Chaperone-Like Activity of a Bacterioferritin Comigratory Protein from Thermococcus kodakaraensis KOD1.

Peroxiredoxins (Prxs) are ubiquitous and conserved proteins that can catalyze the reduction of inorganic and organic hydroperoxides to protect against damage by reactive oxygen species. In this study, a Prx subfamily member, and specifically a bacterioferritin comigratory protein from hyperthermophilic Thermococcus kodakaraensis KOD1 (TkBcp), was overexpressed, purified and characterized. Based on the conserved cysteine (Cys) residues in its amino acids sequence, TkBcp can be grouped into 1-Cys Prx family. Size exclusion chromatography analysis showed that TkBcp exists in three oligomeric forms: 700 kDa, 70 kDa, and 20 kDa. The peroxidase function was found to predominate in the lowmolecular- weight (MW) form, whereas the high-MW complex has the chaperone function. Oxidative reagents caused the protein structure of TkBcp to shift from low-MW form to high-MW complexes, whereas reducing reagents caused a shift in the reverse direction. Furthermore, the high-MW form of TkBcp preferred to tightly bind DNA. The relationship of TkBcp with other homologs was also examined.

Personalized approach to growth hormone treatment: clinical use of growth prediction models.

The goal of growth hormone (GH) treatment in a short child is to attain a fast catch-up growth toward the target height (TH) standard deviation score (SDS), followed by a maintenance phase, a proper pubertal height gain, and an adult height close to TH. The short-term response variable of GH treatment, first-year height velocity (HV) (cm/year or change in height SDS), can either be compared with GH response charts for diagnosis, age and gender, or with predicted HV based on prediction models. Three types of prediction models have been described: the Kabi International Growth Hormone Study models, the Gothenburg models and the Cologne model. With these models, 50-80% of the variance could be explained. When used prospectively, individualized dosing reduces the variation in growth response in comparison with a fixed dose per body weight. Insulin-like growth factor-I-based dose titration also led to a decrease in the variation. It is uncertain whether adding biochemical, genetic or proteomic markers may improve the accuracy of the prediction. Prediction models may lead to a more evidence-based approach to determine the GH dose regimen and may reduce the drug costs for GH treatment. There is a need for user-friendly software programs to make prediction models easily available in the clinic.

A comparison of different definitions of growth response in short prepubertal children treated with growth hormone.

BACKGROUND: How to define poor growth response in the management of short growth hormone (GH)-treated children is controversial. AIM: Assess various criteria of poor response. SUBJECTS AND METHODS: Short GH-treated prepubertal children [n = 456; height (Ht) SD score (SDS) ≤-2] with idiopathic GH deficiency (IGHD, n = 173), idiopathic short stature (ISS, n = 37), small for gestational age (SGA, n = 54), organic GHD (OGHD, n = 40), Turner syndrome (TS, n = 43), skeletal dysplasia (n = 15), other diseases (n = 46) or syndromes (n = 48) were evaluated in this retrospective multicenter study. Median age at GH start was 6.3 years and Ht SDS -3.2. RESULTS: Median [25-75 percentile] first-year gain in Ht SDS was 0.65 (0.40-0.90) and height velocity (HtV) 8.67 (7.51-9.90) cm/year. Almost 50% of IGHD children fulfilled at least one criterion for poor responders. In 28% of IGHD children, Ht SDS gain was <0.5 and they had lower increases in median IGF-I SDS than those with Ht SDS >0.5. Only IGHD patients with peak stimulated growth hormone level <3 µg/l responded better than those with ISS. A higher proportion of children with TS, skeletal dysplasia or born SGA had Ht SDS gain <0.5. CONCLUSION: Many children respond poorly to GH therapy. Recommendations defining a criterion may help in managing short stature patients.

Insulin-like growth factor-I in growth and metabolism.

Deficiency of insulin-like growth factor-I (IGF-I) results in growth failure. A variety of molecular defects have been found to underlie severe primary IGF-I deficiency (IGFD), in which serum IGF-I concentrations are substantially decreased and fail to respond to GH therapy. Identification of more patients with primary or secondary IGFD is likely with investigative and diagnostic progress, particularly in the assessment of children with idiopathic short stature. Diagnosis of IGFD requires accurate and reliable IGF-I assays, adequate normative data for reference, and knowledge of IGF-I physiology for proper interpretation of data. Recombinant human IGF-I (rhIGF-I) treatment improves stature in patients with severe primary IGFD, and has also been shown to improve glycaemic control and insulin sensitivity in patients with severe insulin resistance. Ongoing studies of patients receiving rhIGF-I will allow further evaluation of the clinical utility of this treatment, with concurrent increase in our understanding of IGF-I and conditions of IGFD.

Effects of fat supplementation on postprandial GIP, GLP-1, ghrelin and IGFBP-1 levels: a pilot study on adolescents with type 1 diabetes.

AIMS: To compare the responses of GIP, GLP-1, ghrelin and IGFBP-1 between meals with different fat and energy content in adolescents with type 1 diabetes (T1DM) and to relate them to gastric emptying and glycaemia. METHODS: On different days and in a random order, 7 adolescents with T1DM ingested a high- and low-fat meal (fat content: 38 and 2 g, energy content: 640 and 320 kcal, respectively). At normoglycaemia, the same prandial insulin dose was given at both meals and to all subjects. Postprandial blood samples were taken repeatedly over 4 hours. Gastric emptying was estimated by the paracetamol absorption method. RESULTS: The area under the curve (AUC) for GIP(0-240 min) and for GLP-1(0-120 min) was larger, but smaller for relative ghrelin(0-240 min), after the high-fat meal (p = 0.002, 0.030 and 0.043, respectively). IGFBP-1 decreased significantly, but not differently, after the meals. Larger GLP-1 secretion correlated with slower gastric emptying (p = 0.029) and higher fasting ghrelin levels correlated with lower postprandial glycaemia (p = 0.007). CONCLUSION: In adolescents with T1DM, the postprandial responses of GIP, GLP-1 and ghrelin, but not that of IGFBP-1, depend more on meal size than on insulin.

An archaeal NADH oxidase causes damage to both proteins and nucleic acids under oxidative stress.

NADH oxidases (NOXs) catalyze the two-electron reduction of oxygen to H2O2 or four-electron reduction of oxygen to H2O. In this report, we show that an NADH oxidase from Thermococcus profundus (NOXtp) displays two forms: a native dimeric protein under physiological conditions and an oxidized hexameric form under oxidative stress. Native NOXtp displays high NADH oxidase activity, and oxidized NOXtp can accelerate the aggregation of partially unfolded proteins. The aggregates formed by NOXtp have characteristics similar to beta-amyloid and Lewy bodies in neurodegenerative diseases, including an increase of beta-sheet content. Oxidized NOXtp can also bind nucleic acids and cause their degradation by oxidizing NADH to produce H2O2. Furthermore, Escherichia coli cells expressing NOXtp are less viable than cells not expressing NOXtp after treatment with H2O2. As NOXtp shares similar features with eukaryotic cell death isozymes and life may have originated from hyperthermophiles, we suggest that NOXtp may be an ancestor of cell death proteins.

Hexameric ring structure of a thermophilic archaeon NADH oxidase that produces predominantly H2O.

An NADH oxidase (NOX) was cloned from the genome of Thermococcus profundus (NOXtp) by genome walking, and the encoded protein was purified to homogeneity after expression in Escherichia coli. Subsequent analyses showed that it is an FAD-containing protein with a subunit molecular mass of 49 kDa that exists as a hexamer with a native molecular mass of 300 kDa. A ring-shaped hexameric form was revealed by electron microscopic and image processing analyses. NOXtp catalyzed the oxidization of NADH and NADPH and predominantly converted O(2) to H(2)O, but not to H(2)O(2), as in the case of most other NOX enzymes. To our knowledge, this is the first example of a NOX that can produce H(2)O predominantly in a thermophilic organism. As an enzyme with two cysteine residues, NOXtp contains a cysteinyl redox center at Cys45 in addition to FAD. Mutant analysis suggests that Cys45 in NOXtp plays a key role in the four-electron reduction of O(2) to H(2)O, but not in the two-electron reduction of O(2) to H(2)O(2).

Lack of sex differences in the IGF-IGFBP response to ultra endurance exercise.

The insulin-like growth factor (IGF)-IGF binding proteins (BP) and the pituitary-gonadal axes were investigated during ultra endurance exercise in 16 endurance-trained athletes (seven women). Median duration of the race was 6.3 days. Although food and drink were ad libitum, energy balance was negative. Blood samples were drawn before (PRE), at the end of (END) and 24 h after (POST24h) the race. Serum concentrations of total IGF-I (t-IGF-I) and free IGF-I (f-IGF-I) decreased by 33 (SD 38)% and 54 (19)%, respectively. The decrease in t-IGF-I appeared to be associated to the total energy deficit during the race. At END, the IGFBP-3 fragmentation and IGFBP-1 were increased but these changes did not predict changes in f-IGF-I. An increase in POST24h IGFBP-2 levels in women was the only sex difference. Testosterone was decreased by 67 (12)% in the men and estradiol became undetectable in the women without any detectable increase in LH and/or FSH. In conclusion ultra endurance exercise results in similar IGF-IGFBP responses in men and women reflecting a catabolic state. IGFBP-2 was the only exception, with increased levels in women after exercise. A concomitant decrease in gonadal hormones was not related to endocrine changes in the IGF-IGFBP axis but may be related to local changes in IGF-I expression.

Effects of fat supplementation on glycaemic response and gastric emptying in adolescents with Type 1 diabetes.

AIMS: To compare the glycaemic response to meals with different fat content in adolescents with Type 1 diabetes mellitus (T1DM) and to investigate associations with gastric emptying. METHODS: In this randomized, cross-over study, paired results were obtained from seven adolescents with T1DM who ingested on different days two meals with the same carbohydrate and protein content, but different fat and energy content (2 and 38 g fat, 320 and 640 kcal, respectively). Paracetamol was mixed into the meals and gastric emptying was estimated by the paracetamol absorption method. All subjects were normoglycaemic and given 7 IU insulin aspart at commencement of ingestion. Postprandial blood samples were taken during 4 h. RESULTS: The areas under the curves for plasma glucose and serum paracetamol concentrations were larger after the low-fat than after the high-fat meal during the first 2 h (P = 0.047 and P = 0.041, respectively). The difference between meals in time-to-peak in glucose and paracetamol concentrations did not reach statistical significance (high-fat vs. low-fat meal: 210 min (120-240) vs. 120 min (50-240), P = 0.080 and 120 min (75-180) vs. 60 min (60-120), P = 0.051, respectively). Changes in glucose concentrations correlated with simultaneous changes in paracetamol concentrations (P < 0.001). CONCLUSIONS: For the first time, we have shown that the initial glycaemic response is reduced after a meal with higher compared with a meal with lower fat content in adolescents with T1DM given a rapid-acting insulin analogue preprandially. The type and dose of preprandial insulin may need adjustment to the fat content of the meal to reach postprandial normoglycaemia.

Interstitial IGF-I in exercising skeletal muscle in women.

OBJECTIVE: To study interstitial IGF-I concentrations in resting and exercising skeletal muscle in relation to the circulating components of the IGF-IGF binding protein (IGFBP) system. DESIGN AND METHODS: Seven women performed endurance exercise with 1 leg (Ex-leg) for 1 h. The resting leg (Rest-leg) served as a control. IGF-I was determined in microdialysate (MD) and was compared with veno-arterial (v-a) concentrations of circulating IGF-IGFBP components. RESULTS: Median (range) basal MD-IGF-I was 0.87 (0.4-1.5) microg/l or 0.4 (0.2)% of total-IGF-I (t-IGF-I) determined in arterial serum and in the same concentration range as free dissociable IGF-I (f-IGF-I). Rest-leg MD-IGF-I decreased, reaching significance after exercise. Ex-leg MD-IGF-I was unchanged during exercise and declined after exercise at the level of significance (P = 0.05). There was a release of f-IGF-I from the Ex-leg into the circulation at the end of and shortly after exercise. A small but significant increase in circulating IGFBP-1 was detected at the end of exercise and IGFBP-1 increased further after exercise. Although interleukin-6 (IL-6) has been associated with IGFBP-3 proteolysis, the circulating molecular forms of IGFBP-3 remained unchanged in spite of an IL-6 release from the muscle compartment. CONCLUSIONS: Circulating IGFBP-1 is related to interstitial IGF-I in resting muscle although the temporal relationship may not be simple. Further studies should explore the role of local release of IGF-I and its impact on IGF-I activity during contraction.

A 30-kDa fragment of insulin-like growth factor (IGF) binding protein-3 in human pregnancy serum with strongly reduced IGF-I binding.

Proteolytic cleavage of insulin-like growth factor (IGF) binding protein (IGFBP)-3 during pregnancy is likely to have both IGF-dependent and -independent effects on maternal, placental and fetal growth and metabolism. A 30-kDa proteolytic IGFBP-3 fragment was isolated from third trimester pregnancy human serum and identified by N- and C-terminal amino acid sequence analysis and mass spectrometry to correspond to residues 1-212 of the parent protein. This fragment is the dominating IGFBP-3 immunoreactive species in pregnancy serum. The 30-kDa fragment was also detected in serum of non-pregnant women where it coexists with intact IGFBP-3. Using biosensor technology, (1-212)IGFBP-3 was found to have 11-fold lower affinity for IGF-I compared to intact IGFBP-3, while a 4-fold decrease in affinity was found for IGF-II. Tests with des(1-3)IGF-I suggest fast binding of IGF-I to the N-terminal region of IGFBP-3 and similar affinity to a slow binding site in the C-terminal region.

Local changes in the insulin-like growth factor system in human skeletal muscle assessed by microdialysis and arterio-venous differences technique.

IGF-I plays a direct role in whole body glucose homeostasis primarily by stimulating skeletal muscle glucose uptake. IGF-I is also involved in exercise induced muscle hypertrophy. Knowledge regarding local changes in muscle IGF-I bioavailability and its regulation by IGFBPs at rest and during exercise is limited. We have therefore explored changes in total IGF-I levels as well as circulating IGFBP levels and their post-translational modifications over an exercising leg. For the first time we have determined IGF-I levels in exercising skeletal muscle microdialysate in an attempt to assess local IGF-I bioavailability. Eighteen healthy young men performed one legged knee-extension exercise during 45min. Blood samples were taken from the femoral artery and vein of the exercising leg. No significant differences between arterial and venous concentrations of total IGF-I or IGFBP-1 were detected over the leg at any time. IGF-I concentrations increased significantly during exercise in the artery but not in the vein. Total IGFBP-1 increased after exercise in both artery and vein. The increase in non-plus less phosphorylated forms of IGFBP-1 was less pronounced and did not reach statistical significance. The proportion of fragmented IGFBP-3 (IGFBP-3 proteolysis) assessed by Western immunoblotting did not change significantly during or after exercise. Although optimization and validation of IGF-I determinations in muscle microdialysate (md) will be required, our first results using this technique demonstrate a significant 2-fold increase in mdIGF-I collected during and after exercise. We conclude that determination of A-V-differences appears to be of limited value in the assessments of local muscle change in the IGF-system. A substantial release of IGF-I during short time is required to detect significant change in the large circulating store of IGF-I. We suggest that an optimized and validated microdialysis technique for determination of local IGF-I may be advantageous in future studies.

Acute interleukin-6 infusion increases IGFBP-1 but has no short-term effect on IGFBP-3 proteolysis in healthy men.

Human conditions of elevated interleukin-6 (IL-6) and transgenic mice overexpressing IL-6 have increased proteolytic degradation of insulin-like growth factor binding protein (IGFBP)-3. In addition, IL-6 alters the hepatic expression of insulin-like growth factor-I (IGF-I) and the IGFBPs in vitro. The aim of the present study was to investigate whether moderately elevated IL-6 levels have short-term effects on circulating IGF-I, IGFBP-1 and IGFBP-3 proteolysis in vivo. Healthy men received a 3-h IL-6 (n = 6) or saline (n = 6) infusion and blood samples were collected prior to and up to 8 h after the start of infusion. Free IGF-I, total IGF-I, IGFBP-1, insulin and cortisol were measured using immunoassays. Serum IGFBP-3 proteolysis was analyzed by Western immunoblot and by in vitro degradation of (125)I-IGFBP-3. We found that IL-6 concentrations reaching approximately 100 pg/ml significantly increased IGFBP-1 after the end of infusion in the absence of changes in insulin. In addition, plasma levels of cortisol were increased in response to IL-6 during and after infusion compared to saline. There was no effect of IL-6 on IGFBP-3 proteolysis, total IGF-I or free dissociable IGF-I. These data suggest that moderately elevated levels of IL-6 such as in the post-operative state or after exercise may contribute to increased levels of IGFBP-1. Although this study does not exclude that high levels and/or prolonged exposure to IL-6 may induce IGFBP-3 proteolysis in sepsis or chronic inflammatory disease, it suggests that IL-6 released from exercising skeletal muscle is not directly involved in proteolysis of circulating IGFBP-3.

Lack of insulin-like growth factor binding protein-3 protease activation by venous cannulation.

Disruption of the endothelium activates thrombogenic and fibrinolytic enzymes that cleave insulin-like growth factor binding protein-3 (IGFBP-3) in vitro. The aim of the present human study was to determine whether blood sampling, i.e., venous stasis and cannulation increase IGFBP-3 proteolysis before and/or after surgery by activating these enzymes. Serum samples obtained immediately after cannulation were compared with samples obtained from a previously inserted venous catheter. Cannulation did not increase serum IGFBP-3 proteolytic activity pre- and post-operatively, as determined by in vitro degradation of 125I-IGFBP-3. Furthermore, there was no effect on in vivo IGFBP-3 fragmentation assessed by western immunoblot. In addition, a standardized venous stasis did not affect IGFBP-3 proteolytic activity or fragmentation. Comparison of IGFBP-3 proteolytic activity before and after surgery demonstrated a significant post-operative increase. However, this could not be demonstrated immediately after the initial cannulation, due to a large individual variation at this time-point before surgery.

A double-blind, randomized trial of IV immunoglobulin treatment in acute optic neuritis.

OBJECTIVE: To investigate if IV immunoglobulin (IVIG) treatment in the acute phase of optic neuritis (ON) could improve visual outcome and reduce MRI disease activity 6 months after onset of ON. METHODS: Sixty-eight patients with ON were randomized within 4 weeks from onset of symptoms. Thirty-four patients were randomized to IVIG 0.4 g/kg body wt, and 34 patients were randomized to placebo. Infusions were given at days 0, 1, 2, 30, and 60. Contrast sensitivity, visual acuity, and color vision were measured at baseline and after 1 week, 1 month, and 6 months. Pattern reversal visual evoked potential studies and gadolinium-enhanced MRI were performed at baseline and after 1 and 6 months. Clinical relapses during follow-up were recorded. RESULTS: There was no difference in the primary outcome, contrast sensitivity after 6 months, between patients randomized to treatment with IVIG or placebo. In addition, there was no significant difference in the secondary outcome measures, improvement in the visual function measures and MRI, at any time during follow-up. At baseline, a significantly higher number of patients in the IVIG group had one or more enhancing lesions on MRI and IVIG-treated patients had a significantly higher number of enhancing lesions on MRI than patients treated with placebo. No difference was found in number of patients with one or more enhancing lesions or number of enhancing lesions in subsequent scans between treatment groups. Number of relapses was equal in the two treatment groups during follow-up. CONCLUSIONS: There was no effect of IV immunoglobulin (IVIG) on long-term visual function following acute optic neuritis, nor was there an effect of IVIG treatment in reducing latency on visual evoked potentials and thus preserving function of axons of the optic nerve.

Effect of celecoxib on Ca2+ movement and cell proliferation in human osteoblasts.

In human osteoblasts, the effect of the widely prescribed cyclooxygenase-2 inhibitor celecoxib on intracellular Ca(2+) concentrations ([Ca(2+)](i)) and cell proliferation was explored by using fura-2 and the tetrazolium assay, respectively. Celecoxib at concentrations greater than 1microM caused a rapid rise in [Ca(2+)](i) in a concentration-dependent manner ( EC 50= 10 microM). Celecoxib-induced [Ca(2+)](i) rise was reduced by 90% by removal of extracellular Ca(2+), and by 30% by l-type Ca(2+) channel blockers. Celecoxib-induced Mn(2+)-associated quench of intracellular fura-2 fluorescence also suggests that celecoxib-induced extracellular Ca(2+) influx. In Ca(2+)-free medium, thapsigargin, an inhibitor of the endoplasmic reticulum Ca(2+)-ATPase, caused a monophasic [Ca(2+)](i) rise, after which the increasing effect of celecoxib on [Ca(2+)](i) was greatly inhibited. Conversely, pretreatment with celecoxib to deplete intracellular Ca(2+) stores totally prevented thapsigargin from releasing more Ca(2+). U73122, an inhibitor of phoispholipase C, abolished histamine (an inositol 1,4,5-trisphosphate-dependent Ca(2+) mobilizer)-induced, but not celecoxib-induced, [Ca(2+)](i) rise. Pretreatment with phorbol 12-myristate 13-acetate and forskolin to activate protein kinase C and adenylate cyclase, respectively, partly inhibited celecoxib-induced [Ca(2+)](i) rise in Ca(2+)-containing medium. Separately, overnight treatment with 1-100microM celecoxib inhibited cell proliferation in a concentration-dependent manner. These findings suggest that in human osteoblasts, celecoxib increases [Ca(2+)](i) by stimulating extracellular Ca(2+) influx and also by causing intracellular Ca(2+) release from the endoplasmic reticulum via a phospholiase C-independent manner. Celecoxib may be cytotoxic at higher concentrations.

Human plasma-derived mannose-binding lectin: a phase I safety and pharmacokinetic study.

Mannose-binding lectin (MBL) is an important component of innate immunity that can bind to certain sugar residues on the surface of many types of pathogenic micro-organisms. On binding, MBL generates opsonic activity mainly through activation of the complement system. Genetically determined MBL deficiency is very common and can be associated with increased susceptibility to a variety of infections, especially in children and immunosuppressed individuals. The potential benefits of MBL reconstitution therapy therefore need to be evaluated. We have carried out a phase I safety and pharmacokinetic study on 20 MBL-deficient healthy adult volunteers. The MBL was prepared from plasma of nonremunerated, voluntary Danish donors tested and found negative for hepatitis B surface antigen, antibodies to human immunodeficiency virus (HIV) and hepatitis C virus. Each volunteer received a total of 18 mg of MBL in three 6 mg doses given intravenously, once weekly over a period of 3 weeks. The volunteers were closely monitored at the University Hospital in Reykjavik for 8 h after each infusion and daily thereafter for 5 days after each infusion. No adverse clinical or laboratory changes were observed in any of the 20 participants, and frequent measurements did not reveal any signs of infusion-associated complement activation. No antibodies to MBL, HIV or hepatitis viruses were observed 24 weeks after the last infusion. Serum MBL levels increased up to normal levels (1200-4500 ng/ml) immediately after each infusion, but the half-life of the infused MBL was highly variable, ranging from 18 to 115 h (mean 69.6). It is concluded that infusion of purified MBL as prepared by Statens Serum Institut (SSI) is safe. However, adults have to be given at least 6 mg twice or thrice weekly for maintaining protective MBL levels assumed to be about 1000 ng/ml.

Free insulin-like growth factor I: are we hunting a ghost?

During the last decade, there has been an increasing number of publications reporting concentrations of free dissociable insulin-like growth factor I (IGF-I) in serum or plasma. The goal for attempting to measure free IGF-I in a serum sample in vitro has been to obtain information about the bioactivity of IGF-I in target tissues, and thus relate a measurable parameter to biological responses such as longitudinal growth or glucose disappearance rate. In this review, the serum free IGF-I approach is placed into a physiological perspective. In addition, methodological aspects are discussed and suggestions for the validation of free IGF-I assays are presented.

Mutation of three critical amino acids of the N-terminal domain of IGF-binding protein-3 essential for high affinity IGF binding.

The N-terminal domain is conserved in all members of the IGF-binding protein superfamily. Most recently, studies have demonstrated the importance of an IGF-binding protein N-terminal hydrophobic pocket for IGF binding. To examine more critically the amino acids important for IGF binding within the full-length IGF-binding protein-3 protein while minimizing changes in the tertiary structure, we targeted residues I56, L80, and L81 within the proposed hydrophobic pocket for mutation. With a single change at these sites to the nonconserved glycine there was a notable decrease in binding. A greater reduction was seen when both L80 and L81 were substituted with glycine, and complete loss of affinity for IGF-I and IGF-II occurred when all three targeted amino acids were changed to glycine. Furthermore, the ability of the IGF-binding protein-3 mutants to inhibit IGF-I-stimulated phosphorylation of its receptor was a reflection of their affinity for IGF, with the lowest affinity mutants having the least inhibitory effect. These studies, thus, support the hypothesis that an N-terminal hydrophobic pocket is the primary site of high affinity binding of IGF to IGF-binding protein-3. The mutants provide a tool for future studies directed at IGF-dependent and IGF-independent actions of IGF-binding protein-3.