Transport and Metabolism of Vitamin B , in Rabbit Brain and Choroid Plexus

In uitro, isolated choroid plexuses, the anatomical location of the system that transports vitamin I~,$ (BJ from blood into cerebrospinal fluid, and brain slices concentrated :‘H from media containing [3H]pyridoxine. Within these tissues, most of the intracellular [:‘H]pyridoxine was converted to and retained as the phosphorylated [“HIB,; vitamers. However, isolated choroid plexuses and brain slices were unable to concentrate “H from media containing [“H,“2P]pyridoxine-P when dephosphorylation of the [3H,Y2P]pyridoxine-P in the media was prevented. In uiuo, 2, h after the intraventricular injection of 0.8 nmol of [3H,J’P]pyridoxine-P, most of the recovered [“HIB,, in brain and cerebrospinal fluid, although phosphorylated, was not labeled with 32P. Based on this and other evidence, it is concluded that the nonphosphorylated B,, vitamers are the principal forms transported across brain and choroid plexus cell membranes. The mechanism by which brain slices and isolated choroid plexuses concentrate extracellular [“Hlpyridoxine depends on pyridoxal kinase, the enzyme that phosphorylates all three nonphosphorylated B,; vitamers, since (a) there was an excellent correlation between phosphorylation and accumulation, (h) the concentration of 4’-deoxypyridoxine to inhibit the pyridoxal kinase activity from each tissue was the same as the concentration to inhibit accumulation of [“HI& by intact tissue, and Cc) pyridoxine itself was not concentrated. These and previous results show that the regulation of accumulation of nonphosphorylated B,; vitamers is one mechanism by which the intracellular vitamin B,; levels in brain and choroid plexus are homeostatically maintained.

In uitro, isolated choroid plexuses, the anatomical location of the system that transports vitamin I~,$ (BJ from blood into cerebrospinal fluid, and brain slices concentrated :'H from media containing [3H]pyridoxine. Within these tissues, most of the intracellular [:'H]pyridoxine was converted to and retained as the phosphorylated ["HIB,; vitamers. However, isolated choroid plexuses and brain slices were unable to concentrate "H from media containing ["H,"2P]pyridoxine-P when dephosphorylation of the [3H,Y2P]pyridoxine-P in the media was prevented. In uiuo, 2, h after the intraventricular injection of 0.8 nmol of [3H,J'P]pyridoxine-P, most of the recovered ["HIB,, in brain and cerebrospinal fluid, although phosphorylated, was not labeled with 32P. Based on this and other evidence, it is concluded that the nonphosphorylated B,, vitamers are the principal forms transported across brain and choroid plexus cell membranes. The mechanism by which brain slices and isolated choroid plexuses concentrate extracellular ["Hlpyridoxine depends on pyridoxal kinase, the enzyme that phosphorylates all three nonphosphorylated B,; vitamers, since (a) there was an excellent correlation between phosphorylation and accumulation, (h) the concentration of 4'-deoxypyridoxine to inhibit the pyridoxal kinase activity from each tissue was the same as the concentration to inhibit accumulation of ["HI& by intact tissue, and Cc) pyridoxine itself was not concentrated.
These and previous results show that the regulation of accumulation of nonphosphorylated B,; vitamers is one mechanism by which the intracellular vitamin B,; levels in brain and choroid plexus are homeostatically maintained. were purified on 4-cm Amberlite CG-120 columns ( Fig. 1) (1, 2, 9).
Pyridoxal kinase was assayed on supernatants (22,000 x g for 15 min) of homogenates of brain slices or pooled choroid plexuses (10, 11). The following conditions for the assay were chosen to measure pyridoxal kinase activity at a pH approximately equal to the intracellular pH of brain and choroid plexus cells (7, 12). Briefly, after homogenization of about 10 mg of tissue in 0.5 ml of artificial CSF (2), 0.45 ml of supernatants was added to 0.35 ml of 75 mM potassium phosphate buffer (pH 6.5) containing 1 ymol of ATP, 0.5 pmol of MgC&, 50 nCi of 13Hlpyridoxine and, in some samples, various amounts of I'-deoxypyridoxine. After a 5-min (choroid plexus) or lo-min (brain) incubation at 37" in a metabolic shaker (to stay on the linear part of the phosphorylation curve), the reactions were stopped with 0. brain slices (weighing about 30 mg), or salinewashed rabbit red blood cells were incubated at 37" in a metabolic shaker in either of two media containing [:'H]pyridoxine or ["H,"'Plpyridoxine-P with methods previously described in detail (21. One medium consisted of artificial CSF with 5 rn~ glucose; the second medium consisted of 5 rnM glucose in 50% artificial CSF and 50% 110 rnM phosphate (sodium salt; pH 7.2) which we term "phosphate-CSF" (2). The later medium prevents enzymatic dephosphorylation of the phosphorylated B,; vitamers in the medium (2, 13). At the end of the incubation, the choroid plexuses, after weighing, and brain slices were homogenized in chilled 0.12 M metaphosphoric acid, and the content of 3H and :J2P in the homogenates and medium was determined (2). The red blood cells were decolorized before assaying for :'H and 32P (2). Tissue to medium ratios were obtained by dividing the "H disintegrations per min per g of tissue by the :'H disintegrations per min per ml of medium.

Injections-
To study the clearance and metabolism of intraventricularly injected pyridoxine-P, 1.35 @I!i of ["H,:'2Plpyridoxine-P (0.8 nmol) and, as a reference substance, 0.90 &i of [Hlmannitol (which is transported in the central nervous system by simple diffusion and is not metabolized (1)) in 0.1 ml of artificial CSF were injected into the left lateral ventricle of sodium pentothal-anesthetized rabbits by methods previously described (1). After 2 h the rabbit was killed, and CSF, the choroid plexus, and the left and right half of the brain were removed.
As rapidly as possible, the tissues were weighed.
The choroid plexus was then homogenized in 1.2 ml of chilled 0.12 M metaphosphoric acid, and each half-brain was homogenized in 5.0 ml of chilled 0.25 M metaphosphoric acid. One milliliter of CSF was added to 0.2 ml of chilled 1.0 M metaphosphoric acid. Then, the "H and alp content of the tissue homogenates was assayed.  (14). The CSF was ultrafiltered through the filter with 5-pound pressure with 95% N, and 5% COn at 23". Ultrafiltrate (0.1 ml) was collected after the first 0.05 ml was collected and discarded (14). As a control, 0.1 I*.M [:'H,'lYPlpyridoxine-P in artificial CSF was ultrafiltered in an identical fashion. The fraction of ["HJB, in rabbit CSF that was bound was equated to the concentration of "H in the ultrafiltrate divided by the mean concentration of "H in the parent CSF (before and after the filtration) divided by the same ratio in the control (14). A similar binding study was performed with a mixture of 50% rabbit plasma and 50% 110 rnM phosphate buffer (pH 7.3; sodium salt) which contained 0.05 PM [:'H,"zPlpyridoxine-P. The ability of 0.5 ml of freshly withdrawn rabbit CSF (or artificial CSF) to dephosphorylate pyridoxine phosphate was tested by adding 50 pmol of ["H,"'PJpyridoxine phosphate to each sample. The solutions were incubated in a metabolic shaker at 37" for 60 min under 95% 0, and 5% CO,. At the end, 0.5 ml of chilled 0.24 M metaphosphoric acid was added to each sample and the percentage of dephosphorylated ["H,:"PJpyridoxine-P was measured (1, 2).

RESULTS
In vitro, the accumulation and tissue forms of ["HlB,: within brain slices incubated in media containing ["H,"2Plpyridoxine-P are shown in Table I  of right brain, and 3 ml of CSF was calculated (1).
brain was dephosphorylated during the 2 h in vivo. A substantial proportion of the dephosphorylated [3H]BG was then rephosphorylated with nonradioactive phosphate (Table IV). When it became apparent from both indirect (2) and direct experiments (Tables I and II) that phosphorylated vitamers were not accumulated by brain slices and choroid plexus as well as the nonphosphorylated vitamers, the mechanism of PHlpyridoxine accumulation was further investigated. F'yridoxine was used because, as noted above, all three nonphosphorylated vitamers have comparable ability to depress [3H]pyridoxine transport and accumulation by brain slices and isolated choroid plexuses (2) and because, unlike pyridoxal, pyridoxine does not bind nonspecifically to proteins (16). Several mechanisms for 13HlBG accumulation were possible; [3H]pyridoxine could be transported into brain slices or choroid plexuses by active transport and then be phosphorylated and converted to pyridoxal (or pyridoxamine) phosphate (17), or, alternatively, pyridoxine could be transported into cells by simple or facilitated diffusion and then be "trapped" intracellularly by phosphorylation via pyridoxal kinase (4,18,19). We have previously presented data favoring the later interpretation (2). One criterion for demonstrating active transport of a solute involves showing that the intracellular concentration of the free solute is higher in the tissue than in the medium (19). This was not the case with pyridoxine (Tables I and II).* Therefore, we tested the possibility that 5'-deoxy-PHlpyridoxine, a nonmetabolizable analogue of pyridoxine (4, 191, might be concentrated by brain slices and choroid plexuses in vitro.
Secondly, we tested the possibility that unlabeled 5'deoxypyridoxine might depress 13Hlpyridoxine accumulation by brain slices and choroid plexus in vitro.
A significant intracellular accumulation of free 5'-deoxy13H]pyridoxine coupled with the inhibition of [3Hlpyridoxine transport by carrier 5'-deoxypyridoxine would favor active transport of pyridoxine (19); a lack of intracellular concentration of 5'deoxy[3Hlpyridoxine and a lack of inhibition of [3H]pyridoxine transport would be inconclusive since 5'-deoxypyridoxine might have minimal or no affinity for the postulated carrier. Incubation of brain slices and choroid plexuses for 30 min in artificial CSF containing 48 nM 5'-deoxy[3H]pyridoxine showed that 5'-deoxy13Hlpyridoxine was not accumulated to levels above the medium concentration in either tissue (data not shown). Moreover, 100 PM 5'-deoxypyridoxine (in the medium) did not inhibit the accumulation of [3H]B, by brain slices and choroid plexuses with 60 nM 13Hlpyridoxine in the medium (data not shown).
The ability of 4'-deoxypyridoxine to inhibit both accumulation of [3H]B, and pyridoxal kinase activity from brain slices and choroid plexuses is shown in Fig. 2. In brain and choroid plexus, about 2 and 5 to 15 FM 4'-deoxypyridoxine, respectively, inhibited both processes by 50%. This result argues  Fig. 3. [3HlB, accumulation by brain slices was depressed by three different inhibitors; addition of pyridoxine (2) or 4'deoxypyridoxine to the medium, or preincubation in pyridoxal azine (2). The correlation between the total T/M and phosphorylated T/M is shown in Fig. 3 Fig. 4 (19). The K, was 0.34 KM with a Y,,, = 2.7 pmol kg-' (30 min)-'. However, the Y,,, may slightly underestimate the true Y,,, as the saturable uptake at 30 min was only 1.82 times the 15-min uptake. In choroid plexus, the K, was 7.0 PM with a Y,,, = 40.0 pmol kg-' (30 min))'.

DISCUSSION
The principal findings reported herein are 1) that intact pyridoxine 5'-phosphate does not readily enter rabbit choroid plexuses and brain in vivo or in vitro or red blood cells in vitro and 2) that trapping of pyridoxine, pyridoxal, or pyridoxamine, after phosphorylation by pyridoxal kinase, is the probable mechanism by which the majority of vitamin B,, is accumulated by brain cells and choroid plexuses. Before discussing these results, several technical matters require comment. First, metaphosphoric acid was used to deproteinize the tissue homogenates because perchloric acid did not quantitatively maintain the phosphorylated B,, vitamers intact.' Second, the assumption that no ["'P]Pyr-P was formed inside tissues incubated in artificial CSF containing lXH,"jP1pyridoxine-P may be in error. The formation of any ["LPl-Pyr-P would decrease the amount of ["H,R"P]Pyr-P in the tissues in Tables I and II due to the methods of separation and calculation. Thus, the percentages of [:'H,:"P]Pyr-P in Tables  I and II are upper estimates since the concentration, if any, of [XLPlPyr-P in the tissues is unknown. It should be noted that most OLP organic phosphates (as glucose phosphates), if formed, would behave as orthophosphate in the separations described above and would not alter the results in Tables I  and II. Third, the rationale for using ["Hlpyridoxine prepared by an exchange procedure and for separating the phosphorylated vitamers from the nonphosphorylated vitamers has been discussed in detail (1, 2).
In rabbit brain slices, choroid plexuses, and especially red blood cells, intact pyridoxine-P was accumulated by the tissues less than 25% as well as pyridoxine (with comparable concentrations of BI; in the medium) (Tables I and II). Moreover, a significant amount of the [3H,"2Plpyridoxine-P within the choroid plexus and brain is in the extracellular space of these tissues (2). These results are consistent with previous measurements of the inhibitory potency of the B,, vitamers on ["Hlpyridoxine transport (2). In human red cells, ["H]B, has been reported to achieve T/M ratios of 6 after 1 h of incubation in medium containing 0.25 pM [3H]pyridoxine (16). However, phosphorylated B,, vitamers have been reported to enter human red blood cells by simple diffusion and to achieve T/M ratios of less than one (15, 21). These results are consistent with our findings in rabbit red blood cells.
In viuo, 2 h after the intraventricular injection of ["H,3'P1pyridoxine-P, the majority of the ["HIB, in both CSF and brain had been dephosphorylated and then rephosphorylated (Tables III and IV). The site of the dephosphorylation is uncertain, but rabbit brain slices and choroid plexuses (Tables  I and II) and, less effectively, CSF can all dephosphorylate [3H,32Plpyridoxine-P. Then brain and chroid plexus, assuming comparable behavior in uivo as in vitro (Tables I and II), could transport and phosphorylate (accumulate) ["Hlpyridoxine intracellularly.
A small amount of the [3H,3'P1pyridoxine phosphate may enter the tissues directly (Tables I and II (Tables  I and II) and in viuo depends on pyridoxal kinase (Tables  III and IV). The evidence that these [:'H]B,; accumulation processes depend on pyridoxal kinase is 3-fold. (a) There is an excellent correlation between phosphorylation and uptake (Fig. 3) Previously, the transport of ["Hlpyridoxine into brain slices and choroid plexuses has been shown to be depressed significantly by cold (2") and 1.0 mM dinitrophenol (2). Iodo-acetate (1 mM) and the substitution of N, for 0, coupled with the omission of glucose from the medium depressed the accumulation only slightly (2). However, valinomycin (1.0 pM) and NaF (10 mM) in the medium had no effect on I:'HIB,, accumulation by brain slices.' These studies are not conclusive in separating active transport from trapping because both pyridoxal kinase and probably active transport require ATP, and, therefore, depletion of ATP could affect either process.