Nerve Growth Factor Receptor from Rabbit Sympathetic Ganglia Membranes RELATIONSHIP BETWEEN SUBFORMS*

The receptor for nerve growth factor (NGF) was purified from Triton X-100 extracts of sympathetic ganglia membranes by affinity chromatography on NGF-Sepharose. Elution of purified receptor was ac-complished at pH 5 in the presence of 1 M NaCl. Sodium dodecyl sulfate gel electrophoresis of the purified iodinated receptor showed three major bands at M, = 126,000, M, = 105,000, and M, = 81,000. Affinity labeling of the purified receptor using lZ6I-NGF and the photoreactive agent N-hydroxysuccinimidyl-p-azi-dobenzoate resulted in two major cross-linked com- plexes corresponding to M, = 135,000 and M, = 110,000. This labeling pattern is similar to that ob- served with sympathetic ganglia membranes (Ma-ssague, J., Guillette, B. J., Czech, M. p., Morgan, c . J., and Bradshaw, R. A. (1981) J. Biol. Chem. 256, 9419-9424) and indicates that these two forms do not arise from the cross-linking procedure. Reaction of the photoaffinity labeled NGF receptors with increasing amounts of trypsin resulted in a progressive decrease in the high molecular weight complex with a concomi- tant increase in the low molecular weight form. When the larger complex was isolated by electroelution from a sodium dodecyl sulfate gel and treated with trypsin, a species corresponding to M, = 100,000 was gener-ated. These observations are best explained by a pre- cursor-product relationship

Some progress has been made in the molecular characterization of NGF receptors. Hydrodynamic measurements of the detergent-solubilized NGF receptor of superior cervical ganglia first revealed an asymmetric membrane protein of approximately 135 kDa (14). Covalent cross-linking of the NGF receptor to lZ51-NGF in the same tissue (15), as well as sensory neurons (16,17) and PC12 cells (17), yielded two major complexes corresponding to molecular masses of 100-1 10 and 143 kDa (neurons) or 158 kDa (PC12 cells). Western blots of mouse melanoma membrane proteins identified two NGF receptor species of 100 and 130 kDa (18). In contrast, chemical cross-linking (19,20), characterization of the purified receptor (21), and immunoprecipitation with monoclonal antibodies (22) revealed a single major NGF receptor species of 75-85 kDa in human melanoma cells (A875). However, A875 melanoma cells are not responsive to NGF, in contrast to the other tissues that have been examined.
Steady-state binding of T -N G F to sensory neurons (9), sympathetic neurons (7), and PC12 cells (11) demonstrated the presence of two classes of NGF receptors differing in their affinity for NGF. In PC12 cells, the two NGF receptor populations differ in their rates of dissociation of bound NGF, trypsin sensitivity, and solubility in Triton X-100 (23). Hosang and Shooter (17) identified the 158-kDa-cross-linked species of PC12 cells as the high affinity, trypsin-resistant form of the NGF receptor (from which bound NGF dissociates slowly) and the 100-kDa species as the low affinity, trypsinresistant NGF receptor (showing fast dissociation). This identification is supported by the observation that cells displaying both kinetic classes of NGF receptors show both cross-linked complexes (15,17), whereas cells displaying only low affinity receptors show little or none of the 143-158-kDa cross-linked complex (17,19).
The relationship between the two major NGF receptor species of neurons and PC12 cells is speculative. Massague et al. (15) suggested a conversion of the 143-kDa labeled receptor into the 112-kDa species by limited proteolysis, based on peptide mapping experiments. On the other hand, Hosang and Shooter (17) observed that the relative amount of labeling of the two complexes in PC12 cells did not vary in the presence of protease inhibitors and suggested that the 158-kDa species was not converted to the 100-kDa species, at least by a common protease.
There are several possible relationships between the two NGF receptor species: 1) the larger species could be the result of cross-linking of the smaller receptor with another protein; 2) the smaller entity could be a proteolytic product of the larger one; or 3) the two species could be genetically unrelated. Since the identification of the receptor subspecies has been dependent on covalent cross-linking experiments, it is also possible that the differences are introduced during this modification. The experiments presented in this report were designed to test these possibilities and to ascertain the nature of the receptor responsible for biological activity. We report the purification of NGF receptor from superior cervical ganglia and the biochemical characterization of the purified receptor by direct iodination, photoaffinity labeling, and limited proteolytic digestion. Our results strongly suggest that both receptor forms occur in situ and that they share a precursorproduct relationship in keeping with previous observations (15).

MATERIALS AND METHODS
Reagents-NGF was purified from adult male mouse submandibular glands, obtained from Biolab, St. Paul, MN, according to Bocchini and Angeletti (24). '"I-NGF was prepared by a modification (15) of the Bolton-Hunter technique (25). Superior cervical ganglia were obtained freshly frozen from Pel-Freez. Cyanogen bromideactivated Sepharose was purchased from Pharmacia. NGF was coupled to the activated Sepharose according to the manufacturer's instructions. '251-Bolton-Hunter reagent was obtained from New England Nuclear. Na'? was from ICN. High purity, low peroxide Triton X-100 and N-hydroxysuccinimidyl-p-azidobenzoate (HSAB) were from Pierce. Phenylmethanesulfonyl fluoride, pepstatin, leupeptin, chloramine T, 8-octyl glucoside, poly-L-aspartic acid, and diphenylcarbamyl chloride-treated trypsin (Type XI) were from Sigma. Electrophoresis reagents were from Bio-Rad. All other reagents were of the highest quality available.
Purification of NGF Receptors-Microsomes were prepared from superior cervical ganglia as described previously (5). A mixture of protease inhibitors, 1 mM phenylmethanesulfonyl fluoride, 1 mM EDTA, 0.1 mM pepstatin, and 0.1 mM leupeptin, was present throughout the microsome preparation. T o solubilize NGF receptors, the microsomes, at a protein concentration of 1 mg/ml in HEPES (10 mM) saline butter, pH 7.5 (HBS), with protease inhibitors, were incubated with 1% (v//v) Triton X-100 for 1 h a t 0 "C followed by centrifugation at 100,000 x g for 1 h a t 4 "C. The supernatant fraction was diluted to adjust the final detergent concentration to 0.25%. NGF receptors were purified from the supernatant fraction by affinity chromatography on NGF-Sepharose. The Triton X-100 extract was incubated in batches with NGF-Sepharose for 12 h a t 4 'C with gentle stirring on an orbital shaker (for an extract corresponding to 20-200 superior cervical ganglia, 1 g of NGF-Sepharose was used). The NGF-Elutton Volume ( m l ) FIG. 1. Affinity chromatography of rabbit sympathetic ganglia NGF receptor on an NGF-Sepharose column. Superior cervical ganglia membranes were solubilized with Triton X-100 and applied to the column of NGF-Sepharose as described under "Material and Methods." The column was washed with HBS containing 0.1% Triton X-I00 and a mixture of protease inhibitors (see text for details). At A , the column was washed with 1 M NaCl in the initial buffer; a t R, the column was briefly eluted with HBS containing 0.1% Triton X-100 and 1 mg/ml NGF (A) or 1 mg/ml cytochrome c (0) or with no added protein (0); at C, the column was washed with HBS containing 1% 8-octyl glucoside; a t D, the column was eluted with 50 mM sodium acetate, pH 5.0, containing 1 M NaCl and 1% 8-octyl glucoside. Fractions (0.5 ml) were collected throughout the chromatography and assayed for NGF binding as described in the text.
Sepharose beads were poured into a small column and washed successively with 2 column volumes of 0.1% Triton X-100 in HBS with protease inhibitors, 6 column volumes of 1 M NaCl in 0.1% Triton X-100 in HBS, and finally 10 column volumes of 1% 8-octyl glucoside in HBS. The purpose of the last wash was to eliminate the Triton X-100 and the protease inhibitors which interfere with the subsequent iodination of the purified receptors. The receptors were desorbed from NGF-Sepharose with a sodium acetate (50 mM) buffer, pH 5, containing 1 M NaCl and 1% P-octyl glucoside. Fractions, 0.5 ml, were collected and immediately neutralized with 0.2 M NaOH. Aliquots (10-100 pl) of the fractions were assayed for '"I-NGF binding activity according to Costrini and Bradshaw (26). The NGF receptorcontaining fractions were pooled and kept a t -70 "C. In order to check the specificity of the affinity matrix for NGF receptor, the affinity chromatography was performed as described above except that an additional wash with buffer containing unlabeled NGF was introduced prior to the desorption step between the second wash (1 M NaCl in buffer, pH 7.5) and the last one (removal of Triton X-100). The buffer used consisted of 3-5 ml of HBS containing I mg/ ml NGF (or cytochrome c ) and 0.1% Triton X-100. lodimtion of NGF Receptors-Aliquots, 50 pl, of NGF-Sepharose eluate containing "'1-NGF binding activity were incubated with 100 pCi of Na"'1 and 200 pg of chloramine T for 5 min. The reaction was quenched with 500 pg of potassium metabisulfite and 500 pg of sodium iodide. T o eliminate excess reagents, the reaction mixture was diluted with 1 ml of HBS containing 1% 8-octyl glucoside and 0.2% poly-^aspartic acid (15,000 average molecular weight) as a carrier, and concentrated to 50 pl with a Centricon-30 microconcentrator (Amicon).
Photoaffinity Labeling Protocol-The purified NGF receptors were labeled with 1251-NGF using the photoactivatable cross-linker HSAB by a modification of the method of Massague et al. (15). Aliquots of the eluate of the NGF-Sepharose column, 10-50 pl, were incubated for 60 min at 23 "C with 0.8 nM lZ5I-NGF in the presence or absence of 0.4 p~ unlabeled NGF in a total volume of 200 pl (dilutions up to volume were made with HBS containing 0.1% Triton X-100). At the end of the incubation period, HSAB, freshly dissolved in dimethyl sulfoxide, was added a t 50 p~ final concentration. The reaction mixture was incubated for 4 min on ice in the dark and then transferred to a quartz cuvette and photolyzed for 8 min a t 4 "C using a short wave-length lamp (UV mineralight lamp UVS-11). The reaction was stopped by adding 10 volumes of 10 mM Tris, p H 7.0, with 0.1% Triton X-100. Excess reagents were in part eliminated by concentrating the reaction mixture to 50 pl with a Centricon-30, diluting the concentrate to 1 ml with 0.28% SDS and concentrating again. The final concentrate was retained for electrophoresis. When samples of photoaffinity cross-linked receptor were treated with trypsin, the above protocol was followed up to the first concentration. Digestion with a suitable trypsin concentration was performed on aliquots of the first concentrate at 23 "C for 30 min. The reaction was quenched by addition of 1% SDS and 50 mM dithiothreitol followed by boiling for 1 min. Samples treated in this fashion were directly analyzed by SDS electrophoresis.
Sodium Dodecyl Sulfate Polyacrylamide Gel Electrophoresis-Iodinated and photoaffinity labeled NGF receptors were diluted with an equal volume of sample buffer (20 mM Tris, pH 6.8, 2% SDS, 2 mM dithiothreitol, 15% sucrose, 0.01% bromphenol blue) and boiled for 1 min prior to electrophoresis. Gel electrophoresis was performed according to Laemmli (27), using a 3% stacking gel and a 6% separating gel. After electrophoresis, the gels were stained in 0.025% Coomassie Blue, 25% 2-propanol, 10% acetic acid, and destained in 10% acetic acid. Autoradiograms were obtained from the dried gels after exposure to Kodak X-Omat AR film, using a DuPont Lightning Plus enhancing Electroelution of High Molecular Weight Affinity Labeled NGF Receptor-After electrophoresis of the photoaffinity labeled NGF receptors, the polyacrylamide SDS gel was briefly stained and the area of the gel where the 130-135-kDa NGF receptor species was expected to be found was cut out of the gel and sliced into small pieces. Electrophoretic elution of the photolabeled NGF receptor species was performed using the electrophoretic tank and the elution cell designed by Hunkapiller et al. (28). The gel pieces were transferred to the bottom of an elution cell on top of a membrane with a 50,000 molecular weight cutoff and covered with 2% SDS in 0.4 M sodium bicarbonate. The mixture was carefully overlayered and the elution cell filled with running buffer (0.2% SDS in 0.01 M sodium bicarbonate). Electroelution was performed at 70 V for 24 h, during which the running buffer was recycled. The receptor species was recovered by precipitation with 9 volumes of ice-cold methanol after addition of a few micrograms of bovine serum albumin as carrier Trypsin,pg/ml protein. The pellet was dissolved in 0.1% Triton X-100 in HBS. Half of the receptor solution was submitted to trypsin digestion and the other half was reserved as a control.

Purification of NGF Receptor from Sympathetic
Ganglia-NGF receptors were purified from Triton X-100 extracts of rabbit superior cervical ganglia microsomes. During the purification, the receptor was detected with a soluble '*'I-NGF binding assay making use of the differential precipitation of the NGF-receptor complex and unbound NGF with polyethylene glycol (26). Binding of the Triton X-100-solubilized NGF receptors to NGF-Sepharose was performed in batches thus allowing 90% of the binding activity of the Triton extracts to be adsorbed (data not shown). After adsorption, the beads of NGF-Sepharose were packed in a column and eluted (Fig. 1). No binding activity was released from the column by the initial wash with the high salt, pH 7.5 buffer, indicating that the receptor was tightly bound. This step allowed the removal of most contaminating proteins nonspecifically adsorbed to NGF-Sepharose. The column was then extensively washed with a p-octyl glucoside buffer in order to remove the Triton X-100. This detergent exchange was necessary for the later iodination of the eluted NGF receptors (see below). NGF receptor was desorbed with a sodium acetate buffer, pH 5, containing 1 M NaCl and 1% P-octyl glucoside. Approximately 30% of the binding activity of the Triton X-100 extracts was recovered in the sodium acetate eluate of the NGF-Sepharose. The overall loss of binding activity during the purification procedure is probably due in part to an effectively irreversible association of some of the NGF receptors with the NGF-Sepharose and in part to the progressive loss of binding activity of the receptors observed a t 4 "C.
NGF receptor was totally desorbed by a wash containing NGF, as demonstrated by the absence of binding activity in the subsequent elution with the sodium acetate buffer (Fig.  1). On the other hand, cytochrome c was unable to displace the receptor, indicating that adsorption of NGF receptor to the NGF-Sepharose was specific.  The purification achieved by the method described above could not be quantitated because there was too little protein recovered in the NGF receptor-containing fractions of the eluate for a protein concentration measurement. Quantitation of recovered preiodinated proteins could not be done because of the presence of Triton X-100 in the original extract, and detergents which would have allowed iodination (/3-octyl glucoside, CHAPS, Tween 20) were not as efficient as Triton X-100 in solubilizing the NGF receptor from sympathetic ganglia membranes2 Analysis of the proteins eluted in the NGF receptor-containing fractions by direct iodination (see below) indicated that substantial purification was achieved.

B C D E F G H
Iodination of the Purified NGF Receptors-The components present in the NGF-Sepharose eluate with '2sII-NGF binding activity were iodinated and submitted to SDS-polyacrylamide gel electrophoresis and autoradiography (Fig. 2). Three components were visible corresponding to molecular weights of 126,000, 105,000, and 81,000. No distinct bands were seen in the lower molecular weight region, but a smear was present which was also observed in the iodinated pattern of NGF-

R. N. Kouchalakos, unpublished observations.
Sepharose fractions devoid of binding activity, albeit to a lesser extent. The smear presumably corresponds to contaminants leaching from the NGF-Sepharose continuously. Since the NGF receptors recovered in the eluate of the affinity column were present in extremely small amounts, any lowlevel protein contaminants could easily be labeled equally or even to a greater extent than the receptors themselves. The two larger bands correspond in size to authentic components of the NGF receptor (see below).
Photoaffinity Labeling of the Purified NGF Receptors-More definitive identification of the receptor species present in the NGF-Sepharose eluate was made by photoaffinity cross-linking of "'1-NGF (Fig. 3). Two major components were specifically labeled corresponding to approximate molecular masses of 135 and 110 kDa and a minor component corresponding to 84 kDa. A fourth band was nonspecifically labeled and was also present in cross-linking reaction mixtures containing excess unlabeled NGF (Fig. 3, lane B) and reaction mixtures devoid of NGF receptors (data not shown). It may correspond to a cross-linked aggregate of "'1-NGF itself since the photoaffinity labeling was performed on NGF receptors in solution, and the labeled receptors were not separated from cross-linking artifacts. The two major specifically labeled species, after correction for the ligand molecular mass (13 kDa), correspond well to the two major components of the purified NGF receptors visualized by direct iodination (Fig. 2). The minor cross-linked species may correspond to the 81-kDa component of the eluate of the NGF-Sepharose column, but its labeling intensity indicated a much lower affinity for "'1-NGF than the other NGF receptor components. The same specific labeling pattern of the major species was also observed with superior cervical ganglia membranes (15); indicating that the NGF receptor species present in the membrane are solubilized by Triton X-100 and co-purify on NGF-Sepharose. The minor 84-kDa cross-linked species was not visible in the labeling pattern from membranes ( E ) , suggesting that it could also be an artifact of the cross-linking procedure in solution. Alternatively, cross-linking to receptors in the membrane may not have been sensitive enough to detect this component.
Conversion between the Two Major Photolabeled NGF Receptor Species-NGF receptors cross-linked to "'1-NGF were submitted to trypsin digestion with increasing amounts of enzyme, followed by SDS gel electrophoresis and autoradiography (Fig. 4A). The amount of the high molecular weight cross-linked NGF receptor species (M, = 130,000 in this gel) decreased relative to the smaller major cross-linked species ( M , = 105,000) when the trypsin concentration was increased from 0 to 1.0 pg/m1.3 At higher concentrations of trypsin, the larger species totally disappeared, whereas the smaller species remained visible, albeit that it decreased as well, after digestion with 20 pg/ml trypsin (Fig. 4B). These results suggest that the 105-kDa cross-linked NGF receptor species is a proteolytic product of the 130-kDa species. In contrast, the minor 83-kDa cross-linked species remained invariant, suggesting that it was not a proteolytic product of the larger species (Fig. 4A), although the small amounts present initially render this judgement tentative in nature.
In order to observe directly the conversion of 130-kDa species to the 105-kDa form, the larger cross-linked NGF As observed previously (15). the relative amounts of the two species of NGF receptor are variable from preparation to preparation, and the sample utilized in the experiment contained only trace amount of the lower molecular weight form; that may have occurred because it was purified from freshly prepared membranes as opposed to membranes stored at -70 "C. receptor species was electroeluted from an SDS gel, submitted to limited trypsin digestion, and electrophoresed again on an SDS gel (Fig. 5). Indeed, a species corresponding to approximately 100 kDa was generated, suggesting that the 100-110-kDa NGF receptor species observed by direct iodination (Fig.  2) and photoaffinity labeling (Fig. 3) was derived from the 130-135-kDa receptor species by proteolysis.

DISCUSSION
A simple method has been devised to isolate NGF receptors from superior cervical ganglia membranes in substantially purified form. This is the first report of NGF receptor purification from a natural target, i.e. tissue responsive to NGF in uiuo. Triton X-100 was the most efficient of the detergents tested to solubilize the NGF receptor.' The purification consisted of a single round of affinity chromatography on NGF-Sepharose, from which NGF receptors were easily desorbed by decreasing the pH of the eluting buffer to 5.0. Elution at pH 5.0 in the presence of a high concentration of salt from an affinity matrix has also been used successfully in the purification of insulin-like growth factor I1 receptor (29) and the insulin receptor (30).
The photoaffinity labeling pattern of the purified NGF receptors with lZ5I-NGF using the reagent HSAB displayed two major cross-linked NGF receptor species corresponding to molecular masses of 130-135 and 100-110 kDa. This pattern was similar to the one obtained with intact membranes of sympathetic ganglia where two major species of 143 and 112 kDa were observed (15). Clearly, the main receptor species of the membranes were solubilized and purified. The small difference in the estimation of the molecular weight of the larger form could reflect some proteolysis during solubilization and purification in spite of the presence of protease inhibitors; however, electrophoresis of the cross-linked purified NGF receptors and receptors cross-linked in the membrane on the same SDS gel showed that the respective NGF receptor species co-migrated. The difference in the estimated molecular weights reported in these studies herein and previously (15) probably reflects minor technical differences in the SDS gels used.
The photoaffinity labeling patterns of receptors from membranes and in a purified form did differ in the minor crosslinked species observed. Sympathetic ganglia membranes displayed a minor 300-kDa band which was absent in the purified receptors, suggesting that this band could have corresponded to the cross-linking of one of the NGF receptor species with another protein which did not co-purify on NGF-Sepharose. Alternatively, it could have been a dimer of the 143-kDa NGF receptor species which dissociated during the Triton X-100 solubilization. On the other hand, the purified cross-linked NGF receptors displayed a minor 84-kDa species which was not detected in the membrane sample. Whether the 84-kDa species represented a genuine, albeit low affinity, NGF receptor component or simply the artifactual cross-linking of lZ5I-NGF with the bovine serum albumin used as carrier protein remains to be established.
The studies presented here allow a better understanding of the relationship between the two main NGF receptor species observed after cross-linking. Direct iodination of the purified NGF receptors revealed two species of 126 and 105 kDa which correspond, respectively, to the 130-135-and 100-110-kDa photoaffinity labeled forms. The larger NGF receptor species, therefore, is apparently a single-chain component since the direct isolation experiments eliminate the possibility that the 130-135-kDa species results from cross-linking of the 100-110-kDa species with either dimeric (or even larger aggregates) of NGF or another protein. However, they do not rule out the possibility of a second, noncovalently associated protein that does not dissociate in either the ionic or nonionic detergents utilized in these experiments.
Several lines of evidence support the possibility that the 100-110-kDa NGF receptor species is a proteolytic product of the larger molecule: 1) progressive trypsin digestion of the photoaffinity labeled NGF receptors results in a decrease in the amount of the 130-135-kDa species concomitant with an increase in the 100-110-kDa band, as judged by the labeling intensity (Fig. 4A); 2) the amount of the 130-135-kDa species was greater with NGF receptors purified from freshly prepared membranes than membranes stored at -70 "C (compare Figs. 3 and 4A); and 3) trypsin treatment of the purified 130-135-kDa species electroeluted from SDS gels produced a 100-kDa fragment (Fig. 5). These data together with the previously reported observation of four proteolytic fragments common to the two NGF receptor species (15) strongly suggest a precursor-product relationship for the two NGF receptor species.
Characterizations of NGF receptors from various sources have led to the description of a number of forms summarized schematically in Fig. 6. These forms can be conveniently have led to the description of a number of forms summarized schematically in Fig. 6. These forms can be conveniently grouped into four classes according to their general size: class A, 70 to 81 kDa; class B, 87 to 105 kDa; class C, 120 to 145 kDa; and class D, 190 to 300 kDa. Whether all classes reflect genuine receptor species, as opposed to binding proteins, remains to be proven since all have been identified only through their ability specifically to bind labeled NGF. However, it is significant that class C receptors are consistently present in tissues and cells responsive to NGF yet absent from human melanoma cells which are not stimulated by NGF. Another feature unique to the human melanoma cells is the presence of a uniform population of NGF-binding sites exhibiting low affinity toward NGF (20), in contrast with sympathetic neurons, sensory neurons, and PC12 cells, which all exhibit both high and low affinity NGF-binding sites (7,9,11). Since convincing evidence has been presented that NGF-mediated neurite outgrowth, as well as other responses, occurs through interaction with the high affinity binding sites only (31)(32)(33), it suggests that these correspond to the class C receptors. This coincides with the findings of Hosang and Shooter (17) who found the photoaffinity labeled receptor species of 158 kDa in PC12 cells to be chase-stable at 0 "C and trypsin-resistant, indicating that it is the receptor from which NGF dissociates slowly. In addition, the 158-kDa species is preferentially labeled at low 'T-NGF concentrations (17,34), suggesting that it is also the high affinity NGF receptor, i.e. the receptor mediating NGF actions. Thus, the correlation of class C receptor species with responsiveness to NGF (Fig. 6), together with the identification in PC12 cells of the class C species as the high affinity NGF receptor (17), suggests that these are the biologically relevant NGF receptors.
According to the studies presented herein, class B species are derived from class C by limited proteolysis in sympathetic neurons. However, it is important to note that this classification scheme does not establish that receptors of similar molecular weight may not be entirely different from one organism (or tissue) to another. In this regard, the low molecular weight (class A and B) receptors of the human melanoma cell line A875 may not represent true receptors at all. Alternatively, they may have arisen from the receptor gene by truncation, either of the gene itself or its mRNA, analogous to the erbB protein and the epidermal growth factor receptor (35), rather than from proteolysis of a larger form. In the case of the erbB protein, however, the epidermal growth factorbinding domain has not been retained and thus it is not strictly analogous to the possible situation in A875 cells. The relationship of these melanoma receptors or any of the other species of NGF receptors reported to be in the class C group will require more definitive structual data.