Secretion and N-Linked Glycosylation Are Required for
Prostatic Acid Phosphatase Catalytic and Antinociceptive
Julie K. Hurt1, Brendan J. Fitzpatrick1, Jacqueline Norris-Drouin2, Mark J. Zylka1*
1Department of Cell and Molecular Physiology, Neuroscience Center, University of North Carolina, Chapel Hill, North Carolina, United States of America, 2Center for
Integrative Chemical Biology and Drug Discovery, University of North Carolina, Chapel Hill, North Carolina, United States of America
Secretory human prostatic acid phosphatase (hPAP) is glycosylated at three asparagine residues (N62, N188, N301) and has
potent antinociceptive effects when administered to mice. Currently, it is unknown if these N-linked residues are required
for hPAP protein stability and activity in vitro or in animal models of chronic pain. Here, we expressed wild-type hPAP and a
series of Asn to Gln point mutations in the yeast Pichia pastoris X33 then analyzed protein levels and enzyme activity in cell
lysates and in conditioned media. Pichia secreted wild-type recombinant (r)-hPAP into the media (6–7 mg protein/L). This
protein was as active as native hPAP in biochemical assays and in mouse models of inflammatory pain and neuropathic
pain. In contrast, the N62Q and N188Q single mutants and the N62Q, N188Q double mutant were expressed at lower levels
and were less active than wild-type r-hPAP. The purified N62Q, N188Q double mutant protein was also 1.9 fold less active in
vivo. The N301Q mutant was not expressed, suggesting a critical role for this residue in protein stability. To explicitly test the
importance of secretion, a construct lacking the signal peptide of hPAP was expressed in Pichia and assayed. This ‘‘cellular’’
construct was not expressed at levels detectable by western blotting. Taken together, these data indicate that secretion and
post-translational carbohydrate modifications are required for PAP protein stability and catalytic activity. Moreover, our
findings indicate that recombinant hPAP can be produced in Pichia—a yeast strain that is used to generate biologics for
Citation: Hurt JK, Fitzpatrick BJ, Norris-Drouin J, Zylka MJ (2012) Secretion and N-Linked Glycosylation Are Required for Prostatic Acid Phosphatase Catalytic and
Antinociceptive Activity. PLoS ONE 7(2): e32741. doi:10.1371/journal.pone.0032741
Editor: Maik Behrens, German Institute of Human Nutrition Potsdam-Rehbruecke, Germany
Received November 1, 2011; Accepted January 30, 2012; Published February 28, 2012
Copyright: ? 2012 Hurt et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted
use, distribution, and reproduction in any medium, provided the original author and source are credited.
Funding: This work was supported by grants to MJZ from NINDS (R01NS060725, R01NS067688). The funders had no role in study design, data collection and
analysis, decision to publish, or preparation of the manuscript.
Competing Interests: The authors have read the journal’s policy and have the following conflict: MJZ is listed as an inventor on a patent application that was
licensed for the purpose of developing PAP as a treatment for pain. Patent details: Zylka, M.J. and Vihko, P. Prostatic acid phosphatase for the treatment of pain,
U.S. Provisional patent SN# 61/003,205, filed Nov 15, 2007 to US Patent office. Full U.S. patent # 12/743,110 filed May 14, 2010. Zylka, M.J. and Vihko, P. Prostatic
acid phosphatase for the treatment of pain, International Application # PCT/US08/12849. The corresponding author provides free advice/consultation to the
commercial entity that licensed the patent. In other words, MJZ is not financially compensated for his time. These competing interests do not alter the authors’
adherence to all PLoS ONE policies on sharing materials, methods, and data. The authors are free to share everything described in this submitted manuscript.
* E-mail: firstname.lastname@example.org
Prostatic acid phosphatase (Pap, also known as Acpp) is a single gene
that encodes two extracellularly active enzymes: 1. secretory (S-)
and 2. transmembrane (TM-) PAP . S-PAP is expressed in
prostate epithelial cells and has long been used as a prostate cancer
biomarker . S-PAP was thought to be prostate specific;
however, recent studies revealed that the splice variant (TM-
PAP) was expressed in additional tissues, including salivary gland,
lung, kidney, skeletal muscle and nociceptive (pain-sensing) dorsal
root ganglia neurons [1,3]. While PAP was classically considered
to be a non-specific phosphomonoesterase (E.C. 18.104.22.168) , in vivo
PAP functions as an ectonucleotidase that hydrolyzes extracellular
adenosine 59-monophosphate (AMP) to adenosine . Deletion of
PAP reduces extracellular AMP hydrolysis in nociceptive neurons
and in the dorsal spinal cord . Moreover, S-PAP (injected
intrathecally) has long-lasting (three-day) antinociceptive effects in
mouse models of inflammatory pain and neuropathic pain and
these antinociceptive effects are entirely adenosine A1receptor
(A1R) dependent [3,6]. In addition, S-PAP has enduring (.7 days)
A1R-dependent antinociceptive effects if injected intrathecally
before nerve injury or inflammation . These findings suggest a
recombinant version of human S-PAP could be used as a
treatment for chronic pain or for preemptive analgesia .
The structure and active site of S-PAP has been extensively
characterized from various species . Mammalian S-PAP exists
primarily as a homodimer made up of two 50 kDa subunits .
Mutations that disrupt dimerization eliminate catalytic activity
. S-PAP is present at high concentrations in human semen,
which has facilitated purification and crystallization of the native
protein. Each subunit of the native human enzyme is post-
translationally modified with N-linked carbohydrate residues at
three asparagine residues (N62, N188, N301) . The crystal
structure of recombinant rat S-PAP, produced in insect cells, was
also solved and contains N-linked carbohydrates at two of these
three conserved asparagine residues (N62 and N301 but not N188)
PAP is classified as a histidine phosphatase because the catalytic
residue in the enzyme-substrate intermediate is H12 . Site-
directed mutagenesis of H12 and amino acids in the enzyme active
PLoS ONE | www.plosone.org1 February 2012 | Volume 7 | Issue 2 | e32741
site (R11, R15, H257, and D258) revealed important roles for
these residues in catalysis [11,15]. In addition, X-ray crystallo-
graphic studies revealed interactions between active site amino
acids and a known PAP inhibitor, L-(+)-tartrate [16,17,18].
Though the breadth of mutational analysis led to significant
understanding of residues important for catalysis, it is currently
unknown if secretion or N-linked glycosylation is required for the
stability and activity of PAP. In the present study, we expressed
recombinant human S-PAP (r-hPAP) in the methylotrophic yeast
species, Pichia pastoris X33. Our goals were to determine if
biologically active r-hPAP enzyme could be produced in Pichia—a
eukaryotic species that is used to generate recombinant proteins
for clinical applications [19,20]. In addition, we sought to
determine if each of the known N-linked carbohydrate residues
were important for hPAP stability and biological activity.
In vitro activity of recombinant hPAP from Pichia pastoris
To determine if recombinant human S-PAP (r-hPAP) could be
produced in Pichia, we stably integrated a wild-type version of S-
PAP into the Pichia genome under the control of the methanol-
inducible AOX1 promoter. Fourty-eight hours after induction with
1% methanol, conditioned medium and crude cell pellets were
collected for western blotting and enzyme assays. On denaturing
western blots, we observed a single 50 kDa immunoreactive band
in conditioned media (the secreted fraction) from r-hPAP
transformants but not in media from untransformed P. pastoris
X33 controls (Figure 1A). This 50 kDa band corresponded to the
known molecular weight of native S-PAP , and indicated that
r-hPAP was secreted into the medium by Pichia. r-hPAP was also
detected in the lysate (intracellular fraction) at the 48 hour time
point (Figure 1A), reflective of full-length protein within secretory
organelles (signal peptide-bearing proteins are not trafficked to the
cytoplasm). Lysates from untransformed X33 cells and r-hPAP
transformants also contained a cross-reactive (non-PAP specific)
38 kDa band.
The secreted fraction from r-hPAP transformants contained L-
(+)-tartrate-sensitive phosphatase activity that was equivalent in
activity to native, semen-derived hPAP (Figure 1B). In contrast,
neither the intracellular nor secreted fraction from untransformed
P. pastoris X33 cells had measureable phosphatase activity
(Figure 1B, Table 1). Taken together, these data indicate that r-
hPAP is secreted from Pichia in a catalytically active form.
N-linked glycosylation of recombinant hPAP in Pichia
To determine if this secreted Pichia-derived protein was
glycosylated, we treated cell supernatants containing r-hPAP with
the enzyme N-glycosidase F (PNGase F) for 24 hr. Following
treatment, the molecular weight of r-hPAP was reduced to
,40 kDa (Figure 2A). The phosphatase activity of PNGase-
treated r-hPAP was equivalent to untreated r-hPAP (Figure 2B),
indicating removal of solvent accessible carbohydrates from the
mature, fully processed protein does not affect activity.
Critical role for N-linked residues in hPAP expression
Native hPAP is glycosylated at three asparagine residues (N62,
N188, N301) . To determine if these N-linked residues were
important for protein expression, enzyme activity or secretion, we
generated single mutants of each residue, three double mutants,
and one triple mutant (see methods). We confirmed that each r-
hPAP mutant was stably integrated into the AOX1 locus by colony
PCR (Figure S1).
After inducing protein expression, the r-hPAP (N62Q) mutant
ran at a lower molecular weight than wild-type r-hPAP (Figure 3A,
B), likely due to the loss of a single high-mannose glycan at residue
62 (approximately 2 kDa) . The N62 mutant was secreted and
expressed at similar levels as wild-type r-hPAP (Figure 3B), although
this N62 mutant was significantly less active in comparison to wild-
type r-hPAP (61% relative activity) (Figure 3C, D, Table 1). The r-
hPAP (N188Q) and (N62, N188Q) double mutant were also
secreted (Figure 3B), although these mutants were expressed at very
low levels and had correspondingly low levels of activity (24% and
concentrations of L-(+)-tartrate (Figure S2A–B). Strikingly, none of
the hPAP mutants containing the N301Q mutation were expressed
inPichia(Figure 3Aand3B),an observation that wasconfirmed with
enzyme assays (Figure 3C,D, Table 1), suggesting an essential role
for this N-linked residue in protein expression and stability.
hPAP secretion is critical for activity in vitro
Since proteins are glycosylated while in transit through the
secretory pathway, and mutation of a single N-linked residue
Figure 1. Pichia-derived r-hPAP is secreted and catalytically
active. (A) Western blot of crude cell lysates (Lys) and supernatants
(Sup) from P. pastoris X33 untransformed controls and from P. pastoris
X33 expressing r-hPAP. Blot probed with anti-hPAP antiserum. (B)
DiFMUP fluorometric enzyme assay with concentrated supernatants
(secreted fractions) in comparison to native hPAP from human semen.
0.625 mg total protein used per reaction. Data are plotted as an average
of duplicate trials 6 standard deviation (SD).
Significance of N-Linked Glycans in Secretory hPAP
PLoS ONE | www.plosone.org2February 2012 | Volume 7 | Issue 2 | e32741
eliminated hPAP expression, we hypothesized that hPAP expres-
sion would also be dependent on secretion. To test this hypothesis,
we generated a version of r-hPAP that was otherwise identical to
wild-type r-hPAP except for removal of the a-factor signal peptide
(-SP). This r-hPAP(-SP) clone was stably integrated into the Pichia
AOX1 locus (Figure S1). However, this mutant was not detectably
expressed or active (Figure 4A, B), suggesting that hPAP must
transit through the secretory pathway for proper expression and
Recombinant hPAP has antinociceptive activity
Given the potential use of r-hPAP as a therapeutic in humans,
we next sought to determine if purified r-hPAP protein had
antinociceptive activity in preclinical models of pain. To test this,
we purified r-hPAP to homogeneity (see methods; typical yield
0.5 mg/L supernatant). We then measured noxious thermal
sensitivity before and after injecting pure r-hPAP (250 mU)
intrathecally into wild-type and A1R2/2mice. In parallel, native
hPAP (250 mU) and an equivalent amount of purified r-hPAP
(N62Q, N188Q) were injected into additional groups of WT and
A1R2/2mice. We observed long-lasting antinociceptive activity
(three days) in WT mice following a single injection of each protein
(Figure 5). Area under the curve (AUC) calculations revealed
comparable antinociceptive activity with native hPAP and purified
r-hPAP in wild-type mice (232629.7 and 282620.7, respectively).
The r-hPAP (N62Q, N188Q) protein was significantly less
effective (AUC 149624.5), approximately 53% of the AUC
determined for r-hPAP (Figure 5 and Table 1), consistent with
reduced activity of this protein in vitro (Table 1). None of the
proteins tested had any effect in A1R2/2mice (Figure 5), consistent
with our previous work showing that the antinociceptive effects of
PAP were entirely dependent on adenosine A1receptor activation
[3,6]. These antinociceptive effects were not due to reduced motor
function, as r-hPAP did not affect performance of wild-type or
A1R2/2mice on the rotarod test (Figure 6), a quantitative measure
of balance and mobility.
We next measured noxious thermal and mechanical sensitivity
before and after inflaming one hindpaw of wild-type and A1R2/2
mice with Complete Freund’s adjuvant (CFA, Figure 7). In this
inflammatory pain model, we found that r-hPAP (250 mU, single
intrathecal injection) had long-lasting (three day) thermal anti-
hyperalgesic and mechanical anti-allodynic effects in WT mice but
not A1R2/2mice (inflamed paw, Figure 7A, B).
In a separate group of wild-type and A1R2/2mice, we surgically
cut two of the three branches of the sciatic nerve to model
neuropathic pain (Figure 8). Six days later, mice were injected
intrathecally with r-hPAP (250 mU). We found that r-hPAP had
long-lasting (three day) thermal anti-hyperalgesic and mechanical
anti-allodynic effects WT mice but not A1R2/2mice (injured paw,
Figure 8A, B). Taken together, these data reveal that r-hPAP, like
Table 1. Phosphatase activity of r-hPAP integrants.
Cell Lysates Secreted Fractions
(± SD)Rel. Activity
(± SD) Rel. Activity
P. pastoris X33 1.0 (0.03)
r-hPAP12.2 (0.28)1 109.5 (1.44)1 282 (21)
r-hPAP (N62Q)7.8 (0.29) 0.6467.0 (2.78)0.61 n.t.
r-hPAP (N188Q)3.3 (0.16)0.2726.2 (0.45) 0.24n.t.
r-hPAP (N301Q)1.1 (0.01)
r-hPAP (N62Q, N188Q)2.5 (0.05) 0.2017.6 (1.09)0.16149 (25)
r-hPAP (N62Q, N301Q)0.8 (0.03)
,0.01 1.8 (0.05)
r-hPAP (N188Q, N301Q) 1.0 (0.03)
,0.01 1.9 (0.01)
r-hPAP (N62Q, N188Q, N301Q)1.0 (0.06)
,0.01 1.7 (0.11)
r-hPAP (-SP)0.9 (0.02)
,0.01 2.0 (0.02)
#RFU at 10 min DiFMUP assay timepoint, normalized to total protein.
*AUC of antinociceptive effects shown in Figure 5.
n.t., not tested.
Figure 2. r-hPAP is glycosylated when expressed in Pichia. (A)
Western blot of concentrated r-hPAP secreted fraction after incubation
at 37uC for 24 h with or without 1000 U PNGase. Blot probed with anti-
hPAP antiserum. (B) DiFMUP fluorometric enzyme assay using equal
amounts of untreated and PNGase-treated r-hPAP. Data are plotted as
an average of duplicate trials 6 SD.
Significance of N-Linked Glycans in Secretory hPAP
PLoS ONE | www.plosone.org3February 2012 | Volume 7 | Issue 2 | e32741
native hPAP from semen, has long-lasting antinociceptive effects in
two preclinical models of chronic pain .
PAP catalytic activity requires N-linked glycosylation sites
Native hPAP is post-translationally modified with N-linked
glycans at three solvent exposed asparagine residues: N62, N188
and N301 [12,18]. To understand the role each N-linked residue
plays in protein expression and activity, we generated a series of r-
hPAP N-glycosylation point mutants and stably introduced these
clones into a defined region of the Pichia genome. Three of these
mutants [r-hPAP (N62Q), r-hPAP (N188Q) and r-hPAP (N62Q,
N188Q)] were expressed and active (albeit less active than wild-
type r-hPAP) in the cellular fraction and secreted fraction,
suggesting key elements of the tertiary and quaternary r-hPAP
structure were conserved in these mutants. Of these N-linked
residues, N62 is farthest from H12 in the active site of hPAP, N188
is at an intermediate distance, while N301 is closest to the active
site (25.7 A˚, 20.2 A˚and 12.3 A˚between the alpha carbons of each
residue, respectively) (Figure 9). Correspondingly, the N62Q
mutant had the smallest effect on protein expression and catalytic
activity when mutated (61% activity relative to r-hPAP) (Figure 3,
Table 1). Mutation of N188 resulted in a greater loss of activity
(24% activity relative to r-hPAP) (Table 1), while the hPAP
(N62Q, N188Q) double mutant activity was further reduced (16%
activity relative to r-hPAP) (Table 1). These data thus suggest that
mutation of N-linked residues closer to the active site have a
greater effect on protein activity and stability.
Indeed, no stable protein product was observed in the cellular or
secreted fraction in any r-hPAP integrant bearing the N301Q
mutation. N301 is closely oriented with several polar residues in
the enzyme active site, namely two arginine residues (R11 and
R15) and the catalytic residue, H12 (Figure 9) . The glycoside
moiety of N301 may thus form hydrogen-bonds with amino acids
in or near the active site to direct proper protein folding. In
contrast, the distances between these N-linked residues and H12 of
the neighboring subunit in the homodimer is far greater (26.9 A˚
for N62, 43.8 A˚for N188, and 38.9 A˚for N301). This suggests the
N-linked glycans likely contribute to intrasubunit folding and
stability as opposed to intersubunit stability.
Our findings could explain why a previous attempt to express
active recombinant PAP in E. coli proved unsuccessful , as E.
coli do not glycosylate proteins on asparagine residues . Post-
translational modifications like N-linked glycosylation help direct
protein folding and protein secretion in the endoplasmic reticulum
. Without these modifications, misfolded proteins are not
carried through the secretory pathway, but are instead rapidly
degraded. This likely explains why we saw reduced amounts of
PAP or no PAP in our various N-linked mutants. Likewise, acid
phosphatase from S. cerevisiae aggregated in the endoplasmic
reticulum and was not secreted when N-linked glycosylation was
In contrast, when carbohydrates were removed from the fully-
folded and functional r-hPAP protein, we saw no loss of stability or
activity in vitro (Figure 2B). This is consistent with a previous report
showing that removal of N-linked glycans from fully-folded native
Figure 3. Expression and activity of N-linked glycosylation mutants. Western blots of (A) crude cell lysates and (B) crude secreted fractions
from the indicated P. pastoris X33 integrants. Blots probed with anti-hPAP antiserum. Equivalent amounts of total protein were loaded in each lane.
(C, D) DiFMUP fluorometric enzyme assays of (C) crude cell lysates and (D) crude secreted fractions. Data are plotted as an average of duplicate trials
Significance of N-Linked Glycans in Secretory hPAP
PLoS ONE | www.plosone.org4 February 2012 | Volume 7 | Issue 2 | e32741
PAP did not diminish catalytic activity, although stability was
reduced at high pH .
Our findings also have implications for studies describing a
‘‘cellular’’ version of PAP that reportedly dephosphorylates amino
acid residues in the cytoplasm [27,28,29]. Given that proteins are
glycosylated when in transit through the secretory pathway, that
PAP requires N-linked glycosylation for activity, and that a
‘‘cellular/cytoplasmic’’ version of PAP lacking a signal peptide was
not expressed or active (Figure 4), our data reveal that it is not
possible for PAP, when expressed by cells, to be active in the
cytoplasm. Moreover, despite extensive sequencing of transcripts
from the mouse and human genome, including numerous
transcripts encoding S- and TM-PAP in dbEST, there are no
transcripts encoding a cellular isoform of PAP (i.e., a transcript
lacking a signal peptide). As a result, any study suggesting that PAP
directly dephosphorylates residues within the cytoplasm must be
reinterpreted as indirect [27,28,29].
Recombinant S-PAP is as effective as native S-PAP in
preclinical models of chronic pain
We previously found that native hPAP, purified from human
semen, and recombinant mouse PAP, purified using a baculovirus
expression system, had long-lasting A1R-dependent antinocicep-
tive effects in mouse models of inflammatory pain and neuropathic
pain [3,6,7]. While the baculovirus expression system is ideal for
generating protein for experimental purposes, this system has only
recently been used to make recombinant protein for clinical
applications [30,31]. In contrast, Pichia pastoris is a well
characterized eukaryotic expression system and has been used to
make pharmaceutical-grade recombinant proteins for diverse
applications [19,20]. Here, we expressed r-hPAP under aerobic
conditions in Pichia and obtained protein yields (crude yield 6–
7 mg hPAP/L culture) that were significantly higher than when r-
hPAP was produced in a different yeast species, Saccharomyces
cerevisiae (crude yield 0.5 mg hPAP/L culture) . Protein yields
are typically much higher in Pichia under fermentation conditions
(grams of protein/L) , so it may be possible to further increase
yields of r-hPAP under fermentation conditions.
We found that Pichia-derived r-hPAP was as effective as native
hPAP in vivo and had A1R-dependent antinociceptive effects in
preclinical models of inflammatory pain and neuropathic pain.
Purified r-hPAP (N62Q, N188Q) also had antinociceptive effects
that lasted for three days, although the magnitude of this effect was
reduced (53% relative to purified r-hPAP) (Table 1). Note that the
same amount of total protein was used in these experiments
although the enzyme activity differed. Remarkably, when
combined with our previous work showing that heat denaturation
Figure 4. Expression and activity of r-hPAP lacking signal
peptide(-SP). (A) Western blot of crude cell lysates from untrans-
formed X33 cells and transformants expressing hPAP with (r-hPAP) or
without (-SP) a signal peptide. Blot probed with anti-hPAP antiserum.
Equivalent amounts of total protein were loaded in each lane. (B)
DiFMUP fluorometric enzyme assay with the indicated crude cell lysates.
Data are plotted as an average of duplicate trials 6 SD.
Figure 5. r-hPAP has antinociceptive properties in vivo. Antinociceptive properties of native hPAP, r-hPAP and r-hPAP (N62Q, N188Q) injected
intrathecally into wild-type (n=10) and A1R2/2mice (n=10). Equivalent unit amounts (250 mU/mouse) of native hPAP and r-hPAP were injected.
Equivalent protein amounts (0.21 mg/mL) of r-hPAP and r-hPAP (N62Q, N188Q) were injected. Paired t-tests were used to compare responses at each
time point to baseline (BL). *p,0.05, **p,0.005, ***p,0.0005. Data are plotted as means 6 standard error of the mean (SEM).
Significance of N-Linked Glycans in Secretory hPAP
PLoS ONE | www.plosone.org5 February 2012 | Volume 7 | Issue 2 | e32741
eliminates PAP catalytic activity and antinociceptive activity ,
there appears to be a direct relationship between catalytic activity
and the magnitude of antinociceptive activity. It should thus be
possible to biochemically tune the A1R-dependent antinociceptive
effects of PAP by introducing mutations that increase or decrease
catalytic activity. Ultimately, our study highlights a critical role for
post-translational modifications in regulating catalytic activity and
provides a way to generate large amounts of recombinant hPAP
for future clinical trials.
Materials and Methods
Secretory hPAP was amplified without its signal peptide (nt.
187–1248, GenBank accession # NM_001099.4) from human
placenta cDNA with Phusion DNA polymerase (New England
Biolabs). The amplified product was subcloned into the Pichia
pastoris expression vector pPICZaA (Invitrogen) in-frame with the
AOX1 promoter and a-factor signal peptide. We refer to this clone
as recombinant wild-type hPAP (r-hPAP) because the a-factor
yeast signal peptide is removed upon secretion, generating a
protein that is identical in sequence to wild-type S-PAP. r-hPAP(-
SP), a ‘‘cytoplasmic/cellular’’ version of PAP that lacks a signal
peptide, contains a 59 methionine codon (atg) in-frame with hPAP
(nt. 187–1248, GenBank accession # NM_001099.4) subcloned
into pPICZB. All constructs were confirmed by DNA sequencing.
QuikChangeH multi site-directed mutagenesis (Stratagene) was
used to mutate Asn62, Asn188, and Asn301 to glutamine according
to the manufacturer’s protocol. The following oligonucleotide
primers were used for mutagenesis (with the mutagenic nucleotides
in capital letters): Asn62, 59-gaaagagatatagaaaattcttgCaGgagtcctataaacat-
gaacagg; Asn188, 59-ctttatattgtgagagtgttcacCaGttcactttaccctcctgggccac; and
were treated with the restriction enzyme DpnI (Stratagene,
10 U) at 37uC for 90 min, and transformed into chemically
competent XL10-Gold cells. Plasmids were confirmed by DNA
sequencing. Seven mutants were prepared in total: three single
mutants [hPAP (N62Q), hPAP (N188Q), and hPAP (N301Q)],
three double mutants [hPAP (N62Q, N188Q), hPAP (N62Q,
N301Q), and hPAP (N188Q, N301Q)], and one triple mutant
[hPAP (N62Q, N188Q, N301Q)].
Transformation in P. pastoris
Each r-hPAP construct was linearized with PmeI, purified using
QIAquick PCR Purification Kit (Qiagen) and transformed into
electrocompetent Pichia pastoris X33 cells according to the Easy-
SelectTMPichia Expression Kit protocol (Invitrogen). The hPAP
constructs were stably integrated into the Pichia genome at the
AOX1 locus, with transcription driven by the methanol-inducible
AOX1 promoter. To confirm integration, genomic DNA was
isolated from Zeocin-resistant colonies by treating with lyticase
(25 U in 10 mL water, Sigma-Aldrich) for 10 min at 30uC followed
by 10 min at 280uC. Positive transformants with the Mut+
phenotype were identified by colony PCR with the 39 Pichia primer
(59-gcaaatggcattctgacatcc) and 59 Pichia primer (59-gactggttccaattgacaagc)
with TITANIUMTMTaq DNA polymerase (Clontech). The
Figure 7. Antinociceptive effects of r-hPAP in chronic inflammatory pain model. (A, B) CFA was injected into one hindpaw (CFA-arrow) of
wild-type (n=10) and A1R2/2mice (n=10). r-hPAP (250 mU) was intrathecally injected 1 day later (r-hPAP-arrow). Inflamed and non-inflamed
(control) hindpaws were tested for (A) thermal and (B) mechanical sensitivity. Data are plotted as means 6 SEM. Paired t-tests were used to compare
responses at each time point between genotypes, same paw comparisons. *p,0.05, **p,0.005, ***p,0.0005.
Figure 6. r-hPAP does not affect balance or motor function in
mice. Rotarod tests with wild-type (n=10) and A1R2/2mice (n=10)
24 h before (3 iterations, separated by 40 s) and 24 h after (2 test
iterations, separated by 40 s) intrathecal injection of r-hPAP (250 mU/
mouse). No significant differences between genotypes or treatment.
Data are plotted as means 6 SEM. The same mice shown were also
tested for thermal sensitivity, to confirm that hPAP injections were
successful (as evidenced by a significant thermal antinociceptive effect
in wild-type mice injected with hPAP, data not shown).
Significance of N-Linked Glycans in Secretory hPAP
PLoS ONE | www.plosone.org6 February 2012 | Volume 7 | Issue 2 | e32741
predicted molecular weight for each DNA species was as follows:
the endogenous AOX1 alcohol oxidase gene (2.2 kb), hPAP with
the a-factor signal peptide (1.65 kb) and hPAP without the signal
peptide (1.39 kb).
r-hPAP protein expression
A single clone from each P. pastoris X33 r-hPAP integrant (and
an X33 untransformed control) was used to inoculate BMGY
media (Invitrogen) at 28uC with shaking at 250 rpm. Cells were
grown to a final OD600=4–5. Cells were harvested by
centrifugation at 3,840 rcf for 5 min., and the supernatant was
discarded. The cell pellet was resuspended to a final OD600=1.0
in BMMY media (Invitrogen, 1% methanol). Cells were grown at
28uC with shaking in baffled flasks for 48 h, adding methanol to a
final concentration of 1% (v/v) every 24 h. The cells were
harvested by centrifugation at 3,840 rcf for 5 min. at 4uC, and the
secreted (supernatant) and cellular fractions (cell pellet) were
separated. Cells were lysed with acid-washed glass beads (Sigma-
Aldrich) according to the EasySelectTMPichia Expression Kit
protocol (Invitrogen). Total protein content in each crude
supernatant and cell lysate was measured using the Bio-Rad
Protein samples were analyzed by SDS-PAGE with 4–15%
gradient Tris-HCl polyacrylamide gels (BioRad), loading crude
secreted protein in the supernatant (0.025 mg/mL) or total protein
in the cell lysate (0.25 mg/mL). Proteins were transferred to a
nitrocellulose membrane and were probed with primary hPAP
antiserum (1:10,000; Sigma #P5664) followed by anti-rabbit
IRDye800 secondary antibody (1:20,000; Rockland #611-731-
127). Blots were imaged at 800 nm using the Licor Odyssey
Isolation and purification of hPAP
To generate protein for in vivo testing, P. pastoris X33 were
transformed with r-hPAP (200 mL BMGY), were grown to mid-
log phase (OD600=5) and then were induced with 1% methanol in
1 L BMMY media at 28uC for 72 h with shaking. Cells were
harvested by centrifugation, and secretory hPAP was precipitated
with ammonium sulfate as described . The resulting protein
pellet was resuspended in 10 mM sodium acetate, pH 5.3
(15 mL/L original culture volume) and dialyzed against 16 L of
Figure 9. Location of N-linked asparagines residues relative to
the active site of hPAP. The x-ray crystallographic structure depicts
the essential active site residue H12 and the three N-linked residues
(highlighted in yellow) in one subunit of native hPAP. Distances were
calculated in PyMOL between the alpha carbon of each amino acid.
Structure coordinates from PDB #1ND6 .
Figure 8. Antinociceptive effects of r-hPAP in neuropathic pain model. (A, B) The sural and common peroneal branches of the sciatic nerve
were ligated and then transected (injure-arrow) in wild-type (n=10) and A1R2/2mice (n=10). Six days later, r-hPAP (250 mU) was injected
intrathecally. Injured and non-injured (control) hindpaws were tested for (A) thermal and (B) mechanical sensitivity. Data are plotted as means 6 SEM.
Paired t-tests were used to compare responses at each time point between genotypes, same paw comparisons. *p,0.05, **p,0.005, ***p,0.0005.
Significance of N-Linked Glycans in Secretory hPAP
PLoS ONE | www.plosone.org7 February 2012 | Volume 7 | Issue 2 | e32741
50 mM sodium acetate, pH 5.3. Following dialysis, the crude
ammonium sulfate precipitate was concentrated to 2–5 mL total
volume in centrifuge filter devices (10,000 MW cutoff; Amicon).
The concentrated r-hPAP sample was applied to a HiLoad 26/60
Superdex 200 size-exclusion column (GE Healthcare) at a flow
rate of 2 mL/min in 25 mM Tris, pH 7.5 with 150 mM NaCl.
Fractions containing r-hPAP were combined (15 mL) and dialyzed
against 3 L of 25 mM Bis-Tris, pH 6.5. Following dialysis, r-hPAP
was purified to homogeneity by MonoQ anion-exchange chro-
matography (5/50 GL) in 25 mM Bis-Tris, pH 6.5 (1 mL/min
flow rate, elute in 125–150 mM NaCl). The purified recombinant
protein was dialyzed in 0.9% saline, pH 5.6. The r-hPAP (N62Q,
N188Q) mutant was purified using the same method. All dialysis
and purification steps were performed at 4uC.
Conditioned media containing secreted r-hPAP was concen-
trated with several rounds of high-speed buffer exchange in 10,000
MW cutoff centrifuge filters (Amicon) with 100 mM sodium
acetate, pH 5.3. Total protein concentration was determined
following buffer exchange with the BioRad protein assay (BioRad).
Reactions (50 mL total volume) containing the concentrated r-
hPAP secreted fraction (2 mg total protein) and G7 reaction buffer
(50 mM sodium phosphate, pH 7.5, New England Biolabs) were
treated with PNGase F (1000 U, New England Biolabs) for 24 h at
37uC. An untreated control (no PNGase F) was incubated at 37uC
for 24 h in parallel. Aliquots were removed after 24 h for analysis
by SDS-PAGE and the DiFMUP activity assay.
DiFMUP activity assay
Phosphatase activity of each r-hPAP protein sample was
monitored using the EnzChek Phosphatase Assay Kit (Invitrogen).
Samples (100 mL total volume) containing 25 mL hPAP protein
(0.625 mg crude secreted protein in the supernatant, or 6.25 mg
total protein in the crude cell lysate) were diluted with 25 mL of
reaction buffer (100 mM sodium acetate, pH 5.3) in a black, clear-
bottom 96-well plate (Corning). The reaction was initiated with
addition of a fluorogenic substrate, 6,8-difluoro-4-methylumbelli-
feryl phosphate (DiFMUP, 50 mL, 100 mM final concentration).
The fluorescence in each well was recorded (ex: 390 nm, em:
510 nm) every 30 s over a 60 min. total assay time. All samples
were assayed in duplicate with and without a known acid
phosphatase inhibitor, L-(+)-tartrate (70 mM). Relative activity
measurements were determined in the initial (linear) phase of the
timecourse. The ‘‘endpoint’’ fluorescence was monitored at
10 min for each reaction, and the activity was calculated with
the units RFU/mg protein. Activity for each mutant was also
calculated relative to the activity of r-hPAP.
All procedures and behavioral experiments involving vertebrate
animals were approved by the Institutional Animal Care and Use
Committee at the University of North Carolina at Chapel Hill. All
experiments were performed as previously described with male
mice during the light phase, raised under a 12:12 light:dark cycle
. C57BL/6 mice (2–4 months in age) were purchased from
Jackson Laboratories, and A1R2/2mice were backcrossed to
C57BL/6J mice for 12 generations. All mice were acclimated to
the experimenter, the room and the experimental apparatus for 3–
5 days prior to behavioral testing. Thermal sensitivity was
monitored using the Hargreaves method, where the radiant heat
source was calibrated to elicit a paw withdrawal reflex of
approximately 10 sec in naı ¨ve mice (cutoff time of 20 sec).
Mechanical sensitivity was measured with semi-flexible tips
attached to an electronic Von Frey apparatus (IITC Life Science).
Mice were intrathecally injected (5 mL) with native hPAP (Sigma),
r-hPAP, or r-hPAP (N62Q, N188Q) using acute lumbar puncture
without anesthesia . Each hPAP protein sample was dialyzed
against 0.9% saline, pH 5.6. To maintain antinociceptive activity,
we found that it was essential to store this dialyzed hPAP protein at
280uC in single-use aliquots and thaw just prior to injection. The
activity of purified r-hPAP protein was characterized with the
DiFMUP substrate in comparison to a working stock of 50 U/mL
native hPAP. For the purified r-hPAP enzyme, 50 U/mL
corresponds to a protein concentration of 0.21 mg/mL. The N-
glycosylation mutant, r-hPAP (N62Q, N188Q), was also diluted to
0.21 mg/mL. Complete Freund’s adjuvant (20 mL) was injected
under the glabrous skin to inflame one hindpaw. Spared nerve
injury (SNI) was used to model neuropathic pain . Motor
function was measured by the rotarod performance test (cutoff
time of 300 sec). Baseline measurements were collected with 3 test
iterations, separated by 40 sec for each mouse. Trials were again
performed 24 hr following injection of r-hPAP, with 2 iterations
separated by 40 sec for each mouse.
Pichia AOX1 locus. Agarose gel with ethidium bromide staining
of r-hPAP integrants into the AOX1 gene. The number of correctly
targeted, Mut+colonies relative to the total number of colonies
screened for each clone are shown below each lane.
L-(+ +)-tartrate inhibition of r-hPAP mutants.
(A, B) DiFMUP fluorometric enzyme assays of (A) crude cell
lysates and (B) crude secreted fractions with L-(+)-tartrate
(70 mM). Data are plotted as an average of duplicate trials 6 SD.
r-hPAP clones are stably integrated into the
We would like to thank the Mouse Behavioral Phenotyping Laboratory at
UNC (Sheryl Moy and Randal Nonneman) for performing the rotarod test.
Conceived and designed the experiments: JKH MJZ. Performed the
experiments: JKH BJF JND. Analyzed the data: JKH MJZ. Contributed
reagents/materials/analysis tools: JKH JND. Wrote the paper: JKH MJZ.
acid phosphatase is not a prostate specific target. Cancer Res 67: 6549–6554.
Gutman AB, Gutman EB (1938) An ‘‘acid’’ phosphatase occurring in the serum
of patients with metastasizing carcinoma of the prostate gland. J Clin Invest 17:
Zylka MJ, Sowa NA, Taylor-Blake B, Twomey MA, Herrala A, et al. (2008)
Prostatic acid phosphatase is an ectonucleotidase and suppresses pain by
generating adenosine. Neuron 60: 111–122.
4.Ostrowski WS, Kuciel R (1994) Human prostatic acid phosphatase: selected
properties and practical applications. Clin Chim Acta 226: 121–129.
Street SE, Walsh PL, Sowa NA, Taylor-Blake B, Guillot TS, et al. (2011) PAP
and NT5E inhibit nociceptive neurotransmission by rapidly hydrolyzing
nucleotides to adenosine. Mol Pain 7: 80.
Sowa NA, Vadakkan KI, Zylka MJ (2009) Recombinant mouse PAP has pH-
dependent ectonucleotidase activity and acts through A(1)-adenosine receptors
to mediate antinociception. PLoS ONE 4: e4248.
Significance of N-Linked Glycans in Secretory hPAP
PLoS ONE | www.plosone.org8 February 2012 | Volume 7 | Issue 2 | e32741
7. Sowa NA, Street SE, Vihko P, Zylka MJ (2010) Prostatic acid phosphatase Download full-text
reduces thermal sensitivity and chronic pain sensitization by depleting
phosphatidylinositol 4,5-bisphosphate. J Neurosci 30: 10282–10293.
Zylka MJ (2011) Pain-relieving prospects for adenosine receptors and
ectonucleotidases. Trends Mol Med 17: 188–196.
Hassan MI, Aijaz A, Ahmad F (2010) Structural and functional analysis of
human prostatic acid phosphatase. Expert Rev Anticancer Ther 10: 1055–1068.
10. Lee H, Chu TM, Li SS, Lee CL (1991) Homodimer and heterodimer subunits of
human prostate acid phosphatase. Biochem J 277(Pt 3): 759–765.
11. Porvari KS, Herrala AM, Kurkela RM, Taavitsainen PA, Lindqvist Y, et al.
(1994) Site-directed mutagenesis of prostatic acid phosphatase. Catalytically
important aspartic acid 258, substrate specificity, and oligomerization. J Biol
Chem 269: 22642–22646.
12. Jakob CG, Lewinski K, Kuciel R, Ostrowski W, Lebioda L (2000) Crystal
structure of human prostatic acid phosphatase. Prostate 42: 211–218.
13. Schneider G, Lindqvist Y, Vihko P (1993) Three-dimensional structure of rat
acid phosphatase. Embo J 12: 2609–2615.
14. Van Etten RL (1982) Human prostatic acid phosphatase: a histidine
phosphatase. Ann N Y Acad Sci 390: 27–51.
15. Ostanin K, Saeed A, Van Etten RL (1994) Heterologous expression of human
prostatic acid phosphatase and site-directed mutagenesis of the enzyme active
site. J Biol Chem 269: 8971–8978.
16. Lindqvist Y, Schneider G, Vihko P (1993) Three-dimensional structure of rat
acid phosphatase in complex with L(+)-tartrate. J Biol Chem 268: 20744–20746.
17. LaCount MW, Handy G, Lebioda L (1998) Structural origins of L(+)-tartrate
inhibition of human prostatic acid phosphatase. J Biol Chem 273: 30406–30409.
18. Ortlund E, LaCount MW, Lebioda L (2003) Crystal structures of human
prostatic acid phosphatase in complex with a phosphate ion and alpha-
benzylaminobenzylphosphonic acid update the mechanistic picture and offer
new insights into inhibitor design. Biochemistry 42: 383–389.
19. Tolner B, Smith L, Hillyer T, Bhatia J, Beckett P, et al. (2007) From laboratory
to Phase I/II cancer trials with recombinant biotherapeutics. Eur J Cancer 43:
20. Murasugi A (2010) Secretory expression of human protein in the Yeast Pichia
pastoris by controlled fermentor culture. Recent Pat Biotechnol 4: 153–166.
21. Morris MF, Waheed A, Risley JM, Van Etten RL (1989) Carbohydrate removal
fails to eliminate the heterogeneity of human prostatic acid phosphatase. Clin
Chim Acta 182: 9–20.
22. Risley JM, Van Etten RL (1987) Structures of the carbohydrate moieties of
human prostatic acid phosphatase elucidated by H1 nuclear magnetic resonance
spectroscopy. Arch Biochem Biophys 258: 404–412.
23. Wacker M, Linton D, Hitchen PG, Nita-Lazar M, Haslam SM, et al. (2002) N-
linked glycosylation in Campylobacter jejuni and its functional transfer into E.
coli. Science 298: 1790–1793.
24. Gething M-J (1985) Protein transport and secretion. In: Gething M-J, ed.
Current communications in molecular biology. Cold Spring Harbor: Cold
Spring Harbor Laboratory.
25. Mrsa V, Barbaric S, Ries B, Mildner P (1987) Role of glycosylation in secretion
of yeast acid phosphatase. FEBS Lett 217: 174–179.
26. Reljic R, Barbaric S, Ries B, Buxton R, Hughes RC (1992) Expression,
glycosylation and secretion of yeast acid phosphatase in hamster BHK cells.
Glycoconj J 9: 39–44.
27. Meng TC, Lin MF (1998) Tyrosine phosphorylation of c-ErbB-2 is regulated by
the cellular form of prostatic acid phosphatase in human prostate cancer cells.
J Biol Chem 273: 22096–22104.
28. Veeramani S, Yuan TC, Chen SJ, Lin FF, Petersen JE, et al. (2005) Cellular
prostatic acid phosphatase: a protein tyrosine phosphatase involved in androgen-
independent proliferation of prostate cancer. Endocr Relat Cancer 12: 805–822.
29. Chuang TD, Chen SJ, Lin FF, Veeramani S, Kumar S, et al. (2010) Human
prostatic acid phosphatase, an authentic tyrosine phosphatase, dephosphorylates
ErbB-2 and regulates prostate cancer cell growth. J Biol Chem 285:
30. Cox MM, Hollister JR (2009) FluBlok, a next generation influenza vaccine
manufactured in insect cells. Biologicals 37: 182–189.
31. Baxter R, Patriarca PA, Ensor K, Izikson R, Goldenthal KL, et al. (2011)
Evaluation of the safety, reactogenicity and immunogenicity of FluBlok(R)
trivalent recombinant baculovirus-expressed hemagglutinin influenza vaccine
administered intramuscularly to healthy adults 50–64 years of age. Vaccine 29:
32. Vihko P, Kontturi M, Korhonen LK (1978) Purification of human prostatic acid
phosphatase by affinity chromatography and isoelectric focusing. Part I Clin
Chem 24: 466–470.
33. Fairbanks CA (2003) Spinal delivery of analgesics in experimental models of
pain and analgesia. Adv Drug Deliv Rev 55: 1007–1041.
34. Shields SD, Eckert WA, 3rd, Basbaum AI (2003) Spared nerve injury model of
neuropathic pain in the mouse: a behavioral and anatomic analysis. J Pain 4:
Significance of N-Linked Glycans in Secretory hPAP
PLoS ONE | www.plosone.org9 February 2012 | Volume 7 | Issue 2 | e32741