| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
ARTICLES |
Institute of Reproduction and Development (L.M.F., A.E.O.’C., D.J.P., D.M.d.K.) and the Departments of Physiology (G.J.) and Biochemistry and Molecular Biology (M.T.W.H.), Monash University, Clayton, Victoria 3168; and the Department of Physiology, University of New England (J.R.M.), Armidale, New South Wales 2351, Australia
Address all correspondence and requests for reprints to: Dr. David M. de Kretser, Institute of Reproduction and Development, Monash Medical Center, 246 Clayton Road, Clayton, Victoria 3168, Australia. E-mail: david.de.kretser{at}med.monash.edu.au
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
2-macroglobulin, UTMP was unable to neutralize the
bioactivity of activin in a bioassay based on the capacity of activin
to inhibit the proliferation of an MPC-11 plasmacytoma cell line. The
high concentrations of this protein in uterine fluid during pregnancy
and its ability to bind activin suggest that UTMP may act as a low
affinity, high capacity binding protein to sequester activin in the
local uterine environment. |
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Recent data indicate that some of these bioactive variants circulate in serum (9, 10, 11), raising the question of how their pluripotent actions are limited to specific sites. In part this is achieved by the presence of a specific binding protein, follistatin (FS) (12, 13, 14), which can neutralize the majority of the actions of activin A. The possibility of other activin-binding proteins with the capacity to regulate the actions of these pluripotent molecules has not been extensively explored.
Activin A is produced by the placenta and fetal amnion and chorion preparations (15, 16, 17) and has recently been suggested to have a role in the initiation of parturition in humans (18). Characterization methods based on one- and two-dimensional SDS-PAGE (19, 20, 21) have shown that both bovine and ovine amnion and allantois produce a large variety of proteins, including other members of the transforming growth factor-ß family (22).
In contrast to primates, the allantoic sac in the sheep is a significant compartment during pregnancy. We have previously shown that ovine allantoic fluid is a rich source of immunoactive and bioactive activin A (23), with yields far exceeding those obtained from amniotic fluid (24). Using a newly developed RIA (9) for activin A that removes interference from FS, together with a rat pituitary cell bioassay (25), we found changes in the ratio of immunoactive to bioactive activin during its purification from allantoic fluid, suggesting the presence of activin-binding proteins, other than FS, in allantoic fluid. The following study details the isolation and characterization of one of these proteins together with binding studies detailing its interaction with activin A.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
These investigations were approved by the Monash University standing committee on ethics in animal experimentation and conform to the National Health and Medical Research Council/Commonwealth Scientific and Industrial Research Organization/Australian Agricultural Council Code of Practice for the Care and Use of Animals for Experimental Purposes.
Fractionation
The initial two steps of the purification protocol are based on
those used for the purification of activin A from ovine amniotic fluid
(24). Briefly, 750 ml of the allantoic fluid pool, adjusted to pH 6.0
with 0.5 M phosphate buffer, were mixed with 300 ml yellow
dye HE4R (ICI, Melbourne, Australia) coupled to Fractogel TSK HW-65F
(Merck & Co., Inc., Darmstadt, Germany) in 0.05
M phosphate buffer, pH 6.0, and incubated for 2.5 h in
a shaking water bath before filtering through a sintered glass funnel.
The gel was incubated with 1.0 liter 0.05 M phosphate
buffer, pH 6.0, and again filtered through a sintered glass funnel. The
gel was then mixed with 500 ml 0.4 M potassium chloride
(KCl)/0.05 M phosphate buffer, pH 6.0, returned to the
column (5 x 25 cm), and eluted at a flow rate of 8 ml/min. A
final high salt wash of 250 ml 1 M KCl/4 M
urea/0.05 M phosphate buffer, pH 7.0, was incubated with
the gel overnight at room temperature, then eluted (UA) the following
day at a flow rate of 8 ml/min, followed by a further elution (UB) of
500 ml with the same buffer. The UB dye eluate fraction was diluted 1/8
(vol/vol) with 2.3 M KCl in 0.05 M phosphate
buffer, pH 7.0, to produce a solution containing 2 M
KCl/0.5 M urea. This solution (3.9 liters) was loaded onto
a 75 ml phenyl-Sepharose (Pharmacia Biotech, Uppsala,
Sweden) hydrophobic interaction column (3.5 x 25 cm) overnight at
room temperature at 4 ml/min. The column was then washed with 150 ml 2
M KCl/0.05 M phosphate buffer, pH 7.0, followed
by 150 ml 0.05 M phosphate buffer, pH 7.0, at 8 ml/min
(wash PS P1). The column was incubated overnight with 3 bed vol 25%
acetonitrile (ACN) in 0.05 M phosphate buffer, pH 7.0, and
the activin-containing fraction was eluted at 8 ml/min.
All fractions from these two purification steps were subjected to SDS-PAGE analysis followed by ligand blot analysis, as detailed below, to determine which fractions contained activin-binding proteins. The most prominent of these occurred in the phosphate buffer wash (PS P1) before elution of bound activin A off the phenyl-Sepharose column, and this protein was targeted for isolation.
The PS P1 fraction was initially subjected to reverse phase HPLC (RP-HPLC) using a 0.1% H3PO4/0–50% ACN linear gradient over 100 min with a C3 semipreparative column (Beckman Coulter, Inc., Berkeley, CA) at a flow rate of 3.0 ml/min, and eluted fractions across the gradient were screened for activin binding capacity on ligand slot blots (see below). Selected peaks were rechromatographed by RP-HPLC using a 0.1% trifluoroacetic acid/0–50% ACN linear gradient over 100 min with a C8 Ultrapore analytical column (Beckman Coulter, Inc., Palo Alto, CA) at a flow rate of 1.0 ml/min, and the activin binding capacity of fractions was again confirmed by ligand blot. A pool of fractions was made based on both activin binding capacity and SDS-PAGE analysis with silver-stained bands. These fractions were further separated by Superose 12HR 10/30 (Pharmacia Biotech) gel filtration in 50 mM PBS, pH 7.4, at a flow rate of 1.0 ml/min. The column had previously been calibrated against known molecular weight markers.
SDS-PAGE
Fractions obtained from the last RP-HPLC step and from the
Superose 12HR 10/30 fractionation were run on 12.5% SDS-PAGE gels (26)
under nonreducing conditions, and protein bands were identified by
silver staining (27).
Activin assays
The ability of uterine milk protein (UTMP) to bind to activin
and interfere in its measurement was assessed using the three activin
assays available in our laboratory.
1) A disequilibrium RIA was previously described by Robertson et al. (28) and was modified by McFarlane et al. (9) by the incorporation of dissociating reagents to remove the interference of FS in the assay. Increasing doses of UTMP of 39–2500 ng/ml were assayed alone and in the presence of a known dose of human recombinant activin A (10 ng/ml). The same dose of human recombinant activin was assayed in the presence of 25 ng/ml of the known activin-binding protein, FS (hrFS288), and all samples were assayed in both the presence and the absence of the dissociating reagents. The standard curves were constructed using the same preparation of human recombinant activin A. The intraassay coefficient of variation was 10.90 ± 0.42% (n = 2).
2) An in vitro bioassay for activin, described previously (29), uses MPC-11 plasmacytoma cells cultured in 96-well plates at a density of 1000 cells/well. In this bioassay, addition of activin A causes a dose-dependent inhibition of proliferation in the MPC-11 cell line. Briefly, cells were cultured for 48 h in the presence of increasing amounts of human recombinant activin A (0.02–20 ng/ml), UTMP (2.44–2500 ng/ml), or the activin-binding protein, FS (5–100 ng/ml). Cells were also cultured with the same doses of UTMP and FS in the presence of 5 ng/ml human recombinant activin. After this culture period, [3H]thymidine was added for a further 24 h, and thymidine incorporation was measured using standard scintillation counting methodologies. This bioassay has a sensitivity of 0.4 ng/ml activin A, an ED50 response of 3.5 ng/ml, and an intraassay coefficient of variation of less than 11%.
3) A specific activin A two-site enzyme-linked immunosorbent assay (30), which incorporates an analyte denaturation and an oxidation step, was used to determine activin A activity in the purified UTMP (doses from 3.91–125 ng/ml) with and without the presence of activin (0.31 ng/ml). Human recombinant activin A was used as a standard in the assay. The sensitivity of the assay was 0.01 ng/ml, and the intraplate coefficient of variation was 9.42%.
Iodination
For RIA, ligand blot and slot blot applications, human
recombinant activin A (2 µg) was iodinated with Iodogen (Pierce Chemical Co., Rockford, IL) using the method described by
McFarlane et al. (31). For the binding studies described
below, iodination conditions were modified (i.e. a shorter
incubation time with Iodogen) to produce an activin tracer with a lower
specific activity. The tracer was further purified on a Sephadex G-75
(Pharmacia Biotech) gel filtration column (2.5 x 90
cm) using 0.01 M PBS (pH 7.4), 0.01% sodium azide, and
0.2% BSA before use.
Ligand blotting and slot blotting
For ligand blots of purification fractions, either 10% or
12.5% SDS-PAGE nonreducing gels were run [using bovine FS (bFS) in
control lanes], and separated proteins were transferred overnight to a
0.45-µm pore size nitro-cellulose membrane (Micron Separations, Inc.,
Westboro, MA) at 60 V, with cooling, in transfer buffer containing 20
mM Tris base, 150 mM glycine, and 20%
(vol/vol) methanol. At the conclusion of the transfer, the membranes
were stained with Ponceau S stain (Sigma Chemical Co., St.
Louis, MO) to identify molecular weight marker proteins and were then
blocked for 1 h with 3% nonfat powdered milk in 20 mM
Tris-HCl (pH 7.2), 2 mM EDTA, 154 mM NaCl, 5
mM benzamidine hydrochloride, 1% Triton X-100, and 0.2%
BSA (Tris ligand buffer). The membrane was rinsed with Tris ligand
buffer and incubated with [125I]activin tracer (100,000
cpm/ml in Tris ligand buffer for 20 h), followed by three 5-min
rinses with Tris ligand buffer. The dried membrane was then exposed to
Bio-Max (Eastman Kodak Co., Rochester, NY) film for 4 days
before development of the film.
For scanning of fractions across RP-HPLC profiles for activin binding capacity, 50 µl from each 1.5-ml fraction were applied to a 0.45-µm pore size nitro-cellulose membrane using a slot blot filtration manifold (PR648, Hoefer, San Francisco, CA). Each sample slot on the membrane was washed three times with 1 ml Tris ligand buffer while in the manifold, then the membrane was removed from the manifold and blocked with 3% nonfat powdered milk, incubated with [125I]activin tracer, and exposed to BioMax film in the same way as for ligand blots. RP-HPLC fractions were pooled for further fractionation based on their capacity to bind [125I]activin tracer.
NH2-terminal amino acid sequencing
One fraction from the final RP-HPLC profile was run, together
with known molecular mass markers, on a 12.5% SDS-PAGE nonreducing
gel, and separated proteins were transferred for 80 min at 190 mA in a
SemiPhor Transfer Unit (TE70, Hoefer) to a polyvinylidene
difluoride Immobilon PSQ (Millipore Corp.,
Bedford, MA) membrane. At the conclusion of the transfer, the membrane
was stained with Coomassie blue R-250 stain (Sigma Chemical Co.) for 30 sec, then destained and dried. The protein band of
interest was excised from the membrane and applied to an on-line
PE Applied Biosystems (Foster City, CA) 470A gas phase
sequencer equipped with an on-line PE Applied Biosystems
120A HPLC for analysis of the phenylthiohydantoin-derived amino
acids.
Binding studies
The ability of the isolated and sequenced protein to bind
activin was further investigated in a series of gel filtration
experiments where [125I]activin tracer (50,000–100,000
cpm) was fractionated on a Superose 12HR 10/30 column in 50
mM PBS (pH 7.4), 0.01% sodium azide, and 0.1% BSA under
the following conditions: 1) alone, 2) after preincubation with 1.5
µg bFS (heterologous pool of 35-, 39-, and 45-kDa forms) overnight at
room temperature, 3) after preincubation with 250 µl (estimated 50
µg) of the activin-binding protein pool (peak fractions from final
Superose 12 HR 10/30 fractionation) overnight at room temperature, and
4) after preincubation of 2 µg cold human recombinant activin A with
250 µl (estimated 50 µg) of the activin-binding protein pool for
1.5 h at room temperature followed by incubation with tracer
overnight at room temperature. The column flow rate was 0.75 ml/min,
and fractions of 0.5 min (0.375 ml) were collected and counted in a
-counter (1261 MultiGamma, LKB-Wallac, Inc., Turku, Finland). The molecular sizes of radioactive peaks
were determined by comparison with elution volumes of known mol wt
calibration standards.
Binding characteristics
Binding interactions between activin and UTMP and between
activin and FS were carried out in real-time using surface plasmon
resonance instrumentation (BIAcore, Pharmacia Biotech)
(32). Briefly, human recombinant activin A (35 µl, 13.3 µg/ml in 5
mM maleate buffer, pH 6) was immobilized onto a CM5 sensor
chip (lot 0328, Pharmacia Biotech) using the
amine-coupling method as previously described (33). All binding
interactions were carried out using HBS buffer (10 mM
HEPES, 150 mM NaCl, and 0.05% surfactant P20, pH 7.4),
with UTMP and FS injections made using HBS buffer containing 0.05%
BSA. Interactions between activin and bFS were carried out at a flow
rate of 10 µl/min and at a range of FS concentrations between
0.25–10 nM (six determinations). Interactions between
activin and UTMP were performed at a flow rate of 3 µl/min and a
range of UTMP concentrations between 1–128 nM (five
determinations). Between runs, the sensor chip surface was regenerated
using 0.25 M acetic acid alone or with 0.25 M
acetic acid and 1 M NaCl. Dissociation binding constants
were determined using BIAevaluation software, whereby both associative
and dissociative phases were analyzed, and the equilibrium dissociation
constant (Kd) was derived over the range of concentrations
tested. To confirm the binding characteristics derived from these
procedures, predictive association and dissociation curves were
generated using BIAsimulation software using the experimental
Kd values, flow rates, and injection times for both FS and
UTMP.
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
, revealing peak activin-binding
activity in fraction 84, which corresponded to the maximum UV-absorbing
peak eluted from the gradient (Fig. 1B). After further fractionation on
Superose 12HR 10/30, silver staining of a 12.5% nonreducing SDS-PAGE
gel (Fig. 2) revealed a single protein
band of approximately 55 kDa. Sequence analysis of fraction 82 from the
final RP-HPLC revealed a single polypeptide NH2-terminal
sequence of EKQQHXQQH. This sequence corresponds to part of the known
sequence of ovine UTMP (34).
|
|
, A and B,
represents the results obtained in the activin bioassay; similar
results were obtained in the RIA and enzyme-linked immunosorbent assay
(data not shown). By comparison, in the same activin RIA and as
demonstrated previously (9), FS interfered in the binding of activin A
in the absence of dissociating reagents, but this interference was
removed in the presence of dissociating reagents (data not shown).
Similarly, in the bioassay and as demonstrated previously (29), the
ability of activin to inhibit the proliferation of the MPC-11 cells was
counteracted by the addition of FS to activin (Fig. 3B).
|
) revealed a peak with an apparent
molecular mass of 12.5 kDa, but SDS-PAGE of the same
[125I]human recombinant activin tracer used in this study
(data not shown) has shown it to be the 25-kDa
ßA-ßA dimer. Similar elution behavior with
radiolabeled activin was obtained by Schneyer et al. (14)
and with unlabeled activin dimer by Kogawa et al. (35) on
Superose 12HR columns. Using bFS as a known activin binding protein
that has been shown to bind in an activin/FS ratio of 1:2 (36), we have
demonstrated that bFS binds to [125I]human recombinant
activin (Fig. 4B), forming a complex of approximately 116 kDa, which we
postulate to be FS-ßA-ßA-FS. Binding of
UTMP to [125I]human recombinant activin (Fig. 4C) results
in the formation of a complex of approximately 76 kDa, which suggests
an UTMP/activin ratio of 1:1 (UTMP-ßA-ßA).
However, formation of the UTMP-activin complex was incomplete
(
50%), despite the addition of an apparent excess of UTMP
(estimated 50 µg) and, further, was blocked by preincubation with an
excess of cold human recombinant activin A (2 µg; Fig. 4D).
|
). Preincubating either UTMP or FS with
excess activin before injection resulted in minimal binding to activin
immobilized on the sensor chip (data not shown).
|
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
UTMPs are members of the serine protease inhibitor (serpin) superfamily of proteins (34), which includes a large number of protease inhibitors and at least two hormone-binding proteins. It is not surprising that our purification protocol has resulted in the isolation of UTMP from allantoic fluid, as UTMPs are secreted from the endometrium of the uterus under the influence of progesterone, and large quantities have previously been purified from fluid in late pregnancy from the contralateral uterine horn of unilaterally pregnant ewes (40). UTMPs can also be secreted from the uterus of nonpregnant sheep after progesterone admini-stration.
Known actions of UTMPs include inhibition of mitogen-induced lymphocyte proliferation and mixed lymphocyte cultures, an action that is not blocked by a neutralizing antibody to transforming growth factor-ß (41). Further, in vivo evidence of the action of these proteins as important regulators of uterine immune function has been presented by Liu and Hansen (42), who demonstrated that UTMPs blocked poly(I)·poly(C)-induced natural killer cell activity and abortion in mice, and by Skopets et al. (43), who demonstrated that UTMPs could inhibit antibody titers to the T cell-dependent antigen, ovalbumin, another member of the serpin superfamily. Skopets et al. (43) also postulated that UTMPs could bind other cytokines due to the high concentrations required for inhibition of lymphocyte proliferation.
Although UTMPs have been classified as members of the serpin superfamily, no known antiprotease activity has yet been shown for these proteins, but the ability to bind activin, which we have demonstrated, suggests a role for UTMP in gestation as a carrier serpin. Other possible roles for UTMPs that have been postulated include direct nutrition of the conceptus, transport of limiting nutrients to the embryo, osmotic support of the feto-placental unit, growth control, inhibition of proteolytic activities released by the embryo or mother, and immunosuppression of the local maternal immune system (44).
Interestingly, our preliminary molecular modeling studies have
suggested that the so-called spur region of UTMP associated with the
canonical serpin motif is also present in domain 2 of FS, suggesting
the potential for a similar binding site. However, our studies have
demonstrated that the binding affinity of UTMP for activin is
considerably lower than that noted for FS, but higher than the
published affinities for another activin-binding protein,
2-macroglobulin (2M) (45). Our results
for the binding affinity of FS (379 ± 51 pM)
correlate well with previously published Kd values for
activin bound to the six isoforms of porcine FS (540–690
pM) (46), rat ovarian FS [590 ± 230 pM
(47) and 130 ± 70 pM (48)], bovine pituitary FS (910
pM) (35), human serum FS (198 ± 69 pM)
(13), and Xenopus FS (159 pM) (49). Similarly,
differences in amounts of activin-binding proteins required to form
complexes in the binding studies (1.5 µg FS compared with 50 µg
UTMP required to complex 100,000 cpm iodinated activin) indicate the
lower binding affinity of UTMP for activin than that of FS. This
conclusion is reinforced by the finding that activin-UTMP complex
formation was incomplete and the formation of this complex was blocked
by preincubation with 2 µg cold activin. In contrast and as
previously reported (13), the activin-FS complex formation is largely
irreversible, as demonstrated by the shallow dissociation curve shown
in Fig. 5A
. The lower affinity of UTMP for activin compared with FS
suggests that, unlike that protein, UTMP is unlikely to neutralize the
bioactivity of activin. This prediction was borne out by the failure of
UTMP to neutralize the bioactivity of activin in our recently developed
activin bioassay based on the capacity of activin to inhibit
proliferation of a plasmacytoma cell line. Although the action of UTMP
has not been assessed in other bioassays for activin, as has FS, we
would expect UTMP to behave similarly in all bioassays. However, we
cannot rule out specific roles of UTMP in other cell systems.
Alternatively, this failure of UTMP to neutralize the bioactivity of
activin may be the result of the UTMP binding to a different site of
the activin molecule than that to which its receptor binds.
Interestingly, studies of activin-FS complexes in biological fluids (14) yielded complexes of 60–70 kDa in human serum (matching the size of activin-hrFS315 complexes) and of 200–300 kDa in follicular fluid (matching the size of activin-hrFS288 complexes). Therefore, to distinguish activin-FS complexes from activin-UTMP complexes in biological fluids such as serum, methods other than gel filtration on Superose 12HR would be necessary to distinguish these complexes of similar sizes.
Although UTMP cannot be classed as a neutralizing binding protein based
on our studies in several assay formats, its affinity, which is higher
than that of 2M, suggests that it may be acting as a low
affinity, high capacity binding protein. A number of studies have
postulated that 2M, although not capable of neutralizing
activin’s activity, could act as a tissue or circulatory reservoir
because of its high concentrations in biological fluids (12, 45, 50).
UTMP is present in milligram per ml amounts in the uterine fluid of
pregnant sheep (40). Given the broad similarities in binding affinity
and concentrations of 2M and UTMP, we propose that UTMP
may constitute a mechanism by which activin is sequestered locally in
the uterine environment during pregnancy.
The initiation of this study arose from the variations in the immunoactive/bioactive ratios noted during the purification of activin from allantoic fluid. However, as UTMP does not appear to affect the actions of activin in an in vitro bioassay, it appears unlikely that UTMP represents the factor causing the change in the immunoactive/bioactive ratios observed in our initial studies. Other proteins present in allantoic fluid may be responsible for these changes, and their nature is being actively examined.
Our studies of the binding of activin to UTMP have been established using two independent techniques, namely the association of iodinated activin to UTMP and the binding of unlabeled activin to UTMP using the BIAcore system. The use of these two independent techniques, with appropriate controls, emphasizes that this phenomenon does not represent a technical artifact, but supports the view that the association of these proteins underlies an important biological function. The precise role of the UTMP-activin interaction in pregnancy remains to be determined. There is evidence that the activin ßA-subunit is expressed in the endometrium during the menstrual cycle and pregnancy in the human (51), and it is possible that a UTMP homolog in humans may play a role in the endometrial physiology of activin. Further studies are necessary to elucidate the physiological role of UTMP in mammalian pregnancy.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
2 These authors contributed equally to this work.
Received March 8, 1999.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
2-macroglobulin. Endocrinology 136:5343–5349[Abstract]
2-Macroglobulin is a binding protein of inhibin and
activin. Endocrinology 132:2038–2050This article has been cited by other articles:
 |
|
 |
|
  S. Tekin, M. B. Padua, A. M. Brad, M. L. Rhodes, and P. J. Hansen Expression and Properties of Recombinant Ovine Uterine Serpin Experimental Biology and Medicine, September 1, 2006; 231(8): 1313 - 1322. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. J. Ellem, J. F. Schmitt, J. S. Pedersen, M. Frydenberg, and G. P. Risbridger Local Aromatase Expression in Human Prostate Is Altered in Malignancy J. Clin. Endocrinol. Metab., May 1, 2004; 89(5): 2434 - 2441. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
Y.-G. Chen, H. M. Lui, S.-L. Lin, J. M. Lee, and S.-Y. Ying Regulation of Cell Proliferation, Apoptosis, and Carcinogenesis by Activin Experimental Biology and Medicine, February 1, 2002; 227(2): 75 - 87. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 W. R. Atchley, T. Lokot, K. Wollenberg, A. Dress, and H. Ragg Phylogenetic Analyses of Amino Acid Variation in the Serpin Proteins Mol. Biol. Evol., August 1, 2001; 18(8): 1502 - 1511. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 K. Tsuchida, K. Y. Arai, Y. Kuramoto, N. Yamakawa, Y. Hasegawa, and H. Sugino Identification and Characterization of a Novel Follistatin-like Protein as a Binding Protein for the TGF-beta Family J. Biol. Chem., December 22, 2000; 275(52): 40788 - 40796. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
W. Tuo, D. Zhuang, D. P. Knowles, W. P. Cheevers, M.-S. Sy, and K. I. O'Rourke PrP-C and PrP-Sc at the Fetal-Maternal Interface J. Biol. Chem., May 18, 2001; 276(21): 18229 - 18234. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. A. Irving, R. N. Pike, A. M. Lesk, and J. C. Whisstock Phylogeny of the Serpin Superfamily: Implications of Patterns of Amino Acid Conservation for Structure and Function Genome Res., December 1, 2000; 10(12): 1845 - 1864. [Abstract] [Full Text]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Endocrinology | Endocrine Reviews | J. Clin. End. & Metab. |
| Molecular Endocrinology | Recent Prog. Horm. Res. | All Endocrine Journals |
