Fucosylation of Cripto Is Required for Its Ability to Facilitate Nodal Signaling

Abstract

O-linked fucose modification is rare and has been shown to occur almost exclusively within epidermal growth factor (EGF)-like modules. We have found that the EGF-CFC family member human Cripto-1 (CR) is modified with fucose and through a combination of peptide mapping, mass spectrometry, and sequence analysis localized the site of attachment to Thr-88. The identification of a fucose modification on human CR within its EGF-like domain and the presence of a consensus fucosylation site within all EGF-CFC family members suggest that this is a biologically important modification in CR, which functionally distinguishes it from the EGF ligands that bind the type 1 erbB growth factor receptors. A single CR point mutation, Thr-88 --> Ala, results in a form of the protein that is not fucosylated and has substantially weaker activity in cell-based CR/Nodal signaling assays, indicating that fucosylation is functionally important for CR to facilitate Nodal signaling.

Full text

Fucosylation of Cripto is required for its ability to facilitate Nodal signaling
Susan G. Schiffer‡**, Susan Foley‡**, Azita Kaffashan‡, Xiaoping Hronowski‡, Anne
Zichittella‡, Chang-Yeol Yeo§, Konrad Miatkowski‡, Heather B. Adkins¶‡, Bruno Damon‡,
Malcolm Whitman§, David Salomon¶, Michele Sanicola‡ and Kevin P. Williams‡§§
From ‡Biogen, Inc., 14 Cambridge Center, Cambridge, Massachusetts 02142, the §Ph.D.
Program in Biological and Biomedical Sciences and the Department of Cell Biology, Harvard
Medical School, Boston, Massachusetts 02115, and the ¶Tumor Growth Factor Section,
Laboratory of Tumor Immunology and Biology, National Cancer Institute, National Institutes of
Health, Bethesda, Maryland 20892
** Contributed equally to this work.
§§ To whom correspondence should be addressed: Biogen, Inc., 14 Cambridge Center,
Cambridge, MA 02142. Tel.: 617-679-3341; Fax: 617-679-3148; E-mail:
kevin_williams@biogen.com.
Running Title: Fucosylation of Cripto is required for function
Copyright 2001 by The American Society for Biochemistry and Molecular Biology, Inc.
JBC Papers in Press. Published on August 10, 2001 as Manuscript M104774200
by guest on November 16, 2017http://www.jbc.org/Downloaded from

Fucosylation of Cripto is required for function
2
SUMMARY
O-linked fucose modification is rare and has been shown to occur almost exclusively within
epidermal growth-factor-like (EGF) modules. We have found that the EGF-CFC family member
human Cripto-1 (CR) is modified with fucose and through a combination of peptide mapping,
mass spectrometry and sequence analysis localized the site of attachment to Thr-88. The
identification of a fucose modification on human CR within its EGF-like domain and the
presence of a consensus fucosylation site within all EGF-CFC family members suggests that this
is a biologically important modification in CR, which functionally distinguishes it from the EGF
ligands that bind the type 1 erbB growth factor receptors. A single CR point mutation,
Threonine-88 to Alanine results in a form of the protein that is not fucosylated and has
substantially weaker activity in cell-based CR/Nodal signaling assays, indicating that
fucosylation is functionally important for CR to facilitate Nodal signaling.
by guest on November 16, 2017http://www.jbc.org/Downloaded from

Fucosylation of Cripto is required for function
3
INTRODUCTION
Human Cripto-1 (CR)1 is the original member of the EGF-CFC gene family, which
includes a group of structurally related proteins that play essential roles in early embryogenesis
during normal development and have been implicated as oncogenes in cell transformation
(reviewed in Refs. 1, 2). The EGF-CFC family members contain two conserved domains; a
variant of the epidermal growth factor (EGF) domain (often called “EGF-like”), and a unique
cysteine-rich domain, CFC , named for the founding members of the family: CR in humans (3),
FRL-1 in Xenopus (4), and Cryptic in mice (5). The EGF-like domain in EGF-CFC proteins
differs from the canonical 3 loop EGF structure in that loop 1 is deleted, loop 2 is truncated and
loop 3 is well conserved (6). Studies on the zebrafish CR ortholog, one-eyed pinhead (oep),
indicated that the C-terminal region of oep may contain a putative GPI-anchorage site (7, 8) that
serves to tether the protein to the membrane, and that removal of the C-terminal stretch generated
a soluble form of oep that was able to partially rescue oep mutant embryos (8). The recent
characterization of murine cripto (9) confirmed the presence of a GPI-modification within the C-
terminal region of this EGF-CFC protein (9), and again that removing this C-terminal stretch of
residues generates Cripto forms that are soluble.
During embryogenesis, EGF-CFC family members are essential for the formation of
mesoderm during gastrulation and cardiomyocyte formation. Genetic studies in zebrafish (8)
define oep as necessary for gastrulation and left-right patterning during development. These
studies have defined an obligatory role for EGF-CFC proteins as “co-factors” for the correct
signaling of the TGFβ family member, Nodal (reviewed in Refs. 10, 11). Murine Nodal also
by guest on November 16, 2017http://www.jbc.org/Downloaded from

Fucosylation of Cripto is required for function
4
plays a major role in gastrulation and regulating left-right asymmetry during mesoderm formation
in early embryogenesis. Phenotypic similarities are exhibited by nodal and cripto null mice
suggesting the Cripto-Nodal signaling pathway defined in zebrafish is likely to be conserved for
at least a subset of Nodal activities (12-14). Cripto–dependent Nodal signaling has been shown to
be mediated by phosphorylated Smad2, Smad4 and the transcription factor FAST2 (15) which in
turn binds to a left-right specific enhancer (ASE) on the nodal gene and another TGF-beta gene
family member, lefty2. Thus Nodal is autoregulated. Although a specific receptor for the EGF-
CFC and Nodal proteins has not yet been identified, activin type I and II receptors have been
genetically implicated in zebrafish (8) and mouse studies (11). In Xenopus, Cripto-dependent
Nodal signaling has been shown to involve ALK4 (Activin receptor-like kinase 4, or ActR-IB)
(16), however a receptor type II partner has yet to be identified.
The O-linked fucose modification is rare and until recently had been shown to occur
exclusively within epidermal growth factor-like (EGF) modules of secreted proteins involved in
blood clotting and clot dissolution such as factor XII (17), factor IX (18), factor VII (19) or
urokinase-type plasminogen activator (uPA) (20). From these studies a consensus site for O-
fucosylation was defined (21) as the sequence C2XXGGS/TC3, which falls within the second and
third cysteines of an EGF-like module. Fucosylation has been implicated in modulating the
function of a number of proteins. For example, O-fucosylation of uPA within its EGF-like
module is critical for signaling through its receptor (20), O-fucosylation of the E-selectin ligand
ESL-1 is required for binding to E-selectin (22), and recent studies on the Notch- Delta/Serrate
by guest on November 16, 2017http://www.jbc.org/Downloaded from

Fucosylation of Cripto is required for function
5
signaling system have implicated the fucosylation state of Notch as important for modulating its
ability to interact more favorably with the Delta ligand versus the Serrate ligand (23-26).
Here we identify by mass spectroscopy and peptide mapping that recombinant human CR
produced as a soluble form in CHO cells is fucosylated. The modification was mapped to a seven
amino acid sequence within the EGF-like domain that fits the fucosylation consensus site (21).
This consensus site is present in all EGF-CFC proteins, but not in the EGF-family member
ligands that bind type I erbB growth factor receptor family members. Mutating Threonine 88 of
human CR to alanine blocks the addition of the fucose modification on the CR protein and this
alteration abolishes CR’s ability to function as a co-factor for Nodal. The significance of the
fucosylation in modulating CR function with respect to CR’s role in development and cancer is
discussed.
by guest on November 16, 2017http://www.jbc.org/Downloaded from

Fucosylation of Cripto is required for function
6
EXPERIMENTAL PROCEDURES
Protein Expression and Purification – Recombinant human Cripto-1 (CR) was expressed
in CHO cells as a C-terminally truncated form fused to human IgG1 hinge and Fc domain. An
expression plasmid (pSGS480) was constructed by sub-cloning a cDNA encoding human CR
residues 1-169, fused in frame after residue 169 to the Fc hinge and CH2CH3 portion of human
IgG1 (CR(¨&-Fc) into the vector pEAG1100. pEAG1100 is a derivative of plasmid pCMV-
Sport-β-gal (GIBCO-BRL Life Technologies), and was made by removing the reporter gene β-
galactosidase NotI fragment from the plasmid. CHO cells in serum free media (CD-CHO media,
Life Technologies) were transiently transfected with plasmid pSGS480 at room temperature for
15 min using DMRIE-C (Life Technologies) cationic lipid plus cholesterol solution. The
transfected cells were grown as a suspension culture in a spinner flask for 8 days at 28 oC. The
conditioned media was clarified by centrifugation, filtered through a 0.2 µm filter, and stored at
–70 oC.
CR(¨&-Fc protein expression was assessed by Western blot analysis. For Western blot
analysis, conditioned media and cells from CR transfected cells were subjected to SDS-PAGE on
4-20% gradient gels under reducing conditions, transferred electrophoretically to nitrocellulose,
and the CR protein detected with a rabbit polyclonal antibody (Ab 1579) raised against a CR 17-
mer peptide (comprising residues 97-113 of human CR)-KLH conjugate (27), or with the anti-
CR mouse monoclonal antibody A10B2.183.
CR(¨&-Fc was purified from the conditioned medium on a Protein A-Sepharose column
(Amersham Pharmacia). Bound protein was eluted with 25 mM sodium phosphate pH 2.8, 100
by guest on November 16, 2017http://www.jbc.org/Downloaded from

Fucosylation of Cripto is required for function
7
mM NaCl. The eluate was neutralized with 0.5 M sodium phosphate pH 8.6, and analyzed for
total protein content from absorbance at 280 nm (Extinction coefficient = 59181 mol-1 cm-1), and
for purity by SDS-PAGE. The eluted protein was filtered through a 0.2 micron filter, and stored
at –70 oC. N-terminal sequencing was carried out on a Perkin-Elmer Applied Biosystems (PE-
ABD) Procise HT sequencer, run in the pulsed liquid mode equipped with an on-line
phenylthiohydantoin analyzer.
The EGF-like domain of human CR comprising residues 75-112, was also expressed as a
Fc fusion protein. An expression plasmid (pSGS422) was constructed by subcloning a cDNA
encoding human VCAM-1 signal peptide (28) fused to human CR residues 75-112 fused in
frame after residue 112 of the hinge and Fc domain of human IgG1 into vector pEAG1100
(CR(EGF)-Fc). Plasmid pSGS422 was transiently transfected into CHO cells and the CR(EGF)-
Fc protein purified from the conditioned media by chromatography on Protein A as described
above for CR(¨&-Fc.
A C-terminally truncated form of CR (CR(¨&WKDWZDVQRWIXVHGWR)FZDVJHQHUDWHGE\
transiently transfecting into CHO a cDNA encoding human CR amino acid residues 1 to 169 as
described above. Cells were grown as above and CR(¨&ZDVSXULILHGIURm the conditioned
media by immunoaffinity chromatography on the anti-CR mAb column A40G12.8 that was
prepared by conjugating 4 mg of the anti-CR mAb A40G12.8 (H. Adkins, S. Schiffer, P.
Rayhorn, D. Salomon, K. P. Williams and M. Sanicola, manuscript in preparation) per ml of
CNBr-activated Sepahrose 4B resin. Bound protein was eluted with 25 mM sodium phosphate
pH 2.8, 100 mM NaCl, and the eluate was neutralized with 0.5 M sodium phosphate pH 8.6.
by guest on November 16, 2017http://www.jbc.org/Downloaded from

Fucosylation of Cripto is required for function
8
Analysis of CR by Mass Spectrometry – Aliquots of CR protein (CR(¨&-Fc or CR(EGF)-
Fc) in 40 mM sodium phosphate pH 7.5, 5% acetonitrile were treated with 5 mU of PNGase F
(Glyko, Inc.) for 18 h at 37 oC. The PNGase F treated sample in 0.5 M guanidine HCl, 25 mM
Tris-Cl pH 8 was reduced with 4 mM DTT for 30 min at 37 oC. The sample was desalted on-line
prior to electrospray mass spectrometry (ESI-MS) analysis using a LC-Packings Ultimate HPLC
interfaced to a Micromass Quattro II triple quadrupole mass spectrometer equipped with an
electrospray ion source (Micromass, Manchester, UK). The protein was desalted on a Vydac C4
guard column at a flow rate of 50 µl/min with a 15-min 5-65% acetonitile gradient in 0.05%
trifluoracetic acid. The mass spectra were acquired by scanning the m/z range 400-2000 in 5
sec/scan. The raw data were deconvoluted using the Micromass MaxEnt program to generate
zero-charge mass spectra. All masses are average unless otherwise noted.
Peptide Mapping using Endoproteinase Lys-C – CR protein (CR(¨&-Fc or CR(EGF)-
Fc) in PBS, 5 mM EDTA was reduced with 5 mM DTT for 6 h at room temperature, and then
treated with 150 milliunits of PNGase F (Glyko, Inc.) per mg of protein for 16 h at 37 oC. Sample
was adjusted to 6 M guanidine hydrochloride, reduced with 10 mM DTT for 35 min at 45 oC, and
then alkylated with 30 mM iodoacetamide for 30 min at 20 oC. The alkylated protein was
precipitated by addition of 40 volumes of ice-cold ethanol. The solution was stored at –20 oC for
1 h and then centrifuged at 14000 g for 12 min at 4 oC. The supernatant was discarded, and the
precipitate was washed twice with ice-cold ethanol. The alkylated protein was resuspended in 1
M urea, 200 mM Tris-HCl, pH 8.5, and digested with endoproteinase (Endo) Lys-C from
by guest on November 16, 2017http://www.jbc.org/Downloaded from

Fucosylation of Cripto is required for function
9
Achromobacter lyticus (WAKO Pure Chemical Industrials, Ltd.) at a 1:10 enzyme:substrate ratio
for 16 h at room temperature. Analytical-scale digests were desalted on-line prior to ESI-MS
analysis using a YMC C18 column (1 x 25 cm) on the Ultimate HPLC. The running conditions
were a 120-min 0-45% acetonitrile gradient in 0.05% TFA at a flow rate of 50 µl/min. The mass
spectra were acquired by scanning the m/z range 300-1900 in 2.05 sec/scan. Preparative digests
were analyzed by reversed-phase (RP)-HPLC using a Waters Alliance System (Waters Corp.,
Milford, MA) equipped with a YMC C18 column (1 mm x 25 cm). Individual peaks were
collected for further analysis. The molecular masses of peptides were determined by matrix-
assisted laser desorption ionization time-of-flight mass spectrometry (MALDI-TOF MS) on a
Voyager-DE STR mass spectrometer (PerSeptive Biosystems). Peptides were sequenced by
Edman degradation on a Perkin-Elmer Applied Biosystems Procise HT Protein Sequencer
equipped with an on-line phenylthiohydantoin analyzer.
Peptide Mapping using Cyanogen Bromide – Non-reduced CR(¨&-Fc protein in PBS, 5
mM EDTA was treated with PNGase F and ethanol precipitated as described above. The pellet
was resuspended in 200 µL of 70% formic acid. Cyanogen bromide (10 M in acetronitrile) was
added to a final concentration of 1 M and the sample incubated in the dark for 24 h at room
temperature. The digest was analyzed by MALDI-TOF MS. Peptides were separated by RP-
HPLC on a Vydac C4 column using the following gradient (Solvent A, 0.1% TFA, Solvent B
0.085% TFA, 75% acetonitrile: 0-20 % B from 0-10 min; 20-75% B from 10-120 minutes; 75-
100% B from 120-130 minutes), and collected fractions analyzed by MALDI-TOF MS and
Edman sequencing as described above. CNBr generated CR peptides purified from this digest by
by guest on November 16, 2017http://www.jbc.org/Downloaded from

Fucosylation of Cripto is required for function
1
0
RP-HPLC, were reduced with 5mM DTT and cleaved with Carboxypeptidase Y (Boehringer
Mannheim). Portions of the digest were analyzed at 10 min intervals by MALDI-TOF MS.
Construction of CR Mutants – Mutagenesis of CR, threonine 88 to alanine (T88A) was
accomplished by spliced overlap extension polymerase chain reaction (SOE PCR) (29,30). The
following mutagenic primers (5’ to 3’) are the top and bottom strands creating the T88A
mutation. Asterisks indicate the mutant (Thr to Ala) codon.
5’ GCCTGAATGGGGGAG*C*C*TGCATGCTGGGATCCTTTTGTGCCTGC 3’
5’ GCAGGCACAAAAGGATCCCAGCATGCAG*G*C*TCCCCCATTCAGGC 3’
In addition to changing the threonine codon ACC to the codon for alanine, GCC, the same
oligonucleotides also insert a silent change to the sequence to introduce a BamH1 site, changing
the glycine 92 codon from GGG to GGA. This new site gave the ability to screen for the mutant
clone. The T88A mutant was constructed in both the full length CR (1-188) (CR T88A) in the
pCS2+ vector (31), and in the C-terminally truncated CR (1-169) (CR(¨&7$7KHODWWHUZDV
purified as for wild type CR(¨&E\LPPXQR-affinity chromatography.
Luciferase reporter Assay for Cripto - Mouse teratocarcinoma F9 cripto-/- cells (9) were a gift
from Dr. Eileen Adamson (The Burnham Institute, La Jolla, CA). The nodal enhancer ASE (n2)7
(15) cDNA was a gift of Dr. Masaharu Seno (Okayama University, Okayama, Japan) and was
used to constuct a luciferase reporter plasmid p(n2)7-lux. F9 cripto-/- cells (6.5 x 105 cells/well)
were transfected using lipofectamine (Bethesda Reasearch Labs.) with equal amounts of FAST2
(15) and p(n2)7-lux cDNA, and in the absence and presence of CR full length (residues 1-188)
by guest on November 16, 2017http://www.jbc.org/Downloaded from

Fucosylation of Cripto is required for function
1
1
wild-type DNA and in the absence and presence of CR T88A mutant full length DNA. Total
DNA transfected per well was always 1.0 µg and control vector DNA (pEAG1100) was used to
bring the total up to 1.0 µg. 48 h following transfection, cells were lysed with LucLite (Packard
Instrument Company), and luciferase activity measured in a luminometer (PE Applied
Biosystems).
Embryo Injection and Phospho-Smad Analysis - Cripto-dependent Nodal signaling was
measured as described previously (16, 32). Xenopus embryos between 2 to 4 cell stages were used
for injection. Synthetic mRNAs were injected into each blastomere in the animal hemisphere.
Constructs were generated with the pCS2+ vector (31). RNAs were transcribed from the following
constructs using the SP6 mMessage mMachine Kit (Ambion): pCS-Nodal (16), pSGS151-CR WT,
pSGS904-CR T88A, pSGS150-CR(¨&
Ectodermal explants were isolated between stages 8 and 9, and harvested when sibling
un-injected embryos reached stage 10. The explants were lysed in a buffer containing 50 mM
Tris-Cl pH 8.0, 150 mM NaCl, 1% Nonidet P-40, 0.5% sodium deoxycholate, 2 mM EDTA, 2x
Complete, EDTA-free protease inhibitor cocktail (Roche Molecular Biochemicals), 4 µg/ml
pepstatin A, 1 mM PMSF, 20 nM calyculin A, 25 mM α-glycerophosphate, 100 mM sodium
fluoride, 2 mM sodium orthovanadate and 10 mM sodium pyrophosphate. After centrifugation,
supernatants were mixed with equal volume of 4x Laemmli loading buffer and subjected to SDS-
PAGE. Western blot analysis was performed as described (32). The following antibodies were
used for Western blot analysis: anti-Smad1/5/8 goat polyclonal antibody (Santa Cruz
Biotechnology), anti-Smad2 mouse monoclonal antibody (clone 18, Transduction Laboratories),
by guest on November 16, 2017http://www.jbc.org/Downloaded from

Fucosylation of Cripto is required for function
1
2
anti-actin mouse monoclonal antibody (clone AC-40, Sigma), and anti-phospho-Smad2 rabbit
polyclonal antibodies (32). For detecting Cripto protein expression, the anti-Cripto Ab 1579
described above was used.
by guest on November 16, 2017http://www.jbc.org/Downloaded from

Fucosylation of Cripto is required for function
1
3
RESULTS
Expression of human CR as a soluble form and demonstration that it is glycosylated – When
we expressed human CR in CHO cells as the full length protein (comprising residues 1-188, Fig.
1), it was produced as an insoluble membrane-associated form that was difficult to purify (data
not shown). Others have reported similar difficulties in mammalian (33), and bacterial (34)
systems. We were able to express human CR as a soluble form by making a C-terminal truncated
form analogous to that generated for the Zebrafish oep (8) and murine Cripto (9). Consequently,
we transiently expressed in CHO cells a C-terminally truncated form of human CR, comprising
residues 1-169 (Fig. 1), as an Fc fusion protein (CR(¨&-Fc) that was efficiently secreted into the
supernatant. CR(¨&-Fc was purified from the conditioned media by chromatography on Protein
A-Sepharose. Edman N-terminal sequencing of CR(¨&-Fc identified a single N-terminus that
starts with Leu-31 (Fig. 1). This is consistent with predictions for processing of the signal peptide
using the SIGNALP program (35). Electrospray mass spectrometry (ESI-MS) data for the
purified CR(¨&-Fc gave a complex broad spectra that was not resolvable, suggesting a complex
carbohydrate pattern. On SDS-PAGE, the purified CR(¨&-Fc migrated as two diffuse bands
(Fig. 2A) with apparent masses of 50 kDa and 52 kDa. After treatment of the reduced protein
with PNGase F, these two bands both shift to lower molecular weights with masses of ~ 45 kDa
and 47 kDa (Fig. 2A), suggesting that both these bands are N-glycosylated. The ESI-MS spectra
of the PNGase F-treated CR(¨&-Fc (Fig. 2B) was less complex than the fully glycosylated
protein but still showed a number of ions with masses ranging from 41116 to 42870 Da (Table I).
Theoretical assignments for the nine most intense ions based on their observed masses are shown
in Table I. Some of these species had measured masses that matched the predicted mass for
by guest on November 16, 2017http://www.jbc.org/Downloaded from

Fucosylation of Cripto is required for function
1
4
CR(¨&-Fc starting at residue 31 and being modified with O-linked glycans (HexNAc-Hex-
NeuAc). Peak A with an observed mass of 41116 Da agrees exactly with the theoretical mass for
residues 31-396. Since antibodies frequently end with a lysine residue, the C-terminal lysine is
often lost due to carboxypeptidase B activity in serum. For CR(¨&-Fc this is reflected in the
protein forms ending at residues 396 (peaks A, B, D, F, G and H; Figure 2B, Table I). However,
some of the glycosylated forms appear to end in the terminal lysine. In contrast, others had
measured masses that did not agree with any available prediction, although the difference in mass
of 146 Da for each species was consistent with the addition of a deoxyhexosyl group, potentially
fucose with average mass = 146.14 Da.
Localization of the 146 Da modification within the human CR sequence – The site of
modification for the 146 Da group within the human sequence was identified by a combination of
peptide mapping and mass spectrometry. PNGase F treated CR(¨&-Fc was reduced, alkylated,
and treated with endoproteinase (Endo) Lys-C. Fig. 3 shows results from a Endo Lys-C peptide
mapping LC-MS analysis of CR(¨&-Fc with a total ion current (TIC) readout. For a complete
digestion, 25 potential Endo Lys-C cleavage products would be predicted and were designated
CR-E# where E1 is the peptide from the N-terminus of CR(¨&-Fc and E25 the peptide from the
C-terminus of CR(¨&-Fc (Table II). Mass data accounting for over 96 % of the protein sequence
could be accounted for from this peptide mapping study (Table II). Within the CR(¨&-Fc
sequence, peptides 31-76 and 133-172 were identified by ESI-MS as containing O-linked
glycosylation. Further peptide mapping/MS analysis localized residues Ser-40 and Ser-161 as the
sites modified with O-linked carbohydrate (Table II). The peak noted with an asterisk (Fig. 3,
by guest on November 16, 2017http://www.jbc.org/Downloaded from

Fucosylation of Cripto is required for function
1
5
Table II), with an observed mass of 4473.1 Da corresponds to a single Endo Lys-C peptide
comprising residues 77-112 being modified with a 146 Da group (calculated average mass for
unmodified 77-112 = 4326.9 Da; Table II). No unmodified peptide 77-112 was observed in the
map indicating that for CR(¨&-Fc expressed in CHO the protein is predominantly occupied with
the deoxyhexosyl group.
The modified Endo Lys-C E2 peptide was isolated in pure form from a preparative digest
using RP-HPLC and analyzed by MALDI-TOF MS. For this peptide, the observed (M + H)+ ion
at m/z at 4470.85 corresponds to the peptide comprising residues 77-112 plus a deoxyhexosyl
group (m/z calculated 4470.89). In the preparative Endo Lys-C digest of CR(¨&-Fc, two
peptides resulting from a non-specific fragmentation of peptide CR-E2 were observed. For the
first peptide (CR-E2a), the observed (M + H)+ ion at m/z 793.41 was in agreement with the
calculated m/z 793.3 for residues 77-82. Identification of peptide CR-E2a was confirmed by
Edman N-terminal sequencing. For the second peptide (CR-E2b), the observed (M + H)+ ion at
m/z at 3699.21 corresponds to residues 83-112 plus a deoxyhexosyl group (m/z calculated
3699.21). This non-specific cleavage event occurred at Cys82/Leu83, a site not predicted to be
susceptible to Endo Lys-C, but may have occurred due to the alkylation of the cysteine with
iodoacetamide.
The EGF-like domain of CR is fucosylated - The peptide sequence containing the 146 Da
modification comprising residues 77-112 spans almost the entire predicted EGF-like domain of
human CR (Fig. 1). We also generated and expressed the EGF-like domain of human CR
comprising residues 75 to 112 (Fig. 1) as a Fc fusion protein (CR(EGF)-Fc). The ESI-MS spectra
by guest on November 16, 2017http://www.jbc.org/Downloaded from

Fucosylation of Cripto is required for function
1
6
of the reduced CR(EGF)-Fc showed a complex spectra consisting of a number of peaks with
masses ranging from 33803 to 31142 Da (data not shown). After PNGase F treatment, the ESI-
MS spectra of CR(EGF)-Fc showed three peaks with molecular masses of 29700 (A), 28846 (B)
and 29974 (C) Da (Fig. 4). Peak A had the predicted mass for CR(EGF)-Fc residues 1 to 265
(29700 Da). Peaks B and C had measured masses that matched the theoretical masses (29846 and
29974 Da respectively) for CR(EGF)-Fc plus 146 Da (C-1 and intact respectively).
Characterization of this expressed fragment confirmed that the deoxyhexosyl modification was
within the EGF-like domain of human CR, while the O-linked glycosylations were outside of the
EGF domain of CR.
To further localize the site of the 146 Da modification, PNGase F-treated CR(¨&-Fc was
cleaved with cyanogen bromide (CNBr) and the digest analyzed by MALDI-TOF MS. One (M +
H)+ ion at m/z at 2252.37 corresponds to a peptide comprising residues 71-90 plus 146 Da. This
peptide was treated with carboxypeptidase Y (Boehringer Mannheim), and this yielded a peptide
with an observed (M +H)+ ion at m/z at 2180.62 corresponding to residues to 71- 89 plus 146 Da
(calculated m/z 2181.48). The result showed that the C-terminal homoserine lactone (Met
converted to homoserine lactone after CNBr cleavage) of peptide 71-90 could be removed
without loss of the extra 146 Da.
By taking the results of the CR(¨&-Fc digests with Endo Lys-C (146 Da modification
localized to residues 83-112), CNBr (146 Da modification localized to residues 71-90), and
carboxypeptidase Y (146 Da modification localized to residues 71-89), and looking for overlaps
of sequence between the fragments we can infer that the 146 Da modification is within the
human CR protein sequence comprising residues 83-89 (amino acids: CLNGGTC). Of these
by guest on November 16, 2017http://www.jbc.org/Downloaded from

Fucosylation of Cripto is required for function
1
7
residues we could infer that the deoxyhexosyl group was attached to Thr-88 for the following
reasons: i) since the cysteines at positions 83 and 89 could be alkylated, they could not carry the
146 Da substituent; ii) since the side chains of leucine and glycine are unreactive they could not
be modified, and iii) since typically Asn followed by Gly can form a cyclic imide (36). Thr is a
typical O-linked glycosylation site, although fucosylation is rare. A motif for a fucosylation
consensus sequence, C2XXGGS/TC3, has been described (21), and the fucosylated CR sequence
fits this motif.
The T88A form of human CR is not fucosylated – To test the role of fucosylation on function,
threonine 88 of human CR was mutated to alanine (T88A), and soluble forms of CR(¨&ZLOG
type (WT) and CR(¨&7$ZHUHH[SUHVVHGWUDQVLHQWO\LQ&+2FHOOV%RWKSURWHLQVZHUH
effectively secreted into the conditioned media and could be purified by immuno-affinity
chromatography using an anti-CR mAb immuno-affinity column. The proteins eluted from the
immuno-affinity column were analyzed by SDS-PAGE/Western under reducing and non-
reducing conditions. The CR(¨&SURWHLQPLJUDWHGDVDGLIIXVHEDQGEHWZHHQDQGN'D
(Fig. 5A). The CR(¨&7$PXWDQWSURWHLQKDGDYHU\VLPLOar pattern to the wild type protein
on SDS-PAGE (Fig. 5A). Under non-reducing conditions both proteins were predominantly
monomeric and there was no evidence for aggregation for either protein (data not shown). The
anti-CR mAb used for the immuno-affinity column recognizes a tertiary conformational epitope
within the EGF-like domain of human CR2 and could be used as a probe to assess folding. As an
immuno-affinity resin this mAb recognized both the wild type and T88A proteins equally well as
judged by % binding, suggesting that the CR protein carrying the T88A mutation is properly
by guest on November 16, 2017http://www.jbc.org/Downloaded from

Fucosylation of Cripto is required for function
1
8
folded. The CR(¨&DQG&5¨&7$SURWHLQVZHUHWUHDWHGZLWK31*DVH)XQGHUUHGXFLQJ
conditions and analyzed by ESI-MS (Fig. 5B, Table III). For wild type CR(¨&WKHWKHRUHWLFDO
assignments for the eight most intense ions based on their observed masses are shown in Table
III A. Some of these species had measured masses that matched the predicted mass for CR(¨&
being modified with O-linked glycans (HexNAc-Hex-NeuAc). Other peaks had masses of +146
Da that fit with the protein being modified with a fucose. Peak A with a measure mass of 15616
Da agrees exactly with the theoretical mass for residues 31-169. For CR(¨&7$SHDNVWKDW
correspond to those observed in the wild type spectra are labeled with the same letter but with a
prime mark (‘) (Fig. 5C, Table III B). In the ESI-MS spectra, peaks that had the 146 Da
modification in the wild type spectra are absent in the T88A spectra (peaks B, D, F and H are
absent) confirming a lack of fucosylation.
Fucosylation of Cripto is required for signaling in F9 cells – The TGFβ family member,
Nodal, has been shown to stimulate Smad2 phosphorylation and a FAST transcription factor
dependent reporter activity only in the presence of Cripto (15, 16). CR activity was assessed in a
mouse F9-derived embryonal carcinoma cell line gene-targeted for inactivation of the cripto
locus, i.e. cripto “null” cells (cripto -/- F9) (9). These cells contain endogenous Nodal, but are
null for cripto and thus null for Cripto-dependent signaling.
As a result of Smad2 phosphorylation induced by TGF-β family member stimulation, the
transcription factor FAST can bind to Smad2 through Smad4 and can activate transcription (37)
by measuring the luciferase activity from a FAST regulatory element-luciferase reporter gene
(15). To demonstrate that cripto -/- F9 cells could be used to monitor CR-dependent Nodal
by guest on November 16, 2017http://www.jbc.org/Downloaded from

Fucosylation of Cripto is required for function
1
9
signaling, the cripto -/- F9 cells were transiently transfected with the FAST responsive element
(n2)7-luc, with or without the cDNA constructs for wild type CR, nodal, and FAST2. We first
demonstrated that activity, as measured by the stimulation of the FAST-luciferase reporter, was
dependent upon CR and FAST2 in the cripto -/- F9 cells. Transfection with both CR wild type
(CR WT) + FAST2 cDNA (Fig. 6, column 4) increased (n2)7-lux reporter activity by 6-fold
compared to CR WT (Fig. 6, column 3) or FAST2 (Fig. 6, column 2) alone. Transfection with CR
WT, FAST2 and nodal cDNA gave no appreciable increase in activity (data not shown) compared
to the result for CR WT and FAST2 (Fig. 6, column 4). F9 cells possess endogenous mouse Nodal
(38) presumably in an amount sufficient to facilitate signaling in our assay, consequently in
subsequent assays we did not transfect Nodal into the cells. In contrast, the activity of CR T88A
(Fig. 6, column 6) was significantly reduced versus CR WT (Fig. 6, column 4). Transfection of
the cripto -/- F9 cells with a cDNA encoding the CR T88A mutant only increased FAST-
dependent (n2)7-lux reporter activity by 1.6-fold compared to control (Fig. 6, column 2) whereas
CR WT gave a 6-fold increase. These assay results indicate that fucosylation of CR at Thr-88 is a
requirement for CR-dependent-Nodal signaling through FAST2 and an activin response element.
CR fucosylation is required for Nodal mediated activation of smad2 in Xenopus embryo - We
also tested whether fucosylation of CR was required for the ability of Nodal to induce Smad2
activation in Xenopus embryos as assessed by Smad2 phosphorylation. Synthetic mRNAs
encoding CR wild type or CR T88A along with Nodal were injected into Xenopus embryos, and
the levels of activated, phosphorylated Smad2 from prospective ectodermal explants (animal
caps) were examined by Western blot analysis using an antibody specific to carboxy-terminal
by guest on November 16, 2017http://www.jbc.org/Downloaded from

Fucosylation of Cripto is required for function
2
0
phosphorylated Smad2 (phospho-Smad2) (16, 32). The phospho-Smad2 signal was greatly
enhanced over endogenous levels in explants injected with Nodal and with either full length CR
(CR WT) or CR(¨& (Fig. 7). Smad2 phosphorylation was increased 2-fold versus Nodal alone
(as determined by densitometry of the bands). Hence, in this assay system both cell-tethered CR
and the soluble C-terminally truncated form are able to facilitate Nodal signaling as assessed by
Smad2 phosphorylation. In the absence of Nodal no signal was seen. A previously reported, weak
signal for Smad2 activation by Nodal is seen in the absence of ectopic CR and is presumably due
to the presence of a Xenopus EGF-CFC protein, FRL-1 (4, 16). Expression of the mutant CR
T88A in Xenopus did not result in any increase of Smad2 activation above basal levels with
Nodal alone (Fig. 7), confirming again that the T88A mutation inactivates Cripto’s ability to
facilitate signaling. If anything phosphorylation was slightly decreased compared to basal (as
determined by densitometry) suggesting that T88A might have the capacity to act as a dominant
negative to endogenous wild type CR. Western blot analysis for CR protein levels (Fig. 7)
demonstrate that both the WT and T88A CR proteins are expressed at comparable levels in this
Xenopus system. Hence, the T88A mutation inactivates the ability of human CR to mediate
Nodal signaling in two model systems (F9 cell based assay and Xenopus explants) through a
Smad2 and FAST2 pathway.
by guest on November 16, 2017http://www.jbc.org/Downloaded from

Fucosylation of Cripto is required for function
2
1
DISCUSSION
We have demonstrated that human Cripto-1 (CR) expressed in a mammalian CHO cell
system is highly glycosylated and possesses an O-linked fucose modification. We determined
that this modification occurs in the EGF-like domain at Thr-88 and that this modification is
required for CR-dependant Nodal signaling through a Smad2 and FAST2 pathway. The site for
the fucose modification found on human CR fits a previously proposed consensus motif for
fucosylation (21) of C2XXGGS/TC3 (Table IV). The fucosylation motif is present in all CR
orthologs so far discovered in the EGF-CFC family, including mouse Cripto (39), chick Cripto
(40), Zebrafish oep (7), Xenopus FRL-1 (4) and mouse cryptic (5) (Table IV), suggesting that
fucosylation is a conserved modification and critical for function. Due to the presence of an EGF-
like domain in CR early reports had suggested a close association between CR and EGF-growth
factor family members. A search for the fucosylation motif in family members of the EGF
growth factor family such as EGF, heregulin-α or transforming growth factor α that bind to the
type I erbB growth factor receptor family members revealed that these proteins do not possess the
consensus site for fucosylation (Table IV). The presence of a fucosylated site in EGF-CFC family
members and its absence in the other EGF ligand family of growth factors, reinforces the
distinctions between EGF-CFC family members and other EGF-like peptide growth factors.
Interestingly, a recent report (41) on Thrombospodin has demonstrated O-fucosylation outside of
an EGF-like domain and within a sequence that does not conform to the reported consensus site.
In two different Cripto-dependent Nodal signaling assays, we demonstrate that the
Thr88Ala mutation in human CR results in a loss of fucosylation and in an impairment of the
biological activity. It will be of interest to ascertain whether a similar loss of biological activity
by guest on November 16, 2017http://www.jbc.org/Downloaded from

Fucosylation of Cripto is required for function
2
2
occurs for this CR fucosylation mutant when this protein is assessed for activity in other assays in
mammary epithelial cells such as MAPK activation, branching morphogenesis and inhibition of
β-casein expression in response to lactogenic hormones where CR has activity (6, 34, 42-45). CR
was able to activate FAST2-dependent signaling in F9 cells, and stimulate Smad2
phosphorylation in Xenopus. In both assays, the activity of the CR T88A mutant was
significantly reduced indicating a critical role for function. Growing evidence suggest that
fucosylation is a biologically significant feature. O-fucosylation of uPA does not affect the
affinity of uPA for binding to its receptor but is critical for signaling (20). Fucosylation of the E-
selectin ligand ESL-1, is essential for binding to E-selectin (22). A more comprehensive
understanding of the role of fucosylation in mediating function has come from recent studies on
the Notch-Delta/Serrate signaling system (reviewed in Ref. 46). O-linked fucose modification on
Notch1 is critical for vertebrate development, and Notch1 was recognized as the first membrane-
associated protein with an O-linked fucose modification (24). Indeed, in one case of the human
disease CADASIL, the mutation in Notch disrupts the O-linked fucosylation site (47) which
influences receptor-ligand interactions. Furthermore, it has been shown that the O-linked fucose
groups on Notch can be elongated by Fringe, a fucose-specific β 1,2 N-acetyl-
glucosaminyltransferase, that modulates the ability of Notch ligands, Delta-1 and Serrate, to
activate Notch (23, 25, 26). Elongation of the O-fucose to a tetrasaccharide increases the ability
of Delta-1 to signal and decreases the ability of Serrate to signal (23, 25, 26). In our studies,
human CR was predominantly modified with O-linked fucose in the monosaccharide form. We
did not observe unmodified CR or CR modified with elongated forms of O-fucose. Di- or tetra-
saccharide elongations of O-fucose have also been reported for factor IX (18). It is apparent that
by guest on November 16, 2017http://www.jbc.org/Downloaded from

Fucosylation of Cripto is required for function
2
3
removing the fucose abrogates function but its is not known whether the elongation of the O-
fucose seen for Notch and factor IX is a general mechanism for enhancing activity. A critical role
for EGF-CFC family members in facilitating signaling by Nodal has been demonstrated in this
study and by others (8, 15, 16, 48-50). We propose that fucosylation of CR regulates Nodal’s
ability to stimulate signaling. It may be that the EGF-like module of CR interacts with Nodal
directly or with a related receptor. Furthermore, it may be that there are elongated fucosylated
forms of CR and that these forms may exhibit enhanced interaction with Nodal or favor the
interaction of CR with a different TGF-β family member. Elongation of the O-fucose groups as a
mechanism for modulating CR or Nodal activity may also exist, and may be a general
mechanism during development to allow competition between different TGFβ family members
for binding to CR.
Mouse Cripto and human CR have also been implicated to play a role in cell
transformation and in the etiology of mouse and human cancers and have been defined as
oncogenes in this context (1, 27, 44, 51-55). It is well known that the altered expression of cell-
surface carbohydrates on glycoproteins is often associated with malignant transformation, which
may be due in part to changes in the levels of glycosyltransferases including fucosyltransferases
and that blocking these enzymes can impair tumorgenesis (56-58). Both the Notch (59) and
Cripto (1) signaling pathways underlie a range of diseases including cancers. Changes in the
fucosylation state of CR may also exist in tumor cells and may be part of a mechanism promoting
cell transformation.
by guest on November 16, 2017http://www.jbc.org/Downloaded from

Fucosylation of Cripto is required for function
2
4
In conclusion, we have demonstrated that the human CR protein is fucosylated and that
this fucosylation is required for Cripto-dependent Nodal signaling, providing important evidence
for understanding the mechanism of action for EGF-CFC proteins.
by guest on November 16, 2017http://www.jbc.org/Downloaded from

Fucosylation of Cripto is required for function
2
5
REFERENCES
1. Salomon, D. S., Bianco, C., and De Santis, M. (1999) Bioessays 21, 61-70
2. Salomon, D. S., Bianco, C., Ebert, A. D., Khan, N. I., De Santis, M., Normanno, N.,
Wechselberger, C., Seno, M., Williams, K., Sanicola, M., Foley, S., Gullick, W. J., and
Persico, G. (2000) Endocr. Relat Cancer 7, 199-226
3. Ciccodicola, A., Dono, R., Obici, S., Simeone, A., Zollo, M., and Persico, M. G. (1989)
Embo. J. 8, 1987-1991
4. Kinoshita, N., Minshull, J., and Kirschner, M. W. (1995) Cell 83, 621-630
5. Shen, M. M., Wang, H., and Leder, P. (1997) Development 124, 429-442
6. Lohmeyer, M., Harrison, P. M., Kannan, S., DeSantis, M., O'Reilly, N. J., Sternberg, M.
J., Salomon, D. S., and Gullick, W. J. (1997) Biochemistry 36, 3837-3845
7. Zhang, J., Talbot, W. S., and Schier, A. F. (1998) Cell 92, 241-251
8. Gritsman, K., Zhang, J., Cheng, S., Heckscher, E., Talbot, W. S., and Schier, A. F. (1999)
Cell 97, 121-132
9. Minchiotti, G., Parisi, S., Liguori, G., Signore, M., Lania, G., Adamson, E. D., Lago, C.
T., and Persico, M. G. (2000) Mech. Dev. 90, 133-142
10. Shen, M. M., and Schier, A. F. (2000) Trends Genet. 16, 303-309
11. Schier, A. F., and Shen, M. M. (2000) Nature 403, 385-389
12. Collignon, J., Varlet, I., and Robertson, E. J. (1996) Nature 381, 155-158
13. Lowe, L. A., Supp, D. M., Sampath, K., Yokoyama, T., Wright, C. V., Potter, S. S.,
Overbeek, P., and Kuehn, M. R. (1996) Nature 381, 158-161
by guest on November 16, 2017http://www.jbc.org/Downloaded from

Fucosylation of Cripto is required for function
2
6
14. Levin, M. (1997) Bioessays 19, 287-296
15. Saijoh, Y., Adachi, H., Sakuma, R., Yeo, C. Y., Yashiro, K., Watanabe, M., Hashiguchi,
H., Mochida, K., Ohishi, S., Kawabata, M., Miyazono, K., Whitman, M., and Hamada, H.
(2000) Mol. Cell 5, 35-47
16. Yeo, C.-Y., and Whitman M. (2001) Mol. Cell 7, 949-957
17. Harris, R. J., Ling, V. T., and Spellman, M. W. (1992) J. Biol. Chem. 267, 5102-5107
18. Harris, R. J., van Halbeek, H., Glushka, J., Basa, L. J., Ling, V. T., Smith, K. J., and
Spellman, M. W. (1993) Biochemistry 32, 6539-6547
19. Kao, Y. H., Lee, G. F., Wang, Y., Starovasnik, M. A., Kelley, R. F., Spellman, M. W.,
and Lerner, L. (1999) Biochemistry 38, 7097-7110
20. Rabbani, S. A., Mazar, A. P., Bernier, S. M., Haq, M., Bolivar, I., Henkin, J., and
Goltzman, D. (1992) J. Biol. Chem. 267, 14151-14156
21. Harris, R. J., and Spellman, M. W. (1993) Glycobiology 3, 219-224
22. Steegmaier, M., Levinovitz, A., Isenmann, S., Borges, E., Lenter, M., Kocher, H. P.,
Kleuser, B., and Vestweber, D. (1995) Nature 373, 615-620
23. Moloney, D. J., Panin, V. M., Johnston, S. H., Chen, J., Shao, L., Wilson, R., Wang, Y.,
Stanley, P., Irvine, K. D., Haltiwanger, R. S., and Vogt, T. F. (2000) Nature 406, 369-375
24. Moloney, D. J., Shair, L. H., Lu, F. M., Xia, J., Locke, R., Matta, K. L., and Haltiwanger,
R. S. (2000) J. Biol. Chem. 275, 9604-9611
25. Bruckner, K., Perez, L., Clausen, H., and Cohen, S. (2000) Nature 406, 411-415
26. Hicks, C., Johnston, S. H., diSibio, G., Collazo, A., Vogt, T. F., and Weinmaster, G.
(2000) Nat. Cell Biol. 2, 515-520
by guest on November 16, 2017http://www.jbc.org/Downloaded from

Fucosylation of Cripto is required for function
2
7
27. Saeki, T., Stromberg, K., Qi, C. F., Gullick, W. J., Tahara, E., Normanno, N., Ciardiello,
F., Kenney, N., Johnson, G. R., and Salomon, D. S. (1992) Cancer Res. 52, 3467-3473
28. Hession, C., Tizard, R., Vassallo, C., Schiffer, S. B., Goff, D., Moy, P., Chi-Rosso, G.,
Luhowskyj, S., Lobb, R., and Osborn, L. (1991) J. Biol. Chem. 266, 6682-6685
29. Horton, R. M., Hunt, H. D., Ho, S. N., Pullen, J. K., and Pease, L. R. (1989) Gene 77, 61-
68
30. Ho, S. N., Hunt, H. D., Horton, R. M., Pullen, J. K., and Pease, L. R. (1989) Gene 77, 51-
59
31. Turner, D. L., and Weintraub, H. (1994) Genes Dev. 8, 1434-1447
32. Faure, S., Lee, M. A., Keller, T., ten Dijke, P., and Whitman, M. (2000) Development
127, 2917-2931
33. Brandt, R., Normanno, N., Gullick, W. J., Lin, J. H., Harkins, R., Schneider, D., Jones, B.
W., Ciardiello, F., Persico, M. G., Armenante, F., and et al. (1994) J. Biol. Chem. 269,
17320-17328
34. Seno, M., DeSantis, M., Kannan, S., Bianco, C., Tada, H., Kim, N., Kosaka, M., Gullick,
W. J., Yamada, H., and Salomon, D. S. (1998) Growth Factors 15, 215-229
35. Nielsen, H., Brunak, S., and von Heijne, G. (1999) Protein Eng. 12, 3-9
36. Allen, G. (1989) in Laboratory Techniques in Biochemistry and Molecular Biology
(Burdon, R. H. and van Knippenberg, P. H., eds.) Vol. 9, pp. 232-233, Elsevier,
Amsterdam
37. Chen, X., Rubock, M. J., and Whitman, M. (1996) Nature 383, 691-696
38. Zhou, X. L., Hu, X. L., and Wu, H. L. (1999) Yi Chuan. Xue. Bao. 26, 142-149
by guest on November 16, 2017http://www.jbc.org/Downloaded from

Fucosylation of Cripto is required for function
2
8
39. Dono, R., Scalera, L., Pacifico, F., Acampora, D., Persico, M. G., and Simeone, A. (1993)
Development 118, 1157-1168
40. Colas, J. F., and Schoenwolf, G. C. (2000) Gene 255, 205-217
41. Hofsteenge, J., Huwiler, K. G., Macek, B., Hess, D., Lawler, J., Mosher, D. F., and Peter-
Katalinic, J. (2001) J. Biol. Chem. 276, 6485-6498
42. De Santis, M. L., Kannan, S., Smith, G. H., Seno, M., Bianco, C., Kim, N., Martinez-
Lacaci, I., Wallace-Jones, B., and Salomon, D. S. (1997) Cell Growth Differ. 8, 1257-
1266
43. Kannan, S., De Santis, M., Lohmeyer, M., Riese, D. J., 2nd, Smith, G. H., Hynes, N.,
Seno, M., Brandt, R., Bianco, C., Persico, G., Kenney, N., Normanno, N., Martinez-
Lacaci, I., Ciardiello, F., Stern, D. F., Gullick, W. J., and Salomon, D. S. (1997) J. Biol.
Chem. 272, 3330-3335
44. Niemeyer, C. C., Persico, M. G., and Adamson, E. D. (1998) Cell Death Differ. 5, 440-
449
45. Wechselberger, C., Ebert, A. D., Bianco, C., Khan, N. I., Sun, Y., Wallace-Jones, B.,
Montesano, R., and Salomon, D. S. (2001) Exp. Cell. Res. 266, 95-105
46. Blair, S. S. (2000) Curr. Biol. 10, 608-612
47. Joutel, A., Corpechot, C., Ducros, A., Vahedi, K., Chabriat, H., Mouton, P., Alamowitch,
S., Domenga, V., Cecillion, M., Marechal, E., Maciazek, J., Vayssiere, C., Cruaud, C.,
Cabanis, E. A., Ruchoux, M. M., Weissenbach, J., Bach, J. F., Bousser, M. G., and
Tournier-Lasserve, E. (1996) Nature 383, 707-710
by guest on November 16, 2017http://www.jbc.org/Downloaded from

Fucosylation of Cripto is required for function
2
9
48. Yan, Y. T., Gritsman, K., Ding, J., Burdine, R. D., Corrales, J. D., Price, S. M., Talbot,
W. S., Schier, A. F., and Shen, M. M. (1999) Genes Dev. 13, 2527-2537
49. Kiecker, C., Muller, F., Wu, W., Glinka, A., Strahle, U., and Niehrs, C. (2000) Mech.
Dev. 94, 37-46
50. Kumar, A., Novoselov, V., Celeste, A. J., Wolfman, N. M., ten Dijke, P., and Kuehn, M.
R. (2001) J. Biol. Chem. 276, 656-661
51. Baldassarre, G., Romano, A., Armenante, F., Rambaldi, M., Paoletti, I., Sandomenico, C.,
Pepe, S., Staibano, S., Salvatore, G., De Rosa, G., Persico, M. G., and Viglietto, G.
(1997) Oncogene 15, 927-936
52. Niemeyer, C. C., Spencer-Dene, B., Wu, J. X., and Adamson, E. D. (1999) Int. J. Cancer
81, 588-591
53. Byrne, R. L., Autzen, P., Birch, P., Robinson, M. C., Gullick, W. J., Neal, D. E., and
Hamdy, F. C. (1998) J. Pathol. 185, 108-111
54. De Angelis, E., Grassi, M., Gullick, W. J., Johnson, G. R., Rossi, G. B., Tempesta, A., De
Angelis, F., De Luca, A., Salomon, D. S., and Normanno, N. (1999) Int. J. Oncol. 14,
437-440
55. Normanno, N., De Luca, A., Salomon, D. S., and Ciardiello, F. (1998) Cancer Detect.
Prev. 22, 62-67
56. Lopez-Ferrer, A., de Bolos, C., Barranco, C., Garrido, M., Isern, J., Carlstedt, I., Reis, C.
A., Torrado, J., and Real, F. X. (2000) Gut 47, 349-356
57. Ito, H., Hiraiwa, N., Sawada-Kasugai, M., Akamatsu, S., Tachikawa, T., Kasai, Y.,
Akiyama, S., Ito, K., Takagi, H., and Kannagi, R. (1997) Int. J. Cancer 71, 556-564
by guest on November 16, 2017http://www.jbc.org/Downloaded from

Fucosylation of Cripto is required for function
3
0
58. Hiller, K. M., Mayben, J. P., Bendt, K. M., Manousos, G. A., Senger, K., Cameron, H. S.,
and Weston, B. W. (2000) Mol. Carcinog. 27, 280-288
59. Fortini, M. E. (2000) Nature 406, 357-358
by guest on November 16, 2017http://www.jbc.org/Downloaded from

Fucosylation of Cripto is required for function
3
1
FOOTNOTES
1The abbreviations used are: CR, Cripto-1; EGF, epidermal growth factor; ESI-MS,
electrospray mass spectrometry, MALDI-TOF MS, matrix-assisted laser desorption ionization
time-of-flight mass spectrometry; RP-HPLC, reversed phase high performance liquid
chromatography; EGF-CFC, epidermal growth factor-Cripto, FRL-1, Cryptic.
2K.P. Williams KP and A. Zichittella A, unpublished data.
3H. Adkins, S. Schiffer, P. Rayhorn, D. Salomon, K. P. Williams and M. Sanicola,
unpublished data.
Acknowledgements: The authors thanks Eileen Adamson (The Burnham Institute, La Jolla, CA)
for providing the F9 Cripto-/- cells, and Dr. Masaharu Seno (Okayama University, Japan) for
providing ASE (n2)7 cDNA . The authors thank Joe Amatucci, Paul Rayhorn, Ray Boynton,
Carmen Young, Dingyi Wen, Olivia Orozco, Mohammad Zafari and Irene Sizing for their
contributions to this work, and Blake Pepinsky for valuable discussions and comments on the
manuscript. This work was supported in part by grants from the NICHD (M.W.).
by guest on November 16, 2017http://www.jbc.org/Downloaded from

Fucosylation of Cripto is required for function
3
2
FIGURE LEGENDS
Fig. 1. Amino acid sequence and schematic representation of human CR(¨& A.
Sequence of C-terminally truncated human CR (comprising residues 1-169). Edman sequencing
of human CR(¨&-Fc purified from CHO cells identified leucine-31 as the first residue of the
mature protein. The signal sequence is in lower case. The EGF-like domain of human CR
comprising residues 75-112 is underlined. Thr 88, the fucosylation site is in bold. B. Schematic
of CR(¨&-Fc. The signal sequence is the shaded box. Peptide mapping and mass spectrometry
identified Ser 40 and Ser 161 as O-linked glycosylation sites, Asn 79 as an N-linked
glycosylation site, and Thr 88 as the fucosylation site. The Fc hinge and CH2CH3 portion of
human IgG1 comprise residues 170 to 397 of CR(¨&-Fc (black box).
Fig. 2. Characterization of the CR(¨&-Fc fusion protein. A C-terminally truncated
form of CR (comprising resides 1-169) was expressed as a Fc fusion protein in CHO cells. The
CR(¨&-Fc was purified by Protein-A chromatography and analyzed by SDS-PAGE and mass
spectrometry. CR(¨&-Fc was treated with PNGase F as described in experimental procedures.
A. Samples were run under reducing conditions on SDS-PAGE as follows; lane a, Benchmark
pre-stained protein ladder (New England Biolabs); lane b, CR(¨&-Fc; lane c, CR(¨&-Fc after
PNGase F treatment. The gel was stained with Coomassie Blue B. The PNGase F-treated
CR(¨&-Fc protein was analyzed by ESI-MS on a triple quadrupole instrument (Quattro II,
Micromass, Manchester, UK).
Fig. 3. Characterization of CR(¨&-Fc by LC-MS. The PNGase F treated CR(¨&-Fc
protein was reduced and alkylated with iodoacetamide, and digested with endoproteinase Lys-C
at a enzyme:protein ratio of 1:10 for 16 h at room temperature. The digests were analyzed by
reversed-phase HPLC on-line with an electrospray Micromass Quattro II triple quadrupole mass
spectrometer. The mass spectra were acquired by scanning the m/z range 300-1900 in 2.05
sec/scan and processed using the Micromass MassLynx data system (total ion chromatogram
from the run is shown). Peaks are labeled with the predicted Endo Lys-C fragment number (E#),
by guest on November 16, 2017http://www.jbc.org/Downloaded from

Fucosylation of Cripto is required for function
3
3
where E1 is the peptide from the N-terminus of CR(¨&-Fc and E25 the peptide from the C-
terminus of CR(¨&-Fc. The asterisk indicates the position of the peptide with the 146 Da
modification.
Fig. 4. Analysis of CR(EGF)-Fc fusion protein by ESI-MS. The EGF-like domain of
Cripto (comprising resides 75-112) was expressed as a Fc fusion protein (CR(EGF)-Fc) in CHO
cells and purified by Protein-A chromatography. The CR(EGF)-Fc was treated with PNGase F
and analyzed by ESI-MS using a triple quadrupole instrument (Quattro II, Micromass,
Manchester, UK).
Fig. 5. The CR T88A mutant is not fucosylated. The wild type CR(¨&DQGPXWDQW
CR(¨&7$SURWHLQVZHUHH[SUHVVHGWUDQVLHQWO\LQ&+2FHOOVDQGSXULILHGE\LPPXQR-affinity
chromatography. A. samples were resolved under reducing conditions by SDS-PAGE in 4-20%
Tris-glycine gels. After electrophoresis, the proteins were electrotransferred from the gel onto
nitrocellulose membrane. CR reactive bands were detected using an anti-human CR mAb
A10B2.18. lane a, wild type CR(¨&lane b, CR(¨&7$%7KH31*DVH)WUHDWHGZLOGW\SH
CR(¨&SURWHLQDQG&WKH31*DVH)WUHDWHGCR(¨&7$SURWHLQZHUHDQDO\]HGby ESI-MS
on a triple quadrupole instrument (Quattro II, Micromass, Manchester, UK). Peaks identical in
both spectra are identified with the same letter label with those in the T88A being further denoted
with a prime mark (‘).
Fig. 6. Fucosylation of CR is required for CR to activate p(n2)7-lux reporter in F9
cells. F9 cripto-/- cells were co-transfected with the FAST reporter p(n2)7-lux, and with either no
addition (control), CR WT cDNA or CR T88A cDNA, in the absence (-) or presence (+) of
FAST2 cDNA (250 ng of each cDNA per well). 48 hr following transfection, cells were lysed
with LucLite (Packard Instrument Company), and luciferase activity measured in a luminescence
counter.
by guest on November 16, 2017http://www.jbc.org/Downloaded from

Fucosylation of Cripto is required for function
3
4
Fig 7. Wild type CR but not the CR T88A mutant activates Smad2 in Xenopus
embryos. Synthetic mRNAs (amount/embryo) encoding CR (1 ng), CR T88A (1 ng), CR(¨&
ng), with or without nodal (500 pg) were injected into Xenopus embryos. Smad2 phosphorylation
was measured from stage 10 ectodermal explants. Activated Smad2 was detected by Western
blot analysis using anti-phospho-Smad2 antibody (a: α-PSmad2). Levels of total Smad2 among
different conditions were also compared (b: α-Smad2). Cytoskeletal actin was used as a loading
control (c: α-Actin). Levels of CR were assessed by Western blotting with the α-CR antibody
1579 (d: α-Cripto).
by guest on November 16, 2017http://www.jbc.org/Downloaded from

Fucosylation of Cripto is required for function
3
5
Table I
Analysis of CR(¨&-Fc by MS
CR(¨&-Fc was treated with PNGase F and analyzed by ESI-MS (see Fig. 2) on a triple
quadrupole instrument (Quattro II, Micromass, Manchester, UK). Average masses were used to
calculate theoretical masses.
Peak Observed Mass
(Da) Theoretical Mass
(Da) Identity of CR(¨&-Fc residues
A 41116 41117 31-396
B 41262 41263 31-396 + 146
C 41391 41391 31-397 + 146
D 41773 41774 31-396 + HexNAcHexNeuAc
E 41919 41920 31-396 + HexNAcHexNeuAc + 146
F 42052 42048 31-397 + HexNAcHexNeuAc + 146
G 42210 42211 31-396 + HexNAcHexNeuAc2 + 146
H 42575 42576 31-396 + HexNAc2Hex2NeuAc2 + 146
I 42870 42868 31-396 + HexNAc2Hex2NeuAc3 + 146
by guest on November 16, 2017http://www.jbc.org/Downloaded from

Fucosylation of Cripto is required for function
3
6
Table II
Peptide mapping analysis of CR(¨&-Fc using ESI-MS
CR(¨&-Fc was treated with PNGase F, digested with Endo Lys-C and fragments
analyzed on-line by LC-MS (see Fig. 3).
Retention
time (min) Observed
Mass (Da) Theoretical
Mass (Da) Cleavage product
namea Assignment
19.8 666.1 666.3 CR-E5 128-132
23.7 574.1 574.3 CR-E23 360-364
37.8 837.5 837.5 CR-E16 277-284
40.6 659.3 659.3 CR-E25 390-396 (C-1)
45.2 2311.1 2311.6 CR-E19 291-310
47.8 1639.2 1638.8 CR-E3 113-126
53.2 1160.7 1160.6 CR-E20 311-320
58.9 1677.2 1676.8 CR-E10 225-238
59.8 3044.9 3045.4 CR-E24 365-389
61.0 4473.1 4326.9 CR-E2 77-112 + 146*
61.0 5985.9 5330.0 CR-E1 c31-76 + HexNAcHexNeuAc
63.0 5329.3 5330.0 CR-E1 31-76
66.0 2972.0 2956.4 CR-E9 199-224 oxidized
69.3 2545.0 2544.7 CR-E21 321-342
70.3 2955.9 2956.4 CR-E9 199-224
73.5 1873.7 1874.1 CR-E22 343-359
79.2 5278.7 4622.3 CR-E6 c133-172 + HexNAcHexNeuAc
81.9 4622.2 4622.3 CR-E6 133-172
86.0 2619.5 2620.2 CR-E7 173-196
91.2 3463.5 3463.9 CR-E11/E12 239-267 (deamidated)
92.4 3462.5 3462.9 CR-E11/E12 239-267
93.5 3234.6 3234.6 CR-E12 241-267 (deamidated)
95.0 3233.4 3233.6 CR-E12 241-267
a25 potential fragments would be predicted from a complete digest. These potential Endo
Lys-C peptide counting from N-terminus were designated CR-E#. 7 potential cleavage products;
CR-E4 (residue 127), CR-E8 (residues 197-198), CR-E11 (residues 239-240), CR-E13 (residues
268-270), CR-E14 (residues 271-272), CR-E15 (residues 273-276), CR-E17 (residues 285-288),
by guest on November 16, 2017http://www.jbc.org/Downloaded from

Fucosylation of Cripto is required for function
3
7
and CR-E18 (residues 289-290) were not observed on the LC-MS analysis presumably due to
their small size.
bMonoisotopic masses have been used for values <1700 Da (italicized), and average
masses have been used >1700 Da.
cContains O-liked glycosylation.
by guest on November 16, 2017http://www.jbc.org/Downloaded from

Fucosylation of Cripto is required for function
3
8
Table III
Analysis of CR(¨&DQG&5¨&7$E\06
CR(¨&DQG&5¨&7$ZHUHWUHDWHGZLWK31*DVH)DQGDQDO\]HGE\(6,-MS (see
Fig. 5) on a triple quadrupole instrument (Quattro II, Micromass, Manchester, UK). Average
masses were used to calculate theoretical masses.
A. W ild type CR(¨&DIWHU31*DVH)WUHDWPHQW
Peak Observed Mass
(Da) Theoretical Mass
(Da) Identity of CR(¨&wild type residues
A 15616 15615 31-169
B 15763 15761 31-169 + 146
C 16273 16272 31-169 + HexNAcHexNeuAc
D 16418 16418 31-169 + HexNAcHexNeuAc + 146
E 16929 16928 31-169 + HexNAc2Hex2NeuAc2
F 17075 17074 31-169 + HexNAc2Hex2NeuAc2 + 146
G 17221 17219 31-169 + HexNAc2Hex2NeuAc3
H 17366 17366 31-169 + HexNAc2Hex2NeuAc3 + 146
B. CR(¨&7$DIWHU31*DVH)WUHDWPHQW
Peak Observed Mass
(Da) Theoretical Mass
(Da) Identity of CR(¨&7$UHVLGXHV
A’ 15586 15585 31-169
C’ 16242 16242 31-169 + HexNAcHexNeuAc
E’ 16899 16898 31-169 + HexNAc2Hex2NeuAc2
G’ 17190 17189 31-169 + HexNAc2Hex2NeuAc3
by guest on November 16, 2017http://www.jbc.org/Downloaded from

Fucosylation of Cripto is required for function
3
9
Table IV
Comparison of sequences of proteins possessing EGF-domains
Protein Sequence of the EGF domaina
Hu CR KELNRTC-------CLNGGTCM----L---GSF-CACPPSFYGRNCEHDVR
Mu Cripto KSLNKTC-------CLNGGTCI----L---GSF-CACPPSFYGRNCEHDVR
Oep AKQSRTC-------CKNGGTCI----L---GSF-CACPKYFTGRSCEYDER
Chick Cripto KELNRTC-------CLNGGTCM----L---GSF-CACPPSFYGRNCEHDVR
Xenopus FRL-1 KKLNRKC-------CQNGGTCF----L---GTF-CICPKQFTGRHCEHERRPA
Mouse cryptic AVPVSRC-------CHNGGTCV----L---GSF-CVCPAYFTGRYCEHDQRRR
Factor VII SDGDQC---ASSPCQNGGSCKDQ—L--QSYI--CFCLPAFEGRNCETHKDDGSA
Hu Heregulin-α SHLVKCAEKEKTFCVNGGECFMVKDLSNPSRYLCKCQPGFTGARCTENVPM
Hu TGF-αb SHFNDCPDSHTQFCFH-GTCRFLVQE---DKPACVCHSGYGVARCEHADLL
Hu EGF NSDSECPLSHDGYCLHDGVCMYIEAL---DKYACNCVVGYIGERCQYRDLKWWELR
Mu Notch1 ENNTPDC---TESSCFNGGTCVDGIN-----SFTCLCPPGFTGSYCQYDVN
Consensus
Sequence CXXGGS/TC
Hu CR T88A KELNRTC-------CLNGG
A
CM----L---GSF-CACPPSFYGRNCEHDVR
a Alignment of EGF-CFC family members with other proteins possessing EGF domains.
The alignment was constructed using Genescape Tools (Curagen Inc.). Dashed lines are gaps
introduced to fit alignment. The underlined sequences correspond to presence of the O-
fucosylation consensus sequence and residues in bold are sites of O-fucosylation. In the Hu CR
T88A sequence the site of mutation is shown in italics.
b Transforming growth factor α.
by guest on November 16, 2017http://www.jbc.org/Downloaded from

by guest on November 16, 2017http://www.jbc.org/Downloaded from

by guest on November 16, 2017http://www.jbc.org/Downloaded from

by guest on November 16, 2017http://www.jbc.org/Downloaded from

by guest on November 16, 2017http://www.jbc.org/Downloaded from

by guest on November 16, 2017http://www.jbc.org/Downloaded from

by guest on November 16, 2017http://www.jbc.org/Downloaded from

by guest on November 16, 2017http://www.jbc.org/Downloaded from

Whitman, David Salomon, Michele Sanicola and Kevin P. Williams
Chang-Yeol Yeo, Konrad Miatkowski, Heather B. Adkins, Bruno Damon, Malcom
Susan G. Schiffer, Susan Foley, Azita Kaffashan, Xiaoping Hronowski, Anne Zichittella,
Fucosylation of Cripto is required for its ability to facilitate Nodal signaling
published online August 10, 2001J. Biol. Chem.
10.1074/jbc.M104774200Access the most updated version of this article at doi:
Alerts:
When a correction for this article is posted• When this article is cited•
to choose from all of JBC's e-mail alertsClick here
by guest on November 16, 2017http://www.jbc.org/Downloaded from