Isolation and characterization of tissue-specific isozymes of glucosephosphate isomerase from catfish and conger.

In teleosts glucosephosphate isomerase exists as two tissue-specific isozymes. Most tissues contain the more acidic liver-type isozyme, while white muscle contains the more basic isozyme; and a few tissues contain both the liver- and muscle-type isozymes as well as a hybird. The isozymes were isolated from catfish liver and muscle and from conger muscle and shown to be homogeneous by polyacrylamide gel electrophoresis, isoelectric focusing, analytical ultracentrifugation, and rechromatography. Both isozymes are of molecular weight 132,000 (S020,w = 7.0 S) and composed of two subunits of Mr approximately 65,000. The muscle and liver isozymes were shown to have distinct isoelectric points (catfish liver = 6.2; muscle = 7.0) and amino acid compositions. Tryptic peptide maps, after S-carboxymethylation and carbamylation, revealed several distinct differences in the primary structures of the isozymes. Although the isozymes could also be distinguished on the basis of their stabilities, most of their basic catalytic properties were found to be similar. A conger was obtained which was heterozygous for the variant allele at the muscle-glucosephosphate isomerase locus. A comparison of the variant conger muscle isozyme with the wild type revealed a single altered peptide, suggesting a point mutation. The structure-function studies, as well as the genetic studies, clearly establish that the two types of isozymes are of independent genetic origin.

In teleosts glucosephosphate isomerase exists as two tissue-specific isozymes. Most tissues contain the more acidic liver-type isozyme, while white muscle contains the more basic isozyme; and a few tissues contain both the liver-and muscle-type isozymes as well as a hybrid. The isozymes were isolated from catfish liver and muscle and from conger muscle and shown to be homogeneous by polyacrylamide gel electrophoresis, isoelectric focusing, analytical ultracentrifugation, and rechromatography. Both isozymes are of molecular weight 132,000 (sz,,+ = 7.0 S) and composed of two subunits of M, approximately 65,000. The muscle and liver isozymes were shown to have distinct isoelectric points (catfish liver = 6.2; muscle = 7.0) and amino acid compositions. Tryptic peptide maps, after S-carboxymethylation and carbamylation, revealed several distinct differences in the primary structures of the isozymes. Although the isozymes could also be distinguished on the basis of their stabilities, most of their basic catalytic properties were found to be similar. A conger was obtained which was heterozygous for the variant allele at the muscle-glucosephosphate isomerase locus. A comparison of the variant conger muscle isozyme with the wild type revealed a single altered peptide, suggesting a point mutation.
The structure-function studies, as well as the genetic studies, clearly establish that the two types of isozymes are of independent genetic origin.
In marked contrast, electrophoretic studies on tissue extracts from teleostean fish (20)(21)(22)(23)(24)(25)(26) have suggested that most fish exhibit tissue-specific isozymes of glucosephosphate isomerase which appear to be the result of independent autosomal gene loci. These unique tissue-specific isozymes thus provide an interesting and unique system for studying structure-function relationships of this enzyme. These isozymes also raise fundamental questions regarding their physiological roles in different tissues, as well as their evolutionary origin. The isolation of these isozymes and the elucidation of the catalytic, chemical, and physical properties of these proteins were undertaken in order to establish the fundamental structural and functional properties. sodium dodecyl sulfate-polyacrylamide gel electrophoresis, and analytical ultracentrifugation studies (see below). In most conger tissues including the liver, kidney, heart, spleen, gills, brain, spinal cord, intestine, stomach, swim bladder, erythrocytes, cornea, and other eye tissues, the more anodal, liver-type, glucosephosphate isomerase isozyme predominated.

MATERIALS
Brain and spinal cord ( Fig. 1B) also contained additional traces of glucosephosphate isomerase activities migrating slightly more anodal than the liver isozyme. Only after prolonged staining could traces of the muscle isozyme be found in any of these tissues. In white .muscle the slower moving muscle-type band predominated, but traces of the liver-type isozyme were also observed occasionally.
Serum and red muscle contained both muscle and liver isozymes and were unique among the conger tissues studied in exhibiting a hybrid band between the two major tissue isozymes.
A previously reported (25) more basic genetic variant of the conger muscle isozyme was observed in 5 out of 55 conger examined from the Plymouth area. A zymogram of red muscle and spinal cord extracts from a conger heterozygous for this variant allele is also shown in Fig. 1B. No fish that were homozygous for the variant have been found so far.

Isolation of Lioer and Muscle Isozymes
In order to compare the structure-function properties of the fish isozymes at the molecular level, it was necessary to obtain the enzymes in homogeneous form. The acidic nature of the fish glucosephosphate isomerases prevented them from being bound to cellulose phosphate, and thus, it was not possible to purify them by the specific substrate elution technique developed for the more basic glucosephosphate isomerase from human (9). The following procedures were developed for the isolation of the glucosephosphate isomerase isozymes from either skeletal muscle or liver of the freshwater catfish and from the skeletal muscle of conger. The extraordinary lability of the conger liver enzyme made isolation difficult, and thus, only limited studies were possible on this isozyme.
Catfish Glucosephosphate Isomerase Purification-Catfish were killed, the tissues removed, minced, and extracted by blending for 30 s with 4 volumes of 10 mM triethanolamine buffer, pH 9.0, at 4". The extracts were centrifuged at 20,000 x g for 60 min, the pellets discarded, and the supernatant solutions filtered through glass wool. After dialysis overnight against Buffer A (10 mM triethanolamine, pH 9.0, containing 1 mM EDTA and 0.1% (v/v) 2-mercaptoethanol), the extracts were subjected to chromatography on DEAE-Sephadex columns (8 x 60 cm) which had been equilibrated in Buffer A. One liter of Buffer A containing 0.01 M NaCl was pumped through the column to elute a large amount of contaminating protein.
For the isolation of the muscle isozyme, a linear sodium chloride gradient of 2 liters with ionic strength ranging from 0.02 to 0.08 was applied, and the isozyme was eluted in a sharp peak at an ionic strength of 0.04. When the liver extract was chromatographed, it was necessary to employ a linear

Homogeneity Studies and Physical Properties
Isoelectric Focusing and Electrophoresis-The glucosephosphate isomerase isozymes were purified further by isoelectric focusing, as shown in Fig. 3. Electrofocusing of the catfish muscle isozyme revealed a major component of glucosephosphate isomerase activity with an isoelectric pH of 7.0 and an active minor component at pH 6.6. Isoelectric focusing of liver glucosephosphate isomerase revealed a major component at pH 6.2 and a minor component at a pH of 5.6. Conger muscle glucosephosphate isomerase electrofocused with an apparent isoelectric pH of 6.4 with two minor components with isomerase activity at pH 6.6 and 6.2. These minor fractions seem to be "pseudoisozymes" (15) caused by sulfhydryl oxidation of the native enzyme, since addition of 2-mercaptoethanol or dithiothreitol greatly reduced the amounts of these minor components.
After isoelectric focusing, the isozymes from catfish muscle and liver and conger muscle were obtained with specific activities of approximately 400 to 450 units/mg and were judged to be homogeneous by a variety of criteria. Rechromatography on DEAE-cellulose, Sephadex G-200, or reisoelectric focusing did not increase the specific activity of the isozymes further. Table I summarizes the results of typical purifications of glucosephosphate isomerase from catfish muscle and liver.
Both the catfish liver and muscle isozymes yielded single bands after polyacrylamide gel electrophoresis in the presence of reducing agents (Fig. 4A). In agreement with the isoelectric focusing experiments, the conger muscle enzyme migrated as a more acidic protein than the catfish muscle enzyme (Fig. 4B). conger isozymes when reducing agents were not present (Fig. 4,  A and B). Molecular Weight and Subunit Studies-Although the catfish muscle and liver isozymes exhibited distinctly different electrophoretic, chromatographic, and electrofocusing properties, the two proteins were found to possess essentially identical molecular weights. When the two isozymes were subjected simultaneously to sedimentation velocity ultracentrifugation (Fig. 5), they sedimented as single symmetrical boundaries with identical sedimentation coefficients. The catfish isozymes and conger muscle isozyme yielded the following relationships: catfish muscle .s&,~ = 7.04 (1 -0.005 C); catfish liver .s'&,~ = 7.05 (1 -0.004 C); conger muscle s$0,w = 7.00 (1 -0.004 C), where C is the protein concentration in milligrams per ml. Sedimentation equilibrium ultracentrifugation of the catfish muscle and liver isozymes yielded weight average molecular weights of 132,000 * 2,000 and 131,000 * 3,000, respectively.
Polyacrylamide gel electrophoresis of catfish liver and muscle and conger muscle in the presence of sodium dodecyl sulfate yielded single, sharp bands with migrations corresponding to molecular weights of 65,000. Moreover, only a single Coomassie blue-staining band was observed after sodium dodecyl sulfate polyacrylamide electrophoresis of a mixture of equal amounts of human glucosephosphate isomerase and the fish isozymes. These results, as well as the ultracentrifugation studies, suggest that the fish glucosephosphate isomerase isozymes are dimers composed of two identical subunits of M, = 65,000.

Stability and Catalytic Properties
Stability Studies-The muscle and liver glucosephosphate isomerase isozymes were distinguished on the basis of their stabilities.
The purified enzymes were incubated under identical conditions of pH, protein concentration, and ionic strength at various temperatures and immediately placed in ice and assayed for remaining catalytic activity.
In addition, the enzymes were incubated at fixed temperatures and the activity monitored as a function of time. In both types of studies the catfish liver isozyme was found to be significantly more stable than the muscle isozyme (Fig. 6). Surprisingly, exactly the reverse situation was found for the conger isozymes. The conger liver isozyme proved to be much more labile than the muscle isozyme. In fact, this particular lability of the conger liver isozyme made isolation and comparative structural studies on the conger liver isozyme difficult. (3) or in the absence of the reducing agent (4) was subjected to electrophoresis for 4 hours at 300 volts. All gels were stained for total protein as described under "Materials and Methods." B, the isolated isomerases were subjected simultaneously to polyacrylamide gel electrophoresis as above at 400 volts for 8 hours and then stained for total protein as described under "Materials and Methods." Gel 1 = catfish muscle; Gel 2 = catfish liver; Gel 3 = conger muscle.

Catalytic
Studies-The catfish liver isozyme exhibited a slightly broader pH optimum in the alkaline region than the muscle isozyme (Fig. 7). For example, at pH 10.5, the liver isozyme exhibited 100% maximal activity, whereas the muscle isozyme exhibited only 82% of its maximal activity. Other kinetic parameters of the isozymes were also measured and are summarized in Table II. Essentially no differences were found in K, values for fructose 6-phosphate or the competitive inhibitors, B-phosphogluconate or erythrose 4-phosphate.
Activation energies from Arrhenius plots of the isozymes from the two species were nonlinear with breaks occurring at 21-23". These results are summarized in Table II.

Chemical Properties
Amino Acid Compositions-When the isolated glucosephosphate isomerases from catfish liver and muscle, and from conger muscle, were subjected to amino acid analysis, a high degree of similarity was observed (Table III). The most distinct differences between the catfish liver and muscle isozymes appeared in the contents of lysine, serine, valine, isoleucine, for one cell. The upper boundary is the muscle isozyme, while the lower is the liver isozyme.
The photograph was taken 14 min after reaching speed. ,\. and phenylalanine. Based on over-all compositions, the catfish liver and muscle enzymes were more similar to each other than either was to the conger muscle isomerase.' Peptide Maps-Tryptic peptide maps of the two catfish isozymes after S-carboxymethylation revealed a large number of peptides, far too many to be resolved by this technique. These results were consistent with the high arginine and lysine content of the enzyme (predicted number of peptides = 67 for muscle and 60 for liver).2 Peptide maps of the catfish muscle and liver isozymes were therefore compared after S-carboxymethylation and carbamylation (Fig. 8). In both cases the number of fluorescamine-reacting peptides was in agreement (* two peptides) with the number predicted from the arginine contents. The over-all peptide maps of the two isozymes were I It is interesting to note that the ratio of acidic to basic amino acids (as determined from total acid hydrolysates) parallels the isoelectric points of the two catfish isozymes, as well as the enzyme from conger muscle, rabbit, and human. The more acidic properties of the catfish liver isozyme (as determined by isoelectric focusing, electrophoresis, and ion exchange chromatography) are consistent with its higher acidic to basic ratio (1.60) as compared to the muscle isozyme (1.37).
%It is of interest to note that both of the catfish and the conger muscle isozymes were digested readily after only S-carboxylation, and essentially no "core" material was observed. This is in contrast to tryptic digestion of rabbit (40) or human (9) glucosephosphate isomerases, where complete digestion could only be achieved after both S-carboxymethylation and carbamylation. properties. However, distinctly different migrations of the other peptides and the relative fluorescence after reaction with fluorescamine clearly indicated a number of differences in the primary structure of the liver and muscle isozymes. Although a comparison of the peptide maps of the conger (see below) and catfish muscle isozymes suggested some over-all structural homology, a much greater structural homology was evident between the two catfish liver and muscle isozymes than between the two muscle isozymes from the two species. Glucosephosphate isomerase was isolated from the white muscle of a conger which was heterozygous for the variant allele shown in Fig. 1B and compared with the wild type conger muscle enzyme. Isoelectric focusing indicated that the variant muscle allozyme was slightly more basic than the wild type muscle isozyme (i.e. apparent p1 values of 6.36, 6.42, and 6.55 for the wild type homodimer, the heterodimer, and the variant homodimer, respectively, were obtained). The variant homodimer was collected separately and compared with the wild type conger muscle isozyme by peptide mapping as described above. A composite tryptic peptide map of the wild type and variant homodimers after carboxymethylation and carbamylation is shown in Fig. 9. A single peptide (e.g. Number 7 and 7A) was observed with altered electrophoretic mobilities.
The altered peptide appears to be more basic and slightly more hydrophobic than the corresponding peptide from the wild type isozyme. These data indicate that the phenotype results from a point mutation. In recent years studies on glucosephosphate isomerase have centered on structural and kinetic properties of the mammalian and yeast enzymes. Genetic studies on glucosephosphate isomerase in a number of mammalian species (3,8,11) are consistent with the hypothesis that the enzyme structure is determined by one or more alleles at a single autosomal locus. It was, therefore, of great interest to find that in diverse teleostean species there are two gene loci for glucosephosphate isomerase which result in isozymes with marked tissue specificity.
phate isomerase and the genetic variant (shown in Fig. 1B). The enzymes were S-carboxymethylated and LY-and c-amino groups carbamylated as described under "Materials and Methods" prior to tryptic digestion. Simultaneous thin layer peptide mapping was carried out, and the plates were sprayed with fluorescamine. The figure represents composite tracings of eight and seven maps of the normal and variant proteins, respectively. The variant protein was identical with the normal protein with the exception of Peptide 7 which was present in the wild type and was replaced by Peptide 7A in the genetic variant. The most highly fluorescent peptides are shown by stippling.
In both the conger and catfish most tissues synthesize the isozyme referred to in this study as the "liver-type," while white muscle was found to possess almost exclusively the more basic "muscle-type" isomerase. The small amounts of livertype isozyme observed in white muscle may be the result of contamination by serum or lymph. Red muscle and serum appear to contain both types of isozymes, as well as hybrids, suggesting that both types of polypeptides may be synthesized in some cells.
such as proteolysis, deamidation, or covalent modification. These data corroborate the genetic studies carried out on tissue extracts from these and other species of teleosts. On the other hand, artifactual multiple electrophoretic forms of fish glucosephosphate isomerase occur if caution is not taken to maintain the enzyme in a reduced state. This situation closely parallels the "pseudoisozymes" observed with glucosephosphate isomerase from rabbit muscle (15), human and rat (16). The faint, rapidly migrating anodal bands of glucosephosphate isomerase activity observed upon starch gel electrophoresis of extracts (Fig. 1) may be such an example. The distinct chemical and physical properties determined The specific activities of the isolated glucosephosphate for the isolated muscle and liver isozymes clearly show that the isomerase isozymes from the fish were approximately 50% two forms of glucosephosphate isomerase are indeed of genetic lower than those found for the enzyme isolated from human or origin and do not represent post-transcriptional modifications rabbit muscle (9,15) or yeast (41). This does not, however, FIG. 8. Tryptic peptide maps of catfish muscle (A) and liver (B) glucosephosphate isomerase isozymes. After Scarboxymethylation and carbamylation, the enzymes were digested with trypsin and peptides mapped on thin layer cellulose as described under "Materials and Methods." The stippled peptides are those which were most highly fluorescent after spraying with fluoresc-J amine.
appear to be the result of heterogeneity, since both liver and muscle isozymes were shown to be homogeneous by a variety of criteria. Likewise, the specific activities of 400 to 500 units/mg do not seem to be the result of partial denaturation during the isolation processes. The two isolation procedures were developed totally independently in our two laboratories with comparable results. While it might be argued that the heat step involved in the purification of the conger glucosephosphate isomerase may have induced some partial denaturation or proteolysis, this cannot be the case for the catfish isozymes. The procedure outlined for the isolation of the catfish isozymes is mild, and recoveries were good at all stages.3 It thus appears that the glucosephosphate isomerases from teleosts do, indeed, possess lower specific activities than the enzyme from the higher mammals thus far studied. The difference in the stability of the muscle and liver isozymes was most striking in the conger where the liver isozyme was much more labile than the muscle isozyme. Crude homogenates of cod similarly have shown a more labile liver isozyme, and Avise and Kitto (23) reported a more labile liver glucosephosphate isozyme in extracts of Astynax mexicanus.
It is not understood why the isolated liver isozyme of the catfish does not show this lability. It is possible that the inactivation of the liver isozyme in crude extracts or partially purified fractions may be due to protease contamination. specific isozymes has not evolved. The teleosts appear to be an exception with tissue-specific, genetically determined isozymes. Avise and Kitto (23) have suggested that the high degree of polyploidy in teleosts is a likely cause of gene multiplicity of glucosephosphate isomerase.* The present study has shown that indeed true isozymes exist in catfish and conger, and that these isozymes exhibit a tissue-specific distribution.
The muscle-and liver-type enzymes clearly have experienced a number of amino acid sequence changes during the course of evolution since gene duplication.
Although the unique physiological functions of each isozyme remain to be elucidated, it is clear that further studies on the two isozymes and the apparent widespread allelic variants of these genes should provide a unique opportunity to assess the metabolic roles of the enzyme in various tissues and to correlate structure-function relationships.
The amino acid compositions and peptide maps clearly indicated that a number of structural differences in the two types of isozymes exist. The catfish liver and muscle isozymes were more similar to each other (amino acid composition coefficient (43) of 0.95) than either isozyme was to the conger muscle protein (composition coefficient of catfish muscle uersus conger muscle = 0.90; and catfish liver uersus conger muscle = 0.92). This was corroborated further by comparison of the tryptic peptide maps from the isozymes of the two species. The conger muscle glucosephosphate isomerase was essentially as different in its amino acid composition from the two catfish enzymes as it was from rabbit enzyme (cc = 0.92) or human enzymes (cc = 0.91). This is perhaps not surprising in view of the probable separation of the Elopomorpha from the Euteleostei as early as the late Jurassic (44). However, the true evolutionary relationships between the enzymes must await sequence studies.

Glucosephosphate
isomerase has long been considered a bifunctional enzyme of glycolysis and gluconeogenesis and present in large excess compared to the rates of metabolic flux. Thus, in most organisms studied (19)