Initiation of Protein O Glycosylation by the Polypeptide GalNAcT-1 in Vascular Biology and Humoral Immunity

Abstract

Core-type protein O glycosylation is initiated by polypeptide N-acetylgalactosamine (GalNAc) transferase (ppGalNAcT) activity and produces the covalent linkage of serine and threonine residues of proteins. More than a dozen ppGalNAcTs operate within multicellular organisms, and they differ with respect to expression patterns and substrate selectivity. These distinctive features imply that each ppGalNAcT may differentially modulate regulatory processes in animal development, physiology, and perhaps disease. We found that ppGalNAcT-1 plays key roles in cell and glycoprotein selective functions that modulate the hematopoietic system. Loss of ppGalNAcT-1 activity in the mouse results in a bleeding disorder which tracks with reduced plasma levels of blood coagulation factors V, VII, VIII, IX, X, and XII. ppGalNAcT-1 further supports leukocyte trafficking and residency in normal homeostatic physiology as well as during inflammatory responses, in part by providing a scaffold for the synthesis of selectin ligands expressed by neutrophils and endothelial cells of peripheral lymph nodes. Animals lacking ppGalNAcT-1 are also markedly impaired in immunoglobulin G production, coincident with increased germinal center B-cell apoptosis and reduced levels of plasma B cells. These findings reveal that the initiation of protein O glycosylation by ppGalNAcT-1 provides a distinctive repertoire of advantageous functions that support vascular responses and humoral immunity.

Full text

MOLECULAR AND CELLULAR BIOLOGY, Dec. 2007, p. 8783–8796 Vol. 27, No. 24
0270-7306/07/$08.00⫹0 doi:10.1128/MCB.01204-07
Copyright © 2007, American Society for Microbiology. All Rights Reserved.
Initiation of Protein O Glycosylation by the Polypeptide GalNAcT-1 in
Vascular Biology and Humoral Immunity
䌤
†
Mari Tenno,
1
Kazuaki Ohtsubo,
1
Fred K. Hagen,
2
David Ditto,
3
Alexander Zarbock,
4
Patrick Schaerli,
5
Ulrich H. von Andrian,
5
Klaus Ley,
6
Dzung Le,
3
Lawrence A. Tabak,
7
and Jamey D. Marth
1
*
Howard Hughes Medical Institute and Department of Cellular and Molecular Medicine, University of California San Diego,
La Jolla, California
1
; Department of Biochemistry and Biophysics, University of Rochester, School of Medicine and Dentistry,
Rochester, New York
2
; Department of Pathology, University of California San Diego, La Jolla, California
3
; Department of
Anesthesiology and Critical Care Medicine, University of Muenster, Muenster, Germany
4
; CBR Institute for Biomedical
Research and Department of Pathology, Harvard Medical School, Boston, Massachusetts
5
; Cardiovascular Research
Center and Department of Biomedical Engineering, University of Virginia, Charlottesville, Virginia
6
; and
Section on Biological Chemistry, National Institute of Diabetes and Digestive and Kidney Diseases,
National Institutes of Health, Department of Health and Human Services, Bethesda, Maryland
7
Received 6 July 2007/Returned for modification 18 September 2007/Accepted 25 September 2007
Core-type protein O glycosylation is initiated by polypeptide N-acetylgalactosamine (GalNAc) transferase
(ppGalNAcT) activity and produces the covalent linkage of serine and threonine residues of proteins. More
than a dozen ppGalNAcTs operate within multicellular organisms, and they differ with respect to expression
patterns and substrate selectivity. These distinctive features imply that each ppGalNAcT may differentially
modulate regulatory processes in animal development, physiology, and perhaps disease. We found that ppGal
NAcT-1 plays key roles in cell and glycoprotein selective functions that modulate the hematopoietic system.
Loss of ppGalNAcT-1 activity in the mouse results in a bleeding disorder which tracks with reduced
plasma levels of blood coagulation factors V, VII, VIII, IX, X, and XII. ppGalNAcT-1 further supports
leukocyte trafficking and residency in normal homeostatic physiology as well as during inflammatory
responses, in part by providing a scaffold for the synthesis of selectin ligands expressed by neutrophils and
endothelial cells of peripheral lymph nodes. Animals lacking ppGalNAcT-1 are also markedly impaired in
immunoglobulin G production, coincident with increased germinal center B-cell apoptosis and reduced
levels of plasma B cells. These findings reveal that the initiation of protein O glycosylation by ppGal
NAcT-1 provides a distinctive repertoire of advantageous functions that support vascular responses and
humoral immunity.
A large fraction of cellular protein O glycosylation is di-
rected to producing a series of core-type O-glycan structures
that begin with the covalent linkage of N-acetylgalactosamine
(GalNAc) to serine and threonine residues of proteins in the
secretory pathway (10, 35, 38, 58, 61). This initial enzymatic
step by a polypeptide GalNAc transferase (ppGalNAcT) is
followed in the Golgi apparatus by the regulated attachment of
other saccharide linkages to this GalNAc residue by other
glycosyltransferases, contributing to the changing O-glycan
repertoire of a given cell (8, 45). Full disclosure of ppGalNAcT
structure and expression was made possible by biochemical
purification of the enzymatic activity, followed by acquisition of
the amino acid sequence (17, 24). Unexpectedly, multiple
cDNAs encoding apparent ppGalNAcT isozymes were discov-
ered, while genetic disruption of ppGalNAcT in the mouse
failed to ablate O-glycan formation in vivo (19, 22, 34, 48, 53,
63, 68). A large family of ppGalNAcTs is now evident among
multicellular organisms analyzed thus far, with mammals ex-
pressing at least 15 ppGalNAcT isozymes encoded by different
genes (9, 55).
The prototypic ppGalNAcT family member is ppGal
NAcT-1, which is expressed at high levels and is found broadly
among most tissues and cell types (17, 24, 30, 66). ppGal
NAcT-1 exhibits a selective preference for some polypeptide
sequences (7, 16, 60), further suggesting that a key function is
likely provided by its activity among cells of intact organisms.
The ppGalNAcT family of glycosyltransferases initiates core-
type O-glycan formation, primarily consisting of core 1, core 2,
core 3, and core 4 O-glycan subtypes that may be elaborated
further by glycan linkages, contributing to selectin ligand-de-
pendent control of leukocyte trafficking, regulation of CD8
⫹
T-cell apoptosis, and sensitivity to colitis and preventing the
onset of Tn syndrome (11, 27, 42, 59, 65). Other endogenous
functions have been established among invertebrate model
organisms bearing diminished levels of various core-type O-
glycans (1, 46, 54, 56).
Comparative analysis of ppGalNAcT members further indi-
cates that not all are created equal. Besides considerable dif-
ferences in tissue and cell type expression patterns, substrate
specificities among polypeptides of ppGalNAcT isozymes are
also significantly varied (30, 36, 40, 66). Preferences for dis-
similar polypeptide sequences have been observed along with
the existence of a hierarchical process in protein O glycosyla-
tion that reflects the influence of adjacent core-type O-glycan
* Corresponding author. Mailing address: Howard Hughes Medical
Institute, University of California San Diego, 9500 Gilman Drive,
MC0625, La Jolla, CA 92093. Phone: (858) 534-6526. Fax: (858) 534-
6724. E-mail: jmarth@ucsd.edu.
† Supplemental material for this article may be found at http://mcb
.asm.org/.
䌤
Published ahead of print on 8 October 2007.
8783

linkages in substrate recognition by some ppGalNAcTs (5, 20,
28, 53). These findings together indicate that functional redun-
dancy among ppGalNAcTs is likely limited and that each may
provide key physiologic roles in vertebrates, reflecting the sig-
nificant level of orthologous ppGalNAcT gene sequence conser-
vation throughout speciation. Thus far, however, mammalian
models of ppGalNAcT deficiency, including ppGalNAcT-4, -5,
and -13 deficiencies, lack obvious physiologic manifestations
linked with decreased O-glycan formation and indicative of vital
endogenous roles (22, 55, 68). Further investigations are war-
ranted, however, and mutations in the human GALNT3 gene,
encoding ppGalNAcT-3, have been described for familial tumoral
calcinosis, likely reflecting reduced protein O glycosylation (4, 26,
29, 57). Discovering biologically relevant roles of ppGalNAcTs by
using intact model organisms should ultimately reveal why natural
selection has maintained this family of glycosyltransferases during
the evolution of multicellular organisms.
In order to detect physiologic activities linked to ppGal
NAcT-1 in a mammalian model system, we generated and
characterized intact mice lacking ppGalNAcT-1. We found
that although ppGalNAcT-1 deficiency is generally tolerated
and does not cause infertility, substantial defects occur in the
formation of selectin ligands, resulting in altered innate and
adaptive immune cell trafficking. In addition, ppGalNAcT-1
supports O-glycoprotein expression among a subset of blood
coagulation factors, such that its deficiency results in a mod-
erate to severe bleeding disorder. A key role of ppGalNAcT-1
in promoting adaptive immunity is also evident by an increase
in germinal center (GC) B-cell apoptosis, leading to reduced
plasma B-cell numbers and a substantial reduction in immu-
noglobulin G (IgG) antibody isotype abundance. These phe-
notypic and mechanistic findings indicate that distinctive phys-
iologic activities are controlled by ppGalNAcT-1 in
comparison with other similarly studied glycosyltransferase de-
ficiency states and establish evidence of advantageous biologic
purpose linked to the maintenance of ppGalNAcT-1 within the
mammalian genome.
MATERIALS AND METHODS
ppGalNAcT-1 gene mutagenesis. Isolation of mouse ppGalNAcT-1 genomic
DNA and construction of a targeting vector bearing Cre/loxP recombinant sig-
nals were accomplished using a human cDNA probe according to previously
described procedures (41). Mice bearing the loxP-flanked exon 3 were bred with
Zp3-Cre transgenic mates, as described previously (47). Genotyping was per-
formed using Southern blotting and PCR according to previously published
procedures (41). The wild-type ppGalNAcT-1 allele was detected as a 300-bp
fragment by using primers P1 (5⬘-TCATCACAGTGTCTACCATGGCTG
GAG) and P3 (5⬘-GATCTGATGACCTGTTGTGGACACCTG), while the tar-
geted F[tkneo] allele was detected using P1 and P2 (5⬘-TTCCAGGACAGCCA
GGGCTACACAGAG), yielding a 550-bp fragment.
ppGalNAcT enzyme activity. Tissues were extracted and analyzed as described
previously (18). Enzyme assays for ppGalNAcT-1 activity were performed at
37°C for 1 to 2 h, using the peptide acceptor PRFQDSSSKAPPPLPSPSRLPG in
a final volume of 25 ␮l containing 50 ␮M UDP-GalNAc (77,000 cpm
14
C),5mM
AMP, 10 mM MnCl
2
, 40 mM cacodylate, pH 6.5, 40 mM ␤-mercaptoethanol, and
0.1% Triton X-100. Products were characterized by anion-exchange chromatog-
raphy and evaluated by reverse-phase high-performance liquid chromatography.
Subcloning and expression of ppGalNAcT-1 cDNAs. Isolation of cDNAs from
wild-type and ppGalNAcT-1 exon 3 deletion mice was achieved by reverse
transcription-PCR amplification, using mouse kidney total RNA reverse tran-
scribed with a first-strand cDNA synthesis kit (Clontech). The luminal region of
ppGalNAcT-1 was amplified using the PCR primers Mlu-mT1 (5⬘CACACGCG
TTGCCTGCTGGTGACGTTCTAGAGCTAGT) and Bam-mT1 (5⬘ATGCGG
ATCCAGCCCAGTCAATCCTTCCTT) to incorporate an MluI cloning site
into the stem region of mouse ppGalNAcT-1. This MluI-BamHI cDNA fragment
was cloned into the MluI-BamHI sites of the mammalian expression vehicle
pIMKF3. Three independent ppGalNAcT-I cDNA clones bearing wild-type and
deletion alleles were isolated and characterized. The DNA sequence was ac-
quired for all clones. Expression of the recombinant enzymes was achieved by
transient transfection of COS7 cells, using Lipofectamine (Life Technologies).
All six constructs were transfected in duplicate. Enzyme assays for ppGalNAcT
activity were performed at 37°C under standard assay conditions, with a final
reaction volume of 25 ␮l containing 500 ␮M EA2 peptide (PTTDSTTPAPTTK),
50 ␮M UDP-GalNAc (20,000 cpm
14
C), 10 mM MnCl
2
, 40 mM cacodylate, pH
6.5, 40 mM ␤-mercaptoethanol, and 0.1% Triton X-100.
Bleeding time. Mice were anesthetized by a mixture of 3% isoflurane with
oxygen in an induction chamber and then maintained with a nose cone in a warm
brass cone restraint. Tails of horizontally restrained mice were transected 2 mm
from the tip with a new razor blade and then immersed vertically ⬃2 cm below
the surface of 37°C saline. The time until bleeding stopped for at least 10 seconds
was recorded. If the bleeding continued at 10 min, the tail was withdrawn from
saline and the tip cauterized to stop bleeding.
Hematology. Mice were anesthetized by a mixture of 3% isoflurane with
oxygen in an induction chamber and maintained with a nose cone outside the
induction chamber. Tails of mice were transected 2 mm from the tip with a new
razor blade. One hundred microliters of blood was allowed to drip into EDTA-
containing polypropylene microtubes (Becton Dickinson, NJ). Blood in tubes
was immediately mixed well to ensure proper anticoagulation and kept at room
temperature until analysis (within 4 hours). Blood cell counts with leukocyte
differential and platelet counts were performed in duplicate on a Hemavet 850FS
multispecies hematology system (Drew Scientific, CT) programmed with mouse
hematology settings. A whole blood smear was prepared from each sample and
Wright stained for manual viewing.
Serum chemistry. Blood was collected from anesthetized mice as described for
hematology, but without EDTA. Blood was allowed to clot for 3 to5hatroom
temperature and centrifuged in a serum separator tube for 5 min at 7,000 ⫻g.
Serum was removed and analyzed with a Beckman CX-7 automated chemistry
analyzer.
Coagulation. Up to 1 ml of blood was quickly withdrawn from the heart of a
mouse anesthetized as described for hematology, using a 1-ml plastic syringe with
a 25-gauge needle containing 30 ␮l buffered citrate (0.06 ml/liter sodium citrate
plus 0.04 ml/liter citric acid). Blood was added to a plastic tube containing
sufficient additional citrate to achieve a final ratio of 9 parts whole blood to 1 part
citrate. Blood was immediately mixed well and centrifuged twice at 2,000 relative
centrifugal force for 15 min to obtain citrated platelet-poor plasma. Plasma
samples were aliquoted and frozen at ⫺80°C until being processed for coagula-
tion analysis. von Willebrand factor (VWF) analysis was performed as described
previously (12). Factor VIII and other factors were analyzed as described else-
where (62).
Flow cytometry. Single-cell suspensions from the spleen, thymus, lymph nodes,
and bone marrow were prepared, and red blood cells were removed by ammo-
nium chloride lysis. Antibodies to CD3 (2C11), CD8 (53-6.7), CD40 (HM40-3),
CD62L (MEL-14), B7.2 (GL1), and Gr1 (RB6-8C5) were purchased from eBio-
science. Antibodies to CD4 (RM4-5), CD11a (2D7), CD11b (M1/70), CD18
(C71/16), CD19 (1D3), CD24 (M1/69), CD44 (IM7), CD45R/B220 (RA3-6B2),
activated caspase-3, Fas/CD95, GL7, IgM (R6-60.2), IgG1 (A85-1), IgG2a/2b
(R2-40), IgG3 (R40-82), I-A
b
(AF6-120.1), and P-selectin glycoprotein ligand 1
(PSGL-1) (2PH1) were acquired from Pharmingen. Cells were incubated in the
presence of various antibodies (above) in fluorescence-activated cell sorter buffer
(2% fetal calf serum [FCS] in phosphate-buffered saline [PBS]) for 20 min at 4°C.
For E- or P-selectin binding, cells were treated with 0.5 ␮g/ml of Fc block
(anti-CD32/16; Pharmingen) and then incubated with anti-Gr1 antibody as de-
scribed previously, with and without the addition of 5 mM EDTA for 30 min (33).
Cells were washed and incubated with a goat anti-human fluorescein isothiocya-
nate (FITC)-conjugated secondary antibody (Sigma). Peanut agglutinin (PNA)
lectin (Vector) binding was accomplished as previously described (59). Activated
caspase-3 analysis by cytometry was done by permeabilizing cells using BD
Cytofix/Cytoperm solution for 20 min at 4°C, followed by washing cells twice with
PermWash solution (Pharmingen). Data were analyzed on a FACSCalibur flow
cytometer using Cellquest software (Becton Dickinson).
Histology. Frozen sections of lymph nodes or spleen were cut at 5 ␮m, air
dried, fixed in acetone, and incubated with biotinylated anti-CD4, biotinylated
anti-CD8, anti-CD45R/B220, anti-GL7, anti-activated caspase-3, anti-FDC-M1
(Pharmingen), anti-CD68 (Serotec), and PNA. After being washing, sections
were incubated with streptavidin-FITC and goat anti-rat rhodamine-conjugated
secondary antibody. For L-selectin binding, sections were incubated with an
L-selectin–IgG chimera and with MECA-79 antibody and then incubated with
8784 TENNO ET AL. MOL.CELL.BIOL.

goat anti-human IgG FITC-conjugated secondary antibody and goat anti-rat
rhodamine-conjugated secondary antibody. DNA fragmentation was measured
by terminal deoxynucleotidyltransferase-mediated dUTP-biotin nick end label-
ing (TUNEL) assay (Promega) according to the manufacturer’s instructions. The
mean fluorescence was analyzed using a MetaMorph system (Universal Imaging
Corporation).
Lymphocyte trafficking. Homing assays were carried out with 2.5 ⫻10
7
cells
isolated and incubated with the CellTracker probe CMFDA (Molecular Bio-
probes) prior to injection into the tail vein as previously described (33). Lym-
phoid organs were harvested1hor24hafter injection, and T and B lymphocytes
positive for CMFDA were measured by flow cytometry.
Perfused microflow chamber. We used microflow chambers as previously
described (67). Briefly, microflow chambers were constructed from 20- by
200-␮m rectangular glass capillaries with a length of 30 mm (VitroCom, Moun-
tain Lake, NJ). The capillary was placed between two microscope coverslips, and
the ends were attached to heparinized polyethylene PE 50 tubing (inner diam-
eter, 0.58 mm; outer diameter, 0.965 mm) (Becton Dickinson, Sparks, MD). The
capillaries were coated with P-selectin (20 ␮g/ml) or E-selectin (30 ␮g/ml) for 2 h
and blocked for 1 h using 10% casein (Pierce Chemicals, Dallas, TX). The
chamber was connected at one side to PE 10 tubing and inserted into the carotid
artery, and a PE 50 tube on the other side was used to control the wall shear
stress (5.94 dynes/cm
2
). Microscopy was conducted using a Zeiss Axioskop mi-
croscope (Carl Zeiss, Inc., Thornwood, NY) with a saline immersion objective
(SW 20/0.5). Images were recorded with a 3CCD color video camera (model
DXC-390; Sony Corporation, Japan) connected to a Panasonic S-VHS recorder.
The chamber was perfused with blood for 6 min before one representative field
of view was recorded for 1 min.
B-cell activation and antibody production. B lymphocytes were purified from
splenocytes by using a Dynal MPC magnet system (Dynal Biotech). Equivalent
numbers of B cells of each genotype (1 ⫻10
5
) were cultured in complete RPMI
1640 medium containing ␤-mercaptoethanol (0.1 mM), 10% FCS, and L-glu-
tamine with the indicated concentrations of goat F(ab⬘)
2
anti-mouse IgM anti-
serum (Jackson) or lipopolysaccharide (LPS; Sigma). Proliferative capacity was
measured by cellular incorporation of [
3
H]thymidine (2.5 ␮Ci per well) during
the last 16 h of a 72-h assay period.
Mice were bled to obtain preimmune sera and subsequently immunized by
intraperitoneal injection of 100 ␮g of dinitrophenyl (DNP)-keyhole limpet he-
mocyanin (KLH) (Calbiochem) in Freund’s complete adjuvant or 10 ␮gof
DNP-Ficoll (Biosearch) in PBS. Serum was collected at the indicated times, and
anti-DNP titers were determined by enzyme-linked immunosorbent assay, using
plates coated with 20 ␮g of DNP-bovine serum albumin and blocked with 10%
FCS in PBS. Mice receiving the DNP-KLH antigen were boosted at the indicated
times with the same amount of antigen in Freund’s incomplete adjuvant. Sera
were diluted to various concentrations and analyzed using anti-mouse isotype-
specific antibodies conjugated to alkaline phosphatase (IgM and IgA [Sigma] or
IgG1, IgG2a, IgG2b, and IgG3 [Pharmingen]). Optical densities at 405 nm
(OD
405
) were obtained using a microplate reader (Molecular Devices). Results
shown in Fig. 7 comprise the indicated serum dilutions in the linear range for the
OD
405
values obtained.
Peritonitis. Mice were administered 1 ml of 0.2% casein in PBS by intraper-
itoneal injection. At the indicated times, animals were sacrificed and their peri-
toneal cavities were subjected to lavage with 10 ml of ice-cold PBS containing 1%
bovine serum albumin and 0.5 mM EDTA. Red blood cells were removed by
ammonium chloride lysis. Peritoneal cell exudates were incubated with anti-Gr-1
antibody and the pan-macrophage marker antibody F4/80 and analyzed by flow
cytometry.
Statistical analysis. Data were plotted as means ⫾standard errors of the
means, and Student’s ttest was used to calculate Pvalues, unless otherwise
indicated.
RESULTS
Germ line mutagenesis of polypeptide GalNAcT-1 disrupts
O-glycosylation activity. A genomic DNA clone encompassing
exon 3 of the mouse gene encoding ppGalNAcT-1 was iso-
lated, characterized by sequence analysis, and used in con-
structing a gene-targeting vector for conditional mutagenesis
(Fig. 1A; see Fig. S1 in the supplemental material). Incorpo-
ration of loxP recombination signals flanking exon 3 and of
selectable markers permitted modular excision by Cre re-
combinase activity, resulting in the presence of type 1 (⌬;
deleted) and type 2 (F; loxP-flanked) alleles (Fig. 1B).
Genomic DNA samples from embryonic stem (ES) cells
bearing the targeted F[tkneo] allele were first characterized
for the retention of all three loxP sites (Fig. 1C). Following
Cre transfection, ES cell subclones bearing the expected
type 1 and type 2 alleles were isolated. Mice bearing loxP-
flanked exon 3 in the germ line (ppGalNAcT-1
F
) were gen-
erated from ES cell clone 3-3 and were bred with Zp3-Cre
transgenic mates (47) to produce offspring bearing the
ppGalNAcT-1
⌬
allele (Fig. 1D). The ppGalNAcT-1
⌬
allele
was then crossed into the C57BL/6NHsd inbred mouse
background for at least six generations prior to phenotypic
analysis. A minor but consistent reduction in the frequency
of pups homozygous for the ppGalNAcT-1
⌬
allele was ob-
served; however, those born developed to adulthood without
overt physiologic abnormalities comparable to wild-type (wt/
wt) and heterozygous (wt/⌬) littermates.
ppGalNAcT-1 enzymatic activity in extracts of multiple tis-
sues of wt/wt and ⌬/⌬mice at 8 weeks of age was measured
using a peptide substrate that is O glycosylated preferentially
by ppGalNAcT-1. Enzymatic activity was detected among mul-
tiple tissues from wild-type animals. In contrast, mice homozy-
gous for the ppGalNAcT-1
⌬
allele lacked detectable ppGal
NAcT-1 activity among all tissues surveyed, with the exception
of the brain (Fig. 1E). The remaining enzyme activity in brain
samples may reflect the expression of a closely related ppGal
NAcT isozyme (22, 68). Deletion of exon 3 results in prema-
ture translational termination (Fig. 1F). Further confirmation
that exon 3 deletion abolishes ppGalNAcT-1 enzyme activity
was obtained upon expression of cDNA clones derived from
mice bearing the wt/wt or ⌬/⌬genotype. Significant enzymatic
activity towards the peptide substrate in COS7 cell extracts was
observed upon high-level expression of wild-type ppGal
NAcT-1 cDNA, while significant activity was not detected
upon similar RNA expression of a truncated ppGalNAcT-1
cDNA structure isolated from mice bearing the ⌬/⌬genotype
(Fig. 1G and data not shown). These findings reveal that de-
letion of ppGalNAcT-1 exon 3 results in an enzymatic null
mutation.
Altered hemostasis and increased bleeding in ppGalNAcT-
1-deficient mice. During routine tissue biopsies and surgical
manipulations of ppGalNAcT-1-deficient mice, there was a
significant increase in bleeding and hemorrhaging. We mea-
sured a threefold increase in bleeding time among mice lacking
ppGalNAcT-1 (Table 1). Moreover, while the prothrombin
time was normal, the activated partial thromboplastin time
(APTT) was slightly prolonged by ppGalNAcT-1 deficiency.
This suggested that ppGalNAcT-1 deficiency might alter coag-
ulation factor levels or other hemostatic components of the
blood, such as VWF and platelets. Coagulation factor activity
levels in ppGalNAcT-1-deficient mice were indeed altered,
with a slight to moderate decrease in most coagulation factors,
including those that contribute to bleeding time and APTT
measurements, such as factors V, VII, VIII, IX, X, and XII. In
ppGalNAcT-1 deficiency, the VWF antigen level increased
approximately 50%. The VWF produced by ppGalNAcT-1-
deficient mice had a normal capacity to bind factor VIII but a
slightly reduced (by 10%) capacity to bind collagen (data not
shown). However, there was no apparent difference in PNA
VOL. 27, 2007 O GLYCOSYLATION IN VASCULAR AND IMMUNE BIOLOGY 8785

binding to VWF. Platelet counts (Fig. 2A) and platelet aggre-
gation in response to various stimuli, including arachidonic
acid, collagen, ADP, thromboxane analog, and calcium
ionophore, were similar in both wild-type and null mice
(data not shown). Whether these changes reflect a de-
creased rate of production or a decrease in half-life remains
to be established.
ppGalNAcT-1 function in leukocyte homeostasis and O gly-
cosylation. We carried out further initial phenotypic screening,
including assessments of organ function by analysis of serum
chemistry. While those measurements were within the normal
range among ppGalNAcT-1-deficient mice (see Table S1 in the
supplemental material; data not shown), alterations among
white blood cell populations were detected (Fig. 2A). The
number of lymphocytes in circulation was slightly increased,
while eosinophils and basophils were substantially reduced. No
change was seen in the circulating levels of Gr1
⫹
neutrophils.
Similarly, platelet and red blood cell counts and the hemoglo-
bin level were unaffected. The mean platelet volume, as well as
red blood cell morphology and hemoglobin content, was also
normal (data not shown).
O glycosylation initiated by ppGalNAcTs can be detected in
part by the PNA lectin, which binds to the unsialylated core 1
O-glycan (44). Cytometric measurements of PNA binding at
the cell surfaces of intact viable leukocytes revealed a signifi-
cant decrease in O glycosylation among Gr1
⫹
neutrophils and
B220
⫹
B cells, while no significant diminution occurred on the
surfaces of CD3
⫹
T cells (Fig. 2B). These findings indicate that
a loss of protein O glycosylation, as detected by PNA binding,
occurs among some but not all leukocyte cell types.
FIG. 1. ppGalNAcT-1 mutagenesis and loss of enzyme activity by deletion of exon 3. (A) Construction of ppGalNAcT-1 targeting vector for
subsequent Cre-loxP recombination. (B) Cre recombination results in the deletion of exon 3, producing the null allele (type 1), or the flanking of
exon 3 by loxP sites to generate the floxed allele (type 2). Restriction enzyme sites for panels A and B are indicated as follows: A, ApaI; B, BamHI;
R, EcoRV; X, XhoI; Sp, SpeI; N, NotI. (C) Genomic Southern blot analysis of targeted ES cell clones (6-6, 1-4, and 3-3), using the loxP probe.
(D) Genomic Southern blotting of tail DNAs, using a genomic probe, indicated the presence of germ line type 1 and type 2 mutations in the gene
encoding ppGalNAcT-1. (E) Enzyme activity was assayed with total protein extracts from various tissues by using the peptide substrate
PRFQDSSSKAPPPLPSPSRLPG. O-glycosylated products were profiled by anion-exchange chromatography and evaluated by reverse-phase
high-performance liquid chromatography. Data are represented as means ⫾standard deviations (SD) for three separate experiments. SG,
salivary gland; Spl, spleen; Thy, thymus. (F) Polypeptide GalNAcT-1 cDNA structures sequenced from the kidney tissues of mice bearing
either the wild-type or deleted (⌬) allele. (G) ppGalNAcT enzyme activity towards the EA2 peptide (PTTDSTTPAPTTK) in COS7 cell
extracts following transfection of cDNAs expressed by the pIMKF3 vector. Data are represented as means ⫾SD for three separate
experiments. WT, wild type.
8786 TENNO ET AL. MOL.CELL.BIOL.

ppGalNAcT-1 contributes to E- and P-selectin ligand for-
mation and neutrophil recruitment during inflammation. Cir-
culating Gr-1
⫹
neutrophils were further analyzed for the levels
of E- and P-selectin ligands at the cell surface by flow cytom-
etry. Remarkably, the relative reduction in neutrophil cell sur-
face O-glycans judged by PNA lectin binding was accompanied
by a fourfold decrease in E-selectin ligand expression and a
fivefold reduction in P-selectin ligand levels (Fig. 3A). No
changes occurred in the expression of other cell surface mol-
ecules involved in cell adhesion, including CD11, CD18, and
L-selectin as well as the selectin counterreceptors CD24 and
PSGL-1 (Fig. 3B).
To investigate the effect of reduced selectin ligand expression
on the adhesive properties of neutrophils, E- and P-selectin-me-
diated leukocyte rolling was measured using an autoperfused flow
chamber (67). The number of ppGalNAcT-1-deficient leuko-
cytes rolling on P-selectin was reduced to approximately 50%
of the normal level, whereas the number of cells rolling on
E-selectin was reduced to approximately 30% of the normal
level, at 5.94 dynes/cm
2
(Fig. 3C).
The impact of this deficiency in E- and P-selectin ligands
upon neutrophil recruitment to an inflammatory stimulus was
explored in a model of acute peritonitis. A significant decrease
in intraperitoneal neutrophil recruitment reflecting a three- to
fivefold reduction in Gr-1
⫹
cell number resulted from the
absence of ppGalNAcT-1 (Fig. 3D). This finding is consistent
with the observed reductions of E- and especially P-selectin
ligand levels with diminished leukocyte rolling, indicating that
protein O glycosylation contributed by ppGalNAcT-1 plays a
significant role in the formation of selectin ligands that operate
in the innate immune inflammatory response among activated
endothelial cells.
Selectin ligand formation by ppGalNAcT-1 promotes lym-
phocyte homeostasis, L-selectin ligand formation, and traf-
ficking to lymph nodes. Elevated levels of lymphocytes in the
blood of ppGalNAcT-1-deficient mice reflected an increase in
both T- and B-lymphocyte numbers (Fig. 4A). Moreover, re-
duced total cellularity was detected among the lymph nodes,
while no changes in cellularity occurred within the spleen,
thymus, bone marrow, and Peyer’s patch tissues (Fig. 4B).
Mesenteric, cervical, and axillary lymph nodes were reduced in
total cellularity by an average of 60%, while the inguinal lymph
node lost 90% of the normal cell number and, in some cases,
was anatomically absent. Characterization of cell populations
within the lymph nodes revealed deficiencies in both T and B
lymphocytes, although B-lymphocyte numbers were most se-
verely reduced, by 50 to 95% (Fig. 4C). T-lymphocyte cellu-
larity was also reduced, but not as markedly, with the exception
of the inguinal lymph node tissue (Fig. 4D).
Lymph node architecture, comprising cellular and molecular
determinants, was examined among lymphocyte subpopula-
tions and endothelial cells that express selectin ligands. A re-
duced anatomic size and reduced numbers of lymph node
follicles were observed for ppGalNAcT-1 deficiency, with no
apparent change in the follicular trafficking pattern of B and T
cells (Fig. 5A). Levels of L-selectin receptors on the surfaces of
FIG. 2. Altered blood lymphocyte levels with reduced protein O
glycosylation. (A) Leukocyte counts were obtained from blood of wild-
type and ppGalNAcT-1-deficient mice. Cell numbers are expressed as
means per ␮l of whole blood ⫾standard errors of the means (SEM)
among 21 mice, including littermates of the indicated genotypes. WBC,
white blood cell; Neu, neutrophil; Lym, lymphocyte; Mon, monocyte;
Eos, eosinophil; Bas, basophil; RBC, red blood cell; Plt, platelet. An
unpaired ttest indicated significance (***,P⬍0.001; *,P⬍0.05).
(B) Reduced protein O glycosylation on the cell surfaces of circulating
leukocytes, detected by PNA lectin binding. Reduced PNA binding
was observed among Gr1
⫹
neutrophils and B220
⫹
B cells. In contrast,
close-to-wild-type levels of O-glycans detected by PNA binding levels
were noted among unactivated circulating CD3
⫹
T cells. WT, wild
type.
TABLE 1. Hemostasis and coagulation
Parameter
Value
a
Wild-type
mice
ppGalNAcT-1-
deficient mice
Time (s)
Bleeding 75 ⫾100 239 ⫾257***
Prothrombin 10.3 ⫾0.4 10.5 ⫾0.3
APTT 26.8 ⫾2.0 29.6 ⫾3.6*
Protein level (% of
wild-type BL/6 level)
Antithrombin 136 ⫾17 123 ⫾16
Protein C 85 ⫾15 84 ⫾22
Protein S 153 ⫾102 166 ⫾112
Plasminogen 77 ⫾11 66 ⫾14*
Alpha-2 antiplasmin 121 ⫾26 107 ⫾17
Factor II 109 ⫾36 103 ⫾19
Factor V 124 ⫾17 105 ⫾17*
Factor VII 88 ⫾17 71 ⫾16*
Factor VIII 104 ⫾28 70 ⫾15*
Factor IX 88 ⫾15 65 ⫾6**
Factor X 102 ⫾21 86 ⫾17*
Factor XI 84 ⫾14 69 ⫾21
Factor XII 105 ⫾16 85 ⫾14*
VWF 204 ⫾64 296 ⫾136***
a
Results are expressed as means ⫾SD (***, P⬍0.001; **, P⬍0.01; *, P⬍
0.05). The number of samples studied was 14 for wild-type mice and 10 for ppGal
NAcT-1-deficient mice, except for bleeding time analysis. For that analysis, the
number of mice studied was 39 for wild-type mice and 34 for ppGalNAcT-1-
deficient mice.
VOL. 27, 2007 O GLYCOSYLATION IN VASCULAR AND IMMUNE BIOLOGY 8787

T and B cells from various compartments were analyzed and
found to be normal (Fig. 5B). In contrast, L-selectin ligands
that are normally expressed on high endothelial venules
(HEVs) were significantly reduced in the absence of ppGal
NAcT-1 among mesenteric and inguinal lymph nodes (Fig.
5C). The inguinal node was the most severely affected node
compared with mesenteric lymph node tissue. Approxi-
mately 10% of normal L-selectin ligand expression re-
mained among inguinal lymph nodes, while close to 60% of
normal L-selectin ligand levels remained among mesenteric
lymph nodes (Fig. 5D). Interestingly, a significant decrease
in MECA-79 antibody binding occurred only among ingui-
nal HEVs, while normal MECA-79 levels were expressed
among mesenteric lymph nodes. The degrees of L-selectin
ligand deficiency among different lymphoid aggregates were
proportional to the degrees of reduced lymphocyte cellular-
ity observed, implicating reduced L-selectin ligand levels in
lymph node HEVs as being responsible for reduced cellu-
larity. Attempts to further investigate leukocyte adhesion by
intravital microscopy were unsuccessful due to the extensive
bleeding and hemorrhaging that occurred during surgery, as
noted above. Nevertheless, the efficacy of lymphocyte hom-
ing could be measured.
Isolated T and B cells from wild-type mice were covalently
labeled using CMFDA and injected into the tail veins of syn-
geneic wild-type and ppGalNAcT-1-deficient littermate recip-
ients. Levels of CMFDA-labeled lymphocytes among lymphoid
aggregates were analyzed 1 h and 24 h following transfer.
CMFDA-labeled B and T lymphocytes were detected in all
lymphoid tissues surveyed within 1 h. However, a significant
reduction in B-cell homing was observed among ppGalNAcT-
1-deficient recipients, with the exception of Peyer’s patch tissue
(Fig. 6A). At 24 h posttransplantation, the most severe B-cell
homing deficit continued to involve the inguinal lymph nodes
of recipients, followed in severity by homing to axillary, cervi-
cal, and mesenteric lymph nodes, with a normal homing fre-
quency for Peyer’s patch tissue (Fig. 6B). Lesser but significant
deficits in T-cell homing were also observed and occurred with
a similar profile, with the most significant decrease in homing
to the inguinal lymph nodes. These findings reveal that the
reduced cellularity among lymph nodes is proportional to
the severity of the lymphocyte homing defect and tracks with
the degree of L-selectin ligand deficiency among the HEVs of
ppGalNAcT-1-deficient recipients.
ppGalNAcT-1 supports IgG production in the humoral im-
mune response. Circulating levels of Ig isotypes IgG1, IgG2a,
and IgG3 were significantly reduced in the sera of ppGal
NAcT-1-deficient mice, while normal levels of IgM and IgA
FIG. 3. Reduced selectin ligand expression on ppGalNAcT-1-deficient neutrophils attenuates selectin-mediated rolling and inflammation in a
model of acute peritonitis. (A) Reduced E- and P-selectin ligands among circulating Gr1
⫹
cells, detected by selectin-IgM chimera binding and flow
cytometry. Loss of selectin-IgM chimera binding in the presence of EDTA is also shown. (B) Expression levels of adhesion molecules (CD11a,
CD11b, CD18, and CD62L) and selectin counterreceptors (CD24 and PSGL-1) on Gr1
⫹
cells in blood, analyzed by flow cytometry. The results
shown in panels A and B are representative of three experiments with separate littermates. (C) Numbers of rolling leukocytes on P-selectin and
E-selectin were reduced in ppGalNAcT-1-deficient mice, analyzed using blood-perfused microflow chambers at 5.9 dynes/cm
2
. Data presented are
means ⫾SEM for at least six chambers, using cells from three separate littermates of the indicated genotypes. *,P⬍0.05 compared to ppGal
NAcT-1. FOV, field of view. (D) Peritoneal cells were collected, and Gr1
⫹
neutrophils were counted prior to (0 h) or after (2 or 4 h) intraperitoneal
injection of casein to initiate acute peritonitis. Data are presented as means ⫾SEM for eight mice representing littermates of the indicated
genotypes (***,P⬍0.001). WT, wild type.
8788 TENNO ET AL. MOL.CELL.BIOL.

were detected (Fig. 7A). It seemed unlikely that this effect was
due to a reduction in L-selectin function, as such animals have
been reported to express normal to elevated Ig levels and to
exhibit a robust humoral immune response (2, 49). We further
noted that B-cell surface expression of major histocompatibil-
ity complex class II and costimulatory molecules CD40, CD44,
and B7.2 was unaltered in the absence of ppGalNAcT-1 (Fig.
7B). Moreover, B-cell proliferation responses to LPS or anti-
body-mediated IgM cross-linking were also unaffected (Fig.
7C). In contrast, upon immunization with the T-cell-indepen-
dent antigen DNP-Ficoll, ppGalNAcT-1-deficient mice failed
to induce a significant IgG antibody titer in the presence of
normal titer increases involving IgM and IgA isotypes (Fig.
7D). Similarly, a deficit of anti-DNP IgG antibody production
was detected following immunization with the T-cell-depen-
dent antigen DNP-KLH (Fig. 7E).
ppGalNAcT-1 deficiency attenuates GC formation upon im-
munization. The development of B cells that can express IgG
isotypes takes place within GCs of follicles within peripheral
lymphoid tissues in response to antigenic exposure. The GC
microenvironment is essential for Ig isotype class switching and
affinity maturation in producing plasma B cells that provide the
circulating repertoire of IgG antibodies. Following immuniza-
tion of ppGalNAcT-1-deficient mice, we noted a threefold
reduction in the numbers of detectable GCs within the white
pulp of the spleen and lymph nodes compared with the num-
bers for wild-type littermates (Fig. 8A). Localization of B and
T cells and expression of FDCM1
⫹
dendritic cells appeared
unaltered among follicles of ppGalNAcT-1-deficient mice (Fig.
8B). Although GC B cells were detected with the antibody
GL7 and decreased IgD expression resulted as expected (not
shown), another marker of GC B cells, the PNA lectin (31),
could not detect GC B cells in ppGalNAcT-1-deficient tis-
sues (Fig. 8C). This finding indicates that ppGalNAcT-1 is
essential for synthesizing O-glycans among GC B cells that
are typically detected by PNA binding. In addition, splenic
B-cell levels as well as GL7
⫹
GC B-cell numbers were sig-
nificantly decreased in ppGalNAcT-1 deficiency at 8 days
postimmunization (Fig. 8D). These results indicate that
ppGalNAcT-1 deficiency significantly decreases O-glycans
on GC B cells that normally bear PNA lectin binding deter-
minants and further diminishes the number of GC B cells,
with reduced GC formation.
Increased apoptosis of GC B cells in ppGalNAcT-1 defi-
ciency. The reductions in frequency of GCs and total GC B
cells in ppGalNAcT-1-deficient mice did not appear to be due
to impaired B-cell proliferation. Either by ex vivo B-cell stim-
ulation assays or by in situ analysis of GC expression of the
nuclear proliferation marker Ki-67, B cells lacking ppGal
NAcT-1 appeared unaltered in their responses to stimuli that
lead to cell proliferation (Fig. 7C and data not shown). How-
ever, a reduction in GC B cells was associated with an in-
creased frequency of CD68
⫹
tingible-body macrophages. Cel-
lular markers of apoptosis were significantly induced in GCs
from ppGalNAcT-1-deficient mice. Activated caspase-3 and
DNA fragmentation detected by TUNEL were greatly ele-
vated and often colocalized with CD68
⫹
macrophages (Fig. 9A
to C). The GC has a polar two-compartment structure, includ-
ing a dark zone (centroblasts) and a light zone (centrocytes).
The dark zone is proximal to the T-cell area and contains
rapidly dividing Ig
⫺
B cells called centroblasts, while the light
zone contains Ig
⫹
centrocytes, which undergo selection based
on the affinity of their Ig receptor for the antigen. Following
immunization, centroblasts and centrocytes were measured as
previously described (64). An alteration in the frequency of
viable centroblasts and centrocytes was observed, reflecting a
significant reduction in centrocyte number (Fig. 9D). This was
accompanied by significant increases in activated caspase-3
abundance among centrocytes, and especially centroblasts,
during ppGalNAcT-1 deficiency (Fig. 9E).
IgG abundance was significantly reduced among GCs in
ppGalNAcT-1-deficient tissues, in the context of increased
apoptosis among IgG-positive cells (Fig. 10A). Moreover, re-
duced numbers of IgG
⫹
B lymphocytes were observed in the
spleens of immunized ppGalNAcT-1-deficient mice (Fig. 10B).
In addition, the frequency and number of plasma B cells were
also markedly reduced (Fig. 10C). These findings establish an
essential role of ppGalNAcT-1 in promoting plasma B-cell
abundance and IgG formation by moderating the apoptosis of
centrocytes bearing IgGs among GC B cells.
FIG. 4. Leukocytosis involving T and B cells is associated with a
reduction in lymph node cellularity. (A) Circulating numbers of T cells
(CD3
⫹
) and B cells (CD19
⫹
) were significantly increased in ppGalNAcT-
1-deficient mice. (B) Leukocyte abundance was reduced among dis-
tinct peripheral lymphoid aggregates, including inguinal lymph nodes
(ILN), axillary lymph nodes (ALN), cervical lymph nodes (CLN), and
mesenteric lymph nodes (MLN). No significant change occurred
among the Peyer’s patch (PP), spleen (Spl), thymus (Thy), and bone
marrow (BM) compartments. (C) B-lymphocyte (CD19
⫹
) numbers
were significantly reduced among multiple lymphoid aggregates.
(D) T-lymphocyte (CD3
⫹
) numbers were reduced most significantly
among inguinal lymph nodes. Data are means ⫾SEM for 12 litter-
mate pairs of the indicated genotypes. An unpaired ttest indicated
significance (***,P⬍0.001; **,P⬍0.01; *,P⬍0.05). WT, wild
type.
VOL. 27, 2007 O GLYCOSYLATION IN VASCULAR AND IMMUNE BIOLOGY 8789

DISCUSSION
A significant proportion of nascent vertebrate proteins tran-
siting the secretory pathway are O glycosylated by the action of
one or more members of a conserved family of ppGalNAcTs.
Different genes encode the various ppGalNAcTs among all
multicellular organisms surveyed, and thus it appears likely
that important and often nonoverlapping physiologic activities
will be attributed to individual members of this glycosyltrans-
ferase family (34). The production and study of intact organ-
isms that lack individual ppGalNAcTs, including nematodes,
insects, rodents, and primates, reveal remarkable cell type
specificity in the phenotypes and pathologies arising from these
deficiencies. Although ppGalNAcT-1 was the first member of
the ppGalNAcT family to be identified and characterized be-
cause it is highly expressed among many cell types, the biolog-
ical roles of this glycosyltransferase in mammalian develop-
ment and physiology have remained unknown. By engineering
an inheritable germ line mutation in the mouse gene that
disables ppGalNAcT-1 function, we have established a ppGal
NAcT-1 deficiency state for initial and future investigations
into the biology of core-type protein O glycosylation contrib-
uted by this highly conserved glycosyltransferase. Our analyses
thus far have shown that ppGalNAcT-1 markedly supports the
activities of glycoproteins that contribute to blood coagulation,
selectin-mediated leukocyte trafficking and inflammation, and
GC B-cell apoptosis in modulation of the humoral immune
response. Although decreased B-cell trafficking to lymph nodes
and increased apoptosis do not appear to be connected mech-
anistically, they are likely collaborative in the observed reduc-
tion of humoral immunity. The focus of ppGalNAcT-1 func-
FIG. 5. Tissue-selective deficiency in L-selectin ligand and MECA-79 epitope expression among peripheral lymph nodes. (A) Histological
analysis of inguinal (ILN) and mesenteric (MLN) lymph nodes incubated with CD4-FITC, CD8-FITC, and B220-Rho and visualized by
fluorescence microscopy. Magnification, ⫻100. (B) Unaltered expression levels of L-selectin on CD19
⫹
and CD3
⫹
lymphocytes in spleen, blood,
and inguinal and mesenteric lymph nodes, as indicated by flow cytometry. Dotted lines represent background cell fluorescence using isotype-
specific control antibodies. (C) Frozen sections of inguinal and mesenteric lymph nodes stained with L-selectin–IgG chimera or MECA-79
antibody. Magnification, ⫻400. (D) Relative levels of L-selectin ligands and MECA-79 expression on HEVs were obtained by quantifying
fluorescent signals from serial and parallel tissue sections of inguinal lymph nodes (ILN) and mesenteric lymph nodes (MLN), using deconvolution
microscopy and MetaMorph software analysis (see Materials and Methods). WT, wild type.
FIG. 6. ppGalNAcT-1 supports lymphocyte homing to specific
lymph nodes. CMFDA-labeled lymphocytes (2.5 ⫻10
7
) obtained from
wild-type (WT) mice were injected into the tail veins of recipients of
the indicated genotypes. Lymphoid aggregates, as denoted in the leg-
end to Fig. 4, were harvested 1 h (A) or 24 h (B) after injection.
CMFDA
⫹
T and B lymphocytes were quantified by flow cytometry.
Data are means ⫾SEM for eight mice of each genotype. An unpaired
ttest indicated significance (***,P⬍0.001; **,P⬍0.01; *,P⬍0.05).
8790 TENNO ET AL. MOL.CELL.BIOL.

tion upon selected cell types and glycoproteins enables further
studies on the precise mechanisms by which the initiation of
protein O glycosylation by ppGalNAcT-1 plays an advanta-
geous role in vascular and immune responses. Although ad-
vantageous roles of ppGalNAcT-1 in physiology are evident,
they appear to be dispensable for normal development and
reproduction among laboratory-raised animals (32, 37). More-
over, inhibition of ppGalNAcT-1 activity may have therapeutic
potential for some pathogenic syndromes involving increased
thrombosis, chronic inflammation, and immunologic diseases
of B lymphocytes.
In modulating blood coagulation levels, mild to moderate
decreases in circulating levels of factors V, VII, VIII, IX, X,
and XII, resulting in a prolonged APTT and bleeding time,
were evident in ppGalNAcT-1 deficiency. The reduced level of
factor VIII in the presence of an elevated level of VWF was a
surprising result, since elevated levels of VWF tend to raise
factor VIII levels. We found that VWF produced by ppGal
FIG. 7. Normal antigen receptor activation contrasts with attenuated antibody production due to loss of ppGalNAcT-1. (A) Serum Ig levels
among 8-week-old naive mice (n⫽16). Points represent measurements for individual animals. The median Ig levels are depicted as horizontal bars
(means ⫾SEM are indicated). An unpaired ttest indicated significance (***,P⬍0.001; *,P⬍0.05). (B) The expression levels of activation
markers (CD44, B7.2, and I-A
b
) and CD40 were analyzed by flow cytometry on CD19
⫹
lymphocytes derived from the spleen (Spl), peripheral
lymph nodes (PLN), and mesenteric lymph nodes (MLN). The results shown are representative of three separate experiments. Dotted lines
represent the fluorescence of cells stained using an isotype-specific control antibody. (C) B lymphocytes were isolated and stimulated by antibody
to IgM or LPS. The proliferation response was measured by [
3
H]thymidine incorporation. Data are presented as means ⫾SEM for three mice
of the indicated genotypes. (D) Anti-DNP antibody levels produced in response to immunization with 10 ␮g of the T-cell-independent antigen
DNP-Ficoll, measured at the indicated times. (E) Anti-DNP antibody levels measured before and subsequent to a secondary immunization (arrow)
using 100 ␮g of the T-cell-dependent antigen DNP-KLH. Multiple dilutions of sera were assayed, with the results shown reflecting the linear
range of responses (OD
405
measurements) (data not shown). Data are presented as means ⫾SEM for eight mice of the indicated genotypes.
WT, wild type.
VOL. 27, 2007 O GLYCOSYLATION IN VASCULAR AND IMMUNE BIOLOGY 8791

NAcT-1-deficient mice has a normal capacity to bind factor
VIII but has a slightly reduced (by 10%) capacity to bind
collagen (data not shown). Typically, elevated VWF levels are
not associated with increased bleeding times. Thus, it is likely
that the reductions in multiple coagulation factors, namely,
factors V, VII, VIII, IX, and X, contribute to the hemostatic
defects seen in the null mice. Since O glycosylation is not
linked at present to cellular mechanisms that control protein
synthesis and maturation in the secretory pathway, it is possible
that ppGalNAcT-1 modulates binding activities, postsecretion
proteolysis, and turnover among various O glycoproteins in
circulation. Platelet homeostasis, in contrast, appeared to be
unaffected by ppGalNAcT-1 deficiency. Besides the presence
of normal platelet counts in circulation, no difference in ag-
gregation or activation occurred in response to various stimuli,
including arachidonic acid, collagen, ADP, thromboxane ana-
log, and calcium ionophore (data not shown). Nevertheless,
the prolonged bleeding time may reflect both altered coagula-
tion factor levels and a reduced level of P-selectin ligand for-
mation, which may diminish platelet function in hemostasis.
ppGalNAcT-1 further contributes to selectin ligand synthe-
sis, supporting the formation of a significant proportion of
ligands for E- and P-selectins on Gr-1
⫹
neutrophils in circu-
lation as well as L-selectin ligands produced among lymph
node HEVs. In comparison to the case for mice lacking core 2
GlcNAcT-1 (11), there remained slightly higher levels of both
E- and P-selectin ligands on neutrophils lacking ppGal
NAcT-1. This may explain the absence of neutrophilia in
ppGalNAcT-1-deficient mice, which exists among animals
lacking core 2 GlcNAcT-1. This implies that core 2 GlcNAcT-
1-dependent E- and P-selectin ligand synthesis can occur at
sites of protein O glycosylation initiated by other ppGalNAcT
isozymes expressed in neutrophils. Nevertheless, the loss of
ppGalNAcT-1 has a surprisingly severe impact on neutrophil
recruitment in an acute peritonitis model of endothelial in-
flammation, with a degree of reduction in neutrophil influx
comparable to that seen in animals lacking a larger apparent
proportion of selectin ligands in the absence of either core 2
GlcNAcT-1 or fucosyltransferase (FT) VII.
The absence of ppGalNAcT-1 impairs T- and especially
B-lymphocyte homing among peripheral lymphoid aggregates,
including axial, cervical, mesenteric, and inguinal lymph nodes,
resulting in a reduced number of lymphocytes among periph-
eral lymph node tissues. These findings were proportional to
the degree of L-selectin ligand deficiency in vivo among the
HEVs of various peripheral lymphoid aggregates. L-selectin
ligand synthesis is controlled by the actions of glycosyltrans-
ferases and sulfotransferases, which construct 6-sulfo-sialyl
Lewis
X
(sLe
X
) on glycan branches of several HEV-resident
sLe
X
counterreceptors, including GlyCAM-1, CD34, podoca-
lyxin, Sgp200, endoglycan, and MAdCAM-1 (14, 43). Collab-
orative and differential roles for each class of enzyme have
been observed in selectin ligand formation, involving lympho-
cyte homing and retention among peripheral lymph nodes (15,
21, 23). Glycoproteins bearing L-selectin ligands are typically
O glycosylated and have mucin-like domains for multiple O-
GalNAc linkages that act as the initial scaffolds. L-selectin
ligands can be constructed on core 1- and core 2-type O-glycan
branches, the former of which also harbor the MECA-79 an-
tigen and are prominent in the absence of core 2 GlcNAcT-1
FIG. 8. Diminished GC development and B-cell expansion upon
immunization of ppGalNAcT-1-deficient mice. (A) Splenic follicles
and GCs from mouse spleens prior to and 8 days subsequent to im-
munization with DNP-KLH were detected with hematoxylin and eosin
(H.E.) and with GL7 antibody. WP and RP indicate white pulp and red
pulp, respectively. Images were magnified 100⫻. Small arrows within
white pulp denote GC B cells. The percentage of white pulp bearing
GCs and the size of GCs were measured. (B) GC B cells in the spleen
were further analyzed in situ 14 days after immunization with DNP-
KLH. T- and B-cell zones remained intact, and follicular dendritic cells
(FDCM1
⫹
) were present at normal levels. (C) GC markers include
GL7 antibody binding among B220
⫹
B cells, while PNA binding de-
pendent upon O glycosylation was substantially decreased in ppGal
NAcT-1-deficient mice. Magnification (B and C), ⫻400. (D) Reduced
frequency of splenic GC B cells (B220
⫹
GL7
⫹
) as well as reduced total
splenic B220
⫹
B-cell number among ppGalNAcT-1-deficient mice 8
days after immunization with DNP-KLH. Results shown are represen-
tative of data obtained from analyses of three to six littermates of the
indicated genotypes. WT, wild type.
8792 TENNO ET AL. MOL.CELL.BIOL.

(65). Interestingly, normal levels of MECA-79 expression were
detected among mesenteric lymph nodes of ppGalNAcT-1-
deficient mice in the presence of a moderate reduction in
L-selectin ligands, implying that mesenteric L-selectin ligands
generated by ppGalNAcT-1 may lack the MECA-79 determi-
nant. While glycosyltransferases other than ppGalNAcT-1 con-
tribute to L-selectin ligand formation, L-selectin ligands pro-
duced within the inguinal lymph node are particularly
dependent upon ppGalNAcT-1. The diminished level of L-
selectin ligands is likely involved in the reduction of lympho-
cyte homing and abundance among various lymphoid aggre-
gates during ppGalNAcT-1 deficiency.
Selective differences in the homing abilities of T and B cells
have been described, independent of their site of origin, among
the spleen, peripheral lymph nodes, mesenteric lymph nodes,
and Peyer’s patches (50). L-selectin expression levels appear to
be a key contributor to this effect, as T lymphocytes express a
twofold higher level of L-selectin than do B lymphocytes and
are more efficient at homing to peripheral lymph nodes (50,
52). The absence of L-selectin expression on lymphocytes
markedly reduces both T- and B-lymphocyte homing and lo-
calization among peripheral lymph nodes, as expected, but
does not significantly alter lymphocyte numbers in circulation
or among mesenteric lymph nodes, the spleen, or thymus tissue
(2). Other distinctions are apparent. Splenomegaly is observed
in L-selectin deficiency, but not among ppGalNAcT-1-defi-
cient mice. The lesser defect in cellularity observed among
mesenteric lymph nodes in ppGalNAcT-1 deficiency may re-
FIG. 9. Loss of O-glycans produced by ppGalNAcT-1-deficient mice increases GC macrophage infiltration and B-cell apoptosis. (A) Abun-
dances of tingible-body macrophages (CD68
⫹
) and markers of cell apoptosis, including activated caspase-3 and TUNEL, were increased in ppGal
NAcT-1 deficiency among GC B cells analyzed in situ 14 days after immunization with DNP-KLH. Magnification, ⫻1,000. (B) CD68, activated
caspase-3, and TUNEL measurements obtained by quantifying fluorescent signals. (C) Percentage of CD68
⫹
macrophage colocalization with
apoptotic cell markers, including activated caspase-3 and TUNEL. (D) Centroblast (CB) and centrocyte (CC) populations among splenic GC B
cells (B220
⫹
GL7
⫹
) measured 8 days after immunization with DNP-KLH by forward-scatter (FSC) flow cytometry. (E) Increased abundance of
activated caspase-3 detected among permeabilized centroblasts and centrocytes from ppGalNAcT-1-deficient mice by flow cytometry. Analyses
were accomplished 8 days after immunization with DNP-KLH. Experiments shown are representative of data obtained from three to six littermates
of the indicated genotypes. WT, wild type.
VOL. 27, 2007 O GLYCOSYLATION IN VASCULAR AND IMMUNE BIOLOGY 8793

flect the additional role of MAdCAM-1 as a ligand for the
␣4␤7 integrin expressed on lymphocytes (6).
Multiple glycosyltransferases have been identified directly
as participants in L-selectin ligand synthesis in vivo and pre-
viously included ST3Gal-IV, core 2 GlcNAcT-1, core 1
GlcNAcT, FT IV, and FT VII. Their individual and combined
contributions modulate selectin ligand synthesis differentially
among various cell types in vivo, often markedly reducing lym-
phocyte homing activity to lymph nodes (11, 13, 15, 25, 33, 36).
The absence of both FT IV and FT VII results in a profound
loss of lymphocyte numbers within peripheral lymph nodes,
and animals lacking these factors have impaired T-cell-medi-
ated immune responses (25, 39). Among core 2 GlcNAcT-1-
deficient mice, a lesser reduction in B-cell (28%) than in T-cell
numbers was reported for peripheral lymph nodes (15). The
loss of ppGalNAcT-1 by itself has a remarkably severe impact
on both leukocyte homing to HEVs and homeostatic periph-
eral lymph node cellularity, unlike other single glycosyltrans-
ferase deficiency states analyzed thus far. This may be ex-
plained in part by the more proximal position of ppGalNAcT-1
action in the pathway of core-type O glycosylation required for
selectin ligand synthesis. While the formation of selectin li-
gands may continue in part among N-glycans (36), or perhaps
via O glycosylation by other ppGalNAcT isozymes, core 1 and
core 2 O-glycan branching cannot compensate in the absence
of the ppGalNAcT-1-dependent linkage of GalNAc to serine
or threonine.
The essential role of L-selectin ligands in lymphoid homing
and lymph node homeostasis contrasts with findings regarding
B-cell immune function. Selectin ligand expression does not
correlate with Ig levels in circulation, nor does L-selectin de-
ficiency dampen humoral immunity in assays of antigen-spe-
cific antibody production upon immunization. Humoral im-
mune responses are, in fact, increased in the absence of
L-selectin (49). These findings nevertheless imply that L-selec-
tin likely has additional physiologic roles distinct from the
effects of binding to its sLe
X
glycan ligand. Our findings among
ppGalNAcT-1-deficient mice revealed a significant humoral
immune deficit that appeared to be independent of B-cell
immune responses and altered leukocyte homing due to re-
duced L-selectin ligand levels. No alterations in B-cell immune
responsiveness to LPS or IgM antigen receptor stimulation
were observed in ppGalNAcT-1 deficiency, along with contin-
ued GC B-cell proliferation, and GC B-cell apoptosis was
significantly increased. Moreover, B-cell trafficking and abun-
dance in the spleen were normal, while reduced GCs associ-
ated with apoptotic B cells were nevertheless also observed in
this tissue.
The abundance of circulating IgG was significantly reduced
among ppGalNAcT-1-deficient mice, both prior to and subse-
quent to immunization with either T-cell-independent or T-
cell-dependent antigens. This deficit did not extend to IgM or
IgA, which were present at normal abundances. Centroblast
and centrocyte apoptosis was elevated, with significantly de-
creased centrocyte numbers in ppGalNAcT-1-deficient GCs
and a significant reduction of IgG expression and plasma B-cell
abundance. Some IgG production continued and a reduced
number of plasma cells remained, implicating a robust modu-
latory effect of ppGalNAcT-1 on this regulatory and differen-
tiation process in the B-cell-mediated humoral immune re-
sponse. ppGalNAcT-1 normally diminishes apoptotic signaling
in GCs during the development of IgG-secreting plasma B
cells. Yet levels of Fas at the B-cell surface were normal in
ppGalNAcT-1 deficiency, and no alterations in the expression
of other cell surface glycoproteins were as yet detected (data
FIG. 10. Reduced IgG and plasma B-cell abundances during ppGal
NAcT-1 deficiency. (A) Splenic GCs analyzed 14 days after immuni-
zation with DNP-KLH, indicating a reduced abundance of IgG as well
as colocalization of the apoptosis marker activated caspase-3 with
IgG-expressing B cells. Magnification, ⫻1,000. (B) Reduced frequency
and number of viable IgG
⫹
IgM
⫺
B220
⫹
B cells in the spleens of
ppGalNAcT-1-deficient mice 8 days after immunization with DNP-
KLH. (C) Reduced frequency and number of splenic plasma B cells
(CD138
⫹
) in ppGalNAcT-1-deficient mice, measured 8 days after im-
munization with DNP-KLH. Experiments shown are representative of
data obtained from analyses of three to six littermates of the indicated
genotypes.
8794 TENNO ET AL. MOL.CELL.BIOL.

not shown). Whether GC B-cell apoptosis in ppGalNAcT-1
deficiency is due to altered O glycosylation among B cells,
dendritic cells, or possibly stromal cells that support the mi-
croenvironment is currently unknown. Such studies are neces-
sary to ultimately identify the glycoprotein(s) involved and to
further delineate the precise mechanisms by which ppGal
NAcT-1-dependent protein O glycosylation modulates GC B-
cell apoptosis and thereby the humoral immune response
in vivo.
The initiation of protein O glycosylation in vivo by ppGal
NAcT-1 translates into vital roles in vascular biology, hemo-
stasis, and humoral immunity. Core-type protein O glycosyla-
tion is generally known to support mucin structure and func-
tion on endothelial cell surfaces, protecting cells against stress
and pathogen infection (3, 51). Furthermore, it was recently
shown that a loss of core 3 O-glycans in mice increases their
susceptibility to colitis and colorectal tumors (1). Core-type O
glycosylation is often produced following a stepwise and hier-
archical pattern of peptide modification by multiple ppGal
NAcT isozymes in generating a high-density array of O-glycans
that assist in tissue hydration and can compete to block recep-
tors carried by mucosal pathogens. Such physiologic roles,
which have yet to be probed, may indeed be modulated by
ppGalNAcT-1 and perhaps other ppGalNAcTs that normally
act subsequently but fail to O glycosylate otherwise native
glycopeptide sequences. Although ppGalNAcT-1 deficiency is
likely compensated for to some degree in the intact mouse by
other ppGalNAcT family members, the endogenous physio-
logic functions that we can now assign to ppGalNAcT-1 in
blood coagulation, leukocyte trafficking, and humoral immu-
nity are clearly dependent upon the substrate specificity and
expression of this glycosyltransferase. The biologic activities of
protein O glycosylation and the mechanisms by which this
abundant posttranslational modification regulates cellular bi-
ology can be further resolved by experimental approaches that
incorporate genetic, biochemical, and physiologic analyses of
intact organisms.
ACKNOWLEDGMENTS
We thank Kurt Marek, James Yousif, and Yan Wang for expert
assistance.
This research was partly funded by NIH grant DK48247 (J.D.M.)
and by German Research Foundation grant AZ 428/2-1 (to A.Z.).
J.D.M. is supported as an investigator of the Howard Hughes Medical
Institute.
REFERENCES
1. An, G., B. Wei, J. M. McDaniel, T. Ju, R. D. Cummings, J. Braun, and L. Xia.
2007. Increased susceptibility to colitis and colorectal tumors in mice lacking
core 3-derived O-glycans. J. Exp. Med. 204:1417–1429.
2. Arbones, M. L., D. C. Ord, K. Ley, H. Ratech, C. Maynard-Curry, G. Otten,
D. J. Capon, and T. F. Tedder. 1994. Lymphocyte homing and leukocyte
rolling and migration are impaired in L-selectin-deficient mice. Immunity
1:247–260.
3. Bansil, R., E. Stanley, and J. T. LaMont. 1995. Mucin biophysics. Annu. Rev.
Physiol. 57:635–657.
4. Barbieri, A. M., M. Filopanti, G. Bua, and P. Beck-Peccoz. 2007. Two novel
nonsense mutations in GALNT3 gene are responsible for familial tumoral
calcinosis. J. Hum. Genet. 52:464–468.
5. Bennet, E. P., H. Hassan, U. Mandel, E. Mirgorodskaya, P. Roepstorff,
J. Burchell, J. Taylor-Papadimitriou, M. A. Hollingsworth, G. Merkx, A.
Geurts van Kessel, H. Eiberg, R. Steffensen, and H. Clausen. 1998. Cloning
of a human UDP-N-acetyl-␣-D-galactosamine:polypeptide N-acetylgalac-
tosaminyltransferase that complements other GalNAc-transferases in com-
plete O-glycosylation of the MUC1 tandem repeat. J. Biol. Chem. 273:
30472–30481.
6. Berlin, C., E. L. Berg, M. J. Briskin, D. P. Andrew, P. J. Kilshaw, B.
Holzmann, I. L. Weissman, A. Hamann, and E. C. Butcher. 1993. Alpha 4
beta 7 integrin mediates lymphocyte binding to the mucosal vascular ad-
dressin MAdCAM-1. Cell 74:185–195.
7. Brockhausen, I., D. Toki, J. Brockhausen, S. Peters, T. Bielfeldt, A. Kleen, H.
Paulsen, M. Meldal, F. Hagen, and L. A. Tabak. 1996. Specificity of O-
glycosylation by bovine colostrums UDP-GalNAc:polypeptide ␣-N-acetyl-
galactosaminyltransferase using synthetic glycopeptide substrates. Glyco-
conj. J. 13:849–856.
8. Brockhausen, I. 2000. Biosynthesis of the O-glycan chains of mucins and
mucin-type glycoproteins, p. 313–328. In B. Ernst, G. W. Hart, and P. Sinay
(ed.), Carbohydrates in chemistry and biology, vol. 3. Wiley-VCH, New
York, NY.
9. Cheng, L., K. Tachibana, H. Iwasaki, A. Kameyama, Y. Zhang, T. Kubota, T.
Hiruma, K. Tachibana, T. Kudo, J.-M. Guo, and H. Narimatsu. 2004. Char-
acterization of a novel human UDP-GalNAc transferase, ppGalNAc-T15.
FEBS Lett. 566:17–24.
10. Elhammer, A., and S. Kornfeld. 1986. Purification and characterization of
UDP-N-acetylgalactosamine:polypeptide N-acetylgalactosaminyltransferase
from bovine colostrum and murine lymphoma BW5147 cells. J. Biol. Chem.
261:5249–5255.
11. Ellies, L. G., S. Tsuboi, B. Petryniak, J. B. Lowe, M. Fukuda, and J. D.
Marth. 1998. Core 2 oligosaccharide biosynthesis distinguishes between se-
lectin ligands essential for leukocyte homing and inflammation. Immunity
9:881–890.
12. Ellies, L. G., D. Ditto, G. G. Levy, M. Wahrenbrock, D. Ginsburg, A. Varki,
D. Le, and J. D. Marth. 2002. Sialyltransferase ST3Gal-IV operates as a
dominant modifier of hemostasis by concealing asialoglycoprotein receptor
ligands. Proc. Natl. Acad. Sci. USA 99:10042–10047.
13. Ellies, L. G., M. Sperandio, G. H. Underhill, J. Yousif, M. Smith, J. J.
Priatel, G. S. Kansas, K. Ley, and J. D. Marth. 2002. Sialyltransferase
specificity in selectin ligand formation. Blood 100:3618–3625.
14. Fukuda, M., N. Hiraoka, T. O. Akama, and M. N. Fukuda. 2001. Carbohy-
drate-modifying sulfotransferases: structure, function, and pathophysiology.
J. Biol. Chem. 276:47747–47750.
15. Gauguet, J. M., S. D. Rosen, J. D. Marth, and U. H. von Andrian. 2004. Core
2 branching beta1,6-N-acetylglucosaminyltransferase and high endothelial
cell N-acetylglucosamine-6-sulfotransferase exert differential control over B-
and T-lymphocyte homing to peripheral lymph nodes. Blood 104:4104–4112.
16. Gerken, T. A., C. Tep, and J. Rarick. 2004. Role of peptide sequence and
neighboring residue glycosylation on the substrate specificity of the uridine
5⬘-␣-N-acetylgalactosamine:poly-peptide N-acetylgalactosaminyltransferases
T1 and T2: kinetic modeling of the porcine and canine submaxillary gland
mucin tandem repeats. Biochemistry 43:9888–9900.
17. Hagen, F. K., B. Van Wuyckhuyse, and L. A. Tabak. 1993. Purification,
cloning, and expression of a bovine UDP-GalNAc:polypeptide N-acetyl-
galactosaminyltransferase. J. Biol. Chem. 268:18960–18965.
18. Hagen, F. K., C. A. Gregoire, and L. A. Tabak. 1995. Cloning and sequence
homology of a rat UDP-GalNAc:polypeptide N-acetylgalactosaminyltrans-
ferase. Glycoconj. J. 12:901–909.
19. Hagen, F. K., K. G. Ten Hagen, T. M. Beres, M. Balys, B. C. Van
Wuyckhuyse, and L. A. Tabak. 1997. cDNA cloning and expression of a
novel UDP-GalNAc:polypeptide N-acetylgalactosaminyltransferase. J. Biol.
Chem. 272:13843–13848.
20. Hanisch, F.-G., S. Muller, H. Hassan, H. Clausen, N. Zachara, A. A. Gooley,
H. Paulsen, K. Alving, and J. Peter-Katalinic. 1999. Dynamic epigenetic
regulation of initial O-glycosylation by UDP-N-acetylgalactosamine:peptide
N-acetylgalactosaminyltransferases. J. Biol. Chem. 274:9946–9954.
21. Hemmerich, S., A. Bistrup, M. S. Singer, A. van Zante, J. K. Lee, D. Tsay, M.
Peters, J. L. Carminati, T. J. Brennan, K. Carver-Moore, et al. 2001. Sulfa-
tion of L-selectin ligands by an HEV-restricted sulfotransferase regulates
lymphocyte homing to lymph nodes. Immunity 15:237–247.
22. Hennet, T., F. K. Hagen, L. A. Tabak, and J. D. Marth. 1995. T-cell-specific
deletion of a polypeptide N-acetylgalactosaminyl-transferase gene by site-
directed recombination. Proc. Natl. Acad. Sci. USA 92:12070–12074.
23. Hiraoka, N., H. Kawashima, B. Petryniak, J. Nakayama, J. Mitoma, J. D.
Marth, J. B. Lowe, and M. Fukuda. 2004. Core 2 branching beta1,6-N-
acetylglucosaminyltransferase and high endothelial venule-restricted sulfo-
transferase collaboratively control lymphocyte homing. J. Biol. Chem. 279:
3058–3067.
24. Homa, F. L., T. Hollander, D. J. Lehman, D. R. Thomsen, and A. P. Elhammer.
1993. Isolation and expression of a cDNA clone encoding a bovine UDP-
GalNAc:polypeptide N-acetylgalactosaminyltransferase. J. Biol. Chem. 268:
12609–12616.
25. Homeister, J. W., A. D. Thall, B. Petryniak, P. Maly, C. E. Rogers, P. L.
Smith, R. J. Kelly, K. M. Gersten, S. W. Askari, G. Cheng, et al. 2001. The
alpha(1,3)fucosyltransferases FucT-IV and FucT-VII exert collaborative
control over selectin-dependent leukocyte recruitment and lymphocyte hom-
ing. Immunity 15:115–126.
26. Ichikawa, S., K. W. Lyles, and M. J. Econs. 2005. A novel GALNT3 mutation
in a pseudoautosomal dominant form of tumoral calcinosis: evidence that
VOL. 27, 2007 O GLYCOSYLATION IN VASCULAR AND IMMUNE BIOLOGY 8795

the disorder is autosomal recessive. J. Clin. Endocrinol. Metab. 90:2420–
2423.
27. Ju, T., and R. D. Cummings. 2005. Protein glycosylation: chaperone muta-
tion in Tn syndrome. Nature 437:1252.
28. Kato, K., H. Takeuchi, N. Miyahara, A. Kanoh, H. Hassan, H. Clausen, and
T. Irimura. 2001. Distinct order of GalNAc incorporation into a peptide with
consecutive threonines. Biochem. Biophys. Res. Commun. 287:110–115.
29. Kato, K., C. Jeanneau, M. A. Tarp, A. Benet-Pages, B. Lorenz-Depiereux,
E. P. Bennett, U. Mandel, T. M. Strom, and H. Clausen. 2006. Polypeptide
GalNAc-transferase T3 and familial tumoral calcinosis. Secretion of fibro-
blast growth factor 23 requires O-glycosylation. J. Biol. Chem. 281:18370–
18377.
30. Kingsley, P. D., K. G. Ten Hagen, K. M. Maltby, J. Zara, and L. A. Tabak.
2000. Diverse spatial expression patterns of UDP-GalNAc:polypeptide N-
acetylgalactosaminyltransferase family member mRNAs during mouse de-
velopment. Glycobiology 10:1317–1323.
31. Kraal, G., I. L. Weissman, and E. C. Butcher. 1982. Germinal centre B cells:
antigen specificity and changes in heavy chain class expression. Nature 298:
377–379.
32. Lowe, J. B., and J. D. Marth. 2003. A genetic approach to mammalian glycan
function. Annu. Rev. Biochem. 72:643–691.
33. Maly, P., A. Thall, B. Petryniak, C. E. Rogers, P. L. Smith, R. M. Marks,
R. J. Kelly, K. M. Gersten, G. Cheng, T. L. Saunders, et al. 1996. The
alpha(1,3)fucosyltransferase Fuc-TVII controls leukocyte trafficking through
an essential role in L-, E-, and P-selectin ligand biosynthesis. Cell 86:643–
653.
34. Marth, J. D. 1996. Complexity in O-linked oligosaccharide biosynthesis
engendered by multiple polypeptide N-acetylgalactosaminyltransferases.
Glycobiology 6:701–705.
35. McGuire, E. J., and S. Roseman. 1967. Enzymatic synthesis of the protein-
hexosamine linkage in sheep submaxillary mucin. J. Biol. Chem. 242:3745–
3755.
36. Mitoma, J., X. Bao, B. Petryniak, P. Schaerli, J. M. Gauget, S. Y. Yu, H.
Kawashima, H. Saito, K. Ohtsubo, J. D. Marth, K. H. Khoo, U. H. von
Andrian, J. B. Lowe, and M. Fukuda. 2007. Critical functions of N-glycans in
L-selectin-mediated lymphocyte homing and recruitment. Nat. Immunol.
8:409–418.
37. Nehrke, K., F. K. Hagen, and L. A. Tabak. 1998. Isoform-specific O-glyco-
sylation by murine UDP-GalNAc:polypeptide N-acetylgalactosaminyltrans-
ferase-T3, in vivo. Glycobiology 8:367–371.
38. Ohtsubo, K., and J. D. Marth. 2006. Glycosylation in cellular mechanisms of
health and disease. Cell 126:855–867.
39. Piccio, L., B. Rossi, L. Colantonio, R. Grenningloh, A. Gho, L. Ottoboni,
J. W. Homeister, E. Scarpini, M. Martinello, C. Laudanna, D. D’Ambrosio,
J. B. Lowe, and G. Constantin. 2005. Efficient recruitment of lymphocytes in
inflamed brain venules requires expression of cutaneous lymphocyte antigen
and fucosyltransferase-VII. J. Immunol. 174:5805–5813.
40. Pratt, M. R., H. C. Hang, K. G. Ten Hagen, J. Rarick, T. A. Gerken, L. A.
Tabak, and C. R. Bertozzi. 2004. Deconvoluting the functions of polypeptide
N-alpha-acetylgalactosaminyltransferase family members by glycopeptide
substrate profiling. Chem. Biol. 11:1009–1016.
41. Priatel, J. J., M. Sarkar, H. Schachter, and J. D. Marth. 1997. Isolation,
characterization and inactivation of the mouse Mgat3 gene: the bisecting
N-acetylglucosamine in asparagine-linked oligosaccharides appears dispens-
able for viability and reproduction. Glycobiology 7:45–56.
42. Priatel, J. J., D. Chui, N. Hiraoka, C. T. Simmons, K. B. Richardson, D. M.
Page, M. Fukuda, N. Varki, and J. D. Marth. 2000. The ST3Gal-I sialyl-
transferase controls CD8⫹T cell homeostasis by modulating O-glycan bio-
synthesis. Immunity 12:273–283.
43. Rosen, S. D. 2004. Ligands for L-selectin: homing, inflammation, and be-
yond. Annu. Rev. Immunol. 22:129–156.
44. Sastry, M. V., P. Banarjee, S. R. Patanjali, M. J. Swamy, G. V. Swarnalatha,
and A. Surolia. 1986. Analysis of saccharide binding to Artocarpus integri-
folia lectin reveals specific recognition of T-antigen (beta-D-Gal(1-3)D-
GalNAc). J. Biol. Chem. 261:11726–11733.
45. Schachter, H., and I. Brockhausen. 1989. The biosynthesis of branched
O-glycans. Symp. Soc. Exp. Biol. 43:1–26.
46. Schwientek, T., E. P. Bennett, C. Flores, J. Thacker, M. Hollmann, C. A.
Reis, J. Behrens, U. Mandel, B. Keck, M. A. Schafer, et al. 2002. Functional
conservation of subfamilies of putative UDP-N-acetylgalactosamine:polypep-
tide N-acetylgalactosaminyltransferases in Drosophila, Caenorhabditis elegans,
and mammals. One subfamily composed of l(2)35Aa is essential in Drosophila.
J. Biol. Chem. 277:22623–22638.
47. Shafi, R., S. P. Iyer, L. G. Ellies, N. O’Donnell, K. W. Marek, D. Chui, G. W.
Hart, and J. D. Marth. 2000. The O-GlcNAc transferase gene resides on the
X chromosome and is essential for embryonic stem cell viability and mouse
ontogeny. Proc. Natl. Acad. Sci. USA 97:5735–5739.
48. Sorensen, T., T. White, H. H. Wandall, A. K. Kristensen, P. Roepstorff, and
H. Clausen. 1995. UDP-N-acetyl-␣-D-galactosamine:polypeptide N-acetyl-
galactosaminyltransferase: identification and separation of two distinct trans-
ferase activities. J. Biol. Chem. 270:24166–24173.
49. Steeber, D. A., N. E. Green, S. Sato, and T. F. Tedder. 1996. Humoral
immune responses in L-selectin-deficient mice. J. Immunol. 157:4899–4907.
50. Stevens, S. K., I. L. Weissman, and E. C. Butcher. 1982. Differences in the
migration of B and T lymphocytes: organ-selective localization in vivo and
the role of lymphocyte-endothelial cell recognition. J. Immunol. 2:844–851.
51. Tabak, L. A. 1995. In defense of the oral cavity: structure, biosynthesis, and
function of salivary mucins. Annu. Rev. Physiol. 57:547–564.
52. Tang, M. L., D. A. Steeber, X. Q. Zhang, and T. F. Tedder. 1998. Intrinsic
differences in L-selectin expression levels affect T and B lymphocyte subset-
specific recirculation pathways. J. Immunol. 160:5113–5121.
53. Ten Hagen, K. G., D. Tetaert, F. K. Hagen, C. Richet, T. M. Beres, J.
Gagnon, M. M. Balys, B. VanWuyckhuyse, G. S. Bedi, P. Degand, and L. A.
Tabak. 1999. Characterization of a UDP-GalNAc:polypeptide N-acetyl-
galactosaminyltransferase which displays glycopeptide N-acetylgalactosami-
nyltransferase activity. J. Biol. Chem. 274:27867–27874.
54. Ten Hagen, K. G., and D. T. Tran. 2002. A UDP-GalNAc:polypeptide
N-acetylgalactosaminyltransferase is essential for viability in Drosophila
melanogaster. J. Biol. Chem. 277:22616–22622.
55. Ten Hagen, K. G., T. A. Fritz, and L. A. Tabak. 2003. All in the family: the
UDP-GalNAc:polypeptide N-acetylgalactosaminyltransferases. Glycobiology
13:1R–16R.
56. Tian, E., and K. G. Ten Hagen. 2007. A UDP-GalNAc:polypeptide N-
acetylgalactosaminyltransferase is required for epithelial tube formation.
J. Biol. Chem. 282:606–614.
57. Topaz, O., D. L. Shurman, R. Bergman, M. Indelman, P. Ratajczak, M.
Mizrachi, Z. Khamaysi, D. Behar, D. Petronius, V. Friedman, I. Zelikovic, S.
Raimer, A. Metzker, G. Richard, and E. Sprecher. 2004. Mutations in
GALNT3, encoding a protein involved in O-linked glycosylation, cause of
familial tumoral calcinosis. Nat. Genet. 36:579–581.
58. Van den Steen, P., P. M. Rudd, R. A. Dwek, and G. Opdenakker. 1998.
Concepts and principles of O-linked glycosylation. Crit. Rev. Biochem. Mol.
Biol. 33:151–208.
59. Van Dyken, S. J., R. S. Green, and J. D. Marth. 2007. Structural and
mechanistic features of protein O glycosylation linked to CD8
⫹
T-cell
apoptosis. Mol. Cell. Biol. 27:1096–1111.
60. Wandall, H. H., H. Hassan, E. Mirgorodskaya, A. K. Kristensen, P. Roep-
storff, E. P. Bennett, P. A. Nielsen, M. A. Hollingsworth, J. Burchell, J.
Taylor-Papadimitriou, and H. Clausen. 1997. Substrate specificities of
three members of the human UDP-N-acetyl-␣-D-galactosamine:polypeptide
N-acetylgalactosaminyltransferase family, GalNAc-T1, -T2, and -T3. J. Biol.
Chem. 272:23503–23514.
61. Wang, Y., J. L. Abernathy, A. E. Eckhardt, and R. L. Hill. 1992. The acceptor
substrate specificity of porcine submaxillary UDP-GalNAc:polypeptide N-
acetylgalactosaminyltransferase is dependent on the amino acid sequences
adjacent to serine and threonine residues. J. Biol. Chem. 267:12709–12716.
62. Wang, Y., J. Tan, M. Sutton-Smith, D. Ditto, M. Panico, R. M. Campbell,
N. M. Varki, J. M. Long, J. Jaeken, S. R. Levinson, A. Wynshaw-Boris, H. R.
Morris, D. Le, A. Dell, H. Schachter, and J. D. Marth. 2001. Modeling
human congenital disorder of glycosylation type IIa in the mouse: conserva-
tion of asparagine-linked glycan-dependent functions in mammalian physi-
ology and insights into disease pathogenesis. Glycobiology 11:1–20.
63. White, T., E. P. Bennett, K. Takio, T. Sorensen, N. Bonding, and H. Clausen.
1995. Purification and cDNA cloning of a human UDP-N-acetyl-␣-D-galac-
tosamine:polypeptide N-acetylgalactosaminyltransferase. J. Biol. Chem. 270:
24156–24165.
64. Xiao, Y., J. Hendriks, P. Langerak, H. Jacobs, and J. Borst. 2004. CD27 is
acquired by primed B cells at the centroblast stage and promotes germinal
center formation. J. Immunol. 172:7432–7441.
65. Yeh, J. C., N. Hiraoka, B. Petryniak, J. Nakayama, L. G. Ellies, D. Rabuka,
O. Hindsgaul, J. D. Marth, J. B. Lowe, and M. Fukuda. 2001. Novel sulfated
lymphocyte homing receptors and their control by a Core1 extension beta
1,3-N-acetylglucosaminyltransferase. Cell 105:957–969.
66. Young, W. W., D. R. Holcomb, K. G. Ten Hagen, and L. A. Tabak. 2003.
Expression of UDP-GalNAc:polypeptide N-acetylgalactosaminyltransferase
isoforms in murine tissues determined by real-time PCR: a new view of a
large family. Glycobiology 13:549–557.
67. Zarbock, A., C. A. Lowell, and K. Ley. 2007. Spleen tyrosine kinase Syk is
necessary for E-selectin-induced ␣L␤2 integrin-mediated rolling on intercel-
lular adhesion molecule-1. Immunity 26:773–783.
68. Zhang, Y., H. Iwasaki, H. Wang, T. Kudo, T. B. Kalka, T. Hennet, T. Kubota,
L. Cheng, N. Inaba, M. Gotoh, et al. 2003. Cloning and characterization of
a new human UDP-N-acetyl-alpha-D-galactosamine:polypeptide N-acetyl-
galactosaminyltransferase, designated pp-GalNAc-T13, that is specifically
expressed in neurons and synthesizes GalNAc alpha-serine/threonine anti-
gen. J. Biol. Chem. 278:573–584.
8796 TENNO ET AL. MOL.CELL.BIOL.