Tài liệu Báo cáo khoa học: Mammalian Gup1, a homolog of Saccharomyces cerevisiae glycerol uptake/transporter 1, acts as a negative regulator for N-terminal palmitoylation of Sonic hedgehog doc

Mammalian Gup1, a homolog of Saccharomyces cerevisiae
glycerol uptake/transporter 1, acts as a negative regulator
for N-terminal palmitoylation of Sonic hedgehog
Yoichiro Abe
1
, Yoshiko Kita
1
and Takako Niikura
1,2,
*
1 Department of Pharmacology, Keio University School of Medicine, Tokyo, Japan
2 Department of Neurology, Georgetown University, Washington, DC, USA
Sonic hedgehog (Shh), a member of the vertebrate
Hedgehog (Hh) family [1–4], is an extracellular
secreted signaling molecule that is involved in embry-
onic patterning and organogenesis (for example, in the
dorsal–ventral polarity of the spinal cord and in the
anterior–posterior polarity in the limb bud) in a con-
centration-dependent manner [5].
Shh is initially translated as a precursor protein of
 45 kDa. After excision of the signal sequence, it
undergoes automatic cleavage to release a biologically
Keywords
Gup1; hedgehog acyltransferase;
membrane-bound O-acyltransferase;
palmitoylation; Sonic hedgehog
Correspondence
Y. Abe, Department of Pharmacology,
Keio University School of Medicine,
35 Shinanomachi, Shinjuku-ku,
Tokyo 160-8582, Japan

Fax: +81 3 3359 8889
Tel: +81 3 5363 3750
E-mail:
*Present address
Department of Neurology, Georgetown
University, Washington, DC, USA
(Received 21 August 2007, revised 9
November 2007, accepted 20 November
2007)
doi:10.1111/j.1742-4658.2007.06202.x
Mammalian glycerol uptake ⁄ transporter 1 (Gup1), a homolog of Saccharo-
myces cerevisiae Gup1, is predicted to be a member of the membrane-
bound O-acyltransferase family and is highly homologous to mammalian
hedgehog acyltransferase, known as Skn, the homolog of the Drosoph-
ila skinny hedgehog gene product. Although mammalian Gup1 has a
sequence conserved among the membrane-bound O-acyltransferase family,
the histidine residue in the motif that is indispensable to the acyltransferase
activity of the family has been replaced with leucine. In this study, we
cloned Gup1 cDNA from adult mouse lung and examined whether Gup1
is involved in the regulation of N-terminal palmitoylation of Sonic hedge-
hog (Shh). Subcellular localization of mouse Gup1 was indistinguishable
from that of mouse Skn detected using the fluorescence of enhanced green
fluorescent protein that was fused to each C terminus of these proteins.
Gup1 and Skn were co-localized with an endoplasmic reticulum marker,
78 kDa glucose-regulated protein, suggesting that these two molecules
interact with overlapped targets, including Shh. In fact, full-length Shh
coprecipitated with FLAG-tagged Gup1 by immunoprecipitation using
anti-FLAG IgG. Ectopic expression of Gup1 with full-length Shh in cells
lacking endogenous Skn showed no hedgehog acyltransferase activity as
determined using the monoclonal antibody 5E1, which was found to recog-

nize the palmitoylated N-terminal signaling domain of Shh under denatur-
ing conditions. On the other hand, Gup1 interfered with the palmitoylation
of Shh catalyzed by endogenous Skn in COS7 and NSC34. These results
suggest that Gup1 is a negative regulator of N-terminal palmitoylation of
Shh and may contribute to the variety of biological actions of Shh.
Abbreviations
CHO, Chinese Hamster ovary; CM, conditioned medium; EGFP, enhanced green fluorescent protein; ER, endoplasmic reticulum; GRP78,
78-kDa glucose-regulated protein; Gup1, glycerol uptake ⁄ transporter 1; HHAT, hedgehog acyltransferase; HRP, horseradish peroxidase;
IP, immunoprecipitation; IRES, internal ribosome entry site; MBOAT, membrane-bound O-acyltransferase; Shh, sonic hedgehog; Shh-N,
N-terminal signaling domain of Shh without cholesterol modification; Shh-Np, autoprocessed N-terminal signaling domain of Shh;
TRITC, tetramethylrhodamine isothiocyanate.
318 FEBS Journal 275 (2008) 318–331 ª 2007 The Authors Journal compilation ª 2007 FEBS
active N-terminal signaling domain of  19 kDa [6–11],
which is followed by the addition of cholesterol to its
C-terminal Gly residue, a process catalyzed by the
C-terminal catalytic domain [12]. This autoprocessed
N-terminal signaling domain of Shh (Shh-Np [9]) is also
palmitoylated at its N-terminal Cys residue by the
hedgehog acyltransferase (HHAT) called Skn [13,14], a
homolog of the Drosophila skinny hedgehog (also called
sightless, central missing,orrasp) [15–18] gene product,
in an amide-linked manner [14]. These unique lipid
modifications greatly reduce the diffusibility of Shh-Np
and tether it to the cellular membrane. However, they
are necessary to regulate the movement of the protein
to form the proper concentration gradient. The critical
role of cholesterol modification in the movement of
Hh protein has been demonstrated in both vertebrates
and invertebrates [9,19–23]. Palmitoylation is also
involved in the regulation of movement of Shh-Np in

developing mouse embryos. Loss of long-range signal-
ing of Shh protein was observed in both Skn null mice
and gene-targeted homozygous mice harboring one
nucleotide substitution on the Shh locus, from which
palmitoylation-deficient C25S-Shh is produced [13].
One explanation for the loss of long-range signaling of
the nonpalmitoylated Shh-Np in vivo is its inability to
form a diffusible multimeric Hh protein complex
[13,24,25].
In addition to its role in the movement of Hh pro-
tein, palmitoylation is also implicated in the activity of
Hh protein in both vertebrates and invertebrates. It is
indispensable for the activity of Hh in Drosophila
[15–18,26]. Similarly, palmitoylation is also required
for the induction of rodent ventral forebrain neurons
[27]. Interestingly, in contrast to Drosophila, nonpalmi-
toylated Shh-Np is significantly potent in some tissue,
for example, in chick embryo neural plate explants and
mouse limbs [13,26,28]. Moreover, even in Drosophila
tissue, nonpalmitoylated mouse Shh-Np retains some
signaling activity [16]. These findings indicate that both
palmitoylated and nonpalmitoylated mammalian
Hh proteins can act as signaling molecules. It is nota-
ble that while cholesterylation of Hh protein is an
intramolecular event catalyzed by its own C-terminal
domain, palmitoylation is an intermolecular event cat-
alyzed by Skn. Therefore, while all Shh-Np certainly
possess cholesterol adduct to their C-terminal regions,
palmitoylation of Shh-Np might be controllable. In
fact, only 30% of Shh-Np was observed to be palmi-

toylated in a mammalian cell line transfected with full-
length human Shh [14]. Thus, it is possible that, in
addition to palmitoylated Shh-Np, nonpalmitoylated
Shh-Np is also produced in vertebrates in vivo,
and that a combination of palmitoylated and non-
palmitoylated Shh-Np contributes to cell fate specifica-
tion during development.
Mammalian glycerol uptake ⁄ transporter 1 (Gup1) is
described in the National Center for Biotechnology
Information gene database as a homolog of Saccharo-
myces cerevisiae Gup1 [29] from its sequence homol-
ogy. It has also been found to have sequence
homology to Drosophila skinny hedgehog gene product
and to mammalian Skn [13,30]. The function of the
mammalian Gup1 is still unclear. However, it has a
motif characteristic of the membrane-bound O-acyl-
transferase (MBOAT) superfamily [31], like Drosoph-
ila skinny hedgehog gene product and mammalian Skn,
as well as yeast Gup1 [32]. One strange thing that has
been observed, however, is that in mammalian Gup1,
the highly conserved His residue in the motif indis-
pensable to the acyltransferase activity of the MBOAT
superfamily has been replaced with a Leu residue.
Therefore, it is possible that Gup1 has some function
related to the post-translational modification of the
mammalian hedgehog family, although it may have no
acyltransferase activity. In this work we examined
whether mammalian Gup1 has a role in regulating the
palmitoylation of Shh, by using a novel technique,
developed in this study, for detecting the palmitoylated

N-terminal fragment of Shh.
Results
Monoclonal antibody 5E1 recognizes the
N-terminal signaling domain of Shh with
palmitoylation under denaturing conditions
To understand the behavior of N-terminally palmitoy-
lated Shh in mammalian systems, we established
Chinese Hamster ovary (CHO) cell clones stably
expressing full-length mouse Shh either in the presence
or absence of mouse Skn (Y. Abe, Y. Kita & T. Niik-
ura, unpublished results). While screening the clones,
we found that 5E1, a monoclonal antibody raised
against the N-terminal domain of rat Shh expressed in
insect cells [33], recognized Shh-Np in the lysate by
western blotting only when the clones were transfected
with both Shh and Skn. 5E1 has been reported not to
work well under denaturing conditions such as western
blotting [34], whereas it has been shown to block bind-
ing of Shh to its receptor Patched and consequent sig-
nal transduction in vivo and in vitro [33]. Therefore,
5E1 is believed to recognize a particular conformation
of the N-terminal signaling domain of Shh [34]. Our
observation, however, raises the possibility that Skn
has some function that protects Shh-Np from disrupt-
ing the 5E1 epitope, even under denaturing conditions.
Y. Abe et al. A negative regulator for palmitoylation of Shh
FEBS Journal 275 (2008) 318–331 ª 2007 The Authors Journal compilation ª 2007 FEBS 319
To test this possibility, we first transiently transfected
full-length mouse Shh cDNA into several cell lines,
including CHO, HeLa, COS7 and NSC34 cells

(Fig. 1). In our examination so far, the majority of the
Shh was autoprocessed and 19-kDa Shh-Np was pre-
dominantly detected in the lysate of all lines using
another anti-Shh IgG, H-160, which was raised against
the N-terminal portion (amino acids 41–200) of
human Shh (Fig. 1A–D, lane 2). Consistent with the
previous report, 5E1 failed to recognize Shh-Np in the
lysate of CHO cells (Fig. 1A, lane 2), although it rec-
ognized full-length Shh (Fig. 1A, lanes 2 and 6). This
was also the case with Shh-Np in the lysate of HeLa
cells (Fig. 1B). Remarkably, 5E1 recognized Shh-Np in
the lysate of COS7 and NSC34 cells, even under dena-
turing conditions (Fig. 1C and D, lane 2). The differ-
ence between CHO ⁄ HeLa and COS7 ⁄ NSC34 cell lines
in the reactivity of 5E1 with Shh-Np was attributed to
the existence of endogenous Skn in the latter lines, as
determined by RT-PCR analysis (Fig. 2), suggesting
that Skn affects the reactivity of Shh-Np with 5E1,
regardless of cell type. To confirm this, we transfected
full-length Shh together with FLAG-tagged mouse Skn
into these lines. As expected, 5E1 efficiently recognized
Shh-Np in the lysate of all lines under this experimen-
tal condition without affecting the level of Shh-Np
(Fig. 1A–D, lane 3). Ectopic expression of Skn led to
a reduction in the amount of Shh-Np secreted into the
conditioned media (CM) from all lines (Fig. 1A–D,
lane 3), suggesting increased hydrophobicity of the
protein, probably as a result of palmitoylation
catalyzed by Skn. Similar results were obtained by
using monoclonal anti-Shh N-terminal fragment,

clone 171018 (data not shown). As this antibody also
acts as a neutralizing antibody, it probably recognizes
an epitope overlapping with that of 5E1.
The expression of truncated Shh lacking the C-termi-
nal domain [Shh (1–198)] results in an N-terminal sig-
naling domain of Shh without cholesterol modification
at its C terminus (Shh-N). Using H-160, Shh-N protein
was detected in the lysate of these cell lines, transiently
transfected with Shh (1–198) cDNA, at a level compara-
ble to that of Shh-Np recovered from cells transfected
with full-length Shh (Fig. 1A–D, lane 4). However, 5E1
did not recognize Shh-N in the lysate of all four lines
examined (Fig. 1A–D, lane 4), reflecting the less effi-
cient palmitoylation of Shh-N compared with Shh-Np,
as previously reported [14]. As seen in cells transfected
with full-length Shh, the co-expression of FLAG-tagged
mouse Skn resulted in greatly reduced secretion of Shh-
N into the CM and in the efficient recognition of Shh-N
in lysate by 5E1 under denaturing conditions, without
affecting the amount of Shh-N (Fig. 1A–D, lane 5).
To examine in greater detail whether the effect of
the expression of Skn on the 5E1 epitope of Shh-Np
under denaturing conditions is a result of palmitoyla-
tion at the N terminus of Shh, we substituted Ser or
Ala for the Cys25 of full-length Shh, and transiently
transfected these mutants into COS (Fig. 3A) and
NSC34 (Fig. 3B) cells. These mutants were expressed
at a level comparable to that of wild-type protein, as
determined using H-160 (Fig. 3A,B, lanes 3–6). As
expected, neither C25S-Shh-Np nor C25A-Shh-Np was

recognized by 5E1 (Fig. 3A,B, lanes 3 and 5), whereas
wild-type Shh-Np clearly was (Fig. 3A,B, lane 1). In
the presence of exogenously transfected Skn, C25A-
Shh was not recognized by 5E1 (Fig. 3A,B, lane 6).
These results indicate a strong correlation between the
N-terminal palmitoylation of Shh-N(p) and the reactiv-
ity of 5E1 with Shh-N(p). Unexpectedly, C25S-Shh-Np
retained the 5E1 epitope when Skn was exogenously
overexpressed (Fig. 3A,B, lane 4). Considering that
Skn is a member of the MBOAT superfamily [32], it is
possible that excess Skn transferred an acyl group onto
the hydroxyl group of the N-terminal Ser of C25S-
Shh-Np, although the efficiency seems much lower
than that for wild-type Shh-Np. To confirm this possi-
bility, we labeled COS7 cells with [
3
H]palmitic acid
and examined whether the radioactivity is incorporated
into C25S-Shh-Np, as observed in wild-type Shh
(Fig. 3C, lanes 1 and 2). As expected, we detected a
band corresponding to C25S-Shh-Np, as well as full-
length C25S-Shh, only when Skn was co-expressed
(Fig. 3C, lane 4).
In COS7 cells, a band migrating more slowly than
Shh-N and strongly recognized by 5E1 was observed
when Shh (1–198) alone was expressed (Fig. 3A,
lane 7, asterisk). This species was not prominently
observed in lysate from NSC34, CHO, or HeLa cells.
Thus, there may be a third post-translational modifica-
tion of the N-terminal signaling domain of Shh specific

to COS7 cells affecting the 5E1 epitope.
Gup1 acts as a negative regulator for N-terminal
palmitoylation of Shh
Mammalian Gup1 has been described in the gene data-
base cited above as a homolog of the S. cerevisiae GUP1
gene product, based on its sequence homology. Align-
ment of mouse and yeast Gup1 protein sequences
using the blastp program with BLOSUM62 as a
matrix [35] showed that these two proteins are
21% identical. However, the same program showed
that mouse Gup1 is more closely related to both mou-
se Skn (28%) and Drosophila skinny hedgehog gene
product (25%). These values were comparable to the
A negative regulator for palmitoylation of Shh Y. Abe et al.
320 FEBS Journal 275 (2008) 318–331 ª 2007 The Authors Journal compilation ª 2007 FEBS
AB
CD
Fig. 1. Expression of Shh protein in transiently transfected mammalian cell lines. CHO (A), HeLa (B), COS7 (C) and NSC34 (D) cells were
transiently transfected with pIRES2-EGFP (IG, lane 1) as a vector control, pCAG-Shh ⁄ CMV-IRES-EGFP (Shh-IG, lane 2), pCAG-Shh ⁄ CMV-Skn-
FLAG-IRES-EGFP (Shh-SF-IG, lane 3), pCAG-Shh (1–198) ⁄ CMV-IRES-EGFP (Shh-N-IG, lane 4), pCAG-Shh (1–198) ⁄ CMV-Skn-FLAG-IRES-EGFP
(Shh-N-SF-IG, lane 5), pCAG-C199A-Shh ⁄ CMV-IRES-EGFP (C199A-Shh-IG, lane 6), or pCAG-C199A-Shh ⁄ CMV-Skn-FLAG-IRES-EGFP (C199A-
Shh-SF-IG, lane 7). Construction of these plasmids is described in detailed in the Experimental procedures. Forty-eight hours after transfec-
tion, both conditioned media (indicated as CM) and cell lysates (50 lg) were collected and subjected to western blotting followed by probing
with anti-Shh N-terminal domain H-160, anti-Shh N-terminal domain 5E1, anti-EGFP, or anti-actin IgG. Both full-length Shh and the N-terminal
fragment of Shh are indicated by arrows. The C199A mutation blocks autocatalytic cleavage of Shh, resulting in production of only full-length
Shh.
Y. Abe et al. A negative regulator for palmitoylation of Shh
FEBS Journal 275 (2008) 318–331 ª 2007 The Authors Journal compilation ª 2007 FEBS 321
identity between mouse Skn and Drosophila skinny
hedgehog gene product (29%). Sequence alignment

and calculation of hydrophobicity using several
programs revealed that both proteins have a similar
structure, with a signal sequence and at least nine
transmembrane domains (Fig. 4). In addition, the open
reading frame of both Skn and Gup1 genes consists
of 11 exons; each corresponding exon of the two
genes is similar in size (Fig. 4), suggesting that
these two genes evolve from the same origin. The
transcript of Gup1 was detectable in E9.5 mouse
embryo (Fig. 2, lane 6), in which Shh transcript is also
detected [2]. These facts prompted us to examine
whether Gup1 is involved in regulating N-terminal
palmitoylation in the mammalian hedgehog family,
including Shh.
We cloned Gup1 cDNA by RT-PCR from adult
mouse lung poly (A)
+
RNA and first examined its sub-
cellular localization by transiently expressing Gup1,
whose C terminus was fused to enhanced green fluores-
cent protein (Gup1-EGFP), as well as EGFP-tag-
ged Skn (Skn-EGFP) in HeLa cells, which express little
endogenous Skn or Gup1 (Fig. 2, lane 5). Consistent
with a previous report [13], Skn–EGFP (Fig. 5A,C)
localized on the endoplasmic reticulum (ER), as deter-
mined by immunofluorescent staining of 78-kDa
glucose-regulated protein (GRP78) (Fig. 5B,C). This
was also the case with Gup1-EGFP (Fig. 5D,F), which
was co-localized with GRP78 (Fig. 5E,F). These obser-
vations imply that these two proteins interact with

overlapped targets. As the intensity of the fluorescence
of these proteins was almost the same, the level of
expression of these proteins was presumed to be simi-
lar. To confirm this, we probed western blots of lysate
extracted from COS7 cells, transiently transfected with
each plasmid containing cDNA encoding these
proteins, with anti-GFP IgG. We detected a band, with
a molecular mass of  60 kDa, in the lysate of cells
transfected with Gup1-EGFP (Fig. 5G, lane 3). How-
ever, we did not detect a band corresponding to that of
Gup1-EGFP in the lysate of cells transfected with
Skn-EGFP (Fig. 5G, lane 2). Instead, a larger smear,
which was also seen in cells transfected with
Gup1-EGFP, was observed (Fig. 5G, lane 2 and 3). It
is well known that hydrophobic membrane-bound pro-
teins are often aggregated in the SDS sample buffer
when the lysate is boiled. Therefore, we subjected the
samples to western blotting without boiling. As
expected, we observed double bands, ranging from 60
to 70 kDa, and disappearance of the larger smear in
lysates of both Skn-EGFP-transfected cells and
Gup1-EGFP-transfected cells (Fig. 5G, lanes 5 and 6).
Under this experimental condition, the level of the
expression of these two proteins was almost the same
(Fig. 5G, lanes 5 and 6). Nevertheless, we sometimes
observed a decrease in the intensity of the expected
bands and appearance of a large smear, even in an
unboiled sample of cells transfected with Skn-EGFP
(data not shown), implying that Skn is more hydro-
phobic than Gup1. The expression of FLAG-tag-

ged Gup1 in several cell lines, as detected by western
blotting of unboiled samples using anti-FLAG IgG as
a probe, revealed two major bands with molecular
masses of  45 and 40 kDa (Fig. 6A–C, lane 3). The
expression of Skn-FLAG was undetectable in some
lines (Fig. 6A,C, lane 2) even when the samples
were not boiled. However, in COS7 cells transfected
with Skn-FLAG, a band with a molecular mass
of  40 kDa was detected when probed with anti-
FLAG IgG (Fig. 6B, lane 2). These observations
suggest that Skn without EGFP is more hydrophobic
than Skn with EGFP.
As we observed that Gup1 is localized on the ER,
we examined whether Gup1 can interact with Shh
by immunoprecipitation. We transiently expressed
full-length Shh, together with Gup1-FLAG or Skn-
FLAG, in COS7 cells and immunoprecipitated these
proteins using anti-FLAG IgG. As expected, both full-
length Shh and the N-terminal fragment of Shh were
coprecipitated with Skn-FLAG (Fig. 7A, upper panel,
lane 5), whereas none of the fragment of Shh was
detected in immunoprecipitate from cells transfected
with Shh and empty vector (Fig. 7A, upper panel,
lane 4). Full-length Shh also coprecipitated with Gup1-
FLAG, indicating an interaction between Gup1 and
Shh (Fig. 7A, upper panel, lane 6).
Fig. 2. Expression of Skn and Gup1 transcripts in mammalian cell
lines. Total RNA extracted from CHO (lane 1), NSC34 (lane 2),
COS7 (lane 3), HEK293 (lane 4), HeLa (lane 5) and mouse embry-
onic day 9.5 (E9.5) embryo (lane 6) was subjected to RT-PCR analy-

sis to detect expression of Skn and Gup1 in these cells. PCR
products were separated by agarose gel electrophoresis followed
by staining with ethidium bromide. A fragment of glyceraldehyde-3-
phosphate dehydrogenase (GAPDH) was amplified as an internal
control.
A negative regulator for palmitoylation of Shh Y. Abe et al.
322 FEBS Journal 275 (2008) 318–331 ª 2007 The Authors Journal compilation ª 2007 FEBS
We further assessed whether Gup1 interacts with
Skn. We expressed Gup1-EGFP in COS7 cells, together
with either Skn-FLAG or empty vector, and subjected
the cell lysates to immunoprecipitation using anti-
FLAG IgG followed by western blotting using anti-
GFP IgG (Fig. 7B). We observed a band recognized by
anti-GFP IgG only in precipitate from cells cotrans-
fected with Gup1-EGFP and Skn-FLAG, suggesting an
interaction between Gup1 and Skn (Fig. 7B, lane 4,
arrowhead). In this series of experiments, Skn-FLAG
was barely detectable in both inputs (Fig. 7A, lower
panel, lane 2, and data not shown) and immunoprecipi-
tates (Fig. 7A, lower panel, lane 5, and data not
shown) with horseradish peroxidase (HRP)-conjugated
anti-FLAG IgG, probably because of the tendency of
Skn to be aggregated in SDS sample buffer, as demon-
strated in Figs 5 and 6. Similarly, the band recognized
with anti-GFP Ig in precipitate from cells cotransfected
with Gup1-EGFP and Skn-FLAG was much larger
than expected (Fig. 7B, lane 4, arrowhead). It is
probably aggregated Gup1-EGFP, formed as a result
of boiling to elute proteins from the immunocomplex.
Although Gup1 is predicted to be a member of the

MBOAT superfamily, the His residue indispensable to
MBOAT activity is replaced by Leu (Fig. 4, asterisk).
To examine whether Gup1 has HHAT activity, we trans-
fected Gup1 cDNA together with full-length Shh cDNA
into CHO cells. As shown in Fig. 6A, Shh-Np in the
lysate of CHO cells was not recognized by 5E1 (lane 3),
demonstrating that Gup1 has no HHAT activity.
Next, we examined whether Gup1 affects the palmi-
toylation of Shh-Np in cells expressing endoge-
nous Skn, such as COS7 and NSC34 cells, by
expressing full-length Shh in the presence of Gup1-
FLAG in these cells. Co-expression of Shh with
Gup1-FLAG resulted in a reduction of the total
amount of Shh-Np, determined using H-160 in the
A
B
C
Fig. 3. Requirement of Cys
25
of Shh for Skn-dependent retention
of the 5E1 epitope on the N-terminal fragment of Shh in western
blotting. COS7 (A) and NSC34 (B) cells were transiently transfected
with pCAG-Shh (lanes 1 and 2), pCAG-C25S-Shh (lanes 3 and 4),
pCAG-C25A-Shh (lanes 5 and 6) or pCAG-Shh (1–198) (lanes 7 and
8) together with either pFLAG-CMV5a (lanes 1, 3, 5 and 7) as a
vector control or pCMV-Skn-FLAG (lanes 2, 4, 6 and 8). Cellular pro-
teins (50 lg) were subjected to western blotting, followed by prob-
ing with anti-Shh N-terminal domain H-160, anti-Shh N-terminal
domain 5E1, or anti-actin IgG. Both full-length Shh and the N-termi-
nal fragment of Shh are indicated by arrows. The asterisk indicates

a band in the lysate of COS cells transfected with both pCAG-
Shh (1–198) and pFLAG-CMV5a (A, lane 7), migrating more slowly
than Shh-N and strongly recognized with 5E1. (C) COS7 cells were
transiently transfected with pCAG-Shh ⁄ CMV-IRES-EGFP (lane 1),
pCAG-Shh ⁄ CMV-Skn-FLAG-IRES-EGFP (lane 2), pCAG-C25S-
Shh ⁄ CMV-IRES-EGFP (lane 3), or pCAG-C25S-Shh ⁄ CMV-Skn-FLAG-
IRES-EGFP (lane 4). Twenty-four hours after transfection, cells
were labeled with [9,10-
3
H]palmitic acid for 24 h. Then the cells
were lysed and Shh was immunoprecipitated with 5E1 followed by
SDS-PAGE. Dried gel was exposed to an X-ray film to visualize
radiolabeled Shh.
Y. Abe et al. A negative regulator for palmitoylation of Shh
FEBS Journal 275 (2008) 318–331 ª 2007 The Authors Journal compilation ª 2007 FEBS 323
Fig. 4. Comparison of mouse Gup1 and mouse Skn. Mouse Gup1 (Mo Gup1) and mouse Skn (Mo Skn) were aligned based on amino acids
in their sequences conserved between them, indicated with grey boxes. Amino acids identical among mouse Gup1, mouse Skn and the Dro-
sophila skinny hedgehog gene product are indicated in red. The numbers at the right of the alignment indicate the position in the sequence.
Arrowheads above the alignment indicate the positions of introns in the encoding genes. The putative signal sequence is shown on a black
background. The putative transmembrane domains were estimated on the basis of hydrophobicity calculated using seven programs:
TMHMM,
TMPRED, HMMTOP, PSORT II, SOSUI, TOPPRED and PREDICTPROTEIN. The range of hydrophobic regions predicted by more than three programs listed
above is indicated with lines on each sequence. Within each range of hydrophobic regions, the overlapped part recognized as a putative
transmembrane domain by all the programs is represented by a thick bar. The position of the His residue in the MBOAT motif indispensable
to the activity is indicated with an asterisk.
A negative regulator for palmitoylation of Shh Y. Abe et al.
324 FEBS Journal 275 (2008) 318–331 ª 2007 The Authors Journal compilation ª 2007 FEBS
lysate of both COS7 and NSC34 cells, to 73.1% and
67.1%, respectively, as compared with that in cells
cotransfected with full-length Shh and empty vector

(Fig. 6B,C, lane 3, and Fig. 6D,E, solid column). The
levels of modified Shh-Np in COS7 and NSC34 cells,
as detected using 5E1, which is expected to recognize
palmitoylated Shh-Np, were further reduced to 6.4%
and 10.7%, respectively, as compared with control
cells, suggesting that the expression of Gup1 inhibits
palmitoylation of Shh-Np catalyzed by endoge-
nous Skn in these cells (Fig. 6B,C, lane 3, and
Fig. 6D,E, open column). It seemed that the overex-
pression of Skn-FLAG in these cells slightly increased
the level of palmitoylated Shh-Np, although the differ-
ence was not statistically significant (Fig. 6D,E).
Taken together, these observations strongly suggest
that mammalian Gup1 acts as a negative regulator of
the N-terminal palmitoylation of Shh.
Discussion
In this report, we found that mammalian Gup1, a mem-
ber of the MBOAT superfamily bearing sequence simi-
larity to HHAT, acts as a negative regulator of
N-terminal palmitoylation of Shh. Several reports have
demonstrated the critical role of N-terminal palmitoyla-
tion of Hh protein for its activity in Drosophila [15–
18,26]. Drosophila Hh protein without palmitoylation
not only loses its activity but also obstructs endogenous
Hh signaling in vivo [26]. By contrast, mammalian Shh
without palmitoylation can act in some tissues
[13,26,28]. Analysis of both Skn knockout and C25S-
Shh knockin mice revealed that the responsiveness to
nonpalmitoylated Shh-Np varied among tissues [13].
Thus, it is possible that while palmitoylated Hh-Np is

the only signaling molecule in Drosophila, both palmi-
toylated and nonpalmitoylated Shh-Nps act as signaling
molecules in mammals, and combining these molecules
produces a variety of effects on developing organs and
tissues. If this were the case, the proportion of these
molecules would have to be controlled precisely. In the
present study, mammalian Gup1 was found to interact
with full-length Shh, as determined by immunoprecipi-
tation (Fig. 7), and to inhibit the N-terminal palmitoy-
lation of Shh-Np in multiple mammalian cell lines, as
determined by western blotting using 5E1 as a probe
(Fig. 6). These results, as well as structural similarity to
both mammalian Skn and Drosophila skinny hedgehog
gene product (Fig. 4) and subcellular localization of
these proteins (Fig. 5), strongly suggest that mamma-
lian Gup1 may be involved in such a mechanism. It is
not clear how Gup1 decreases the N-terminal palmitoy-
lation of Shh. Although the N terminus of the mature
signaling domain of Shh is a Cys residue, N-terminal
palmitoylation is not S-palmitoylation but N
a
-palmitoy-
lation [14]. Other than the hedgehog family, Ga
s
is the
only example that undergoes N
a
-palmitoylation in
vertebrates, to our knowledge [36]. How the N
a

-palmi-
toylation of Ga
s
is regulated also remains unclear.
However, we assume that mammalian Gup1 competes
with Skn for Shh to prevent palmitoylation rather
than catalyzing depalmitoylation of Shh because
other known N
a
-acylations, namely N
a
-acetylation
and N
a
-myristylation, are irreversible [37,38]. Further
in vitro analyses are necessary to determine whether
Gup1 can depalmitoylate Shh-Np.
G
ABC
DEF
Fig. 5. Subcellular localization of mouse Skn and Gup1. (A–F) To
visualize the subcellular localization of mouse Skn (A–C) and Gup1
(D–F), EGFP was fused to the C terminus of these proteins and
expressed in HeLa cells. Forty-eight hours after transfection, the cells
were fixed, permeabilized and stained with an ER marker (78-kDa
glucose-regulated protein) followed by TRITC-labeled secondary anti-
body. The fluorescence of EGFP (A, C, D and F, green) and TRITC
(B, C, E and F, red) was observed using a confocal microscope.
Scale bar, 25 lm. (G) COS7 cells were transiently transfected with
pIRES2-EGFP (lanes 1 and 4) as a vector control, pCMV-Skn-EGFP

(lanes 2 and 5) and pCMV-Gup1-EGFP (lanes 3 and 6). Lysates were
extracted from these cells, and boiled (lanes 1–3) or unboiled
(lanes 4–6) samples (50 lg) were subjected to western blotting fol-
lowed by probing using monoclonal anti-GFP IgG.
Y. Abe et al. A negative regulator for palmitoylation of Shh
FEBS Journal 275 (2008) 318–331 ª 2007 The Authors Journal compilation ª 2007 FEBS 325
We also demonstrated that 5E1 recognizes the N-ter-
minal fragment of Shh under denaturing conditions
when its N terminus is palmitoylated. The property of
this antibody is useful in identifying the state of N-ter-
minal lipid modification of the protein. Although 5E1
is an antibody that recognizes an epitope in the N-ter-
minal signaling domain of Shh overlapping with the
Patched-binding region and acts as a good neutralizing
antibody [33], it was previously reported not to work
under denaturing conditions such as western blotting
[34]. However, we found that 5E1 worked in western
blotting when Shh was co-expressed with Skn (Figs 1,3
and 5), suggesting that palmitoylation at the N-termi-
nal Cys residue of the N-terminal signaling domain of
Shh protects the protein from disruption of the 5E1
epitope, even under denaturing conditions. One expla-
nation for this phenomenon might be that the palmi-
tate itself constitutes the epitope when the N-terminal
fragment of Shh is denatured. However, our results
also showed that both full-length Shh (Fig. 1, lanes 2
and 6) and Shh-N with unknown modification in
COS7 cells (Fig. 3A, lane 7, asterisk) were also recog-
nized by 5E1 in western blotting, although they were
not expected to undergo palmitoylation under those

transfection conditions. Therefore, palmitoylation may
not be a component of the 5E1 epitope but may influ-
ence the structure of the 5E1 epitope under denaturing
conditions. Crystal structure analysis revealed that the
residues Pro
42
, Lys
46
, Arg
154
, Ser
157
, Ser
178
and Lys
179
are located close to each other on the surface of the
mouse Shh-N protein and are essential for Shh-N to
bind both Patched and 5E1 [34,39,40]. Among the resi-
dues, Ser
178
at least is found to be included in the 5E1
epitope [39]. In addition, mouse Shh-N lacking the
N-terminal 25 amino acids [Shh (50–198)] loses the
ability to bind not only Patched but also 5E1 in
immunoprecipitation [34]. These observations indicate
the requirement of the N-terminal region of the Shh-
N, including Pro
42
and Lys

46
, for recognition by 5E1
in immunoprecipitation. Therefore, there arise two
possibilities. One is that the N-terminal region, includ-
ing Pro
42
and Lys
46
, may constitute the epitope for
5E1, but that under denaturing conditions it is dissoci-
ated from the other parts of the protein, probably the
C-terminal region, including Arg
154
, Ser
157
, Ser
178
and
Lys
179
. In this case, palmitoylation at the N terminus
of Shh-N(p) may support the N-terminal region being
located near other amino acids on the C-terminal
AB C
DE
Fig. 6. The effect of Gup1 on N-terminal
palmitoylation of Shh-Np. CHO (A), COS7
(B) and NSC34 (C) cells were transiently
transfected with pCAG-Shh (lanes 1–3),
together with either pFLAG-CMV-5a (lane 1)

as a vector control, pCMV-Skn-FLAG
(lane 2), or pCMV-Gup1-FLAG (lane 3).
Cellular proteins (50 lg) were subjected to
western blotting, using polyclonal anti-
Shh N-terminal domain H-160, monoclonal
anti-Shh N-terminal domain 5E1, or
monoclonal anti-FLAG IgG. The intensity of
the signals obtained from the western blot
analysis was quantified using
QUANTITY ONE
software (Bio-Rad). The effect of Skn or
Gup1 on the amount of total Shh-Np,
determined with H-160 (solid column), and
on the amount of modified Shh-Np, deter-
mined with 5E1 (open column), in COS7 (D)
and NSC34 (E) cells was expressed as the
ratio of the intensity of the band of Shh-Np
to that from control cells transfected with
Shh and empty vector in the same blot.
Values were the mean ± SD of three
independent experiments. The difference
between total Shh-Np and modified Shh-Np
was determined using a paired t-test.
*, P < 0.01; NS, not significant. The level of
Shh-Np in control cells is shown by the
dotted line.
A negative regulator for palmitoylation of Shh Y. Abe et al.
326 FEBS Journal 275 (2008) 318–331 ª 2007 The Authors Journal compilation ª 2007 FEBS
region to form the 5E1 epitope, even under denaturing
conditions. The other possibility is that the N-terminal

portion of Shh-N(p) may not constitute the 5E1
epitope but may contribute to stabilization of the 5E1
epitope located within the C-terminal portion of
Shh-N; when palmitoylated, the N-terminal portion
would retain activity, even under denaturing condi-
tions. To understand, in full, the 5E1 epitope under
denaturing conditions, extensive analyses will be
required. One clue may come from identifying the
post-translational modification of Shh-N seen in COS7
cells transfected with Shh (1–198) alone (Fig. 3A,
lane 7, asterisk).
Experimental procedures
Plasmid construction
The EcoRI–NcoI fragment of mouse Shh cDNA (kindly
provided by A. P. McMahon) was subcloned between the
EcoRI and the SmaI sites of pEGFP-N3 (Clontech, Moun-
tain View, CA, USA). Then, it was excised with SpeI,
which was blunted with Klenow fragment, and with XhoI,
and was inserted between an XhoI and the SwaI sites of
pCALNLw, resulting in pCAG-Shh. pCALNLw vector is a
6.6-kbp plasmid derived from a cosmid vector, pA-
xCALNLw (Takara, Shiga, Japan), by digestion with SalI
followed by self-ligation.
Mouse Skn was cloned from the total RNA of embry-
onic day 9.5 mouse embryo by reverse transcription using
the avian myeloblastosis virus (AMV) reverse transcriptase
first-strand synthesis kit (Life Sciences, Inc, St Petersburg,
FL, USA) followed by PCR with Taq DNA polymerase
(Promega, Madison, WT, USA) using the primers
5¢-CACACTACACTGGGAAGCAGAG ACTCCAGC-3¢

and 5¢-AGCTGGCCCAGCAGCCATACACAGTTAAAG-
3¢. The cDNA was subcloned into the EcoRV site of pBlue-
script SK(+) (Stratagene, La Jolla, CA, USA) and
sequenced using an automated sequencer (ABI-PRISM310
Genetic Analyzer; Perkin-Elmer Applied Biosystems, Foster
AB
Fig. 7. Gup1 interacts with both full-length Shh and Skn. (A) COS7 cells were transiently transfected with pCAG-Shh (lanes 1–6), together
with pFLAG-CMV-5a (vector) (lanes 1 and 4), pCMV-Skn-FLAG (Skn-F) (lanes 2 and 5), or pCMV-Gup1-FLAG (Gup1-F) (lanes 3 and 6). Cells
were lysed with IP buffer, as described in the Experimental procedures, and subjected to immunoprecipitation using anti-FLAG IgG. Then,
the samples were boiled and subjected to western blot analysis using either anti-Shh N-terminal IgG H-160 or HRP-conjugated anti-FLAG IgG
(lanes 4–6). Some of the lysate (1 ⁄ 20 volume) was unboiled and also subjected to western blot analysis as input (lanes 1–3). Both full-length
Shh and the N-terminal fragment of Shh are indicated by arrows. (B) COS7 cells were transiently transfected with pCMV-Gup1-EGFP (Gup1-
G), together with pFLAG-CMV-5a (vector) (lanes 1 and 3), or pCMV-Skn-FLAG (Skn-F) (lanes 2 and 4). Cells were lysed with IP buffer and
subjected to immunoprecipitation using anti-FLAG IgG. Then, samples were boiled and subjected to western blot analysis using anti-GFP IgG
as a probe (lanes 3 and 4). Some of the lysate (1 ⁄ 20 volume) was unboiled and also subjected to western blot analysis as input (lanes 1 and
2). Immunoglobulin G heavy and light chains (IgG-H and IgG-L, respectively) are indicated by arrows. Putative Gup1–EGFP is indicated by the
arrowhead.
Y. Abe et al. A negative regulator for palmitoylation of Shh
FEBS Journal 275 (2008) 318–331 ª 2007 The Authors Journal compilation ª 2007 FEBS 327
City, CA, USA). A BamHI site was introduced immediately
before the termination codon for addition of the FLAG-tag
to the C terminus of Skn by PCR using primers 5¢-AA
GCTTCCGGAGGCTGCTAGAGAC-3¢ and 5¢-GGATC
CAAGAACTGTGTATGTCTG-3¢. The 1.6-kbp full-
length Skn cDNA, whose termination codon was changed
to a BamHI site, was inserted between the SalI and the
BamHI sites of pFLAG-CMV5a (Sigma, St Louis, MO,
USA) (pCMV-Skn-FLAG, SF) or between the XhoI and
the Bam HI sites of pEGFP-N3 (pCMV-Skn-EGFP). The
cDNA encoding FLAG-tagged Skn was then excised by

digestion with BglII and ScaI and inserted between the
BglII and the SmaI sites of the p-internal ribosome entry
site (IRES2)-EGFP (IG) vector (Clontech), resulting in
pCMV-Skn-FLAG-IRES-EGFP (SF-IG).
Mouse Gup1 was cloned from the poly (A)
+
RNA of
adult mouse lung by RT-PCR using primers 5¢-CTCG
AGGCCATGGGCATCAAGACAGC-3¢ and 5¢-GGATC
CCTCCAGCTTCTCTCTGTCCTGC-3¢, which remove
the termination codon and add a XhoI and a BamHI
site to the 5¢ and 3¢ ends, respectively. It was then sub-
cloned into pGEM-T vector (Promega) for sequencing.
Then, the cDNA was excised with XhoI and BamHI and
subcloned between the HindIII and the BamHI sites of
pFLAG-CMV5a (pCMV-Gup1-FLAG) or between the
XhoI and the BamHI sites of pEGFP-N3 (pCMV-Gup1-
EGFP).
The EcoRI site of pCAG-Shh, and the AseI site of both
pIRES2-EGFP and pCMV-Skn-FLAG-IRES-EGFP, were
changed to a SalI site by linker ligation. Subsequently,
pCAG-Shh was digested with SalI and the fragment of
 5.5 kbp, containing CAG promoter, full-length
Shh cDNA and rabbit globin polyadenylation signals, was
inserted into the SalI site of pIRES2-EGFP and pCMV-
Skn-FLAG-IRES-EGFP, resulting in pCAG-Shh ⁄ CMV-
IRES-EGFP and pCAG-Shh ⁄ CMV-Skn-FLAG-IRES-
EGFP, respectively.
C25S and C199A mutations of Shh were introduced by
PCR using primers 5¢-CCTGCAGCAGCGGCAGGCA

AGGTTATATAG-3¢ and 5¢-GGGCCCA
GAGGCCAGG
CCGGGGCACACCAG-3¢, and primers 5¢-GGCATGC
TGGCTCGCCTGGCTGTGGAAGCA-3¢ and 5¢-GGAT
CCTGGGAAA
GCGCCGCCGGATTTGGC-3¢, respec-
tively. Shh (1–198) (Shh lacking the C-terminal catalytic
domain) was constructed by PCR, which changed the
codon TG
T corresponding to the Cys
199
to TGA (Stop)
and added an EcoRV site immediately after the stop
codon using a sense primer 5¢-GGCATGCTGGCT
CGCCTGGCTGTGGAAGCA-3¢ and an antisense pri-
mer 5¢-AAGCTTGATATC
TCAGCCGCCGG ATTTGGC- 3¢.
C25A mutation of Shh was introduced by a QuikChange
site-directed mutagenesis kit (Stratagene) using primers
5¢-CCGGGCTGGCCGCTGGCCCCGGCAGGGG-3¢ and
5¢-CCCCTGCCGGGGCCAGCGGCCAGCCCGG-3¢.
Cell culture and transient transfection
CHO cells were cultured in Ham’s F12 supplemented with
10% fetal bovine serum, 50 unitsÆmL
)1
of penicillin and
50 lgÆmL
)1
of streptomycin. COS7, HeLa, HEK293 and
NSC34 [41] cells were maintained in Dulbecco’s modified

Eagle’s medium supplemented with 10% fetal bovine serum,
50 unitsÆmL
)1
of penicillin and 50 lgÆmL
)1
of streptomycin.
CHO (5 · 10
5
cells ⁄ dish), NSC34 (2 · 10
5
cells ⁄ dish),
COS7 (2 · 10
5
cells ⁄ dish) and HeLa (2 · 10
5
cells ⁄ dish)
cells seeded onto 60-mm dishes were transfected with each
plasmid using Lipofectamine reagent (Invitrogen, Carlsbad,
CA, USA). The transfected cells were used 48 h after trans-
fection unless otherwise indicated.
RT-PCR analysis
Expression of Skn and Gup1 was determined by two-step
RT-PCR, as described above, from total RNA extracted
using Isogen (Nippon Gene, Tokyo, Japan). For Skn, PCR
was performed at 94 °C for 1 min, 65 °C for 1 min and
72 °C for 1 min (35 cycles). For Gup1, PCR was performed
at 94 °C for 1 min, 65 °C for 1 min and 72 °C for 3 min
(40 cycles). Primers used were Skn, 5¢-CTGCGTGAGCAC
CATGTTCA-3¢ and 5¢-TCTCCACAGTGACTCCCAGC-3¢;
and Gup1, 5¢-GCACAATGGGCCCATGGTACCTGC-3¢

and 5¢-GGATCCCTCCAGCTTCTCTCTGTCCTGC-3¢.
These primer sets were designed based on the mouse seque-
nce and were compatible with human species. As an inter-
nal control, glyceraldehyde-3-phosphate dehydrogenase was
amplified using the primers 5¢-TCCACCACCCTGTTGCT
GTA-3¢ and 5¢-ACCACAGTCCATGCCATCAC-3¢ (25 cycles
at 94 °Cfor1min,65°C for 1 min and 72 °C for 1 min).
Western blot analysis
Cells were washed twice with NaCl ⁄ P
i
(PBS) and lysed with
lysis buffer containing 20 mm Tris-HCl (pH 7.5), 1 mm
EDTA, 1% Triton X-100 and Complete
TM
protease inhibi-
tor cocktail tablets (Roche Diagnostics, Indianapolis, IN,
USA). Each sample was analyzed using a bicinchoninic acid
protein assay kit (Pierce, Rockford, IL, USA), and 30–50 lg
of cellular protein was subjected to SDS-PAGE, followed by
transfer to polyvinylidene difluoride membrane (Pall Life
Sciences, East Hills, NY, USA) and blocking with 10% skim
milk (Becton-Dickinson, Franklin Lakes, NJ, USA) in
NaCl ⁄ P
i
containing 0.1% Tween 20 (Wako, Osaka, Japan).
Signals were detected with enhanced chemiluminescence
reagents (GE Healthcare Bio-Sciences, Piscataway, NJ). As
for the analysis of the secreted N-terminal signaling domain
of Shh, 20 lL of CM was diluted with an equivalent volume
of 2· SDS-PAGE sample buffer and subjected to SDS-

PAGE, followed by western blotting as described above.
A negative regulator for palmitoylation of Shh Y. Abe et al.
328 FEBS Journal 275 (2008) 318–331 ª 2007 The Authors Journal compilation ª 2007 FEBS
To quantify the amount of total and modified 19-kDa
Shh-Np, the same blots were probed with both H-160 and
5E1, respectively. The intensity of the signal corresponding
to Shh-Np was quantified using quantity one software
(BioRad, Hercules, CA, USA).
Antibodies used were monoclonal anti-Shh N-terminal
fragment (5E1, 1 : 2000; DSHB); rabbit anti-Shh N-termi-
nal (H-160, 1 : 2000; Santa Cruz Biotechnology Inc, Santa
Cruz, CA); monoclonal anti-GFP (1E4, 1 : 750; MBL,
Nagoya, Japan); HRP-conjugated monoclonal anti-FLAG
(M2, 1 : 3000; Sigma); rabbit anti-actin (1 : 5000; Sigma);
HRP-conjugated goat anti-mouse IgG (1 : 5000; BioRad);
and HRP-conjugated goat anti-rabbit IgG (1 : 5000; Bio-
Rad).
Immunoprecipitation
Transfected COS cells on 60-mm dishes were washed twice
with NaCl ⁄ P
i
and lysed with immunoprecipitation (IP) buf-
fer containing 25 mm Tris-HCl (pH 7.5), 150 mm NaCl,
1% Nonidet P-40, and Complete
TM
protease inhibitor
cocktail tablets (Roche Diagnostics). Then, the lysate was
incubated with 4 lg of anti-FLAG IgG overnight at 4 °C,
followed by the addition of 20 lL of Protein G–Sepha-
rose 4 Fast Flow beads (GE Healthcare Bio-Sciences). Two

hours later, the immunocomplexes were washed four times
with the IP buffer. Then, the beads were boiled in 20 lLof
2 · SDS-PAGE sample buffer, and the eluted samples were
subjected to western blot analysis as described above.
Isotope labeling with [
3
H]palmitic acid
COS7 cells seeded onto 60-mm dishes at a density of
2 · 10
5
cells ⁄ dish were transiently transfected with pCAG-
Shh ⁄ CMV-IRES-EGFP, pCAG-Shh ⁄ CMV-Skn-FLAG-
IRES-EGFP, pCAG-C25S-Shh ⁄ CMV-IRES-EGFP, or
pCAG-C25S-Shh ⁄ CMV-Skn-FLAG-IRES-EGFP. Twenty-
four hours after transfection, cells were incubated in
Dulbecco’s modified Eagle’s medium with 10% fetal bovine
serum containing 200 lCiÆmL
)1
of [9,10-
3
H]palmitic acid
(Perkin Elmer) for 24 h and lysed with IP buffer. Shh was
immunoprecipitated from the lysate as described above
using 5E1 and subjected to SDS-PAGE. The gel was fixed
with isopropanol ⁄ water ⁄ acetic acid (25 : 65 : 10, v ⁄ v ⁄ v) for
30 min and treated with Amplify Fluorographic Reagent
(GE Healthcare Bio-Sciences) for 30 min. Then, the gel was
dried and exposed to an X-ray film at )80 °C for 14 days.
Confocal microscopy
HeLa cells were seeded onto six-well plates at a density of

8 · 10
4
cells ⁄ well and transfected with either pCMV-Skn-
EGFP or pCMV-Gup1-EGFP. Forty-eight hours after
transfection, the cells were fixed with 4% paraformaldehyde
and EGFP fluorescence was observed using an LSM510
laser-scanning confocal microscope (Carl Zeiss, Oberko-
chen, Germany). ER was visualized after staining with
rabbit anti-GRP-78 Ig (1 : 100; Sigma) followed by tetra-
methylrhodamine isothiocyanate (TRITC)-labeled swine
anti-rabbit IgG (1 : 100; DAKO, Glostrup, Denmark).
Structural analysis
The putative signal sequence of mouse Skn and Gup1 was
calculated using signalp 3.0 [42]. Hydrophobicity of the
proteins was calculated using seven programs: tmhmm,
tmpred, hmmtop, psort ii, sosui, toppred, and predict-
protein [43–49].
Acknowledgements
The authors thank Drs Masato Yasui, Sadakazu Aiso
and Masaaki Matsuoka for support; Dr Andrew
P. McMahon for providing the full-length mou-
se Shh cDNA; Dr Neil Cashman for providing NSC34
cells; Dr Tomohiro Chiba for preparation of embryonic
day 9.5 mouse embryos; Dr Dovie Wylie and Ms Tak-
ako Hiraki for expert assistance; and all members of the
Departments of Pharmacology and Anatomy at Keio
University for cooperation. The monoclonal anti-
Shh IgG (5E1) developed by Dr Thomas M. Jessell was
obtained from the Developmental Studies Hybridoma
Bank developed under the auspices of the NICHD and

maintained by the University of Iowa Department of
Biological Sciences (Iowa City, IA 52242, USA). This
work was supported, in part, by grants from Japan
Society for the Promotion of Science Grant-in-Aid for
Scientific Research (C) 16590845 (YA), 17590893 (YK),
17590894 (TN), Keio Gijuku Academic Development
funds (YA and TN), National Grant-in-Aid for the
Establishment of High-Tech Research Center in a
Private University (YA), and the Nakabayashi Trust
for ALS Research (TN). The authors gratefully dedicate
this article to late Professor Ikuo Nishimoto.
References
1 Riddle RD, Johnson RL, Laufer E & Tabin C (1993)
Sonic hedgehog mediates the polarizing activity of the
ZPA. Cell 75, 1401–1416.
2 Echelard Y, Epstein DJ, St-Jacques B, Shen L, Mohler
J, McMahon JA & McMahon AP (1993) Sonic hedge-
hog, a member of a family of putative signaling mole-
cules, is implicated in the regulation of CNS polarity.
Cell 75, 1417–1430.
3 Krauss S, Concordet J-P & Ingham PW (1993) A func-
tionally conserved homolog of the Drosophila segment
Y. Abe et al. A negative regulator for palmitoylation of Shh
FEBS Journal 275 (2008) 318–331 ª 2007 The Authors Journal compilation ª 2007 FEBS 329
polarity gene hh is expressed in tissues with polarizing
activity in zebrafish embryos. Cell 75 , 1431–1444.
4 Roelink H, Augsburger A, Heemskerk J, Korzh V,
Norlin S, Ruiz i Altaba A, Tanabe Y, Placzek M,
Edlund T, Jessell TM et al. (1994) Floor plate and
motor neuron induction by vhh-1, a vertebrate homo-

log of hedgehog expressed by the notochord. Cell 76,
761–775.
5 Ingham PW & McMahon AP (2001) Hedgehog signal-
ing in animal development: paradigms and principles.
Genes Dev 15, 3059–3087.
6 Bumcrot DA, Takada R & McMahon AP (1995) Prote-
olytic processing yields two secreted forms of
Sonic hedgehog. Mol Cell Biol 15, 2294–2303.
7 Chang DT, Lo
´
pez A, von Kessler DP, Chiang C, Si-
mandl BK, Zhao R, Seldin MF, Fallon JF & Beachy
PA (1994) Products, genetic linkage and limb patterning
activity of a murine hedgehog gene. Development 120,
3339–3353.
8 Lee JJ, Ekker SC, von Kessler DP, Porter JA, Sun BI
& Beachy PA (1994) Autoproteolysis in hedgehog pro-
tein biogenesis. Science 266, 1528–1537.
9 Porter JA, von Kessler DP, Ekker SC, Young KE, Lee
JJ, Moses K & Beachy PA (1995) The product of
hedgehog autoproteolytic cleavage active in local and
long-range signaling. Nature 374, 363–366.
10 Porter JA, Ekker SC, Park W-J, von Kessler DP,
Young KE, Chen C-H, Ma Y, Woods AS, Cotter RJ,
Koonin EV et al. (1996) Hedgehog patterning activity:
role of a lipophilic modification mediated by the car-
boxy-terminal autoprocessing domain. Cell 86, 21–34.
11 Roelink H, Porter JA, Chiang C, Tanabe Y, Chang
DT, Beachy PA & Jessell TM (1995) Floor plate and
motor neuron induction by different concentrations of

the amino-terminal cleavage product of Sonic hedgehog
autoproteolysis. Cell 81, 445–455.
12 Porter JA, Young KE & Beachy PA (1996) Cholesterol
modification of Hedgehog signaling proteins in animal
development. Science 274, 255–259.
13 Chen M-H, Li Y-J, Kawakami T, Xu S-M & Chuang
P-T (2004) Palmitoylation is required for the production
of a soluble multimeric Hedgehog protein complex and
long-range signaling in vertebrates. Genes Dev 18, 641–
659.
14 Pepinsky RB, Zeng C, Wen D, Rayhorn P, Baker DP,
Williams KP, Bixler SA, Ambrose CM, Garber EA,
Miatkowski K et al. (1998) Identification of a palmitic
acid-modified form of human Sonic hedgehog. J Biol
Chem 273, 14037–14045.
15 Amanai K & Jiang J (2001) Distinct roles of Central
missing and Dispatched in sending the Hedgehog signal.
Development 128, 5119–5127.
16 Chamoun Z, Mann RK, Nellen D, von Kessler DP,
Bellotto M, Beachy PA & Basler K (2001) Skinny
hedgehog, an acyltransferase required for palmitoyla-
tion and activity of the Hedgehog signal. Science 293
,
2080–2084.
17 Lee JD & Treisman JE (2001) Sightless has homology
to transmembrane acyltransferases and is required to
generate active Hedgehog protein. Curr Biol 11, 1147–
1152.
18 Micchelli CA, The I, Selva E, Mogila V & Perrimon N
(2002) rasp, a putative transmembrane acyltransferase,

is required for Hedgehog signaling. Development 129,
843–851.
19 Chen Y & Struhl G (1996) Dual roles for Patched in
sequestering and transducing Hedgehog. Cell 87, 553–
563.
20 Lewis PM, Dunn MP, McMahon JA, Logan M, Martin
JF, St-Jacques B & McMahon AP (2001) Cholesterol
modification of Sonic hedgehog is required for long-
range signaling activity and effective modulation of sig-
naling by Ptc1. Cell 105, 599–612.
21 Li Y, Zhang H, Litingtung Y & Chiang C (2006) Cho-
lesterol modification restricts the spread of Shh gradient
in the limb bud. Proc Natl Acad Sci USA 103, 6548–
6553.
22 Burke R, Nellen D, Bellotto M, Hafen E, Senti K-A,
Dickson BJ & Basler K (1999) Dispatched, a novel ste-
rol-sensing domain protein dedicated to the release of
cholesterol-modified Hedgehog from signaling cells. Cell
99, 803–815.
23 Tian H, Jeong J, Harfe BD, Tabin CJ & McMahon AP
(2005) Mouse Disp1 is required in sonic hedgehog-
expressing cells for paracrine activity of the cholesterol-
modified ligand. Development 132, 133–142.
24 Zeng X, Goetz JA, Suber LM, Scott WJ Jr, Schreiner
CM & Robbins DJ (2001) A freely diffusible form of
Sonic hedgehog mediates long-range signalling. Nature
411, 716–720.
25 Goetz JA, Singh S, Suber LM, Kull FJ & Robbins DJ
(2006) A highly conserved amino-terminal region of
Sonic hedgehog is required for the formation of its

freely diffusible multimeric form. J Biol Chem 281,
4087–4093.
26 Lee JD, Kraus P, Gaiano N, Nery S, Kohtz J, Fishell
G, Loomis CA & Treisman JE (2001) An acylatable
residue of Hedgehog is differentially required in Dro-
sophila and mouse limb development. Dev Biol 233,
122–136.
27 Kohtz JD, Lee HY, Gaiano N, Segal J, Ng E, Larson
T, Baker DP, Garber EA, Williams KP & Fishell G
(2001) N-terminal fatty-acylation of sonic hedgehog
enhances the induction of rodent ventral forebrain neu-
rons. Development 128, 2351–2363.
28 Williams KP, Rayhorn P, Chi-Rosso G, Garber EA,
Strauch KL, Horan GSB, Reilly JO, Baker DP, Taylor
FR, Koteliansky V et al. (1999) Functional antagonists
of sonic hedgehog reveal the importance of the N termi-
nus for activity. J Cell Sci 112, 4405–4414.
A negative regulator for palmitoylation of Shh Y. Abe et al.
330 FEBS Journal 275 (2008) 318–331 ª 2007 The Authors Journal compilation ª 2007 FEBS
29 Holst B, Lunde C, Lages F, Oliveira R, Lucas C &
Kielland-Brandt MC (2000) GUP1 and its close homo-
logue GUP2, encoding mutimembrane-spanning pro-
teins involved in active glycerol uptake in
Saccharomyces cerevisiae. Mol Microbiol 37, 108–124.
30 Feng J, White B, Tyurina OV, Guner B, Larson T &
Lee HY (2004) Synergistic and antagonistic roles of the
Sonic hedgehog N- and C-terminal lipids. Development
131, 4357–4370.
31 Hofmann K (2000) A superfamily of membrane-bound
O-acyltransferases with implications for Wnt signaling.

Trends Biochem Sci 25, 111–112.
32 Bosson R, Jaquenoud M & Conzelmann A (2006)
GUP1 of Saccharomyces cerevisiae encodes an O-acyl-
transferase involved in remodeling of GPI anchor. Mol
Biol Cell 17, 2636–2645.
33 Ericson J, Morton S, Kawakami A, Roelink H & Jessell
TM (1996) Two critical periods of Sonic Hedgehog sig-
naling required for the specification of motor neuron
identity. Cell 87, 661–673.
34 Fuse N, Maiti T, Wang B, Porter JA, Hall TMT, Leahy
DJ & Beachy PA (1999) Sonic hedgehog protein signals
not as a hydrolytic enzyme but as an apparent ligand
for Patched. Proc Natl Acad Sci USA 96, 10992–10999.
35 Henikoff S & Henikoff JG (1993) Performance evalua-
tion of amino acid substitution matrices. Proteins 17,
49–61.
36 Kleuss C & Krause E (2003) Ga
s
is palmitoylated at
the N-terminal glycine. EMBO J 22, 826–832.
37 Linder ME & Deschenes RJ (2003) New insights into
the mechanisms of protein palmitoylation. Biochemistry
42, 4311–4320.
38 Glozak MA, Sengupta N, Zhang X & Seto E (2005)
Acetylation and deacetylation of non-histone proteins.
Gene 363, 15–23.
39 Pepinsky RB, Rayhorn P, Day ES, Dergay A, Williams
KP, Galdes A, Taylor FR, Boriack-Sjodin PA &
Garber EA (2000) Mapping Sonic hedgehog-receptor
interactions by steric interference. J Biol Chem 275,

10995–11001.
40 Tanaka Hall TM, Porter JA, Beachy PA & Leahy DJ
(1995) A potential catalytic site revealed by the 1.7-
Angstrom crystal structure of the amino-terminal sig-
nalling domain of Sonic hedgehog. Nature 378, 212–
216.
41 Cashman NR, Durham HD, Blusztajn JK, Oda K, Ta-
bira T, Shaw IT, Dahrouge S & Antel JP (1992) Neuro-
blastoma · spinal cord (NSC) hybrid cell lines resemble
developing motor neurons. Dev Dyn 194 , 209–221.
42 Bendtsen JD, Nielsen H, von Heijne G & Brunak S
(2004) Improved prediction of signal peptides: Sig-
nalP3.0. J Mol Biol 340, 783–795.
43 Sonnhammer EL, von Heijne G & Krogh A (1998) A
hidden Markov model for predicting transmembrane
helices in protein sequences. Proc Int Conf Intell Syst
Mol Biol 6, 175–182.
44 Hofmann K & Stoffel W (1993) TMbase-A database
of membrane spanning protein segments. Biol Chem
Hoppe-Seyler 374, 166.
45 Tusna
´
dy GE & Simon I (2001) The HMMTOP trans-
membrane topology prediction server. Bioinformatics
17, 849–850.
46 Horton P & Nakai K (1997) Better prediction of pro-
tein cellular localization sites with the k nearest neigh-
bors classifier. Proc Int Conf Intell Syst Mol Biol 5,
147–152.
47 Hirokawa T, Boon-Chieng S & Mitaku S (1998)

SOSUI: classification and secondary structure predic-
tion system for membrane proteins. Bioinformatics 14,
378–379.
48 von Heijne G (1992) Membrane protein structure pre-
diction: hydrophobicity analysis and ‘positive inside’
rule. J Mol Biol 225, 487–494.
49 Rost B, Yachdav G & Liu J (2004) The PredictProtein
Server. Nucleic Acids Res 32 (Web Server issue), W321–
W326.
Y. Abe et al. A negative regulator for palmitoylation of Shh
FEBS Journal 275 (2008) 318–331 ª 2007 The Authors Journal compilation ª 2007 FEBS 331