Pyridoxine-derived B6 Vitamers and Pyridoxal5’-Phosphate-binding Proteins in Cytosolic and Nuclear Fractions of HTC Cells*

The nuclear fraction of rat hepatoma-derived HTC cells contained approximately 8% of the total cellular pyridoxal5’-phosphate. HTC cells were able to metab- olize [3H]pyridoxine to coenzymatically active pyridoxal 5’-phosphate and pyridoxamine 5’-phosphate. As HTC cells did not have any demonstrable pyridox- ine-5’-phosphate oxidase activity, the conversion of pyridoxine to pyridoxal 5’-phosphate must have taken place by a nonconventional route. The ratio of pyri- doxal 5’-phosphate to pyridoxamine 5’-phosphate in the nonnuclear fraction of HTC cells was approxi- mately l:l, whereas in the nuclear fraction it was approximately 17~1, indicating that there was selec-tive acquisition of pyridoxal 5’-phosphate by the nu- cleus. With the aid of a monoclonal antibody specific for the 5’-phosphopyridoxyl group, it was shown that there was one major pyridoxal 5’-phosphate-binding protein in a sodium dodecyl electrophoresis

The subcellular distribution of pyridoxal 5'-phosphate (PLP)' and the major cytosolic PLP-binding proteins in rat liver have been examined by Bosron et al. (1). The fraction of the total PLP found in the nuclei of liver cells was reported to be approximately 21% in the case of rats fed a diet adequate in vitamin Bs and 39% in the case of rats fed a diet deficient in vitamin Bs; the mitochondrial plus lysosomal fraction  (6). Nuclear and chromatin preparations were sonicated before electrophoresis. Two-dimensional gel electrophoresis was carried out by the O'Farrell method (7) essentially as described by Adams (8); the second dimension had a 5% stacking gel and a 10% running gel. Monoclonal antibody E6(2)2, which is specific for the P-Pxy group, has been described previously (6), as has its application for the identification of P-Pxy proteins on Western blots by a horseradish peroxidase-dependent immunoblot procedure (6,9,10). In the present study, PLP-depleted human serum (6,ll) was replaced by PLP-depleted goat serum in the initial blocking solution. In addition, both E6 (2) (Du Pont-New England Nuclear). X-ray film was exposed to dried gels and Western blots at -70 "C until clear signals were discernible upon development (4-5 weeks). Assays-PLP was quantified as described previously (12). The activity of PNP oxidase (EC 1.4.3.5) in cytosolic fractions was determined by a sensitive radiochemical method developed by Langham et al. (13). Pyridoxine kinase (EC 2.7.1.35) activity in cytosolic fractions was measured by our modification (12) of the method of Karawya and Fonda (14). The various Bg vitamer forms were separated by ion-exchange chromatography, as described previously (5), using the procedure of Lumeng and Li (15) (5). For hepatomas grown in Go, PLP has been found to be in the range of 14-22 ng/mg of protein; for liver tissue obtained from rats fed pyridoxine-sufficient and pyridoxine-deficient diets, PLP contents were 37.0 f 2.5 and 19.8 + 2.7 ng/mg of protein, respectively (12,(17)(18)(19). When HTC cells were grown in pyridoxine-free medium for six passages, their PLP concentration fell to 4.9 + 0.9 ng/mg of protein (n = 5). These cells apparently grew perfectly well under these conditions and must have met their vitamin Bs requirements from the vitamer forms available in the 5% fetal bovine serum in the medium. Lipson et al. (11) have reported that the PLP content of human fibroblasts was maintained at approximately 4 ng/ mg of protein during four passages in vitamin B6 (pyridoxal)free Eagle's minimal essential medium containing 10% fetal bovine serum-an observation similar to our own.
Activities of Pyridoxine-metabolizing Enzymes in HTC Cells-PNP oxidase activity in the cytosolic fraction from HTC cells grown in F-12 medium was measured as described previously (5) by a sensitive radiochemical method that uses N-5'-P-Pxy-[3H]tryptamine as substrate (13). As was the case with McA-RH7777 cells (5), there was no demonstrable PNP oxidase activity in HTC cytosols. Pyridoxine kinase activity in HTC cytosolic fractions was determined by a radiochemical method (12) to be 39% of that found in rat liver; the comparable figure for McA-RH7777 cells was 36% (5).

Metabolism of rH]Pyridoxine by HTC
Cells-HTC cells were passed four times in vitamin Bs (pyridoxine)-free F-12 medium so as to deplete endogenous stores of pyridoxine. Cells were then grown in medium containing [3H]pyridoxine (0.126 NM; 0.176 &i/ml).
Nuclear and nonnuclear fractions were analyzed for [3H]pyridoxine-derived Bs vitamer forms by ion-exchange chromatography.
Of the total radioactivity applied to ion-exchange columns in three replicate experiments, the average recovery in the form of Bg vitamers was 92% in the case of the nuclear fraction and 94% in the case of the nonnuclear fraction. The results are provided in Table  I (5). On the other hand, HTC cells grown in suspension culture with Eagle's minimal esssential medium for suspension culture, which contains pyridoxal, did not grow well when pyridoxal was replaced by pyridoxine; in addition, ["Hlpyridoxine was very poorly converted to coenzymatically active BF vitamer forms. Thus, HTC cells grown in suspension appear to require pyridoxal, a vitamer form that can readily be converted to PLP by the action of a kinase found in these hepatoma cells in an amount equal to 39% of that found in normal rat liver, as noted above.
PLP-binding Proteins in HTC Cells-A monoclonal antibody, designated E6(2)2, directed against the P-Pxy group enables the detection on Western blots of P-Pxy proteins formed by the action of reducing agents such as sodium borohydride or sodium cyanoborohydride on P-pyridoxylidene proteins (see "Experimental Procedures" and references therein). The discriminating power of E6 (2)2, as used in the present modification of the Western immunoblot detection method for P-Pxy proteins, is shown in Fig. 1. Lane 1 shows endogenous PLP-binding proteins in rat liver. Lane 2 demonstrates the reagent properties of PLP, i.e. its ability to derivatize proteins nonspecifically when it is present in excess (4). Lane 3 reveals that failure to link PLP to PLP-binding proteins reductively and covalently will result in loss of PLP during SDS-PAGE.
It is apparent from Fig. 1 that a positive reaction is dependent on the presence of PLP, the addition of a reducing agent such as sodium borohydride, and monoclonal antibody E6(2)2.
The patterns of P-Pxy proteins obtained after reduction with sodium cyanoborohydride of nuclear and nonnuclear fractions obtained from HTC cells are shown in Fig. 2. The major finding from the patterns provided in Fig. 2 is that there is one very prominent P-Pxy protein in nuclei obtained from HTC cells, having an apparent molecular mass after SDS-PAGE in the range of 50-55 kDa. This protein is in the nucleoplasmic extract obtained after hypotonic lysis of nuclei and not in the chromatin fraction, as shown in lanes B and C. This band has been found in the nuclei of HTC cells grown on plates in F-12 medium (n = 4), in the nuclei of HTC cells grown in suspension culture in Eagle's minimal essential medium for suspension culture (n = 4), and in rat liver nuclei (n = 1). many fewer bands in the nonnuclear fraction obtained from HTC cells compared with the nonnuclear fraction obtained from normal rat liver. It has been demonstrated previously that hepatomas have far fewer detectable P-Pxy proteins than normal rat liver and that the pattern of P-Pxy proteins in hepatomas resembles that found in fetal rat liver (19)(20)(21). A rationalization for the differences between normal liver and hepatomas with respect to vitamin BG metabolism and requirements has been proposed (22). HTC cells were grown in medium containing ["Hlpyridoxine. Fig. 3 provides radioautographs of a sodium cyanoborohydride-reduced SDS-PAGE-resolved cytosolic extract from such cells and a sodium cyanoborohydride-reduced SDS-PAGE-resolved nucleoplasmic extract (panel A, lanes I and 2, respectively) and a corresponding Western blot (panel B).
As noted above, we have shown that PLP is by far the major ["Hlpyridoxine-derived Bs vitamer in the nucleus of HTC cells and the only one capable of being covalently linked to nuclear proteins by reduction with sodium cyanoborohydride. It is apparent that the ["Hlpyridoxine-derived radioactive bands shown in Fig. 3 are the same as the bands detected by monoclonal antibody E6(2)2, which is specific for P-Pxy proteins.
Thus, two completely independent experimental methods indicate the presence of one major PLP-binding B6 Vitamers and PLP-binding Proteins in HTC Cells protein in the nucleoplasmic extract. Of some interest is the fact that there is a significant radioactive band that moves with the dye front in lane 2 of panel A but which is not found in the corresponding lane of the Western blot. We infer that this particular band in the cytosolic extract is comprised of free ["Hlpyridoxine-derived PLP and/or radioactive PNP, obtained by sodium cyanoborohydride reduction of ["Hlpyridoxine-derived PLP. Free PLP and PNP would be expected to be retained in the matrix of a dried polyacrylamide gel but would be expected to be lost during electroelution in a Western blot. Evidently then, there are PLP-containing proteins in cytosolic extracts of HTC cells whose PLP may not be accessible to sodium cyanoborohydride reduction. Such PLP would subsequently dissociate from proteins when subjected to the denaturing conditions of SDS-PAGE and as a small negatively charged species, would migrate with the dye front. A clear example of such a protein is PLP-containing phosphorylase b, which can be reduced to P-Pxy phosphorylase b only under conditions that are denaturing and that unfold the protein so as to make the PLP accessible to borohydride reduction (23). A second possibility is that there are PLP-containing proteins in cytosolic extracts of HTC cells whose PLP becomes reduced to diffusible PNP rather than to protein-bound P-Pxy groups. Like PLP, PNP would migrate with the dye front. Thus, it is clear that monoclonal antibody E6(2)2, which is specific for covalently bound P-Pxy residues in proteins, does not necessarily detect all of the PLP-containing proteins found in a sample. On the other hand, the nucleoplasmic extract does not contain an analogous radioactive band found in the dye front on the dried gel, and so one can infer that most, if not all, of the PLP found in the nucleus of HTC cells is protein bound, accessible to reduction by sodium cyanoborohydride, and predominantly associated with one band.

Detection of PLP-binding
Proteins in the Nucleoplasmic Extract of HTC Cells following Isoelectric Focusing and SDS-PAGE- Fig. 4 is a composite of two isoelectric focusing experiments (panels A and C) and the corresponding Western blots (panels B and D). The Western blots demonstrate the presence of one major PLP-binding protein in the nucleoplasmic extract of HTC cells. The protein of interest focused at the extreme basic end of the tube gel when the pH gradient was in the range of 5.0-6.5 (Fig. 4, panel B). Accordingly, ampholytes having a higher pH range were used in the focusing step (panels C and D). Panel D reveals that the major P-Pxy protein in the nucleoplasmic extract had a relatively high p1, which placed it at a considerable distance from the majority of the proteins in the nucleoplasmic extract, most of which focused at the acidic end of the tube gel when the pH range in the tube gel was from 6.0 to 7.4. This property should be of considerable use in the purification of this protein.
Evidence for the Existence of Higher Molecular Mass Forms of the Major PLP-binding Protein in the Nucleoplasmic Extract of HTC Cells-When SDS-PAGE gels were run in the absence of mercaptoethanol, we observed the presence of bands whose mobility corresponded approximately to what one would expect for dimeric species. Attempts to convert completely the 50-55 kDa form to the higher molecular mass form by incubation overnight in the absence of mercaptoethanol and in the presence of oxidized glutathione (5 mM in phosphate-buffered saline, pH 7.4) were not completely successful (Fig. 5). Nevertheless, it is apparent from Fig. 5 that under these conditions, there were bands that approximately corresponded to dimers of the 50-55-kDa form. One potential candidate protein is ornithine decarboxylase, a dimeric PLPdependent enzyme that has been found in nucleoplasmic extracts ( Procedures." c , 6.6.6.5.6.6,6.7.6.6.7...7.1.7.1.7.2.7 ,,.6,3.3,,.6,~~.6.l.6.2.~4,6.5.6.4.6.~, 97.4 -66.2 -42.7 -31.0-2I.5-14.4-D ,6.6.6.S.6.6.6.7.6.6.7..,7.1.7.1.7.2.7.q, by guest on July 10, 2020 . Nevertheless, with the information in hand, one cannot definitively rule out at this time an ornithine decarboxylase variant in the nucleoplasmic extract of HTC cells having an unusually high isoelectric point. DISCUSSION A fraction of the PLP found in rat liver and rat hepatoma cells resides in the nucleus. There is nothing known about its function. The results reported in this study establish that HTC cells contain pyridoxine-derived PLP in the nucleus, that PLP comprises the only significant Bg vitamer form in the nucleus, and that there is one particularly prominent PLP-binding protein in the nucleoplasmic fraction of nuclear preparations obtained from HTC cells. In addition, we have shown the presence of a corresponding protein in rat liver nuclei. One possibility is that PLP happens to leak into the nucleus somehow, and its presence there is of no physiological significance. If this were the case, one would expect PLP to be primarily and nonspecifically associated with lysine-containing histones in the nucleus. Along these lines, Pal and Christensen (28) have reported that the pellet obtained from broken Ehrlich ascites tumor cells became intensely yellow when incubated with PLP, no doubt a consequence of Schiff base formation between PLP and lysine side chains. However, there does not appear to be any significant association of PLP with the histones found in the nucleus of HTC cells grown in media containing pyridoxine.
Rather, the major PLP-binding protein is found in the nucleoplasmic extract and has an apparent molecular mass of 50-55 kDa after SDS-PAGE (Fig.  2). In addition, one would expect that PLP would be the least likely Bs vitamer form to diffuse across the nuclear membrane, as it is generally accepted that free phosphorylated forms of Bs vitamers do not readily cross membranes of mammalian cells (29). Indeed, PLP has been used as a labeling reagent for the external surface of cell membranes owing to the fact that it will spontaneously react with the side chains of lysine residues that are accessible to it on the surface of cell membranes (30). Accordingly, it is reasonable to infer that the presence of PLP in the nucleus is of physiological significance and that it gets there by a specific mechanism(s). What possible role(s) might PLP have in the nucleus of a cell? There is considerable evidence that PLP might affect steroid hormone activity by altering the interaction of steroid receptor complexes with DNA, chromatin, and nuclei (31-38). As an example, DiSorbo and Litwack have shown that rat hepatoma cells, when grown in the presence of 5 mM pyridoxine, have a significantly decreased glucocorticoid-dependent induction of tyrosine aminotransferase; the reverse was the case when the medium was depleted of pyridoxine (35). Similar findings have been reported for the glucocorticoid-dependent induction of HeLa cell alkaline phosphatase (36). Majumdar et al. (39) have shown that the addition of PLP to an incubation medium containing mouse mammary gland explants resulted in a significant inhibition of both the binding of dexamethasone to nuclear steroid receptor as well as of dexamethasone-stimulated casein mRNA synthesis. In addition to its effects on steroid receptors, PLP has been very effectively used as a site-specific affinity-labeling reagent for other DNA-binding proteins, e.g. the gene 5 DNA unwinding binding protein from bacteriophage fd (40) and for nucleotide and polynucleotide-binding sites on enzymes such as Escherichia coli DNA polymerase I (41) and the recBC enzyme of E. coli (42). It is assumed that the pyridine ring of PLP mimics the base portion of a nucleotide, with a phosphate group being common to both PLP and a nucleot,ide. Thus, there is an excellent chemical rationale for the possibility that PLP might act as a small molecular modulator of protein-DNA interactions.
It must be emphasized that the physiological significance of observations that are dependent on nonphysiological concentrations of Bs vitamers remain problematical because of the reagent properties of B, vitamer forms, particularly PLP (4). It is important to note that the findings reported in the present study were made under experimental conditions employing no excess amount of vitamin. We have no information at this time as to how PLP gets into nuclei of HTC cells or rat liver. One possibility is that it gains entry by a conventional route involving transport in an unphosphorylated form across the nuclear membrane followed by phosphorylation and binding to nuclear protein(s) (43). A second is that PLP is transported intact across the nuclear membrane, as demonstrated for PLP acquisition by mitochondria (44). A third is that it enters the nucleus only in association with a specific protein(s).
It is intriguing that the form and substance of the present report resemble in many ways those published recently in this journal for a nuclear retinoic acid-binding protein (mass, 55-60 kDa) found in the human leukemia cell line HL60 (45). Since retinoic acid, a vitamin A derivative, is a potent inducer of terminal differentiation of HL60 cells, it was postulated that the retinoylated nuclear protein may be involved in regulation of retinoic acid effects at the gene level. The possibility that PLP may exert similar effects in the nucleus via the protein we have identified in the nucleoplasmic extract of HTC cells is of considerable interest. In conclusion, this is the first substantive report concerned with PLP in the nucleus of cells. Of particular interest is the finding that there is one major PLP-binding protein in the nucleoplasmic fraction of nuclei obtained from HTC cells and rat liver. The purification and characterization of this protein are in progress.