Cellular Localization and Function of the Antiviral Protein, OvineMx1 (oMx1): I. Ovine Mx1 Is Secreted By Endometrial EpithelialCells Via an ‘Unconventional’ Secretory PathwayKoji Toyokawa, Stacy J. Carling*, Troy L. Ott
Department of Animal and Veterinary Science, Center for Reproductive Biology, University of Idaho, Moscow, ID, USA
Introduction
The myxovirus resistance, or Mx, proteins are 70–
80 kD intracellular antiviral proteins induced by vir-
uses and type I interferons (IFNs).1 Mx proteins are
well characterized as antiviral proteins, and form an
important early line of defense against virus infec-
tion.2 However, cytoplasmic rat Mx33 and human
MxB (Mx2 homolog) show no antiviral activity.4
Human MxB is postulated to be involved in normal
cellular functions such as nuclear import and cell
cycle progression.4 In fact, it is becoming increas-
ingly clear that Mx proteins may possess important
roles in normal cellular functions.
Interest in the roles of Mx1 proteins outside the
antiviral response comes from studies showing that
Keywords
Antiviral, Mx1, pregnancy, secretion, uterus
Correspondence
Troy L. Ott, PhD, PAS, Department of Dairy
and Animal Science, Pennsylvania State
University, 324 Henning Bldg., University Park,
PA 16802, USA.
Email: [email protected]
*Department of Human Physiology, Penning-
ton Biomedical Research Center, Baton Rouge,
LA, USA.
Submitted May 1, 2006;
revised August 16, 2006;
accepted October 13, 2006.
Citation
Toyokawa K, Carling SJ, Ott TL. Cellular
localization and function of the antiviral
protein, ovine Mx1 (oMx1): I ovine Mx1 is
secreted by endometrial epithelial cells via an
‘unconventional’ secretory pathway. Am J
Reprod Immunol 2007; 57:13–22
doi:10.1111/j.1600-0897.2006.00444.x
Problem
Embryonic loss is a major contributor to infertility. Understanding fac-
tors contributing to embryonic loss will aid in development of technol-
ogies to improve/regulate fertility in animals and humans.
Method of study
We tested the hypothesis that the antiviral protein, ovine Mx1 (oMx1),
is secreted by uterine epithelial cells. Uterine flushes were obtained from
cyclic and early pregnant ewes and examined for levels of oMx1 protein.
The pathway for ovine Mx1 secretion in ovine glandular epithelial
(oGE) cells was determined using brefeldin A (BFA), an inhibitor of the
conventional secretory pathway. Effects of BFA were determined using
b2-microglobulin (b2MG) as a marker for the conventional secretory
pathway, and interferon stimulated gene 15 (ISG15) and Galectin-1
(Gal-1) as markers for the unconventional secretory pathways.
Results
Ovine Mx1 protein levels were low in uterine flushes from cyclic ewes
and levels increased in pregnant ewes after D 15. Ovine GE cells secre-
ted oMx1 in response to interferon and secretion was not reduced by
BFA, suggesting oMx1 was secreted via an unconventional secretory
pathway. b2MG secretion was reduced by BFA, whereas ISG15 and
Gal-1 were not.
Conclusion
This is the first report that the antiviral protein, oMx1, is secreted and
provides evidence that secretion occurs via unconventional secretory
pathway(s).
ORIGINAL ARTICLE
American Journal of Reproductive Immunology 57 (2007) 13–22 ª 2007 The Authors
Journal compilation ª 2007 Blackwell Munksgaard 13
oMx1 protein is upregulated in the endometrium in
response to conceptus derived-IFNs during early
pregnancy in domestic ruminants (cattle and sheep)
and swine.5–8 In addition, ovine Mx1 (oMx1) was
recently shown to be upregulated in peripheral
blood leukocytes of pregnant ewes9and cows.10 All
three of these species have conceptuses (embryo and
associate extraembryonic membranes) that secrete a
type I IFN during early pregnancy.11
In addition to Mx proteins, conceptus-derived
IFNs induces a number of interferon stimulated
genes (ISGs) in the uteri of cattle and sheep. These
include signal transducers and activators of transcrip-
tion 1 (STAT1) and STAT2,12 interferon regulatory
factor1 (IRF-1) and IRF-2,13b2-microglobulin
(b2MG),14 interferon stimulated genes 15 (ISG15),15
granulocyte chemotactic protein-2 (GCP-2),16 and
2¢,5¢-oligoadenylate synthetase (OAS).17 In fact, it is
now clear that a large number of ISGs are regulated
in the endometrium during early pregnancy in rumi-
nants.18 A definitive role for any of these genes in
establishment and maintenance of pregnancy has
yet to be determined, but is an area of active investi-
gation.18
Among known IFNs-induced genes in the uterus,
b2MG and ISG15 were detected in uterine secretions
(uterine flushings).19,20 In addition, b2MG was
shown to be secreted in cell culture by conventional
secretory pathways involving the endoplasmic reti-
culum-golgi-trans-golgi network,21 whereas ISG15
was secreted from both non-immune and immune
cells utilizing the poorly characterized unconven-
tional secretory pathways.20
The majority of extracellular proteins are secreted
via the ER/Golgi-dependent secretory pathway; also
termed the classical secretory pathway.22 Proteins
secreted via this pathway typically contain a 13–30
amino acid hydrophobic signal peptide at their N-
terminus.23 These signal peptides are cleaved before
completion of protein transport across the membrane
of the endoplasmic reticulum.24 However, a number
of secreted proteins do not contain these signal pep-
tides including fibroblast growth factor-2 (FGF-2),25
galectin-1 (Gal-1),26 and ISG15.15,20 These secretory
proteins are exported from cells in the absence of an
intact ER/Golgi transport system27 via ‘unconven-
tional’ secretory pathways.28 Proteins that are secre-
ted via unconventional secretory pathways have
some common features; including translation on free
ribosomes, protein localization throughout the cyto-
plasm,28 the absence of hydrophobic signal
sequence,23 and insensitivity to drugs that block
classical secretion such as brefeldin A (BFA)29 and
monensin.30 Although many secreted proteins are
now known to be secreted via unconventional secre-
tory pathways, the precise mechanisms involved in
the unconventional secretory pathways remain
poorly described.
Recent results utilizing uterine gland knockout
(UGKO) ewes have highlighted the key role that
uterine gland secretions play in support of pre-
attachment conceptus development.31 Mx proteins
are large GTPases that belong to the superfamily of
mechanochemical enzymes of which dynamin is the
prototype.32 Members of this family are molecular
motors that perform cellular work. For example,
dynamin is involved in the final pinching off process
of endocytosis33 and cytokinesis.34 In addition,
human MxA was recently shown to be able to bind
and tubulate lipids in a GTP-dependent fashion.35
On the basis of these results and structural similarit-
ies with dynamins, we postulate that oMx1 is
involved in intracellular protein trafficking or secre-
tion in the uterine secretory epithelium. However, a
role for oMx1 outside the antiviral response has not
been established. Likewise, it is not known whether
oMx1 protein is itself secreted.
In studies presented here our objective was to
determine if oMx1 protein was secreted into the
uterine lumen and to address the pathway by
which it was secreted using a uterine epithelial-
derived cell line. We hypothesized that oMx1
expression would be increased during early preg-
nancy in ovine uterine flushings. On the basis of
its apparent lack of a secretion signal sequence, we
postulated that oMx1 secretion would be mediated
via the ER/Golgi-independent, or ‘unconventional’
secretory pathways.
Materials and methods
Animals and Experimental Design
Mature ewes of primarily Rambouillet breeding were
observed daily for estrus behavior in the presence of
vasectomized rams. Following two estrous cycles of
normal length (16–18 days), ewes were randomly
assigned to either cyclic or pregnant status. Ewes
assigned to pregnant status were bred to intact rams
at estrus, and 12 and 24 hr later. At estrus ewes
were randomly assigned to treatment and under-
went ovariectomy–hysterectomy (n ¼ 3 ewes/day)
TOYOKAWA ET AL.
American Journal of Reproductive Immunology 57 (2007) 13–22 ª 2007 The Authors
14 Journal compilation ª 2007 Blackwell Munksgaard
on day 15 of the estrous cycle (day 0 ¼ estrus) and
days 15 or 17 of pregnancy (day 0 ¼ mating). At
surgery, 20 mL of sterile saline was used to lavage
the uterine lumen of each ewe. Uterine flushings
were then recovered and frozen at )80�C until
assayed. Pregnancy was confirmed by the presence
of apparently normal conceptuses in the uterine flu-
shes. All procedures using animals were review and
approved by the Animal Care and Use Committee
(Animal Use Protocol 7–286).
Western Blot Analysis
Total protein (8 lg) samples were dissolved in 1X
sample buffer [4X: 7.5 mL dH2O/760 mg Tris base/
2 g sodium dodecyl sulfate (SDS)/10 mL glycerol/pH
to 6.8/5 mL 2-bmercaptoethanol/300 lL bromphenol
blue]. Proteins were separated on a 15% SDS–poly-
acrylamide gel electrophoresis (PAGE) gel with 8%
stacking gel in 1X electrode buffer (10X: 30.3 g Tris
base/144.2 g glycine/10 g SDS/pH to 8.3/add dH2O to
1.0 L) at a constant current 70 mA for approximately
2 hr. The proteins in the gel were transferred to
nitrocellulose membranes (Protran BA83; pore
size ¼ 0.2 lm, Schleicher & Schuell, Keene, NH,
USA) in a Mini-Protean II Cell apparatus (Bio-Rad
Laboratories, Hercules, CA, USA) at a constant 70 V
for 90 min with an ice pack. Non-fat milk was used
as a blocker during detection of oMx1 and beta-actin,
and bovine serum albumin (BSA; Fisher Scientific,
Pittsburgh, PA, USA) was used as a blocker during
detection of b2MG, ISG15 and Gal-1. After blocking
non-specific binding with 5% non-fat milk or BSA
for 2 hr, immunoblotting was performed with tris-
buffered saline tween-20 (TBST) (20 mL 1 m Tris pH
7.5/137 mL 1 m NaCl/500 lL Tween 20/dH2O to
1.0 L). Immunoblotting was performed with 1:1000
dilution of rabbit polyclonal human b2MG antiserum
(Abcam, Cambridge, MA, USA), 1:1000 dilution of
an amino terminal rabbit polyclonal ovine Mx1 anti-
serum36 (90618 bleed no. 2, prepared by Multiple
Peptide Systems, San Diego, CA, USA) or 1:2000
dilution of rabbit polyclonal bovine ISG15 antiserum
(Provided by Dr Thomas R. Hansen, Colorado State
University) or 1:500 dilution of goat anti-human
Gal-1 (R&D, MN) in TBST including 2% BSA at 4�Covernight. Following washing with TBST for 5 min
four times on a rocker, the membranes were then
incubated with TBST including 2% non-fat milk or
BSA and 1:100,000 dilution of goat anti-rabbit
immunoglobulin G (IgG)-horseradish peroxidase
conjugate (Pierce, Rockford, IL, USA) or rabbit anti-
goat IgG-horseradish peroxidase conjugate on a
rocker at room temperature for 1 hr. Following
washing with TBST four times for 10 min on a
rocker, the membranes was incubated with Super
Signal� West Femto Maximum Sensitivity Substrate
chemiluminescence kit (Pierce) to detect immuno-
reactive proteins for 10 min at room temperature.
Detection of the chemiluminescence signal was per-
formed and quantified using the Bio-Rad Fluor-S
Multiimager system (40-mm lens, F-stop ¼ 2.7,
exposure ¼ 360 s) and Quantity One software (Bio-
Rad Laboratories).
Immunofluorescence
Immortalized ovine glandular epithelial (oGE)
cells37(300,000 cells/well) were cultured in 6-well
plates (Coster� 3516; Corning Inc., Corning, NY,
USA) for 24 hr in Dulbecco’s Modified Eagle’s Med-
ium (15.63 g/L DMEM; Sigma, St Louis, MO, USA)
with 10% FBS under 5% CO2 at 38.5�C. Cells were
plated on three cover slips per well in 6-well plates.
Cells on the cover slips were collected at 0, 1, 2, 3,
4, 8, and 72 hr after IFNs treatment (10,000 U/mL;
provided by Fuller W. Bazer, Texas A&M Univer-
sity). Cells on the cover slips were rinsed with 1 mL
of 1X phosphate-buffered saline (PBS) (10X: 80 g
NaCl/2 g KCl/14.4 g Na2HPO4/2.4 g KH2PO4/dH2O
to 1.0 L/adjust pH to 7.2). Following washing, cells
were fixed with 1 mL of 3% formaldehyde (Poly-
sciences Inc., Warrington, PA, USA) for 10 min on a
rocker and were washed with 1 mL of 1X PBS for
5 min twice on a rocker. Non-specific binding was
blocked with 1 mL of Base (1X PBS/0.05% Triton X-
100/2% BSA) overnight at 4�C. Following overnight
blocking with Base, cells were incubated in 1 mL of
1X PBS-2% BSA including 1:1000 dilution of an
amino terminal rabbit polyclonal ovine Mx1 anti-
serum and 1:200 dilution of mouse monoclonal to
58 K Golgi protein (Abcam) for approximately 6 hr
at room temperature. For oMx1 staining, pre-
immune rabbit serum (Multiple Peptide Systems) at
a dilution of 1:1000 was used as a negative control.
Cells were also incubated in 1X PBS–2% BSA with-
out immune serum as a negative control (no pri-
mary Ab control). After incubation with primary
antibodies, cells were washed with Base three times
for 5 min on a rocker at RT. Following washing,
cells were incubated in 1 mL of 1X PBS–2% BSA
including 1:2000 dilution of goat anti-rabbit Alexa
UTERINE SECRETION OF THE MYXOVIRUS RESISTANCE PROTEIN MX1
American Journal of Reproductive Immunology 57 (2007) 13–22 ª 2007 The Authors
Journal compilation ª 2007 Blackwell Munksgaard 15
488 (Molecular Probe Inc., Eugene, OR, USA) and
goat anti-mouse Alexa 555 (Molecular Probe Inc.)
for 1 hr at room temperature. Nuclei of cells were
stained by Hoechst reagent (Sigma) for 5 min on a
rocker. After washing cells with 1X PBS, each cover
slip was whole-mounted in a mounting medium
(Vector, Burlingame, CA, USA) on a glass slide
sealed by nail polish. Immunofluorescence signals
were detected using an Eclipse E 1000 fluorescence
photomicroscope (Nikon, Tokyo, Japan) and Meta-
Morph version 6.1 (Universal Imaging Corpora-
tionTM, Downingtown, PA, USA).
Amino Acid Sequences
Prediction of the signal peptide cleavage sites in
amino acid sequences was analyzed using SignalP
3.0.38 Amino acid sequences of ovine b2MG, oMx1,
ISG15 and Gal-1 were obtained from NCBI (acces-
sion numbers AY549962, X66093, AF152103 and
AY694970, respectively).
Cell culture and brefeldin A treatment
Immortalized oGE cells (1,200,000 cells/dish)37 were
cultured in Ten-twenty-nineTM Petri dishes (Falcon,
100 X 15 mm style, Becton Dickinson Labware,
Franklin Lakes, NJ, USA) for 24 hr in DMEM with
10% FBS under 5% CO2 at 38.5�C. Following 24 hr
culture, cells (approximately 80% confluent) were
further cultured for another 24 hr in DMEM with
10% FBS in the presence of IFNs (10,000 U/mL).
The medium was replaced at 24 hr after IFNs treat-
ment with DMEM in the absence of FBS or phenol
red (Sigma). Cells were randomly assigned to con-
trols or treatments. Controls were non-treatment
(NT) without IFNs or BFA, IFN with IFNs plus BFA
at 0.01, 0.025 or 0.05 lg/mL (BFA 0.01, BFA 0.025,
and BFA 0.05). Conditioned media were collected
12 hr later. At harvest, cells were approximately
95% confluent and showed no visual evidence of
toxicity. Each experiment was replicated three inde-
pendent times with three dishes per treatment
within each experiment. Conditioned media were
centrifuged at 10,000 · g for 10 min to remove cell
debris and concentrated using Amicon� Ultra-4 cen-
trifugal Filter Devices (Millipore, Billerica, MA, USA)
according to manufacturer’s directions. Cell lysates
from the same dish were also collected using Mam-
malian Protein Extraction ReagentTM (MPERTM;
Pierce) to detect intracellular levels of b2MG, oMx1,
ISG15, and Gal-1. Protein concentrations in condi-
tioned media and cell lysates were determined by
BCA protein assay (Pierce). Total protein (40 lg)
from concentrated conditioned media or cell lysates
was used for Western blot assay to determine levels
of b2MG, oMx1, ISG15 and Gal-1 in conditioned
media or cell lysates.
Data Analysis
For effects of BFA on b2MG in conditioned media,
data were analyzed using the GLM procedures of the
Statistical Analysis System (SAS, version 8.0; SAS
Institute, Cary, NC, USA). The model tested the
main effects of treatment (NT, IFN, IFN + BFA),
experiment (three independent replicates) and their
interaction on levels of b2MG in conditioned media.
Experiment was not a significant source of variation
and was dropped from the model for all subsequent
analyses. Following a significant F-test (P < 0.05),
the following set of orthogonal contrasts were tested;
NT versus IFN and NT versus IFN + BFA. Because of
problems associated with heterogenous variances
that could not be alleviated with data transforma-
tion, effects of BFA on oMx1, ISG15 and Gal-1 in
condition media were determined using the Bonfer-
roni procedure (SAS, version 8.0). This approach is
considered conservative and holds the experiment-
wise error rate and alpha. Protein levels are reported
as least squares means ± pooled standard error of
the mean (pSEM).
Results
Ovine Mx1 Expression in Uterine Flushes from
Cyclic and Pregnant Ewes
Uterine flushes from cyclic and pregnant ewes were
examined for presence of oMx1 protein (Fig. 1).
Ewe 1 2 3 1 2 3 1 2 3
D15PD15C D17P PC
oMx1
Fig. 1 Western blot analysis of oMx1 protein in uterine flushes from
cyclic and pregnant ewes. D, day; C, cyclic; P, pregnant; PC, positive
control of consisting of proteins from D15P ovine endometrium; n,
three animals for each status and day. Ovine Mx1 was detected
approximately at 80 kDa in ovine uterine flushes.
TOYOKAWA ET AL.
American Journal of Reproductive Immunology 57 (2007) 13–22 ª 2007 The Authors
16 Journal compilation ª 2007 Blackwell Munksgaard
Ovine Mx1 was detected at approximately � 80 kDa
in ovine uterine flushes from all three day-17 preg-
nant ewes, whereas oMx1 expression in uterine
flushes from day 15 cyclic ewes was relatively low,
but detectable. Ovine Mx1 levels in flushes from day
15 pregnant ewes were intermediate but variable.
Time Course Induction of oMx1 in oGE Cells
To examine the induction and localization of intra-
cellular oMx1, we conducted time course experi-
ment in oGE cells treated with IFNs using
immunofluorescence microscopy (Fig. 2). At 0 hr,
expression of oMx1 was very low to undetectable.
Through 1 and 2 hr after induction, expression of
oMx1 increased around the nucleus. Ovine Mx1
protein localization started to expand towards the
periphery of the cells at 3–4 hr after induction.
There was no apparent co-localization of oMx1 with
the Golgi apparatus (stained red in Fig. 2). Strong
expression of oMx1 was observed in the cytoplasm
at 8 hr after induction and this staining pattern per-
sisted through 72 hr after induction. Negative con-
trols included use of pre-immune serum and no
primary antibody in cells collected at 8 hr after treat-
ment with IFNs.
Prediction of the Signal Peptide Cleavage Sites in
Amino Acid Sequences from b2MG, oMx1, ISG15
and Gal-1
Detection of a signal peptide cleavage site(s) in
amino acid sequences from ovine b2MG, oMx1,
ISG15 and Gal-1 was accomplished using SignalP
3.0.38b2MG contained a conventional secretion sig-
nal peptide and was predicted (probability 1.000) to
be a classically secreted protein (Fig. 3a). The predic-
ted cleavage site was between amino acids 20 and
21 in the ovine b2MG sequence. In contrast, oMx1,
ISG15 and Gal-1 did not contain a signal peptide
and are predicted to be secreted via non-classical
secretory pathways (Fig. 3b–d).
Effects of BFA on b2MG, oMx1, ISG15, and Gal-1
Levels in Secretion
Ovine glandular epithelial cells were treated with dif-
ferent concentrations of BFA to examine effects on
oMx1 secretion; a protein that was previously not
known to be secreted (Fig. 4). b2MG is an IFNs-
stimulated protein known to be secreted via the
classical pathway that is a component of uterine
secretions. ISG15 is an IFNs-stimulated protein in
uterine secretions that is secreted via an unconven-
tional secretory pathway. Gal-1 is an unconvention-
ally secreted protein that is not known to be induced
by IFNs. Western blot analysis detected b2MG in
medium conditioned by oGE cells treated with IFNs.
As expected, secretion of the 11.8 kDa b2MG was
blocked by BFA in a dose-dependent manner with-
out affecting intracellular levels of b2MG. Ovine
Mx1, ISG15, and Gal-1 were detected at approxi-
mately � 80 kDa, � 17 kDa and � 15 kDa, respect-
ively, and their levels in oGE-conditioned medium
were not reduced by BFA. As expected, levels of
oMx1 and ISG15 in cell lysates were increased in
0 hr
4 hr
No 1A
8 hr 72 hr Preimmune
1 hr 2 hr 3 hr
Fig. 2 Immunofluorescence analysis of the
time course for induction of oMx1 in inter-
feron-s (IFNs)-treated oGE cells. Green, red,
and blue show expression of oMx1, Golgi, and
nuclei of oGE cells, respectively. Pre-immune
serum and no primary antibody were used as
negative controls. Magnification used was
100·.
UTERINE SECRETION OF THE MYXOVIRUS RESISTANCE PROTEIN MX1
American Journal of Reproductive Immunology 57 (2007) 13–22 ª 2007 The Authors
Journal compilation ª 2007 Blackwell Munksgaard 17
response to IFNs, whereas Gal-1 was constitutively
expressed and not affected by IFNs treatment. Levels
of oMx1, ISG15 and Gal-1 in secretions were
increased by BFA in a dose-dependent manner with
no changes in intracellular levels of these proteins.
Levels of b2MG, oMx1, and ISG15 in conditioned
media were assayed in triplicate in three independ-
ent experiments for b2MG, oMx1, ISG15, and Gal-1
in two independent experiments (Fig. 5). IFNsinduced a modest increase in amount of b2MG in
oGE-conditioned culture media compared to NT
(1.5-fold increase; P ¼ 0.03). Levels of b2MG in oGE
cells treated with 0.05 lg/mL BFA were reduced
compared to NT alone (50% reduction; P < 0.01).
IFN alone did not increase oMx1 (P ¼ 0.18), ISG15
(P ¼ 0.6) or Gal-1 (P ¼ 0.86) in oGE-conditioned
media compared with NT at 12 hr. In contrast, levels
of oMx1, ISG15, and Gal-1 in conditioned media
were dramatically enhanced by BFA treatment com-
pared with NT (6.4-fold, 14-fold, and 16-fold
increase P < 0.01, respectively).
Discussion
Results presented here provide the first evidence
that the myxovirus resistance protein, oMx1, is a
secreted protein. We showed that oMx1 was pre-
sent in uterine flushes from pregnant ewes and
Fig. 3 Prediction of the signal peptide clea-
vage sites in amino acid sequences from
b2MG, (a) oMx1 (b), ISG15 (c) and Gal-1 (d)
using SignalP 3.0. C (red), S (green) and Y
(blue)-scores indicate cleavage site, ‘signal
peptideness’ and combined cleavage site
predictions, respectively.
BFA µg/mLIFN
β2MG
oMx1
Secretions
Cell lysates
Secretions
Cell lysates
Secretions
Cell lysates
Secretions
Cell lysates
ISG15
Galectin-1
0– + + + +
0 0.01 0.025 0.05
Fig. 4 Western blot analysis of the effects of brefeldin A (BFA) on
b2MG, oMx1, ISG15 and Gal-1 levels in secretions from oGE cells at
12 hr after IFNs treatment. Concentrations of BFA used in this study
were 0, 0.01, 0.025, and 0.05 lg/mL. Secretions and cell lysates within
each treatment were derived from the same dish. Each lane shows
representative secretions and cell lysates from the same sample.
TOYOKAWA ET AL.
American Journal of Reproductive Immunology 57 (2007) 13–22 ª 2007 The Authors
18 Journal compilation ª 2007 Blackwell Munksgaard
confirmed that observation by demonstrating that
oMx1 is secreted by an ovine uterine epithelial-
derived cell line in vitro.39 In the present study, we
used a well-characterized inhibitor of the classical
(ER–Golgi–TGN) secretory pathway, BFA, to deter-
mine if oMx1 was secreted via the classical secretory
pathway. Results clearly showed that secretion of
oMx1, ISG15, and Gal-1 were not inhibited by BFA.
This is consistent with their lack of a classical secre-
tion signal sequence at their amino termini and their
cellular localization. These results strongly suggest
that all three are secreted via unconventional secre-
tory pathways.
The fact that levels of oMx1 protein in uterine flu-
shes increased between days 15 and 17 of early
pregnancy suggests that it may be a bona fide secre-
tory protein. This is further supported by the fact
that oGE cells secreted oMx1 in vitro and that levels
increased in response to IFNs. However, it remains a
possibility that levels of oMx1 increased in uterine
flushes as a result of lysis of epithelial cells during
the process of forming the synepitheliochorial pla-
centa that is characteristic of domestic ruminants.40
It is also possible that immune cells could be present
in the uterine lumen and have contributed to
increased levels of oMx1 in uterine flushes; how-
ever, there are no reports on the presence of free
immune cells in uterine flushes during this period of
early pregnancy. Furthermore, we have not been
able to detect oMx1 mRNA in conceptus tissue (S.J.
Carling, unpublished observations). Therefore, the
simplest explanation of the present results is that the
uterine epithelium in vivo, like we showed in vitro, is
secreting oMx1 into the uterine lumen. The pres-
ence of oMx1 protein in endometrial secretions begs
the question of whether it affects conceptus growth
and development during the peri-implantation per-
iod.
We further wanted to address the mechanism of
oMx1 secretion by determining whether oMx1 was
secreted via ‘classical’ or ‘unconventional’ pathways.
Proteins secreted via unconventional pathways have
a number of features in common. First of all, uncon-
ventionally secreted proteins tend to be localized
throughout the cytoplasm.28 Our immunofluores-
cence analysis demonstrated that upon IFNs treat-
ment oMx1 was distributed diffusely throughout the
cytoplasm. Among other Mx proteins, human
MxA,41,42 rat Mx2 and Mx343 and canine Mx1 and
Mx244 are also localized in the cytoplasm. Secondly,
unconventionally secreted proteins lack a hydropho-
bic signal peptide.23 Based on predictions (SignalP
3.0), oMx1, ISG15 and Gal-1 do not contain a classi-
cal secretion signal peptide, whereas b2MG does (see
Fig. 3). In agreement with this analysis, secretion of
b2MG is known to occur in an ER/Golgi/TGN-
dependent fashion,22 whereas ISG15 is secreted via
an unconventional secretory pathway.28 Finally,
inhibitors of the ER/Golgi/TGN-dependent secretion
such as BFA do not block unconventional secretory
processes.28 We observed in the present study that
BFA treatment resulted in higher secretion levels of
two IFNs-stimulated proteins, oMx1 and ISG15
without affecting intracellular levels of these
700
(a) (b)
(c) (d)
700
500
300
806040200
600
500
400
300
200
100
180070
6050
6420
1000
400300200100M
ean
leve
ls o
f IS
G15
prot
ein
(CN
T*m
m2)
Mea
n le
vels
of B
2MG
prot
ein
(CN
T*m
m2)
Mea
n le
vels
of o
Mx1
prot
ein
(CN
T*m
m2)
Mea
n le
vels
of G
al-1
prot
ein
(CN
T*m
m2)
0
0NT IFN NT IFNIFN+BFA0.05 IFN+BFA0.05
NT IFN IFN+BFA0.05 NT IFN IFN+BFA0.05
Fig. 5 Quantitation of effects of brefeldin A
(BFA) on b2MG, (A) oMx1 (B), ISG15 (C) and
Gal-1 (D) levels in secretions from oGE cells.
Each experiment was replicated three inde-
pendent times with three replicate wells per
treatment within each experiment (n ¼ 9 for
b2MG, oMx1, ISG15, n ¼ 6 for Gal-1). NT, no
treatment; IFN, interferon tau alone
(10,000 AVU/mL); and IFN + BFA0.05, BFA at
0.05 lg/mL. Levels of b2MG in conditioned
culture media treated with IFN + 0.05 lg/ml
BFA were reduced compared with NT (� 50%
reduction; P < 0.01). In contrast, levels of
oMx1, ISG15, and Gal-1 in conditioned media
were increased (6.4-fold, 14-fold and 16-fold
increase, respectively; P < 0.01) in IFN + BFA
compared with NT, but were not different
between NT and IFN.
UTERINE SECRETION OF THE MYXOVIRUS RESISTANCE PROTEIN MX1
American Journal of Reproductive Immunology 57 (2007) 13–22 ª 2007 The Authors
Journal compilation ª 2007 Blackwell Munksgaard 19
proteins. This effect of BFA to decrease secretion via
classical secretory pathways while increasing secre-
tion of unconventionally secreted proteins has been
reported previously.45 The fact that b2MG levels in
secretions were reduced argues against this resulting
from toxic effects of IFN s or BFA. In fact, we
(unpublished data) and others37 have used 10-fold
higher concentrations of IFNs in this cell line with-
out any evidence of toxicity. In addition, our tran-
sient oMx1 knockout model also demonstrated that
levels of b actin in cell lysates were similar in the
absence and presence of IFNs at the same concentra-
tion utilized in these experiments.39
Galectin-1 is not thought to be stimulated by IFNsand it also increased in secretions in response to
BFA treatment. This is particularly interesting in
light of preliminary results from Yeast 2-hybrid
screens in our laboratory that identified Gal-1 as a
binding partner for oMx1 (J. Welker and T.L. Ott,
unpublished data). Thus, we speculate that oMx1
and Gal-1 may interact in vivo in the uterus although
this interaction remains to be confirmed in vivo.
In the experiments examining the effects of BFA
on secretion, IFNs did not increase levels of oMx1,
and ISG15 in secretions at 12 hr after treatment
(Fig. 4). We chose 12 hr to minimize any potential
toxic effects of BFA on the oGE cells. While levels of
these proteins in the secretions were apparently
increasing they did not achieve significance in the
short 12 hr incubation period. Whereas, it is clear
from the Western blot (Fig. 4) that oMx1 and ISG15
were beginning to accumulate in conditioned media
in response to IFNs, and oMx1 levels were
approaching significance, longer incubation times
are need to detect these effects. In contrast, all three
unconventionally secreted proteins increased at
12 hr in conditioned culture media in response to
IFN + BFA treatment. This indicates that BFA stimu-
lated or enhanced unconventional secretory path-
way(s). Other reports showed a similar phenomenon
for IL-b secretion in the NIH3T3 cell line.45 The
mechanism(s) whereby BFA increases secretion of
unconventionally secreted proteins is not known.
Conclusion
Results from this study provide support for the
hypothesis that oMx1 is a secretory protein in
the ovine uterus and a component of endometrial
secretions in ewes. Furthermore, our results sup-
port the hypothesis that oMx1 is secreted via an
unconventional secretory pathway. Results from
transient oMx1 knockdown using siRNA suggest that
oMx1 may regulate secretion of other unconven-
tionally secreted proteins.39
Acknowledgements
This study was supported in part by USDA NRICGP
grant 2002-02398 and NIH NCRR grant P20-
RR15587-01 to T.L.O.
References
1 Horisberger MA, Gunst MC: Interferon-induced
proteins: identification of Mx proteins in various
mammalian species. Virology 1991; 180:185–190.
2 Horisberger MA, Staeheli P, Haller O: Interferon
induces a unique protein in mouse cells bearing a
gene for resistance to influenza virus. Proc Natl Acad
Sci U S A 1983; 80:1910–1914.
3 Meier E, Kunz G, Haller O, Arnheiter H: Activity of
rat Mx proteins against a Rhabdovirus. J Virol 1990;
64:6263–6269.
4 King CM, Raposo G, Lemmon AM: Inhibition of
nuclear import and cell-cycle progression by mutated
forms of the dynamin-like GTPase MxB. Proc Natl
Acad Sci U S A 2004; 24:8957–8962.
5 Charleston B, Stewart HJ: An interferon-induced Mx
protein: cDNA sequence and high-level expression in
the endometrium of pregnant sheep. Gene 1993;
137:327–331.
6 Ott TL, Yin J, Wiley AA, Kim HT, Gerami-Naini B,
Spencer TE, Bartol FF, Burghardt RC, Bazer FW:
Effects of the estrous cycle and early pregnancy on
uterine expression of Mx protein in sheep (Ovis
aries). Biol Reprod 1998; 59:784–794.
7 Hicks BA, Etter SJ, Carnahan KG, Joyce MM, Assiri
AA, Carling SJ, Kodali K, Johnson GA, Hansen TR,
Mirando MA, Woods GL, Vanderwall DK, Ott TL:
Expression of the uterine Mx protein in cyclic and
pregnant cows, gilts, and mares. J Anim Sci 2003;
81:1552–1561.
8 Johnson GA, Joyce MM, Yankey SJ, Hansen TR,
Ott TL: The Interferon Stimulated Genes (ISG) 17 and
Mx have different temporal and spatial expression in
the ovine uterus suggesting more complex regulation
of the Mx gene. J Endocrinol 2002; 174:R7–R11.
9 Yankey SJ, Hicks BA, Carnahan KG, Assiri AM, Sinor
SJ, Kodali K, Stellflug JN, Ott TL: Expression of the
antiviral protein Mx in peripheral blood mononuclear
cells of pregnant and bred, non-pregnant ewes.
J Endocrinol 2001; 170:R7–R11.
TOYOKAWA ET AL.
American Journal of Reproductive Immunology 57 (2007) 13–22 ª 2007 The Authors
20 Journal compilation ª 2007 Blackwell Munksgaard
10 Gifford CA, Racicot K, Clark DS, Austin KJ, Hansen
TR, Lucy MC, Davies CJ, Ott TL: Regulation of
interferon stimulated genes in peripheral blood
leukocytes in pregnant and bred, non-pregnant dairy
cows. J Dairy Sci 2006 (in press)
11 Roberts RM, Liu L, Alexenko A: New and atypical
families of type I interferons in mammals:
comparative functions, structures, and evolutionary
relationship. Prog Nucleic Acid Res Mol Biol 1997;
56:287–325.
12 Stewart DM, Johnson GA, Vyhlidal CA, Burghardt
RC, Safe SH, Yu-Lee L-Y, Bazer FW, Spencer TE:
Interferon tau activates multiple STAT proteins and
has complex effects on interferon-responsive gene
transcription in ovine endometrial epithelial cells.
Endocrinology 2001; 142:98–107.
13 Spencer TE, Ott TL, Bazer FW: Expression of
interferon regulatory factors one (IRF-1) and two
(IRF-2) in the ovine endometrium: effects of
pregnancy and interferon tau. Biol Reprod 1998;
58:1154–1162.
14 Vallet JL, Barker PJ, Lamming GE, Skinner N,
Huskisson NS: A low molecular weight endometrial
secretory protein which is increased by ovine
trophoblast protein-1 is a b2-Microglobulin-like
protein. J Endocrinol 1991; 130:R1–R4.
15 Austin KJ, Carr AL, Pru JK, Hearne CE, George EL,
Belden EL, Hansen TR: Localization of ISG15 and
conjugated proteins in bovine endometrium using
immunohistochemistry and electron microscopy.
Endocrinology 2004; 145:967–975.
16 Teixeira MG, Austin KJ, Perry DJ, Dooley VD,
Johnson GA, Francis BR, Hansen TR: Bovine
granulocyte chemotactic protein-2 is secreted by the
endometrium in response to interferon-tau (IFN-tau).
Endocrine 1997; 6:31–37.
17 Mirando MA, Short EC Jr, Geisert RD, Vallet JL,
Bazer FW: Stimulation of 2¢ 5¢-oligoadenylate
synthetase activity in sheep endometrium during
pregnancy, by intrauterine infusion of ovine
trophoblast protein-1, and by intramascular injection
of recombinant bovine interferon-a1. J Reprod Fertil
1991; 93:599–607.
18 Bazer FW, Spencer TE: Methods for studying
interferon tau stimulated genes. Methods Mol Med
2006; 122:367–380.
19 Vallet JL, Barker PJ, Lamming GE, Skinner N,
Huskisson NS: A low molecular weight endometrial
secretory protein which is increased by ovine
trophoblast protein-1 is a b2-microglobulin-like
protein. J Endocrinol 1991; 130:R1–R4.
20 D’Cunha J, Ramanujam S, Wagner RJ, Witt PL,
Knight E Jr, Borden EC: In vitro and in vivo secretion
of human ISG15, an IFN-induced immunomodulatory
cytokine. J Immunol 1996; 157:4100–4108.
21 Dargemont C, Dunon D, Salamero J, Deugnier MA,
Davoust J, Thiery JP: Overproduction and secretion of
beta 2-microglobulin by a rat thymic epithelial cell
line that expresses MHC class I heavy chain. J Cell Sci
1991; 98:559–565.
22 Palade G: Intracellular aspects of the process of
protein synthesis. Science 1975; 189:347–358.
23 Von Hejine G: Signal sequences: the limits of
variations. J Mol Biol 1985; 184:99.
24 Blobel G, Dobberstein B: Transfer of proteins across
membranes. I. Presence of proteolytically processed
and unprocessed nascent immunoglobulin light
chains on membrane-bound ribosomes of murine
myeloma. J Cell Biol 1975; 67:835–851.
25 Mignatti P, Morimoto T, Rifkin DB: Basic fibloblast
growth factor, a protein devoid of secretory signal
sequence, is released by cells via a pathway
independent of the endoplasmic reticulumn-Goligi
complex. J Cell Physiol 1992; 15:81–93.
26 Wilson TJ, Firth MN, Powell JT, Harrison FL: The
sequence of the mouse 14 kDa b-galactosidase-
binding lectin and evidence for its synthesis on free
cytoplasmic ribosomes. Biochem J 1989; 261:847–852.
27 Kuchler K, Thorner J: Secretion of peptides and
proteins lacking hydrophobic signal sequences: The
role of adenosine triphosphate-driven membrane
translocators. Endocr Rev 1992; 3; 499–514.
28 Nickel W: Unconventional secretory routes. Direct
protein export across the plasma membrane of
mammalian cells. Traffic 2005; 6:607–614.
29 Misumi Y, Misumi Y, Miki K, Takatsuki A, Tamura G,
Ikehara Y: Novel blockade by brefeldin A of
intracellular transport of secretory proteins in
cultured rat hepatocytes. J Biol Chem 1986;
261:11398–11403.
30 Tartakoff AM: Perturbation of vesicular traffic with
the carboxylic ionophore monensin. Cell 1983;
32:1026–1028.
31 Gray CA, Burghardt RC, Johnson GA, Bartol FF,
Tarleton BJ, Wiley AA, Bazer FW, Spencer TE:
Developmental biology of uterine glands. Biol Reprod
2002; 65:1311–1323.
32 Praefcke GJ, McMahon HT: The dynamin superfamily:
universal membrane tubulation and fission molecules?
Nat Rev Mol Cell Biol 2004; 5:133–147.
33 Shupliakov O, Low P, Grabs D, Gad H, Chen H, David
C, Takei K, De Camilli P, Brodin L: Synaptic vesicle
endocytosis impaired by disruption of dynamin-SH3
domain interactions. Science 1997; 276:259–263.
34 Thompson HM, Skop AR, Euteneuer U, Meyer BJ,
McNiven MA: The large GTPase dynamin associates
UTERINE SECRETION OF THE MYXOVIRUS RESISTANCE PROTEIN MX1
American Journal of Reproductive Immunology 57 (2007) 13–22 ª 2007 The Authors
Journal compilation ª 2007 Blackwell Munksgaard 21
with the spindle midzone and is required for
cytokinesis. Curr Biol 2002; 12:2111–2117.
35 Accola MA, Huang B, Masri AA, McNiven MA: The
antiviral dynamin family member, MxA, tubulates
lipids and localizes to the smooth endoplasmic
reticulum. J Biol Chem 2002; 24:21829–21835.
36 Johnson GA, Joyce MM, Yankey SJ, Hansen TR, Ott
TL: The interferon stimulated genes (ISG) 17 and Mx
have different temporal and spatial expression in the
ovine uterus suggesting more complex regulation of
the Mx gene. J Endocrinol 2002; 174:R7–R11.
37 Johnson GA, Burghardt RC, Newton GR, Bazer FW,
Spencer TE: Development and characterization of
immortalized ovine endometrial cell lines. Biol Reprod
1999; 61:1324–1330.
38 Bendtsen JD, Nielsen H, von Heijine G, Brunak S:
Improved prediction of signal peptides: SignalP 3.0.
J Mol Biol 2004; 340:785–795.
39 Toyokawa K, Leite F, Ott TL: Cellular localization and
function of the antiviral protein, ovine Mx1 (oMx1):
II. The oMx1 protein is a regulation of secretion in an
ovine glandular epithelial cell line. Am J Reprod
Immunol 2007; 57:23–33.
40 Wooding FB, Hobbs T, Morgan G, Heap RB, Flint AP:
Cellular dynamics of growth in sheep and goat
synepitheliochorial placentomes: an autoradiographic
study. J Reprod Fertil 1993; 98:275–283.
41 Zurcher T, Pavlovic J, Staeheli P: Mechanism of human
MxA protein action: variants with changed antiviral
properties. EMBO Journal 1992; 11:1657–1661.
42 Ponten A, Sick C, Weeber M, Haller O, Kochs G:
Dominant-negative mutants of human MxA protein:
domains in the carboxy-terminal moiety are
important for oligomerization and antiviral activity.
J Virol 1997; 71:2591–2599.
43 Johannes L, Arnheiter H, Meier E: Switch in antiviral
specificity of a GTPase upon translocation from the
cytoplasm to the nucleus. J Virol 1993; 67:1653–1657.
44 Nakamura T, Asano A, Okano S, Ko JH, Kon Y,
Watanabe T, Agui T: Intracellular localization and
antiviral property of canine Mx proteins. J Interferon
Cytokine Res 2005; 25:169–173.
45 Wingren AG, Bjorkdahl O, Labuda T, Bjork L,
Andersson U, Gullberg U, Hedlund G, Sjogren HO,
Kalland T, Widegren B, Dohlsten M: Fusion of a
signal sequence to the interleukin-1 beta Gene directs
the protein from cytoplasmic accumulation to
extracellular release. Cell Immunol 1996; 169:226–237.
TOYOKAWA ET AL.
American Journal of Reproductive Immunology 57 (2007) 13–22 ª 2007 The Authors
22 Journal compilation ª 2007 Blackwell Munksgaard