THE JOURXAL OF 13ror,o~rcar. CHEMISTRY Vol. 247, 30. 6, Issue of Marsh 25, pp. 1880-1882, 19i2

Printed in U.S.A.

Biosynthesis of Vitamin B,*

(Received for publication, July 25, 1971)

ROBERT E.HILL,FREDERICK J. ROWELL, RAM NATH GUPTA,AND IAN D. SPENSER

From the Department of Chemistry, McMaster University, Hamilton, Ontario, Canada

SUMMARY

The biosynthesis of vitamin Bs was studied in an Esche-

richia coli mutant, blocked between pyridoxol and pyridoxal. Specific incorporation of glycerol, pyruvate, serine, and glu- cose was shown by chemical degradation of labeled samples of pyridoxol isolated from the medium of cultures which had been incubated with various 14C-labeled radiomers of these substrates.

The observed pattern of incorporation of 14C shows that pyridoxol is derived from 3 triose units. One of these is in- corporated by way of pyruvate, as a Z-carbon unit at the oxida- tion level of acetaldehyde. The other 2 triose units are incorporated intact.

A hypothetical sequence for the biogenesis of pyridoxol is advanced on the basis of the tracer evidence. FLDeoxyxylu- lose l-phosphate and a branched chain 8-carbon sugar are postulated as intermediates.

Past attempts to identify the primary precursors of vitamin Bs and to define the steps of its biosynthesis have failed to yield clear-cut results. Neither investigations with organisms con- taining metabolic blocks (3-17) nor tracer studies (10, 17-25) have led to firm conclusions. Inferences drawn by different investigators were, in some cases, mutually contradictory.

m-Alanine, L-aspartate, L-cysteine, m-glutamate, and DL-

leucine each promoted the growth of a pyridoxol auxotroph of Eschericlria coli (7). Resting cells of a soil microorganism, identified as a strain of Flavobacterium, generated approximately 3 times as much pyridoxol on incubation with glycerol together with one of L-leucine, L-aspartate, L-glutamate, or L-valine, as on incubation wit,h glycerol alone, or with glycerol in the presence of other amino acids, e.g. L- or u-alanine, L-cystine, L-serine, or glycine (9, 10, 17).

;ln early view, att’ractive on structural grounds, that alanine might serve as a precursor of pyridoxol, was based on the ob- servation that, in Streptococcus faecalis and in Lactobacillus casei, D-alanine replaced the vitamin as a growth factor (3, 4). This view had to be abandoned when it was shown that cells grown with n-alanine did not synthesize the vitamin (5), and that vitamin BB was required for formation of D-alanine, rather than the reverse (6).

* This investigation was supported by a grant from the Na- tional Research Council of Canada. Preliminary accounts of parts of this work have appeared (1, 2).

Of a series of a-amino acids whose effectiveness as a sole carbon source for the yeast, Candida albicans, was investigated, aspar- tate and glutamate induced maximal accumulation of pyridoxol (1 I). Of the Krebs cycle acids tested, malic acid and citric acid were the most effective in supporting pyridoxol production (12). With a mixture of glyoxylate and acetate as the sole carbon source, the growth rate of the yeast as well as the yield of pyri- doxol was high, whereas glyoxylate alone did not support growth. These results were interpreted to indicate that the glyoxylate cycle played a part in the biosynthesis of pyridoxol from dicarbox- ylic acids (13).

Glycolaldehyde, together with either glycine or serine, replaced vitamin Bs in promoting the growth of two Be-requiring strains of E. co&, and supported the synthesis of the vitamin (8). On the basis of a recent study of serine-requiring mutants of E. coli it was suggested that it was 3-phosphoserine, rather than serine, which may be a precursor of pyridoxol (14-16), and that activity from [3-‘4C]serine might be incorporated by way of l-carbon intermediates. However, label from [14C]formic acid or from [methyl-‘4Clmethionine did not enter the product in this system. Nor did [‘4C]aspartic acid serve as a precursor (23).

Radioactivity from [2-14C]succinic acid entered pyridoxal in the yeast, C. albicans (22), and label from [3-14C]serine, [l-14C]- alanine, and [F4C]- and [2-14C]acetic acid was reported to be incorporated into pyridoxamine in Candida utilis (18) (but see below). No attempt was made to determine the sites of labeling in these samples. Pyridoxamine, isolated from cultures of C. utilis which had been incubated with [l ,3-14C]glycerol, [3-‘4C]pyr- uvic acid, [4-14C]aspartic acid, [F4C]succinic acid, [14C]formic acid, [7-‘4Clnicotinic acid (18), [U-14C]aspartic acid,’ [7cr-14C]tryp- tophan,’ or [6-14C]quinolinic acid’ failed to show activity beyond the limits of confidence of the determination.

In another yeast, Saccharomyces jragilis, addition of aspartic acid, tryptophan, or of a mixture of glycine and glycolaldehyde spared the incorporation of label from [IV4C]glucose into pyri- doxol. Act.ivity from [V4C]aspartic acid and [1-‘“Cl- and [2-14Clacetate entered the product, whereas label from [14C]for- mate and [methyZ-14C]methionine did not (24).

In the strain of Flavobacterium, already referred to (see above), label from nL-[2-14C]aspartic acid entered pyridoxamine. The product was shown, by partial degradation, to be nonrandomly labeled. Its aminomethyl (C-4’) and hydroxymethyl (C-5’) carbon atoms were essentially free of activity (19). Surprisingly, on the other hand, it was reported very recently by the same authors that in the pyridosol, derived from [2-14C]aspartic acid in the same organism, these 2 carbon atoms contained 507, of the total activity (25).

1 S. B. Stanley (1969) Ph.D. thesis, Smith College, Mass.

1869

by guest on August 11, 2019

http://ww

w.jbc.org/

Dow

nloaded from

1870 Biosynthesis of Vitamin B, Vol. 247, X0. 6

Ho

x

0

CH3 0

C02H

c C02H

OH

OH

SCHEMES 1 TO 3. Biogenetic hypotheses of the origin of pyri- doxol, based on structural relations (26-28).

Label from L-[U-KJleucine was reported to be efficiently in- corporated into pyridoxal synthesized in this bacterium, whereas radioactivity from nL-[1-‘4C]leucine did not enter the product (10, 17, 20). Pyridoxol, obtained from the experiment with L-[U-YJleucine contained little activity at C-4’ and C-5’ (25).

Radioactivity from [%]glycerol enters pyridoxal in this or- ganism. When [2-‘4C]glycerol served as the precursor, C-4’ and C-5’ of the product were again devoid of label (20,21,25). With [I, 3-14C]glycerol as the substrate, however, the 2 carbon atoms, C-4’ plus C-5’, accounted for approximately one-half of the activity of the intact pyridoxal (21, 25).

On the basis of these incorporation data it was suggested that the pyridoxol nucleus arises by combination of a glycerol unit (yielding the fragment C-5’,-5,-B) with a Cg unit related to leucine (e.g. &hydroxy-p-methylglutaryl-CoA) or to aspartic acid (?). ,4 structural formulation of this scheme was not attempted (20, 25).

Other biogenetic schemes, advanced without experimental support, entirely on the basis of structural relations, envisaged the pyridoxol skeleton to be derived from alanine, succinate, and a l-carbon unit (Scheme 1) (26), from pyruvate, ammonia, and the isoleucine precursor, o(, p-dihydroxy+methylvaleric acid (Scheme 2) (27), or from a ketopentose, glyceraldehyde, and a nitrogen source (Scheme 3) (28).

The major difficulty in t,esting these schemes or other hypo- thetical models2 for the biogenesis of the pyridoxine skeleton,

2 Thus in Scheme 1 the Cl unit might be derived from formate, C-3 of serine, or C-2 of glycine or glyoxylate; the C3 unit from pyruvate, serine, a triose, or even from aspartate or oxalacetate (by loss of C-4 bv decarboxylation). In Scheme 2, the C3 unit mfght originate from alanine-or from any of the C3 precursors al- readv mentioned under Scheme 1. and the branched C5 unit might be d”erived from mevalonic acid, isoleucine, leucine, or valiie. Again, in Scheme 3, the CL unit might be related to citric acid, and the C3 unit might arise by decarboxylation of a C4 precursor such as aspartic acid, oxalacetic acid, or homoserine.

which can be arbitrarily constructed on the basis of structural relations of putative precursors to the product, and by analogy with the known origin of other ,naturally occurring pyridine derivatives (29), is the minute quantity of vitamin Bs (40 to 200 ng per mg of dry cell material) (30,31) which is present in systems that produce it. Success in biosynthetic tracer studies of the origin of Be is thus contingent on a high radiochemical yield in the incorporation of radioactive tracer, such that the specific activity of the product, even after many-fold dilution with inac- tive carrier, is still high enough for degradation studies.

The choice of carrier is dictated mainly by the relative nbun- dance in the experimental organism of the various forms of vita- min Be (32). Since in yeasts pyridoxamine appears to be the most abundant form of the vitamin (33,34) and since its isolation from the yeast C. utilis has been reported in detail (18), our early experiments were carried out with this yeast. Samples of pyr- idoxamine hydrochloride isolated from C. utilis (A.T.C.C. 9950) (18) after incubation in the presence of [1,3-YJglycerol or of DL-[3-‘4C]alaninf3 appeared to maintain a high and constant level of radioactivity after repeated recrystallization. However, when these purified samples were subjected to sublimation in a high vacuum as a final purification step, the sublimed samples of pyridoxamine hydrochloride so obtained were totally inactive. In each case all activity remained associated with minute amounts of nonvolatile impurities. C. utilis thus appeared to be unsuitable for a study of vitamin B6 biosynthesis.

In 1965 to 1966, W. B. Dempsey described a mutant of E. coli B (BG-2) (35, 36) blocked at the oxidation step between pyridoxol 5’-phosphate and pyridoxal 5’-phosphate. This mutant is ca- pable of synthesizing pyridoxol 5’-phosphate and pyridosol, but does not grow unless pyridoxal is supplied. When a growing culture is deprived of pyridoxal, growth ceases, but under these conditions the organism continues, for about 3 hours, to produce pyridoxol5’-phosphate and pyridoxol, at a rate 4 times that found in wild type E. coli B. The newly synthesized pyridosol and pyridoxol 5’-phosphate are excreted into the culture medium (37).

The tracer experiments, here described, made use of this E. co& B mutant.3 Labeled substrate was added to cultures at the onset of pyridoxal starvation and incubation was continued for 4 to 5 hours. Samples of labeled pyridoxol, isolated b) carrier dilution from the culture medium, were chemically de- graded by unambiguous steps in such a way that the distribu- tion of 14C within the carbon skeleton could be deduced.

The pattern of incorporation of 1% which was observed shows that pyridoxol is derived from 3 triose units. Two of these are incorporated intact, yielding t’he CP units of pyridoxol, C-3,-4,-4’ and (X-5,-6, respectively. The third, which yields the Cz unit, C-2’,-2, is incorporated via pyruvate, as a 2-carbon unit at t’he oxidation level of acetaldehyde.

MATERIALS Ar;D METHODS

Organism

The organism used in this investigation was a pyridosol auxo-

troph of E. coli, obtained from ‘vlr. B. Denlpsey.3 Its original

designation, E. coli B-BG-2 (36) was recently changed and t’he mutant has now been reclassified as E. coli B-WG2 (38). The mutant grows well on a medium supplemented with ppridosal,

but cannot utilize pyridoxol.

3 We are greatly indebted to Dr. W. B. Dempsey for supplying an isolate of E. coli B mutant WG2.

by guest on August 11, 2019

http://ww

w.jbc.org/

Dow

nloaded from

Issue of March 25, 1972 R. E. Hill, F. J. Rowe& R. N. Gupta, and I. D. Xpenser 1871

Stock cultures were maintained on monthly slants of minimal out. In fourteen of these (Experiments 1 to 5, 7 to 12, 14 to 16) medium containing glycerol (0.5%) as a carbon source, supple- the 14C-labeled substrate was dissolved in sterile distilled water mented with pyridoxal hydrochloride (6 x 10-T M). These (10 ml), and this solution was added to the culture fluid by means slants were incubated for about 48 hours at 37” and then stored of a disposable syringe. Mixtures of YXabeled compounds at 4”. were used in the two remaining experiments. In Experiment

Media 13 a mixture of nn-isoleucine and m-alloisoleucine was used. In

All bacterial stocks were maintained on a minimal medium Experiment 6 intermolecularly doubly labeled sodium [l , 3-14C&

containing glycerol (0.570) as the carbon source, and inorganic pyruvate of known distribution of activity was used which was

salts (KH2P04, 7 g; KSHPOI, 3 g; (NH&S04, 1 g; MgS04, 0.1 g; prepared as follows.

CaC&, 0.01 g, made up to 1000 ml with glass-distilled water) (39), Sodium [I , S-14Cz]pyruvate (Experiment 8)-Samples of sodium

supplemented with pyridoxal hydrochloride (final concentration [lJ4C]pyruvate (150 &i) and sodium [3-r4Clpyruvate (150 pC!i)

6 x 1OV M). This medium will be referred to as supplemented were each dissolved in sterile distilled water (10 ml). Samples

glycerol (0.5 %) medium. When solid medium was required (6 x 10 ~1) were withdrawn from each solution for liquid scin-

Bacto agar (20 g) was added. Pyridoxal solutions were sterilized tillation counting. The solutions were then mixed and made up to a volume of 25 ml with sterile water. This solution of inter-

by filtrat,ion. In all t,racer experiments the same salts medium was used, but

molecularly labeled [l , 3-14Cz]pyruvate was added to the culture

the glycerol concentration was reduced to 0.2 or 0.1% (see fluid in Experiment 6.

Table I), and pyridoxal supplementation was omitted. These Two samples (2 x 100 ~1) were withdrawn from the solution

media will be called unsupplemented glycerol (0.2%) medium of intermolecularly labeled [I, 3-14Cz]pyruvate. Each 100 ~1

and unqupplemented glycerol (0.1%) medium, respectively. was added to a solution containing sodium pyruvate (110 mg) in

To each of these solutions were added In one experiment (Experiment 1) glycerol was replaced by

distilled water (100 ml).

glucose (0.27J and in another (Experiment 14) by glucose 20 ml of a solution of phenylhydrazine hydrochloride, which

(0.1%). These media will be referred to as unsupplemented had been prepared from 0.25 ml of phenylhydrazine, made up to

The reaction mixture was glucose (0.27,) and (0.1%) medium, respectively.

50 ml with 0.1 M hydrochloric acid. allowed to stand overnight, when the crystals of [l , 3-14Cz]pyr-

Reisolation Procedure uvic acid phenylhydrazone were filtered off and dried. For

Mut.ant stocks were reisolated before each tracer experiment. Kuhn-Roth oxidation this doubly labeled material was diluted

Stock cells were suspended in sterile distilled water (10 ml) and with unlabeled carrier as follows.

the absohne count of this suspension was estimated by means of [I ,3-r4C2]Pyruvic acid phenylhydrazone (11 mg) was dis-

a hemocyt’ometer. Successive dilutions in sterile distilled water solved in aqueous sodium hydroxide (0.1 M, 1 ml) and this so-

yielded a final suspension containing approximately 200 cells lution was added to a solution of unlabeled pyruvic acid phenyl-

per ml. Samples of this suspension (0.5 ml) were spread on hydrazone (120 mg) dissolved in aqueous sodium hydroxide

plates of supplemented glycerol (0.5%) medium. The plates (0.1 111, 8 ml). Water (25 ml) was added, the solution was

were incubated at 37” to yield single colonies of approximately thoroughly mixed and filtered and water (33 ml) was added to

2-mm diameter. the filtrate. Hydrochloric acid (0.1 M, 18 ml) was added and

Two series of plates were prepared, one of supplemented the mixture was allowed to stand overnight. Crystals of the

glycerol (0.5%) medium, the other of unsupplemented glycerol diluted [I, 3-14CZ]pyruvic acid phenylhydrazone (123 mg) were

(0.57,) medium. Each plate was allotted 20 inoculation sites, filtered off and dried. A portion of this product (50 mg, specific

80 sites were used on each medium. Both media were inoculated activity, (3.47 =t 0.02) x lo4 cpm mmole-I), was subjected to

at a given site with cells from a single colony. The plates were Kuhn-Roth oxidation (by a method analogous to that described

incubat,ed (37”) and inspected for growth after 36 hours. Cells below for the Kuhn-Roth oxidation of pyridoxol hydrochloride)

from supplemented sites, which had no growth at the correspond- and the acetic acid so obtained was converted (see below) into

ing unsupplemented site, were subcultured onto slants of supple- acetyl-cu-naphthylamide (specific activity, (1.65 & 0.01) X lo4

mented glycerol (0.5%) medium. These slants were incubated cpm mmole-*; relative specific activity (intact [I, 3-14Cz]pyruvic

(37”, 24 hours) before use in a tracer experiment. acid phenylhydrazone = 100 f l), 48 f 1 To).

Growth Experiment Incubations in Presence of Labeled Xubstrates

Supplemented glycerol (0.5%) medium (250 ml) was placed Freshly isolated cells from a 24.hour slope were used to inocu-

in a l-lit.er Erlenmeyer flask. The medium was inoculated with late two 250-ml samples of supplemented glycerol (0.5%) medium

freshly isolated cells and the culture incubated on a rotary in l-liter Erlenmeyer flasks. The cultures were incubated on a

shaker (3i”, 400 rpm). The optical density of the culture was rotary shaker (36”, 400 rpm) until the optical density indicated

measured by means of a Klett Summerson photoelectric color- that growth was well into the exponential phase (approximately

imeter (model 800-3, fitted with a green filter) at zero time and 12 hours). The cells were harvested by centrifugation for 10

every hour until a constant value was observed and stationary min at 7000 x g were then washed with sterile distilled water

phase was thus evident. Distilled water was used as a blank in (3 X 100 ml).

every measurement. The resultant growth curve was used to The cells were divided into four equal parts and each was

determine the period of logarithmic growth in all subsequent resuspended in 250 ml of unsupplemented glycerol (0.2%)

incubations. medium in l-liter Erlenmeyer flasks (Experiments 2 to 7, 9 to

Labeled Compounds 13). The solution of radioactive tracer was equally divided among these flasks which were then incubated for 6 hours on a

The labeled compounds used in individual tracer experiments rotary shaker (36O, 400 rpm). are listed in Table I. Sixteen tracer experiments were carried In Experiments 8, 15, and 16 the unsupplemented glycerol

by guest on August 11, 2019

http://ww

w.jbc.org/

Dow

nloaded from

(0.2y0) medium was replaced by unsupplemented glycerol and the pyridoxol band was identified by its characteristic blue (O.lyO) medium. In Experiment 1 unsupplemented glucose fluorescence under ultraviolet light. The band was removed (0.2%) medium and in Experiment 14 unsupplemented glucose from the plate and the pyridoxol extracted from the silica gel (0.1%) medium were used. In Experiment 6 four 250-ml sam- by stirring overnight at 40” in anhydrous methanol (40 ml). ples of supplemented glycerol (0.5%) medium were inoculated The resultant slurry was filtered through a fine sintered glass and incubated, the harvested cells were then resuspended in filter. The filtrate was reduced to a small volume (1 to 2 ml) six 330.ml samples of unsupplemented glycerol (0.2%) medium. and pyridoxol hydrochloride (10 to 20 mg) was added, toget.her Incubation ronditions remained the same. with a few drops of hydrochloric acid (0.1 M). On addition of

Isolation of Pyridoxol anhydrous ether pyridoxol hydrochloride crystallized.

The product was repeatedly recrystallized from the same sol- Work-up of Bacterial Cultures-The contents of the culture vent system until successive crystal batches did not change in

flasks from each of the incubation experiments with labeled sub- specific radioactivity. In several of the experiments (e.g. Ex- strate were centrifuged 10 min at 7000 x g, and the supernatant periments 2, 4, 6, 8, 14 to 16) a further quantity of inactive solution decanted. pyridoxol hydrochloride was added as a carrier at this stage.

Cells remaining in the combined decanted solution were re- Final purification was effected by sublimation at 125-130” and moved by filtration (0.2 pm membrane filter, Nagle Co.). The 2 x 10e3 mm. The yield of sublimed pyridoxol monohydro- filtrate was concentrated to 200 ml in uacuo on a rotary evapora- chloride, melting point 205-206” (with decomposition) (reported tor and the concentrated solution was subjected to acid hydroly- melting point 204-205” (with decomposition) (40, 41) ; 206-208” sis. Sulfuric acid (1 M, about 80 ml) was added until the final (42)), corresponded, in general, to 70 to 80% of the total weight pH of the solution was 1.5, and the mixture was autoclaved of inactive carrier (Table I) which had been added. The sub- (121’, 3 hours). The hydrolysate was lyophilized and the limed product was used for chemical degradation. residue stored at 4”.

Chromatography-The lyophilized residue was dissolved in Systematic Degradation of %-Labeled

potassium acetate-acetic acid buffer (0.2 M, pH 4.5, 200 ml) and Samples of Pyridoxol

the solution was filtered through a fine sintered glass filter. Each of the labeled samples of pyridoxol, isolated from the Pyridoxol hydrochloride (5 mg) was added to the filtrate as a tracer experiments (Table I), was degraded in order to deter- carrier, and this solution was applied to a cation exchange mine radioactivity at individual carbon atoms. The methods column, which had been prepared as follows: Dowex 50-X8 employed for this purpose for the most part utilized known (200 to 400 mesh) was washed, in succession, with water, hydro- chemical reactions, which were adapted to small scale work and chloric acid (3 M), water, potassium hydroxide (6 M), and water, improved yields. until free of fine particles. This material was loaded into a column (15 X 1.5 cm) which was then washed with potassium Carbon Atoms 2’ and 2 as Acetic Acid (Scheme 4)

hydroxide (6 M), followed by water, until the pH of the eluate Kuhn-Roth Oxidation of Pyridoxol Hydrochloride (cf. Reference was approximately 9. @-Pyridoxol hydrochloride (60 mg) was dissolved in dilute

Elution of the column was carried out by stepwise increase in sulfuric acid (10 ml, 10% v/v), and chromium trioxide (2 g) was pH, with the following buffer sequence: potassium acetate added. A slow stream of nitrogen was then passed through the (0.2 $-acetic acid (0.2 M), pH 5.0, 100 ml; potassium acetate solution which was slowly distilled, while its volume was main- (0.2 @-acetic acid (0.2 M), pH 5.5, 100 ml; potassium acetate tained by repeated addition of 5-ml portions of water. Over (0.2 i&acetic acid (0.2 M), pH 6.0, 50 ml; boric acid (0.2 M)- approximately 4 hours 80 ml of distillate were collected. The potassium chloride (0.2 M) (200 ml), adjusted to pH 6.6 by addi- distillate, containing acetic acid, was titrated to pH 7 with tion of sodium hydroxide (0.2 M). sodium hydroxide (0.1 N), and the solution was then evaporated

Fractions (10 ml) were collected and assayed by ultraviolet at 90”, yielding sodium acetate (24 mg, 72’%). spectroscopy in order to determine the position of pyridoxol in Acetyl-ol-naphthylamide (C-2’,-@ (44)-A portion of the sodium the elution sequence. Pyridoxol was eluted in the first six acetate (approximately 5 mg) was dissolved in water (1 ml) and fractions following the addition of the final buffer (pH 6.6) a solution containing a slight molar excess of a-naphthylamine

A second Dowex 50-X8 (200 to 400 mesh) column (6 x 1 cm) hydrochloride (15 mg) in water (1 ml) was added, followed by was prepared and washed with hydrochloric acid (0.1 M) until I-ethyl-3-(3-dimethylaminopropyl)carbodiimide (approximately the eluate was acidic. The pyridoxol fractions from the first 40 mg). Acetyl-a-naphthylamide, which solidified on stirring, column were pooled, evaporated under reduced pressure, and was filtered off, washed with water, recrystallized from a mixture redissolved in hydrochloric acid (0.1 M, 50 ml). Pyridoxol of benzene and hexane, and sublimed at 80-90” and 10m3 mm. hydrochloride (15 mg) was added to this solution which was then Yield was 5 mg. Melting point was 158-160” (reported melting applied to the second column. The column was eluted with point 159-160” (44)). distilled water until the eluate was neutral, and then with dilute Schmidt Reaction of Kuhn-Roth Acetate and Conversion of Re- ammonia (3 %) . sulting Methylamine (C-Z?‘) into 2,&Dinitro-N-methylaniline-A

The ammoniacal eluate (10 ml) was evaporated under reduced portion of the sodium acetate (approximately 15 mg) was heated pressure and the solid residue dissolved in a small volume of with sodium azide (50 mg) and concentrated sulfuric acid (1 ml) anhydrous methanol (2 ml). This solution was applied to a on the steambath in a flask attached to a system of gas traps. preparative plate for thin layer chromatography (thickness 2 The carbon dioxide which evolved was passed by means of a mm, Silica Gel G according to Stahl). Development was for stream of nitrogen into potassium hydroxide solution (10%). 15 cm in the solvent system tert-butyl alcohol-methyl ethyl When gas evolution had ceased, the potassium hydroxide solu- ketone-ammonia (0.880)-water (4 : 3 : 2 : 1). The plate was dried tion in the traps was replaced by methanol (15 ml) to which

1872 Biosynthesis of Vitamin B6 Vol. 247, No. 6

by guest on August 11, 2019

http://ww

w.jbc.org/

Dow

nloaded from

Issue of March 25, 1972 R. E. Hill, F. J. Rowe& R. N. Gupta, and I. D. Xpensev 1873

2,4-dinitrofluorobenzene (0.1 ml) and sodium bicarbonate (0.5 g) had been added. The acidic reaction mixture was cooled to near freezing point in a Dry Ice-acetone bath, basified (pH > 12) by addition of potassium hydroxide solution (lo%), and was then heated 3 hours on the steambath while the syst,em was being flushed with nitrogen.

The methanol solutions from the traps were combined and evaporated. The residue was dissolved in ether (5 ml), and the ether solution washed with water (3 x 1 ml), and concentrated. 2,4-Dinitro-N-methylaniline, which crystallized on standing, was purified by sublimation at 120-140” and 3 x lop3 mm. Yield was 8 mg. Melting point was 176-177” (reported melting point 177” (45)).

Carbon Atoms 4’ and 4, together with Carbon Atoms 2’ and 2, as Acetic Acid (Scheme 5)

2,4-Dimethyl-S-hydroxy-Lhydroxymethylpyridine (4’-Deoxypyri- dozol) (2) (46)-A mixture of pyridoxol hydrochloride (60 mg) and anhydrous hydrazine (95%, 5 ml) was refluxed 18 hours with exclusion of moisture. The solution was evaporated in vucuo at loo”, and the residue was dissolved in a mixture of methanol and tetrahydrofuran, from which hydrazine hydro- chloride crystallized. The filtrate was evaporated, the residue dissolved in tetrahydrofuran, and a solution of hydrogen chloride in anhydrous ether was added. 4’-Deoxypyridoxol hydrochlo- ride which precipitated was recrystallized from methanol-ether. Yield was 45 mg (81%). Melting point was 267-269” (reported melting point 267-268” (47) ; 274” (with decomposition) (46)).

Kuhn-Roth Oxidation of Q’-Deoxypyridoxol-4’-Deoxypyridoxol hydrochloride (42 mg) was oxidized, by the method described above for the Kuhn-Roth oxidation of pyridoxol hydrochloride. The sodium acetate (22 mg, 60%) so obtained was converted into acetyl-cu-naphthylamide (C-2’,-4’;2,-4) and into 2,4-dini- t,ro-Wmethylaniline (C-2’;4’).

Carbon 5’ as Benzoic Acid (Scheme 6)

3,4’-0-Isopropylidenepyridoxol (3) (48)-Anhydrous hydrogen chloride was passed through a cooled (0”) stirred suspension of pyridoxol hydrochloride (50 mg) in dry acetone (5 ml) until the solid had dissolved (approximately 5 min). The solution was kept 3 hours at -lo”, anhydrous ether (10 ml) was then added, and the mixture was kept at -10” for a further 30 min. The product, 3,4’-0-isopropylidenepyridoxol hydrochloride, was washed with ether and dried. Yield was 55 mg (94%). Melt- ing point was 209-211” (reported melting point 217-218” (with decomposition) (49) ; 210-212” (with decomposition) (48)). The hydrochloride was converted into the free base by precipitation with sodium bicarbonate. Yield was 44 mg (87%). Melting point was 109-110” (reported melting point 108-109’ (50), lll- 112” (48), 113-115” (49)).

5-Formyl-S- hydroxy-4-hydroxymethyl-S,k’-O-isopropylidene-Z- methylpyridine (3,4’-0-Isopropylideneisopyridoxal) (4) (5l)- 3,4’-0-Isopropylidenepyridoxol (40 mg) was carefully dried in vacua and dissolved in anhydrous pyridine (5 ml) containing chromium trioxide (vacuum-dried, 31 mg). The mixture was heated slowly, under exclusion of moisture, so that it reached reflux temperature after 30 min, and was then refluxed 90 min. The mixture was cooled to room temperature and filtered, the filtrate was evaporated in vacua, the residue was dissolved in saturated aqueous sodium bicarbonate (15 ml) and the solution extracted with chloroform (3 x 15 ml). The chloroform extract

was dried (anhydrous potassium carbonate) and evaporated, and the residue distilled at 70-80” and 3 x 1O-3 mm, yielding 3,4’- 0-isopropylideneisopyridoxal as a colorless liquid which crystal- lized on standing. Yield was 29 mg (70%). Melting point was 61-62” (reported melting point 62-63” (52), 64-65” (51)).

S- Hydroxy-4- hydroxymethyl-5-a- hydroxybenzyl-3,4’- 0-isopro - pylidene-2- methylpyridine (3,4’- 0- Isopropylidene- 5’-phenylpyri- doxol) (5)-3,4’-0-Isopropylideneisopyridoxal (25 mg) was dis- solved in anhydrous tetrahydrofuran (2 ml), and phenyl mag- nesium bromide (60 mg) in tetrahydrofuran (1 ml) was added. The solution was refluxed 2 hours and dilute hydrochloric acid (0.1 M, 5 ml) was then added. The acid solution was extracted with chloroform (3 x 25 ml) and was then carefully basified with saturated sodium bicarbonate solution. The basic solution was extracted with chloroform (3 x 25 ml) and the extract dried (anhydrous potassium carbonate) and evaporated. The residue was sublimed in vacua at 100-120” and 3 x 10e3 mm to give the 5’.phenyl derivative as a colorless solid, melting point 159-160”. Yield was 28 mg (80%). Mass spectrum: M+ at m/e 285; infrared: yCHC13 3600, 3010, and 872 cm-l; nuclear magnetic resonance (CDC13): 6 1.41 (s, 6H), 2.32 (s, 3H), 4.62 (m, 3H), 5.69 (s, lH), 7.24 (s, 5H), 7.84 (s, 1H) ppm.

Calculated: C 71.56, H 6.71, N 4.91 Found: C 71.80, H 6.72, N 4.76

Benzoic Acid (C-5’) by Oxidation of the 5’-Phenyl Derivative- The 5’-phenyl derivative (25 mg) was suspended in aqueous potassium permanganate (250 mg in 10 ml) and the mixture was heated 12 hours at 100” with stirring. Excess permanganate was destroyed by addition of sodium bisulfite and sulfuric acid (lo%, v/v). The colorless solution was extracted with ether (3 x 25 ml), the ether extract was washed with water (10 ml), dried (anhydrous sodium sulfate), and evaporated. The residue was sublimed at 60-70” and 3 x 10V3 mm, to yield benzoic acid, 9 mg (70%), melting point 121-122”.

Carbon 4’ (Scheme 7)

3.0-Methylpyridoxol (6) (41, 53)-Pyridoxol hydrochloride (120 mg) in methanol (5 ml) was added to a solution of diazo- methane in ether (20 ml). After 12 hours the solution was evaporated, the residue was dissolved in methanol (1 ml), and ether was added. A small amount of solid which precipitated, presumably N-methylpyridoxol betaine (54)) was filtered off. The filtrate was evaporated and the residue sublimed at lOO- 120” and 3 x 10-a mm, to yield 0-methylpyridoxol. Yield was 65 mg (61%). Melting point was 88-90” (reported melting point 89.5-90” (53); 101-102” (41)).

S-Methoxy-2-methylpyridine-4,5-dicarboxylic Acid (7) (41, 43, 55)-3-0-Methylpyridoxol (50 mg) was added to aqueous po- tassium permanganate (184 mg in 10 ml) and the mixture stirred 10 hours at room temperature. Excess permanganate was de- stroyed by addition of methanol at 6&70”, the mixture was filtered, and the filtrate was concentrated in vucuo to approxi- mately 5 ml. The solution was applied to a column (1 x 10 cm) of anion exchange resin (Dowex I-formate), the column developed with aqueous formic acid (0.25 M, 40 ml), and the eluate dis- carded. The desired product was then eluted with aqueous formic acid (5 M). Elution was continued until the eluate no longer showed absorption at 280 nm. The eluate was evaporated

by guest on August 11, 2019

http://ww

w.jbc.org/

Dow

nloaded from

1874 Biosynthesis of Vitamin Bg Vol. 247, n-o. 6

i7~ vucuo and the residue crystallized from hot water, yielding weighed to the nearest microgram on a Micro Gramatic balance colorless crystals of 3-methoxy-2-methylpyridine-4,5dicarbox- (Mettler Instrument Corp.)) was weighed on aluminium plan- ylic acid. Yield was 43.5 mg (61’%). Melting point was 219- chettes (diameter 3.2 cm) and then dissolved in suitable solvents. 221” (with decomposition) (reported melting point 209-210” Aqueous sucrose solution (10% w/v) was used for pyridoxol {with decomposition) (41) ; 222” (with decomposition) (55)). hydrochloride. Dimethylformamide containing collodion (1 “ic)

S-Methoxy-6-methylpyridine-5-carboxylic acid (8) (55)-A sus- was found to be suitable for all other substances. Even distri- pension of 3-methoxy-2-methylpyridine-4,5-dicarboxylic acid bution of the samples was ensured by the lens tissue technique, (40 mg) in nitrobenzene (4 ml) was heated under nitrogen at and the planchettes were thoroughly dried before counting. 190”, until gas evolution had ceased (approximately 1 hour). Liquid scintillation counting was carried out on a Mark I The product which crystallized on heating was washed with Liquid Scintillation Computer (Nuclear Chicago Corp., model benzene and ether and was then sublimed at 150-160” and 3 x 6860). Aqueous samples were dispersed with the aid of “Aqua- 1O-3 mm to give 3-methoxy-2-methylpyridine-5-carboxylic acid sol” (10 ml) (New England Nuclear Corp.). as a colorless solid. Yield was 13 mg (40%). Melting point Confidence limits shown in the tables are standard deviation was 223-224” (reported melting point 224” (55)). Mass spec- of the mean. trum: M+ at m/e 167; nuclear magnetic resonance ((CD&SO) : 6 2.40 (s, 3H), 3.86 (s, 3H), 7.62 (d(J = 0.5), lH), 8.52 (d(J = RESULTS AND CRITERIA FOR THEIR INTERPRETATION

O.5), 1H) ppm. The samples of pyridoxol hydrochloride which were isolated

Determination of Radioactivity by the carrier dilution technique from the medium of cultures of the E. co& B mutant WG2 after incubation with various 14C-

Radioactivity was assayed, on samples of finite thickness on labeled substrates, maintained radioactivity after vigorous puri- aluminium planchettes, with a low background gas flow Geiger fication in all but 2 of the 16 tracer experiments. Details of comlting system (Nuclear Chicago Corp., model 4342). The these experiments are presented in Table I. counting efficiency for 14C was approximately 30%. Corrections It should be noted that the specific activity of pyridoxol hy- for background and self-absorption were applied. Samples for drochloride given in Table I is that of the purified sample isolated eount,ing were prepared as follows. Solid material (0.5 to 3 mg, by carrier dilution and not of the original biosynthesized product.

TABLE I Incorporation of labeled substrates into pyridoxol

Experi- ment No.

1

2 15 16 14

3 4

5

B

8 7

11 10

13

9 12

.I-

-

Substrate

[1,3-14C]Glycerolb [l, 3-W]Glycerolc

[1,3-W]Glycerolb [2-W]Glycerolb n-[l-14C]Glucoseb Sodium [3-W]pyruvateb Sodium [3-W]pyruvateb

Sodium [2-W]pyruvateb Sodium [l, 3-Wz]pyruvated

from Sodium[l-‘%]pyruvateb

I Sodium[3-W]pyruvateb oL-[3J4C]Serineb Sodium [2J4C]acetateb

oL-[3-14C]Aspartic acid8 DL-[2J4C]Leucinec DL-[3,3’ 4 5-14C4]Isoleucinec , I ~~-[3,3’,4,5-~~C4]Alloisoleucine~ Sodium [W]formateb L-[methyZ-14C]Methionineb

Nominal total

activity

??Ci ?XCi/?WdC

0.1 15.4

0.5 (10) 0.023 1.0 (15.3) 0.092 1.0 (10.5) 0.091

0.5 (3.0) 0.087 0.1 35.6 0.1 35.6

0.1 31.7 0.3 30.8 0.15 27.2

0.15 35.6 0.2 5.90 0.1 54.7 0.1 1.9 0.1 48.0

0.1 160 0.1 160 0.1 36 0.2 53.6

Culture conditions

General carbon Total Carrier Source volume added

70 w/n Glucose (0.2)

Glycerol (0.2) Glycerol (0.1) Glycerol (0.1)

Glucose (0.1) Glycerol (0.2) Glycerol (0.2)

Glycerol (0.2) Glycerol (0.2)

Glycerol (0.1) Glycerol (0.2) Glycerol (0.2)

Glycerol (0.2)

Glycerol (0.2 )

Glycerol (0.2) Glycerol (0.2)

liters

1

1 1 1 1

1 1 1 2

1

1 1 1

1

1

1

Pyridoxol hydrochloride

_ g Q%?%/??R%OlC

0.04 1.78 f 0.05 0.12 2.38 f 0.06 0.66 1.67 f 0.03 1.0 1.09 f 0.03 1.0 0.46 f 0.01

0.04 1.20 f 0.03 0.09 0.54 f 0.01 0.04 0.72 f 0.02

0.08 1.05 f 0.02

0.15

0.04 0.04 0.03

0.05

0.04

0.03

0.29 f 0.01

0.36 f 0.01 0.25 f 0.01 0.31 f 0.02

Inactive

0.16 f 0.01

Inactive

3.5

2.8 5.4 5.3 4.5

2.3 2.4 1.4 1.4

1.1

0.7 1.0’ 0.9’

0.3

a Total activity recovered = specific activity of product (counts mine1 mmole-l) X carrier added (milligrams) f molecular weight of product (milligrams mmole-1).

b Amersham-Searle. c Commissariat a 1’Energie Atomique, France. d For determination of distribution of activity see “Materials and Methods.”

e Calbiochem. f Calculated on the assumption that only the L enantiomer of the substrate is utilized.

by guest on August 11, 2019

http://ww

w.jbc.org/

Dow

nloaded from

Issue of March 25, 1972 R. E. Hill, F. J. Rowe& R. N. Gupta, and I. D. Spenser 1875

ExpeGxnt

3 [3-‘4CJPyruvic acid 4 [3-l%]Pyruvic acid 5 [2-W]Pyruvic acid 6 [l , 3-Wz]Pyruvic acid’ 8 nL-[3J4C]Serine 7 [2J4C]Acetic acid

11 nL-[3-14C]Aspartic acid 10 ~~-[2-‘~C]Leucine 9 [14C]Formic acid

-

_

TABLE II &?ecijic activity of pyridoxol and its degradation products. C-2’ and C-2, derived from various substrates

Substrate Pyridoxol” SAd

1.20 f 0.03 0.54 f 0.01 0.72 i 0.02 1.05 f 0.02 0.29 f 0.01 0.36 f 0.01 0.25 f 0.01 0.31 f 0.02 0.16 f 0.01

--

Acetic acidb (C-Z’,C-2)

T SAd

1.15 f 0.03

0.68 rt 0.02 0.94 f 0.02 0.27 f 0.01 0.21 zlz 0.01 0.15 f 0.01 0.08 f 0.004 0.04 f 0.002

RSAe

96 f 3

94 f 3 90 * 2 92 f 4 58 zk 2 59 f 2 24 rt 2 24 rt 2

MethylamineC (C-2’)

SAd

0.47 * 0.01

0.87 zt 0.02

RSAC

88 f 2

84 f 2

a Assayed as pyridoxol hydrochloride. b Assayed as acetyl-a-naphthylamide. c Assayed as 2,4-dinitro-N-methylaniline. d SA = specific activity (counts min-’ mmole-l) X 10-‘. e RSA = relative specific activity (per cent) (pyridoxol = 100). f For distribution of activity within this sample see “Materials and Methods.”

For calculation of the latter value an accurate assay of the amount of pyridoxol generated in each experiment (approxi- mately 150 f 50 pg) would have been required. Since knowl- edge of the pyridoxol content of the cultures was not crucial to the present investigation and since the requisite microbiological assay is time-consuming and tedious, such determinations were not carried out. The specific activity of the newly biosyn- thesized pyridoxol is therefore unknown, and without this value meaningful radiochemical yields cannot be calculatedP

Another index of incorporation efficiency is available, however. Since the weight of added carrier was at least two orders of magnitude larger than the approximate weight of pyridoxol originally present in the culture, and since the specific activity of the labeled pyridoxol isolated after carrier dilution is inde- pendent of the chemical yield obtained in the isolation, the total activity (counts min-l) entering the product from a given sub- strate (i.e. carrier added (milligrams) x specific activity of product (counts min-l mmole-l) /molecular weight of pyridoxol hydrochloride (205.5 mg mmole-l)) serves as an approximate measure of the relative efficiency of different substrates as pyr- idoxol precursors. On this basis glycerol (Experiments 1, 2, 15, 16) and glucose (Experiment 14) are the most efficient precursors (total activity recovered in the product per 0.1 mCi of substrate: 2.8 x lo3 to 5.4 x lo3 cpm). Pyruvate (Experiments 3 to 6) and serine (Experiment 8) fall into a second group (1.1 X lo3 to 2.4 X 1Oa cpm). From acetate (Experiment 7), leucine (Experiment lo), aspartate (Experiment 11) or formate (Experi- ment 9) little activity entered the product ( <l.O x lo3 cpm). The product from the experiments with isoleucine (Experiment 13) and with methionine (Experiment 12) was inactive.

Distribution of label within the product was investigated (Table II) by partial degradation (Scheme 4) of the radioactive samples of pyridoxol by Kuhn-Roth oxidation. The acetic acid (isolated as the a-naphthylamide) so obtained represents C-2’ plus C-2. Further degradation of this acetic acid by a Schmidt reaction yields C-2’ of pyridoxol as methylamine (isolated as the

4 For this reason it is unnecessary, and indeed undesirable, to report specific activities of product and substrates in identical units.

2,4-dinitrophenyl derivative). Random incorporation of label into pyridoxol would have supplied one-eighth of its total activity to each of its carbon atoms, and would have resulted in the re- covery of one-quarter of the total activity in the Kuhn-Roth acetate (C-2’,-2) and one-eighth in the methylamine (C-2’) obtained therefrom.

The Kuhn-Roth acetate derived from the pyridoxol from the incubations with [2-Wlleucine (Experiment 10) and formate (Experiment 9) did indeed contain one-quarter of the total activity. Lack of material precluded further degradation of the acetate samples.

The result of Kuhn-Roth oxidation showed that distribution of label in all other samples of pyridoxol was nonrandom. The acetic acid (C-2/,-2) isolated from the samples of pyridoxol de- rived from pyruvate (Experiments 3, 5, and 6) and serine (Es- periment 8) contained 90% or more of the activity of the intact vitamin. Further degradation of the acetic acid to methyla- mine showed that almost all the activity of the pyridoxol derived from [3-Wlpyruvate (Experiment 4) and from [l , 3J4Cz]pyruvate (Experiment 6) was confined to C-2’. The pyridoxol derived from [2-Wlacetate (Experiment 7) or from [3-14C]aspartate (Experiment 11) contained approximately 60% of its total activity within the Cz unit, C-2’,-2.

More extensive degradation of the samples of pyridoxol de- rived from [l , 3-14C]glycerol (Experiments 1, 2, 15), [2-14C]glyc- erol (Experiment 16), and [1-14C]glucose (Experiment 14) per- mitted assay of activity not only at C-2’ and C-2 (Scheme 4), but also at C-4’ (Scheme 7), C-5’ (Scheme 6), and, somewhat indirectly,6 at C-4. The results of t,hese degradations are pre-

6 The yield of acetate on Kuhn-Roth oxidation of 4’-deoxypyri- doxol (2) was found to be 1.2 to 1.4 eq per mole (c.f. “Materials and Methods,” 60% yield in a typical experiment) (Scheme 5). It follows that at least a small fraction of this acetate must be de- rived from the C!z fragment, C-4’,-4. The actual contribution made by the fragment, C-4/,-4, to the isolated acetate can be deduced from Table V. Since virtually all activity of serine-de- rived pyridoxol resides in the CZ unit, C-2’, -2 (as shown by Kuhn- Roth oxidation of this pyridoxol), the CP unit, C-4’, -4, of this sample contains little, if any, activity. Since the specific activity of the acetate isolated by degradation of the 2,4-dimethylpyridine derivative 2, obtained from serine-derived pyridoxol, was less

by guest on August 11, 2019

http://ww

w.jbc.org/

Dow

nloaded from

18i6 Biosynthesis of Vitamin Bs Vol. 247, No. 6

Kuhn-Roth ’ :H,&~,H

Schmidt

oxidation reaction + ~33~~13~

isolated as isolated as

1 acetyl-a- 2,4-dinitrophenyl

naphthylamide derivative

SCHEME 4

CHzOH Kuhn-Roth

oxidation

Schmidt reaction

:: CHNHz

1 2

SCHEMES

1 3 4 5

SCHEME 6

1 6 7 8

SCHEME 7

SCHEMES 4 TO 7. Degradation of pyridoxol permitting isolation of C-2’, C-4’, C-5’, C-2, and C-4.

sented in Tables III and IV. Pyridoxol derived from [I, 3-14C]- glycerol contained approximately one-fifth of its activity at each of C-2’, C-4’, and C-5’, whereas C-2 and C-4 were essentially without label (Tables III and IV). Pyridoxol derived from [2J4C]glycerol, on the other hand, contained approximately one-third of its activity at each of C-2 and C-4, while C-2’, C-4’, and C-5’ were inactive, within experimental error (Table IV). The entire activity of the sample of pyridoxol derived from n-[1-i%]glucose is accounted for in terms of the activities of C-2’, C-4’, and C-5’ (Table IV). Each of C-2’ and C-4’ contain approximately 37% of the total activity, whereas C-5’ accounts for 26% of label.

than 667, of that of the intact vitamin (Table V), at least one-third of the acetate, obtained in this way, must have been derived from the C2 unit, C-4’, -4. It is evident that if the labeling pattern in the Cz unit, C-4’,-4, of a given pyridoxol sample is significantly different from that of the CZ unit, C-2’, -2, the specific activities of the samples of acetate and methylamine obtained from it by Schemes 4 and 5 must differ. Conversely, if these specific activi- ties do not differ significantly, the labeling patterns of the Cs units, C-4’, -4 and C-2/,-2, must be similar. Since degradations are available permitting individual assay of activity at C-2’, C-2, and C-4’, it follows that activity at C-4 can be deduced.

DISCUSSION

General-When deprived of exogenous pyridoxal an exponen- tially growing culture of E. coti B mutant WG2 ceases to grow but continues to generate pyridoxol at an enhanced rate for 3 to 4 hours. This newly formed pyridoxol is shed into the culture medium (37).

These characteristics made the organism a potentially valuable tool in the investigation of the biosynthesis of pyridoxol by tracer methods, and in the search for the primary precursors of the vitamin, in particular. Since, in the absence of added pyr- idoxal, there is no growth, little if any radioactive label should be dissipated into macromolecules if tracer is added at the onset of pyridoxal starvation. Since the duration of metabolic activity following pyridoxal deprivation is short, randomization of label should be minimal. Since pyridoxol is synthesized at an unusu- ally high rate, the chances of incorporation of label into the desired product are enhanced. Since pyridoxol is excreted into the culture medium, labeled product would be removed from the site of metabolic activity. Lastly, the isolation of pyridoxol from the medium is far less tedious than its isolation from intact cells.

by guest on August 11, 2019

http://ww

w.jbc.org/

Dow

nloaded from

1877 Issue of March 25, 1972 R. E. Hill, F. J. Rowe& R. N. Gupta, and I. D. Spewer

TABLE III

Distribution of activity in pyridoxol derived from [i,S-“C]glycerol. Relative specific activity at C-d’ and C-2

Product Experiment 1 Experiment 2

(carbon atoms of pyridoxol) SA” RSAb SA RSA SA RSA

Pyridoxolc (all). 1.78 rt 0.05 100 f 3 2.38 f 0.06 100 f 2 1.12 f 0.03d 100 f 3 Acetic acide (C-2’,-2). 0.39 f 0.01 22 f 1 0.51 f 0.01 21 f 1 Methylaminef (C-2’). 0.20 f 0.01 18 f 1

a SA = specific activity (counts min-’ mmole-I) X lo+.

b RSA = relative specific activity (per cent) (pyridoxol = 100). c Assayed as pyridoxol hydrochloride.

d Obtained from the sample of pyridoxol, specific activity (2.38 f 0.06) X lo4 cpm mmole-r (Experiment 2), by further dilution with inactive pyridoxol.

6 Assayed as acetyl-a-naphthylamide. f Assayed as 2,4-dinitro-N-methylaniline.

TABLE IV

Distribution of activity in pyridoxol, derived from glycerol and glucose

Product (carbon atoms of pyridoxol) [1,3-WjGlycerol Experiment 15

Pyridoxolc (all). ............................... Acetic acid” (C-2’,-2). .........................

Methylaminea (C-2’) ...........................

4’-Deoxypyridoxol (2) (all). ...................

Acetic acid” (C-2’,-4’,-2,-4). ................... Methylaminee (C-2’, -4’). .......................

Pyridoxolc (all). ............................... 5’.Phenyl derivative (5) (all). .................. Benzoic acid (C-5’). ............................

Pyridoxolc (all). ...............................

0-Methylpyridoxol (6) (all). ................... 4,5-Dicarboxylic acid (7) (all) .................. 5-Monocarboxylic acid (8) (all but C-4’) ........

- SA5 RSAb

1.67 f 0.03 100 f 2 See Table III

See Table III

1.63 f 0.04 0.36 f 0.01

0.31 f 0.01

1.67 f 0.03 1.64 f 0.02

0.37 f 0.01

0.94 f 0.01’

0.93 f 0.02 0.94 f 0.03 0.73 f 0.02

I -

98 f 3 22 f 1 19 f 1

100 f 2 98 f 2 22 SC 1

100 f 1

99 f 2 100 f 3 78 f 2

/

Substrate

[2-W]Glycerol Experiment 16

SA” RSAb

1.09 f 0.03 0.38 f 0.01

0.03 rt 0.004

1.10 f 0.03 0.35 f 0.01

1.09 f 0.03 1.08 f 0.05

0.005 f 0.002

0.76 f 0.04g

0.79 f 0.03

100 f 3 35 f 1

2.4 r!z 0.4

101 f 4 32 zk 1

100 f 3 99 f 5

0.5 f 0.2

100 f 5

103 f 6

I n-[l-“ClGlucose Experiment 14

SA”

0.46 f 0.01 0.18 f 0.01

0.17 f 0.004

0.17 i 0.004 0.16 i 0.004

0.46 i 0.01

0.12 f 0.01

0.36 f 0.01” 0.36 i 0.01 0.23 f 0.01

RSAb

100 f 2 39 f 2

36 f 1

37 f 1

36 f 1

100 f 2

26 f 1

100 f 4

100 f 5 62 f 3

u SA = specific activity (counts min-’ mmole-r) X 10-4.

b RSA = relative specific activity (per cent) (pyridoxol = 100). c Assayed as pyridoxol hydrochloride.

d Assayed as acetyl-cr-naphthylamide. 6 Assayed as 2,4-dinitro-N-methylaniline. f Obtained from the sample of pyridoxol, specific activity (1.67 f 0.03) X 104 cpm mmole-r (Experiment 15), by further dilution

with inactive pyridoxol.

Q Obtained from the sample of pyridoxol, specific activity (1.09 f 0.03) X lo4 cpm mmole-i (Experiment 16), by further dilution with inactive pyridoxol.

* Obtained from the sample of pyridoxol, specific activity (0.46 f 0.01) X lo4 cpm mmole-’ (Experiment 14) by methylation, fol-

lowed by dilution with inactive 0-methylpyridoxol.

The preliminary experiments confirmed these expectations.

Radioactive pyridoxol, which maintained high and constant activity after repeated crystallization and resublimation, was obtained when [l ,3-i4C]glycerol served as the sole carbon source (Experiment 2), as well as when [l ,3-14C]glycerol served as the tracer while glucose served as the major carbon source (Experi- ment 1).

Having found conditions for incorporation of activity into pyridoxol from a labeled substrate which, at this early stage of

the investigation, was regarded solely as a general carbon source, entry of label from a series of putative specific precursors (c.f.

Schemes 1 to 3) was then examined (Experiments 3, 7 to 13). As a result of these experiments several of the substrates were dismissed from further consideration, since, on the basis of the criteria already discussed (see “Results and Criteria for Their Interpretation”), incorporation of their activity was minimal and random (Tables I and II). Thus, hypotheses demanding l-carbon units (e.g. Scheme 1) (methionine (Experiment 12, no

by guest on August 11, 2019

http://ww

w.jbc.org/

Dow

nloaded from

1578 Biosynthesis of Vitamin B6 Vol. 247, h-o. 6

TABLE V

Degradation of pyridoxol derived from DL-[S-Wlserine (Experiment 8)

Product (carbon atoms of pyridoxol) SAa SA RSA

Pyridoxolc.. 0.29 f 0.01 100 f 3 0.118 f 0.004” 100 f 3 Acetic acid6 (C-

2’,-2). _. ._.. 0.27 f 0.01 92 f 4 4’-Deoxypyri-

doxol (all). . 0.117 f 0.005 100 f 5 *teetic acid’ (C-

2’,-4’,-2;4). 0.074 f 0.002 63 f 3

n SA = specific activity (counts mine1 mmole-I) X 10-4.

b RSA = relative specific activity (per cent) (pyridoxol = 100). c Assayed as pyridoxol hydrochloride.

d Obtained from the sample of pyridoxol, specific activity (0.29 f 0.01) X lo4 cpm mmole-I, by further dilution with inactive pyridoxol.

e Assayed as acetyl-cr-naphthylamide.

incorporation), formate (Experiment 9, low and random incor-

poration)) or branched chain substrates (e.g. Scheme 2) (leucine

(Experiment 10, low and random incorporation) ; isoleucine

(Experiment 13, no incorporation)e) as precursors appeared to be

eliminated.

Role of Pyruvate-Of the substrates whose incorporation was nonrandom (Table II) only pyruvate approached glycerol in its radiochemical efficiency of incorporation into the product. Since, even more significantly, the activity of pyridoxol, derived from [3J4C]pyruvate, was located entirely within the CZ unit, C-2’,-2, (Experiment 3, Table II) as predicted by several bio-

genetic hypotheses (Schemes 1 and 2, c.j. Footnote 2) which feature pyruvate or a closely related Cs unit as the hypothetical precursor of the a-carbon fragment, C-2’,-2,-3, of pyridoxol, a more detailed examination of the mode of incorporation of the carbon chain of pyruvate into pyridoxol was mandatory. Two further experiments strengthened the evidence in support of a direct biosynthetic relationship of pyruvate with pyridoxol (Table II). Almost the entire activity7 derived from [2-l%]- pyruvate (Experiment 5), like that from [3-Wlpyruvate (Experi- ment 3), was located within the Gil unit, C-2/,-2, of pyridoxol. Since almost all the activity derived from [3-‘4Clpyruvate was confined to C-2’ (Experiment 4), it can be concluded that the activity from [2-14C]pyruvate resides at C-2 of pyridoxol. Thus, the C-methyl group (C-2’) is supplied by the methyl group (C-3) of pyruvate, and C-2 of pyridoxol by the carbonyl carbon atom (C-2) of pyruvate. It is likely that an intact Ct unit is involved.

It remained to examine whether the carboxyl carbon (C-l)

6 A third branched chain substrate, mevalonic acid, was re- moved from immediate contention on the basis of inferences drawn from the result of the experiment with its progenitor, acetate. The low level of incorporation of activity from [2-‘4Clacetate into pyridoxol (Experiment 7) and the preferential incorporation of this activity into the CZ unit, C-2’, -2, (Table II), which on struc- tural grounds cannot be part of a mevalonate-derived unit (c.f. Scheme 2 and Footnote 2) made it very improbable that mev- alonate served as a direct precursor of thk vitamin.

i The observed deviation from the theoretical value of the relative specific activity of the degradation products (5 to 10%) is presumably due to a background of low random scatter of label throughout pyridoxol, as a consequence of metabolic turnover of the labeled substrates during incubation.

of pyruvate serves as a source of C-3 of pyridoxol, i.e. whether the intact 3-carbon chain of pyruvate enters the vitamin (Schemes 1 and 2). Since a chemical degradation of pyridoxol, capable of extracting C-3, has not yet proved practicable, an experiment with [l-Wlpyruvate would have yielded an incon- clusive result. Instead, a doubly labeled sample, [l ,3-YJpyr- uvate, containing 48 =t 1% of its label at the methyl group (C-3) and 52 f 1% at the carboxyl group (C-l), was used as the tracer (Experiment 6). Intact incorporation of this sample would have led to pyridoxol containing 48% of its total label at C-2’ and 52% at C-3. On Kuhn-Roth degradation such a sample of pyridoxol would have yielded acetic acid (C-2/,-2) and thence methylamine (C-2’) containing 48% of the activity of vitamin. In fact (Table II) the acetic acid which was ob- tained contained 90% of the activity of pyridoxol, most of which7 was located at C-2’ and was therefore derived (cf. Experiments 3 and 4) from the methyl group of the doubly labeled precursor. It follows that the carboxyl group of this pyruvate had not been incorporated into pyridoxol.

Thus, it is a 2-carbon unit, corresponding to the methyl and the carbonyl group of pyruvate, which supplies the 2-carbon unit, C-2’,-2, of pyridoxol. The route, which leads from pyruvate into this 2-carbon fragment remains to be established. A work- ing hypothesis, consistent with all available experimental evi- dence, will be presented later in this paper.

Incorporation of Glycerol-The samples of pyridoxol obtained from the experiments with [l, 3-W]glycero18 (Experiment 1 and 2) were among those showing the highest radiochemical recovery. Kuhn-Roth oxidation of these samples yielded acetic arid (C-2’, -2), accounting for 22 and 21aj,, respectively, of the activity of the intact vitamin (Table III). These values were sufficiently close to that expected for random distribution of activity (25y’) to appear to indicate that incorporation of label from glycerol may have been nonspecific, as expected for a general carbon source. Further degradation of the Kuhn-Roth acetate proved that such an inference was entirely unwarranted. Almost all of this activity (18%) was confined to the C-methyl group (C-2’) (Table III), showing that incorporation of glycerol had been nonrandom.

The distribution, within pyridoxol, of activity derived from [1,3-W]glycerol as well as from [2-14C]glycerol was therefore examined in greater detail (Experiments 15 and 16)) by means of more extensive controlled chemical degradation of the labeled samples (Schemes 5 to 7). The distribution of radioactivity established by the results of these degradation experiments (Table IV) is summarized in Figs. 1 and 2. It appears that, of

8 The samples of [1,3-l%]glycerol which were used had been prepared by a chemical synthesis (59) (J. R. Catch, Radiochemical Centre, Amersham, England, personal communication) which places radiocarbon at only one of the primary carbon atoms of each labeled molecule. However, since glycerol is a prochiral compound (60) whose enzymic phosphorylation takes place exclu- sively at the pro-R-hydroxymethyl group, metabolic phosphoryla- tion of the labeled sample of glycerol must yield an equimolar mixture of (2R)-[l-W]- and (2R)-[3-Wlglycerol 3-phosphate (stereospecific numbering (60)) (’ z.e. L-w[~-~~C]- and L-o(-[~-~%] glycerophosphate) (i).

1 CHZOH HO s

Y

H 2=

8 HzOP

6)

by guest on August 11, 2019

http://ww

w.jbc.org/

Dow

nloaded from

Issue of March 25, 1972 R. E. Hill, F. J. Rowe& R. N. Gupta, and I. D. Spenser 18i9

22+1 I ;H20H

Fig.1 3.451.0

o/5+0.2

HO ;H20H

Fig.2 3221 FIG. 1. Observed distribution of activity (per cent) of pyridoxol

derived from [1,3-l%]glycerol (Tables III and IV). FIG. 2. Observed distribution of activity (per cent) of pyridoxol

derived from [2-‘%]glycerol (Table IV).

the 8 carbon atoms of pyridoxol, five (of which three are C-2’, C-4’, and C-5’, each of which contains approximately 20’% of the total activity) are derivable from the terminal carbon atoms of glycerol (Fig. 1), and three (of which two are C-2 and C-4, each of which contains approximately 33% of the total activity) from the central carbon atom of this substrate (Fig. 2). The data lead to the inference that 3 3-carbon units related to glyc- erol are implicated in the biosynthesis of pyridoxol and that, on route to the product, one of these units loses a terminal car- bon atom. From the mode of incorporation of pyruvate (see above) it follows that the truncated glycerol unit supplies C-2’ and C-2 of pyridoxol. The labeling pattern of glycerol-derived pyridoxol and the structural correspondence of precursor units and product, deduced from the experiment.al data, are shown in Fig. 3.

Glycerol enters the glycolytic pathway by way of L-cr-glycero- phosphate, phosphodihydroxyacetone and n-glyceraldehyde 3- phosphate and is hence convertible into pyruvate. Glycerol can thus serve as the source of the pyruvate-derived carbon atoms of pyridoxol (C-2’,-2). Since the final step on this route from glycerol into pyruvate, the dephosphorylation of phos- phoenolpyruvate, is essentially irreversible and since reconver- sion of pyruvate into triose phosphate by alt.ernative pathways is unlikely to be important under the chosen experimental con- ditions, pyruvate cannot replace glycerol as a complete source of t,he carbon skeleton of pyridoxol. Since [Wlpyruvate thus leads to singly labeled pyridoxol whereas [lJC]glycerol leads to

4’

%H,OH L

3824

2,65 1

HO EH20H

36>H3q NJ Fig.4 2.622. I

FIG. 3. The mode of incorporation of glycerol into pyridoxol. Sites of activitv derived from 11.3~‘4Clglvcerol (A, A) (relative specific activitGm200jC) and from [2k]g!ycerol 10, 0) (relative specific activity -33%), shown by degradation (A, 0) or inferred (A, 0).

FIG. 4. Observed distribution of activity (per cent) of pyridoxol derived from [1-Wlglucose (Table IV).

multiply labeled product, the lower recovery of radioactivity in the pyruvate-derived samples (see “Results and Criteria for Their Interpretation”) is explained.

The distribution of label within the pyridoxol sample derived from nL-[3-14C]serine (Experiment 8) closely resembles that of the pyruvate-derived samples (Table II). Since serine is con- vertible into pyruvate (56, 57) by an enzymic process which is essentially irreversible (58) it follows that entry of serine into pyridoxol takes place via pyruvic acid. The low radiochemiral yield (Experiment 8, Table I) is consistent with this conclusion. The enzymes catalyzing the conversion, in E. coli, of L- and D-

serine into pyruvate are known to require pyridoxal phosphate (56). This is the cofactor which is limiting in a pyridoxal-lees mutant under conditions of pyridoxal deprivation, the present experimental conditions.

Biosynthesis of Pyridoxol: Working Hypothesis

The tracer evidence suggests that glycerol or closely related compounds are the primary precursors of pyridoxol. Whereas glycerol itself lacks the reactivity demanded by the condensat.ion reactions which must occur in the construction of the 8-carbon skeleton of pyridoxol from 3-carbon precursors, the t.riose phos- phates, phosphodihydroxyacetone, and n-glyceraldehyde 3-phos- phate, would appear to be likely substrates of such reactions.

A hypothetical biosynthetic sequence which is consistent with

by guest on August 11, 2019

http://ww

w.jbc.org/

Dow

nloaded from

Biosynthesis of Vitmnin Bg

,

PHOSPHODIHYDROXY .+ ACETONE

CYJP

GLYCEROL -

N! * CO2H

* A CH, 0

PYRUVIC ACID

ACET- ALDEHYDE

5-DEOXY p-xv LULOSE I -PHOSPHATE

BGLYCERALDEHYDE 3-PHOSPHATE

PYRIDOXOL

f

OP

f

OP -P

OP

HO CH20H I

o&op ~ Ho&opbH+ :+H+, Ho&op eHpo;:gop

NH3 NH2

SCHEME 8. Biosynthesis of pyridoxol. Structural correspondence of glycerol, pyruvate, and pyridoxol (box) and hypothetical bio- synthetic sequence.

all available experimental data and is also mechanistically rational is presented in Scheme 8. In the scheme the predicted labeling pattern of pyridoxol, as well as that of the three postu- lated primary precursors, phosphodihydroxyacetone, n-glycer- aldehyde 3-phosphate, and the pyruvate-derived 2-carbon unit (shown at the oxidation level of acetaldehyde), when derived from chemically labeled glycerol: is indicated. (The interrelation- ship of the precursor units, through the Embden-Meyerhof pathway, is not shown.)

The distribution of activity of glycerol-derived pyridoxol, deduced from the experimental data (Fig. 3), corresponds to that, demanded by the hypothesis (Scheme 8).

An early intermediate predicted by the scheme is a 5-deoxy- pentulose phosphate, originating by condensation of acetaldehyde and phosphodihydroxyacetone. Depending on the mode of this postulated condensation, the compound would be either a l-phosphate (as shown in Scheme 8) or a 3-phosphate (1). 5-Deoxypentulose 3-phosphates appear to be unknown. The intermediacy of such a compound would nevertheless be attrac- tive. The irreversible loss of the S-phosphate group at a later stage of the sequence, during the conversion of an enol phosphate to a ketone (c.f. the conversion of phosphoenolpyruvate into pyruvate) would explain why pyridoxol diphosphate has not been detected.g

9 It should be noted in passing that the present hypothesis predicts that, contrary to the most common currently held view (61), pyridoxol5’-phosphate serves as the precursor of pyridoxol, rather than vice versa.

Formation of 5-deoxypentulose l-phosphate on the other hand, by aldolase-catalyzed condensation of acet,aldehyde and phos- phodihydroxyacetone, is known to occur in a variety of tissues, e.g. muscle and yeast (62). Since, in aldolase-catalyzed reac- tions, the hydroxy groups astride the newly formed carbon-car- bon bond are almost invariably trans to each other (63), the two possible products of this reaction are 5-deoxy-n-xylulose l-phos- phate or 5-deoxy-r-xylulose l-phosphate. Aldolase preparations from pea seedlings (64) and from human erythrocytes (65) yield the n-enantiomer. An aldolase from an E. coli mutant catalyzes formation of the L-enantiomer (66). The n-enantiomer is arbi- trarily shown in Scheme 8.

The experiment with n-[l-i4C]glucose (Experiment 14) pro- vides further evidence in support of the biosynthetic hypothesis outlined in Scheme 8 and, in particular, serves to distinguish between the two possible modes of entry of phosphodihydroxy- acetone. Glycolysis of n-[l-14C]glucose leads, via n-[1J4C]fruc- tose 1 ,6-diphosphate, to [l-Wldihydroxyacetone l-phosphate and unlabeled n-glyceraldehyde S-phosphate. Equilibration of the two triose phosphates, catalyzed by triose phosphate isom- erase, eventually leads to n-[3-14C]glyceraldehyde 3-phosphate, which on further degradation yields [3-14C]pyruvate and hence [2WJacetaldehyde. Incorporation of these fragments into pyridoxol (Scheme 9), would place activity from [l-14C]glucose at C-2’ (via acetaldehyde), at C-5’ (via n-glyceraldehyde 3-phos- phate) and at either C-4’ or at C-3 (via phosphodihydroxyace- tone), depending on whether the l-phosphate or the 3-phosphate of 5-deoxyxylulose had served as intermediate. Three carbon

by guest on August 11, 2019

http://ww

w.jbc.org/

Dow

nloaded from

Issue of March 25, 1972 R. E. Hill, F. J. Rowell, R. N. Gupta, and I. D. Spenser 1881

1 -14C - FRUCTOSE 1,6- DI PHOSPHATE

PHOSPHODIHYDROXY xr ACETONE

=wp

1‘ CO2H 3

d4C-D-

GLUCO:E PYRUVIC CO2

5-DEOXY ACET-

BGLYCERALDEHYDE

ACID ALDEHYDE o-XY LULOSE 3-PHOSPHATE I -PHOSPHATE

4’ **CH20H

PYRIDOXOL

SCHT;ME 9. Biosynthesis of pyridoxol. Mode of incorporation of activity from n-[l-14C]glucose.

atoms of the pyridoxol sample derived from [I-14C]glucose, C-2’,

C-5’, and either C-4’ or C-3, would thus account for its entire

activity. The observed distribution of label (Fig. 4) agrees with t,his prediction: within experimenbal error all activity is accounted for in terms of the activities of C-2’, C-4’, and C-5’. The pres- ence of activity at C-4’ is consistent with the intermediacy of 5deosyxylulose l-phosphate, but not with that of the 3-phos- phate.

Although the location of label from D-[1-%]glucose is entirely according to prediction, the quantitative distribution of activity among the predicted posit,ions is puzzling. Equilibration of label among the glycolytic C3 units would lead to pyridoxol carrying 33% of its total act,ivity at each of C-2’, C-4’, and C-5’. This was not observed (Fig. 4). Since glycolytic breakdown of n-[lJ4C]glucose yields, in the first inst,ance, labeled phospho- dihydroxyacetone and unlabeled glyceraldehyde 3-phosphate, and since the equilibrium of the triose phosphate isomerase reac- tion favours the former, equilibration of label need not take place in a short term incubation. Pho;phodihydroxyacetone would t,hen show a higher specific activity than glyceraldehyde 3-phosphate, and the site of pyridoxol derived from [l-‘*Cl- dihydroxyacetone l-phosphate (C-4’) would be expected to carry a greater share of the total activity of the intact vitamin than the site derived from D-[3-14C]glyceraldehyde 3-phosphate (C-5’). This is indeed observed (Fig. 4).

Since pyruvate arises glycolytically from phosphoglyceralde- hyde, activity at C-2’ of pyridoxol should be similar to that at C-5’. In fact (Fig. 4) the observed activity at C-2’ resembles that of C-4’ and is unlike that of C-5’. Taken at face value this result appears to indicate that the pyruvate-derived 2-carbon unit which serves as the progenitor of the Cz unit, C-2/,-2, of pyridoxol, is derived from phosphodihydroxyacetone or directly from a hesose rather than from phosphoglyceraldehyde.

This and other aspects of the present results require further study. Although the observed distribution of activity from acetate and from aspartate (60% of the total activity of pyridoxol within the C2 unit, C-2’,-2) (Experiments 7 and 11, Table II) may be rationalized on the basis of reconversion of these sub&rates into phosphoenolpyruvate by conventional metabolic processes, the actual mode of incorporation of acetate and aspartate into pyridoxol remains to be clarified experimentally. The identity of the 2-carbon unit serving as the building block of the pyruvate- derived CZ fragment, C-2’,-2, of pyridoxol, assumed to be at the oxidation level of acetaldehyde on mechanistic grounds and on the basis of the relatively poor incorporation of acetate, remains to be established. Direct evidence bearing on the intermediacy of 5-deoxyxylulose l-phosphate and on the existence of the hypothetical branched chain B-carbon sugar has yet to be ob-

Nonetheless, the foundations have been laid to an understand- ing, long overdue (c.f. 61, 67, 68), of the process which leads from primary precursors to vitamin Bs.

1. 2.

3.

HILL, R. E., AND SPENSER, I. D. (1970) Science 169, 773 HILL, R. E., GUPTA, R. N., ROWELL, F. J., AND SPENSER, I.D.

(1971) J. Amer. Chem. Sot. 93, 518 SNELL, E. E., AND GUIR~RD, B. M. (1943) Proc. Nut. Acad.

Sci. U. S. A. 29, 66 4. 5.

SNF,LL, E. E. (1945) J. Biol. Chem. 168, 497 HOLDEN, J. T., FURMAN, C., .~KD SNELL, E. E. (1949) J. Biol.

Chem. 178, 789 6. HOLDEN, J. T., AK-D SNELL, E. E. (1949) J. Biol. Chem. 178,

799 7. 8. 9.

10.

11.

AMANO, T. (1958) Vitamins (Kyoto) 14, 595 MORRIS, J. G. (1959) J. Gen. Microbial. 20, 597 SUZUE, R., AND HAKUN.~, Y. (1963) Vitamins (Kyoto) 28, 108 SUZUE, R., HAI~UN~, Y., H~Y~ISHI, O., Y.~MMVIOTO, S. AXD

OGATA, K. (1964) V&m&s (Kyoto) 29, 569 SATO, K., SUZUKI, T., AND SAHASHI, Y. (1965) Vitumins (Kyoto)

31, 356 12. SATO, K., SUZUKI, T., AND SAH~SHI, Y. (1965) Vitamins (Kyoto)

31, 359 13. SATO, K., SUZUKI, T., AND SAHASHI, Y. (1966) Vitamins (Kyoto)

33, 355 14. DEMPSEY, W. B. (1969) Biochem. Biophys. Res. Commun. 37,89 15. DEMPSEY, W. B. (1969) J. Bacterial. 100, 1114 16. DEMPSEY, W. B., AND ITOH, H. (1970) J. Bucteriol. 104, 658 17. SUZUE, R., AND HARUNA, Y. (1970) J. Vituminol. (Osaka) 16,

154 18. LUNAN, K. D., AND WEST, C. A. (1963) Arch. Biochem. Biophys.

REFERENCES

101, 261 19. SUZUE, R., H,\I~uNA, Y., AND HAYAISHI, 0. (1963) Seilcuguku

(Tokyo) 36, 617 20. S&TJE,“R. (1964) Bull. Inst. Chem. Res. Kyoto Univ. 42, 288 21. HAYAISHI, 0. (1965) Vitamins (Kyoto) 31, 320 22. SATO, K., SUZUKI, T., AND SAHASHI, Y. (1966) Vitamins (Kyoto)

33, 357 23. DEMPSEY, W. B. (1970) Fed. Proc. 29, 343; (1970) Biochem.

Biophys. Actu 222, 686 24. SCHROER, R. A., AND FRIEDEN, E. H. (1970) Fed. Proc. 29, 926 25. SUZUE, R., AND HARUNA, Y. (1970) J. Vituminol. (Osaka) 16,

161 26. COHEN, A., HA~ORTH, J. W., ‘END HUGHES, E. G. (1952) J.

Chem. Sot. 4374 27. DALGLIISSH, C. E. (1960) in W. UM~REIT AND H. MOLITOR

(Editors), Proceedings of Fourth International Congress of Biochemistry, Vienna, 1958. Vol. 11, p. 32, Pergamon Press, London

28. LEETE, E. (1963) in P. BERNFI~LD (Editor), Biogenesis of natural compounds, Ed. 1, p. 748, Pergamon Press, Elmsford, New York

29. SPENSER, I. D. (1968) in M. FLORKIN AND E. M. STOTZ (Editors), Comprehensive biochemistry, Vol. 20, p. 231, Elsevier Pub- lishing Co., Amsterdam

30. NURMIKKO, V., AND RAUNIO, R. (1961) Actu Chem. &and. 16, 856

t,ained 31. DEMPSEY, W. B. (1967) J. Bucteriol. 93, 1179

by guest on August 11, 2019

http://ww

w.jbc.org/

Dow

nloaded from

Biosynthesis of Vitamin B6 Vol. 247, Xo. 6

32. 33.

(1970) J. Biol. Chem. 246, 4229 RABINOWITZ, J. C., AND SNELL, E. E. (1948) J. Biol. Chem. 176,

1157 34. 35. 36.

FUKUI, S. (1953) Anal. Chem. 25, 1884 DEMPSEY, W. B. (1965) J. Bacterial. 90, 431 DEMPSEY, W. B., AND PACHLER, P. F. (1966) J. Bacterial. 91,

642 37. 38. 39.

DEMPSEY, W. B. (1966) J. Bacterial. 92, 333 DEMPSEY, W. B. (1969) J. Bacterial. 97, 1403 RICKENBERG, H. V., YANOFSKY, C., AND BONNER. I). M.

(1953) J. Bacterial. 66, 683 40. 41.

42.

KUHN, R., AND WENDT, G. (1938) Ber. 71, 1118 STILLER, E. T., KERESZTESY, J. C., AND STEVENS, J. R. (1939)

J. Amer. Chem. Sot. 61, 1237 HARRIS, S. A., AND FOLI~ERS, K. (1939) J. Amer. Chem. Sot.

61, 1245 43. 44.

KUHN, R., AND WENDT, G. (1939) Ber. 72, 305 LEETE, E., GREGORY, H., AND GROS, E. G. (1965) J. Amer.

Chem. Sot. 67, 3475 45. GLAZER, J., HUGHES, E. D., INGOLD, C. K., JAMES, A. T.,

46. 47. 48. 49. 50. 51.

JONES, G. T., AND ROBERTS, E. (1950) J. Chem. Sot. 2657 TABORSKY, R. G. (1961) J. Org. Chem. 26, 596 HARRIS, S. A. (1940) J. Amer. Chem. Sot. 62, 3203 KORYTNYK, W., AND WIEDEMAN, W. (1962) J. Chem. sot. 2531 COHEN, A., AND HUGHES, E. G. (1952) J. Chem. Sot. 4384 BADDILEY, J., AND MATHIAS, A. P. (1952) J. Chem. Sot. 2583 KORYTNYK, W., KRIS, E. J., AND SINGH, R. P. (1964) J. Org.

Chem. 29,574 52.

53.

BROOKS, H. G., JR., LAAKSO, J. W., AND METZLER, D. E. 68. SAUBERLICH, H. E. (1968) in W. H. SEBRELL, JR. .~ND R. S. (1966) J. Heterocyc. Chem. 3, 126 HARRIS (Editors), The vitamins, Vol. ,%‘, 2nd Ed, p. 31,

KUHN, R., AND WENDT, G. (1938) Ber. 71, 1534 Academic Press, New York

54.

55.

56.

57.

58.

59.

60.

61.

62.

63.

64.

65. 66.

67.

HARRIS, S. A., WEBB, T. J., AP\‘D FOLI~ERS, K. (1940) J. Amer. Chem. Sot. 62, 3198

PALM, D., SMUCKER, A. A., AND SNELL, E. E. (1967) J. Org. Chem. 32, 826

GREENBERG, D. M. (1961) Metabolic pathways Vol. 2, p. 89; (1969) Vol. 3, p. 120, Academic Press, New York

WOOD, W. A., AND GUNSALUS, I. C. (1949) J. Biol. Chem. 181, 171

GREENBERG, D. M. (1961) Metabolic pathways, Vol. 2, p. 174, Academic Press. New York

GIDEZ, L. I., AND’KARNOVS~Y, M. L. (1952) J. Amer. Chem. Sot. 74, 2413

BENTLEY, R. (1969) Molecular asymmetry in biology, Vol. 1, p. 171, 175, Academic Press, New York; (1970) Vol. 2, p. 224

GOODWIN, T. W. (1963) The biosynthesis of vitamins and related compounds, p. 159, Academic Press, New York

MEYERHOF, O., LOHMANN, K., AND SCHUSTER, P. (1936) Bio- them. 2. 286, 301

RUTTER, W. J. (1961) in P. D. BOYER, H. LARDY, AND K. MYR- B&X (Editors), The enzymes, Vol. 5, p. 341, Academic Press, New York

GORIN, P. A. J., HOUGH, L., AND JONES, J. K. N. (1953) J. Chem. Sot. 2140

BARTLETT, G. R. (1968) Biochim. Biophys. Acta 166, 254 CHIU, T. H., AND FEINGOLD, D. S. (1965) Biochem. Biophys.

Res. Commun. 19, 511 SNELL, E. E. (1960) in W. UMBREIT AND H. MOLITOR (Editors),

Proceedings of Fourth International Congress of Biochemistry, Vienna, 1958, Vol. 11, p. 250, Pergamon Press, London

by guest on August 11, 2019

http://ww

w.jbc.org/

Dow

nloaded from

Robert E. Hill, Frederick J. Rowell, Ram Nath Gupta and Ian D. Spenser6Biosynthesis of Vitamin B

1972, 247:1869-1882.J. Biol. Chem. 

  http://www.jbc.org/content/247/6/1869Access the most updated version of this article at

 Alerts:

  When a correction for this article is posted• 

When this article is cited• 

to choose from all of JBC's e-mail alertsClick here

  http://www.jbc.org/content/247/6/1869.full.html#ref-list-1

This article cites 0 references, 0 of which can be accessed free at

by guest on August 11, 2019

http://ww

w.jbc.org/

Dow

nloaded from