The insulin-like growth factor (IGF) binding site of bovine insulin-like growth factor binding protein-2 (bIGFBP-2) probed by iodination.

The insulin-like growth factor (IGF) binding site of bovine insulin-like growth factor binding protein 2 (bIGFBP-2) has been probed by chemical iodination. Tyrosyl residues of bIGFBP-2 were reacted by chloramine T-mediated iodination. The modification patterns of free bIGFBP-2 and bIGFBP-2 associated with insulin-like growth factor II (IGF-II) were compared by tryptic mapping using electrospray mass spectrometry and N-terminal sequencing. The presence of bound IGF-II resulted in protection of tyrosine at position 60 from iodination measured by the relative loss of tyrosine specific fluorescence and the incorporation of the radioisotope 125I. In addition, the pattern of iodine incorporation of bIGFBP-2 was not different whether IGF-I or IGF-II was the protective ligand. bIGFBP-2, when iodinated alone sustained a 8-fold loss of binding affinity for IGF-I and a 4-fold loss in binding affinity for IGF-II. In contrast, bIGFBP-2 iodinated while complexed with either IGF-I or IGF-II retained the same binding affinity for IGF-I or IGF-II as non-iodinated bIGFBP-2. We conclude that tyrosine 60 lies either in a region of bIGFBP-2 which directly interacts with both IGF-I and IGF-II or lies in a region of bIGFBP-2 which undergoes a conformational change that is important for IGF binding. Furthermore, iodination of tyrosine residues at positions 71, 98, 213, 226, and 269 has no detectable impact on binding of bIGFBP-2 to the IGFs.

The insulin-like growth factors (IGF-I and IGF-II) 1 are polypeptide mitogens which play diverse roles in development and metabolism across a wide range of vertebrate species (recently reviewed in Ref. 1). While the mitogenic actions of IGFs are mediated via specific type-I IGF receptor interactions, the bio-availability and localization of IGFs are considered to be largely determined by a family of proteins known as the insulin-like growth factor binding proteins (IGFBPs) (1)(2)(3). Detailed molecular structures of both IGF-I (4 -6) and IGF-II (7,8) have been determined. Moreover, some of the structural elements which are involved in IGFBP binding have been determined by a number of approaches which include chemical modification (9) and the generation of IGF-I and IGF-II mutants (10 -14). In contrast, there are no published structures for the IGFBPs and furthermore, the residues of IGFBPs which are important for IGF binding are largely unknown.
Most of the studies which address structural and functional aspects of the IGFBP family have not sought to directly identify the sites to which IGFs bind. Rather, researchers have focused mainly on issues such as IGFBP phosphorylation (15), IGFBP proteolysis (16 -19), and the functional significance of motifs identified in the primary sequences of IGFBPs, for example, the Arg-Gly-Asp sequence of IGFBP-1 (20).
There is evidence to suggest that both the N-and C-terminal regions of IGFBPs are important for complex formation with IGFs (21,22). Notably, the strongest regions of sequence homology are found in the N-and C-terminal cysteine-rich regions of these molecules. Included in this sequence homology is the conserved alignment of 16 cysteine residues, 10 in the N-terminal and 6 in the C-terminal third of IGFBPs. While all of the cysteine residues of IGFBPs are believed to participate in disulfide bonds (23), it remains to be shown whether all IGFBPs share the same disulfide-bridging pattern. In contrast, the respective middle regions of IGFBP sequences show greater heterogeneity.
We have used chemical iodination to characterize the binding interaction between IGFs and bIGFBP-2. We have achieved this by identifying tyrosine residues of bIGFBP-2 which are iodinated when this molecule is not bound to IGFs, but on the other hand are protected from iodination when either IGF-I or IGF-II is bound. The primary structure of bIGFBP-2 contains six tyrosine residues which are spread throughout the amino acid sequence of this molecule. This study has shown that an approximate 5-fold protection against modification specifically at Tyr 60 occurs when bIGFBP-2 is iodinated while in a complex with either IGF-I or IGF-II. Finally, iodination at Tyr 60 leads to a significant reduction in the affinity of IGFs for bIGFBP-2, indicating that Tyr 60 may be one of a number of residues which play an important role in the bIGFBP-2 binding reaction with the IGFs.

EXPERIMENTAL PROCEDURES
Materials-Recombinant bIGFBP-2 was transiently expressed in the COS-1 (ATCC: CRL 1650) monkey kidney cell line and purified from medium conditioned by the transfected cells by a method adapted from Szabo et al. (24). Receptor grade IGF-I and IGF-II were the kind gift of GroPep Pty. Ltd. (Adelaide, Australia). Radiolabeled 125 I-IGF-I and 125 I-IGF-II peptides were kindly provided by Spencer Knowles (CRC for Tissue Growth and Repair, Adelaide, Australia). Modified, sequencing grade trypsin was purchased from Boehringer Mannheim (North Ryde, New South Wales, Australia). HPLC columns were purchased from Brownlee Laboratories (Santa Clara, CA) and Pharmacia Pty. Ltd. (Sydney, Australia). Carrier-free Na 125 I was purchased from Amersham International (Sydney, Australia). Pre-siliconized tubes (Sorenson BioScience, Inc., Salt Lake City, UT) were used for reaction vessels and for the collection of fractions during chromatography. All HPLC was carried out using Waters 510 solvent pumps, a Waters 490 4-channel absorbance detector (Millipore-Waters, Lane Cove, N.S.W) and a Perkin Elmer LS4 fluorescence spectrometer (Scoresby, Victoria, Australia). The Waters Maxima software package was used to control solvent gradients and for data collection. HPLC-grade acetonitrile was purchased from Merck (Kilsyth, Victoria, Australia) and trifluoroacetic acid from Sigma-Aldridge (Castle Hill, N.S.W., Australia). All other reagents were analytical grade.
Optimizing Iodination Conditions-At the scale of 0.1 nmol of bIG-FBP-2, a series of iodinations were performed by the chloramine-T method (9). The amount of Na 125 I (specific radioactivity 50 nCi/nmol) was maintained at 4 nmol/reaction and the molar ratio of chloramine-T/Na 125 I per reaction was varied between 0.125 and 1.25. Two stock solutions were prepared, one containing bIGFBP-2 and the other containing both bIGFBP-2 and IGF-II in the following manner. In two separate tubes, 0.8 nmol of lyophilized bIGFBP-2 was resuspended in 350 l of iodination buffer (250 mM sodium phosphate, pH 6.5, 0.05% (v/v) Tween 20). Forty microliters of 10 mM acetic acid was added to the bIGFBP-2 stock solution and 3.2 nmol of IGF-II in 40 l of 10 mM acetic acid was added to the bIGFBP-2/IGF-II stock solution (resulting in a 4-fold molar ratio of IGF-II to bIGFBP-2). The vials were left at room temperature for 120 min to allow the complex between bIGFBP-2 and IGF-II to form. Then 36 nmol of NaI was added to each reaction mixture in 10 l of 5 mM NaOH. This was the equivalent of a 40-fold molar excess of NaI over bIGFBP-2. After mixing, both of the reaction mixtures were divided into 8 ϫ 50-l (0.1 nmol) aliquots of bIGFBP-2. Iodination was initiated by the addition of 5 l of the appropriate chloramine-T solution (prepared in a 10-fold serial dilution in 8 steps from 0.5 to 0.05 mM with water) followed by an incubation at room temperature for 45 s. Each reaction was quenched by the addition of 20 nmol of sodium metabisulfite in 10 l of water. The 16 reaction mixtures were stored at Ϫ20°C prior to HPLC gel filtration at acid pH.
HPLC Gel Filtration at Acid pH-Reaction products were thawed and acidified by the addition of 50 l of 2 M acetic acid to dissociate the IGF⅐bIGFBP-2 complex. Iodinated bIGFBP-2 was separated from free iodide and iodinated IGF (when present) by gel filtration on a TSK G3000SW Ultrogel HPLC column (7.5 ϫ 600 mm) equilibrated with a solution containing 150 mM NaCl, 10 mM HCl, and 0.05% (v/v) Tween 20. The column was eluted at a flow rate of 0.5 ml/min and 1-ml fractions were collected for ␥-counting. Protein was detected by absorbance at 215 nm and tyrosine fluorescence (excitation 275 nm, emission 305 nm). The amount of bIGFBP-2 recovered from the column was estimated by the integrated area of the 215-nm absorbance peak using the extinction coefficient ⑀ 0 ϭ 6.5 ϫ 10 6 volt⅐s⅐mg Ϫ1 (25) and standardized with accurately quantified IGF-II standards (GroPep Pty. Ltd.). The amount of iodine incorporated was estimated by the radioactivity associated with the bIGFBP-2 peak, adjusting for the known 125 I specific activity. The elution times of bIGFBP-2, IGF-II, and radioactive iodide were established by independent injection of the respective components.
Iodinations of bIGFBP-2 for Peptide Mapping-bIGFBP-2 (3.9 nmol) was dissolved in 255 l of iodination buffer and divided into three equal 1.3-nmol aliquots. Either IGF-I or IGF-II (5 nmol in 40 l of 10 mM acetic acid) was added to one aliquot, while 40 l of 10 mM acetic acid was added to the other two bIGFBP-2 solutions. The reaction mixtures were incubated at room temperature for 120 min to allow the complex between IGF and bIGFBP-2 to form. Then 52 nmol of NaI (specific radioactivity 2 nCi/nmol) was added to each of the three bIGFBP-2 solutions. Sodium metabisulfite, 200 nmol in 10 l of water, was added to one of the bIGFBP-2 solutions to prevent iodination in this tube. Following the addition of 40 nmol of chloramine-T in 10 l of water (corresponding to a chloramine-T/NaI molar ratio of 0.75) to each tube, the iodination reactions proceeded as described above. The iodinations of bIGFBP-2 alone and bIGFBP-2 in complex with IGF-I or IGF-II were then stopped by the addition of 200 nmol of sodium metabisulfite in 10 l of water. All of the reaction mixtures were acidified by the addition of 100 l of 4 M acetic acid and the bIGFBP-2 species were then purified by HPLC gel filtration as described above and the bIGFBP-2 peaks were lyophilized (Speed-Vac, Savant, Farmingdale, NY).
Tryptic Mapping of bIGFBP-2-Iodinated bIGFBP-2 species were subjected to tryptic peptide mapping as follows. Each of the lyophilized bIGFBP-2 species was denatured and reduced in 1 ml of S-carboxy-methylation buffer (8 M urea, 0.6 M guanidine HCl, 20 mM dithiothreitol, 50 mM Tris-HCl, pH 8.5) and the cysteine thiols were blocked by reaction with 69 mol of iodoacetic acid in 200 l of 1 M Tris-HCl, pH 8.5. The reaction was allowed to proceed in the dark at room temperature for 15 min and subsequently, the S-carboxymethylated bIGFBP-2 species was purified on a C4 reverse-phase HPLC cartridge (Brownlee Aquapore BU300 C4, 7 m particle size, 300-Å pore size, and dimensions of 2.1 ϫ 100 mm). The protein was eluted over a linear gradient of 10 -50% acetonitrile (v/v) in 0.1% (v/v) trifluoroacetic acid over 20 min, at a flow rate of 0.5 ml/min. The bIGFBP-2 peak was collected and concentrated to approximately 20 l by solvent evaporation under vacuum (Speed-Vac, Savant), then diluted to 1 ml with digestion buffer (250 mM Tris-HCl, pH 8.5, 20 mM CaCl 2 ). Digestion was initiated by the addition of 2 g of trypsin in 2 l of 10 mM acetic acid and the mixture was incubated at 37°C for 12 h after which a further 2 g of trypsin was added and the digestion continued for an additional 12 h. This corresponded to a final enzyme to substrate ratio of 1:10 (w/w) and a total reaction time of 24 h. Due to the relatively high enzyme to substrate ratio and extended reaction times, the self-digestion products of trypsin were also determined in a control incubation of trypsin in digestion buffer alone, under the same conditions. The digestions were stopped by the addition of 10 l of trifluoroacetic acid. The digested bIGFBP-2 peptides were separated by reverse-phase HPLC (using the same cartridge as described above) at 40°C with a linear gradient of acetonitrile from 0 to 50% (v/v) in 0.1% (v/v) trifluoroacetic acid over 50 min at 0.5 ml/min. The absorbance of the peptide backbone bonds was monitored at 215 nm, aromatic residues at 275 nm and tryptophan at 295 nm. Non-iodinated tyrosyl peptides were detected by their specific tyrosine fluorescence (excitation 275 nm, emission 305 nm) while iodinated tyrosyl peptides were detected at 295 nm and by detection of ␥-radiation as tyrosine iodination abolished tyrosine fluorescence (26). Eluate was collected in 0.25 min fractions starting with the first fluorescent peak.
Peptides which contained tyrosyl peptides were subjected to N-terminal sequencing. Edman degradation was carried out automatically on an Applied Biosystems model 470A gas-phase Sequencer with a 900A Control/Data Analysis Module. Mono-iodotyrosine and di-iodotyrosine were identified in the sequencer chromatograms (27) and were confirmed by their radioactivity. The separate tryptic peptides which encompass the six tyrosine residues of bIGFBP-2 (Tyr 60 , Tyr 71 , Tyr 98 , Tyr 213 , Tyr 226 , and Tyr 269 ) will be referred to subsequently as the peptides P1, P2, P3, P4, P5, and P6, respectively, and the iodinated derivatives will be referred to as Px* (mono-iodotyrosyl derivative) and Px** (di-iodotyrosyl derivative), where (x) refers to the peptides 1-6.
The positions of P1, P2, P3, P4, P5, P6 and the iodinated derivatives of these peptides in the elution profile of the tryptic map were established by N-terminal sequencing. The identities of tyrosine containing peptides were confirmed by electrospray mass spectroscopy (EMS) on a Perkin Elmer Psi-ex triple quadrupole mass spectrometer by Yoji Hawasaka at the Australian Research Council EMS unit, Adelaide.
Iodinations of bIGFBP-2 and IGF Binding Assays-bIGFBP-2 was modified as described in the large-scale iodinations except that the radioactive iodine isotope 125 I was omitted from the reaction. Again, bIGFBP-2 was iodinated alone and in a complex with both IGF-I and IGF-II. The iodinated bIGFBP-2 species were then purified by HPLC gel filtration as described above and the incorporation of iodine was confirmed by the loss of tyrosine-specific fluorescence during the chromatography. The relative affinities of the respective iodinated bIG-FBP-2 species for both 125 I-labeled IGF-I and IGF-II were determined by charcoal binding assay, as described previously (24).

RESULTS
We have used IGF-mediated protection against chemical modification of the tyrosine residues of bIGFBP-2 to investigate the association between bIGFBP-2 and IGF-I or IGF-II at the molecular level. The 6 tyrosine residues of bIGFBP-2 are distributed throughout the amino acid sequence and so enable the IGF-binding site to be localized to the N-terminal, middle, or C-terminal regions of this molecule. In this study, we have identified tyrosine residues which are protected from chloramine T-mediated iodination when IGFs are bound as outlined in Fig. 1.
Optimization of Iodination Conditions-When the concentration of NaI was held at 40-fold excess over bIGFBP-2, an increase in the molar ratio of chloramine-T/NaI in the iodination reaction resulted in an increase in the amount of iodine which became incorporated into bIGFBP-2 ( Fig. 2a). This figure also shows that when bIGFBP-2 was iodinated in a complex with IGF-II, the molar incorporation of iodine into bIGFBP-2 was reduced. A linear increase in iodine incorporation was observed for both IGF-protected bIGFBP-2 and unprotected bIGFBP-2 over the chloramine-T/NaI molar ratio range of 0.125 and 0.75. However, when this ratio was increased beyond 0.75, the subsequent increase in iodine incorporation into both protected and free bIGFBP-2 was less, on a molar basis. For IGF-protected bIGFBP-2 this change in the reactivity of the bIGFBP-2 tyrosine residue population was observed to occur when approximately 5 mol of iodine had been incorporated per bIGFBP-2 molecule compared with approximately 8 mol of iodine for bIGFBP-2 alone. These results suggest that the number of reactive tyrosine residues of bIGFBP-2 was reduced when IGF-II was bound.
The tyrosine fluorescence of the intact bIGFBP-2 molecule was a useful measure of the degree of iodination of the tyrosine residues. The chloramine T-dependent increase in the iodination of bIGFBP-2 was seen by a corresponding drop in tyrosinespecific fluorescence (Fig. 2b). As the ratio of chloramine-T/NaI per reaction was increased between 0.125 and 0.5, there was a rapid loss of bIGFBP-2 associated tyrosine fluorescence. However, negligible further losses in the tyrosine fluorescence of the iodinated bIGFBP-2 species were observed when this ratio was increased beyond 0.5. Fig. 2b shows that the tyrosine fluorescence ultimately dropped to approximately 10% of the unmodified molecule when bIGFBP-2 was iodinated alone. In contrast, bIGFBP-2 which had been iodinated while protected by IGF-II still retained 40% of the tyrosine fluorescence of the unmodified molecule. Both sets of data outlined above, which are representative examples of four independent experiments, suggest that ordinarily reactive tyrosine residues of bIGFBP-2 were protected from iodination when IGF-II was bound.
A 0.75 molar ratio of chloramine-T/NaI was chosen to generate iodinated bIGFBP-2 at a scale which enabled the modification pattern of iodinated IGF-associated and free bIGFBP-2 to be determined by tryptic mapping. These reaction conditions are indicated by arrows in Figs. 2, a and b.
Identification of IGF/IGFBP Binding Sites-Complete tryptic digestion of bIGFBP-2 would theoretically generate 32 frag-ments ranging in size from single amino acids up to a peptide of 45 residues (Fig. 3). Trypsin was chosen as the proteolytic mapping enzyme because in theory it could liberate each tyrosine residue in a discrete peptide fragment. Following tryptic digestion, peptides of bIGFBP-2 were separated by reversephase HPLC (Fig. 4). Autodigestion of trypsin (Fig. 4d) generated a background of 7 fluorescent peaks. Digestion of noniodinated bIGFBP-2 ( Fig. 4a) resulted in the generation of 9 tyrosine containing peptides in addition to those generated by trypsin autodigestion as detected by tyrosine fluorescence. The pattern of digestion shown in Fig. 4 is representative of the many experiments performed (data not shown). N-terminal sequence analysis and EMS identified the fluorescent tryptic peptides of bIGFBP-2 to be (in order of elution) P1, P6, P2, and P3, ϩP6, P4ϩP5ϩ, P4ϩP5, ϩP4ϩP5, where ϩ indicates an intact trypsin-sensitive bond, as defined in Fig. 3. The tyrosyl peptides P4, P5, and P6 were all identified as partial digestion fragments while the peptides P2 and P3 were not resolved by this chromatography. The mass of the peptide which was identified as P3 by N-terminal sequencing was between 15 and 16 mass units greater than the predicted mass. One possible explanation for this observation is the previously documented (28,29) oxidation of methionine by chloramine-T. The EMS and N-terminal sequencing results are summarized in Table Ia.
The loss of tyrosine fluorescence due to iodination can clearly be seen when the tryptic map of non-iodinated bIGFBP-2 ( Fig.  4a) is compared with the tryptic tyrosine fluorescence maps of bIGFBP-2 iodinated ether alone (Fig. 4b) or complexed with IGF-II (Fig. 4c). Table Ib  ual tyrosine fluorescence of bIGFBP-2 which was iodinated in a complex with IGF-II or alone. An additional fluorescent peak (A) was present in the iodinated bIGFBP-2 maps but was absent in the non-iodinated bIGFBP-2 map. However, this peak was most probably an artifact of the iodination reaction as both N-terminal sequence and EMS analysis showed conclusively that this peak was not a peptide. The order of tyrosine reactivity when calculated in terms of the residual tyrosine fluorescence (Table Ib) changed from P6 Ͼ P1 Ͼ P2 and P3 Ͼ P4ϩP5 when bIGFBP-2 was modified alone to P2 and P3 Ͼ P6 Ͼ P4ϩP5 Ͼ P1 when bIGFBP-2 was modified in a complex with IGF-II. However, the major difference in the iodination pattern of both free bIGFBP-2 and IGF-II associated bIGFBP-2 was the large residual tyrosine fluorescence of the peptide P1 in the map of IGF-II associated bIGFBP-2 (Fig. 4c). When quantified, the residual fluorescence of this peptide was 94% of the tyrosine specific fluorescence yielded by the same peptide isolated from non-iodinated bIGFBP-2 (Fig. 4a) and was 5.7fold more intense than for the same peptide from bIGFBP-2 which had been iodinated free of IGF-II (Fig. 4b). The amount of non-iodinated P6 peptide was very low in all of the iodinated bIGFBP-2 tryptic maps which indicated that this residue was freely available for iodination, although complex formation with IGF-II afforded a 2.5-fold protection against modification at this site (Fig. 4c) in comparison to bIGFBP-2 modified alone (Fig. 4b). The non-iodinated tyrosyl peptides P2 and P3 were 2.8 times more abundant in the peptide map of bIGFBP-2 which had been modified alone (Fig. 4b) compared with bIG-FBP-2 which was iodinated with IGF-II bound (Fig. 4c). The additive residual tyrosine specific fluorescence of partially digested peptides which contained P4 and P5 were similar (Fig.  4, b and c) regardless of IGF association prior to the iodination reaction. The tyrosine fluorescence associated with P4ϩP5 peptides was significant, accounting for 36% of the total remaining fluorescence of the non-iodinated tyrosyl peptides.
The tryptic peptides of bIGFBP-2 which had incorporated iodine were also directly identified by their 125 I radioactivity. Fig. 5 shows an alignment of the tryptic map fractions of bIGFBP-2 iodinated free (a) and bIGFBP-2 iodinated as a complex with IGF-II (b) which contain 125 I radioactivity. Fractions which contained iodinated peptides were characterized by both N-terminal sequencing and EMS as summarized in Table IIa. Tyrosine containing peptides were the only peptides of bIG-FBP-2 which were found to incorporate iodine under our reaction conditions. As shown in Fig. 5a, the order of elution of iodinated tyrosyl peptides of bIGFBP-2 was determined to be P1*, P1**, P6*, P6**, and P3*; and P3**, P2*, P2**, ϩP6**, (ϩP4ϩP5)**; and (P4ϩP5)*, (P4ϩP5)**, and (P4ϩP5)***, where parentheses indicate that the locations of the iodine atoms in these peptides were not characterized. The incorporation of iodine was observed to increase the hydrophobicity of tyrosine containing peptides as can be seen in the longer retention times of modified peptides ( Table IIa). The exceptions to this were the mono-iodotyrosyl (*) and di-iodotyrosyl (**) derivatives of P3, which did not exhibit significantly increased retention times, presumably due to the larger size of this tyrosyl peptide (45 residues). Edman degradation of the iodinated, partially digested derivatives of the peptides P4 and P5 indicated that the tyrosine residue in P4 (Tyr 213 ) and not P5 (Tyr 226 ) was the major site of iodine incorporation in these peptides.
It is immediately evident on comparison of panels a and b in Fig. 5 that the peptide P1 was shielded from iodination when IGF-II was bound. Quantification of the iodinated peptides helped to decipher the impact of IGF-association on the modification of other bIGFBP-2 tyrosyl peptides. The extent of iodine incorporation at each tyrosine residue could be calculated as a fraction of the total radioactivity incorporated into bIG-FBP-2 thereby allowing the relative reactivity of each tyrosine residue to be compared, as summarized in Table IIb. These data also show that the association of IGF-II with bIGFBP-2 did not significantly reduce the incorporation of iodine into peptides P2 and P3, or P6. In fact, there was a 1.5-fold increase in the modification of the peptides P2, P3, and P6 when IGF was associated with bIGFBP-2. The tyrosine-specific fluorescence data described above also showed that an increase in the degree of modification of the peptides P2 and P3 occurred when bIGFBP-2 was iodinated in a complex with IGF. A slight protection against modification in either P4 or P5 was observed when bIGFBP-2 was iodinated in a complex with IGF-II (Table  IIb). The peptide P1 accounted for 15% of the total iodine incorporation when bIGFBP-2 was iodinated alone. In contrast, the labeling of this peptide was reduced by a factor of 4.3 to only 3.5% of the total iodine incorporation when bIGFBP-2 was iodinated in a complex with IGF. This result is also in strong agreement with the tyrosine fluorescence data reported above.
Presented in Fig. 5b is the profile of 125 I radioactivity in tryptic peptides of bIGFBP-2 iodinated while in a complex with IGF-II. The same modification pattern was observed when the protective ligand was IGF-I (data not shown).
Determination of the Biological Activity of Iodinated Forms of bIGFBP-2-The abilities of non-radioactive iodinated bIG-FBP-2 species to bind radiolabeled IGF-I or IGF-II were compared with each other and non-iodinated bIGFBP-2. Fig. 6 shows the binding of radiolabeled IGF-I (a) and IGF-II (b), respectively, by increasing amounts of non-iodinated and iodinated bIGFBP-2 species. All of the iodinated bIGFBP-2 species FIG. 5. The tryptic map of iodinated bIGFBP-2, 125 I radioactivity. Shown are the radioactivity chromatograms of bIGFBP-2 iodinated free (a) and associated with IGF-II (b). Peaks which were shown by N-terminal sequencing and mass spectroscopy to correspond to modified tyrosyl peptides of bIGFBP-2 are identified in chromatogram a. Peptides with both one (*) and two (**) iodine atoms were identified. Iodinated derivatives of tyrosyl peptides P4, P5, and P6 were also identified as partial digestion products as indicated (ϩ).

TABLE I
A summary of bIGFBP-2 tryptic peptides identified by tyrosine specific fluorescence a, bIGFBP-2 tryptic peptides which contained tyrosyl residues were detected by their tyrosine-specific fluorescence (excitation 275 nm, emission 305 nm) following reverse-phase HPLC. Tyrosyl peptides (P1-P6 according to their position in the bIGFBP-2 primary structure) were characterized by both Edman degradation and EMS. The peptides P2 and P3 were not resolved. Tyrosine containing partial digestion products in the C-terminal half of bIGFBP-2 were identified, the uncleaved trypsin sensitive bonds are shown (ϩ). A period (.) indicates complete sequencing of a peptide to the C terminus. b, the residual fluorescence of each tyrosyl peptides after iodination represents the fraction of non-iodinated tyrosine in the maps of bIGFBP-2 iodinated in a complex with IGF-II or free. Therefore, the fold-protection which IGF-II binding afforded to each tyrosyl residue of bIGFBP-2 is the ratio of the residual tyrosine fluorescence of peptides from bIGFBP-2 iodinated with IGF-II bound over the residual tyrosine fluorescence of peptides from bIGFBP-2 iodinated free. were observed to bind the same maximal percentage of radiolabeled IGF-I or IGF-II tracer as non-iodinated bIGFBP-2. However, when bIGFBP-2 was iodinated in the absence of IGF, the half-maximal binding concentration of this molecule was increased from 2.3 to 17.9 ng for IGF-I and 0.8 to 3.2 ng for IGF-II, when compared with the half-maximal binding concentration of non-iodinated bIGFBP-2. In contrast, bIGFBP-2, iodinated while in complex with either IGF-I or IGF-II, showed half-maximal binding equal to non-iodinated bIGFBP-2.

DISCUSSION
Iodination is a very sensitive chemical modification technique which, under mildly acidic conditions, specifically labels tyrosine residues (30). Iodination has been used in this study as a structural and functional probe to investigate the interaction between bIGFBP-2 and its ligands IGF-I and IGF-II, at a molecular level. Previously, this approach has been successfully used to identify tyrosine residues of IGF-I which are important for association with the type-I IGF receptor (31,32) and tyrosine residues of both IGF-I and IGF-II which are important for association with bIGFBP-2 (9). The primary structure of bIGFBP-2 contains six tyrosine residues. The tyrosine residues Tyr 60 and Tyr 71 are located in the cysteine-rich Nterminal region, Tyr 98 is located in the middle region, and the three remaining tyrosine residues Tyr 213 , Tyr 226 , and Tyr 269 lie in the cysteine-rich C-terminal region of the molecule (33). Iodine incorporation into the bIGFBP-2 molecule was measured both directly by 125 I associated ␥ radiation counting and by monitoring the loss of tyrosine specific fluorescence (26).
In this study, a range of iodination conditions of both free and IGF-associated bIGFBP-2 were investigated, all of which clearly showed that the presence of IGF-I or IGF-II bound to bIGFBP-2 directly modified iodination of bIGFBP-2 tyrosine residues. It could be argued that the additional tyrosine residues of IGF-II which are present in the "complex" iodination reaction simply titrate the reactive iodous ion and therefore nonspecifically reduce bIGFBP-2 modification. However, iodinations were performed at reaction conditions in which the concentration of the reactive iodous ion was sufficiently high to label all of the available tyrosine residues of bIGFBP-2, irrespective of the presence of IGF-II. This was evident when further incorporation of iodine into bIGFBP-2 did not result in a further loss of the tyrosine fluorescence associated with this molecule (Fig. 2b). Whereas bIGFBP-2 had lost 90% of tyrosine fluorescence before a stable fluorescence minimum was reached, IGF-II associated bIGFBP-2 had only lost a maximum of 60% of its original tyrosine specific fluorescence. Therefore these results cannot be explained by the presence of additional tyrosine residues of IGF-II titrating the available iodous ion thus leading to an apparent drop in bIGFBP-2 labeling.
Tryptic mapping established that all of the tyrosine residues of bIGFBP-2 (Tyr 60 , Tyr 71 , Tyr 98 , Tyr 213 , Tyr 226 , and Tyr 269 in peptides P1, P2, P3, P4, P5, and P6, respectively) were readily iodinated when modified alone. Both mono-and di-iodotyrosyl derivatives of each tyrosyl peptide, with the exception of P5 were identified. We therefore conclude that all of the tyrosine residues of bIGFBP-2 with the exception of Tyr 226 are solvent accessible and thus are potential sites for interaction with IGF-I or IGF-II.
Indeed, complex formation with IGF-II changed the availability of tyrosine residues for iodination, predominantly in the N-terminal region of bIGFBP-2. The major difference in the iodination patterns of IGF-protected and free bIGFBP-2 was iodination at Tyr 60 , which was reduced by a factor of 5 when bIGFBP-2 was iodinated in a complex with IGF-II (Fig. 4, b and  c, and Fig. 5, a and b). In contrast, Tyr 71 was more reactive when IGF-II was bound to bIGFBP-2 compared with bIGFBP-2 iodinated free. Furthermore, these trends in tyrosine reactivity remained consistent when the iodinations were repeated at  125 I radioactivity a, iodinated tryptic peptides of bIGFBP-2 were separated by reverse-phase HPLC and then identified by their I radioactivity. Tyrosyl peptides (designated P1-P6 according to their position in the bIGFBP-2 primary structure) were then characterised by both Edman degradation and EMS. Peptides which had incorporated one or two iodine atoms are indicated by * and **, respectively, whereas ,22 indicates the presence of an intact trypsin sensitive bond. A period at the end of a peptide indicates that it was sequenced to the C terminus. b, the radioactivity of each peptide is expressed as a percentage of the total radioactivity recovered in the peptide map. The radioactivity of related peptides (e.g. P1* and P1**) or peptides which were not resolved (e.g. P6** and P3**) are considered together. The fold-protection is the ratio of the percent iodine incorporation into given tyrosyl peptides of bIGFBP-2, when iodinated free divided by when iodinated with IGF-II bound. chloramine-T/NaI molar ratios of 0.125, 0.375, and 1.25 (data not shown). There are two likely explanations for IGF-II mediated protection of bIGFBP-2 at Tyr 60 . First, this residue may lie in a region of bIGFBP-2 over which IGF-II binds, thus preventing modification at this site. Second, Tyr 60 may become less accessible for iodination following a conformational rearrangement of the bIGFBP-2 tertiary structure on IGF-II binding. In contrast, we conclude that the increased availability of Tyr 71 for iodination when IGF-II was bound may be due a change in the conformation of this residue which rendered it more reactive.
The tyrosine labeling pattern of bIGFBP-2 was the same regardless of whether IGF-I or IGF-II was used as the protective ligand. When the binding ability of non-iodinated bIG-FBP-2 was compared with iodinated unprotected bIGFBP-2 and iodinated bIGFBP-2 protected by IGF-I or IGF-II, we observed that iodination did not reduce the IGF-binding ability of bIGFBP-2 if this modification was carried out while IGF was bound. This suggests that the residues of bIGFBP-2 which were available for iodination when IGF was bound (Tyr 71 , Tyr 98 , Tyr 213 , and Tyr 269 ) are not involved in the binding interaction between IGF and bIGFBP-2. Therefore it is likely that Tyr 71 , Tyr 98 , Tyr 213 , and Tyr 269 all occur in regions of the IGFBP structure which does not closely interact with IGF in an exclu-sive manner. In contrast, bIGFBP-2 which was iodinated alone sustained a reduction in binding affinity for both IGF-I and IGF-II as a result of this modification (Fig. 6, a and b). Therefore, not only was Tyr 60 rendered less accessible for iodination by IGF-binding, iodination at this site resulted in the loss of binding affinity for both IGFs. We conclude therefore that Tyr 60 lies in a functionally significant region of bIGFBP-2. Despite protection from iodination at Tyr 60 , the binding capacity of all iodinated bIGFBP-2 species were the same (Fig. 6, a  and b). This suggests that Tyr 60 and the other reactive tyrosine residues of bIGFBP-2 do not play an essential role in the bIGFBP-2/IGF interaction.
Complex formation with both IGF-I and IGF-II produced the same tyrosine protection pattern when bIGFBP-2 was iodinated. Yet when bIGFBP-2 was iodinated alone, the loss in binding affinity for IGF-I was 2-fold more severe than that observed for IGF-II. This result was of interest considering the difference in relative affinity of bIGFBP-2 for IGF-II over IGF-I, which has been reported to be 4-fold (11) and as high as 20-fold (24,34). Therefore, Tyr 60 may influence the binding of IGF-I more strongly than IGF-II. Our results therefore support the notion that bIGFBP-2 possesses distinct but overlapping binding sites for IGF-I and IGF-II.
Tyr 60 is located in the cysteine rich N-terminal region of bIGFBP-2, a region in which all IGFBP family members exhibit strong sequence homology (35). Alignment of the known amino acid sequences of the IGFBP family shows that there is high degree of homology in the vicinity of Tyr 60 , which is consistent with conserved structure or function. Furthermore, tyrosine residues are found at the position equivalent to Tyr 60 of bIG-FBP-2 in all IGFBP family members with the sole exceptions being the IGFBP-1 family members, where an alanine exists at this position (35). As an alanine residue exists at this site in IGFBP-1, it is not likely that Tyr 60 is an exclusive determinant of IGF binding. Furthermore, when Ala 60 of IGFBP-1 was mutated to Thr, there was no gross loss of IGF binding affinity which could be detected by ligand blotting (36). Therefore, the other highly conserved residues adjacent to Tyr 60 , such as Thr 61 may also be important for the interaction between IGFBPs and IGFs.
The strong sequence conservation in both the cysteine rich N-and C-terminal regions of IGFBPs and the IGF-binding properties of proteolytic fragments which account for little more than either the N-or C-terminal regions of these molecules has lead to the prediction that there are at least two independent domains in the IGFBP tertiary structure. Furthermore, it has been predicted that both the N-and C-terminal domains act in a cooperative fashion to bind IGF (35). Previous studies have been carried out to identify the minimal binding domain of IGFBP-1. These include deletion analysis of the N terminus (36) and C terminus (37) as well as random mutagenesis within the N-and C-terminal regions of this molecule (36,37). The mutation of Cys to Tyr at position 38 in the N-terminal region of IGFBP-1 resulted in the loss of IGF binding ability (36). However, this mutant was also observed to aggregate, presumably through non-native disulfide formation. Therefore it is likely that the loss of IGF binding was due largely to the disruption of the IGFBP-1 structure. Similarly, deletion mutants of IGFBP-1 with abolished IGF binding activity may have resulted from a similar disruption of structure through the loss of native disulfide bonds in the putative Nand C-terminal domains (36,37).
In conclusion, this study has identified a region located in the putative N-terminal domain of bIGFBP-2 which includes Tyr 60 , that is important for the association of this molecule with IGF-II and IGF-I. To minimize the structural disruption which FIG. 6. The IGF binding assays of iodinated and non-iodinated bIGFBP-2 species. Increasing amounts of non-iodinated bIGFBP-2 (É); IGF-I protected, iodinated bIGFBP-2 (Ⅺ); IGF-II protected, iodinated bIGFBP-2 (E), and unprotected, iodinated bIGFBP-2 (å) were incubated with either radiolabeled IGF-I (15,000 cpm) (a) or radiolabeled IGF-II (6000 cpm) (b). The unbound IGF was separated from the bIGFBP-2 bound (24) and then quantified by measuring the 125 I radioactivity. These results are the mean of triplicate determinations and error bars show standard deviation values when greater than the size of the symbols. may be introduced by the random substitution of amino acids, bIGFBP-2 was iodinated both free and in complex with IGF-I or IGF-II as a probe for the binding interaction. Iodination at Tyr 60 and not Tyr 71 , Tyr 98 , Tyr 213 , or Tyr 269 lead to an 8-and 4-fold reduction in affinity for IGF-I and IGF-II, respectively. These results are further proof to suggest that major determinants of IGF binding reside in the N-terminal region of IGFBPs and they also raise implications for the generation of iodinated IGFBPs for radioimmunoassay and ligand blot analysis. We therefore recommend that IGFBPs should be iodinated while in a complex with IGF to subsequently retain their full affinity for IGFs.