variant call format and VCFtools. Bioinformatics

27, 2156–2158.

30. Bainbridge, M.N., Wiszniewski, W., Murdock, D.R., Friedman,

J., Gonzaga-Jauregui, C., Newsham, I., Reid, J.G., Fink, J.K.,

Morgan, M.B., Gingras, M.C., et al. (2011). Whole-genome

sequencing for optimized patient management. Sci. Transl.

Med. 3, 87re3.

31. de Ligt, J., Boone, P.M., Pfundt, R., Vissers, L.E., Richmond, T.,

Geoghegan, J., O’Moore, K., de Leeuw, N., Shaw, C., Brunner,

H.G., et al. (2013). Detection of clinically relevant copy num-

ber variants with whole-exome sequencing. Hum. Mutat. 34,

1439–1448.

32. Keerthikumar, S., Raju, R., Kandasamy, K., Hijikata, A., Rama-

badran, S., Balakrishnan, L., Ahmed, M., Rani, S., Selvan, L.D.,

Somanathan, D.S., et al. (2009). RAPID: Resource of Asian

Primary Immunodeficiency Diseases. Nucleic Acids Res. 37

(Database issue), D863–D867.

33. Al-Herz, W., Bousfiha, A., Casanova, J.L., Chapel, H., Conley,

M.E., Cunningham-Rundles, C., Etzioni, A., Fischer, A.,

Franco, J.L., Geha, R.S., et al. (2011). Primary immunodefi-

ciency diseases: an update on the classification from the inter-

national union of immunological societies expert committee

for primary immunodeficiency. Front. Immunol. 2, 54.

34. Cheung, S.W., Shaw, C.A., Yu, W., Li, J., Ou, Z., Patel, A., Yat-

senko, S.A., Cooper, M.L., Furman, P., Stankiewicz, P., et al.

(2005). Development and validation of a CGH microarray

for clinical cytogenetic diagnosis. Genet. Med. 7, 422–432.

35. Boone, P.M., Bacino, C.A., Shaw, C.A., Eng, P.A., Hixson, P.M.,

Pursley, A.N., Kang, S.H., Yang, Y., Wiszniewska, J., Nowakow-

ska, B.A., et al. (2010). Detection of clinically relevant exonic

copy-number changes by array CGH. Hum. Mutat. 31, 1326–

1342.

36. Boone, P.M., Campbell, I.M., Baggett, B.C., Soens, Z.T., Rao,

M.M., Hixson, P.M., Patel, A., Bi, W., Cheung, S.W., Lalani,

S.R., et al. (2013). Deletions of recessive disease genes: CNV

contribution to carrier states and disease-causing alleles.

Genome Res. 23, 1383–1394.

37. Lausch, E., Keppler, R., Hilbert, K., Cormier-Daire, V., Nikkel,

S., Nishimura, G., Unger, S., Spranger, J., Superti-Furga, A.,

and Zabel, B. (2009). Mutations in MMP9 and MMP13 deter-

mine the mode of inheritance and the clinical spectrum of

metaphyseal anadysplasia. Am. J. Hum. Genet. 85, 168–178.

4

38. Arnold, K., Bordoli, L., Kopp, J., and Schwede, T. (2006). The

SWISS-MODEL workspace: a web-based environment for

protein structure homology modelling. Bioinformatics 22,

195–201.

39. Arnold, K., Kiefer, F., Kopp, J., Battey, J.N., Podvinec, M.,

Westbrook, J.D., Berman, H.M., Bordoli, L., and Schwede, T.

(2009). The Protein Model Portal. J. Struct. Funct. Genomics

10, 1–8.

40. Fang, W., Du, T., Raimi, O.G., Hurtado-Guerrero, R., Marino,

K., Ibrahim, A.F., Albarbarawi, O., Ferguson, M.A., Jin, C.,

and Van Aalten, D.M. (2013). Genetic and structural

validation of Aspergillus fumigatus N-acetylphosphoglucos-

amine mutase as an antifungal target. Biosci. Rep. 33. Pub-

lished online September 4, 2013. http://dx.doi.org/10.1042/

BSR20130053.

41. Maruyama, D., Nishitani, Y., Nonaka, T., Kita, A., Fukami, T.A.,

Mio, T., Yamada-Okabe, H., Yamada-Okabe, T., and Miki, K.

(2007). Crystal structure of uridine-diphospho-N-acetylglu-

cosamine pyrophosphorylase from Candida albicans and cat-

alytic reaction mechanism. J. Biol. Chem. 282, 17221–17230.

42. Kang, P., Mechref, Y., and Novotny, M.V. (2008). High-

throughput solid-phase permethylation of glycans prior to

mass spectrometry. Rapid Commun. Mass Spectrom. 22,

721–734.

43. Lacey, J.M., Bergen, H.R., Magera, M.J., Naylor, S., and

O’Brien, J.F. (2001). Rapid determination of transferrin iso-

forms by immunoaffinity liquid chromatography and electro-

spray mass spectrometry. Clin. Chem. 47, 513–518.

44. Tegtmeyer, L.C., Rust, S., van Scherpenzeel, M., Ng, B.G.,

Losfeld, M.E., Timal, S., Raymond, K., He, P., Ichikawa, M.,

Veltman, J., et al. (2014). Multiple phenotypes in phosphoglu-

comutase 1 deficiency. N. Engl. J. Med. 370, 533–542.

45. Milewski, S., Chmara, H., and Borowski, E. (1986). Antibiotic

tetaine—a selective inhibitor of chitin and mannoprotein

biosynthesis in Candida albicans. Arch. Microbiol. 145,

234–240.

46. Elorza, V., Mormeneo, S., Garcia de la Cruz, F., Gimeno, C.,

and Sentandreu, R. (1989). Evidence for the formation of

covalent bonds between macromolecules in the domain of

the wall of Candida albicans mycelial cells. Biochem. Biophys.

Res. Commun. 162, 1118–1125.

47. Coman, D., Bostock, D., Hunter, M., Kannu, P., Irving, M.,

Mayne, V., Fietz, M., Jaeken, J., and Savarirayan, R. (2008).

Primary skeletal dysplasia as a major manifesting feature in

The A

an infant with congenital disorder of glycosylation type Ia.

Am. J. Med. Genet. A. 146, 389–392.

48. Drijvers, J.M., Lefeber, D.J., deMunnik, S.A., Pfundt, R., van de

Leeuw, N., Marcelis, C., Thiel, C., Koerner, C., Wevers, R.A.,

and Morava, E. (2010). Skeletal dysplasia with brachytelepha-

langy in a patient with a congenital disorder of glycosylation

due to ALG6 gene mutations. Clin. Genet. 77, 507–509.

49. Kranz, C., Basinger, A.A., Gucsavasx-Caliko�glu, M., Sun, L.,

Powell, C.M., Henderson, F.W., Aylsworth, A.S., and Freeze,

H.H. (2007). Expanding spectrum of congenital disorder of

glycosylation Ig (CDG-Ig): sibs with a unique skeletal

dysplasia, hypogammaglobulinemia, cardiomyopathy, geni-

tal malformations, and early lethality. Am. J. Med. Genet. A.

143A, 1371–1378.

50. Foulquier, F., Vasile, E., Schollen, E., Callewaert, N., Rae-

maekers, T., Quelhas, D., Jaeken, J., Mills, P., Winchester, B.,

Krieger, M., et al. (2006). Conserved oligomeric Golgi complex

subunit 1 deficiency reveals a previously uncharacterized

congenital disorder of glycosylation type II. Proc. Natl. Acad.

Sci. USA 103, 3764–3769.

51. Morava, E., Zeevaert, R., Korsch, E., Huijben, K., Wopereis, S.,

Matthijs, G., Keymolen, K., Lefeber, D.J., De Meirleir, L., and

Wevers, R.A. (2007). A common mutation in the COG7

gene with a consistent phenotype including microcephaly,

adducted thumbs, growth retardation, VSD and episodes of

hyperthermia. Eur. J. Hum. Genet. 15, 638–645.

52. Ng, B.G., Kranz, C., Hagebeuk, E.E., Duran, M., Abeling, N.G.,

Wuyts, B., Ungar, D., Lupashin, V., Hartdorff, C.M., Poll-The,

B.T., and Freeze, H.H. (2007). Molecular and clinical character-

ization of a Moroccan Cog7 deficient patient. Mol. Genet.

Metab. 91, 201–204.

53. Foulquier, F., Amyere, M., Jaeken, J., Zeevaert, R., Schollen, E.,

Race, V., Bammens, R., Morelle, W., Rosnoblet, C., Legrand,

D., et al. (2012). TMEM165 deficiency causes a congenital dis-

order of glycosylation. Am. J. Hum. Genet. 91, 15–26.

54. Mariappa, D., Sauert, K., Marino, K., Turnock, D., Webster, R.,

van Aalten, D.M., Ferguson, M.A., and Muller, H.A. (2011).

Protein O-GlcNAcylation is required for fibroblast growth fac-

tor signaling in Drosophila. Sci. Signal. 4, ra89.

55. Harms, H.K., Zimmer, K.P., Kurnik, K., Bertele-Harms, R.M.,

Weidinger, S., and Reiter, K. (2002). Oral mannose therapy

persistently corrects the severe clinical symptoms and

biochemical abnormalities of phosphomannose isomerase

deficiency. Acta Paediatr. 91, 1065–1072.

merican Journal of Human Genetics 95, 96–107, July 3, 2014 107

The American Journal of Human Genetics, Volume 95

Supplemental Data

PGM3 Mutations Cause a Congenital Disorder

of Glycosylation with Severe Immunodeficiency

and Skeletal Dysplasia

Asbjørg Stray-Pedersen, Paul H. Backe, Hanne S. Sorte, Lars Mørkrid, Niti Y. Chokshi,

Hans Christian Erichsen, Tomasz Gambin, Katja B.P. Elgstøen, Magnar Bjørås, Marcin

W. Wlodarski, Marcus Krüger, Shalini N. Jhangiani, Donna M. Muzny, Ankita Patel,

Kimiyo M. Raymond, Ghadir S. Sasa, Robert A. Krance, Caridad A. Martinez, Shirley M.

Abraham, Carsten Speckmann, Stephan Ehl, Patricia Hall, Lisa R. Forbes, Else

Merckoll, Jostein Westvik, Gen Nishimura, Cecilie F. Rustad, Tore G. Abrahamsen, Arild

Rønnestad, Liv T. Osnes, Torstein Egeland, Olaug K. Rødningen, Christine R. Beck,

Baylor-Johns Hopkins Center for Mendelian Genomics, Eric A. Boerwinkle, Richard A.

Gibbs, James R. Lupski, Jordan S. Orange, Ekkehart Lausch, and I. Celine Hanson

                                                                Figure S1

Man$Glc$

Glycosyla)on$Disorder$

Fru262PO4$NH3$Glutamine$

Glutamate$

Glucosamine262PO4$AcetylCoA$

CoA$

GlcNAc262PO4$

GlcNAc212PO4$

ADP$

ATP$

Glucosamine$

GlcNAc$

ADP$

ATP$

UTP$

PPi$

UDP2GlcNAc$

UDP$

N2Acetylmannosamine$

ManNAc262PO4$PEP$

UDP2GalNAc$

Neu5Ac$CTP$

CMP2Neu5Ac$NADPH$

Fuc$

Fuc212PO4$

GTP$

GDP2Fuc$

NADP+$

GDP242keto262$deoxygalactose$

GDP242keto262$deoxymannose$

Dol2P2Man$

Dol2P$GDP2Man$

Man212PO4$Man262PO4$PEP$

KDN$CTP$

CMP2KDN$

PPi$

Pi$

Pi$ PPi$

GDP$

ATP$

ADP$

N2Acetylneuraminic262PO4$

KDN292PO4$NADP+$

NADPH$

NADPH$

PPi$

ADP$

ATP$

GTP$

PPi$

PPi$

NADP+$

CMP2Neu5Gc$

GalNAc212PO4$GalNAc$ADP$ATP$ UTP$

Glc262PO4$

ADP$

ATP$

Glc212PO4$UTP$

PPi$

UDP2Glc$

UDP2$Gal$Dol2P2Glc$

Dol2P$

UDP$NADH$ NAD$

6PG$

NADPH$ NADP+$

UDP2GlcA$

UDP2Xylose$

CO2$

PGM1$

PGM3$

Glc212PO4$

UDP2Glc$

Gal212PO4$Gal$

ADP$ATP$ ATP$

ADP$

Pi$

Glc262PO4$ Glc$+$PO4$Endoplasmic$Re)culum$

G6PT1$ G6PC3$

CP$

CP$

CP$

CP$

CP$

Pi$

PPi$

UTP$

PMM2$

DPM1$

GNE$

GNE$

GALT$

GALE$

GALK1$

MPI$

GFAT$

Figure S1. Monosaccharide metabolism in mammals and human glycosylation disorders. Biosynthesis and interconversion of monosaccharides. Figure modified from the figures published by Freeze, H.H.,1,2 and presented here with permission from Freeze, H.H and J. Biol. Chem. with copyright © 2013,

by the American Society for Biochemistry and Molecular Biology.

Enzymes marked red, nucleotide sugars (activated forms of monosaccharides) in light blue boxes, and monosaccharide substrates with proven treatment

effects or potential disease modifying effects are bold and marked blue. Symbols and abbreviations: CP, Control points; Dol, Dolichol; PDM1, Dolichyl-

phosphate mannosyltransferase 1 (EC 2.4.1.83); Fuc, Fucose; G6PT1, Glucose-6-phosphate translocase subunit 1; G6PC3, Glucose-6-phosphatase 3 (EC

3.1.3.9); Gal, Galactose; GalNAc, N-acetyl-galactosamine; GALE, UDP-galactose-4-epimerase (EC 5.1.3.2); GALK1, Galactokinase 1 (EC 2.7.1.6); GALT,

Galactose 1 phosphate uridyltransferase (EC 2.7.7.12); Glc, Glucose; GlcA, Glucoronic acid; GlcNAc, N-acetyl-glucosamine; GFPT1 (GFAT), Glutamine

fructose 6-amidotransferase (EC 2.6.1.16); GNE, UDP-N-acetyl-glucosamine 2-epimerase/N-acetyl-mannosamine kinase (EC 3.2.1.183/EC 2.7.1.60); KDN,

2-keto-3-deoxy-D-glycero-D-galactonononic acid (synonym 2-Keto-3-deoxynononic acid); Man, Mannose; ManNAc, N-acetyl-mannosamine; MPI,

Mannosephosphate isomerase (EC 5.3.1.8); Neu5Ac, N-acetyl-neuraminic acid; Neu5Gc, N-glycolyl-neuraminic acid; PEP, Phosphoenolpyruvate; PGM1,

Phosphoglucomutase 1 (EC 5.4.2.2); PGM3, Phosphoglucomutase 3 (EC 5.4.2.3); PMM2, Phosphomannomutase 2 (EC 5.4.2.8); 6PG, 6-phosphogluconate.

Figure S2

p.D239H p.N246S p.N246Kfs*7

p.Q451R

Figure S2. Alignment in different species. Sequences of PGM3 homologs were aligned using the alignment feature of UniProt using the aligner clustalo. The alignment shows that all three reported

disease-causing missense mutations are highly conserved in the selected species, p.N246 completely conserved through yeast.

Figure S3  

A PGM3 GPRC6A PGM3

B

1.2 Mb deletion

UBE3D PGM3 ME1 SNAP91

 

 

Figure S3. Genomic regions with absence of heterozygosity (AOH) predicted from the exome data,  with focus on chromosome 6 (lower panel). A) P1, B) P2. Blue dots and lines indicate AOH. Lowest panel: Prediction of copy number variant (CNV) deletion from exome data in P2. CoNIFER3 was used

as a bioinformatics tool comparing the P2 data (red) with other samples (black).

Footnote Figure S3 A lower panel: Regarding AOH and GPRC6A, P1 has skeletal abnormalities in addition to immunodeficiency, which were not observed in

P2. The initial hypothesis was that the skeletal dysplasia was an additional trait under the assumption of a potential digenic model. A homozygous GPRC6A

variant (Table S5) detected in WES in P1 was regarded as a candidate, since reduced bone mass and impaired osteoblast function is reported in GPRC6A

null mice.4 The analysis of P3, who also had the same skeletal phenotype, allowed us to exclude GPRC6A as a contributing gene because this child was

proven wild-type for GPRC6A, and thus conclude that the skeletal phenotype is indeed due to autosomal recessive inherited mutations in PGM3.

rela

tivePGM3

mRN

A

CTL

patien

t 30.0

0.5

1.0

1.5

Control Patient 3

Figure S4

Figure S4. Quantitative real-time PCR of PGM3 mRNA

Amount of PGM3 mRNA detected in P3’s blood cells relative to control sample. Mean ± SEM of column Control: 1,040 ± 0.03055 (N=3), Mean ± SEM of

column P3: 3617 ± 0.04262 (N=3), Difference between means 0,6783 ± 0.05244, p = 0.002, two-tailed unpaired t test. Primer sequences in Table S5

Abbreviation: SEM, Standard Error of the Mean.  

Table S1. Immunological evaluations and neutrophil counts P1

4 weeks 5 weeks 6 weeks 7 weeks 8 weeks 9 weeks 10 weeks 11 weeks 12 weeks 13 weeks Mean of 4-13 weeks

1 year of age, (6 months

post HSCT)

Age matched references* 0-3 months

Age matched references*

1-2 years B CD19+ (cells/microL) 418 380 489 694 477 NA NA NA NA 398 476 1540 300-2000 720-2600 T

CD3+ (cells/microL) 367 325 623 734 664 367 325 623 734 752 551 3316 2500-5500 2100-6200 CD4+ (cells/microL) 193 204 435 492 499 193 204 435 492 592 374 2333 1600-4000 1300-3400 CD8+ (cells/microL) 67 52 75 91 57 NA NA NA NA 55 66 805 560-1700 620-2000

NK CD16/CD56+ (cells/microL) 327 214 227 355 152 NA NA NA NA 179 242 411 170-1100 180-920

Neutrophils (cells/microL) NA 300 300 200 300 200 1100 600 500 300 <500 7100 1500-8000 1500-8000

*Normal vales (10th-90th percentiles) for lymphocytes obtained from Shearer et al JACI Nov 2003.5

Abbreviations: NA, Not available

Table S2. Immunological evaluations and neutrophil counts P2

3 months 14 months 27 months 35 months 4¼ years 5 years 6 years

(2 months post HSCT)

Age matched references* 0-3 months

Age matched references*

1-2 years

Age matched references*

2-6 years

Age matched references* 6-12 years

B CD19+ (cells/microL) 114 32 2 3 1 4 126 300-2000 720-2600 390-1400 270-860 T

CD3+ (cells/microL) 318 352 312 326 97 311 1154 2500-5500 2100-6200 1400-3700 1200-2600 CD4+ (cells/microL) 253 293 239 217 65 50 865 1600-4000 1300-3400 700-2200 650-1500 CD8+ (cells/microL) 30 32 61 103 28 243 284 560-1700 620-2000 490-1300 370-1100

NK CD16/CD56+ (cells/microL) 144 123 111 69 31 17 137 170-1100 180-920 130-720 100-480

Neutrophils (cells/microL) 700 NA 190 230 450 450 1660 1500-8000 1500-8000 1500-8000 1500-8000

*Normal vales (10th-90th percentiles) for lymphocytes obtained from Shearer et al JACI Nov 2003.5

Abbreviations: NA, Not available

Table S3. Immunological evaluations and neutrophil counts P3

Day 5 3 weeks

3 months (20 µg/kg G-CSF)

4.5 months (20 µg/kg G-CSF)

Age matched references* 0-3 months

Age matched references* 3-6 months

B CD19+ (cells/microL) 138 764 228 198 300-2000 430-3000 T

CD3+ (cells/microL) 478 1190 347 455 2500-5500 2500-5600 CD4+ (cells/microL) 277 699 315 365 1600-4000 1800-4000 CD8+ (cells/microL) 71 188 11 42 560-1700 590-1600

NK CD16/CD56+ (cells/microL) 150 688 89 399 170-1100 170-830

Neutrophils (cells/microL) 0 0 898 1066 1500-8000 1500-8000

*Normal vales (10th-90th percentiles) for lymphocytes obtained from Shearer et al JACI Nov 2003.5

Abbreviations: NA, Not available

Table S4. Whole exome sequencing statistical summary for P1-3

Individual P1 P2 P3 Total captured regions size 52Mb 52Mb 52Mb

% of captured regions with coverage>10 93.78 91.75 92.81 Average coverage 149x 96x 98x

% of bases covered by 1x 96.89 96.09 96.18 % of target hits 98.5 98.32 98.29

Total numbers of SNPs 99,531 112,575 98,751 Total numbers of INDELs 8,533 12,398 8,945

N rare* homozygous (n confirmed by Sanger) 15 (2) 20 (1) 18 (3)

N rare* compound heterozygous (n confirmed by Sanger) 11 pairs (0) 10 pairs (0) 8 pairs (1)

N X linked (n confirmed by Sanger) 20 het,1 hom (0) 9 hemi (0) 12 hemi (0) N shared genes with deleterious variants

(n confirmed by Sanger) 1 (1) 1 (1) 1 (1)

* Rare defined as < 0.003 in ESP, ARIC, and BCM-CMG.

Abbreviations: ESP, NHLBI GO Exome Sequencing Project (ESP) server; ARIC, Arterosclerosis Risk In Communities contains WES results from ~4000

individuals; BCM-CMG, in-house database for Baylor College of Medicine-Center for Mendelian Genomics >200 individuals.

Table S5. Sequence of the primers used for Sanger verification, and qPCR of the detected variants.

Primer name RefSeq transcript Primer sequence (5’-3’) Variants and regions of interest

PGM3_ex6F NM_015599.2

CCTTTGTAGGCTTCTTGCAGTGG [chr6.hg19:g.83891505T>C, c.737A>G, p.Asn246Ser]; [chr6.hg19:g.83891505dupT; c.737dupA, p.Asn246Lysfs*7]; [chr6.hg19:g.83891527C>G; c.715G>C, p.Asp239His] PGM3_ex6R TGTGGCAGAGCCAGGAATAGC

PGM3 qPCR rqf NM_015599.2 CACATGAAGTGAGCTTGGCA PGM3 exon 13 PGM3 qPCR rqr CAATCCAGTAGGCTGCATTG PGM3_ex11F NM_015599.2 ATTTGTTTCCCCATTTGCAG [chr6.hg19:g.83881669T>C; c.1352A>G; p.Gln451Arg] PGM3_ex11R TGTCAGTGAGATATAATGAGAATTGG GPRC6A_ex6F NM_148963.2 GCATTTGGCACCATGCTGGGC [chr6.hg19:g.117113762_117113766delinsGGTAATTTCCT;

c.2320_2324delinsAGGAAATTACC;p.Lys774_Tyr775delinsArgLysLeuPro] GPRC6A_ex6 CCAGTGAAGCAGGACTCAGGGC

 

 

Table S6. Matrix-assisted laser desorption ionization-time of flight mass spectrometry  quantification of O-linked glycan profiles in two individuals with CDG-PGM3

Individual Age Therapy T antigen (micromol/L)

Sialyl-T antigen (micromol/L)

Ratio T/ST

P1 1m Pre HSCT 0.55 12.2 0.05 P1 2m Pre HSCT 0.57 17.7 0.03 P1 2y Post HSCT 0.53 17.7 0.03 P2 6y Post HSCT 0.55 16.3 0.03

Reference ranges 0.22 - 1.4 11.7 - 31.4 0 - 0.06 Symbols and abbreviations: HSCT, Hematopoietic stem cell transplantation; IgG, Immunoglobulin G, m, months; T/ST, T antigen/sialyated-T antigen; y, years.

Supplemental References

1. Freeze,H.H. (2013). Understanding human glycosylation disorders: biochemistry leads the charge. J. Biol. Chem. 288 , 6936-6945.

2. Freeze,H.H., and Elbein,A.D. (2009). Glycosylation Precursors. In Essentials of Glycobiology, A.Varki, R.D.Cummings, J.D.Esko, and et al, eds. (Cold

Spring Harbor (NY): Cold Spring Harbor Laboratory Press), pp. 47-61.

3. de Ligt,J., Boone,P.M., Pfundt,R., Vissers,L.E., Richmond,T., Geoghegan,J., O'Moore,K., de Leeuw,N., Shaw,C., Brunner,H.G., et al. (2013).

Detection of clinically relevant copy number variants with whole-exome sequencing. Hum. Mutat. 34 , 1439-1448.

4. Pi,M., Zhang,L., Lei,S.F., Huang,M.Z., Zhu,W., Zhang,J., Shen,H., Deng,H.W., and Quarles,L.D. (2010). Impaired osteoblast function in GPRC6A null

mice. J. Bone Miner. Res. 25 , 1092-1102.

5. Shearer,W.T., Rosenblatt,H.M., Gelman,R.S., Oyomopito,R., Plaeger,S., Stiehm,E.R., Wara,D.W., Douglas,S.D., Luzuriaga,K., McFarland,E.J., et al

(2003). Lymphocyte subsets in healthy children from birth through 18 years of age: the Pediatric AIDS Clinical Trials Group P1009 study. J. Allergy

Clin. Immunol. 112 , 973-980.