| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
GROWTH FACTORS-CYTOKINES-ONCOGENES |
Department of Pediatric Oncology, Otto von Guericke University (P.V., B.H.), Magdeburg 39112, Germany; Department of Pediatrics, Oregon Health Sciences University (Y.O., R.G.R.), Portland, Oregon 97201; and Department of Scientific Computing, Novo Nordisk A/S (R.M.S.), Måløv DK-2760, Denmark
Address all correspondence and requests for reprints to: Dr. Peter Vorwerk, M.D., Department of Pediatric Oncology, Otto von Guericke University, Emanuel Larisch Weg 17-19, D-39112 Magdeburg, Germany. E-mail: . peter.vorwerk{at}medizin.uni-magdeburg.de
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Binding of insulin to IGFBP-3 and its N- and C-terminal fragments and of IGF-I and IGF-II to the structurally related proteins mac25 and connective tissue growth factor was also investigated. Weak insulin binding to full-length IGFBP-3 could be demonstrated in a few experiments, but we found that binding of IGF-I, IGF-II, and insulin to mac25 or connective tissue growth factor was below the detection limit of the biosensor instrument.
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
IGFBP-3 is the major serum IGFBP and transports 70–90% of the circulating IGFs (5). In target cell systems it inhibits IGF actions, but also, under specific conditions, potentiates IGF action or exerts IGF-independent effects (5, 6). Proteolysis of IGFBP-3 was initially demonstrated in human pregnancy serum, in which circulating IGFBP-3 was found primarily in low mol wt forms (7, 8). The proteolytic fragments were shown to bind IGF with lower affinities, thereby increasing availability of IGF to target receptors. Subsequent studies in different biological fluids demonstrated that limited proteolysis is not restricted to IGFBP-3, but also occurs in other IGFBP-species, IGFBP-1 through -5 (2, 3). As for IGFBP-3, the resulting fragments have a decreased affinity for IGFs and, therefore, more easily release IGF to the target receptors. Furthermore, various IGFBP fragments are capable of direct stimulatory or inhibitory action at the target cells (9, 10, 11, 12, 13).
All members of the proposed IGFBP superfamily preserve the N-terminal cysteine-rich domain, including the IGFBP motif GCGCCXXC, but vary in the intermediate region and C-terminal domain of the protein. Only the high affinity IGFBPs also share homology in the cysteine-rich C-terminal domain, which has led to the hypothesis that the conserved N-terminal domain contains the main IGF-binding activity, forming, together with the C-terminal conserved region, the high affinity IGF-binding activity in IGFBPs (1, 13, 14, 15). With the expression and purification of a variety of recombinant IGFBPs, structurally related proteins mac25 and CTGF, and well defined fragments or mutants, our understanding of structure-function relationships of the members of the proposed IGFBP superfamily has improved (12, 13, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25). However, exact measurement of the binding properties of these various proteins, and especially their proteolytic fragments, has proved difficult. Different methods have been used to estimate the binding affinities: conventional binding assays using radiolabeled IGFs, affinity cross-linking with radiolabeled ligands, gel filtration or Western ligand blots using radiolabeled or biotinylated ligands (12, 13, 14, 17, 20, 21, 26, 27, 28, 29, 30, 31, 32). However, it has been difficult to determine accurate affinities of IGFs for IGFBPs using those assays due to several limiting factors, such as the quality of labeled ligands and proteins, the effect of labeling iodine on affinity, and the limitation of the measurable range of the binding affinity.
In recent years biosensor instruments using surface plasmon resonance technology have been increasingly used to study biomolecular interactions. We (33, 34, 35) and others (15, 36, 37, 38, 39) have used this technology to study IGFBP interaction with IGFs or their analogs. In the present study our goal was to make a detailed analysis, using a BIACORE biosensor instrument (Biacore AB, Uppsala, Sweden), of the binding of IGF-I and IGF-II to recombinant IGFBP-3 and fragments of this molecule and also to the structurally related proteins mac25 and CTGF.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Immobilization of peptides on the sensor chip
All immobilizations were carried out at 25 C using an amine coupling procedure (33) with a constant flow rate of 5 µl/min. A more thorough description of the immobilization procedure can be found elsewhere (41). Equal volumes of 0.1 M N-hydroxysuccinimide and 0.1 M N-ethyl-N'-(3-diethylaminopropyl)carbodiimide were mixed by the BIACORE system’s robotics and injected over the surface of the sensor chip to activate the carboxymethylated dextran. For coupling to the sensor chip, peptides were injected over the activated surface in a 10-mM sodium acetate solution. A solution of 1 M ethanolamine was then passed over the surface to deactivate the remaining active carboxyl groups and to wash out nonspecifically bound protein.
Combinations of peptide concentration, pH during coupling, activation time, and coupling time, respectively, giving a stable surface with satisfactory amounts of the coupled peptide, were as follows: rhIGFBP-3: 10 µg/ml, pH 4.5, 3 min, 7 min; rhIGFBP-398–264: 2 µg/ml, pH 4.5, 7 min, 7 min; rhIGFBP-31–97: 20 µg/ml, pH 4.0, 10 min, 10 min; IGF-I: 4 µg/ml, pH 4.5, 7 min, 7 min; and IGF-II: 1 µg/ml, pH 4.0, 7 min, 7 min. All ethanolamine deactivation steps were run for 7 min. Immediately after the immobilization procedure, HBS buffer (10 mM HEPES; 150 mM NaCl; 3.4 mM EDTA; and 0.005% P20, pH 7.4) was flowed over the sensor chip surface for a minimum of 2 h to allow the surface to stabilize.
Quality control of immobilized surfaces
In repeated experiments for each peptide, concentration and coupling time were varied to give a range of immobilized peptide coupled to the sensor chip, to check for possible mass transport effects, which are expected to be most pronounced at high immobilization levels and high binding affinities (42, 43), and also to control for dependence of fitted binding parameters or of apparent stoichiometry of binding on immobilization level. We calculate the apparent stoichiometry as the ratio of the maximum achievable analyte bound, measured in the instrument’s standard resonance units (RU), to the theoretical maximum, where the theoretical maximum is defined as: immobilized RU x molecular weight (analyte)/molecular weight (immobilized peptide), where immobilized RU is the amount of peptide immobilized on the sensor chip, also expressed in RU. The apparent stoichiometry gives a good indication of the degree of inactivation of protein due to immobilization and the steric hindrance of binding due to overcrowding of immobilized molecules on the surface. Immobilization levels in our experiments were kept low both to give as high apparent stoichiometries as possible and to reduce steric hindrance effects. Typical apparent stoichiometries for immobilized rhIGFBP-3 were 0.8–0.9, whereas those for immobilized IGF-I were about 25% and even lower for IGF-II. Note that a low apparent stoichiometry will still give accurate binding results if the immobilized molecules are all either fully active or fully inactive.
Mass transport effects (42, 43) were observed when measuring binding of rhIGFBP-3 to immobilized IGF-I or IGF-II, and these effects could not be fully eliminated even at the high flow rates (50 µl/min) used in our experiments. Mass transport effects were not apparent for any of the other tested molecules. In the case of rhIGFBP-3 binding to immobilized IGF-I, analysis of sensorgram data using models including or lacking mass transport effects showed only minor differences in estimated binding parameters. These differences were especially minimal in global analyses (i.e. simultaneous analysis of multiple sensorgrams at different analyte concentrations) (44, 45). None of the experiments showed a systematic change in binding parameters with immobilization level.
Kinetic assays on the BIACORE
All experiments were carried out at 25 C, with a constant flow rate of 50 µl/min HBS buffer. This high flow rate was chosen to minimize mass transport effects (46). Purified analyte was diluted to various concentrations in HBS buffer using the system robotics, and the solution was injected over the peptides coupled to the chip surface for 5 min (association phase), followed by 10-min flow of HBS buffer alone (dissociation phase). In most experiments the binding phases were preceded by a 10-min wash with HBS buffer alone to allow the surface to equilibrate with the buffer. Bound analyte was removed from the coupled peptide by flowing a solution of 100 mM HCl over the surface for 3 min. This treatment regenerated the surfaces efficiently without any apparent damage. We observed, however, that IGF-I surfaces immobilized at very low levels tended to be somewhat unstable. In these cases the surface was regenerated using 50 mM HCl to minimize any damaging effects of the regeneration procedure.
In most experiments, a standard intermediate concentration of analyte was first injected over the surface to provide a reference sensorgram. Then a series of varying concentrations of analyte was injected, followed by a second reference injection, a repeat of the concentration series, and a final reference injection. With this procedure the immobilized surface could be monitored for loss of activity, and test results checked for reproducibility.
Solution affinity assays
Solution affinity assays are designed to measure the equilibrium affinity of two molecules in solution, using the BIACORE instrument as a probe to measure the free concentration of one of the molecules (47). In these experiments rhIGFBP-3 or one of its N- or C-terminal fragments was immobilized on a sensor chip to provide an active surface for measuring the free concentration of IGF-I or IGF-II in a mixture of IGF and binding protein flowing over the surface. The surface was calibrated by flowing IGF at different (free) concentrations over the surface, recording a sensorgram for each concentration. Mixtures of IGF and binding protein or fragment at various concentrations were flowed over the same surface, and the calibration curve was used to estimate the free IGF remaining in solution for each of the mixed samples. The equilibrium affinity constant, Ka, was calculated using this estimate plus the known total concentrations of IGF and binding protein, assuming a one-site binding model.
Data analysis
Kinetic analyses were carried out using the BIAEvaluation 3.0 program (Biacore AB), assuming a one-site binding model. If the surface was judged stable by inspection of the repeated series of injected analyte concentrations, all sensorgrams in an experiment were analyzed simultaneously (global analysis). If a surface showed some instability, sensorgrams were analyzed individually, and the results pooled. In all cases, special attention was paid to the fitted Rmax, which gives the amount of analyte bound at saturation and is a therefore a measure of the apparent stoichiometry of binding. We have observed that even a visually good fit of a binding model to sensorgram data sometimes gives an unreasonable value of Rmax, especially when analyzing single sensorgrams. Binding parameter estimates were rejected in such cases.
Model fitting was first done allowing the BIAEvaluation program to estimate the bulk refractive index effect for each sensorgram. Sensorgrams were then inspected visually to obtain another estimate of these values based on the step up in the sensorgram at the start of the association phase and the step down at the start of the dissociation phase (33). If the fitted estimates did not correspond to the visual estimates, the fitted estimates were rejected, and the fit was redone with the refractive index values fixed using the visual estimates. In a few experiments the surface baseline signal (that is, the signal seen when buffer alone flows over the surface) drifted slowly downward through time (baseline drift), evident in the 10-min preassociation wash and at the end of the dissociation phase. In these cases, the downward drift rate was estimated by fitting a straight line to the 10-min wash, and the sensorgram was corrected for this drift during the fitting.
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Figure 1
shows the binding of IGF-I to rhIGFBP-3 with either rhIGFBP-3 immobilized (412 RU; Fig. 1A) or IGF-I immobilized (256 RU; Fig. 1B). These sensorgrams show the typical protocol used in these studies: a 10-min wash to allow the surface to equilibrate with the buffer, followed by a 10-min association phase and a 10-min dissociation phase. The short downward spikes seen in the sensorgrams are disturbances resulting from opening and closing of valves in the BIACORE instrument’s flow system. A sharp rise in signal at the beginning of the association phase and a corresponding sharp drop at the beginning of the dissociation phase are due to a small refractive index difference between buffer alone and buffer containing the analyte protein.
|
, A and B, the concentration of analyte was 20 nM. The difference in maximum analyte bound (75 RU IGF-I vs. 200 RU rhIGFBP-3) is due to the difference in mol wt. The linearity of the initial portion of the association phase of the sensorgram in Fig. 1B indicates the presence of mass transport effects. However, analyses of these experiments with or without including mass transport in the model gave comparable results, which were also in agreement with results obtained with rhIGFBP-3 immobilized and IGF-I as analyte (e.g. Fig. 1A).
Figure 2 shows the results of similar experiments using the binding protein fragments rhIGFBP-31–97 and rhIGFBP-398–264. For each of the fragments, the shapes of the sensorgrams were similar whether the fragment or IGF-I was immobilized. The rhIGFBP-31–97 and rhIGFBP-398–264 sensorgrams had different general shapes because both the kinetic association and dissociation constants were slower for rhIGFBP-398–264 than for rhIGFBP-31–97 (see Table 1), although these differences were not statistically significant over the entire set of experiments. Note that there was a downward or upward drift in signal during the initial wash stage in the experiments with binding protein fragments immobilized. This was a general observation during these studies; that is, sensor chip surfaces immobilized with binding protein fragment tend to be more unstable than those immobilized with full-length rhIGFBP-3, or with IGF-I.
|
|
is shown a comparison of the measured binding affinity for IGF-I binding to full-length rhIGFBP-3 or to the binding protein fragments, either with binding protein immobilized or with IGF-I immobilized. The data are plotted with a logarithmic ordinate, and the results show that the affinity of IGF-I for either rhIGFBP-31–97 or rhIGFBP-398–264 is approximately 1000-fold lower than the affinity for full-length rhIGFBP-3. Although the mean affinity for rhIGFBP-31–97 measured in these experiments was higher than that for rhIGFBP-398–264, the differences were not statistically significant. All differences between full-length rhIGFBP-3 and the fragments were significant.
|
also shows that experiments in either immobilization orientation give similar measured affinities. However, because of apparent self-aggregation of rhIGFBP-3 and the C-terminal fragment (see Discussion), the results with IGF-I immobilized on the sensor chip were considered less reliable.
Whenever possible, binding parameter estimates were obtained from global fitting to multiple sensorgrams at different analyte concentrations. An example of such an experiment is shown in Fig. 4. In this experiment rhIGFBP-3 was immobilized at 726 RU, and IGF-I at concentrations from 0.5–50 nM was flowed over the immobilized surface. Because of the high affinity of IGF-I for rhIGFBP-3, binding is affected by mass transport limitation, which is most pronounced at lower analyte concentrations. The entire set of sensorgram data was fitted using a model that includes mass transport effects, and for comparison was refitted using an ordinary one-site (Langmuir) binding model. For this experiment, the computed binding affinities for the two models were 5.6 x 109 and 5.8 x 109 M-1, respectively. The respective kinetic association constants were 8.3 x 105 and 6.4 x 105 M-1s-1, and the dissociation constants were 1.4 x 10-4 and 1.2 x 10-4 s-1. Other global experiments also showed a close correspondence between the results using the two different models, indicating that the estimated binding parameters, especially the equilibrium constants, are not especially sensitive to mass transport effects when global fitting is used.
|
shows sensorgrams for IGF-I and IGF-II as well as a biotin-labeled IGF-I, all at 50 nM, binding to the same immobilized rhIGFBP-3 surface. These sensorgrams were taken from a larger experiment designed for global analysis. The top two sensorgrams are those for IGF-I and IGF-II. Note that in the association phase the IGF-I signal is higher than that for IGF-II, which is due solely to a slightly higher refractive index of the IGF-I-containing buffer. In the dissociation phase, where both surfaces are exposed to the same buffer, the IGF-II signal is slightly higher, indicating a higher affinity than for IGF-I. The affinities for IGF-I and IGF-II in this experiment, calculated by global analysis, were 2.1 x 109 and 3.0 x 109 M-1, respectively.
|
gives the summary of the calculated binding parameters for IGF-I and IGF-II binding to immobilized rhIGFBP-3 or its N- or C-terminal fragments. The table only includes results from experiments using immobilized binding protein and fragments, as the experiments with the immobilized IGFs were judged to be less reliable, primarily due to possible self-association of binding proteins (see Discussion). None of the parameter differences between the two binding protein fragments was statistically significant. All differences between full-length rhIGFBP-3 and the fragments were significant, except for those indicated.
In this experiment we also tested the binding of biotin-labeled IGF-I to immobilized rhIGFBP-3. The lowest of the curves in Fig. 5 shows the binding of biotin-IGF-I to rhIGFBP-3. In this experiment, using global analysis, the affinity of biotin-IGF-I was calculated to be 1.1 x 109 M-1.
Binding of human insulin to rhIGFBP-3 and its N- and C-terminal fragments
In this study we tested the binding of human insulin to rhIGFBP-3 and its fragments, with either binding protein or insulin immobilized on the sensor chip. In most experiments the observed binding was very weak or absent in either configuration, and reasonable estimates of binding parameters could not be made, nor could comparisons be made of the relative affinities of insulin binding to full-length rhIGFBP-3 or its fragments. Apparent binding was observed in a few experiments, and by comparing this binding with control runs using IGF-I, we could estimate equilibrium Ka values less than 106 M-1. Figure 6A shows an example of the weak binding of human insulin (2 µM) to immobilized rhIGFBP-3. For this sensorgram, the equilibrium Ka was estimated to be less than 105 M-1, assuming that insulin binds with the same stoichiometry as IGF-I.
|
, B and C, shows examples of tests of mac25 and CTGF binding to immobilized IGF-I.
Self-association of IGFBP and its fragments
Most of our experiments using immobilized IGFs gave consistent results, with calculated binding parameters very similar to those for the experiments with IGFBP immobilized (see Fig. 3). However, because of potential inaccuracies in binding parameter determinations due to the two-dimensional geometry of binding, we also attempted to measure equilibrium binding affinities using solution affinity assays on the BIACORE. In an example of one such experiment, IGF-I at 20 nM was incubated with varying concentrations (2–200 nM) of rhIGFBP-3 for 15–20 min at 25 C, after which the solution was passed over an rhIGFBP-3-immobilized sensor chip to record a sensorgram. Because rhIGFBP-3 in solution should compete for free IGF-I, the binding of IGF-I to the immobilized rhIGFBP-3 was expected to decrease at higher rhIGFBP-3 concentrations. Instead, the signal increased at higher concentrations. The most straightforward interpretation of this was that the rhIGFBP-3 was itself binding to the rhIGFBP-3 immobilized surface, and therefore solution assays could not be performed in this system.
We then investigated directly the binding of rhIGFBP-3 and its fragments to each other. Figure 7 shows an example of binding of rhIGFBP-3 or rhIGFBP-398–264 to a surface immobilized with rhIGFBP-398–264, in which binding is clearly evident. From experiments using all combinations of immobilized and solution proteins, we observed that rhIGFBP-3 and rhIGFBP-398–264 bind weakly to themselves and to each other (apparent Ka, 
8–30 x 106 M-1; that is, more than 100-fold lower affinity than that of IGF-I for rhIGFBP-3), whereas rhIGFBP-31–97 binds very weakly or not at all to itself and to the other proteins (Ka, 0–3 x 106 M-1). Because rhIGFBP-3 and rhIGFBP-398–264 would be expected to also aggregate in the flow-through solution in these assays, these results are approximate.
|
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The functional significance of different glycosylation forms of recombinant human IGFBP-3 was studied by Firth et al. (18). It could be shown that different glycosylation patterns (e.g. fully glycosylated vs. nonglycosylated IGFBP-3) do not alter IGF binding to the binding protein in Western ligand blots. Therefore, the difference in binding affinity of nonglycosylated Escherichia coli-expressed IGFBP-3 and baculovirus-expressed glycosylated IGFBP-3 fragments is not expected to be due to their glycosylation status.
Recently, Galanis et al. (38), using a BIACORE instrument, measured equilibrium affinities of 1.03 x 107 and 6.16 x 107 M-1 for IGF-I and IGF-II binding, respectively, to an N-terminal IGFBP-31–88 fragment, compared with our measurements of about 0.6 x 107 M-1 for both IGF-I and IGF-II. Interestingly, they found that the 1–88 N-terminal fragment could not be immobilized using the amine coupling method, and that a 165–264 C-terminal fragment could be immobilized, but was inactive in BIACORE assays. In contrast, both our 1–97 N-terminal and 98–264 C-terminal fragments could be immobilized and retained activity, and both were also active in binding to immobilized IGF-I.
Our measurements of the binding of IGF-I and IGF-II to full-length rhIGFBP-3 indicate equilibrium affinities in the range of 4–5 x 109 M-1. These estimates are lower than those measured using 125I-labeled IGFs, where affinities between 8 x 109 and 152 x 109 M-1 have been reported (quoted and summarized in Ref. 36). We had observed a similar difference in an earlier study of IGF analog binding to IGFBP-3 (33), but it is not clear whether this is due to the different conditions of binding, for example, changes in the ligand due to immobilization or the two-dimensional binding geometry, or to differences in analytical methods.
Wong et al. (36), using a BIACORE instrument, calculated an affinity of 18.4 x 109 M-1 for IGF-I binding to hrIGFBP-3. These results are apparently the average of separate analyses of individual sensorgrams, whereas our results for IGF-I binding to immobilized rhIGFBP-3 are derived from global analyses (44, 45) (see Fig. 4), in which a binding model is fitted simultaneously to multiple sensorgrams at different analyte concentrations. Global analyses should be more accurate, because they assume a common value in all sensorgrams for Rmax, the maximum analyte bound at saturation, whereas analyses of single sensorgrams from the same experiment can yield different apparent values of Rmax, which is unrealistic.
Galanis et al. (38) also determined a high affinity (1 x 1011 M-1) for full-length IGFBP-3 binding to immobilized IGF-I, compared with our estimate of 7 x 109 M-1 under similar conditions. Their high affinity was primarily due to a very low kinetic kd (1 x 10-6 s-1 compared with 3 x 10-4 s-1 in our experiments). One possible explanation for the difference is that Galanis et al. (38) conducted their assays at a flow rate of 5 µl/min, whereas ours were performed at 50 µl/min. It is well known (50, 51) that for high affinity interactions such as that between IGF-I and IGFBP-3, there can be significant rebinding of analyte during the dissociation phase, especially at low flow rates, leading to erroneously low dissociation rates and higher affinities. In tests of IGF-I binding to rhIGFBP-3 using flow rates of 5–10 µl/min and low analyte concentrations, we observed that analyses of individual sensorgrams often yielded unreasonably high Rmax values, low kd, and apparent affinities well over 1011 M-1. Hence, all of our later experiments were performed at high flow rates, and global analyses were used to minimize this type of error.
Her, and in our previous study (33), our data show that IGF-II binds rhIGFBP-3 with a similar, perhaps slightly higher, affinity than does IGF-I, which was also observed in other studies in the literature (52). Our measurements of binding of biotinylated IGF-I to rhIGFBP-3 using global analysis show that ka is reduced by a factor of 3, and kd is increased by a factor of 1.4 relative to nonbiotinylated IGF-I. This results in a 4-fold reduction in the apparent equilibrium affinity of biotinylated IGF-I relative to nonbiotinylated IGF-I. These are relatively small differences, indicating that biotinylation of IGF-I does not greatly interfere with its binding to IGFBPs. It is of note that labeling with iodine resulted in an affinity change of labeled IGFs to IGFBP-3 (53). Thus, the biotinylated IGF-I could be a better alternative to [125I]IGF-I in experiments requiring labeled hormone.
We attempted to measure equilibrium binding affinities using a solution binding assay (47) to minimize errors in kinetic determinations due to the two-dimensional binding geometry in the BIACORE instrument, for example, from mass transport limitation or steric hindrance (46). However, we observed that rhIGFBP-3 and rhIGFBP-398–264 self-aggregate, making solution assays impossible and raising questions about the accuracy of binding results obtained using binding proteins in the flow solution and immobilized IGFs. With increasing binding protein concentration, self-aggregation should increase both in solution and on the sensor chip surface, introducing errors in the binding analysis. Therefore, results using immobilized binding proteins, for example, as shown in Table 1, should be more reliable than those using immobilized IGFs (36, 38) unless binding protein concentrations are kept well below the Kd for self-aggregation (10-7 M by our estimates).
On the basis of evidence that activation of the insulin receptor is inhibited by IGFBPs, and that mac25 is able to bind insulin (13), we tested the binding of human insulin to rhIGFBP-3 and to recombinant mac25 and CTGF molecules using the BIACORE instrument. We also tested insulin binding to the IGFBP recombinant fragments rhIGFBP-31–97 and rhIGFBP98–264. Most experiments could not confirm binding, but in a few experiments the binding of insulin to rhIGFBP-3 was detectable, indicating that insulin may bind IGFBP-3 with low affinity, as we had previously observed (33). We could not detect binding of insulin to the IGFBP-related proteins, or to the rhIGFBP-3 fragments, in these biosensor assays.
We tested the binding of IGF-I and IGF-II to the mac25 and CTGF molecules, with either these molecules or the IGF molecules immobilized, at analyte concentrations up to 500 nM. In no case could we detect binding using the biosensor instrument. These results were unexpected in light of the evidence for binding of both IGF-I and insulin to the binding protein molecules from studies using radiolabeled IGF molecules (13, 20). However, as binding affinities less than about 105 or 106 M-1 are difficult to detect using the biosensor, we cannot rule out that there is weak binding below the detection limit of the instrument. There is an indication from affinity cross-linking studies that IGF-I binding affinity for mac25 may be higher than these values (13, 40) so it remains to be determined whether our inability to detect binding on the biosensor instrument is because of differences in binding conditions or because cross-linking experiments are, in fact, measuring extremely low binding affinities.
|
Footnotes |
|---|
Abbreviations: CTGF, Connective tissue growth factor; IGFBP, IGF binding protein; rhIGFBP-31–97, recombinant human IGF binding protein-31–97; Rmax, amount of analyte bound at saturation in RU; RU, resonance units.
Received August 30, 2001.
Accepted for publication January 2, 2002.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
This article has been cited by other articles:
 |
|
 |
|
  T. Sitar, G. M. Popowicz, I. Siwanowicz, R. Huber, and T. A. Holak Structural basis for the inhibition of insulin-like growth factors by insulin-like growth factor-binding proteins PNAS, August 29, 2006; 103(35): 13028 - 13033. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. Lopez-Bermejo, J. Khosravi, J. M. Fernandez-Real, V. Hwa, K. L. Pratt, R. Casamitjana, M. M. Garcia-Gil, R. G. Rosenfeld, and W. Ricart Insulin Resistance Is Associated With Increased Serum Concentration of IGF-Binding Protein-Related Protein 1 (IGFBP-rP1/MAC25). Diabetes, August 1, 2006; 55(8): 2333 - 2339. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 G. J. Allan, E. Tonner, M. Szymanowska, J. H. Shand, S. M. Kelly, K. Phillips, R. A. Clegg, I. F. Gow, J. Beattie, and D. J. Flint Cumulative Mutagenesis of the Basic Residues in the 201-218 Region of Insulin-Like Growth Factor (IGF)-Binding Protein-5 Results in Progressive Loss of Both IGF-I Binding and Inhibition of IGF-I Biological Action Endocrinology, January 1, 2006; 147(1): 338 - 349. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. Beattie, K. Phillips, J. H Shand, M. Szymanowska, D. J Flint, and G. J Allan Molecular recognition characteristics in the insulin-like growth factor (IGF)-insulin-like growth factor binding protein -3/5 (IGFBP-3/5) heparin axis J. Mol. Endocrinol., February 1, 2005; 34(1): 163 - 175. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. J Headey, K. S Leeding, R. S Norton, and L. A Bach Contributions of the N- and C-terminal domains of IGF binding protein-6 to IGF binding J. Mol. Endocrinol., October 1, 2004; 33(2): 377 - 386. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. Lam, R. N. van der Geest, N. A.M. Verhagen, F. A. van Nieuwenhoven, I. E. Blom, J. Aten, R. Goldschmeding, M. R. Daha, and C. van Kooten Connective Tissue Growth Factor and IGF-I Are Produced by Human Renal Fibroblasts and Cooperate in the Induction of Collagen Production by High Glucose Diabetes, December 1, 2003; 52(12): 2975 - 2983. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. Lopez-Bermejo, J. Khosravi, C. L. Corless, R. G. Krishna, A. Diamandi, U. Bodani, E. M. Kofoed, D. L. Graham, V. Hwa, and R. G. Rosenfeld Generation of Anti-Insulin-Like Growth Factor-Binding Protein-Related Protein 1 (IGFBP-rP1/MAC25) Monoclonal Antibodies and Immunoassay: Quantification of IGFBP-rP1 in Human Serum and Distribution in Human Fluids and Tissues J. Clin. Endocrinol. Metab., July 1, 2003; 88(7): 3401 - 3408. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
L. D. Payet, X.-H. Wang, R. C. Baxter, and S. M. Firth Amino- and Carboxyl-Terminal Fragments of Insulin-Like Growth Factor (IGF) Binding Protein-3 Cooperate to Bind IGFs with High Affinity and Inhibit IGF Receptor Interactions Endocrinology, July 1, 2003; 144(7): 2797 - 2806. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. H. Shand, J. Beattie, H. Song, K. Phillips, S. M. Kelly, D. J. Flint, and G. J. Allan Specific Amino Acid Substitutions Determine the Differential Contribution of the N- and C-terminal Domains of Insulin-like Growth Factor (IGF)-binding Protein-5 in Binding IGF-I J. Biol. Chem., May 9, 2003; 278(20): 17859 - 17866. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||
, Binding protein or fragment immobilized, IGF-I in the flow-through buffer; 
, IGF-I immobilized, binding protein or fragment in the flow-through buffer. Error bars show SDs. All differences between parameters for rhIGFBP-3 and those for the fragments are statistically significant, but differences between the two fragments are not.
