Assignment of the gene for cytosolic alanine aminotransferase (AAT1) to human chromosome 8

The segregation of human cytosolic alanine aminotransferase (AAT1) and the individual human chromosomes has been studied in 27 secondary and tertiary rat hepatoma‐human (liver) fibroblast hybrids. The staining solution used to visualize AAT activity on starch gels was specific for AAT since it was visualized only when all components of the stain were present. The locus for human AAT1 has been assigned to chromosome 8.

Cytosolic and mitochondrial AAT activities have been found in mammals (Swick, Barnstein & Stange, 1965). These enzymes play a key role in gluconeogenesis via the glucose-alanine cycle (Felig, 1975). In muscle, pyruvate is transaminated by the cytosolic enzyme (AAT1) to alanine (DeRosa & Swick, 1975), which is then released into the circulation for uptake by the liver. In the hepatocyte, alanine is converted to pyruvate primarily by mitochondrial AAT (Dieterle et al. 1978), and the hepatic pyruvate is then metabolized to glucose.
The cytosolic enzyme has been characterized in rat liver (Gatehouse et al. 1967 ;Matsuzawa & Segal, 1968) pig heart (Saier & Jenkins, 1967) and human liver (Hopper & Segal, 1964;Kamoda et al. 1980). I n addition, the cytosolic enzyme is present in human kidney, muscle, pancreas and erythrocytes (Maclagan, 1970;Kamoda et al. 1980). The specific activity of the cytosolic enzyme varies markedly with the tissue source, the hepatic enzyme having the highest specific activity (Kamoda et al. 1980). Although human erythrocytes have only 10% of the hepatic specific activity (Maclagan, 1970), the erythrocyte enzyme has been the subject of extensive genetic studies. Chen & Giblett (1971) were the first to find a polymorphism for human cytosolic erythrocyte AAT. To date, a t least 11 different alleles have been identified a t the erythrocyte AAT locus (Chen et al. 1972;Kompf & Ritter, 1979). It should also be noted that the serum AAT activity has proven to be a useful indicator of hepatocellular injury (Combes & Schenker, 1969).
The previous demonstration that rat hepatoma cells expressed cytosolic AATl and that the activity was inducible by dexamethasone (Lee & Kenney, 1970;Sparkes & Weiss, 1973) suggested that somatic cell hybrids made from these hepatoma cells and human fibroblasts derived from liver might be useful for the chromosomal localization of the human cytosolic AATl gene. We report herc the assignment of the structural gene for cytosolic AATl to human chrornosomc 8 using rat hepatoma-human (liver) fibroblast hydrids. These hybrids expressed the human and rat AAT isozymes a t high levels without glucocorticoid induction. Since a previous report (van Someren et al. 1974b) had suggested that the staining solution used to visualize AAT also detected other enzymes, the specificity of the stain was investigated in both hybrid cells and other tissues.

C'ell lines and hybrid cell formation
The rat hepatoma cell line, FU5AH (HGPRT-), was initially established by Dr Mary Weiss (Schneider & Weiss, 1971 ;Weiss & Chaplain, 1971) and was generously provided by Dr Marcello Siniscalm. Fresh human foetal liver was minced, treated with 0.05 yo collagenase, washed in serum free Dulbecco's Medium (DMEM, GIBCO, Grand Island, NY) and then the individual cella were established in culture using standard techniques. The cells were grown in culture for several weeks before fusion, at which time the majority of these cells appeared fibroblastic in morphology. Hybrids were produced by the fusion of rat hepatoma FU5AH cells and the foetal human (liver) fibroblasts. For fusion, the cells were mixed together in serum free DMEM containing 47 yo polyethylene glycol (molecular weight 6000) and 15 Yo dimethylsulphoxide for 2 niin (Shay, 1977). Additional serum free medium was added, and the cells were washed twice before plating in DMEM containing 10 yo foetal calf serum (GIBCO), 10 mM hypoxanthine, 40 ,UM aminopterin, 16 ,UM thymidine and 5 PM ouabain. A single colony was observed after several weeks in the selective medium. This colony was isolated, expanded and subcloned by placing 100 cells in 25 em2 plastic tissue culture flasks. Individual clones, usually one from each flask, were isolated with cloning rings. Several of these secondary clones were further subcloned to give tertiary clones.

Preparation of cell extracts
Confluent hybrid cells were removed from 75 cm2 plastic tissue culture flasks using a rubber policeman. The cells were washed twice with Dulbecco's phosphate buffered saline (GIBCO) and after centrifugation a t 200 g, the cells were either used immediately or frozen a t -70 "C. For enzyme analysis, the cells were suspended in 25-1OOpl of water, subjected to one cycle of freezing ( -70 "C) and thawing (37 "C), centrifuged at 14000 g for 20 min, and the supernatant was removed €or electrophoretic studies. Fresh rat or fresh frozen human liver was homogenized in an equal volume of water or 0-1 M Tris-HC1 buffer, pH 7-5, centrifuged a t 14000 g for 20 min and then the supernatant was removed for electrophoresis.

Analysis of AAT
AAT activity in tissues and hybrid cells was analysed by starch gel electrophoresis. The electrophoretic system was adapted from the method used for separation of phosphoglycolate phosphatase . The bridge buffer was TEMM (0.1 M Tris-HCl/O*Ol M disodium ET)TA/Ol M maleate/0.01 M-MgCl,) adjusted to pH 7.2 with NaOH. The gel buffer was a 1 to 10 dilution of the bridge buffer. I n some gels 2-mercaptoethanol(O.1 ml of 0.1 M mercaptoethanol per 100 ml) was added to the gel buffer. No differences were observed in the separation of the  In the stain containing only 2-oxoglutaric acid and NADH, ammonium chloride (10 mg) was added.
In some experiments the initial staining was followed by a second staining step. Methyl thiazolyl tetrazolium (MTT; 5 mg) and meldola blue (MLB; 100 pg) in 25 ml of 0.5 M Tris-HC1 buffer, pH 7.5, and 25 ml of 2 yo agar were applied to the surface of the gel after incubation with the above stain. All chemicals were from Sigma Chemical Co. (St. Louis, MO) except for MLB which was obtained from Boehringer-Mannheim (Indianapolis, IN).

Identi$cation of human chromosomes in hybrids
The human chromosomes in the hybrid cells were identified by karyotypc and/or enzyme marker analyses. Metaphase spreads were stained first with Giemsa-11 (Friend et al. 1976), destained and then restained with quinacrine hydrochloride for Q banding (Breg, 1972).
The enzyme markers examined are listed in Table 1. In many cases, the published electrophoretic conditions were modified as indicated ( Table 1) for optimal separation of the rat and human isozymes. All the enzymes were separated by starch electrophoresis, except for glutathione reductase, aconitase, and glucose-6-phosphate dehydrogenase which were separated by cellulose acetate electrophoresis. The staining systems used were from the published methods referenced in Table 1 or from Harris & Hopkinson (1976).

Speci$city of AAT staining system
The complete staining system used to visualize AAT on gels contained alanine, 2-oxoglutarate, LDH and NADH. I n this system, pyruvate generated by AAT was converted to lactate by LDH, and fluorescent NADH was converted concurrently to non-fluorescent NAD. Under ultraviolet light, AAT activity (i.e. the newly generated NAD) appeared as dark bands against a light background. The specificity of this stain was examined with samples of human liver, human erythrocytes, rat liver and rat hepatoma FU5AH cells. With the complete stain (Figs 1 and  2, gel A), human erythrocytes had only one activity band, indicating that the donor of the erythrocytes was homozygous a t the AAT locus. Human liver had two bands; one (labelled H i ) was a strong band which migrated to the same position as the erythrocyte isozyme, and the second was a much weaker band (H2) which migrated more anodally. FU5AH hepatoma cells had one activity band ( R l ) which migrated to the same position or slightly morc anodally than the minor human liver band (H2), and a second band (R2) which migrated close to the anode. Rat liver had only a single activity band (Rl). In some gels, an additional weak band was visible in rat liver and hepatoma cells a t the origin after prolonged incubation (Fig. 1, gel A).
No bands were seen when 2-oxoglutarate was omitted from the stain. Exclusion of alanine or both alanine and LDH from the staining solution (Fig. 1 , gels B and D), resulted in the detection of a weak band in human liver which migrated to the same position as band H2 when visualized with the complete stain; no such band was detected in human erythrocytes with the modified stains. When only LDH was omitted (Fig. 1, gel C), a single activity band (3) which migrated more cathodally than band H1 was seen in erythrocytes; human liver had the more cathodal band 3 as well as the H2 band. Both rat liver and FU5AH cells had a band in the same position as band R1 when alanine, LDH or both were absent ( Fig. 1 and Fig. 2). They also had an additional band which migrated to the same position as 3 in the absence of LDH ( Fig. 1, gel C). Band R2 was present with equal intensity under all staining conditions, while all other rat and human activity bands were much weaker and took longer to visualize than the activity bands seen with the complete stain.
Some gels were also stained with MTT and MLB after staining with the AAT stain. Besides the bands observed under ultraviolet light, additional bands which migrated closer to the anode were observed (not shown). Similar bands also were seen on gels stained only with MTT and MLB, suggesting that these were superoxide dismutase.  AAT activity in cultured human fibroblasts was examined and no AAT activity was observed even when extracts from 2.5 x los cells were applied to the gel.

Analysis of AAT in rat-human hybrids
When visualized with the complete stain, 23 out of 27 secondary and tertiary subclones had an activity band which co-migrated with human liver AAT ( H l ) (Figure 2, gel A). Of these 23 hybrids, 4 also expressed a band which migrated to the same position as the H2 and R1 activity bands seen in human and rat liver, respectively. I n one hybrid, the H2/Rl band consistently was equal in intensity to the human H1 band while in the other hybrids, the H2/R1 band was weaker than the human activity. No intermediate bands were detected which would correspond to human-rat AAT hybrid proteins. Nineteen of the 23 hybrid clones had either an extremely faint band or no activity a t the H2/Rl position. The four hybrids which were negative for the human isozyme had a strong band which migrated to the same position as the H2/R1 band.
When the hybrids were stained without alanine (Fig. 2, gel B), no band corresponding to the human AAT isozyme H1 was observed in any of the hybrids. However, a faint activity which migrated in the H2/R1 position was seen in all hybrids examined, whether the H2/R1 band was present or absent when visualized by the complete stain.
The segregation of human cytosolic AAT and the individual human chromosomes in 14 secondary and 13 tertiary subclones is shown in Table 2. The human chromosome with the highest degree of concordant segregation with the expression of human AATl is chromosome 8. Only one out of 14 secondary and none of the 13 tertiary hybrids were discordant. A high degree of concordant segregation also was observed with chromosome 6.

DISCUSSION
Our studies on the specificity of the AAT staining methods were consistent with the strong human activity band (designated H1) in both human liver and erythrocytes being cytosolic AAT. I n support of this was the fact that the H1 activity was present only with the complete stain and the fact that erythrocytes lack the mitochondrial enzyme. The weaker, more anodal band (H2) in human liver may be the mitochondrial activity; however, the presence of this band in human liver in the absence of alanine and/or LDH suggests that this band is not an AAT isozyme. The bands seen with the stain lacking LDH did not co-migrate with AAT and were much weaker in intensity, indicating that they also were not AAT isozymes.
Rat liver and hepatoma cells both had an activity band ( R l ) which was visible with both the complete stain and staining solutions lacking alanine and/or LDH. The R1 band was, however, markedly more intense with the complete stain. It appears that both rat AAT and an alanine and/or LDH independent enzyme(s) migrated to the same position on the starch gels. It is reasonable to assume that the R1 band seen with the complete stain is AAT since it is the only band present in rat liver, and it is much stronger in intensity than the other bands.
The bands seen in both rat and human sources in the absence of alanine alone or both alanine and LDH, could be glutamate dehydrogenase (GLUD, EC 1 . 4 . 1 . 3 ) since this enzyme can be visualized with a staining solution containing 2-oxoglutarate, NADH and an ammonium source (Nelson et al. 1977). The LDH used in our stain was in ammonium sulphate solution and ammonium chloride was added to the stain lacking LDH and alanine. Van Someren et al. (1974b) also found that LDH activity was visualized when gels were stained with 2-oxoglutarate and NADH in the presence or absence of exogenous LDH. The bands detected in our gels in the absence of exogenous LDH may also be LDH since similar bands were seen on a starch gel stained with only pyruvate and NADH (Harris & Hopkinson, 1976). The most anodal band, R2, seen in the rat hepatoma was not identified. lntensely staining bands were present in the rat-human hybrids with the complete stain. The most cathodal band, H1, present in some hybrids was designated AATl since it co-migrated with human liver and erythrocyte AAT, and it appeared only with the complete stain. ldentifieation of the more anodal band in the hybrids was more difficult since rat AAT (band R1) and human liver activity H2 migrated to the same position. I n addition, the hybrids also expressed a weak band a t the H2/R1 position when the stain lacked alanine. Clearly, further studies are necessary to identify electrophoretic and biochemical property differences between H2 and R1.
In the majority of hybrids positive for Hl(AAT), little or no activity was seen in the H2/R1 position. This finding suggested that when human AATl is expressed, rat AAT and the rat alanine independent activity (Rl) are not expressed, or the human activity H2 is not expressed, or that human AAT and H2 activity segregate independently.
It is of interest to note that in hybrids where both the H1 and H2/Rl bands were present,  Matsuzawa & Segal, 1968) and the human erythrocyte enzyme also is dimeric since three isozymes have been seen on gels in erythrocytes from individuals heterozygous at the AAT locus (Chen & Giblett, 1971). Therefore, AAT heterodimers, similar to those seen with other dimeric enzymts in hybrid cells, were expected but were not observed. This finding suggested that either the clones are heterogeneous with rat AAT expressed in some cells and human AAT in the others, or that the H2/R1 band observed in certain hybrids positive for human AAT was not rat AAT. Based on these studies, we have assigned the locus for cytosolic AATl to human chromosome 8. Giblett et at. (1978) had suggested, on the basis of linkage analysis, that AATl and soluble glutamate oxaloacetate transaminase (GOT) were weakly linked and thus on chromosome 10. Four out of 22 of our hybrids were discordant for AAT and GOT. Kielty & Povey (1981) have also reported the assignment of AATl to human chromosome 8 using rat-human hybrids. However, Wijnen & Meera Khan (1981) have assigned the locus to human chromosome 16. In our hybrids, 8 of 14 secondary clones were discordant for AATl and chromosome 16. Jf the locus coding erythrocyte AATl is also on chromosome 8, it can be deduced by exclusion mapping to be located in the region 8q13 + 8qter (Cook, Jeremiah & Buckton, 1981).