| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Experimental Studies |
School of Animal and Microbial Sciences, University of Reading (R.J.W., C.F.K., P.J.L.), Whiteknights, Reading RG6 6AJ; and the Department of Rheumatology, Battle Hospital (J.D.), Reading RG30 IAG, United Kingdom
Address all correspondence and requests for reprints to: Dr. R. J. Woods, School of Animal and Microbial Sciences, University of Reading, Whiteknights, P.O. Box 228, Reading, United Kingdom RG6 6AJ.
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Here we present evidence for a C-terminally truncated form of the native binding protein in the plasma of subjects suffering from rheumatoid arthritis, which may parallel the in vitro truncation.
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The catabolism and half-life of circulating CRF-BP and the fate of the complex formed between CRF-BP and CRF (9) are unknown. It is recognized that as a 37-kDa protein, CRF-BP is small enough to pass through the glomerular membranes (10). However, the complex formed between monomeric CRF-BP and CRF is capable of dimerization (9), and with a resultant mass of 80 kDa, its filtration fraction, like that of albumin, would be extremely small. It is unlikely, therefore, that the rapid disappearance of CRF-BP from the plasma after iv bolus injection of CRF (11) occurs via kidney clearance, and we have postulated that it is mediated by other tissue systems. Dimerization may, therefore, facilitate the uptake and clearance of CRF, after which the binding protein moiety would be returned to the blood, possibly in an altered form, or, alternatively, all may be catabolized.
We have observed by RIA that bolus injection of CRF into human subjects results in the 50% reduction of binding protein levels in plasma and simulates the changes in plasma CRF-BP levels that occur in the last few weeks of pregnancy (11). The return to the circulation of CRF-BP, after discharge of hCRF ligand in a form that cross-reacts poorly in the current RIA, may go some way to explain the rapid fall in CRF-BP levels after iv injection of CRF, and the fall that occurs in late pregnancy that accompanies rising concentrations of placental CRF being secreted into the circulation. These negative changes in plasma CRF-BP are reversed after CRF administration is discontinued and after parturition (12, 13).
Interaction with peptide ligands and dimerization both change the conformation of CRF-BP (12), and removal of ligand is known to reverse dimerization (9), but we suspect that the protein may not return to its original conformation. We have investigated the existence of CRF-BP isoforms using denatured recombinant CRF-BP.
A conformational change in CRF-BP may also be involved in generating another form of heterogeneity observed in both recombinant CRF-BP and in the native, circulating form. There are increasing numbers of reports of cleavage of proteins in plasma that result in modification of their bioactivity. PRL (14) and insulin-like growth factor-binding proteins 3 and 4 (IGFBP-3 and IGFBP-4) can undergo cleavage in vivo and in vitro, although the proteases responsible have not yet been isolated (15, 16, 17).
In a preliminary study we have found that the determination of CRF-BP levels in human plasma samples from arthritic patients by RIAs using antibodies that react with different epitopes can lead to different values being obtained for the same plasma. Reported values for circulating CRF-BP differ greatly among laboratories (13, 18, 19); this may be explained by heterogeneity of the protein. We have investigated the possibility that in arthritic subjects, truncation may be a possible explanation.
This report attempts to explain the rapid disappearance of the binding protein after injection of CRF and suggests that intracellular processes designed to remove the ligand may result in temporary denaturation of the protein that is returned to the circulation.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
hCRF and CRF-BP-(25–45) were generous gifts from Dr. J. Rivier (The Salk Institute, La Jolla, CA), and CRF-BP-(298–322) was obtained from Prof. T Suda (University of Hirosaki, Hirosaki, Japan).
Antibodies
Polyclonal rabbit antisera, A and B, raised against the N-terminal peptide CRF-BP-(25–45) were gifts from Dr. Vale, and polyclonal rabbit antiserum raised against the C-terminal peptide CRF-BP-(298–322) was a gift from Prof. Suda. All bound 50% radiolabeled recombinant CRF-BP at a dilution of approximately 1:5000.
A third type of polyclonal antiserum (RABPAb) that binds 50% of radiolabeled recombinant binding protein at a concentration of 1:10,000 was raised in rabbits against intact recombinant binding protein as described previously (12). The resulting antiserum did not interact with radiolabeled synthetic peptides CRF-BP-(25–45) and CRF-BP-(298–322). The addition of a second antibody (SARFc) was used to separate bound and free radiolabels. It consisted of 10% (vol/vol) sheep antiserum raised against the Fc region of rabbit IgG prepared in 0.05 mol/L phosphate buffer at pH 7.4 containing 4% (wt/vol) polyethylene glycol 6000 (Sigma Chemical Co., St. Louis, MO).
Plasma samples
Blood was collected in ethylenediamine tetraacetate from normal controls and patients with rheumatoid arthritis (ARA criteria 1987) (20). The blood was centrifuged, and plasma was removed and stored at -20 C. All rheumatoid arthritis patients were treated with nonsteroidal antiinflammatory drugs. There were 13 control subjects (7 women and 6 men) and 11 arthritic subjects (6 women and 5 men). All samples were collected between 1400–1600 h.
Production of recombinant CRF-BP
Chinese hamster ovary cells were transfected with a complementary DNA sequence isolated from a human liver library (21). CRF-BP was then isolated from the culture medium by affinity chromatography (11) and eluted with freshly prepared buffer, pH 10.5, containing 20% acetonitrile and the protease inhibitors, ethylenediamine tetraacetate (10 mmol/L), iodoacetamide (5 mmol/L), pepstatin (1 µmol/L), and phenylmethylsulfonylfluoride (1 mmol/L). The pH was adjusted immediately to 8.5 by the addition of 100 µL/mL of 0.5 mol/L orthophosphoric acid, pH 3.5. Acetonitrile was removed under a stream of nitrogen without elevation of pH. These preparations were checked for integrity by SDS-PAGE, carried out under reducing conditions according to the method of Schagger and von Jagow (22), both before and after storage at -20 C.
Preparation of standards containing CRF-BP and urea-treated binding protein (UBP)
Immediately after preparation, intact CRF-BP was diluted about 30-fold in pooled sheep serum to give a concentration of 3.48 mg/L. This stock was divided into aliquots of 250 µL and stored frozen at -20 C.
UBP was prepared by the addition of solid urea to 400 µg purified protein in 1 mL acetonitrile-free, neutralized elution buffer to a final concentration of 8 mol/L and incubation for 0.5 h at 37 C. The incubate was also diluted by the addition of sheep serum to a final concentration of 3.48 µg/L and stored as described above. The final concentration of urea in this stock was 60 mmol/L.
Antibody dilution curves
Serial dilutions of anti-N-terminal antibodies A and B of between 2,000- and 32,000-fold were prepared in assay buffer [0.05 mol/L sodium phosphate, pH 7.4, containing 0.5% (wt/vol) BSA], and 200 µL of each were incubated in duplicate at room temperature with 50 µL radioiodinated synthetic peptide, CRF-BP-(25–45) (2 x 104 cpm), or CRF-BP (2 x 104 cpm). Bound labels were precipitated by the addition of 200 µL SARFc reagent, as described above. After 0.5 h, 2 mL saline were added to samples, and all were separated by centrifugation at 5,000g for 25 min. Supernatants were aspirated, and the radioactivity of pellets was determined.
Ligand immunoradiometric assay (LIRMA)
CRF-BP standards were prepared by diluting the stock CRF-BP in the assay buffer described above. To duplicate 100-µL volumes of samples or standards were added 50 µL [125I]hCRF, with an activity of 35 x 106 cpm. All were incubated at room temperature for 0.5 h, after which 50 µL of a 100-fold dilution of anti-CRF-BP antiserum or UBP antiserum in assay buffer were added. After an additional hour, 200 µL SARFc were added, and incubation was continued for 0.5 h. Finally, 2 mL normal saline were added to samples and standards, and all were separated by centrifugation at 5000g for 25 min. Supernatants were aspirated, and the radioactivity of pellets was determined. Concentrations of CRF-BP were automatically computed with reference to the standard curve.
Cross-reaction of UBP with CRF-BP
Cross-reaction of UBP with CRF-BP was determined by RIA, using the rabbit antiserum raised against intact recombinant CRF-BP at a final concentration of 1:12 x 103. Serial dilutions of CRF-BP or UBP standards, ranging from 7.25–464 µg/L, were added in 100 µL assay buffer to 100 µL [125I]CRF-BP (2 x 104 cpm). CRF-BP was radioiodinated by the glucose oxidase-lactose peroxidase method (11). Duplicates of each standard were incubated at 4 C for 16 h with 100 µL antiserum. Assays were separated after the addition of SARFc.
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Figure 1
shows the interaction of N-terminal
antibodies A and B with the synthetic N-terminal peptide against which
they were raised and the N-terminus in intact recombinant CRF-BP. Both
antibodies, A and B, react with the radiolabeled peptide to the same
degree over a range of 2,000- to 32,000-fold dilution.
|
Interaction of CRF-BP and UBP with N- and C-terminal antibodies
The properties of the two antibody preparations, A and B, directed
against the N-terminus of CRF-BP were compared in the LIRMA with that
of an antibody directed against the C-terminus and with that of an
antibody (RABPAb) raised against intact recombinant CRF-BP. For these
experiments, all antibodies were present at a final dilution of 1:400.
The results are summarized in Fig. 2. The capacity of
RABPAb to bind CRF-BP and its attached ligand, radioiodinated hCRF, was
greater than that of the three antipeptide antibodies. The extent to
which radiolabeled ligand complex was bound by C-terminal antibody and
N-terminal antibody A was similar, but at 116 mg CRF-BP/L, the capacity
of the C-terminal antiserum to bind CRF-BP was exceeded, and
precipitation of radiolabel began to decline.
|
. Nevertheless, upon addition of SARFc, all
antibodies precipitated the complex formed by [125I]CRF
and binding protein in a manner directly related to the concentration
of CRF-BP.
These experiments were repeated with binding protein that had
been exposed to 8 mol/L urea. The results are shown in Fig. 3. Once again, RABPAb was the most effective in binding
the radioactive complex, but this time a lower percentage of the total
radioactive ligand was precipitated. The activity of the C-terminal
antibody was similar to that in the previous experiment; that of
N-terminal antibody A was slightly reduced. N-Terminal antibody B,
however, failed to bind to the UBP-ligand complex to any significant
degree.
|
The polyclonal antiserum, RABP Ab, was used at a 12,000-fold
dilution to compare the cross-reactivity of UBP with that of CRF-BP
under RIA conditions. At this dilution it cross-reacted poorly, as
shown in Fig. 4. At 50% of maximum binding of
radiolabel, the cross-reaction of UBP was approximately 20% that of
CRF-BP.
|
Heterogeneity resulting from proteolytic cleavage was also
observed in a series of highly purified preparations of recombinant
CRF-BP. Figure 5 shows that when examined by SDS-PAGE,
in most of the 12 preparations more than 1 protein band was visualized
after staining with Coomassie brilliant blue G. Western blotting for
cleaved preparations with both N- and C-terminal antibodies showed that
a C-terminal fragment of approximately 10 kDa had been excised. Figure 6 (left) shows the remaining 27-kDa fragment
identified by the N-terminal antibody. In Fig. 6 (right),
the C-terminal antibody binds to the intact protein, but the truncated
fragment was not identified by this antibody. The C-terminal peptide of
10 kDa was also not apparent on this gel.
|
|
Figure 7 shows a linear correlation obtained by
plotting the concentrations of active binding protein in the plasma of
healthy control subjects determined by LIRMA using N- and C-terminal
antibodies. Both antibodies give similar values. In contrast, values
for plasma samples taken from arthritic patients did not correlate.
Results obtained with the C-terminal antibody are consistently lower
than those obtained with the N-terminal antibody.
|
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Other differences between the CRF-BP and UBP preparations can be discounted. The concentration of both UBP and CRF-BP used in these experiments was nominally the same, and the functional activities of the preparations were similar, as approximately the same amount of radiolabeled CRF was precipitated after reacting with another antibody raised against the same fragment or with the C-terminal antibody. By its reaction with antibodies to both N- and C-terminals, UBP was shown to be intact and not truncated before or after treatment with urea.
The mechanism by which stability of the isoform is maintained is as yet unknown. Although small proteins with several disulfide bonds are reported to renature particularly rapidly (23) after denaturation, there are no theoretical objections to the existence of two or more stable configurations of a complex polymer with similar free energies yet conserved by the barrier imposed by the activation energy necessary to convert one conformation to another. As a consequence, both domains of the protease, carboxypeptidase Y, will not refold completely after denaturation unless the pro-region is still attached (24). Proteins with multiple conformations in the native state, all of which are bioactive, have also been reported. For both Staphylococcal nuclease (25) and octopine dehydrogenase (26), they may undergo interconversion after repeated denaturation. Holoazurins isolated from Pseudomonas fluorescens and P. aeruginosa appear to exist in three conformational states (27).
Proteins renatured at high concentration can achieve stability by aggregation (28), but there is no evidence for aggregation of UBP, and in our experiments this would be unlikely, as BP containing 8 mol/L urea was diluted rapidly in sheep serum by an excess of 30-fold. Gel filtration of UBP on Sephacryl S-200 results in elution of the protein in the same volume as that of the monomeric CRF-BP (results not shown). Furthermore, addition of CRF to UBP results in the formation of a dimer complex that elutes in the same volume as recombinant CRF-BP dimer, thus demonstrating that the functional activity of UBP is retained.
It is possible that the circulating isoforms would not necessarily be similar to UBP; consequently, specific antibodies to detect them in vivo are currently not available to us. The physiological roles of potential isoforms, including the relative affinities of UBP and CRF-BP for various peptide ligands, remain to be investigated.
In patients with arthritis or septicemia, CRF-BP levels are elevated
(29). The elevated levels found in these active immune states appear to
be caused by increased liver secretion, presumably due to activation of
enhancer elements found in the 5'-flanking region of the CRF-BP gene by
the transcription factors nuclear factor-
ß and interferon-1, which
are known to be implicated in the acute phase response (21). CRF or CRF
immunoreactivity of peripheral origin is reported to be involved in the
initiation of the inflammatory cascade, and CRF receptors are present
on human lymphocytes and monocytes (30) as well as on mouse spleen
macrophages (31). Changes in the levels of CRF-BP will, therefore,
influence free CRF levels, and many of the reported actions of CRF may
be mediated and modulated through endogenous CRF-BP. Heterogeneity of
the binding protein as a result of inflammatory disease might,
therefore, influence the binding and availability of CRF or other
ligands of the type we detected in synovial fluid (29). A second form
of heterogeneity was indeed observed that appears to be attributed to
truncation of the protein, probably toward the C-terminus. Truncation
was also observed to occur in stored preparations of CRF-BP and is
believed to result from an autocatalytic cleavage of the binding
protein (32). It remains to be established whether the cleavage site is
identical for both native and recombinant proteins, but the possibility
is established that, like IGFBP-3 and IGFBP-4, modifications of
physiological significance may occur (15, 16, 17). Two forms of CRF-BP,
resulting from N-terminal cleavage of the 25–45 peptide region, have
been identified in sheep brain (33). These, too, may affect
conformation.
The existence of isoforms and heterogeneity shown by this study serves to illustrate the additional degree of control over hormonal ligand concentrations that may result from the presence of a binding protein and adds a word of caution for the interpretation of RIAs that use antisera raised against synthetic peptide fragments.
Received October 14, 1996.
Revised January 3, 1997.
Accepted February 7, 1997.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
This article has been cited by other articles:
 |
|
 |
|
  R. J. Woods, C. F. Kemp, J. David, I. G. Sumner, and P. J. Lowry Cleavage of Recombinant Human Corticotropin-Releasing Factor (CRF)-Binding Protein Produces a 27-Kilodalton Fragment Capable of Binding CRF J. Clin. Endocrinol. Metab., August 1, 1999; 84(8): 2788 - 2794. [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 |
