1

I1
1.1

.15...
I I

1.1. II..1.1.I.....1I1I.1|.1. 11111111111111.1111.1I.

..

'.
i

1‘

.

1111
:‘ ~1111I I1II
1 .‘..I.11.11111

I 11 1'1 1 1,1111 1:111.“ .- z"... ,._’ ....—‘--.:.'

. 11111111111111111111111|I1 . . I 1 .1. .1
111111111111111111111111111 111 ‘ 11111111111111.1111 1* 1 2.311311

.g 1.11.1 I . . 3' .‘rr
. .11....I 1111 .11. .11. 1.' ...I II11 1I..11I.1.1 1 . .. .. 11111112;11111111111111.111111111111111 3;.
11111111111111111111111111111I 11111 1 1 11 . ‘ 1 ‘ 1,I1'-‘ 1111111111.} 11111111111111.11111 11111111131111 11.111 111 .

1 1’ 11' 11'”1 1111111111 1 ' ' 1 111111111111111111..” I II .1111

1 .1 1 1 11.11 11*1 11.11 $11.11:

. | .11111111111111'1‘ ,.

‘ 111.. 11.. 1.11111

13"?"‘1111‘11‘1311‘1111'111' y}. ""11'1'“

 
  
   
     
   
 

1

  

“1.1.11.1... WW
1.11.11. 11111 1.12.711. .r I. I1 *1

 
  
    
 
 

  

 
    
 

111111 .9 ~
11'1" 1111111111111 11 I ...‘11.:;1’.‘.:‘.,....:;;.W.~v
I I . ‘ 1' 1111111111111 .
111.111111111 11 111111 11111111 1.. .11 111111 1 1:21 3'11"?
111111 1' 1 " I L.
.11.1111I1111I1... 1 1111.111. ...111111.1.11111. I 1 . '1 1111.1.11I11.‘ .16 111111 I 1

1111.1 .1111I 1.1111. . 1.I... 11111 ... 1

1

11 I1 '1'"

 
       

11.1.1. I

I1. . 1 1

Wm“:
M?

I

1 11
.11. 11
11111111. 1111
11111
1.. .1..III|1...
11.11 1111 11.11 1111.1 . 11

.111
1.1111.11111

.q:'—‘--;

111

    
   
       
  
 

 
  
   
 
  

   
  
   

| J 111‘1T11I
11“. i1“. .1
11.1.

M
‘—v
.
_--
‘
—o
.4.
c _\
4- -

1

 

..
‘
W
v...

  
    
 

     
     
       
     

  

   

  
   

     
  
 

'11.
11111111111111111111111.111..
1111111.1111111111111I1.111...1. ..1. 11111111

1.111

1111 I II1I1“ . "1111‘

. a n I'1I-'I1“I‘?' "1:: ' ‘

. v.4...» .I. :I -

( .I 2-1141: I'. ."'. II'I :' _ _
'1 14".“... 1I'11I1111711111-11'. '- l

 

I 11.11
111111111111

.1 I
111111 I11 11.
111111 I

1111111

 
 

 

1.27.1

_'———_—_—‘____

d
tw
.- -
“.3,"

 

LIBRARY

Michigan State
University

 

BIOLOGICAL AVAILABILITY OF VITAMIN E IN LIQUID

6
AND DEHYDRATED MODEL FOOD SYSTEMS AFTER

THERMAL PROCESSING AND STORAGE

BY

Jesse F. Gregory III

A DISSERTATION

Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

DOCTOR OF PHILOSOPHY

Department of Food Science and Human Nutrition

1977

G l C: 3,391

ABSTRACT

BIOLOGICAL AVAILABILITY OF VITAMIN B6 IN LIQUID
AND DEHYDRATED MODEL FOOD SYSTEMS AFTER
THERMAL PROCESSING AND STORAGE

BY

Jesse F. Gregory III

Liquid and dehydrated model food systems were
employed to study the behavior of the B6 vitamers with
respect to the retention of vitamin 36 activity during
thermal processing and storage.

A liquid model system comprised of a neutral
solution of heat stable peptides was utilized to examine
the extent and manner of binding of pyridoxal (PL) and
pyridoxal phosphate (PLP) to peptides and proteins during
retort processing. Isolation and spectrophotometric
examination of the model system peptides after processing
for 20 min at 121°C revealed that about 21% of the PLP
in the model system was peptide bound. Characterization
of the PLP-peptide complexes indicated that about 10% of
the PLP in the model system was bound as a pyridoxylamino
complex, potentially limiting the bioavailability. The
presence of glucose and ascorbic acid in the model system

solutions appeared to weakly inhibit the binding.

Jesse F. Gregory III

The effects of glucose and ascorbic acid on the
retention of available pyridoxamine (PM) were studied
using a similar liquid model system which was fortified
with PM. While losses were low during 20 min treatment
at 121°C in the absence of glucose or ascorbate, 21% of
the total B6 was degraded during processing in the pre-
sence of glucose. Ascorbic acid promoted only a slight
loss, and its effect was not additive with that of glu-
cose. Gel filtration chromatographic examination of
thermally induced interactions between PM and glucose,
ascorbate, or their degradation products revealed no
detectable binding.

Dehydrated model food systems of composition
simulating breakfast cereals were utilized to examine the
relative stability and bioavailability of the B6 vitamers.
Roasting and storage conditions were selected to permit
the estimation of the maximum vitamin 36 degradation
encountered in food processing and storage.

Roasting of the dehydrated model systems for
25 min at 180°C resulted in a loss of 50-70% of each of
the B6 vitamers. During the storage of similar dehy-
drated systems at 37°C after equilibration to a water
activity (aw) of 20.6, varying stability of the B6 vita-
Iners was observed by liquid chromatographic (HPLC) analy-
isis. When packaged in enameled metal cans with a small
lueadspace (TDT cans), the dehydrated model systems

exhibited higher rates of nonenzymatic browning and

Jesse F. Gregory III

degradation of each of the B6 vitamers than the rates
observed with storage in cans with a large headspace.
Half-lives of the vitamers in model systems stored in
TDT cans at 37°C, =0.6 aw, were 35-46 days for PM, PL,
and PLP, and 141 days for pyridoxine (PN). The degrada-
tion of each of the B6 vitamers during storage followed
first order kinetics.

Microbiological and HPLC assays for total vitamin
86 correlated closely with rat bioassay estimates of
available B6 in the dehydrated model systems which had
been roasted 25 min at 180°C or stored for 128 days at
37°C. These results demonstrated that the vitamin B6
remaining after roasting or storage retained full bio-
availability. Vitamin 36 assay by the semiautomated
fluorometric method was not accurate in estimating avail-
able 36 in the dehydrated model systems.

The only significant mechanism identified for
the loss of vitamin 36 activity during storage was the
binding of PLP by the proteins of the dehydrated model
systems in the form of e-pyridoxyllysine. Tests of the
vitamin 36 activity of protein bound pyridoxyllysine
revealed only 50% of the molar potency of PN. In addi-
tion, this complex inhibited the utilization of about 55%

<>f the free PN added to rat diets.

ACKNOWLEDGMENTS

The author wishes to express sincere thanks to
his major professor, Dr. James Kirk, for willing counsel
and support during this research and for his efforts in
the review of this manuscript.

Appreciation is also extended to Dr. J. Robert
Brunner for his willing discussion of the peptide studies
of this research and to Dr. John Gill for helpful advice
concerning the statistical aspects of bioassay interpre-
tation.

The author also expresses grateful appreciation
to the members of his graduate committee, Drs. Loran
Bieber, J. Robert Brunner, Wanda Chenoweth, and Dale
Romsos, for their review of the manuscript.

Grateful acknowledgment is also due to the
Department of Food Science and Human Nutrition, Michigan
State University, for facilities and financial support
during this research. The author also acknowledges
financial support from NIH Training Grant GMO 1818.

Finally, the author extends most sincere thanks
‘to his wife, Helen, for constant inspiration and encourage-

znent throughout his graduate studies.

ii

TABLE OF CONTENTS

Page
LIST OF TABLES O O O O O O O O 0 O O O O Vii
LIST OF FIGURES. O O O O O O O O O O O O x
INTRODUCTION 0 O O O O O O O O O I O O O 1
LITERATURE REVIEW 0 O O O O O O O O O O O 3
VitaIRin BS History 0 o o o o o o o o o o 3
Determination of Vitamin B6 . . . . . . . 5
Extraction Methods . . . . . . . . . . 6
Microbiological Methods. . . . . . . . . 12
Preparative Chromatography. . . . . . . . 15
Bioassay Methods . . . . . . . . . . . 18
Evaluation of Vitamin 36 Bioassays . . . . . 20
Chemical Methods for Determination of Total
Vitamin 36 O O O O O O o o o 0 o o 22
Spectroscopic Methods. . . . . . . . . 22
Analytical Chromatoqraphic Methods . . . . 25
Vitamin B5 Chemistry . . . . . . . . . . 26
Formation of Complexes . . . . . . . . 26
Nonenzymatic Interconversions and Catalysis
by Vitamin 85 . . . . . . . . . . . 32
Vitamin B5 Stability and Bioavailability. . . . 34
EXPERIMENTAL METHODS . . . . . . . . . . . 44
Preparative Methods. . . . . . . . . . . 44
Heat Stable Peptide Mixture . . . . . . . 44
e-Pyridoxyl-L-lysine Synthesis . . . . . . 44
Phosphopyridoxyl-Bovine Serum Albumin . . . . 44
Chemical Methods. . . . . . . . . . . . 44

iii

 

Page
Protein Determination . . . . . . . . . 44
Amino Groups . . . . . . . . . . . . 45
Sulfhydryl Groups . . . . . . . . 45
Fluorometric Determination of Vitamin B5
Compounds . . . . . . . . . . . . 45
Erythrocyte Aspartate Aminotransferase Assay . 46
Hemoglobin Determination . . . . . . . . 47
Physical Methods . . . . . . . . . . . 47
Spectrophotometry . . . . . . . 47
High Performance Liquid Chromatographic Assay
of Vitamin 85 . . . . . . . . . 47
High Performance Liquid Chromatographic Assay
of Pyridoxyllysine . . . . . . . . . 48
High Performance Liquid Chromatographic Assay
of 4-Pyridoxic Acid . . . . . . 48
Discontinuous Polyacrylamide Gel Electro-
phoreSiso O O O O O O O O O I O O 48
Thin Layer Peptide Mapping . . . . . . . 49
Gel Filtration Chromatography . . . . . . 49
Biological Methods. . . . . . . . . . . 50
Microbiological Determination of Total
Vitamin 85 . . . . . 50
Rat Bioassay of Biologically Available
VitMin 36 o o o o o o o o o o o o 50
Experimental Design and Procedures . . . . . 54
Interaction of Pyridoxal and Pyridoxal Phos-
phate with Peptides in a Liquid Food Sys—
tem during Thermal Processing. . . . . . 54
Characterization of Peptides . . . . . 55
Pyridoxal and Pyridoxal Phosphate Binding
Studies . . . . . . . . . . . . 57
Interactions of Pyridoxamine in a Liquid Food
System during Thermal Processing. . . . . 61
Model System Formulation and Treatment. . . 61
Evaluation of Thermal Effects. . . . . . 63
Statistical Analysis. . . . . . . . . 64
Stability and Bioavailability of Vitamin B6
in Dehydrated Model Food Systems:
Roasting Effects . . . . . . . . . . 64

iv

 

Page
Formulation and Treatment of the Model
Systems . . . . . . . . . . . 65
Analysis of Model Systems. . . . . . . . 66
Statistical Analysis . . . . . . . . . 67
Stability and Bioavailability of Vitamin B5 in
Dehydrated Model Food Systems: Storage
Effects. 0 O O O O O C O O O I O O 68
Formulation, Processing, and Storage of Model
Systems . . . . . . . . . . . . . 69
Analysis of Model Systems. . . . . . . . 70
Biological Activity of Phosphopyridoxyl-Bovine
Serum Albumin. . . . . . . . . . . . 73
Analysis of Phosphopyridoxyl-Bovine Serum
Albumin O O I O I O I I O O O O O 74
Determination of Biological Activity . . . . 75
RESULTS 0 O O O O O O O O O O O O O O 77
Interaction of Pyridoxal and Pyridoxal Phosphate
with Peptides in a Liquid Model Food System
during Thermal Processing . . . . . . . . 77
Preparation and Characterization of Heat Stable
Peptides . . . . . . . . . . . 77
Binding of PL and PLP in Liquid Model Systems. . 84
Interactions of Pyridoxamine in a Liquid Model
Food System during Thermal Processing . . . . 93
Stability of Pyridoxamine . . . . . . . 93
Pyridoxamine--Carbohydrate Complex Formation . . 93
Stability and Bioavailability of Vitamin 36 in
Dehydrated Model Food Systems: Roasting
Effects 0 O O O I O I O O I I O O O 95
Comparison of Assay Methods for Total Vitamin 36' 95
Interference in Semiautomated Fluorometric
Vitalnin 86 Assay. o o o o o o o o 104

Biologically Available Vitamin 35 in Roasted
Dehydrated Model Systems . . . . . . . . 105

Stability and Bioavailability of Vitamin 36 in
Dehydrated Model Systems: Storage Effects. . . 110

Page
Moisture Content and Water Activity . . . . . 110
Nonenzymatic Browning . . . . . . . . 112

Kinetics of Vitamin B5 Degradation . . . . . 117
Determination of 4-Pyridoxic Acid in Dehydrated

Model Systems after Storage . . . . . . 127
Determination of Pyridoxyllysine in Dehydrated

Model Systems after Storage . . . . . . 129
Determination of Total Vitamin 85 in Dehydrated

Model Systems after Storage . . . . . . 135

Biologically Available Vitamin 86 in Dehydrated
Model Systems after Storage . . . . . . . 138

Biological Activity of Phosphopyridoxyl-Bovine
serum Albumin I I I I I I I I I I I I l4 4

Analysis of Phosphopyridoxyl-Bovine Serum

Album1n I I I I I I I I I I I I I 144
Pyridoxylamino Group Utilization and Anti-
vitamin Activity. . . . . . . . . . . 145

DISCUSSION I I I I I I I I I I I I I I I 149

Interaction of Pyridoxal and Pyridoxal Phosphate

with Peptides in a Liquid Model Food System

during Thermal Processing . . . . . . . . 149
Interactions of Pyridoxamine in a Liquid Model

System during Thermal Processing . . . . . . 158
Vitamin Ba Assay Methods . . . . . . . . . 159
Stability and Bioavailability of Vitamin 35 in

Dehydrated Model Systems: Roasting Effects . . 162
Stability and Bioavailability of Vitamin B6 in

Dehydrated Model Systems: Storage Effects. . . 164

Water Activity and Nonenzymatic Browning . . . 164
Stability of Vitamin B during Storage . . . '. 168
Degradation Products 0 Vitamin 85 . . . . . 170

Biological Availability of Vitamin B5 in
Dehydrated Model Systems after Storage . . . 171

Biological Activity of Protein Pyridoxylamino
Compounds. . . . . . . . . . . . . . 172

SUMMARY AND CONCLUSIONS . . . . . . . . . . 175
APPENDIX I I I I I I I I I I I I I I I 180

SELECTED BIBLIOGRAPHY . . . . . . . . . . . 198

vi

 

LIST OF TABLES
Table Page
1. Composition of diets for rat bioassays . . . 51

- 2. Formulation of liquid model systems for
study of PL and PLP interactions . . . . 58

3. Formulation of liquid model systems for
study of pyridoxamine interactions. . . . 62

4. Composition of dehydrated model system for
determination of roasting effects . . . . 66

5. Total peptide bound PL and PLP. . . . . . 86

6. Percent of model system PLP bound as a non-
reducible complex (substituted aldamine
or pyridoxylamino compound) at pH 7.0. . . 89

7. Percent of PLP in liquid model systems which
was bound as a pyridoxylamino complex. . . 90

8. Sulfhydryl content of peptides in liquid
model systems after thermal processing . . 92

9. Vitamin 85 in PM-fortified liquid model sys-
tems after thermal processing . . . . . 94

10. Calibration curves for the HPLC assay of
vitamin B6 using 280 nm absorption
detection . . . . . . . . . . . . 98

11. Total vitamin 36 in control and roasted
model systems determined by microbio-
logical and HPLC methods . . . . . . . 100

12. Vitamin 85 in control and roasted model sys-
tems, as determined by the semiautomated
fluorometric assay procedure. . . . . . 101

13. Comparison of fluorometric and microbiological

methods for determination of total vitamin
B6 in control and roasted model systems . . 103

vii

Table Page

14. Characteristics of the emission spectra of
the fluorophores produced from vitamin
36 standards and model system extracts
during the fluorometric assay of vitamin
36 . . . . . . . . . . . . . . 106

15. Rat bioassay of roasted model systems
(Bioassay l). . . . . . . . . . . 107

16. Linear regression parameters of the dose-
response curves. Bioassay l, roasted
model systems . . . . . . . . . . 108

17. Biologically available vitamin 36 in
roasted model systems. . . . . . . . 109

18. Comparison of rat bioassay estimates of bio-
logically available vitamin 36 in roasted
model systems with microbiological and
semiautomated fluorometric assay results
for total vitamin B5 . . . . . . . . 111

19. Moisture content and water activity of
stored model systems . . . . . . . . 113

20. Comparison of browning levels in model sys-
tems stored in TDT and 303 cans for 128
days at 37°C . . . . . . . . . . 116

21. Kinetic data for the loss of the 85 vitamers
from dehydrated model systems during
storage in TDT cans at 37°C. . . . . . 126

22. Vitamin B6 content in dehydrated model sys-
tems after storage for 128 days at 37°C
in TDT and 303 cans . . . . . . . . 128

23. Determination of 4-pyridoxic acid in model
systems stored for 128 days in TDT cans
at 37°C I I I I I I I I I I I I 136

24. Determination of pyridoxyllysine in model
systems from TDT and 303 cans pooled
after storage for 128 days at 37°C . . . 136

25. Microbiological assay of vitamin B6 in model

systems pooled after storage for 128 days
in TDT and 303 cans at 37°C. . . . . . 137

viii

Table
26.

27I

28.

29.

30.

31.

32.

33.

Page

High performance liquid chromatographic assay

of vitamin B5 in model systems pooled

after storage for 128 days in TDT and 303

cans at 37°C . . . . . . . . . . . 137
Semiautomated fluorometric determination of

vitamin B6 in dehydrated model systems

pooled after storage for 128 days at

37°C in TDT and 303 cans . . . . . . . 138
Rat Bioassay 2. Determination of available

vitamin 86 in model systems pooled after

128 days storage in TDT and 303 cans at

37°C, and analysis of biological activity

Of PP-BSA o o o o o o o o o o o o 139
Linear regression parameters of the dose-

response curves. Bioassay 2. . . . . . 141
Biologically available vitamin B5 in model

systems pooled after storage for 128 days

in TDT and 303 cans at 37°C . . . . . . 142
Comparison of semiautomated fluorometric,

microbiological, HPLC, and rat bioassays

for vitamin B in model systems pooled

after 128 days in storage in TDT and 303

cans at 37°C . . . . . . . . . . . 143
Determination of free vitamin 36 in

dialyzed PP-BSA . . . . . . . . . . 146
Biological activity of pyridoxyllysine resi-

dues in PP-BSA o o o o o I o o o o 148

ix

Figure

1.

2.

10.

11.

.12.

LIST OF FIGURES

Chemical forms of vitamin B6 (free bases) .
Complexed forms of pyridoxal . . . . .

Polyacrylamide gel electropherogram of B-
lactoglobulin which was used in isolation
of heat stable peptides . . . . . .

Thin layer peptide map of heat stable pep-
tides . . . . . . . . . . . .

Sephadex G-50 (fine) calibration curve . .

Sephadex G-50 chromatogram of heat stable
peptide mixture . . . . . . . . .

Ultraviolet-visible difference absorption
spectrum of peptides isolated from a PLP-
fortified model system versus those from
the nonfortified blank. . . . . . .

Fluorescence emission spectra of peptides
from liquid model systems with and with-
out PLP fortification . . . . . . .

Sephadex G-10 chromatograms of PM-fortified
liquid model systems after 20 min at
121°C I I I I I I I I I I I I

HPLC chromatogram of standard vitamin 85
free bases. . . . . . . . . . .

Loss of free amino groups in soluble pro-
teins of dehydrated model systems during
storage in TDT cans at 37°C . . . . .

Formation of melanoidin pigments during
storage of dehydrated model systems in
TDT cans at 37°C. . . . . . . . .

Page

27

79

80

82

83

85

88

96

97

114

115

 

 

Figure Page

13. HPLC chromatograms of a recovery sample of
a nonfortified model system stored for
97 days in a TDT can at 37°C . . . . . 120

14. Degradation of PLP in dehydrated model sys-
tems stored in TDT cans at 37°C . . . . 122

15. Behavior of vitamin 86 in a PM-fortified
dehydrated model system stored in TDT
cans at 37°C . . . . . . . . . . 123

16. Degradation of added PN and inherent PL in
dehydrated model systems stored in TDT
cans at 37°C I I I I I I I I I I 125

17. Absorption spectrum of pyridoxyllysine;
difference spectrum of 1 mg/ml PP-BSA
versus 1 mg/ml BSA . . . . . . . . 131

18. HPLC chromatograms of pyridoxyllysine
standard and protein hydrolysate of
PLP-fortified model system. . . . . . 132

19. Fluorescence excitation and emission
spectra of crystalline pyridoxyllysine
standard and collected pyridoxyllysine
chromatographic peak from PLP-fortified
model system protein hydrolysate. . . . 134

xi

INTRODUCTION

Vitamin 36 serves as a cofactor in the metabolism
of all plant and animal species, and, hence, is widespread
in nature. Variation in the 36 requirement among humans
and the possible instability of certain forms of the vita-
min in foods are factors which may contribute to the mar-
ginal nutriture of certain groups of the population.

The stability of vitamin B6 in foods has not been
extensively studied. Previous research has demonstrated
that, under certain processing and storage conditions,
losses of vitamin B6 may occur. The precise behavior of
the various chemical forms of the vitamin, their rates of
degradation, and the identity of their degradation pro-
ducts are largely unknown.

The aldehyde forms of vitamin 36' pyridoxal and
pyridoxal phOSphate, have the capability of forming
several types of complexes with compounds containing
free amino groups. Pyridoxamine may react with aldehydes
to form analogous complexes. Potential reactions involv-
.ing these vitamin B6 compounds have never been investi-

‘gated in food systems. Because these complexes vary

widely in their stability and bioavailability, their
characterization and quantitation in foods would provide
important information on nutritional quality.

Animal bioassay results have indicated that in
certain instances, particularly in the thermal sterili-
zation of dairy products, the content of biologically
available 36 may be less than that predicted by micro-
biological or chemical assay procedures. Whether this
disparity represents an inherent difference in assay
specificity or a true loss of bioavailability of the
vitamin has not been established.

The purposes of this research were: (1) to inves-
tigate the interaction of several of the B6 vitamers with
food proteins and carbohydrates and determine the bio-
logical activity of the complexed forms of the vitamin;
(2) to determine the accuracy and specificity of newly
developed and existing methods for vitamin 36 assay;

(3) to study the stability of the B6 vitamers under
detrimental conditions of processing and storage, such
that maximal losses encountered in foods could be esti-
mated; and (4) to determine the effects of these severe
thermal processing and storage conditions on the biologi-
cal availability of vitamin 36 in foods. It was antici-
pated that the results of these studies would further the
understanding, with respect to vitamin B6, of the nutri-

tional quality of foods.

LITERATURE REVIEW

Vitamin B5 History

 

The chemical and nutritional pr0perties of vita-
min B6 were first recognized when Birch and Gyorgy (1936)
identified a rat acrodynia-preventing factor in rice bran
concentrates. Shortly thereafter, pyridoxine (PN) was
isolated and crystallized independently in five labora-
tories (Lepkovsky, 1938: Keresztesy and Stevens, 1938;
Gyorgy, 1938; Kuhn and Wendt, 1938; Ichiba and Michi,
1938). Coincidentally, this compound proved to be iden-
tical to the unidentified nitrogenous base which had been
previously isolated by Ohadake (1932).

From studies with vitamin BG-requiring micro-
organisms, Snell, 22.21“ (1942), identified a non-pyri-
doxine factor or factors, termed "pseudopyridoxine,"
which possessed vitamin 86 activity. These were subse-
quently identified as pyridoxal (PL) and pyridoxamine
(PM), having aldehyde and amino groups at the C-4 position
of the pyridine ring, respectively (Snell, 1942; Snell,
l944a,b; Harris, 32 31., l944a,b). The structures of

these three vitamers are illustrated in Figure 1.

mz_—2<xoo_m>m

u \
ozo
«IZNIQ

n10

.Ammmmn mmnmv

mm GHEmuH> mo mahom HMUHEmQU .H ousmflm

J<XODE>¢ w2_xoo_m>a

n n
zx, :0 z/. :o
N \ N \
:o :0 oz :0 5 oz
0.10 10.210

am 2:253 10 mzmou 43.510

Studies concerning microbial metabolism demon-
strated the active coenzyme forms of vitamin 36 to be
phosphorylated derivatives of pyridoxal and pyridoxamine
(Bellamy and Gunsalus, 1944; Gale and Epps, 1944; Gun-
salus and Bellamy, 1944). Heyl, 25 21° (1951a,b), showed
that the active coenzyme forms were the 5'-phosphates of

pyridoxal (PLP) and pyridoxamine (PMP).

Determination of Vitamin 85

 

For the accurate determination of vitamin B6 in
foods, all active forms of the vitamin, i.e. pyridoxal,
pyridoxine, pyridoxamine, and their S'-phosphates, must
be measured. All of these vitamin B6 compounds have been
shown to possess approximately equal activity in mammalian
and avian species.

Numerous assay procedures have been developed for
the determination of the active forms of vitamin B6, based
upon chemical, physical, and microbiological methodolo-
gies. All of these analytical procedures require extrac-
tion of the B6 vitamers from the sample matrix since they
are based on determination of the vitamers in solution.

In contrast, several animal bioassay methods have been
developed for the determination of biologically available
vitamin B6. These bioassay procedures measure only bio-
logically available vitamin 36' as opposed to total 86,
because no extraction or hydrolysis procedures are

employed.

Extraction Methods

 

In addition to releasing the free forms of the
vitamin from the sample, the extraction procedure must
recover any complexed forms of the vitamin which are bio-
logically active for accurate quantitation of vitamin B6
activity.

Initially, most of the assay procedures were based
on microbial growth. This necessitated hydrolysis of the
sample to cleave phosphate ester forms of vitamin 86
because most of the organisms did not respond to the
phosphorylated vitamers. Extraction of samples in strong
acids at 12l-130°C proved to be a fairly satisfactory
method, because this treatment resulted in the disinte-
gration of the sample and hydrolysis of the phosphate
derivatives to their respective free base forms.

Atkin, 25 31. (1943), found that optimum extrac-
tion efficiency could be achieved by autoclaving most bio-
logical samples 1 hour at 20 psi in 0.055 N sulfuric acid.
For the assay of cereals, the use of 0.44 N sulfuric acid
provided a more complete extraction. In a rigorous study
of extraction efficiency, Rubin, 22.21' (1947), found that
for most samples, autoclaving in sulfuric acid within the
range of pH 1.7 to 1.8 was most effective. This cor-
responds closely to 0.055 N sulfuric acid.

Extraction methods developed by Fujita, gt 31.

(l955a,b,c), for the chemical assay of vitamin B6 were

based on the use of H2804. For the determination of PN
and PM, a sample homogenate adjusted to pH 4.5 was heated
initially to 80°C for 15 minutes. The supernatants were
then adjusted to 0.6 N with respect to H2804 and auto-
claved 1 hour at 130°C. PL was extracted by direct auto-
clave treatment in 0.1 N H2804 for 1 hour at 130°C.

Studies of autoclave extraction methods by Rabin-
owitz and Snell (1947) demonstrated that 0.055 N HCl
functioned as well as 0.055 N H2804 as an extraction medium
for the assay of vitamin B6 in most samples. Extraction
in an autoclave at 20 psi for 5 hours resulted in the
greatest hydrolysis and release of vitamin B6 in most
cases. Again, cereals required a more rigorous extraction
treatment; autoclaving for 1 hour at 20 psi in 2 N HCl
was found to be satisfactory. Storvick, st 31. (1964),
slightly modified the procedure of Rabinowitz and Snell
(1947) by extracting with 0.055 N HCl at 15 psi for 5
hours. This was found to provide satisfactory extraction
of vitamin B6 in most samples. Toepfer, gt_gl. (1963,
1970), adopted the 0.055 N HCl, 15 psi, 5 hour autoclave
extraction method for their microbiological assay pro-
cedure. For plant products, including cereals, they
determined that autoclaving in 0.44 N HCl at 15 psi for
2 hours was sufficient for complete extraction.

A method for the microbiological determination of

vitamin B6 in milk (Gregory, 1959) was based on extraction

with 0.55 N HCl under flowing steam, rather than autoclav-
ing. Subsequently, Gregory and Mabbitt (1961) demonstrated
the presence of pyridoxamine Phosphate in milk and found
that this vitamer required heat treatment for 3 hours at
125°C in 0.055 N HCl for complete hydrolysis of the phos-
phate ester. Under the same conditions, pyridoxal phos-
phate was totally hydrolyzed in only 30 minutes. There-
fore, extraction of samples containing pyridoxamine phos-
phate for less than 3 hours at 125°C would result in
erroneously low vitamin B6 values by most assay procedures.
In studies concerning the effects of autoclaving
time on apparent vitamin B6 content of blood samples
using several species of assay microorganisms, highly
divergent results have been observed (Benson, st 31.,
1964). Using blood samples which were extracted with
0.055 N HCl for time periods from 30 minutes to 72 hours,

Saccharomyces uvarum (formerly g. carlsbergensis) assays

 

showed a maximum value after 2 to 5 hours of autoclaving,
followed by a sharp decline to almost nondetectable levels

by 10 hours. In contrast, assays with Lactobacillus casei

 

and Streptococcus faecium showed peak values after 10 to
24 hours, with little decline up to 72 hours. These
anomalous results were later explained by the obser-
vations that both 2, gaggi_and S. faecium assays suffer

interference from free and peptide bound D-alanine formed

 

by racemization during extensive autoclaving (Gregory,
1959: Haskell and Wallnofer, 1967; Raines and Haskell,
1968).

Further evidence against the use of long heat
treatments for extraction of vitamin 86 was obtained by
Storvick and Peters (1964). Nonfat dry milk samples
were hydrolyzed in 0.055 N HCl at 121°C for periods up
to 72 hours, then assayed by an S. uvarum method. The
apparent vitamin 86 in the samples increased with increas-
ing hydrolysis time to a maximum at approximately 5 hours,
followed by a progressive decline. After 72 hours at
121°C, only 10-20% of the maximum apparent vitamin B6
remained in the samples. Parallel treatments of pure
vitamin B6 standards, however, demonstrated very little
degradation over the entire 72-hour period. Storvick
and Peters (1964) suggested that extensive heat treatment
during the extraction of biological samples could result
in erroneously low values because of an irreversible
binding of the vitamers to proteins. Extraction at
121°C for 5 hours in 0.055 N HCl provided maximal ex-
traction and hydrolysis of all forms of the vitamin, yet
minimized the possible losses by this mechanism.

Extraction with aqueous buffers, followed by
enzymatic hydrolysis and chemical deproteination, were
investigated in an effort to alleviate the numerous

Iproblems inherent in the autoclave extraction procedures

10

and to shorten the extraction time. Early methods of
enzymatic hydrolysis using digestion with proteolytic
and diastatic enzymes failed largely because of the high
levels of vitamin 86 in the enzyme preparations (Rabino-
witz and Snell, 1947; Hopkins and Pennington, 1947).

A combination of the autoclave extraction pro-
cedure with enzymatic hydrolysis was performed by Barton-
Wright (1971). Samples were autoclaved in 0.055 N HCl
for 1 hour, followed by adjustment to pH 4.5 and treat—
ment with takadiastase. This method was satisfactory
for a variety of biological samples and foods, including
cereals.

For analysis of vitamin B6 in blood using S.
uvarum, Storvick and Peters (1964) compared autoclave
extraction with 0.055 N HCl to extraction by deprotein-
ation with trichloroacetic acid (TCA) without enzymatic
treatment. Results with the autoclave extraction method
were approximately 50% greater than those with the TCA
method. These data demonstrated the need for cleavage
of phosphate esters for the accurate determination of
total vitamin B6 using the S. uvarum method.

Numerous studies of vitamin 36 metabolism have
involved extraction of all forms of the vitamin from
tissue samples, followed by individual analysis of the
vitamers. Columbini and McCoy (1970) demonstrated

95-100% recovery of all 36 vitamers by homogenization

11

of the tissue in cold water, followed by protein precipi-
tation with TCA. Bain and Williams (1960) and Lyon, 3E
311 (1962), developed a similar method employing depro-

teination with perchloric acid. This method provided

extraction of at least 85% of the total vitamin 86 in
tissues (Johansson, SE 31" 1968).
Enzymatic hydrolysis of the vitamin B phosphate

6
esters was readily accomplished by Takanashi, 25.31'

(1970a,b). Potato acid phosphatase treatment of sample
homogenates in pH 4.0 acetate buffer provided quantitative
hydrolysis. TCA was employed for subsequent protein pre-
cipitation. Recoveries of PL and PLP added to biological
samples were 95-100% and 70-90%, respectively.

Recently, a biologically active B-glucoside and
several other conjugates of pyridoxine have been found to
occur naturally in certain cereals (Yasumoto, gE_gl.,
1976, 1977: Tsuji, 25.2l'v 1977). These derivatives,
although found to be readily cleaved by autoclave treat-
ment, were resistant to hydrolysis with takadiastase.
Treatment of buffer extracts with a B-glucosidase prepar-
ation cleaved the linkages, rendering the pyridoxine
available for microbiological or chromatographic analysis.
In addition, these authors observed the presence of phos-
phorylated B6 vitamers in the cereals examined. There-
fore, determination of total vitamin B6 in such a product
would require treatment of aqueous extracts with phospha-

tase and B-glucosidase.

12

Microbiological Methods

 

Numerous methods for assaying vitamin B have been

6
developed employing microorganisms which require an
exogenous source of the vitamin for growth. These methods
are based on the stimulation of microbial growth by the
vitamin B6 in a sample extract added to a basal medium
containing adequate levels of all other nutrients.

The procedure developed by Atkin, SE.E$° (1943),

employing the yeast Saccharomyces uvarum, has been the

 

most widely applied of the microbiological methods.
Numerous modifications of the method have been made to
improve the response of the organism and increase the
sensitivity (Hopkins and Pennington, 1947; Jones and
Morris, 1950; Parrish, 2E.3l°' 1955, 1956; Toepfer and
Lehmann, 1961; Haskell and Snell, 1970).

The primary problem encountered with this pro-
cedure is the differential in growth response of S. uvarum
to PM, PN, and PL. The response to PM has been found to
be significantly lower than to PN or PL, with PM often
exhibiting only 50-60% of the potency of the other vita-
mers (Rabinowitz and Snell, 1948; Parrish, SE 21" 1955;
Gregory, 1959; Woodring and Storvick, 1960; Chin, 1975).
The use of a vigorously growing yeast culture has been
shown to minimize this difference in response (Sauberlich,
1970). Toepfer, 22.21‘ (1961, 1970), circumvented the

;prob1em by developing a method for S. uvarum assay of

13

the individual B6 vitamers after separation by preparative
ion exchange chromatography. This improved the accuracy
of the determination of each vitamer by permitting the
quantitation of each by comparison to its respective
standard curve. Although the method is laborious and
cumbersome, it alleviates the problem of erroneously low
values for total vitamin B6 in products containing high
levels of pyridoxamine.

Another problem encountered in S. uvarum assays
is the inhibition of yeast growth by high salt concen-
tration in sample extracts (Rabinowitz and Snell, 1949).
Presumably, this, too, would be alleviated by preparative
ion exchange chromatography.

A method for the determination of total vitamin 36

in foods using Neurospora sitophilia 299 was developed

 

by Stokes, 2E El“ (1943), and subsequently modified by
Harris (1952) and Hodson (1956). In contrast to S. uvarum

and most other assay organisms, N. sitophilia exhibits

 

equal response on a molar basis to all forms of vitamin
86, including PLP and PMP (Hodson, 1956). This assay
method has not been widely used since it is complex and
requires a 5-day growth period.

Recently, assay methods employing Kloeckera

apiculata have shown considerable promise for the measure-

 

nent of total vitamin B6. Procedures using cup-plate

(Barton-Wright, 1971) and turbidimetric (Daoud, 1973)

l4

analysis have been developed. This organism shows equal
reSponse to PM, PL, and PN (Barton-Wright, 1971; Daoud,
1973), thus obviating the problems encountered with

‘S. uvarum.

Methods for the assay of total vitamin 36 using
other organisms have been less successful. Baker and
Sobotka (1962) reported a procedure based on stimulation
of the growth of Tetrahymena pyriformis. While adequate
for the assay of vitamin B6 of tissues and physiological
fluids containing mainly PM and PL, further application
has been limited by the poor response to PN. A procedure

using Escherichia coli, strain 154-59L, was developed by

 

Diding (1955). Although quite simple, this method suffers
from interference by certain amino acids and shows only
a weak response to PM.

Differential methods for the determination of the
individual vitamers have also been developed. These have
been based on the specific requirements of certain micro-
organisms for the various forms of vitamin B6. The pro-
cedure of Rabinowitz and Snell (1948) was based on the
use of S. uvarum assays for total PN, PM, and PL, £32327

bacillus casei assays for PL, and Streptococcus faecalis

 

 

R for PL and PM. Because PN was determined by difference
it was susceptible to high experimental uncertainty.
Gregory (1959) modified this differential method by

employing Streptococcus faecium ¢ 51 for the determination

15

of PL and PM, since this organism was less sensitive to
interference. As previously mentioned, however, D-alanine,
L-alanine, and certain peptides were later shown to stimu-
late the growth of S. faecium ¢ 51 (Haskell and Wallnofer,
1967; Raines and Haskell, 1968). This has made the dif-
ferential method unsatisfactory for most analyses of

vitamin 86.

Preparative Chromatography

Numerous methods have been developed for the
preparative chromatographic separation of the 36 vitamers.
These methods were developed to facilitate quantitative
analysis of the individual 86 compounds by microbiological,
chemical, or radioisotopic counting methods. In addition,
most of these methods served to remove potentially inter-
fering compounds to some extent.

The ionogenic nature of the B6 vitamers makes them
well suited for separation by ion exchange chromatography.
Peterson and Sober (1954) developed the first method for
the separation of PN, PL, PM, and their respective phos-
phates. The separation was performed on a column using
the weak cation exchange resin, Amberlite XE-64 (H+).

This procedure was never applied to extracts of foods
or biological samples.

Fujita, 22.21' (l955a,b,c) developed a procedure
for the preparative purification of the B6 vitamers prior

to fluorometric analysis. PL was determined following

16

elution from columns of Amberlite IRC-SO (H+) or IR-112
(H+). PN and PM, after deamination to PN, were deter-
mined after elution from Permutit columns. These pro-
cedures represented the first application of preparative
chromatography to the quantitative analysis of vitamin 36'

MacArthur and Lehmann (1959) successfully separated
PN, PM, and PL on a single column containing Dowex AG
50W—X8 (K+) by a stepwise elution system of boiling
potassium acetate and citrate buffers. Because each of
the vitamers eluted in a discrete fraction, the method
was ideally suited as a preparative technique. Toepfer,
SE.2£° (1960), improved the chromatographic resolution
of the free base vitamers by slightly modifying the
elution buffers. Although this ion exchange procedure
has been successfully applied to microbiological analysis
of the B6 vitamers (Toepfer and Lehmann, 1961; Toepfer
and Polansky, 1970; Erin, 1970), the method is cumbersome
because of the use of boiling buffers.

Because of the problems associated with the use
of boiling buffers, attempts were made to develop a room
temperature separation of PL, PN, and PM. Hedin (1963)
proposed a method for the separation of the free base
vitamers using an Amberlite IR—120 (Na+) column, iso-
«cratically eluting with 0.1 M sodium phosphate, pH 6.5.
{The three phosphorylated B6 compounds could be isocrati-

<=ally separated on columns containing the same resin by

17

elution with 0.1 M sodium acetate, pH 5.0. The stepwise
elution methods of Toepfer, 2E 21' (1960), had the advan-
tage of eluting each vitamer as a discrete chromatographic
fraction readily suited for analysis. The isocratic and
continuous gradient methods were limited by the problem

of collection and identification of the fractions contain-
ing the various vitamers. For this reason, neither the
procedure of Hedin (1963) nor that of Storvick, EE.E$°
(1964), has received extensive application.

The separation of PN, PM, PL, and their phosphates
was accomplished by Johansson, SE S1. (1968), using a
single column of Dowex 50-X8 (NHX) with an ammonium
formate gradient. This procedure was modified slightly
by Tiselius (1972) to provide better separation of PLP
and PNP. Excellent resolution of all six vitamers by
this method permitted the study of the metabolic dis-
tribution of vitamin 36 in tissues.

In addition to the ion exchange chromatographic
methods for vitamin B6 separation, several thin layer
techniques have been developed. Both thin layer electro-
phoretic (Ahrens and Korytnyk, 1970; Columbini and McCoy,
1970; Smith and Dietrich, 1971) and thin layer chromato-
graphic (Smith and Dietrich, 1971) methods have been
tdescribed for the preparative separative separation of

all B6 vitamers.

18

Bioassay Methods

 

Animal bioassay techniques for the determination

of biologically available vitamin B have the advantage

6
that no extraction or hydrolysis of the sample is required.
The animal bioassay is the ultimate method for comparison
with microbiological and chemical assay procedures, serv-
ing as a check of their accuracy in the determination of
the vitamin 86 potency of foods. Limitations of the bio-
assay methods are: (l) a relatively large amount of the
test substance is required, (2) the assays are time con-
suming and expensive, and (3) the precision of bioassay
estimates is much poorer than that of the other analyti-
cal methods.

The predominant methods for bioassays with
laboratory rats have been based on the procedures of
Conger and Elvehjem (1941) and Clarke and Lechycka (1943).
Sarma, 23.31' (1946), succeeded in developing a basal
diet which permitted minimal growth in the absence of
vitamin 36' but allowed maximal growth with optimal
levels of the vitamin. Using this basal diet, the bio-
assay was performed with a 2-week depletion period fol-
lowed by a 4-week test period. During the test period,
'the rats were fed the basal diet fortified with graded
:levels of crystalline PN or the test substance. Rat

gyrowth over the 4-week period was used as an indicator

<>f'the dietary B concentration. Numerous modifications

6

19

of the diet composition have been made, largely depend-
ing on the material to be assayed (Linkswiler, ES 31.,
1951; Tomarelli, §£'§£., 1955; Lushbough, 2E 31" 1959;
Davies, 2E.El°' 1959; Richardson, SE.E$°I 1961).

The chick has also been used in vitamin B6 bio-
assays because of its high growth rate and sensitive
nutrient requirements. The procedures of Ott (1946)
and Coates, SE El: (1950), with the modification of
Davies, 22.21' (1959), provided sensitive assays for
biologically available vitamin B6. The recent method
of Yen, EE.E£° (1976), utilizing a chemically defined
basal diet, permits bioassays in less than 2 weeks, thus
markedly shortening the analysis time.

Problems encountered in the interpretation of the
results of rat or chick bioassays have been mainly due to
the role of the intestinal microflora in vitamin B6
utilization. The use of dextrin or starch as the source
of dietary carbohydrate has been shown to stimulate the
intestinal synthesis of vitamin B6’ inducing greater
growth on the nonfortified basal diet. Such growth
markedly reduces the sensitivity of the bioassay. The
presence of soluble sugars, such as sucrose, lactose,
or glucose, in the basal diet was shown to provide much
lower growth rates in the absence of dietary vitamin B6
(Sarma, EE(E£" 1946; Waibel, 2E El" 1952). The bio-

assay results of materials containing high levels of

 

20

complex carbohydrates would therefore be inaccurate if
dietary replacement levels were sufficiently high to
increase intestinal synthesis of vitamin 36.
The relative activities of PN, PM, and PL have
been shown to be approximately equal when administered
orally by medicine dropper or intraperitoneally (Sarma,
EE.E£°' 1946). In contrast, the activity of PN was found
to be consistently greater than that of PM or PL when
mixed in rat or chick diets (Sarma, EE.El" 1946; Waibel,
SE.E£°' 1952; Davies, EE.E£°I 1959). Compared to PN, the
apparent activities of PL and PM were found to be approxi-
mately 60-70% in rats (Sarma, g£_gl., 1946) and 70-80%
in chicks (Davies, 22.31;! 1959). Equal activity of the
three vitamers was observed when the animals were treated
orally with aureomycin (Linkswiler, EE.El" 1952).
These results suggested that under normal conditions

the intestinal microbes preferentially utilize the PL

and PM contained in the diet.

Evaluation of Vitamin B5 Bioassays

 

Most bioassay methods have used animal growth as
an indicator of the content of vitamin B6 in diets con-
taining suboptimal levels of the vitamin. Several func-
tional indices of vitamin 36 status have been discovered
which would increase the specificity of bioassays if

these were used in addition to growth data.

21

Numerous researchers have observed the correlation
between aminotransferase activity of various tissues and
blood fractions and the vitamin B6 nutriture of the
animal (Raica and Sauberlich, 1964; Beaton and Cheney,
1965; Erin and Thiele, 1967; Chen and Marlatt, 1975;
Kirksey, EE.El" 1975; Lee, EE,E£°' 1976; Yen, EE.E£°'
1976). Aspartate aminotransferase (AspAT), also known
as glutamate oxaloacetate transaminase, has been shown
in all tissues to possess much higher activity than
alanine aminotransferase (Beaton and Cheney, 1965; Erin
and Thiele, 1967). Therefore, AspAT activity is gen-
erally considered to be the more sensitive indicator of
vitamin 86 status.

Raica and Sauberlich (1964) examined the AspAT
activity of blood plasma, leukocytes, and erythrocytes
of human subjects, and concluded that the enzyme activity
was not sufficiently sensitive to be used as an indicator
of vitamin B6 nutriture in diverse human populations.
However, in laboratory animals under a controlled dietary
(regimen, erythrocyte AspAT activity has been found to be
a sensitive measure of vitamin B6 status (Beaton and
Cheney, 1965; Erin and Thiele, 1967; Kirksey, EL 21"
1975). Plasma AspAT activity was found to be more sen-
sitive to changes in vitamin 36 status than serum AspAT
activity in chicks; however, growth was even more sen-

sitive than either enzyme activity (Yen, EE.El°' 1976).

 

 

 

 

 

22

The relationship between the stimulation of the
aminotransferase enzymes by in vitro addition of the co-
factor PLP and vitamin B6 deficiency state was first
noted by Raica and Sauberlich (1964). These authors
observed that the extent of stimulation of erythrocyte
AspAT by added PLP was a sensitive indicator of human
vitamin B6 nutriture. Recent detailed studies have con-
firmed the sensitivity and specificity of the in vitro
stimulation method for the estimation of human vitamin B6
status (Bayoumi and Rosalki, 1976; Kishi and Folkers,
1976). The PLP stimulation of the erythrocyte amino-
transferases has also been used to monitor 36 status in
rat feeding studies (Chen and Marlatt, 1975; Kirksey,

SE El., 1975).

Other indicators of vitamin 36 nutriture, such
as 4-pyridoxic acid excretion, tryptophan load tests,
and the vitamin B6 concentration of tissues, blood, and
urine could potentially be used in animal bioassays.
However, the laborious analytical methods tend to pre-
clude the use of these indices because of the large
numbers of animals generally employed.

Chemical Methods£or Determination
of TotaI Vitamin B5

 

Spectroscopic Methods.--Several Spectrophoto-
Imetric methods have been developed for the determination

Of'the various vitamin B6 compounds (Wada and Snell,

23

1961; Soda, EE.§£°' 1969). These methods were of limited
use in food analyses because of their lack of specificity
and sensitivity.

The methods of Fujita, SE.§£° (l955a,b,c), were
the first application of fluorometry to the determination
of vitamin 86 in foods. These methods were based on the
formation of the fluorophore, 4-pyridoxic acid lactone,
by a complex series of reactions for each vitamer.
Hennessy, SE 31. (1960), modified this method for the
determination of PN in bread.

Basic studies by Bonavita (1960) demonstrated the
feasibility of fluorometric analysis of PL by quantitation
of the product of the reaction of potassium cyanide with
PL. Detailed studies by Ohishi and Fukui (1968) con-
firmed that the reaction product of PL and cyanide was
4-pyridoxic acid lactone. The assay of PN and PM could
be performed by their conversion to PL with manganese
dioxide and glyoxylic acid, respectively, prior to
treatment with potassium cyanide (Toepfer, 22.21" 1961;
Polansky, EE.E£°' 1964). These reactions provided a
much simpler method for the formation of the fluorophore
than those employed by Fujita, EE.E£° (l955a,b,c).

Numerous modifications of the 4-pyridoxic acid
lactone method have been made for the analysis of foods,
biological fluids, and tissues (Contractor and Shane,

1968; L00 and Badger, 1969; Takanashi, EE.El" 1970;

24

Takanashi and Tamura, 1970; Masukawa, SE.E£°' 1971; Fied-
lerova and Davidek, 1974; Chin, 1975). Most of these
methods involved preparative ion exchange column chroma-
tography prior to fluorophore formation. Contractor and
Shane (1968) observed that their fluorometric results
correlated well with published microbiological data for
total vitamin B6 in blood, but were lower than the values
obtained by Fujita, SE 3;. (1955b). They suggested that
the method of Fujita, SE 31., may be subject to inter-
ference from nonspecific fluorophores. Studies by Chin
(1975) demonstrated fluorometric values significantly
greater than microbiological results for most foods
analyzed. It was suggested, based on these studies, that
a possible source of interference was the extensive
browning routinely observed following the autoclave
extraction step (Chin, 1975). Notably, the procedure

of Fujita, 22.21’ (l955b), was based on an autoclave
extraction, while that of Contractor and Shane (1968)
used only TCA precipitation. Kraut and Imhoff (1967)
developed a fluorometric method based on entirely dif-
ferent methods of 4-pyridoxic acid lactone formation.
Here, the vitamers were isolated by paper chromatography,
followed by combined chemical and enzymatic treatments.
The determination of vitamin B6 in foods by this fluro-

metric also yielded results which were much greater than

IPublished microbiological data. No studies, to date,

 

 

25

have systematically and conclusively established the
validity of the fluorometric assay procedures with respect
to interfering compounds.

An innovative procedure for the fluorometric
determination of PL and PLP in blood and tissues has been
reported by Srivastava and Beutler (1973). Sample homo-
genates were treated with semicarbazide, forming the
respective semicarbazones of PL and PLP. Each semicarba—
zone was identified and quantitated according to its
characteristic fluorescence spectra. The advantage of
this technique was the ability of the semicarbazide to
react with and release protein bound PL and PLP, thereby
permitting direct fluorometric quantitation after protein
precipitation. This procedure has not been applied to
the assay of the other B6 vitamers and has not been used

with food systems.

Analytical Chromatographic Methods.--Several gas

 

liquid chromatographic methods have been developed for
the quantitative analysis of vitamin B6 compounds in
relatively pure mixtures, such as pharmaceutical prepar-
ations (Sheppard and Prosser, 1970; Korytnyk, 1970;
Williams, 1974). The method of Williams is the most
sensitive, employing electron capture detection of the
heptafluorobutyryl derivatives. No successful appli-
cations of these methods have been made to foods or

biological samples .

 

 

 

26

Williams and Cole (1975) proposed a high per-
formance liquid chromatographic (HPLC) method for the
separation of PN, PM, and PL. The vitamers were separated
by gradient elution from a column of Aminex A-5, requiring
80 minutes per separation. The low specificity of the
ultraviolet absorption detector employed prevented the
application of this method to food systems. HPLC was
applied to the determination of vitamin B6 in foods by
Yasumoto, SE 3;. (1975). The specificity of the detection
was heightened by continuously treating the effluent of
the Aminex A-S column with the diazide of S-dichloro-
analine 2,4-disulfonyl chloride, a chromogenic reagent
for aromatic compounds. Although the separation requires
90 minutes and the instrumentation is complex, this
method is the first successful chromatographic procedure

for analysis of vitamin B6 in foods.

Vitamin B6 Chemistry

 

Formation of Complexes

The ability of PL and PLP to form several types
of complexes with amino acids, amines, peptides, and pro-
teins (Figure 2) has been widely studied. The most widely
studied of these reactions has been the formation and
behavior of the Schiff bases. Metzler (1957) and Lucas,
EELEl° (1962), observed that Schiff base formation occurs
:most readily in neutral or alkaline media. Matsuo (1957a)

observed that PL was much less reactive than PLP, by

27

msomoaosm mEMOm muonmmonm meoowuwm

z m
/ .10
1036 \ o...
«rm
.1.
m
ncaoquo

05523825 .m

 

Z n
/ Io
Icazo \ o:
:0
\xz\ /x/
mm _m

3.032.»: .5
._>._c>;::m .oEEOnxv
mEEoEc 33533 .N

.meooflumm mo mEHom omxmamfiou

.mmxmamfioo
.m musmwm

2/ are

$8 1128..

 

 

 

28

virtue of the existence of PL predominantly in the form
of an internal hemiacetal (Heyl, EE.El" 1951a). Schiff
bases of PL and amino acids were shown to readily chelate
a number of polyvalent metal ions (Matsuo, 1957a). In
addition to the pH-dependence of the rate of formation
and stability of the Schiff bases, several tautomeric
forms have been found to exist in a pH-dependent equil-
ibrium (Christensen, 1958; Arrio-Dupont, 1970).

When PLP or PL forms a Schiff base with a poly-
functional amine, amino acid, peptide, or protein, nucleo-
philic attack at the azomethine linkage can lead to the
formation of a cyclic derivative, the substituted alda-
mine. This reaction has been extensively studied with
amino acids and amines (Heyl, SE.2£°' 1948a; Buell and
Hansen, 1960; Bergel and Harrap, 1961; Mackay and Shep-
herd, 1962; Mackay, 1963; Abbott and Martell, 1970;
O'Leary, 1971; Kierska and Maslinski, 1971; Schonbeck,
SE 21" 1975; Der Garabedian and Der Garabedian, 1976).
Amino, sulfhydryl, and imidazole groups have been shown
capable of reaction with a Schiff base to form the sub-
stituted aldamine. In acid media, these complexes are
highly unstable and readily dissociate via the Schiff
base intermediate (O'Leary, 1971; Kierska and Maslinski,
1971; Buell and Hansen, 1960; Bergel and Harrap, 1961).

Chemical reduction of the double bond of the

azomethine linkage of the Schiff base by a sufficiently

29

strong reducing agent forms a pyridoxylamino compound.
Methods for catalytic hydrogenation (Hey, SE 21" 1948a,
1952) and sodium borohydride reduction (Fischer, 33'3I.,
1958; Severin, SE 31" 1969) have been widely used in

the synthesis of pyridoxylamino compounds. These com-
plexes are highly stable, showing only slight degradation
under hydrolytic conditions including incubation for 24
hours at 100°C in 6 N HCl (Dempsey and Snell, 1963;
Ronchi, EE.E£°' 1969; Anderson, EE.E£°' 1971).

The reaction of PL and PLP with proteins occurs
in a manner identical to the reactions with small mole-
cules such as amino acids, amines, and peptides. The
interaction of these vitamers with protein amino groups
has been recently reviewed by Feeney, EE.E$' (1975).
Numerous enzymes have been shown to be inhibited in vitro
by Schiff base formation between PLP and an active site
amino group. Among them are: glutamate dehydrogenase
(Anderson, SE El., 1966), 6-phosphogluconate dehydroge—
nase (Rippa, EE.El" 1967), hexokinase (Grillo, 1968),
fructose 1,6 diphosphate aldolase (Shapiro, EE.El" 1968),
glyceraldehyde 3-phosphate dehydrogenase (Ronchi, 32 21"
1969). pyruvate kinase (Johnson and Deal, 1970), and
ribonuclease A (Riquelme, SE 21" 1970).

The binding of PLP to serum albumin and glycogen
;phosphorylase probably represent the most thoroughly

«examined vitamin 86-protein interactions. In the case

 

 

 

30

of serum albumin, PLP has been found to bind in strong
preference to PL, with no binding of PN to the protein
(Anderson, SE 21" 1974). Spectrophotometric evidence
and resistance to sodium borohydride reduction suggested
that PLP was bound as a substituted aldamine at the
primary binding site at neutral pH (Dempsey and Chris-
tensen, 1962; Anderson, 33 Ei’! 1971). However, Hilak,
EE.El° (1975), demonstrated that the primary binding of
PLP to serum albumin occurs by Schiff base formation at
a specific apolar binding site. Similarly, initial
studies with glycogen phosphorylase suggested that PLP
was bound as a substituted aldamine complex at neutral
pH (Kent, 2E.El°' 1958; Scheller and Will, 1973). The
existence of a Schiff base at an apolar binding site
was also demonstrated for glycogen phosphorylase (Shal-
teil and Cortijo, 1970; Cortijo, SE 31" 1976).

In addition to its binding as a cofactor in many
enzymatic systems, PLP readily forms a Schiff base with
the amino groups of many other proteins. Typical
examples are studies in which PLP was used as a reagent
for modifying collagen (Page, SE 31" 1968) and erythro-
cyte membrane proteins (Cabantchick, gg,gl., 1975).

The bioavailability of the various complexed
forms of vitamin B6 has not been thoroughly studied.
Presumably, the Schiff base and substituted aldamine

forms would be totally available because of their

31

instability in acid environments, which would be encoun-
tered in the stomach. Pyridoxylamino acids have been
found to possess low, but measurable, vitamin 36 bio-
logical activity. Studies using the microbiological

assay organisms, S. uvarum, N. sitophilia, S. faecalis,

 

and S. EEEEi demonstrated that pyridoxylamino acids pos-
sessed no more than 0.5% of the molar activity exhibited
by PL; however, no significant antivitamin activity was
observed (Snell and Rabinowitz, 1948). In contrast, the
vitamin B6 activity of numerous pyridoxylamines was found
to range from 50 to 100% of that of PN for most of the
compounds tested using rat bioassays (Heyl, 3E Ei’! 1952).
Contrary to the observed vitamin B6 activity of the pyri-
doxylamines, several pyridoxyl-pressor amines were found
to possess very low pressor activity in rats (Heyl, EE.E£°I
1952).

The antivitamin B6 activity of numerous analogues
and complexes has been widely studied and reviewed by
Sauberlich (1968) and Klosterman (1974). Through the
formation of a stable complex with PL or PLP, carbonyl
reagents, such as semicarbazides and hydrazines, may
effectively reduce the in vivo concentration of metaboli-
cally available vitamin B6. McCormick and Snell (1961)
have demonstrated that many of the stable complexes
strongly inhibit pyridoxal phosphokinase, thereby
inhibiting the in vivo formation of PLP. These com-

;Plexes, whether formed in vivo or present in the diet,

32

would interfere with the utilization of dietary vitamin
B6. McCormick and Snell (1961) observed that many of the
pyridoxylamino acids did not possess inhibitory activity,
although the reduced complex, N, N'—bis(pyridoxyl)hydra-

zone, was itself a potent inhibitor.

Nonenzymatic Interconversions and
Catalysis by Vitamin 86

 

Transamination reactions between PL or PLP and
amino acids were first shown to occur nonenzymatically
by Metzler and Snell (1952), producing the respective
d-ketoacid and PM or PMP. Matsuo (1957) reported that
these reactions occurred via a Schiff base intermediate.
In aqueous media, the presence of metal salts was required
(Metzler and Snell, 1952), although the reaction proceded
spontaneously with no metal catalyst in ethanol solution
(Matsuo, 1957). PL and a-amino acid esters readily
reacted in aqueous media in the absence of metal ions
(Cennamo, 1964). These results suggested that the role
of the metal ion was merely to mask the charged a-carboxyl
group of the amino acid. Dixon (1966) reported that
transamination reactions could occur between PL and the
free amino groups of proteins in the presence of metal
ions.

The occurrence of transamination reactions during
the thermal processing of foods has been demonstrated.

The formation of PM and PMP from PL and PLP have been

 

33

shown to occur during milk processing (Hodson, 1956;
Gregory, 1959; Gregory and Mabbitt, 1961; Polansky and
Toepfer, 1969). Transamination during the cooking of
pork was suggested by the data of Polansky and Toepfer
(1969), which indicated that fully cooked ham contained
predominantly PM, while raw pork contained mainly PL.

PL has been shown to catalyze the oxidative
deamination of PM and amino acids in the presence of
metal ions (Ikawa and Snell, 1954; Hamilton and Revesv,
1966; Rotilio, EE.E$" 1970; Hill, 1972). These oxygen-
dependent reactions resulted in the variable oxidation
of PL to 4-pyridoxic acid (Ikawa and Snell, 1954; Hamil-
ton and Revesv, 1966; Rotilio, EE.2l°' 1970), and, there-
fore, represent a potential mechanism for the loss of
vitamin B6 activity in foods.

The catalytic cleavage of hydroxyamino acids
(Metzler, 2E.2i" 1954b) and the B-elimination of serine
and O-phosphoserine (Reiber, 1976) by PL and PLP have
been described. The analogous catalysis of the non-
enzymatic release of hydrogen sulfide and methyl mer-
captan from cysteine and methionine, respectively, has
also been observed (Metzler, EB Ei" 1954a; Gruenwedel
and Patnaik, 1971). These reactions have been found to
require the presence of metal ions. Metzler, EE.2$°
(1954b), demonstrated that a side reaction involving

glycine, formed from threonine cleavage, and PL formed

34

B-pyridoxylserine, a compound with no vitamin 86 activity
for S. uvarum. B-pyridoxylserine formation was observed
to be responsible for a loss of up to 8% of the total PL
in an equimolar mixture of PL and threonine at 100°C in
the presence of metal ions. This compound was also found
as a secondary product of the room temperature B-elimi-
nation reactions involving PLP (Reiber, 1976).

A mechanism for the nonenzymatic oxidation of PN
to PL has been identified by Shane and Snell (1972). In
the presence of ferrous ions and low concentrations of
hydrogen peroxide, PN was found to be readily oxidized

to PL at room temperature.

Vitamin B6 Stability and Bioavailability

 

Vitamin B6 in pure solution has been shown to
retain its activity in the presence of heat, acid, and
alkali, although all of the vitamers are readily destroyed
by strong oxidizing agents (Keresztesy and Stevens, 1938;
Cunningham and Snell, 1945). The irreversible degradation
of vitamin B5 by light, particularly ultraviolet wave-
lengths, has long been recognized (Hochelberg, EE.E£"
1943; Cunningham and Snell, 1945). The oxygen-dependent,
light catalyzed, oxidation of PLP to 4-pyridoxic acid
‘phosphate was reported by Morrison and Long (1958) and
Iheiber (1972). Under the same conditions, PL was con-
v1arted to 4-pyridoxic acid, while PMP was converted to

a1: unidentified product (Reiber, 1972). The sensitivity

35

of vitamin 86 in foods was demonstrated by Hellstrom
(1960) by exposure of milk to sunlight for a 2-hour
period, inducing the loss of 50% of the vitamin.

Although pure solutions of vitamin B6 are resis-
tant to thermal degradation, food processing has been
shown to induce significant losses of biologically avail-
able B6 in certain products. In most cases, the mechanism
of loss has not been identified. The relative stability
and behavior of the individual vitamers have not been
well elucidated.

The roasting of many types of meat has been found
to induce a loss of approximately 50% of the total vitamin
B6, as measured by S. uvarum or rat bioassay (Lushbough,
g£_3£., 1959). The rat bioassay data were observed to be
approximately twice the microbiological values for all
samples, raw and roasted. This discrepancy could not be
satisfactorily explained.

Several studies concerning the effects of thermal
processing during canning have indicated that losses of
low levels of vitamin 36 occur. Losses of no greater
than 20% of the total vitamin 36 have been observed dur-
ing the retorting of several types of beans at 120°C
for 30 to 45 minutes (Everson, EE.3£°' 1964; Raab, 25.3l"
1973; Miller, 32 Si., 1973; Daoud, EE.El" 1977). Similar
losses were observed in the processing of canned beef,

while there was no degradation of the vitamin B6 in

36

canned tomato juice concentrate during retort processing
(Everson, SE 31., 1964). No differences were observed
between conventional and aseptic canning methods on the
retention of vitamin B6 (Everson, SE Si’! 1964).

The effects of processing corn and soybean meal
on the content of biologically available vitamin B6 were
studied by Yen, 25 El' (1976). Mild roasting treatments
slightly increased the content of available B6 in corn,
while significant losses were induced by a more severe
roasting treatment. The processing of soybean meal by
autoclaving decreased the content of available vitamin

B by approximately 30%.

6
Although no systematic studies have been carried

out on the storage stability of vitamin B in foods, a

6
number of experimental findings have been reported.
Harding, EE.E£° (1959), observed a 43% lower content

of total vitamin B6 in canned combat rations stored for
20 months at 38°C compared to those stored at 1°C.
During the storage of canned beef and bean products for
270 days at room temperature, approximately 25% of the
vitamin B6 was degraded in each product, although the
initial rate of loss was greater in the beef (Everson,
EE.E£°' 1964). No loss was observed during the storage
of tomato juice concentrate under the same conditions.

These results suggested a pH effect on the rate of

«degradation of the vitamin. Analyzing by rat bioassay,

37

Richardson, EE.2£' (1961), found little difference in the
rate of loss of vitamin B6 at room temperature in stored
products processed by either retorting or gamma irradi-
ation, while frozen storage generally retarded the
degradation. The net retention of vitamin B6 in products
stored 20 months at room temperature after retort pro-
cessing or irradiation was found to be only 40 to 60% of
that of comparable products stored frozen.

Several researchers have examined the stability
of PN used in the fortification of various cereal pro-
ducts. Bunting (1965) reported the retention of 90 to
100% of the PN added to corn meal and macaroni following
storage for 1 year at 38°C and 50% relative humidity.
Cort, EE.E£° (1976), demonstrated no loss of PN in corn
meal which had been fortified with salts of iron, zinc,
magnesium, and calcium, when stored for 6 months at room
temperature. Storage at 45°C for 12 weeks resulted in a
loss of only 4% of the PN. Anderson, 32 El' (1976),
observed no loss of PN in iron fortified breakfast
cereals during storage for 3 months at 40°C or 6 months
at 22°C.

The stability and bioavailability of vitamin B6
has been studied more extensively in dairy products than
in other foods. Most of this research was prompted by
the occurrence of severe vitamin B deficiency in a group

6
<3f infants which were fed diets composed largely of

38

commercially sterilized infant formulas with no B6
fortification (Coursin, 1954).

Pasteurization and cold storage of milk have been
shown to have little effect on the content of total
vitamin B6 in milk, as measured by S. uvarum (Hassinen,
SE Ei'r 1954). The manufacture of sweetened condensed
milk and the spray drying of milk or infant formula have
been shown to induce a loss of approximately 20% of the
total vitamin B6 (Hassinen, SE.E£°I 1954). Commercial
sterilization of evaporated milk or infant formula was
observed to be much more detrimental, resulting in losses

of 40 to 60%, as measured by S. uvarum. However, using

S. sitophilia as the assay organism, only a small loss

 

of biologically active B6 was observed (Hassinen, 2E.2l"
1954; Gregory, 1959). Hassinen, SE El° (1954), demon-
strated that after the sterilization of evaporated milk,
a rapid loss of apparent vitamin B6 (S. uvarum assay)
proceeds for approximately 1 week, followed by a much
slower rate of loss during subsequent storage. Hodson
(1956) also observed this phenomenon, but found a much
slower rate of apparent degradation when using the S.

sitophilia assay. Storage of evaporated milk at room

 

temperature for 6 months induced the loss of approximately

20% and 40-50% when assayed by S. sitophilia and S. uvarum,

 

respectively (Hodson, 1956).
These unexpected results may be explained by the

distribution of total vitamin B6 among the various

39

vitamers in the milks. While over 80% of the vitamin B6
in raw milk exists as PL and PLP, at least 60% of the
total B6 is in the form of PM and PMP in sterilized
evaporated milk (Gregory, 1959; Polansky and Toepfer,

1969). Because S. uvarum exhibits somewhat limited

response to PM, while S. sitophilia shows no such dis-

 

crepancy, the difference was probably one of assay spe-
cificity.

Hassinen, 25 El' (1954), demonstrated that milk
fortified with PN exhibited little loss of the added
vitamin during sterilization and storage. PL and PM
added in fortification were found to be degraded at the
same rate as the B6 occurring naturally in milk.

6 in
sterilized evaporated milk and infant formula remains

The biological availability of vitamin B

unclear after considerable investigation. Tomarelli,
25 El“ (1955), determined the content of available B6
in milk products by rat bioassay. They found excellent
agreement between bioassay and S. uvarum microbiological
data for raw milk and sterilized infant formula which
had been fortified with PN. However, in nonfortified
sterilized infant formula and sterilized milk, rat bio-
assay data indicated that only 50% of the total vitamin
B6 as measured microbiologically was available to the

rats. In direct opposition is the data of Davies, gg_3£.

(1959). These authors observed that rat and chick

40

bioassays of evaporated milk, even after a 6-month
storage period, gave apparent vitamin B6 values which
were significantly greater than S. uvarum data, although
the three assay methods gave comparable results in the
assay of raw milk. Therefore, Davies, 2E.2i° (1959),
concluded that S. uvarum assays may serve effectively
as an indicator of the maximum levels of degradation of
vitamin B6 in milk products.

In an attempt to explain the losses of available
vitamin B6 during the sterilization of milk, Bernhart,
23 Si. (1960), postulated a reaction between the sulfhydryl
groups of milk proteins and PL, forming bis-4-pyridoxyl
disulfide (Wendt and Bernhart, 1960). This compound was
synthesized by heating a concentrated mixture of cysteine
and PL at neutral pH, and was identified in milk which
had been fortified with high levels of PL prior to steril-
ization. Bernhart, 2E.2l° (1960), demonstrated that this

compound possesses vitamin 36 activity of 20% for S.

uvarum, 20% for rats, and 65% for S. sitophilia, relative

 

to the activity of PN. The formation of bis-4-pyridoxyl
disulfide in thermally sterilized nonfortified milk or
infant formula was never reported. Srncova and Davidek
(1972) obtained evidence of the formation of a complex
involving PL and milk protein sulfhydryl groups. How-
ever, calculations based on their data indicate the

involvement of only 2 to 3% of the PL in their highly

41

fortified reaction mixtures. Furthermore, the chemical
nature of these complexes was never identified. The
possibility of substituted aldamine formation involving
the sulfhydryl groups was not ruled out.

Few studies of vitamin B6 biological availability
have been carried out on nondairy food products. Sarma,
35 Ei' (1947), compared rat bioassay and S. uvarum micro-
biological assay data for a variety of natural materials.
The rat bioassay values were lower than S. uvarum data
in most cases. Most of the products showed a difference
of 10 to 30% between the two assays, with a maximum dif-
ference of 34% for whole wheat. The authors suggested
that the observed differences were due to the slightly
lower activity of PL and PM when mixed in the diet of the
rat.

In studies of the vitamin B6 content of canned
combat rations, the response of rats fed only the rations
suggested that vitamin B6 was limiting, although micro-

biological assays for total vitamin B indicated the

6
presence of sufficient levels of the vitamin (Register,
3E'3£., 1950; Tappan, SE Ei" 1953). The addition of

PN to the rations stimulated rat growth in these studies.
As a possible explanation, Sauberlich (1961) observed

that the addition of PN to a diet adequate in B but

6'
limiting in an essential amino acid, would induce a

growth stimulation. Therefore, the low protein quality

42

in the processed rations may have been responsible for
the apparently low bioavailability of the vitamin.

The chick bioassay results of Yen, SE El' (1976),
suggested that the content of available B6 in corn and
soybean meal, either raw or thermally processed, is con-
siderably less than predicted by published microbiological
data. Similarly, Leklem (1977) demonstrated in a human
study that the bioavailability of vitamin B6 in whole
wheat bread was significantly less than that of white
bread. Toepfer, SE Ei' (1963), observed a slight, but
statistically significant, difference between rat and
microbiological assays for whole wheat flour and nonfat

dry milk, suggesting that the vitamin B in these products

6
may not be entirely available biologically. No differences
were observed between rat and microbiological assays for
dried beef and Lima beans.

Further evidence that naturally occurring vitamin
86 may not be completely utilized biologically was found
by Nelson, 25.21' (1976). The intestinal absorption of
the vitamin B6 in orange juice by humans was significantly
less than that of the pure B6 vitamers in saline solution.
Under the experimental conditions, a 30 cm jejunal seg-
ment absorbed 65% of the vitamin B6 in a saline solution,
with a slightly enhanced uptake induced by the addition
6 in
diluted orange juice was absorbed. Whether these data

of glucose. In contrast, only 30% of the vitamin B

43

represented differences in rate of absorption or indi-
cated true differences in absolute bioavailability could

not be determined.

EXPERIMENTAL METHODS

Preparative Methods

 

Heat Stable Peptide Mixture

 

The isolation of heat stable peptides from a pep-
tic hydrolysate of bovine B-lactoglobulin was performed as

described in the Appendix.

e-Pyridoxyl-L—lysine Synthesis

 

The synthesis of d-acetyl-e-pyridoxy1-L-1ysine was
performed according to the method of Dempsey and Christen-
sen (1962). e-Pyridoxyllysine was formed by a modifica-
tion of‘the procedure of Dempsey and Snell (1963). Details
are presented in the Appendix.

Phosphopyridoxyl-Bovine Serum
Albumin

 

The reductive linkage of pyridoxal phosphate to
bovine serum albumin was performed according to the
method of Dempsey and Christensen, procedure 2 (1962).

Details are presented in the Appendix.

Chemical Methods

 

Protein Determination

 

The concentration of proteins and peptides in

solution was determined by the method of Lowry, EE.E£°

44

45

(1951). Crystalline B-lactoglobulin, 3X (United States

Biochemical Corp.) was employed as the standard.

Amino Groups

Protein and peptide amino groups were determined
by the method of Fields (1972) using 2, 4, 6-trinitro-
benzenesulfonic acid dihydrate from Pierce Chemical Co.
Molar absorptivity coefficients at 420 nm for the trini-
trophenyl derivatives of heat stable peptides, B-lacto-
globulin, and serum albumin were calculated to be 19,650,
18,700, and 18,980 M-lcm-l, respectively. These calcu-
lations were based on the published amino acid composi-
tion of B-lactoglobulin and bovine serum albumin and

the reported molar absorptivity coefficients of TNP-

a-amino, TNP-e-amino, and TNP-thiol groups (Fields, 1972).

Sulfhydryl Groups

 

Peptide sulfhydryl groups were determined by the
procedure of Ellman (1959) using the reported molar

1cm-1 at 412 nm for

absorptivity coefficient of 13,600 M-
the 3-carboxy-4-nitrothiophenol anion. The 3, 3'-dithio-
bis(6-nitrobenzoic acid) was obtained from Eastman
Organic Chemicals.

Fluorometric Determination of
Vitamin B5 Compounds

 

The fluorometric determination of PN, PM, and PL

was performed by the semiautomated procedure of Gregory

46

and Kirk (1977a). All phosphorylated forms were deter-
mined as their free base. The fluorometric procedure is
based on the conversion of each vitamer to the fluoro-
phore, 4-pyridoxic acid lactone. Details are presented
in the Appendix.

Erythrocyte Aspartate Amino-
transferase Assay

 

 

The activity of erythrocyte AspAT was assayed, in
the presence and absence of PLP added in vitro, by a
slight modification of the automated serum glutamic-
oxaloacetic transaminase method of Levine and Hill (1966).
This procedure is based upon the coupling of malate dehy-
drogenase to the AspAT reaction, with fluorometric quan-
titation of the oxidation of reduced nicotinamide adenine
dinucleotide. A standard curve for enzyme activity was
determined by running serial dilutions of a hemolysate
of high activity. The enzyme activity of the standard
hemolysate was determined spectrophotometrically by the
Calbiochem (1974) procedure. After calibration of the
AutoAnalyzer, all unknown hemolysates were assayed by
the modified automated procedure. PLP stimulation of
AspAT activity was determined by adjusting 0.9 ml of
each hemolysate to 0.820 mM with respect to added PLP
by the addition of 0.1 ml of an 8.20 mM aqueous solution
of PLP (Bayoumi and Rosalki, 1976). Following incubation

for 30 minutes at room temperature, the hemolysates were

47

reassayed. The percent stimulation induced by the in
vitro addition of PLP was calculated from the observed

enzyme activity in the presence and absence of PLP.

Hemoglobin Determination

 

Determination of hemoglobin in the hemolysates
was performed according to the automated colorimetric
method described by Technicon (1969). Working standards
were prepared by dissolving bovine hemoglobin (2X crystal-
lized, Type I; Sigma Chemical Co.) in distilled water and

diluting over a range of 4 to 20 mg/ml.

Physical Methods

 

Spectrophotometry

 

Ultraviolet and visible absorption spectra were
determined using a Beckman DK-2A spectrophotometer.
Single wavelength measurements were made using a Gilford
Model 240 spectrophotometer. Fluorescence spectra were
determined using an Aminco Bowman spectrophotofluorometer,
equipped with a xenon lamp.

High Performance Liquid Chromato-
graphic Assay of Vitamin B6

 

 

Analyses were performed using an ALC/GPC-202
liquid chromatograph, equipped with a Model M-6000 pump,
Model U6K septumless injector, and a Model 440 absorbance
detector, which were manufactured by Waters Associates.

Chromatographic separations were performed on a uBondapak

48

C18 column (4mm ID x 30 cm, Waters Associates). Fluor—
escence detection was accomplished using an Aminco Fluoro-
Monitor equipped with an 18 pl flow cell. All B6 vitamers
were determined as their free base. The details of sample
extraction, chromatography, and detection are presented

in the Appendix.

High Performance Liquid Chromatographic
Assay of PyridoxyIlysine

 

e-Pyridoxyllysine in protein acid hydrolysates
was determined by a procedure similar to the HPLC assay
of vitamin 36' Details are presented in the Appendix.

High Performance Liquid Chromatographic
Assay of 49Pyridoxic Adid

 

The determination of 4-pyridoxic acid in dehy-
drated model system extracts was performed by a modifi-
cation of the HPLC method of Gregory and Kirk (1977b)
for the determination of 4-pyridoxic acid in urine. The
modifications and details of the procedure are presented
in the Appendix.

Discggtinuous Polyacrylamide
Gel Electrophoresis

Gel electrophoresis of B-lactoglobulin was per-

formed by a modification of the discontinuous method of

Melachouris (1969). Details are presented in the Appendix.

49

Thin Layer Peptide Mapping

Peptide mapping was performed on cellulose MN-300
(Sigma Chemical Co.) plates by the procedure of McDonald,
22 31. (1976). Peptides (230 pg) were applied to the
20 x 20 cm plate with a cellulose support thickness of
0.50 mm. Electrophoresis was carried out for 2 hours in
a Desaga Heidelberg thin layer electrophoresis apparatus
at a field strength of 20 V/cm, 13 mA, in a pH 4.5 buffer
(pyridine: acetic acid: water; 5:9:986), with running
tap water as the coolant. The peptides were visualized

with ninhydrin spray (von Arx and Neher, 1963).

Gel Filtration Chromatography

Gel filtration chromatography was used both in
preparative and analytical applications. Columns of
various dimensions were prepared with Sephadex G-10,
G-25, and G-50 packings as described by the manufacturer
(Pharmacia Fine Chemicals, 1974). Column effluents were
continuously monitored at 254 or 280 nm with an ultra-
violet absorbance detector (Instrumentation Specialties
Co.). Columns were preserved during storage by equili-
bration in 0.1 M ammonium acetate, pH 7.0, containing

0.1% sodium azide, or 0.05 M KCl-HCl, pH 2.5.

50

Biological Methods

 

Microbiological Determination of
Total Vitamin B5

 

The method of Haskell and Snell (1970) was

employed for the assay of total vitamin B using Sac-

6
charomyces uvarum 4228 (ATCC 9080). Stock cultures

 

were maintained on malt agar slants at 2°C. Sample
extracts were added to the assay tubes in 0.05, 0.10,
0.15, and 0.20 ml aliquots and were assayed in tripli-
cate. A dose-response curve using PN as the standard
was prepared for each assay, covering the range of 0 to
10 ng/tube. The vitamin B6 content of each sample was
calculated from the mean value of ng total B6/m1 extract,
as determined from the dose-response curve.

Rat Bioassay of Biologically
Available Vitamin B6

 

The concentration of biologically available
vitamin B6 in dehydrated model systems and phosphopyri-
doxyl-bovine serum albumin (PP-BSA) was determined by a
modification of the methods of Sarma, EE.El° (1946) and
Linkswiler, EE.3l' (1951).

All diets were of the composition shown in
Table l and were prepared by thoroughly mixing the diet
ingredients in a Hobart mixer. The composition of the
standard diets was identical to that of the basal, with

the exception of the addition of a dry pyridoxine

51

Table 1. Composition of diets for rat bioassays.a

 

Test Diets
. Basal, Standard,
Ingred1ent or PP-BSA Diets (model systems)

 

Stored Roasted

 

Casein (vitamin free,

test)b ' 19.80 18.86 18.85
DL-Methionine, N.R.c.C 0.20 0.20 0.20
Sucrose 60.50 57.44 57.45
Cellulose (Alphacel) 9.40 9.40 9.40
Salt mixd 4.00 4.00 4.00
Vitamin mixe 1.10 1.10 1.10
Corn oil 5.00 5.00 5.00
Model system -- 4.00 5.00

 

aWater added to all diets at a level of 52.1 ml
per 3000 g to all diets in Bioassay 2. A 5 mg/ml solution
of PP-BSA replaced the water in diets designated for PP-
BSA additions.

bICN Pharmaceuticals, Inc. (Bioassay l) and Teklad

Test Diets (Bioassay 2).
cGrand Island Biological Co.

dTeklad Test Diets. Wesson modified Osborne-
Mendel. ZnCO3 added to provide 15 ppm Zn in all diets.

eVitamin mix (ICN Pharmaceuticals, Inc.) provided
(per kg diet): vitamin A, 9900 units; vitamin C, 1100
units; a-tocopheral, 55 mg; choline chloride, 825 mg;
menadione, 25 mg; niacin, 50 mg; riboflavin, 11 mg;
calcium pantothenate, 33 mg; thiamin, 0.22 mg; folic
acid, 0.99 mg; vitamin 312' 0.015 mg.

52

hydrochloride premix (prepared in casein) to a level of
0.25, 0.50, 0.75, and 1.00 ug PN per gram diet. The
test diets were prepared by blending finely powdered
dehydrated model systems into the basal diet. In Bio-
assay 1, roasted model systems were incorporated at a
level of 5% of the diet and in Bioassay 2, stored model
systems were incorporated at a level of 4% of the diet.
Replacement of the basal diet ingredients to maintain
composition with addition of the test systems was based
on the model system proximate composition. The formu-
lation of test diets containing phosphopyridoxyl-bovine
serum albumin was performed by substituting 52.1 ml of
a 5 mg/ml solution of the conjugated protein for the
52.1 ml of water added to each of the other diets. All
diets were stored in sealed containers at 2°C until fed.
Male weanling Sprague-Dawley rats (Spartan
Research Animals, Inc., Haslett, MI.) were randomly
asSigned into groups of 8 to 9 rats each for each bio-
assay. The rats were individually housed in metal cages
with raised wire mesh floors. Water was supplied ad
libitum. The basal diet was fed to all rats for 14 days
to induce a mild state of vitamin B6 depletion. The
rats were weighed at the start of the assay period and
weekly thereafter. Standard (0.00, 0.25, 0.50, 0.75,
and 1.00 pg PN/g diet) and test diets (containing model

systems or PP-BSA) were fed to the respective groups ad

53

libitum for 21 days. Feed consumption was carefully
measured. Spillage was found to be low, and, therefore,
it was neglected.

After the 21-day assay period, the rats were
killed by decapitation. Blood was collected into 16 x
100 mm screw cap test tubes containing several drops of
a 100 U/ml aqueous solution of heparin (Grade 1, sodium
salt; Sigma Chemical Co.). Immediately after collection,
samples were gently mixed with the heparin solution and
then centrifuged in a clinical centrifuge (Damon/IEC
Division) at approximately 3000 rpm. The plasma of
each sample was removed by aspiration, and the erythro-
cytes washed by suspension in cold 0.9% (w/v) saline
solution and recentrifuged. After removal of the saline
wash solution, the packed erythrocytes were stored at
-25°C until assayed. On the day of the AspAT assays,
the tubes containing the erythrocytes were thawed and
10 ml of distilled water added to each. The contents
of each tube were then thoroughly mixed on a vortex
mixer, and cell debris sedimented by centrifugation in
the clinical centrifuge at 3000 rpm. The supernatants
were assayed for AspAT activity and hemoglobin content
without dilution.

The bioassay results were evaluated by comparison
of the response of the animals on the test diets to the

dose-response results of animals fed the standard diets.

S4

Does-response curves were determined by linear regression
techniques (Neter and Wasserman, 1974). The logarithm of
the dietary PN concentration (pg/g) was plotted against
rat growth, growth per gram feed consumed, and percent
PLP stimulation of AspAT. The dietary PN concentration
was coded to permit taking the 10g of 0.00 ug/g by the
addition of l to each standard concentration, as described
by Bliss and White (1967). Because of the nonlinearity
of the log dose-response curves observed for the enzyme
activity data, several other methods of plotting the data
were evaluated. These included: linear dose versus
activity, log dose versus log activity, and reciprocal
dose versus reciprocal activity (Bliss and White, 1967).
The enzyme activity curve providing the highest linear
correlation coefficient was used for quantitation in

each bioassay.

Experimental Design and Procedures

Interaction of Pyridoxal and Pyridoxal
Phosphate with Peptideslin a Liquid
Food System durinnghermal Processing

 

The behavior of PL and PLP in liquid model sys-
tems was examined to determine whether complexes may be
formed which could affect the biological availability of
the vitamin. The aldehyde forms of the vitamin were
examined in this experiment because they have been shown
to be the most reactive of the B6 vitamers (Apderson,

EE.El°' 1974). Peptides, rather than intact proteins,

55

were chosen on the basis of their thermal solubility
properties which facilitated the stabilization of Schiff
bases with sodium borohydride (NaBH4), gel filtration
chromatography to recover the peptides after heat treat-
ment, and Spectrophotometric evaluation of the peptide
bound vitamin. The effects of glucose and ascorbic acid
were also examined. Glucose, through its role in Maillard
browning, may compete with PL or PLP for free amino
groups, thus acting as a binding inhibitor. Ascorbic

acid or certain browning intermediates may be sufficiently
strong reducing agents to promote pyridoxylamino compound

formation by reduction of Schiff base forms of the vitamin.

1. Characterization of Peptides.--The sulfhydryl
content of the heat stable peptides was measured (Ellman,
1959) after denaturation with 8 M urea in 0.1 M potassium
phosphate, pH 8.0. Free amino groups (Fields, 1972) and
peptide (Lowry, st 21., 1951) concentrations were also
determined.

The number of peptide species in the heat stable
mixture was determined by peptide mapping, as previously
described.

Gel filtration chromatography on Sephadex G-10,
G-25, and G-50 columns was employed to estimate the
molecular weight distribution of the peptide mixture.

The Sephadex G-50 column (2.6 x 40 cm) was calibrated

using blue dextran (Pharmacia Fine Chemicals) to

56

determine the void volume, and the following proteins:
trypsin, 23,300 daltons (ICN Pharmaceuticals, Inc.);
ribonuclease A, 13,700 daltons (Worthington Biochemical
Corp.); insulin, monomer = 5,733 daltons (Sigma Chemical
Co.); insulin B-chain, monomer = 3,378 daltons (Sigma
Chemical Co.); and insulin A-chain, monomer = 2,360
daltons (Sigma Chemical Co.). All standards were chro-
matographed individually. Insulin undergoes a pH-depen-
dent aggregation reaction, and, therefore, this protein
could serve as a standard in several polymeric forms.
Oncley, 32 31. (1952), observed that, at neutral pH,
insulin exists as a hexamer, while at pH 2, the dimer
predominates. Hence, when dissolved and chromatographed
in 0.01 N HCl, insulin was used as a molecular weight
standard of 11,466 daltons, where in 0.1 M ammonium
acetate, pH 7.0, insulin was used as a 34,398 dalton
standard. The A- and B-chains of insulin were chroma-
tographed in 0.1 M ammonium acetate, pH 7.0, containing
0.1% 2-mercaptoethanol, maintaining the peptides in their
reduced state. Under these conditions, both assumed

the dimer state. Attempts to dissociate intact insulin
and the two subunits to their respective monomer states
by chromatography in a pH 10.3 carbonate buffér
(Fredericq, 1953), even with 8 M urea, failed. Con-
tamination of the samples with small amounts of zinc, as

(iiscussed by Klostermeyer and Humbel (1966), may have

57

been responsible for the persistent aggregation. Trypsin,
ribonuclease, and the heat stable peptides were dissolved
and chromatographed in 0.1 M ammonium acetate, pH 7.0.

The ratio of the elution volume to the void volume was
plotted against the log of the molecular weight for each
standard. Pyridoxal, L-tryptophan, and sodium azide were
the low molecular weight compounds chromatographed indi-
vidually to determine the internal volume of the Sephadex
G-50 column. The molecular weight range of the peptide
mixture was determined by comparison with the calibration

curve .

2. Pyridoxal and Pyridoxal Phosphate Binding

 

Studies.--

a. Stock Solutions.--Stock solutions of

 

pyridoxal hydrochloride and pyridoxal phosphate mono—
hydrate (ICN Pharmaceuticals, Inc.), L-ascorbic acid
(Sigma Chemical Co.), and D-glucose (Matheson, Coleman,
and Bell) were prepared in 0.1 M ammonium acetate and
adjusted to pH 7.0 with ammonium hydroxide, as follows:

1. Pyridoxal, 500 ug free base/ml
(2.93 mM), was prepared by dissolving 15.1 mg pyridoxal
HCl in 25 m1.

2. Pyridoxal phosphate, 670 ug free
base/m1 (2.93 mM), was prepared by dissolving 18.1 mg

pyridoxal phOSphate monohydrate in 25 ml.

58

3. L-ascorbic acid, 1.50 mg/ml, was pre-
pared by dissolving 37.5 mg in 25 m1.
4. D-glucose, 100 mg/ml, was prepared by
dissolving 5 g in 50 ml.
All stock solutions were stored at 2°C. Vitamin
stock solutions were protected from light and kept no
longer than one week.

b. Liquid Model System Preparation and

 

Treatment.--Model system mixtures were prepared in 16 x

 

100 mm screw cap test tubes by the addition of 0.1 ml of
each appropriate stock solution to 0.7 m1 of the 1.57 mg/ml
peptide solution, as shown in Table 2.

Table 2. Formulation of liquid model systems for study
of PL and PLP interactions.

 

ml Stock Solution

 

 

Treatment
Peptides Ascorbate Glucose PL or PLP
Control 0.7 0 0 0.1
Ascorbate 0.7 0.1 0 0.1
Glucose 0.7 0 0.1 0.1

Glucose &
Ascorbate 0.7 0.1 0.1 0.1

 

Final concentrations of components, when present, were:
peptides, 1.10 mg/ml; PL, 50 ug/ml (0.293 mM); PLP
67 ug/ml (0.293 mM); glucose, 10 mg/ml, ascorbate,

150 ug/ml. All model systems were adjusted to 1.0 ml

59

total volume with 0.1 M ammonium acetate. After mixing,
the samples were autoclaved 20 minutes at 121°C, and upon
removal from the autoclave, cooled to 2°C. A correspond-
ing blank containing no added PL or PLP was prepared for
each sample and was treated in a manner identical to the
sample. Each model system was prepared, treated, and
analyzed in duplicate.

c. Evaluation of Binding to Peptides.--After

 

heating at 121°C for 20 minutes, for the determination of
total peptide bound PL or PLP all model systems were

treated with 3 mg solid NaBH at 2°C. This stabilized

4
the labile Schiff base forms by conversion to pyridoxyl-
amino compounds. Thus, the total bound forms of the
vitamin would represent these reduced Schiff base forms
in addition to any nonreducible substituted aldamine or
pyridoxylamino complexes present. Determination of PL
or PLP bound solely by nonreducible linkages was carried
out by omitting the NaBH4 treatment. All samples were
held at 2°C until chromatographed as described below.

Peptides were isolated from the treated model
system solutions by gel filtration chromatography on a
Sephadex G-10 column (1.6 x 30 cm, Vo = 28 ml). The
entire 1 ml model system solution was chromatographed
using 0.1 M ammonium acetate, pH 7.0. The column

effluent was monitored at 254 nm and the peptides were

recovered from the void volume of the column.

60

The bioavailability of the nonreducible complexes
was estimated by determination of PLP which remained
bound following acidification of the model systems using
the following procedure: after autoclaving the model
systems, they were cooled and adjusted to pH 2.5 with
3 N HCl. Following overnight incubation at 2°C, the pep-
tides were isolated using a Sephadex G-10 column (1.6 x
30 cm) equilibrated and eluted with 0.05 M KCl-HCl,
pH 2.5.

Browning in the model systems was quantitated by
measuring the free amino groups in the peptides collected
from the Sephadex column. The concentration of free
sulfhydryl groups was determined (Ellman, 1959) before
and after gel filtration isolation of the peptides from
the model systems.

The concentration of PL or PLP bound to the pep-
tides was determined spectrophotometrically using the

1cm-l) at 323 nm

molar absorptivity coefficient (5800 M-

for bound pyridoxyl groups (Dempsey and Christensen,

1962). Correction for blank absorbance was made using

an equivalent peptide concentration. Peptide concen-

trations were determined according to Lowry, et 31. (1951).
To minimize photochemical destruction of PL and

PLP, all experimental procedures were run in the absence

of incandescent light and direct sunlight. Two parallel

trials of each binding study were performed. The data

61

were analyzed by analysis of variance employing 2 x 2 and

3 x 2 factorial designs (Neter and Wasserman, 1974).

Interactions of Pyridoxamine in a Liquid
Food System during Thermal Processing

 

 

Using the initial carbonyl-amine condensation
reactions of nonenzymatic browning as a theoretical model,
it was postulated that PM may condense with reducing
sugars to form Schiff bases. Thermal rearrangement of
the sugar-PM complexes could affect the biological
availability of the PM.

This experiment was designed to study the behavior
of PM in a liquid model system at neutral pH. The sta-
bility of PM and its participation of transamination or
the formation of stable complexes during thermal process-
ing were investigated. The effects of glucose and ascorbic
acid on the vitamin were examined, as in the previous
experiment. Peptides were omitted from these model sys-
tems to facilitate analysis by high performance liquid

chromatography.

1. Model System Formulation and Treatment.--

 

Pyridoxamine dihydrochloride (ICN Pharmaceuticals, Inc.),
L-ascorbic acid, and D-glucose stock solutions were pre-
pared in 0.1 M ammonium acetate and adjusted to pH 7.0
with ammonium hydroxide. Pyridoxamine, 500 ug free

base/ml (2.93 mM), was prepared by dissolving 17.8 mg

62

in 25 ml. L-ascorbic acid (1.50 mg/ml) and D-glucose
(100 mg/ml) stock solutions were prepared as previously
described.

Duplicate model system mixtures were prepared in
0.1 M ammonium acetate, as shown in Table 3.

Table 3. Formulation of liquid model systems for study
of pyridoxamine interactions.

 

m1 Stock Solution

 

 

Treatment
Glucose Ascorbate Pyridoxamine
Control 0 0 0.4
Glucose 0.4 0 0.4
Ascorbate 0 0.4 0.4
Glucose & Ascorbate 0.4 0.4 0.4

 

Final concentrations of the components, when present,
were: PM, 50 ug/ml (0.293 mM); glucose, 10 mg/ml; and
ascorbate, 150 ug/ml. All model systems were adjusted

to 4.0 ml total volume with 0.1 M ammonium acetate.
Corresponding blanks containing no PM were prepared for
each sample and treated in an identical manner. Each
model system was autoclaved 20 minutes at 121°C, then
cooled to 2°C. To induce dissociation of any acid labile
complexes, the autoclaved solutions were adjusted to pH

2.5 with 3 N HCl, followed by incubation overnight at 2°C.

63

2. Evaluation of Thermal Effects.--

 

a. Stability of Pyridoxamine.--The concen-

 

tration of PM and the other B6 vitamers in the autoclaved
model systems was determined by the HPLC method for the
quantitative analysis of the B6 vitamers, as described

in the Appendix. The HPLC analysis was performed as
described, with the exception of the omission of the
extraction and enzymatic hydrolysis steps. Aliquots of
the liquid model systems were injected directly into the
HPLC; no TCA was added to these samples. It was found
that the absence of TCA shortened the retention time of
the vitamers. Therefore, a standard curve was prepared

for the B vitamers under the conditions employed for the

6
analysis of the liquid model system samples. Using this
HPLC procedure, the possible conversion of PM to PN or
PL could be readily detected.

b. Gel Filtration Chromatography.--Two ml

 

aliquots of the PM-fortified liquid model systems and
their blanks were chromatographed on a Sephadex G-10
column (2.6 x 34 cm: Vo = 63 ml). The column was equil-
ibrated and the samples chromatographed with 0.1 M
ammonium acetate, pH 7.0. The absorbance of the effluents
was monitored at 254 nm. Changes in the molecular size
distribution of the model system components versus blanks
‘would provide evidence of the formation of acid stable

complexes.

64

3. Statistical Analysis.--The significance of

 

the effects of glucose and ascorbic acid on the thermal
stability of PM was determined by two—way factorial
analysis of variance. Differences between treatment
means were determined by the Tukey procedure for pair-
wise comparisons (Neter and Wasserman, 1974).
Stability and BiggyailabilityofIZitamin Begin
thydratedModel Food Systems: Roasting
Effects

The results of Yen, st 31. (1976), have indi-
cated that roasting of shelled corn during the production
of animal feeds may reduce the amount of biologically
available vitamin B . Few other studies have examined

6

the stability of vitamin B in low moisture foods during

6
roasting processes.

The purposes of this experiment were: (1) to
determine the extent of destruction and relative sta-
bility of the B6 vitamers during roasting, (2) to compare
the results of microbiological, HPLC, and fluorometric
methods of vitamin 86 determination, and (3) to determine
the bioavailability of vitamin B6 in the roasted model
systems. The severe time-temperature conditions employed
in the processing of the model systems were selected in
order to quantitate the maximum effect that could be

encountered during the commercial roasting of foods

and feeds.

65

l. Formulation and Treatment of the Model
Systems.--The dehydrated model system, formulated to
simulate breakfast cereal composition, was a modification
of that designed by Kirk, 22 31. (1977). The composition
is shown in Table 4. Four model systems were prepared,
differing only in vitamin B6 fortification. Model systems
were fortified with: pyridoxine hydrochloride, pyridoxa-
mine dihydrochloride, or pyridoxal phosphate monohydrate.
A nonfortified blank was also prepared. Each model system
was homogenized as a slurry of 40% solids by mixing all
ingredients in distilled water in a Waring blender.
Vitamin B6 fortification was performed by thoroughly
mixing aqueous solutions of the respective vitamers with
the slurry. Fortification was to a level of approxi-
mately 25% NAS/NRC recommended dietary allowance (0.5 mg)
per ounce.

The slurries were transferred to shallow trays,
cooled to -40°C, and dried in a Virtis Model FFD 42 WS
Freeze-Dryer at a platen temperature of 110°F to a con-
stant absolute pressure of 5 pm Hg. After drying, the
model systems were finely powdered in a Waring blender
and layered on stainless steel sheets at a thickness of
approximately 3 mm. The trays were placed in an oven
at 180°C and held for 25 minutes, after which the model
systems had a golden brown appearance. The roasted model

systems, and aliquots of each taken prior to roasting,

66

were stored in stoppered glass containers at 2°C until

assayed for their vitamin B6 contents.

Table 4. Composition of dehydrated model system for
determination of roasting effects.

 

 

Component Percentage by Weight
Casein, vitamin free testa 5
Corn syrup solidsb 37
Starchc 50
Cellulose (Alphacel)d 7
Sodium chloride 1

 

a . .
General Biochemicals, Inc.

bMalto-Dextrin (D. E. = 15). Provided approxi-
mately 5% reducing sugar. American Maize.

cFood grade powdered starch. A. E. Staley Mfg. Co.

dICN Pharmaceuticals, Inc.

2. Analysis of Model Systems.--The effects of

 

roasting on the stability of vitamin 36 in the model
systems was determined by comparison of assay results
from roasted and unroasted system samples. In order to

compare the specificity of the vitamin B assay pro-

6
cedures, the model systems were assayed by microbiological
(Haskell and Snell, 1970), fluorometric (Gregory and

Kirk, 1977a; Appendix), and HPLC (Appendix) methods.

67

Biologically available vitamin B in the roasted

6
model systems was determined by rat bioassay. Rat growth,
growth per gram feed consumed, erythrocyte AspAT activity,
and PLP stimulation of AspAT were employed as indicators
of vitamin B6 status. Each roasted model system was

incorporated into the test diet at a level of 5% by

weight (Table l).

3. Statistical Analysis.--The presence of sig-

 

nificant differences among the rat bioassay estimates of
available vitamin B6 using the various criteria of B6
status was tested by analysis of variance, employing a
randomized complete block design.

For each model system, completely randomized dis-
tribution analysis of variance (ANOVA) was used to com-
pare estimates by any one of the bioassay criteria with
the microbiological and fluorometric assay data. An
ANOVA was run repeatedly for each model system in which
the microbiological and fluorometric assay data were
tested against the bioassay data based on each of the
response parameters. This repeated analysis for each
model system was necessitated because of the lack of
independence between the bioassay estimates; that is, for
each model system, each of the four bioassay estimates
was derived from the response of the same group of rats.

Therefore, to achieve a family error rate of a = 0.05,

the critical probability employed in each of the four

68

ANOVAs for each model system was a/4 = 0.0125. All sta-

tistical procedures were described by Neter and Wasserman

(1974).

Stability and Bioavailability of Vitamin B5
ififflefiyarafed Mbdél FOOG'SysEém§:
Storage Effects

 

 

 

Previous research has demonstrated that naturally
occurring vitamin B6 in foods may undergo degradation
during prolonged storage (Harding, 33 31., 1959; Richard-
son, gt_gl., 1961; Everson, 25 $1., 1964). The storage
stability of vitamin B6 in PN fortified cereal products
has also been examined (Bunting, 1965; Cort, 22 31.,
1976; Anderson, 25 31., 1976). No studies to date have
quantitatively examined the relative stability of PN,

PM, and PLP in foods. Although several studies have
demonstrated the occurrence of transamination reactions
during thermal processing, the possible occurrence of
interconversions among the B6 vitamers during storage has
not been examined. Similarly, the products of vitamin B6
degradation in foods have not been identified.

In this study, the stability of various forms of
vitamin B6 was examined in dehydrated model food systems.
These systems were equilibrated to a water activity of
approximately 0.65 and stored in metal cans at 37°C:
conditions which have been previously shown by numerous
researchers to produce high rates of vitamin degradation

and nonenzymatic browning. Storage under these conditions

69

should give maximum rates of B6 degradation that may be
encountered in food systems during storage. The experi—
mental objectives were: (1) to determine the kinetics
of degradation and/or interconversion of PL, PM, PLP,
and PN during storage, (2) to relate the vitamin degra-
dation to the rate of nonenzymatic browning, (3) to
determine the effect of headspace oxygen content in the
container on the stability of the B6 vitamers, (4) to
identify the degradation products of the B6 vitamers,
(5) to compare fluorometric, microbiological, and HPLC
assay results for vitamin B6 in the model systems, and
(6) to determine the bioavailability of vitamin B6 in

the model systems after the storage period of 128 days.

1. Formulation, Processing, and Storage of

Model Systems.--The dehydrated model food system was

 

identical in composition to that previously employed
(Table 4), with the exception that bovine serum albumin
(essentially fatty acid-free; Sigma Chemical Co.) and
bovine B-lactoglobulin (United States Biochemical Corp.)
were substituted for the casein at levels of 4% and 1%
of the total solids, respectively. These proteins were
selected on the basis of their high solubility which
facilitated their isolation from the model systems
after storage. This permitted determination of protein

bound forms of vitamin BG. Maillard browning pigments

70

and losses of free amino groups in the soluble proteins
served as convenient indicators of nonenzymatic browning
in the systems.

The model systems were formulated, fortified, and
freeze dried as previously described. Equilibration was
performed by the method of Palnitkar and Heldman (1971),
in which air, conditioned to a relative humidity of 65%
by an Aminco-Airs unit, was forced for 24 hours over
pieces of the freeze dried model systems of approximately
1 cm thickness. The systems were equilibrated on the
adsorption branch of the water sorption hysteresis loop.

After equilibration, aliquots of the model systems
were quickly packaged in 208 x 006 thermal death time
(TDT) and 303 x 406 (303) enameled cans to prevent trans-
fer of water vapor and oxygen into or out of the container
(Kirk, 25 21., 1977). Each can, irrespective of size,
contained approximately 10 to 12 g of the appropriately
fortified model system. The use of large (303) and
small volume containers permitted storage under atmos-
pheres differing greatly in total oxygen content. All
canned samples were stored in a constant temperature

incubator at 37 t 1°C.

2. Analysis of Model Systems.--
a. Physical Measurements.--The moisture con-
tents of the equilibrated model systems were determined

by the method of Kirk, 25 El. (1977). Samples were dried

71

in a vacuum oven at a vacuum of 28 inches Hg at 37°C.
The use of a cold trap between the oven and the vacuum
pump facilitated water vapor transfer from the samples.
Displacement of water vapor from the oven was accelerated
by slow admission of air at a rate of approximately
20 ml/min after drying by bubbling through sulfuric acid.
The moisture content, expressed on a dry weight basis, was
determined by sample weight loss after each had reached
constant weight.

The water activity of each system was measured
directly by the vapor pressure manometric method of
Taylor (1961), as described by Sood and Heldman (1974).

b. Determination of Nonenzymatic Browning.--

 

The progression of nonenzymatic browning during storage
was evaluated by monitoring losses of protein free amino
groups and formation of protein bound melanoidin pigments.
The soluble protein was extracted by mixing 1 g of each
model system with 15 ml 0.05 M KCl-HCl, pH 2.5, and soni-
cating 30 minutes in a 65 watt ultrasonic cleaning bath
(Fisher Scientific Co.). The mixtures were centrifuged
20 minutes at 6000 rpm in a Beckman Model J-21C centrifuge,
rotor JA-20. Ten ml aliquots of the supernatants were
placed in 0.39 inch diameter dialysis tubing and dialyzed
3 days against 4 1 of 0.05 M KCl-HCl, pH 2.5, with two
changes of the buffer. To remove any fine particulates,

the dialyzed extracts were filtered through Whatman 42

72

filter paper. The protein concentrations were determined
(Lowry, 95 31., 1951), and the absorbance of each extract
was measured against a buffer reference at 420 nm. The
absorbance at 420 nm per mg protein/ml was used as an
index of nonenzymatic browning. The free amino group
concentrations were determined in the soluble proteins
(Fields, 1972) to permit evaluation of the loss of a
reactant in the browning reaction.

c. Rate of Vitamin B5 Degradation.--The con-

 

centration of vitamin 36 in the dehydrated model systems
was determined at time = 0 (1 day after canning) and at
subsequent intervals by the HPLC method previously
described. Model systems which were stored in TDT cans
were assayed approximately every 3 weeks over the 128-day
storage period, while the model systems packaged in 303
cans were assayed at the end of storage.

d. Vitamin B5 Degradation Products.--The

 

postulated degradation products, 4-pyridoxic acid and
pyridoxyllysine, were determined by the HPLC methods
described in the Appendix. Analysis was performed at
the end of the 128-day storage period.

e. Bioavailability of Model System Vitamin

 

B5.--Prior to incorporation into the bioassay test diets,
the model systems which were stored in TDT and 303 cans
were blended, yielding a mixture comprised of approximately

60% 303 and 40% TDT can stored model systems. These

73

mixtures were assayed for their vitamin B6 contents by
microbiological (Haskell and Snell, 1970), fluorometric
(Gregory and Kirk, 1977a; Appendix), and HPLC methods
previously described. The biologically available vitamin
B6 in the model systems was determined by the rat bioassay
methods as previously described.

f. Statistical Methods.--In these studies,

 

vitamin 36 degradation rate constants and their errors
were determined by linear regression methods. The sig-
nificance of differences between assay methods for
vitamin B6 was tested by analysis of variance. Statis-
tical analysis of the rat bioassay data was performed as
described for the bioassay of the roasted model systems
(c. f. pages 67-68). All procedures were described by
Neter and Wasserman (1974).

Biological Activity of Phosphopyridoxyl-
Bovine Serum Albumin

 

Although previous research has demonstrated vary-
ing vitamin B6 activities for several pyridoxylamino
compounds, the activity of protein bound derivatives has
never been investigated. No antivitamin activity has
ever been reported for the pyridoxylamino acids examined.
McCormick and Snell (1961), in an extensive study of
pyridoxal phosphokinase, reported that while most pyri-
doxylamino acids had no inhibitory activity, other pyri-

doxylamino compounds were potent inhibitors. Because

74

pyridoxal phosphokinase is essential for the conversion
of dietary vitamin B6 to the metabolically active phos-
phorylated derivatives, their data suggested that certain
pyridoxyamino compounds may act as B6 antivitamins.

The purposes of this experiment were: (1) to
determine the apparent vitamin B6 content of phospho-
pyridoxyl-bovine serum albumin (PP-BSA) using the micro-
biological and fluorometric assay procedures based on
autoclave extraction, (2) to determine the availability
of protein bound pyridoxylamino groups for utilization
by the rat, and (3) to determine whether the presence

of PP-BSA in the rat diet affects the utilization of

dietary pyridoxine.

1. Analysis of Phosphopyridoxyl-Bovine Serum
Albumin.--PP-BSA was prepared as previously described.
Pyridoxylamino residues were quantitatively determined
from the ultraviolet-visible absorption difference spec-
trum of a 1 mg/ml solution of PP-BSA in 0.1 M ammonium
acetate, pH 7.0, against a 1 mg/ml solution of bovine
serum albumin (Sigma Chemical Co.) in the same buffer.
Calculations were based on the molar absorptivity coef-
ficient at 323 nm for protein bound pyridoxyl groups of

lam-l

5800 M- (Dempsey and Christensen, 1962). Pyridoxyl-
amino groups were also quantitated using the HPLC method
for the determination of pyridoxyllysine in protein

hydrolysates, as previously described.

75

The PP-BSA preparation was examined for the
presence of free vitamin B6 contaminants by gel fil-
tration chromatography over a Sephadex G-25 column
(2.6 x 36 cm, Vo = 78 ml) equilibrated and eluted with
0.1 M ammonium acetate pH 7.0. The column effluents
were monitored at 254 nm with an ultraviolet absorption
detector. Trace contamination of free vitamin B6 in
the PP-BSA was quantitated by the HPLC method for vitamin
B6 free bases. The assays were performed as previously
described, with the exception that the sample size of
PP-BSA was 25 mg.

Samples of the PP-BSA preparation were analyzed
by microbiological (Haskell and Snell, 1970) and fluoro-
metric (Gregory and Kirk, 1977a) methods to determine the
extent of release of the protein bound pyridoxyl groups
during the autoclave extraction. Apparent vitamin B6
levels were corrected by subtracting the low level of
free vitamin B6 in the PP-BSA, as determined by HPLC

(buffer extraction, rather than autoclave extraction

methods).

2. Determination of Biological Activity.--The

rat bioassay for biologically available vitamin B was

6
performed as previously described. The bioassay was

run parallel to the assay of stored model systems (Bio-
assay 2), and utilized the same dose-response curves for

quantitation. The biological availability of the bound

76

vitamin 36 in the PP-BSA was determined from the response
of rats fed a test diet with PP-BSA as the sole source of
vitamin 36. Antivitamin B6 activity of the PP-BSA was
determined from the response of rats fed diets contain-
ing the same level of PP-BSA plus 0.25 ug pyridoxine

free base per gram. The extent of utilization of the
free PN was found by comparing the response to the diet
containing the added PN to the response expected from

the dose-response curves. In each diet, the estimated
vitamin B content of the PP-BSA was corrected for free

6

36 contaminants by subtraction of the value determined

by HPLC vitamin B assay.

6

RESULTS

Interaction of Pyridoxal and Pyridoxal Phosphate
with Peptides in a Li uid Model Food System
during Thermal Processing

 

 

 

Preparation and Characterization
of Heat Stable Peptides

 

 

 

The commercially obtained preparation of B-lacto-
globulin, when examined by discontinuous polyacrylamide
gel electrophoresis (Melachouris, 1969), was found to
contain several minor contaminants (Figure 3). Upon
denaturation with 8 M urea, the protein exhibited 1.53
moles of free sulfhydryl (SH) groups per mole of B-lacto-
globulin dimer (36,000 daltons). In the absence of urea,
0.10 moles SH per mole dimer were detected. For the
preparation of heat stable peptides, this protein was
used without further purification.

Heat stable peptides were isolated from a peptic
hydrolysate of B-lactoglobulin with a yield of approxi-
mately 5%. Analysis of the isolated peptides revealed
1.26 umoles of free amino groups per mg. No free SH
groups could be detected in the presence or absence of
urea. Mapping of the peptide mixture revealed a high
degree of heterogeneity: at least 17 peptide species were

detected (Figure 4).

77

78

Figure 3. Polyacrylamide gel electropherogram of
B-lactoglobulin which was used in isolation
of heat stable peptides. Slots 1—4 con-
tained 75, 100, 125, and 150 pl of a 5 mg/ml
protein solution.

79

 

Figure 3

80

 

Figure 4. Thin layer peptide map of heat stable peptides.
The origin is designated by x.

81

The molecular weight range of the peptide mixture
was determined by gel filtration chromatography on
Sephadex G-10, G-ZS, and G-SO. The largest fraction of
the heat stable peptides eluted in the void volume of
the Sephadex G-25 column (exclusion limit = 5000 daltons).
A portion of the peptide mixture eluted as an incompletely
resolved shoulder of the void volume peak from the
Sephadex G-25 column, suggesting that this fraction was
comprised of species of molecular weight slightly less
than 5000. A small fraction of the heat stable mixture
eluted from the G-25 column in the total volume of the
column, indicating a molecular weight 5 1000. Further
analysis of the heat stable peptide mixture was performed
on Sephadex G-10. Elution of the peptides in the void
volume of the column indicated that all species had
molecular weights greater than the 700 dalton exclusion
limit.

Confirmation of the molecular weight ranges was
obtained by chromatography on a Sephadex G-50 column
which was calibrated as shown in Figure 5. The chroma-
togram of the heat stable peptide mixture (Figure 6)
indicated that the most prevalent peptide species were

those of approximately 4,500 and 12,500 daltons.

82

.uxou ca conwuomoo mQOfluwccoo Hoods oonmmno
Ioumeouno mums mommocmum .0>H:o sowumunflamo Aocflmv omlu xopmnmom .m shaman

 

    
   
        

O).
\¢>
O.m 9N N.N 0.. e... 0..
1 u q d W N

1 n W

9

1 ¢ m

.353 52.014 0 WI

5.3.... 1 J

.36. 1 M

.9 52.010 1 h mm.

5.3:. 1 .D

. N

.1 O— x.

.025 5.3:. mm
o 4 macs-Scoem .4 m. n...

. om M.

53?... W

1 On MW

.3598: £32..
o¢

 

83

OM

.ououxHE moflummm magnum now: no Emumoumfiouno omlm xmpmnmom

o
>
\m>

9N O.N 0..

.m Guzman

0..

 

 

I a 1

\L

 

 

(wuoez) aouoqlosqv

84

Binding of PL and PLP in
Liquid Model Systems

 

 

Determination of total bound PL and PLP was
carried out using liquid model systems which were treated
with sodium borohydride after autoclave processing for
20 minutes at 121°C. Peptide bound forms of vitamin B6
would thus represent borohydride-reduced Schiff base
derivatives plus nonreducible complexes such as substi-
tuted aldamines and pyridoxylamino compounds. The pep-
tides were isolated from the autoclaved liquid model sys-
tems by chromatography on a column of Sephadex G-10. The
presence of peptide bound PL or PLP was demonstrated by
the UV-visible difference absorption spectra of each
peptide fraction versus its respective blank. Figure 7
is a difference spectrum recording typical of those
observed for all model systems. The absorption maxima
at 323-325 and 253-254 nm are characteristic of the
absorption spectra of pyridoxylamino and substituted
aldamine derivatives (Dempsey and Christensen, 1962;
Kierska and Maslinski, 1971; O'Leary, 1971) and demon-

strated the presence of peptide bound vitamin B in the

6
heat treated liquid model systems. Quantitation of the

levels of total peptide bound vitamin B (Table 5)

6
indicated that PLP was much more reactive than PL

(P < .001), as had been previously reported. Factorial
analysis of the data showed that glucose, ascorbic acid,

and their interaction did not significantly affect the

85

.o.h mm .mumumom sasaossm z H.o u ummmnm .xcman ooflmauuou
loo: 0:» Eonm omonu momuo> Eoumam H0005 owovfla vowmwuHOMImnm o Bonn

 

 

 

ooUMHOmH moowumom mo Esuuoomm coeumuomnm monouommwo oHnHmH>IuoHow>muuHD .5 «Human
3550:2963
ave owe .N can can cow ecu
. . . . . 1 . .290.
1 . .
.u_o av
O.
8.
O
J
a.
D
.u
«u
.8
.ouo.
.ono.

86

levels of bound PL or PLP (P < .05). Inspection of these

data suggests, however, that the presence of ascorbate

may have had a slight inhibitory effect on the net bind-

ing of PL and PLP.

Table 5. Total peptide bound PL and PLP. Data repre-
sents percent of model system B5 present as a

Schiff base, substituted aldamine, or pyri-
doxylamino complex at pH 7.0.

 

 

 

 

PL PLPb
Treatmenta
1 2 1 2
Control 8.1 3.0 27.8 20.0
Ascorbate 3.0 3.9 ' 15.1 9.5
Glucose 8.9 6.1 23.9 22.3
Glucose & Ascorbate 10.4 9.9 26.3 15.7

 

aEffects of glucose, ascorbate, and their inter-
action were not significant (P < .05).

bPLP was significantly more reactive (P < .001).

Nonreducible complexed forms of vitamin B6, sub-
stituted aldamine and pyridoxylamino derivatives, were
determined spectrophotometrically after isolation of the
peptides from heated liquid model systems which were not
treated with borohydride. It was presumed that the
labile Schiff base forms would dissociate during peptide
isolation, leaving only the stable nonreducible com-
plexes. In determining the nonreducible complexes, only
PLP was studied because of the low reactivity observed

with PL in the previous studies. UV-visible difference

87

spectra were determined on the peptide solutions from
the autoclaved model systems which had not undergone
NaBH4 reduction. In all cases, the observed spectra
were qualitatively identical to Figure 7, indicating
the presence of peptide bound nonreducible complexes.
The absence of an absorption maximum at 410-430 nm
indicated that little or no PLP remained bound as a
Schiff base (Kent, 33 31., 1958). Further evidence
for the presence of peptide bound PLP was found in the
fluorescence emission spectra of the recovered peptides
when excited at 280 nm. Churchich (1965) observed that
the emission of protein tryptophan residues is markedly
quenched by bound PLP through a radiationless energy
transfer process. The emission spectra of the peptides
of each model system and the respective blanks exhibited
this quenching effect. Typical emission spectra are
shown in Figure 8. These spectrophotometric results
confirmed the formation of nonreducible peptide bound
vitamin B6 complexes during thermal processing.
Quantitation of the peptide bound nonreducible
complexes at neutral pH (Table 6) demonstrated that they
comprised approximately 12% of the total model system
PLP, or about 60% of the total bound PLP. The degree
of binding was not significantly affected by the pre-
sence of glucose and/or ascorbate (P < .05), although

the variability of the data may have precluded the

 

88

fl PEPTIDES
/

 

 
   

PEPTIDES WITH
BOUND PLP

 

  

l l l l I

 

7O
60
8
c 50
03
C)
U)
93 40
O
3
LL
a, 30
.2
§
11 20
CE
IO
0
Figure 8.

 

300 350 400 450 500
WavelengtMnm)

Fluorescence emission spectra of peptides from
liquid model systems with and without PLP
fortification. Buffer = 0.1 M ammonium
acetate, pH 7.0. Excitation wavelength = 280 nm.

89

detection of any slight effects of these factors. Because
of the similarity of the UV-visible spectra of the non-
reducible complexes, these data could not be used to
differentiate between the acid labile substituted alda-
mine and the acid stable pyridoxylamino derivatives
(Dempsey and Christensen, 1962; Anderson, 3E 31., 1971;
O'Leary, 1971).

Table 6. Percent of model system PLP bound as a non-

reducible complex (substituted aldamine or
pyridoxylamino compound) at pH 7.0.

 

 

 

a Trial
Treatment
1 2
Control 12.3 16.1
Ascorbate 11.1 7.0
Glucose 10.7 17.1
Glucose & Ascorbate 17.5 2.9

 

aEffects of glucose, ascorbate, and their inter-
action were not significant (P < .05).

Direct measurement of pyridoxylamino complexes
was accomplished after peptide isolation under conditions
which would induce dissociation of the acid labile com-
plexes. Each model system was adjusted to pH 2.5 after
autoclaving, and the peptides were isolated by gel fil-
tration chromatography using 0.05 M KCl-HCl, pH 2.5, as
the eluent. In this acidic medium, the absorption maxi-

mum of the pyridoxylamino groups shifted from 323-325 nm

90

to 295 nm. Therefore, prior to quantitation, the pH of
the peptide fractions was adjusted to 7.0 with 0.4 M
KZHPO4. At pH 7.0, the UV-visible difference spectra

of peptides from PLP-fortified systems versus those from
the corresponding blanks were qualitatively identical to
that of Figure 7. These results confirmed that PLP was
bound to the peptides during the autoclave treatment as

a highly stable pyridoxylamino complex. Quantitation of
the pyridoxylamino groups (Table 7) indicated a mean value
of 9.7 i 6.4% of the PLP in the model systems was bound

in a state which may affect the bioavailability. These
results indicated that essentially all of the PLP isolated
as nonreducible derivatives at pH 7.0 was comprised of
acid stable pyridoxylamino complexes.

Table 7. Percent of PLP in liquid model systems which
was bound as a pyridoxylamino complex.

 

 

 

a Trial
Treatment
1 2
Control 15.9 19.7
Ascorbate 12.6 7.6
Glucose 10.4 2.4
Glucose & Ascorbate 1.1 8.0

 

aEffects of glucose, ascorbate, and their inter-
action were not significant (P < .05).

The presence of added glucose and/or ascorbic

acid had no significant effect on the extent of

91

pyridoxylamino complex formation (P < .05), although
variability in the data may have masked any slight
effects of these factors.

The data suggest that the presence of glucose
and/or ascorbate may partially retard the formation of
pyridoxylamino complexes (Table 7) through browning
reactions which may compete for free amino groups. As
a measure of the extent of browning beyond the initial
carbonyl-amine condensation complexes, free amino groups
were measured (Fields, 1972) in the peptides recovered
at pH 2.5. Mean values of 0.93 i 0.06 and 1.26 i 0.07
umoles of free amino groups per mg peptide were found
in the presence and absence of glucose, respectively,
indicating a 26% loss. The presence of ascorbic acid
showed no effect on the concentration of free amino
groups.

Because of the slight inhibitory effect observed
for ascorbic acid on the levels of total bound vitamin B6
(Table 5), and the previously reported involvement of
SH groups in a possible binding mechanism of PL or PLP
to proteins (Bernhart, g£_gl., 1960), the levels of free
SH groups in the various liquid model systems were
examined. The content of free SH groups in the peptides
of the heated model systems was determined directly on
the model system solutions, and on the peptides recovered

after Sephadex G-10 gel filtration. In this manner, the

92

reaction of the Ellman reagent with nonpeptide SH com-
pounds, such as HZS released from the peptides during
heat treatment, could be detected by comparison. The
results of Table 8 demonstrated that the presence of
ascorbic acid significantly increased the levels of pep-
tide free SH groups (P < .01), probably by reduction of
intramolecular disulfide bonds and hindrance of their
reoxidation. By comparison, the effect of glucose on
SH levels was weak, although significant in the case
of the isolated peptides (P < .01). The consistent dif-
ferences between direct analysis and analysis of the
isolated peptides indicated that a DTNB-reactive small
molecule, probably H28, was released from the peptides
during thermal processing.
Table 8. Sulfhydryl content of peptides in liquid model
systems after thermal processing. Analysis

either performed directly or after peptide iso-
lation by gel filtration chromatography.a

 

 

 

 

Direct Sephadex 6310
Treatment Analysis Isolation
1 2 1
Control 7.7 10.1 2.1 1.0
Ascorbate 27.5 31.5 8.6 10.2
Glucose 5.7 4.5 4.8 7.7
Glucose & Ascorbate 24.8 27.0 16.6 15.8

 

anmoles SH/mg peptides.

bEffect of ascorbate was significant (P < .01).

cEffects of glucose and ascorbate were signifi-
cant (P < .01). The interaction in either analysis was
not significant (P < .05).

93
Interactions of Pyridoxamine in a Liquid Model
iFood System durin§:Thermal Processing

Stability of Pyridoxamine

The concentration of the B6 vitamers in liquid
PM-fortified model systems after thermal processing 20
minutes at 121°C was performed by the HPLC method. Losses
of PM during thermal processing were quite small, except
in the case of the model systems containing glucose
(Table 9). The presence of glucose in the model systems
resulted in significantly lower levels of PM after the
heat treatment (P < .01), with an observed decrease of
24% of the PM added in fortification. No effect of the
presence of ascorbic acid was detected. Partial conver-
sion of PM to PL could be detected, accounting for 1 to
3% of the added PM in the model systems containing glu-
cose and/or ascorbate. No conversion of PM to PN was
observed. The results of this study indicated that the
losses of total vitamin B6 in the PM-fortified liquid
model systems were generally low, with the maximum being
about 21%.

Pyridoxamine--Carbohydrate Complex
Formation

 

Gel filtration chromatography on Sephadex G-10
revealed no detectable changes in the molecular size of
the fractions containing PM in the PM-fortified model

systems, as indicated by a constant elution volume

94

Table 9. Vitamin B6 in PM-fortified liquid model systems
after thermal processing. Fortification prior
to processing was at a level of 50 ug/ml.al

 

 

 

 

U9 BG/ml
Treatment Perggggage
PM PL Total
Control 48.5 ND 48.5 3.0
Ascorbate 45.6 0.6 46.2 7.6
Glucose 38.0 1.5 39.5 21.0
Glucose & Ascorbate 41.8 1.6 43.3 13.2
aMeans of duplicate determinations. ND = not

detectable.

bEffect of glucose on lowering PM and increasing
PL was significant (P < .01). Ascorbate significantly
increased PL levels, but did not affect the concentration
of PM (P < .05). The glucose x ascorbate interactions
were not significant (P < .05).

95

(Figure 9). Thus, complexes containing PM were not
observed by this procedure.
Stability and Bioavailability of Vitamin B5 in

Dehydfafea'ModeI—Food'systems:
Roasting Effects

 

 

 

 

Compgrison of Assay Methods
for Total Vitamin 85

 

 

The assay dehydrated roasted and nonroasted model
systems provided the first application of the HPLC method
for determination of vitamin B6 in complex systems con-
taining low levels of the vitamin. The chromatographic
procedure, based on the methods of Horvath, 23 31. (1977),
involved the separation of the 86 vitamers by reversible
"solvophobic" interaction with an octadecylsilica sta-
tionary phase. A chromatogram of the vitamin B6 standards
is shown in Figure 10. The theoretical plate counts for
PM, PL, and PN were approximately 6850, 3850, and 3000,
respectively. The linearity of response with the 280 nm
absorbance detector, reproducibility of the retention
times, and high sensitivity of the method is demonstrated
by the calibration data of Table 10. Typical recovery
values for the B6 vitamers added to the dehydrated model
systems prior to extraction were 75% for PM and PLP and
95% for PN.

The HPLC assay of vitamin B in the nonroasted

6
model systems demonstrated excellent precision and cor-

related well with the expected concentrations based on

96

o o .o.h mm .oumuoom Eowcoeem z H.o
u Hommom .zm was h.H u >\ > an xmom one .UoHNH um awe ow nouns
mEoumhm H0005 oflavwa wowufluHOMIZm mo mawumouoeouno calm xoomnmom .m ousmam

 

 

 

 

o>
\o>
0.» ON o.~ m... 0..
22.33
+ I
3820
3836/1111 <

’
b

 

 

33.534

.228 <

(muegapauoqlosqv v

 

4O

RECORDER RESPONSE

Figure

 

 

10.

97

 

 

 

 

 

 

 

 

o I 2 3 4 5 6
TIME (min)

HPLC chromatogram.of standard vitamin B
free bases. Detection =_280 nm absorption;
flow = 2.0 ml/min: 1700.psig: injection =
50 pl containing 0.100 mg of each vitamer
in 0.2 M potassium.acetate in 5% TCA, as
performed with model system extracts.

98

Table 10. Calibration curves for the HPLC assay of vita-
min B6 using 280 nm absorption detection.
Flow = 2.0 ml/min, 1700 psig, 0.02 AUFS.

 

 

 

 

 

 

 

Pk ht = peak height; tr = retention time.
PM PL PN
ug/SO ul tr pk ht tr pk ht tr pk ht
(min) (mm) (min) (mm) (mm) (mm)
0.125 2.85 223 4.35 196 5.20 116
0.100 2.95 181 4.45 159 5.30 93
0.075 2.90 144 4.35 125 5.20 73
0.050 2.90 103 4.40 88 5.25 54
0.025 2.85 60 4.35 51 5.20 33
0.0125 2.90 29 4.45 22 5.30 15
Mean 2.89 4.39 5.24
Std. Dev. 0.04 0.05 0.05
Slope (mm/ug/Soul) 1683 i 49 1505 i 45 861 i 31
Y-intercept (mm) 14.7 i 3.7 9.5 i 3.4 8.4 i 2.3
Correlation
Coefficient 0.9983 0.9982 0.9975

 

99

the 25% RDA per ounce (15 09/9) fortification (Table 11).
Attempts to determine the vitamin B6 in the roasted model
systems were unsuccessful because the many UV-absorbing
peaks from soluble browning compounds obscured the chrOma-
tographic peaks of the B6 vitamers. Comparison of the
HPLC and microbiological results for vitamin B6 in the
nonroasted model systems (Table 11) exhibited excellent
1correlation, with the exception of the markedly lower
microbiological value observed for the PM-fortified model
system. The low activity of PM for the assay organism,
S. uvarum (Rabinowitz and Snell, 1948; Parrish, 33 31.,
1955; Gregory, 1959; Woodring and Storvick, 1960; Chin,
1975), was presumably responsible for this discrepancy.

The data from the semiautomated fluorometric
determination of vitamin 36 are presented in Table 12.
Inspection of these fluorometric data for the individual
vitamers indicated reasonably good agreement with the
HPLC results for PN and PL in the nonroasted model sys-
tems (Table 11). By contrast, semiautomated fluorometric
values for PM in the nonroasted systems (Table 12) were
much higher than those determined by HPLC (Table 11). A
comparison of the data for the nonroasted and roasted
model systems indicated that the roasting process degraded
all forms of the vitamin (Table 12). The observed

decrease in PM and corresponding increase in PL in the

100

Table 11. Total vitamin B5 in control and roasted model
systems determined by microbiological and
HPLC methods.arbrc

 

 

 

 

Control Roasted
Fortification
HPLC Micro. Micro.
Nonfortified <:l.0 1.1 i 0.3 1.0 i 0.4
PN 18.7 i 1.5 19.6 i 3.6 9.3 1 0.8
PM 19.8 t 0.1 9.7 i 2.4 6.2 i 1.6
PLP 14.4 i 0.1 13.6 i 0.8 5.1 i 0.5

 

aug total B6/g model system.

bMean and standard deviations. Duplicate deter-
minations.

CThe HPLC assay was performed with 280 nm absorp-
tion detection. This could not be applied to the roasted
model systems because of the extensive browning of the
sample and resulting complexity of the chromatograms.

101

.mGOHumcHEHmuoo oumoHHmHHu

.cowumw>oc oumosmum com com:

a

.Eoummm Hoooa m\mnm

 

 

 

 

~.H H o.HH v.o H m.m H.H H m.H ¢.o H h.m mam
h.a H m.oa ~.a H m.m m.o H H.N ~.H H m.m Sm
v.H H v.va N.H H m.h m.o H H.m v.o H v.v zm
m.H H ~.m o.o H N.H v.H H m.~ H.H H N.H ooflmwuuomcoz
noummom

n.~ H m.mm m.o H a.mH m.~ H H.m v.0 H H.m mam
h.v H «.mm m.o H w.a ~.v H m.vm w.H H h.m 2m
m.v H m.vm m.H H o.m m.v H m.mH H.H H m.mH zm
m.m H o.ma h.o H m.~ o.m H H.mH m.H H v.H ooamwuuomcoz
a

Hmuoe Am 2m 2m :OHHMOHHHuHom

n.m.ouopmooum momma owuuofionosam ooumfiousmwfiom on» an

omcHEHmuoo mm .mfioummm HmGOE voummou can Houucoo 2H mm GHEMHH> .NH manna

102

PM-fortified model system suggested the occurrence of a
thermally induced transamination reaction.
Comparison of semiautomated fluorometric and

microbiological assay data for total B in nonroasted and

6
roasted model systems (Table 13) indicated a similar per-
cent loss by both assay methods. Observed losses during
the roasting process, as assessed by either assay method,
ranged from 53 to 73%, except in the case of the micro-
biological analysis of the PM-fortified system. As
previously mentioned, the possible inaccuracy of the
microbiological assay for PM may have been responsible
for the observed difference. The apparent total vitamin
B6 contents of all roasted and nonroasted model systems
were consistently higher when assayed by semiautomated
fluorometric, rather than microbiological, methods. The
nonroasted and roasted model food systems fortified with
PN and PLP and the roasted PM-fortified systems exhibited
a ratio of fluorometric to microbiological values for
total B6 ranging from 1.65 to 2.16. The ratio rose to
3.63 in the case of the nonroasted PM-fortified system.
The poorest correlation between semiautomated fluoro-
metric and microbiological assays was observed in the
case of the nonfortified model systems, in which the
observed fluorometric to microbiological ratios were

17.3 and 5.2 for nonroasted and roasted systems, respec-

tively. The mean difference between semiautomated

103

.mmmmmm Hmowmoaoanouowfi How mGOHHMGHEHouoU
oumoflamoo can owuuosouosHm How mooHuocfleuouoo oumowamwnu mo amaze

.Eoummm Hmooe m\mm Hmuou mam

 

 

mo H.m w.ma hm o.HH m.m~ mam
mm ~.m h.m on m.cH «.mm 2m
mm m.m m.ma mm v.va m.vm zm
III o.H H.H ms N.m o.m~ onMHuHomcoz
mmoq w coummom Houucou mmoq m woummom Houucou COHHMOHMHuHom

 

 

HMOHmoHownouowz owuuoeouooam Eoumhm Hoooz

 

n.m.msoum>m
HoooE woummou can Houucoo GA on cHEmuH> Hmuou mo coaumsweuouoo
How moonuofi HmowmoHOHQOHOHfi can UHHuoEouoon mo cOmHHmmEOU .mH wanna

104

fluorometric and microbiological data for all nonroasted
model systems was 17.5 i 5.9 pg/g, while the mean dif-
ference for roasted systems was 4.9 t 0.8 ug/g.

Interference intSemiautomated Fluoro—
metric Vitamin B5 Assay

 

 

Thin layer chromatographic studies performed on
the products of the reactions employed to form 4-pyridoxic
acid lactone in the semiautomated fluorometric vitamin
B6 assay (Gregory and Kirk, 1977a: Appendix) indicated
that the preparative ion exchange chromatography was
effective in removing potentially interfering compounds.
Since the poor correlation of data from semiautomated
fluorometric assays with microbiological and HPLC values
suggested the presence of interfering fluorophores,
further studies were undertaken to identify these inter-
fering compounds. The fluorescence emission spectra of
sample extracts treated for fluorometric analysis, and
their assay blanks, were examined in an effort to detect
these interfering compounds. All samples were adjusted
to pH 10.2 with 0.4 M sodium carbonate, 0.2% Brij 35,
prior to spectral analysis. The emission spectrum of
the fluorophore, 4-pyridoxic acid lactone, was calculated
manually by subtraction of the emission spectra of the
assay blanks from those of chemically treated samples.
All samples and blanks were prepared in a manner identi-

cal to that followed in the semiautomated fluorometric

105

assay. The excitation wavelength was 355 nm. The fluoro-
phore emission spectra determined for all model system
extracts were found to be highly similar to that of 4-

pyridoxic acid lactone formed from B standards (Table 14).

6
No significant differences were observed in the emission

maxima, and differences in peak width (full width at half
maximum, FWHM) were not considered to be significant,

in view of the imprecision in the determination of these

spectra. Reagent blank fluorescence was observed in the

case of the PL and PM determinations.

Biologically Available Vitamin B5 in
R53§Ee6“UefiyarEEed Model System§"__

 

The results of the rat bioassay (Bioassay l) for
biologically available vitamin B6 in the roasted model
systems are presented in Table 15. The responses of rats
fed the standard diets containing 0 to 1.0 ug PN/g were
used to generate the dose-response curves used in calcu-
lation of available vitamin B6 in the test diets con-
taining the roasted model systems. The dose-response
curves of best linear fit were determined, and their
regression parameters shown in Table 16.

The estimated concentration of biologically
available vitamin B6 in each of the roasted model systems
is shown in Table 17. Comparison of the data based on
each of the four response parameters indicated no sig-

nificant differences among the estimates for each model

106

 

 

 

 

 

 

 

Table 14. Characteristics of the emission spectra of the
fluorophores produced from vitamin B6 standards
and model system extracts during the fluoro-
metric assay of vitamin B5. Excitation wave-
length = 355 nm. Wavelength values are approx-
imately t 5 nm for maxima and i 10 nm for full
width at half maximum (FWHM).

. . A maxb FWHM

Vitamer S l a Relative

Assayed amp e Fluorescence

(nm) (nm)

Pyridoxamine

PM standard 74 430 78
Nonfortified-C 18 430 (450 shld) 95
Nonfortified-R 13 430 (450 shld) 95
PM-C 73 430 71
PM-R 38 430 73
Reagent blank 3 450 85
Pyridoxal
PL standard 65 430 75
Nonfortified-C 14 430 77
Nonfortified-R 12 430 78
PLP-C 80 430 66
PLP-R 43 430 68
Reagent blank 8 425 105
Pyridoxine
PM standard 76 430 68
PN-C 50 430 71
PN-R 9 430 60
Reagent blank 0 --- --

 

aSample designation:

and roasted model systems.

b

C and R refer to control

Shld refers to the approximate maximum of an
emission peak shoulder.

.cOHuonoo Moosuw o mcH3OHHom coaumm
umou amolam on» How ouoz oosdmcoo comm m\£uzoum can .GOHumEdmcoo comm .nusouwn

.msoum humumwv Mom mumu m .Honnm pumpsmum pom Goose

 

107

 

 

 

 

 

 

 

 

 

~.v H H.vm ~m.o H am.> «Ho.o H Hmm.o m H mmm m H om mam
¢.m H m.nm Hm.H H m¢.> mmo.o H mom.o m H avm n H me So
v.¢ H o.om no.H H om.m moo.o H mmm.o m H mvm v H «m zm
m.m H «.mm Hm.o H m¢.o bHo.o H omm.o «a H HHN w H Hm omHMHuHOMGOZ
msouwxw H0602 omummom «m
m.m H m.m~ oo.a H 5H.¢H oHo.o H mov.o ea H hem m H owa : oo.H
m.~ H m.m¢ mm.o H h~.m moo.o H Hov.o m H mmm m H wma = m>.o
m.m H o.mm ~m.H H mm.m Nao.o H hwm.o ca H Hon m H NoH a om.o
m.m H v.mm mh.o H mH.n mHo.o H wvm.o v H nmm w H mu a mN.o
h.m H o.ao mm.o H oo.h moo.o H mvm.o m H and m H vv m\zm m: oo.o
oumocmum

1». .nm msxosv .m\m. .mc .mc

sowumaaeflum mam >uw>wuo¢ mmEdmsOD moon m ooESmcou nusouw Howe

Bdmmc ouhoousuhnm .m. nuzouw pooh Dom

..H mmmmmon. mEoummm Hoooe woummou mo homomoHn pom .mH manna

5.8

108

.msam> comm on H «o cowuwoom osu an @0000 mums mcowumuucoocoo on» .soxmu mos 00H 0:»

coca .mcwxoowumm cocoa mo GOHumnucoosoo mnmuowo on» mucomoummu pump x Had

a

.coHHouHHo oncommon comm How mucowowmmooo sowumaouuoo
Hmonmfln on» oo©H>oum cowcs omonu ..o.H .mwnmcowumHoH Hmocfla m usonm wuHHHQMHHm>
ummma on» pooH>oum cows3 mGOHuocdm Hammoumou mo>uso oncommoulomoo omonam

 

 

oomp.1 m.m H H.mm1 m.m H m.ow coHumHssHum mam 94mma .zm.
mhsm.+ om.H H Hm.m mm.o H mH.o HHH>HHo< scone 12m.
mmmm.+ mmo.o H omm.o NHo.o H ou~.o ummm m\.m. nusouo 1H + zmc moH
mmwm.+ mH H Hmm H.m H m.HH Am. susoHu 1H + zmc moH
HcmHoHHHmoo
GOHHMHMHHOU mmoam umoououcwlw w ox

 

.H homomon m

.mo>Hoo oncommoulmmoo on» no

.mfioumhm Hooos woummon
muouoEmuwm :onmonoH Hmocwq .oa manna

109

..mo. v m. mEoummm H0008 mnu mo comm How ucmHoMMHv mausmOHuHcmHm no:
mums mHHmHHHo oncommou on» mo comm an on GHEmHH> oHQmHHm>m mo mmumsflummo

.Emumhm Hoooe on» cw mm cHEmuH> oHQmHHm>m on» no cowumaooamo
CH moms moumum HmGOHuHHqu mm mo xomcH may ou mnowou cowumuHHo mucommmmn

.mooum mmmmm Mom mumu m .Houum vumocmum pom cmozm

 

 

 

 

 
 

o.~ H m.¢ m.a H m.¢ w.H H m.m m.o H m.h mam
m.~ H H.m n.~ H m.a m.m H >.v N.H H ~.m 2m
v.~ H w.m m.m H m.~ m.o H o.n >.o H m.m 2m
N.m H m.o m.H H m.o ~.H H ¢.~| m.o H ¢.H mmHMHuHomcoz
coHumHssHHm mam HHH>HHoH mssmcoo 00m m
. . nusouu susouo
Hmomm mumoounumum cowumOHMHuuom

 

Emumam Homo:

o.oc0HHoHHHU oncommom

 

.Emumwm Honofi m\wm oHanHm>m m: Doom
Ioummu mumo m.mEmum>m Hmpoe ooummou :H mm GHEmuH> oaomawm>m waamowmoHon .na magma

110

system (P < .05). Inspection of the standard errors of
each estimate indicated that rat growth generally pro-
vided the greatest precision. This was also suggested
by the high correlation coefficient of the rat growth
dose-response curve.

Statistical comparison of the microbiological

assay data for total vitamin B with each of the rat

6
bioassay estimates (Table 18) indicated no significant
differences (P < .05). The microbiological data cor-
related well with the bioassay estimates based on growth,
growth per gram feed consumed, and PLP percent stimulation
of AspAT. The bioassay estimates based on AspAT activity
exhibited a slightly poorer correlation with the micro-
biological data. In strong contrast, the semiautomated
fluorometric assay results were consistently higher than
all bioassay estimates, with a mean difference of 6.3 i
1.9 ug vitamin B6/g.

Stability and Bioavailability of Vitamin B in
DehydratedModeI"§ystems: Storage Effects

 

Moisture Content and Water
Activity

The dehydrated model systems were equilibrated

 

to the predetermined water activity (aw) under forced air
which had been conditioned to 65% relative humidity.
The precise moisture equilibrium for each fortified model
system stored in TDT and 303 cans was determined by

measuring the aw and moisture content of the systems

111

. .3. v m.
mmmmmOHn umH pom HmonOHOHQOHOHE amen Hoummnm haucmoHMHGmHm mums mumm UHHuoEouoon
“mmumEHumm hmmmmoHo umu com HmonOHoHQOHoHE somsumn wmocmummmwo unmonwcmHm ozo

.Eoumam Hoooe m\mm Hmuou mam

 

o.HH H.m m.¢ m.v m.m m.h mam
m.oa N.m H.m m.H b.v N.m 2m
«.va m.m o.m m.~ o.b m.w zm
N.m o.H m.o m.o v.NI ¢.H omwmwuuomcoz

 

COflUMHDEHHm mam huflkrflfinvdm wwgmcou ”mm.“ m

.m. cusouo

 

 

su3ouu cowumowmwuuom

.Eouooam .HownouoHZ Bdmmc Eoummm Homo:

 

GOHHGHHHU oncommom

 

n.m.mm sHEmuH> Hmuou How muaomou hmmmm OHHHQE
Ionooam omumfiousmHfimm pom HmonOHOHQOHOHE nuH3 mEoummm HmooE mmummou cw
mm cflemuw> maanHm>m madmonoHOHn mo mmumeaumo wmmmmown umu mo cOmHHmQEoo .ma magma

112

after 128 days at 37°C (Table 19). No significant dif-
ferences in moisture content or aw were observed between
fortified model systems after storage in either container.
Therefore, the headspace volume did not affect the moisture
equilibria. Slight differences in the moisture data were
observed for the model systems fortified with the dif-
ferent 86 vitamers. These differences were not statisti-
cally significant (P < .05). The analysis of the water
activity data revealed several significant differences
among the various model systems. The aw of the non-
fortified and PM-fortified systems was found to be sig-

nificantly less than that of the PN- and PLP-fortified

systems (P < .05).

Nonenzymatic Browning

 

Browning in the model systems was determined by
measuring the levels of free amino groups and browning
pigments in the soluble protein fraction. Plots of the
time-dependent loss of amino groups (Figure 11) and the
concomitant formation of melanoidins (Figure 12) demon-
strate that nonenzymatic browning is not a linear func-
tion, indicative of the complex kinetics of the reaction.
The data in Table 20 show the concentration of free amino
groups and browning pigments associated with the soluble
protein fraction of the dehydrated model systems stored
in TDT and 303 cans for 128 days. These data demonstrate

a large difference in the concentration of both amino

113

Table 19. Moisture content and water activity of stored
model systems. Model systems analyzed after
storage for 128 days.a' ,c

 

 

 

 

Model System TDT Cans 303 Cans
Fortification Moisture aw Moisture aw
(%) (%)
Nonfortified 11.12 0.584 10.88 0.582
PN 12.08 0.621 11.37 0.617
PM 11.48 0.571 11.19 0.589
PLP 11.53 0.626 --- ---

 

aAll moisture and water activity determinations
were performed in duplicate. Data represent the mean.
PLP-fortified system stored in 303 cans not analyzed due
to insufficient supply. Moisture expressed on a dry
weight basis.

bNo significant differences between the moisture
content or water activity of equivalent model systems
whether stored in TDT or 303 cans (P < .05).

cMoisture contents not significantly different
among model systems of different fortification. Water
activity of nonfortified and PM-fortified systems was
significantly less than that of PN- and PLP-fortified
systems (P < .05).

114

.m.m mm .Homuaom
z mo.o nuHs oouomuuxm mumz mcHououm .Uobm um ammo Boa cH mmmuoum mawuso
mEoumMm H0005 omumuvmaoo mo mcHououm manoaom :H monoum ocHEm moan mo mmoq .HH ouomwm

seems;
02 0m. 00. 0m 00 ov ON 0

q q q q 4 d

an nu
0’ o 1

 

 

 

ugason 5w / 3 HN - salon: 7’

.Nw. Zn.
0N0. a1:
:10. in.
V00. 0:02

 

: 8:02:13...

115

OS

.oonm um memo Boa GH mswummm
Homofi omumuohnoo mo ommuoum mofluou musmemflm dHoHOGmHmE mo cowumfinom

3322.:

ON. 09 om om ov ow

 

q u a d q q

   

a
mum. and o
in. 2n. 0
won. 282 o

3o 8:02:15“.

 

.NH muomflh
o
H 00.
.1 V
1 b
70
m
.1Nfiu m
II
1 w
AU
M
15. mw
w.
1 “U
II
1 M
00.

116

Table 20. Comparison of browning levels in model systems
stored in TDT and 303 cans for 128 days at 37°C.

 

Model System

Free Amino Groupsa

 

Fortification

Browning Pigmentsb

 

 

 

TDT 303 TDT 303

Nonfortified 0.108 0.111 0.0422 0.0082
PN 0.066 0.103 0.0523 0.0114
PM 0.088 0.117 0.0331 0.0071
PLP 0.075 0.099 0.0527 0.0095

aMean of triplicate determinations. Data = umoles/
mg protein.

bMean of triplicate determinations. Data = A420

nm[mg protein/ml.

117

groups and browning pigments, indicating that the rate
of browning was much greater in the model systems stored
in TDT cans. The data indicate that the model systems
stored in TDT cans browned 4 to 6 times faster than
identical systems packaged in 303 cans. A comparison
of the data for the browning pigment levels and amino
group concentration for each of the fortified model sys-
tems indicated that the difference in free amino groups
between the TDT and 303 can systems was much less than
the difference in pigment formation.

Within model systems of each storage container
(TDT or 303 metal can), the loss of free amino groups
and formation of pigments was strongly a function of
the water activity of the system. The data suggest that
maximum browning would occur at an aw greater than 0.62
in this model system. Numerous investigators have
reported similar values for the dependence of browning
on water activity (Lea and Hannan, 1949: Loncin, gE_§l.,
1968; Warmbier, gt 31., 1976).

Kinetics of Vitamin B6
Dagrsaariafi :’

 

 

The time-dependent loss of the 36 vitamers used
in fortification of the dehydrated model systems was
examined to determine the degradation kinetics during
the relatively severe storage conditions of 37°C at an

aw of about 0.6. In order to determine whether the

118

headspace atmosphere of the storage container affected
the rate of degradation, the model systems stored in
303 cans were assayed at the end of the 128-day storage
period and compared with the TDT can data.

All vitamin B6 analyses in this study were per-
formed by the HPLC method. Initially, all assays were
performed using absorption detection at 280 nm. During
the later stages (after 78 days) of the storage period,
the extensive browning of the model systems precluded
accurate quantitation by this procedure. Final assays
were made using HPLC separation and fluorescence detec-
tion. The HPLC fluorescence calibration curves were
prepared as previously described, and were found to be
linear over the entire range employed. The specificity
of the fluorescence detection method permitted quanti-
tation of the vitamin B6 in all dehydrated model system
samples (Figure 13).

The degradation of all B6 vitamers could be

described by the first order equation:

where k represents the first order rate constant, C rep-

resents reactant concentration, and - 3% is the rate of

change of reactant concentration (Moore, 1962).
Pyridoxal phosphate was found to degrade at a

high rate during storage of the dehydrated model system

119

.mocoommuooam Ho momma
IHOmnm Ed omm n cowuooumo .oohm um cmo Boa m GH mmmo pm How monoum
Emummm Hmpofi ooHMHuH0mcos m mo mamfimm muo>ooou m mo memumoumfiouno qum

.mH muomflm

120

Nd—)%

j

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

p l 1
O O
. H 8 <2 0
GOUSOSGJODIJ among 5
Mel-9:: .
D
187’ 3' ‘
wde‘ j 7
q
l 1 l l l
8 <0 . 8 8
o. 8 8 0. 0.

(wuogz) aouoqlosqv

Figure 13

Time(min.)

121

in a TDT can at 37°C, with the process following first
order kinetics (Figure 14). No conversion of PLP to
PM (PMP) was detected.

The greatest complexity in vitamin B6 degradation
was observed in the PM-fortified system (Figure 15). The
disappearance of PM was rapid and could be described by
first order kinetics. The coincident rapid increase in
PL concentration during the first several weeks of
storage was evidence for a transamination process. No
evidence for a reverse reaction could be obtained.
Therefore, under these conditions, the conversion of
PM to PL was unidirectional. The PL curve (Figure 15),
showing a maximum PL concentration after approximately
35 days of storage, suggests that the rate of PL for-
mation exceeded that of PL degradation. Therefore, the

rate limiting step in the loss of vitamin B in the PM-

6
fortified model system was the degradation of PL.
Accurate determination of the rate of conversion of

PM to PL could not be made because of the limited number
of observations during the phase of rapid conversion.

As a result, the rate of degradation of PM directly to
inactive products could not be determined. The behavior

of the vitamin in the PM-fortified system is most readily

described by the following model:

‘ PL .; Products

 

PM
l,k2 k
Products

122

PL P - Fortified System

30f

 

 

 

IO
8
{a 7
a, 6
1 5
£1
3
2
0 20 40 60 80 IOO I20 I20
Time(Doys)
Figure 14. Degradation of PLP in dehydrated model

systems stored in TDT cans at 37°C.

123

30 "

20 - PM-Fortified System
A-PM
OBPL

71-9/9
01 A 0|me 0

 

 

 

 

0 20 40 so 80 IOO I20 I40
Time(Days)

Figure 15. Behavior of vitamin B5 in a PM-fortified dehy-
drated model system stored in TDT cans at 37°C.

124

During storage of the model system at 37°C at an aw of
about 0.6, k_1 was much smaller than k1. Thus the rate
of loss of PM from the model system was represented by
the sum of the rates of degradation (k2) and conversion

to PL (k1). Calculation of k1 or k cannot be performed

2
from these data.

Of the three vitamers used in the fortification
studies, PN was the most stable. No loss was observed
for the first 58 days of storage, followed by a very
slow degradation which followed first order kinetics
(Figure 16). The rate of loss of PL inherent in the
ingredients of the PN-fortified and nonfortified model
systems (Figure 16) was found to be much more rapid.
During storage of the PN-fortified model system, no
formation of PL or PM could be detected, indicating that
the losses were not due to conversion to the more labile
vitamers.

Linear regression analysis of the PL, PM, PLP,
and PN degradation data confirmed that all of the
observed reactions were compatible with description by
first order kinetics. All first order plots of the data
exhibited correlation coefficients greater than -0.93
(Table 21). Zero and second order rate functions did
not provide better fit for these data. The observed rate
constants for the loss of PL and PLP were similar (0.0155

and 0.0158 day-1, respectively), while that for the net

125

A=PN in PN- Fortified
20 _ System

 

: Inherent PL:

O'Nonfortified System
0: P N - Fortified

-' o

FQIQ
u 4: 01034100 0

 

 

O 20 4O 60 80 IOO IZO |4O

 

Time(Doys)

Figure 16. Degradation of added PN and inherent PL in
dehydrated model Systems stored in TDT cans

at 37°C.

126

.wMHH mammn

.muon OHEnuHHmmOHHEom mo mamaamcm sonmoumoH HmmcHH
an mocHEHouoo .cOHumH>oo oumocmum H ucmumcoo mumu “mono Hmuflmm

 

 

 

 

 

 

 

we humm.| smoo. H mwao. AA + 2m Hmuou mo mmoq

«v hwwm.| ovoo. H mmao. Am moEH0m mo moon

mm mmwm.t mmoo. H come. So no mmoq
mcflfimxomwuwm

vw nmmm.n maoo. H mmao. no mo mmoq
oumnmmonm meOUHHNM

HHH mmmm.1 mooo. H mqoo. .msme mm Ac zm Ho mmoq

we mmmm.l omoo. H Hmao. Am ucouoscw mo mmoq
ocwxoowwNm

av mumm.u nmoo. H onao. Am Hammoncw mo mmoq
omHmwuH0mcoz
.mamov ucowowwmoou .lehmov cowuomom cowumoHMHuHom
IIIIII. COHumHoHHou Emumwm Homo:

QN\H u mx

 

luom mammmm mm GHEmuH>

.muomoooum 04mm on» he oMEHOM

.Uohm um memo Boa cw ommuoum mcHumo mEoumhm

Homoe woumuownoo Eoum mHoEmuH> mm onu mo mmoa on“ How mumm oaumcwx .Hm magma

127

1

disappearance of PM was slightly higher (0.0200 day- ).

The rate constant for the first order degradation of PN
was observed to be 0.0049 day-1.

The analysis of the vitamin 86 contents of the
model systems packaged in 303 cans and stored for 128
days at 37°C revealed significantly greater stability
of the vitamin in these containers than in the TDT cans
(Table 22). The detection of PM after 128 days storage
in the 303 cans was indicative that the rate of con-
version of PM to PL was slower in storage containers
with the large headspace.

Determination of 4-Pyridoxic Acid in
DehydratediModel Systemsafter

Storage
The slow oxidation of PL and PLP to 4-pyridoxic

acid and its 5'-phosphate has been observed in various
liquid nonenzymatic systems (Ikawa and Snell, 1954;
Hamilton and Revesv, 1966; Rotilio, £3 31., 1970). The
formation of 4-PA during the storage of the dehydrated
model systems at 37°C was monitored in these studies.
After 128 days of storage, the model systems stored in
TDT cans were analyzed by HPLC for the presence of 4-PA.
4-Pyridoxic acid was detected only in the PLP-fortified
model system. Only 2.6% of the initially added PLP was
found as 4-PA (or its 5'-phosphate). These results
indicated that 4-PA was a minor degradation product

of PLP, since 88% of the PLP added to the dehydrated

128

Table 22. Vitamin 36 content in dehydrated model systems
after storage for 128 days at 37°C in TDT and
303 cans. HPLC analysis, fluorescence detec-

 

 

 

tion.a'
Model System TDT Cans 303 Cans
F°rt1flcat1°n PN PM PL Total PN PM PL Total
Nonfortified ND ND tr <0.5 ND ND tr <0.5
PN 10.8 ND tr 10.8 14.2 ND tr 14.2
PM ND ND 2.7 2.7 ND 4.3 4.7 9.0
PLP ND ND 2.7 2.7 ND ND 4.8 4.8

 

aug/g model system. Means of duplicate determi-
nations.

bND = not detectable; tr = trace, too low to
quantitate. Trace data not used in calculation of

total 36'

129

model system was lost during the 128-day storage period.
The presence of 4-PA as its lactone could not be deter-
mined by the HPLC assay procedure because this compound
does not fluoresce in the acid environment of the HPLC
mobile phase.

Determination of Pyridoxyllysine in
Dehydrated Model Systems after

Storage

In order to determine the biologically available
vitamin B6 in the model systems by the rat bioassay,
the systems in the TDT and 303 cans were pooled after
storage for 128 days at 37°C. As previously discussed,
researchers have suggested that vitamin B6 derivatives,
such as certain pyridoxylamino compounds, may possess
only partial B6 activity or may act as antivitamins
(Snell and Rabinowitz, 1948; Heyl, 22‘31., 1952; McCor-
mick and Snell, 1961). In order to evaluate the effect
of pyridoxylamino compound formation on the biological
availability of vitamin B6, model system protein hydroly-
sates were quantitatively examined for e-pyridoxyllysine.
The hydrolysates were analyzed solely for e-pyridoxylly-
sine since numerous researchers have observed that essen-
tially all of the PLP bound to proteins by reductive
coupling was attached to the e-amino group of lysine
(Dempsey and Christensen, 1962; Dempsey and Snell, 1963;
Anderson, st 31., 1966; Page, 33 31., 1968; Ronchi, 3E

31., 1969: Anderson, gt_gl., 1971).

130

The synthesized pyridoxyllysine was examined for
purity by several methods. The UV—visible absorption
spectrum of pyridoxyllysine in 0.1 M ammonium acetate,
pH 7.0, demonstrated the characteristic maxima at 323-

325 and 253-254 nm (Figure 17). Thin layer chromatography
on cellulose MN-300 gave a single ninhydrin reactive spot
which exhibited a strong blue fluorescence when irradiated
with ultraviolet light at 254 or 366 nm. Analysis by

HPLC demonstrated a single major peak and two minor con-
taminants (Figure 18). The crystalline pyridoxyllysine
was used as a standard without further purification.

In the analysis of the hydrolysates of the pro-
teins from the pooled TDT and 303 can model systems, only
the PLP-fortified system exhibited detectable levels of
pyridoxyllysine (Figure 18). The identity of the pyri-
doxyllysine chromatographic peak from the PLP-fortified
model system was confirmed by comparison of the exci-
tation and emission spectra of the collected peak with
those of the prepared pyridoxyllysine standard (Figure 19).
Although the collected fraction from the model system
exhibited a shoulder on the excitation peak at approxi-
mately 340 nm, probably the result of incomplete reso-
lution on the chromatogram, the basic spectral features
confirmed the identity of the peak as pyridoxyllysine.
Quantitation of the chromatogram revealed that 4.17 mg

PL/g of the PLP-fortified model system participated in

Hommsm

oocoHoMMHo k........ ochmaamxomHuao mo fisuuooom coHumHomn<

..

.o.h mm .oumuoom EpchEEm z H.o ems muuoomm anon mo

. «mm HE\mE H momuo> amnion HE\mE H mo asuuoomm

 

131

.HH muanm
3550:2263
new on» own 00» ohm Com mum
1 - ion. a n J u 8.
00000 00000000 I 8.
1 o_. W.
a .. u c.
.. .. u o
O O O J
00. 000 o q
.000 D
u
1 9. w
1 ON .

 

L

0N.

132

8533108823 1 23.1 . 2: e3»? 388 eoHHHHHomumqm Ho

 

 

 

 

mummmaouows sHououm mam .dv Unmocmum.ocwmaaahxooflumm mo memumoumeouno qum .ma mummwm
1.55.5:
H o m w m N _ o ‘n c m m.
j 10

10...“...
m < m.
H...
.n
1oN a
H
._, n
H m
o—AI. ion a
.. a
nu
nu
a .9. m

.d

_.1

u .8

 

133

 

.ommsm oHHnoE

one: ~.~ mm H Hommom .......... mummhaouomn cHououm Seaman
Hoooe moHMHHHOMImqm Eoum xmom owsmmumoumEouno onwmwaahxooflumm
mouooaaoo com ¢IIIIIII. mummcmum ocwmhaaaxomwuhm onwaamumhuo
mo muuoomm ..Em. conmHEo mam ..xm. cOHumuHOxm oocmomouooam

.ma ousmflm

134

 

 

 

 

 

1
O O O O 0
IO ¢ '0 N '—

70"
60 '

eoueoselonl g emotea

 

 

250 300 350 400 450 500
Wavelength (nm)

ZOO

Figure 19

135

the formation of pyridoxyllysine (Table 24). Therefore,
the reductive coupling of PLP to protein amino groups
represented a significant mechanism for the loss of bio-
logically available PLP from the model systems during
storage at 37°C at a water activity of about 0.6.

Determination of Total Vitamin B5 in
ggfiydrafed MbdeI Systems affef"—“_"

Storage
After pooling the model systems stored for 128

days in TDT and 303 cans at 37°C, total vitamin B was

6
determined by microbiological, semiautomated fluorometric,
and HPLC methods. The vitamin B6 data obtained by micro-
biological (Table 25) and HPLC (Table 26) methods were in
good agreement. This was not true for the results of the
semiautomated fluorometric assay (Table 27) which did

not correlate well with either HPLC or microbiological
data. As observed in the study concerning the effects

of roasting on vitamin 36’ the semiautomated fluorometric
values were consistently higher than those of either of
the other methods. Analysis of the data for total B6 in
the fortified dehydrated model systems (PN, PM, and PLP)
revealed discrepancies in the semiautomated fluorometric
data ranging from 2.0 to 4.6 fold relative to the results
from the microbiological assay and 1.9 to 3.2 fold rela-

tive to the HPLC assay data. As previously observed, the

greatest relative differences were observed in the analysis

136

Table 23. Determination of 4-pyridoxic acid in model sys-
tems stored for 128 days in TDT cans at 37°C.

 

 

Model System _ a,b,c Percentage of
Fortification pg 4 PA/g Initial B6
Nonfortified ND --—

PN ND ---

PM ND ---

PLP 0.39 i 0.02 2.6 i 0.1

 

aMean and standard deviation. Duplicate determi-
nations.

bDetection limit (signal/noise = 3) approximately
3 ng/50 ul injection, corresponding to 0.08 ug/g model
system.

cND = not detectable.

Table 24. Determination of pyridoxyllysine in model sys-
tems from TDT and 303 cans pooled after storage
for 128 days at 37°C. Data expressed as bound
pyridoxal/g model system.

 

 

Model System a,b,c
Fortification ug Bound PL/g
Nonfortified ND

PN ND

PM ND

PLP 4.17 t 0.03

 

aMean and standard deviation. Duplicate determi-
nations.

bDetection limit (signal/noise = 3) approximately
6 ng/50 ul injection, corresponding to 12 ng/mg protein
or 0.6 ug/g model system.

cND = not detectable.

137

Table 25. Microbiological assay of vitamin B6 in model
systems pooled after storage for 128 days in
TDT and 303 cans at 37°C.a

 

Model System pg Total B6/g

 

Fortification

Nonfortified <0.5

PN 11.9 i 2.1
PM 6.3 t 0.1
PLP 2.9 1 0.6

 

aMean and standard deviation. Duplicate determi—
nations.

Table 26. High performance liquid chromatographic assay
of vitamin B6 in model systems pooled after
storage for 128 days in TDT and 303 cans at

 

 

37°C.a,b,c
Model System
Fortification PN PM PL Total
Nonfortified ND ND tr <0.5
PN 12.8 i 0.1 ND ND 12.8 i 0.1
PM ND 3.3 i 0.4 4.2 i 0.4 7.5 t 0.6
PLP ND ND 4.2 i 0.3 4.2 i 0.3

 

aug/g model system. Mean and standard deviation
of duplicate determinations.

bND = not detectable; tr = trace, too low to
quantitate.

136

Table 23. Determination of 4-pyridoxic acid in model sys-
tems stored for 128 days in TDT cans at 37°C.

 

 

Model System _ a,b,c Percentage of
Fortification ug 4 PA/g Initial B6
Nonfortified ND ---

PN ND ---

PM ND ---

PLP 0.39 i 0.02 2.6 i 0.1

 

3Mean and standard deviation. Duplicate determi-
nations.

bDetection limit (signal/noise = 3) approximately
3 ng/50 ul injection, corresponding to 0.08 ug/g model
system.

CND = not detectable.

Table 24. Determination of pyridoxyllysine in model sys-
tems from TDT and 303 cans pooled after storage
for 128 days at 37°C. Data expressed as bound
pyridoxal/g model system.

 

 

Model System a,b,c
Fortification ”9 Bound PL/g
Nonfortified ND

PN ND

PM ND

PLP 4.17 i 0.03

 

aMean and standard deviation. Duplicate determi-
nations.

bDetection limit (signal/noise = 3) approximately
6 ng/50 ul injection, corresponding to 12 ng/mg protein
or 0.6 ug/g model system.

cND = not detectable.

137

Table 25. Microbiological assay of vitamin 36 in model
systems pooled after storage for 128 days in
TDT and 303 cans at 37°C.a

 

Model System pg Total BG/g

 

Fortification

Nonfortified <0.5

PN 11.9 i 2.1
PM 6.3 i 0.1
PLP 2.9 i 0.6

 

aMean and standard deviation. Duplicate determi-
nations.

Table 26. High performance liquid chromatographic assay
of vitamin Be in model systems pooled after
storage for 128 days in TDT and 303 cans at

 

 

37°C.a,b,c
Model System
Fortification PN PM PL Total
Nonfortified ND ND tr <0.5
PN 12.8 t 0.1 ND ND 12.8 i 0.1
PM ND 3.3 i 0.4 4.2 i 0.4 7.5 i 0.6
PLP ND ND 4.2 i 0.3 4.2 t 0.3

 

aug/g model system. Mean and standard deviation
of duplicate determinations.

bND = not detectable; tr = trace, too low to
quantitate.

138

of the nonfortified systems. Comparing the data for the

semiautomated fluorometric and HPLC assays for the indi—

vidual vitamers, the greatest difference was consistently

observed in the PM fraction, while the PL and PN data

correlated more closely.

Table 27. Semiautomated fluorometric determination of
vitamin B5 in dehydrated model systems pooled

after storage for 128 days at 37°C in TDT and
303 cans.ar

 

Model System

 

Fortification PN PM PL Total

Nonfortified 3.8 i 0.3 8.0 i 1.5 1.7 i 0.3 13.5 i 1.6
PN 12.9 i 0.8 7.5 i 1.6 3.7 i 0.3 24.0 i 1.8
PM 0.6 i 0.5 9.7 i 1.1 6. i . 16.3 i 1.6
PLP 0.5 i . . t 1.8 4.4 i 1.8 13.4 i 2.6

 

aug/g model system.

bMean and standard deviation of triplicate
analyses.

Biologically Available Vitamin B5 in
IRflERfififliiT1Rfifiif1fifififififi??fifififi?“""

-:

Storage

The results of the rat bioassay for available
vitamin B6 in the dehydrated model systems which were
pooled after 128 days storage in TDT and 303 cans at
37°C (Bioassay 2) are presented in Table 28. (The data
for the diets containing PP-BSA will be discussed in
the next section.) The dose-response curves based on

the response of rats fed the 0-1.0 pg PN/g standard

139

 

 

 

 

 

 

 

 

 

 

 

m.m H H.~oH mm.o H ~a.e HHo.o H mpm.o NH H mom m H ooa m\zo on m~.o
+ dmmlmm
o.m H «.moa o~.o H mm.v mao.o H mvm.o h H mom m H mm maco «mmlmm
.HoHo m\qo meson m, m.o. mmmaoo
m.~a H m.ova ma.o H Hm.m mHo.o H mm~.o m H haw m H no mam
h.m H «.moa mm.o H b~.¢ mao.o H am~.o w H Hmm m H me So
m.h H m.mm mm.o H oo.m moo.o H wov.o NH H mom m H mad zm
m.e H ~.oma H~.o H m~.m «wo.o H mm~.o m H mma m H we monwuuomsoz
mEonam Homo: ooHon we
v.~ H m.mm mm.o H mm.m mmo.o H mvc.o m H van m H and = oo.H
o.~ H m.om No.0 H Hm.m mao.o H -¢.o Ha H mum m H had a mh.o
~.h H m.am o>.o H ma.m mmo.o H omm.o ca H «mu m H mm a om.o
o.v H ¢.mHH m¢.o H oo.m mao.o H mmm.o Ha H mam v H me : m~.o
H.m H ~.mHH «H.o H HH.m -o.o H em~.o o H omH N H we o\zo on oo.o
oummamum
.Hc .om o2\oev .o\mo .oc .my
GOHumaseHum mam >HH>HHU< woesmcoo moon m mofismcoo 30H uoHo
mm. ousouo comm no o
Edged ouhoousuhnm
.msoum Hem mumn mum .Houuo oummcmum mam :mmz .«mmlmm mo muH>Huom Hmowmoaown
mo mHmmamcm cam .oobm um mcmo mom mam any cw ommnoum mamo mma Houmm moaoom
mEoumwm Homes cH mm cHemuH> maanHm>m mo GOHumcHEHouoo .m mmmmmOHm umm .mm oanme

140

diets were calculated, and their regression parameters
are presented in Table 29. Log dose-response functions
provided the best linear fit for the growth, growth per
gram feed, and PLP stimulation of AspAT data, while a
double reciprocal plot was employed for the erythrocyte
AspAT activity data.

The bioassay estimates of biologically available
vitamin 86 in the model systems which had been stored
128 days at 37°C are presented in Table 30. The esti-
mates of available B6 by each of the response parameters
were not significantly different in the case of the non-
fortified and PN- and PLP-fortified model systems (P <
.05). However, for the PM-fortified system, the esti-
mates based on growth and growth per gram feed were sig-
nificantly less than those determined by PLP stimulation
of AspAT, and the growth-based estimate was significantly
less than the estimate based on AspAT activity (P < .05).
As previously observed, the precision of the estimates
was generally highest when employing growth and growth
per gram feed as the response criteria.

Generally excellent correlation was observed
between the rat bioassay, microbiological, and HPLC
methods for determination of vitamin B6 in the dehydrated
model systems (Table 31). No significant differences
were observed between the total vitamin B6 in the model

systems, as determined by the microbiological and HPLC

141

.oSHm> some on H «0 GOHuHoom on» an oomoo mums mGOHumuucoocoo may .soxmu mm3 UOH on»

cmsz .oconoHuhm venom mo GOHumHucoocoo anmuoHU on» ucomoumou mumo x HH<o

.coHHouHHo uncommon some now mucoHoHumooo :oHumHoHHoo
HomemHn on» vooH>oum 50H£3 mmonu ..o.H .aHnchHumHoH HmocHH m uoonm mHHHHbmHHm>
ummoH on» oo©H>oum 20Hn3 m:0Huocom ucomoumoh mm>uoo oncommoulomoc moosem

 

 

oaHo.+ Hmo.o H oo~.o mmo.o H ~4o.o H1.>HH>HH0< Hammav H1.H + zoo
amwa.1 H.mH H m.mm~1 H.m H H.mHH ooHHHHoeHHm ago aeoma .H + 2o. ooH
oaaa.+ «Ho.o H ovo.o Hoo.o H om~.o omen o\.m. noono .H + zoo ooH
Hemm.+ o.oH H m.oh~ ~.m H m.me .o. ouzouo .H + zoo ooH
ucoHonmooo
:oHumHoHHoo oQOHm umoououcle H ex

 

m.¢mmumm mo >HH>Huom HmonoHOHn mom mfimumwm Hooofi monoum :H mm oHanHm>m
.N wmmmmOHm .mo>uso oncommmuuomom may mo muouosmnmm sonmonoH HmocHA .mm oHnt

142

..mo. v a. HHH>HHoo Edema co momma
umnu omen mmoH mHucmOHchmHm mmz oumEHumo nu3onm may oHHn3 .:0HumHaEHum mam an omonu
cmnu mmoH MHucmoHHHcmHm mums comm m\nu3oum 0cm £u3oum an moumEHumm .Emumhm oonHuHom

12o moo oH .mEoommm ooHHHHHoouoHo one .1zo .ooHHHHHoHooe oH HooHoHoHo HHHoHoHHHo
ImHm no: mHHouHHo oncommou may no some an on CHEmuH> oHanHm>m mo moumEHummo

.Emummm H0605 on» cH mm :HemuH> oHanHm>m on» no cOHumHsono
CH poms moumum HmGOHuHHusc mm mo xowcH on» on mummon GOHHouHHo uncommomn

.mooum mmmmm mom mumu m .Houuo oumocmum mom cmmzm

 

 

 

 

 

m.~ H m.H ~.H H m.H m.H H ~.m N.H H ~.o mam
v.H H ~.~H m.m H H.m ~.H H n.m m.o H m.v 2m
H.~ H m.mH m.v H m.mH v.H H ~.hH H.H H 5.5H zm
~.H H m.o| o.~ H ~.H| m.o H h.ol h.o H m.ot ooHMHuHOHcoz
ooHuoHoeHum mam >HH>HHoH moeomooo moon o
.m. nusouo sueouo
94mm¢ ouhoounumum

coHumOHMHuHom

cOHHouHHU oncommom Emumzm Home:

o.n

 

m.Eouwa HoooE
m\mm oHanHm>m m1 Hammonmou mumo .Uohm um memo mom 0cm Boa :H mmmo mNH
How ommuoum Houmm ooHoom msmumwm Hocoa cH mm cHemuH> oHanHm>m >HHm0Hm0H0Hm .om oHnme

143

..mo. v a. soummm

Hooofi some How moumfiHumo hmmmmOHn on» Eonm ucmHoMMHc hHusmoHMHcmHm poo ouo3 mmsHm>
qum 0cm HmOHmOHoHQOHOHz .mosHm> Odom mam HmOHmOHOHAOHOHE coosuon monsoHoMMHm Homo

IHMHcmHm oz .moumEHumo Honvo HHm cmnu Houmoum mHucmoHMHsmHm mumn OHHuoEouoon

n

.Emumam Hoooa m\mm :HemuH> Hmuou mam

 

 

 

 

v.mH ~.v m.m m.H m.H ~.m ~.v mam
m.mH m.n m.m ~.oH H.m h.m m.v 2m
o.v~ m.~H m.HH m.mH m.mH m.nH >.nH zm
m.mH m.o m.o m.ot ~.H| >.ol m.c| oonHHHomooz
coHumHofiHum muH>Hu04
mam . . mesmcou mom 5 nu3on0
.eoHoon ego: .HoHooHon gamma 1 . ouono ooHHHoHHHHHoo

Eoumhm Homo:

 

GOHHouHHo oncommmm

 

o.m.oohm Hm memo mom one

any aH ommuoum mwmo mmH Houmm omHoom mfimumhm HoooE GH mm :HemuH> How mammmm
IOHA umu 0cm .oqmm .HmOHmOHOHoouoHE .OHHuoEonosHm moumfiousmHfiom mo :OmHHmmEoo .Hm oHnt

144

methods, and any of the rat bioassay estimates, indicat-
ing no significant loss of bioavailability during the
storage period (P < .05). That is, in spite of the losses
of vitamin B6 during storage of the dehydrated model sys-
tems, the remaining B6 exhibited complete bioavailability.
In contrast to the correlation exhibited between the
microbiological, HPLC, and rat bioassay data, the results
of the semiautomated fluorometric assay were observed to
be consistently higher than the rat bioassay results.

For all model systems, semiautomated fluorometric values
were greater by a mean of 10.6 i 3.0 ug vitamin BG/g.

Biological Activit of Phosphopyridoxyl-
Bovine Serum Albumin

 

Analysis of Phosphopyridox l-
BOVine Serum Albumin

 

The dialyzed, freeze dried, PP-BSA was examined
to determine the concentration of pyridoxylamino residues
prior to analysis of their biological activity by rat
bioassay. The ultraviolet absorption difference spectrum
of PP-BSA against a solution of bovine serum albumin at
neutral pH (Figure 16) permitted measurement of the pyri-
doxylamino groups using their molar absorptivity at
323 nm (Dempsey and Christensen, 1962). Employing this
spectrophotometric method, 2.32 moles of pyridoxylamino
groups were detected per mole of serum albumin (68,000
daltons) in the PP-BSA preparation. This was equivalent

to 7.81 ug bound PLP/mg protein or 5.76 09 bound PL/mg

145

protein. Measurement of pyridoxyllysine residues in a
PP-BSA acid hydrolysate by HPLC confirmed this value,

and indicated that essentially all of the pyridoxylamino
groups of the PP-BSA were in the form of e-pyridoxyllysine
residues.

Chromatography of the PP-BSA on Sephadex G-25
revealed a low concentration of UV-absorbing compounds,
suggesting trace contamination with free vitamin B6.
Identification and quantitation of the contaminants was
performed by semiautomated fluorometric and HPLC assays

for free vitamin B (Table 32). Both fluorometric and

6
HPLC procedures revealed approximately 1 ug total free
BG/mg protein. In view of the greater precision of the
HPLC analysis, the value of 1.06 ug/mg was used in sub-
sequent calculations to quantitate the free vitamin B6
contributed by PP-BSA when added to the rat diets.
Microbiological assay of the PP-BSA revealed an apparent
value of only 0.46 ug/mg (Table 32). These results
suggest incomplete utilization of the free B6 by the
assay organism, possibly as a result of inhibition by
the PP-BSA pyridoxylamino groups.

Pyridoxylamino Group Utilization
and Antivitamin Activity

 

The vitamin B6 activity of protein bound pyri-
doxylamino groups was determined by rat bioassay of

diets with PP-BSA as the sole source of vitamin 86. The

146

.mcoHumcHEHmumo mumOHHmHHu mo GOHumH>mU mnmocmum mam :mm:

a

.chuoum mE\mm cHEmuH> mam

 

 

no.o H III III III HmOHmOHoHQOHOHz
mH.o H vo.o H H~.o MH.o H m~.o oo.o H mv.o OHHumEouosHm
no.o H no.0 H h~.o oz mo.o H mh.o 04mm
monumz
Hmuoe Am So zm HmOHuhHmad

 

n

.m.¢mmnmm omumHmHo GH om GHEmuH> mmuu mo GOHumcHEHmumo .Nm mHnt

147

addition of PP-BSA to the diets was calculated to pro-
vide 0.5 ug bound PL/g diet. In the same study, the
effects of dietary protein bound pyridoxylamino groups

on the utilization of free pyridoxine were determined by
rat bioassay of diets containing PP-BSA at the same level,
as well as an additional 0.25 ug PN/g diet. The results
of the rat bioassay (Bioassay 2) are shown in Table 28.
The vitamin B6 activity of the diets was determined by
comparison of the responses of the rats of each dietary
group to the dose-response curves (Table 29).

The apparent vitamin B content of diets contain-

6
ing PP-BSA and PP-BSA fortified with free PN are shown
in Table 33. These data indicated that approximately

60% of the pyridoxylamino group B in PP-BSA was available

6
for utilization as vitamin B6 by the rats. Each estimate
was corrected for the contribution of the free B6 in the
PP-BSA by subtraction of 0.092 ug/g diet. No significant
differences were observed between the bioassay estimates
based on the various response parameters for either diet
(P < .05) .

Comparison of the observed response of the rats
to the 0.25 ug PN/g diet with the response predicted by
the dose-response curves was used to determine the extent
of utilization of free PN added to the diet. The results

indicated a mean utilization of only 44.5% of the free

PN (Table 33).

148

.mm>Hoo mmcommmulmmoo ms» Eonm omuomoxm umsu ou m\zm m: m~.o omoom
may on mmsommmu om>Hmmno m2» m0 cOmHHmmEoo men on mummmu omNHHHuo zm usmoummm

.zm mmum mo
umnu ou m>HumHmH Am mason :Hmuoum mo >HH>HHom mnu ou mnmwmu >HH>Huom unmoummo

..qum an omcHEHmumo mm. «mmtmm may :H mucmcHEmucoo
momma may we omusoHHucoo umHo m\mm mmum m mmo.o man How mmuomuuoo mumm m\mnn

.msoum Hmm mumu m .Houum oumocmum 0cm cmmzm

 

 

 

 

Hm mo.o H mm.o Hm oo.o H m~.o ooHHmHoeHHm ago gamma

mm mH.o H mH.o mm mo.o H om.o HHH>HHo< gamma

ca oo.o H HH.o No Ho.o H Hm.o ooom o\1ov ooono

mm Ho.o H mv.o so mo.o H om.o ouzouo
oomNHHHuD zm w umHo m\ma o>HH>Huo< m umHo m\m: :oHHmHHHU
ooHo m\zm ma mm.o + Hmmumm mHoo Hanuoo omooomom

 

o.m.HoHo o\qa ooooo ma om.o ooH>oHo 0» mHoHo soon a» oooom
mm3 4mmlmm .«mmlmm cH mmsonmH mCHmmHHaxomHHmm mo muH>Huom HmOHmoHOHm .mm mHnt

DISCUSSION

Model food systems were employed to study the
stability and bioavailability of vitamin B6 as a function
of selected processing and storage conditions. The model
system approach was chosen to permit: (1) ready isolation
of individual system components for analysis, (2) deter-
mination of the effects of various ingredients by varying
the model system composition, and (3) direct study of
the stability and bioavailability of the individual
vitamers because of the low levels of vitamin B6 inher—
ently present. The results obtained in these studies
could readily serve as a basis for further investigations
of the effects of ingredients, processing, and storage
conditions on the stability of vitamin B6 in the model
systems.

Interaction of Pyridoxal and Pyridoxal Phosphate

with Peptidesiin a Liquid Model Food
System duringThermal ProcesSing

 

 

 

Thermal stability of the peptide mixture was
required to facilitate the reduction of Schiff bases
with sodium borohydride, isolation of the model system

peptides by gel filtration chromatography, and analysis

149

150

of peptide bound vitamin B6 by spectrophotometric methods.
The heterogeneity of the peptide mixture, as indicated

by the peptide mapping results and molecular weight range
(700-16,500), was considered to be an advantage of these
peptides as a model for complex mixtures of food proteins.

The UV-visible absorption difference spectra of
each peptide fraction versus its respective nonfortified
blank demonstrated the presence of bound PL and PLP. The
absorption maxima at 323-325 nm and 253-254 nm (Figure 7)
were identical to those observed for pyridoxylamino and
substituted aldamine complexes (Dempsey and Christensen,
1962; Kierska and Maslinski, 1971; O'Leary, 1971). These
difference spectra were also identical to those exhibited
by a wide variety of proteins with PLP bound by reduced
linkages (Dempsey and Christensen, 1962; Anderson, 23 31.,
1966; Page, 23 31., 1968; Ronchi, st 31., 1969; Johnson
and Deal, 1970). Thus, the total bound forms of the
vitamin represented the sum of borohydride-reduced Schiff
base derivatives plus nonreducible complexes.

The examination of peptides recovered at neutral
pH without prior borohydride reduction revealed dif—
ference spectra identical to those obtained in the analy-
sis of total bound PL and PLP. These results confirmed
that nonreducible B6 complexes were formed during the
thermal processing of all liquid model systems. The

absence of an absorption maximum at 410-430 nm indicates

151

that none of the bound PLP derivatives were present as a
Schiff base (Kent, 22 31., 1958). This observation sup-
ports the hypothesis that the labile Schiff base com-
plexes would dissociate during the gel filtration isolation
of the peptides, leaving only the stable nonreducible
forms.

Several researchers have demonstrated that PLP
bound as a Schiff base in specific hydrophobic sites of
glycogen phosphorylase and serum albumin exhibits
absorption spectra similar to those of the nonreducible
complexes. In addition, these forms of bound PLP are
resistant to NaBH4 reduction (Shaltiel and Cortijo, 1970;
Hilak, st 31., 1975; Cortijo, st 31., 1976). However,
in the thermally denatured peptides used in this study,
the existence of such specific apolar binding sites is
highly unlikely.

Further evidence for the formation of nonreducible
complexes was observed in the fluorescence spectra of
the isolated peptides upon excitation at 280 nm. Pro-
teins with reduced pyridoxylamino groups have been shown
to exhibit a decrease in the native fluorescence of tryp-
tophan residues as a result of the radiationless transfer

of electronic energy to the bound B .chromophores

6
(Churchich, 1965). The fluorescence emission spectra
of the peptides from PLP-fortified model systems, when

compared to the spectra of peptides from nonfortified

152

blanks, demonstrated this quenching effect (Figure 8).
Thus the presence of nonreducible PLP-peptide conjugates
was confirmed by two independent spectrophotometric
methods.

In the determination of peptide bound pyridoxyl-
amino compounds, the peptides were isolated from the model
systems at pH 2.5 to induce dissociation of the acid
labile substituted aldamine and Schiff base complexes.
Upon readjustment of the recovered peptide solutions to
pH 7.0, the difference spectra confirmed the presence of
peptide bound PLP as pyridoxylamino derivatives in each
of the liquid model systems, regardless of treatment with
glucose and/or ascorbic acid.

Quantitation of levels of total bound vitamin BG
(Table 5) indicated that PLP was significantly more
reactive than PL (P < .001). The difference in reactivity,
previously observed in many coenzyme-apoenzyme binding
studies, has been attributed to the inhibition of internal
hemiacetal formation by the 5'-phosphate group of PLP
(Heyl, £3 31., 1951). Because of this blocking effect,
the concentration of free aldehyde is much greater in
solutions of PLP than in PL solutions of comparable molar
concentration. Presumably, electrostatic effects between
the phosphate group and cationic sites on the peptides

may also play a role in the greater affinity of PLP. A

153

mean of 20% of the total model system PLP was peptide
bound, compared to only 6.7% of the total PL.
Comparison of the mean value of the data of
Table 5 for total bound PLP (Schiff base, substituted
aldamine, and pyridoxylamino) with that of Table 6 for
PLP bound through nonreducible linkages indicated that,
at neutral pH, approximately 60% of the total bound PLP
existed as a nonreducible complex. These data (Table 6)

suggested that losses of biologically available B may

6
occur through complex formation. The direct determination
of the pyridoxylamino complexes as acid stable, non-
reducible conjugates (Table 7) confirmed the existence

of bound PLP in a form which has limited bioavailability.
Binding in this manner would not represent a major
mechanism for the loss of available vitamin 36’ as
indicated by the fact that only 9.7 i 6.4% of the total
PLP in the systems was involved in these complexes.

The analysis of the effects of glucose and
ascorbic acid on the binding of PL and PLP may have been
hampered by the precision of the data. Statistically,
neither factor along nor in combination significantly
affected the extent or manner of PL or PLP binding
(P < .05) .

Several qualitative effects were apparent which
correlated with chemical analysis of the peptides. Data

shown in Table 5 for the total bound vitamin B indicate

6

154

that, in systems treated only with ascorbic acid, the
extent of binding was partially depressed. Tables 6

and 7 also exhibited this trend. This mechanism,
although unclear, may involve the competition between
ascorbate degradation products and PL or PLP for peptide
free amino groups. Free SH levels in the peptides in
all ascorbate-treated model systems were significantly
higher than in nontreated systems (Table 8). The exper-
imentally determined SH levels in the heat treated pep-
tides, which were comparable to those of denatured milk
serum proteins (Srncova and Davidek, 1972), indicated a
reduction of disulfide bonds during the heat treatment.
These SH data, when correlated with the levels of total
bound PL and PLP (Table 5), suggest that the binding of

these B vitamers was wholely independent of the concen-

6
tration of peptide SH groups. This conclusion is sup-
ported by the kinetic results of Schonbeck, 22 31. (1975),
which demonstrated that the affinity of amino groups
toward PLP was much greater than that of SH groups.

Since the concentration of peptide amino groups was
approximately 50 fold greater than the SH concentration,
the reaction of PL and PLP with amino groups would be
greatly favored in most food systems. In total, these
observations suggest that SH groups would not signifi-

cantly participate in the mechanism of binding by

proteins or peptides.

155

The reaction between PL and milk protein SH
groups, which was proposed to explain the losses of bio-
logically available vitamin B6 in thermally sterilized
milk (Bernhart, 33 31., 1960; Wendt and Bernhart, 1960),
was not supported by the results of the present study
and the findings of Schonbeck, 25,31. (1975). The data
of Srncova and Davidek (1972) demonstrated that, although
SH groups may be involved in the binding of vitamin B6
to milk proteins, they only react with a small fraction
of the total B6 in the system. In addition, studies of
SH reactions have never conclusively demonstrated the
formation of bis-4-pyridoxyl disulfide in milk products,
as postulated by Wendt and Bernhart (1960), except in
milk experimentally fortified with unrealistically high
levels of PL. The findings of Srncova and Davidek, which
demonstrated a small loss of protein free SH groups when
heated in the presence of PL, could be interpreted as
indicating low levels of substituted aldamine formation,
with SH groups acting as the nucleophile attacking the
Schiff base linkage.

Inspection of the data of Tables 6 and 7 suggests
that the presence of glucose may have a weak effect in
inhibiting the formation of nonreducible complexes,
probably by competing for free amino groups. Chemical

analysis of the peptides after heat treatment of the

liquid model systems indicated that the presence of

156

glucose induced a 26% loss of free amino groups. Such
levels of browning in food processing have been widely
reported; for example, Mauron (1960) found a 20% loss
of available lysine in evaporated milk after retort
sterilization. Thus, the level of browning routinely
encountered in food as a result of thermal processing

may retard the loss of available vitamin B by reaction

6
with free amino groups.

At the onset of these studies, it was postulated
that ascorbic acid or reductones formed in browning
reactions during processing may promote the reduction of
Schiff bases to pyridoxylamino complexes. The experi-
mental data, as discussed, contradict this hypothesis.
The data of Table 7 demonstrate that the highest levels
of pyridoxylamino compound formation occurred in control
model systems containing no added ascorbate or glucose.
Based on these experimental results, the mechanism of
Schiff base reduction remains unclear. Conceivably,
the reducing equivalents may originate with the peptides
per se, possibly during the thermally induced degradation
of certain amino acids.

The levels of pyridoxylamino compound formation
(9.7 i 6.4% of the total PLP) observed in these studies
indicate that, under the conditions employed, losses of
available vitamin 36 by complex formation are of little

nutritional significance. Therefore, these data do not

157

support the thermally induced 50% reduction of vitamin B6
bioavailability in sterilized milk products reported by
Tomarelli, gg'gl. (1955). Davies, g£_§l, (1959), observed
an apparent thermal destruction of vitamin B6 in milk
during sterilization; however, they were unable to
demonstrate any losses of bioavailability of the 36
remaining after processing. In fact, they observed

higher values for available B as measured by rat or

6
chick bioassays than were obtained for apparent total
B6 in microbiological (S. uvarum) analyses. Similar
results, i.e. higher rat bioassay than S. uvarum results,
were obtained by Lushbough, 25 2;. (1959), in a study
involving cooked meats. When taken in total, the present
experimental results, along with those of Davies, gE_§£.
(1959), and Lushbough, SE.E£3 (1959), suggest that the
specificity of the microbiological assay method employed
by Tomarelli, 22.3l. (1955), may have artifactually
indicated losses of biological availability of the
vitamin B6 in the sterilized milk products.

As previously discussed, PLP has been found to
react with proteins to a much greater extent than PL,
PM, or PN (Anderson, 2E.Elf’ 1974). Since the losses
of bioavailability due to the binding of PLP were quite
low, it may be inferred that losses of the other vitamers

by binding to proteins during food processing would be

minimal.

158

Interactions of Pyridoxamine in a Liquid

Model System during_THermal Processing

 

 

 

Although the losses of available vitamin B by

6
the binding of PL and PLP to peptides and proteins were
found to be low, the reaction of PM with low molecular
weight carbonyl compounds was postulated as another

possible mechanism for the loss of available B Follow-

6.
ing a thermal process of 20 minutes at 121°C, the vitamin
B6 concentration of PM-fortified liquid model systems was
determined. Peptides were omitted from these systems to

permit HPLC analysis of vitamin B To ensure measure-

6'
ment of all biologically available forms of the vitamin,
the model systems were adjusted to pH 2.5 after auto-
claving to induce the dissociation of any acid labile
complexes. The data of Table 9 indicate that, while
glucose was involved in the partial conversion of PM to
PL, it also participated in the degradation of PM.
Ascorbic acid had a similar effect, although much less
pronounced. These results suggest that the observed
losses of PM proceded through a reaction involving the
carbonyl groups of glucose, ascorbate, or their thermal
degradation products. Efforts to detect PM—containing
complexes by gel filtration chromatography were not
successful. The low levels of PM degraded, in addition

to the poor resolving power of the chromatographic

method, may have precluded their detection.

159

The extent of conversion of PM to PL in these
liquid model systems was very low. Similar results
were obtained by Metzler and Snell (1952) in their
findings that metal ions were required for the catalysis
of transamination in aqueous solution.

In summary, the results of this study have demon-
strated the involvement of glucose and ascorbic acid in
the loss of PM during thermal processing. The losses of
total vitamin B6 in the PM-fortified liquid model sys-
tems under these conditions were quite low. Therefore,
the thermally induced degradation and/or binding of PM
probably does not represent an important pathway for
the loss of biologically available vitamin B6 during the

retort processing of foods.

Vitamin B6 Assay Methods

 

Comparison of the vitamin 36 analyses of the
dehydrated model system studies (Tables 11, 13, 18, and
31) provided an opportunity to determine the accuracy
of the various assay methods (microbiological, HPLC,
semiautomated fluorometric, and rat bioassay). The
data indicate that, in general, the microbiological and
HPLC methods correlated very well with each other, and
were in close agreement with rat bioassay estimates based
on growth and growth per gram feed consumed. .Model sys-
tems used in these studies exhibited no significant

losses of vitamin B6 bioavailability during processing

160

or storage, therefore, the rat bioassay data served to
confirm the validity of the microbiological and HPLC
methods as estimators of biologically available vitamin
B6 in the dehydrated model systems. Unless prior animal
bioassays were performed on other food products, the
accuracy of the microbiological and HPLC methods in
predicting biologically available vitamin B6 could not
be assumed.

In the case of all dehydrated model systems
assayed, the semiautomated fluorometric assay data was
significantly higher than HPLC, microbiological, and
rat bioassay results. The greatest differential between
HPLC and semiautomated fluorometric results was observed
in the PM determination. This was unexpected because
of the theoretically adequate blank to provide correction
for non-B6 fluorophores. In thin layer chromatographic
(Gregory and Kirk, 1977a) and fluorescence spectrophoto-
metric studies (Table 14), no non-B6 interfering com-
pounds could be detected. The initial studies of
Gregory and Kirk (1977a) demonstrated that previous
fluorometric methods based on 4-pyridoxic acid lactone
formation were subject to interference by cyanide-
reactive browning compounds. The present ion exchange
preparative procedure was developed to alleviate this
problem. However, the observed poor correlation with

other assay methods suggests that the semiautomated

161

fluorometric assay is still subject to interference.
The data of Table 13 indicate that the interfering com-
pound(s) possesses heat sensitivity quite similar to
that of PN, PM, and PLP. The fact that the fluorescence
emission spectra of all model system fractions were
identical to that of pure 4-pyridoxic acid lactone sug-
gests that the interfering compound is converted to 4-
pyridoxic acid lactone or a very similar product during
the formation of the fluorophore. The heat sensitivity
and fluorescence spectra indicate that the interfering
compound may be a vitamin B6 derivative which possesses
little or no biological activity. Further research
would be required to identify this compound and alleviate
the problem of lack of specificity of the semiautomated
fluorometric assay.

The results of Bioassays l and 2 (Tables 15 and
28) demonstrate dose-response relationships similar to
those previously reported (Sarma, EE.21°' 1946; Links-
wiler, $2.31f' 1951; Lushbough, EE.El°' 1959; Beaton and
Cheney, 1965; Brin and Thiele, 1967). The positive cor-
relation between feed consumption and dietary vitamin B6
concentration, when using an ad libitum feeding regimen,
was previously observed by Beaton and Cheney (1965) and
Yen, gE_g£. (1976). Thiele and Brin (1966) observed
that rats receiving adequate levels of PN exhibited a

much greater growth response when fed ad libitum than

162

when pair-fed to rats receiving a nonfortified diet.
The growth rate of pair-fed rats was only slightly
greater than that of vitamin B6 deficient rats in their
study. Thus, for quantitation of a bioassay using
growth as a response criterion, an ad libitum design
must be employed. In such a feeding regimen, the
results of the assay could be biased by any preference
for or avoidance of the test diets by the rats. In the
bioassays of this study, the high correlation between
estimates of available B6 based on growth and growth
per gram feed consumed indicated that this was not a
problem (Tables 17 and 30).

Stability and Bioavailability of Vitamin B5

In'DEEYHFEEEH'MEHEI‘SYEEEME:
Roasting Effects

 

 

 

 

The relative stability of the B6 vitamers used
in fortifying the model systems was determined by com-
parison of the total vitamin B6 data of microbiological
and semiautomated fluorometric analyses of control and
roasted model systems. In spite of the high values
obtained by the semiautomated fluorometric assay of the

dehydrated model systems for total vitamin B , the

6
percent loss calculated from the fluorometric assay

correlated well with that determined microbiologically.
Therefore, for relative measurements, the fluorometric

assay appears to be valid. The data of Table 13 indi-

cated that, within the limits of experimental error, no

163

differences in the stability of PN, PM, and PLP could
be detected. All three underwent approximately 50 to
70% degradation as a result of the 25-minute roasting
at 180°C.. The apparent greater stability of PM as
determined microbiologically was presumably due to the
lower response of the organism to the high level of PM
in the nonroasted system, which was lowered somewhat
by partial conversion to PL during the roasting process
(Table 12).

The levels of thermal degradation of vitamin B6
in the roasted model systems were comparable to those
observed for the loss of total B6 in the retort sterili-
zation of evaporated milk, as monitored bny. uvarum
assays (Hassinen, EE.El°' 1954). Hassinen, EE.21° (1954),
also observed that added PN was much more resistant to
thermal degradation than added PL or PM in milk. The
equivalent levels of destruction of PM, PN, and PLP
during the roasting of the model systems, as compared
to the thermal destruction in fluid milk, may have been
due to the fact that moisture levels were much lower in
the dehydrated model system.

The fact that all rat bioassay estimates of
available vitamin B6 were not significantly different
from the microbiological data for total B6 indicates

that the roasting process did not affect the bioavail-

ability of the vitamin (Table 18). As previously

164

discussed, the data indicate that the semiautomated
fluorometric assay procedure was not an accurate measure
of the biologically available vitamin B6 in the model
systems.

Stability and Bioavailability of Vitamin B5

in e y ra e nge gys ems:
Storage Effects

 

Water Activigy and Nonenzymatic
Browning

The observation that the water activity of all
model systems was slightly less than that predicted
(Table 19) suggests that moisture uptake from the air
(65% relative humidity) during the 24-hour equilibration
period was partially incomplete. The slight differences
in water activity of the model systems fortified with
different B6 vitamers may have been a result of dif-
ferences in the rate of moisture adsorption from the air,
possibly due to differences in model system position in
the equilibration chamber.

The slight differences in water activity between
the model systems containing various B6 vitamers strongly
affected the rate of nonenzymatic browning. The high
sensitivity of browning rate to changes in water activity
has been previously demonstrated (Lea and Hannan, 1949;
Labuza, 1975; Warmbier, g£_gl., 1975). The correlation
observed between the browning rate and water activity

of the model systems stored in TDT or 303 cans (Tables 19

165

and 20) demonstrated the sensitivity of the browning in
these systems to changes in water activity.

The rate and extent of browning was examined in
these model systems primarily to provide an index of
quality in addition to vitamin B6 content. Browning
has been found to be the major organoleptic defect
appearing during the storage of many intermediate
moisture foods (Waletzko and Labuza, 1976). The com-
position of low and intermediate moisture food systems,
particularly the type of protein, its ratio to reducing
sugars, the pH of the system, and the presence of humec-
tants have been found to be important factors affecting
the rate of nonenzymatic browning (Lea and Hannan, 1949;
Warmbier, SE g£., 1976; Schnickels, gg'g£., 1976). The
direct applicability of these browning data to food sys-
tems of different composition is of limited value. Sim-
ilarly, reaction kinetic parameters describing the
degradation reactions of the B6 vitamers are not known
(energy of activation, 010’ etc.). Thus, the rate con-
stants for browning and vitamin B6 loss in these studies
cannot be directly applied to the prediction of losses
of nutritional and organoleptic quality under other
storage conditions. The storage conditions used in
this study were selected to provide rates of vitamin
degradation and browning near the maximum for dehydrated

low and intermediate moisture foods.

166

The kinetics of the browning reaction for model
systems stored in TDT cans were examined. The loss of
free amino groups in the various dehydrated model systems
during storage appeared to follow a very complex overall
kinetic function. Plotting the data as a first order
function (Figure 11), several linear processes were
apparent. These results suggest that the progressive
phases of the reaction could each be described by first
order kinetics. Similar results, indicating several
first order processes involved in the disappearance of
browning reactants, have been observed by Warmbier, 25
3;. (1976), and Warren and Labuza (1977). The water
activity of the model systems appeared to affect time
required for transition between the various first order
phases of the loss of free amino groups in this study.
The kinetics observed for the formation of browning pig-
ments in the TDT can model systems were contrary to
those previously reported (Waletzko and Labuza, 1976;
Warmbier, 2E,g£., 1976; Warren and Labuza, 1977). After
an initial induction period of approximately 20 days,
the formation of browning pigments followed complex
kinetics (Figure 12). The zero order rate function
reported for pigment formation in low and intermediate
moisture food systems (Waletzko and Labuza, 1976; Warm-
bier, SE g£., 1976; Warren and Labuza, 1977) was not

observed. This difference cannot be readily explained.

167

The large differences observed between systems
stored in TDT and 303 cans in the extent of browning
during the 128-day storage period (Table 20) was
unexpected. The model systems fortified with various
36 vitamers were not found to vary in water activity
between the TDT and 303 cans. These data indicate that
slight shifts in moisture equilibria were not responsible
for the difference in browning rates. The only storage
variable between the two systems was the volume of the
container headspace. This would imply that the larger
supply of gaseous oxygen in the 303 cans may have played
a role in the inhibition of pigment production. The
extent of browning pigment in the model systems in TDT
cans was 4.6-5.5 times that of the 303 can model systems,
while the difference in final amino group concentration
ranged from only 1.03 to 1.50 fold. This small dif-
ference in free amino groups, compared to the large
disparity in pigment formation, suggests that the locus
of the oxygen effect was in the intermediate stages of
the sequence of browning reactions. This effect of
headspace atmosphere on sugar-amine browning has not
been previously reported. Recently, Waletzko and
Labuza (1976) demonstrated that in an oxygen-free
environment, an intermediate moisture food, which had
been highly fortified with ascorbic acid, exhibited a

slightly lower rate of browning due to the increased

168

stability of ascorbic acid in the absence of oxygen.
Ascorbic acid degradation products readily participate
in the browning mechanism at neutral pH. These data are
not applicable to the model system used in the present
study because of the absence of ascorbic acid fortifi-
cation.

Stability of Vitamin B5 during
Storage

 

The kinetics of the degradation of all forms of

vitamin B in the model systems stored in TDT cans were

6
satisfactorily described by the first order rate function
(Table 21). The kinetics of vitamin B6 degradation in
food systems have not been previously reported. The
stability observed for PN was in agreement with several
studies concerning the storage stability of PN in low
moisture cereal products (Bunting, 1965; Anderson, SE 21"
1976; Cort, ggtg£., 1976). The induction period observed
prior to the slow degradation of PN (Figure 16) suggested
that the PN degradation was dependent upon another
reaction occurring in the model system during storage,

or that an activated state must be reached. Coincident
with the initiation of PN degradation, the most rapid
period of melanoidin formation was observed (Figure 12),

indicating a possible link between the formation of

melanoidins and the degradation of PN.

169

PM, PL, and PLP were much less stable than PN
during storage of the dehydrated model systems at 37°C
(Table 21). The magnitude of the stability differential
between PN and the other vitamers was comparable to the
differential reported for the thermal destruction of
these vitamers during retort processing of milk products
(Hassinen, gg‘g£., 1954). In contrast, the dry roasting
process employed in these studies induced a fairly uni-
form destruction of all of the B6 vitamers. Because the
thermodynamic parameters for the degradation of vitamin B6
were not measured in this study, this difference in rela-
tive stability cannot be explained.

The stabilizing effect of storage in the 303 cans
(Table 22), which had a large headspace volume, cannot
be explained. The fact that nonenzymatic browning pig-
ment formation was retarded in the 303 cans, in addition
to the slower rate of destruction of the B6 vitamers, is
evidence that the degradation of vitamin 86 during
storage may be intimately related to browning. The
greater stability of the 36 vitamers is in contrast to
the results obtained for ascorbic acid and riboflavin
(Dennison, EE.E£°' 1977; Kirk, et al., 1977). Both
ascorbic acid and riboflavin exhibited greater rates of
degradation when the oxygen concentration of the storage
environment was greater than or equivalent to atmospheric

conditions, as opposed to oxygen concentrations less

170

than 0.2 atmospheres. Because of the severe nature of
the storage conditions of 37°C at approximately 0.6 aw,
several predictions concerning the storage stability of
vitamin B6 in low to intermediate moisture foods can be
made. Pyridoxine, whether added or naturally occurring,
is highly stable. The data from this study suggest that
foods fortified with PN would retain high vitamin B6
activity for long periods of time, provided light was
excluded. Foods in which the vitamin B6 was present as
PM, PL, PLP, or PMP occurring naturally would be subject
to a rapid loss of vitamin B6 activity. Fortification
with pyridoxine hydrochloride would be required to main-
tain the nutritional quality of such foods, particularly
in products such as breakfast cereals and those designed
for meal replacement.

Degradation Products of
Vitamin B6

 

 

Little progress was made in identifying the
degradation products of the B6 vitamers used to fortify
the dehydrated model systems. A very small fraction of
the model system PLP was found to be oxidized to 4-PA
(Table 23). None of the other vitamers were converted
to 4-PA to any detectable extent. As mentioned pre-
viously, 4-pyridoxic acid lactone could not be measured
by the HPLC method employed for 4-PA analysis; therefore,

the lactone cannot be ruled out as a possible degradation

171

product. The absence of significant levels of 4-PA
suggests that oxidative degradation to 4-pyridoxic acid
lactone was not a major mechanism for the loss of active
vitamin 36 during storage.

The detection of significant levels of pyri-
doxyllysine associated with the proteins of the stored
PLP-fortified model systems of TDT and 303 cans which
were pooled prior to bioassay (Table 24) indicates that
binding was a major mechanism for the loss of biologically
available B6. Since pyridoxyllysine could not be detected
in hydrolysates of systems fortified with PN or PM, bind-
ing of these vitamers as pyridoxylamino compounds was
not responsible for the losses of available PN, PM, or PL.

Biological Availability of Vitamin B5
in Dehydrafed Model Sysfems affer

Storage

The absence of significant differences between

HPLC and microbiological assays for vitamin B and the

6

rat bioassays for biologically available B model systems

6
after storage for 128 days at 37°C indicated that the

vitamin B6 remaining in the stored model system possessed
full bioavailability. In Spite of the numerous reactions
occurring during the storage of the food systems, includ-
ing nonenzymatic browning and the binding of the vitamers

to other food components, the potency of the remaining

B6 was not affected under these experimental conditions.

172
Bioiogical Activity of Protein
nyidoxylamino Compounds

The pyridoxylamino residues of phosphopyridoxyl-
bovine serum albumin were observed to possess approxi-
mately 60% of the activity of free PN. In addition,
these pyridoxylamino residues inhibited the utilization
of approximately 55% of the free PN which was added to
the rat diet (Table 33). These data indicate that pro-
tein bound pyridoxyllysine, an apparent degradation
product of vitamin BG' possessed both pro— and anti—
vitamin B6 activities. Under the conditions of the bio-
assay of PP-BSA, the inhibition of utilization of the
free PN was counteracted by the partial vitamin B6
activity of the pyridoxyllysine residues.

Whether the negation of pro- and antivitamin B6
activities would occur in food systems which had under-
gone formation of protein bound pyridoxyllysine cannot
be determined. The results of the bioassay of the PLP-
fortified dehydrated model system after 128 days storage
at 37°C provide some evidence for this effect. The PLP-
fortified model system contained approximately equal
molar quantities of free PLP and protein bound pyri-
doxyllysine (Tables 24 and 26) following storage. The
fact that full apparent bioavailability of the vitamin B

6
was observed suggests that the partial availability of

173

B6 in the pyridoxyllysine residues compensated for the
partially inhibited utilization of free PLP in the model
system.

Since the antivitamin B6 activity of the protein
bound pyridoxyllysine residues was not thoroughly studied
at more than one dosage level, the possible dose-dependence
of its effects cannot be predicted. Although the anti-
vitamin activity of these complexes was observed to be
rather weak in comparison to other B6 antimetabolites,
the nutritional effects of higher levels of the pyri-
doxylamino compounds should be investigated. Higher
dosage levels of pyridoxyllysine could conceivably induce
a more severe inhibition of vitamin B6 utilization and/or
metabolic function.

Comparison of the chemical assays for the con-
centration of free vitamin B6 in the PP-BSA preparation
(Table 32) revealed no significant differences between
the results of semiautomated fluorometric assay using
the autoclave extraction technique and the HPLC assay,
which employed aqueous buffer extraction. These data
indicate that the pyridoxylamino compounds were not
hydrolyzed to a significant extent by the autoclave
extraction procedure. These results suggest that most
vitamin B6 assay methods would not suffer interference
due to hydrolysis of pyridoxylamino compounds. Thus,

the formation of pyridoxylamino complexes may be

174

responsible for the loss of available B6 in a food pro-
duct, but their presence would probably not cause a
discrepancy between the total B6 assay values and animal
bioassay values for biologically available vitamin B6.

This would indicate that the apparent difference between
biologically available B6 and total 36 of 50% in sterilized
milk products (Tomarelli, SE g£., 1955), probably cannot

be explained on the basis of pyridoxylamino complexes.

SUMMARY AND CONCLUSIONS

The stability and bioavailability of vitamin 86'
as affected by thermal processing and storage, was
investigated using liquid and dehydrated model food
systems.

Autoclave treatments simulating the retort pro-
cessing of foods were performed on liquid model systems.
In a liquid system of heat stable peptides at pH 7.0,

PLP was found to partially bind to free amino groups,
forming pyridoxylamino complexes of limited bioavail-
ability. The presence of glucose and/or ascorbic acid
weakly inhibited the formation of these complexes. A
mean of approximately 10% of the PLP in the liquid model
system was spectrophotometrically demonstrated to bind
as a pyridoxylamino complex, representing a nutritional
effect of relatively minor consequence. In a similar
neutral solution, PM was found to interact with glucose
and/or ascorbic acid in a manner which lowered its bio-
logical availability. Maximum losses of PM in the liquid

model system were approximately 20%. These results

suggest that in a system containing both the aldehyde

175

176

and amine forms of vitamin 36' an additive loss may
occur. Therefore, losses of available 36 could approach
30% of the total vitamin 36 during the retort processing
of a food system with physical-chemical characteristics
similar to the liquid model system and containing equi-
molar amounts of PLP and PM.

The study of vitamin B stability in dehydrated

6
model food systems during roasting and storage provided
an opportunity to correlate microbiological, semiautomated
fluorometric, and HPLC methods for the quantitation of
vitamin B6. The HPLC and microbiological methods were
found to correlate well in all systems except those con-
taining predominantly PM. In samples containing high
levels of PM, the HPLC method provided greater accuracy.
Although no direct evidence for the presence of interfer-
ing compounds could be found, the semiautomated fluoro-
metric assay procedure yielded apparent vitamin B6 values
for all dehydrated model systems which were significantly
higher than those of the other methods.

The dry roasting of dehydrated model systems at
180°C for 25 minutes, a relatively severe treatment,
was designed to permit an approximation of the maximum
levels of vitamin B6 which could be affected by a roast-
ing process. The three forms of the vitamin examined,

PN, PM, and PLP, were found to undergo between 50 and

70% degradation. Rat bioassay of these roasted systems

177

demonstrated excellent correlation with the microbiological
assay employed for the determination of total B6 in the
systems. The results indicated that the roasting process
induced no losses of bioavailability in the remaining 36'
Semiautomated fluorometric analysis of the roasted model
systems did not give an accurate estimation of biologi-
cally available vitamin 86'

Studies were designed to estimate the maximum
effects of storage on the vitamin B6 in a dehydrated
food system. Storage conditions were selected on the
basis of previous literature data to maximize browning
and vitamin degradation in a normal storage environment.
The degradation of PN, PM, PL, and PLP in model systems
at a water activity of approximately 0.6 at 37°C followed
first order kinetics. A rapid unidirectional conversion
of PM to PL was observed. Calculated half lives were
141 days for PN, and 35-46 days for PM, PL, and PLP when
stored in containers having a limited headspace. Storage
in containers providing a large headspace significantly
lowered the rates of degradation of all B6 vitamers and
nonenzymatic browning.

4-Pyridoxic acid, postulated to be a degradation
product of vitamin B6, was found at very low levels in
PLP-fortified dehydrated model systems. It was not
found in model systems fortified with the other 36

vitamers. Oxidation to 4-PA was not a significant

178

pathway for the loss of vitamin B6 activity during
storage. Approximately 50% of the total PLP which was
lost from the model systems during storage was found to
be bound to lysine residues of proteins as the pyri-
doxylamino compound, pyridoxyllysine. This complex
could only be detected in PLP-fortified model systems.
Other degradation products of the B6 vitamers were not
identified.

Rat bioassay of the dehydrated model systems
after 128 days at 37°C indicated that the remaining 36'
as measured microbiologically or by HPLC assay, was bio-
logically available. The fluorometric values for total
36 were significantly greater and were not an accurate
estimation of available vitamin B6.

To determine the biological activity of protein
bound pyridoxyllysine, a rat bioassay was performed on
diets containing added phosphopyridoxyl-bovine serum
albumin. Approximately 60% of the complexed B6 was found
to be available to the rat, while the remaining pyri-
doxyllysine exerted a partial inhibitory effect on the
utilization of free PN added to the diet. The fact that
full bioavailability was observed for vitamin B6 in the
stored PLP-fortified dehydrated model system, which was
shown to contain approximately equal molar amounts of
protein bound pyridoxyllysine and free PLP, was evidence

that the pro- and antivitamin 36 activities of pyri-

doxyllysine may counteract each other.

179

In conclusion, the results of this study have
indicated that, under the several processing and storage
conditions examined, varying losses of biologically
available vitamin B6 may occur. The nutritional conse-
quences of such losses can be monitored accurately by
microbiological assay procedures. The data also suggest
that the HPLC vitamin B6 assay method could be applied
to accurately determine the content of available B6 in
foods. Fortification with pyridoxine hydrochloride,
the most stable form of the vitamin, would assure mini-
mum losses of biologically active vitamin B6 in food

products subject to extensive nutrient degradation in

processing or storage.

APPENDIX

APPENDIX

Heat Stable Peptides

 

Formation of Peptides

 

Crystalline bovine B-lactoglobulin (3X) was
obtained from United States Biochemical Corporation.
Although discontinuous polyacrylamide gel electropho-
resis revealed several minor contaminants, for this
study, this preparation was used without further purif-
ication.

B-Lactoglobulin (1.5 g) was dissolved in 20 ml
0.1 M ammonium acetate to yield a concentration of
75 mg/ml. Concentrated hydrochloric acid was added to
adjust the pH to 2.0, followed by the addition of 15 mg
porcine pepsin (3X, Nutritional Biochemicals Corp.).
The mixture was incubated at 37°C for 6 hours with
gentle agitation. The reaction was terminated by the
adjustment of the mixture to pH 7.0 with 30% ammonium

hydroxide.

Isolation of Peptides

 

The peptic hydrolysate was divided into two
10 m1 portions and each chromatographed separately on

a Sephadex G-50 (fine) column of 2.6 x 40 cm (Vo = 81 ml)

180

181

using 0.1 M ammonium acetate, pH 7.0, at 1.0 ml/minute.
The column effluent was monitored at 254 nm using an
ultraviolet absorption detector and collected in 10 ml
fractions. Studies indicated that all material eluting
after the void volume was heat stable, i.e. not precip-
itable by autoclaving 20 minutes at 121°C in the presence
of 10 mg/ml glucose. Based on this information, the
heat stable fractions were pooled and lyophilized.
Further purification of the heat stable peptides
was carried out on a Sephadex G-10 column (2.6 x 4.0 cm;
Vo = 78 ml) to remove free amino acids and peptides less
than 700 daltons. The lyophilized peptides were dis-
solved in 20 ml 0.1 M ammonium acetate, pH 7.0, and
10 ml aliquots were chromatographed. The void volume
peaks were collected, mixed, and aliquots stored at

-25°C until used.

Synthesis of e-Pyridoxyl-L-Lysine

d-Acetyl-e-pyridoxyl-L-lysine was synthesized by
the method of Dempsey and Christensen (1962). In 30 ml
anhydrous methanol (Burdick and Jackson), 392 mg potassium
hydroxide, 1.00 g a-acetyl-L-lysine (Sigma Chemical Co.),
and 1.14 g pyridoxal hydrochloride (ICN Pharmaceuticals,
Inc.) were stirred for 15 minutes in the dark at room
temperature. After undissolved solids were removed by
filtration through Whatman 1 filter paper, 50 mg of

platinum dioxide catalyst (ICN Pharmaceuticals, Inc.)

182

was added. The mixture was hydrogenated in a 50 m1
Erlenmeyer flask at atmospheric pressure by bubbling
hydrogen gas for 1 hour, with periodic swirling of the
flask. The catalyst was removed by filtration through
Whatman 1 filter paper. The mixture was adjusted to
approximately pH 6 with dry methanolic HCl using
moistened indicator paper. The dry HCl in methanol was
prepared by bubbling HCl gas (technical; Matheson Gas
Products) through anhydrous methanol. After pH adjust-
ment to 6, the mixture was concentrated to one-third
volume in a rotary evaporator at 50°C. The precipitated
KCl was removed by filtration. The filtrate was adjusted
with dry methanolic HCl to pH 4.4, monitoring directly
with a pH meter. A yellowish precipitate of a-acetyl-
pyridoxyl-L-lysine appeared at this point and was removed
by filtration. To maximize the recovery, a second crop
of crystals was obtained by evaporating the filtrate to
dryness under reduced pressure.

The crystals of a-acetyl-pyridoxyl-L-lysine were
dissolved in 6 N HCl and autoclaved 45 minutes at 121°C
to hydrolyze the d-acetyl moiety (Dempsey and Snell,
1963). After drying in a rotary evaporator, the product
was dissolved in 10 ml distilled water. Salts and free
lysine and PL were removed by chromatography of 3 m1
portions on a column of Sephadex G-10 (2.6 x 34 cm,

Vo = 63 ml). The void volume fractions were pooled,

183

dried under reduced pressure at 50°C and dissolved in
10 ml absolute ethanol. Pale yellow crystals of s-pyri-
doxyl-L-lysine slowly formed upon storage at 2°C, as
described by Dempsey and Snell (1963).
Preparation of Phosphopyridoxyl-
Bovine SerumiAlbumin

Phosphopyridoxyl-bovine serum albumin was pre-
pared by procedure 2 of Dempsey and Christensen (1962).
Fourteen grams of bovine serum albumin (essentially
fatty acid-free; Sigma Chemical Co.) was dissolved in
560 ml of 0.1 M potassium phOSphate, pH 7.5. Pyridoxal
phosphate monohydrate (2.4 mg; ICN Pharmaceuticals, Inc.)
was dissolved in a small volume of water and added to
the albumin solution in an ice bath. The mixture was
stirred at 1 to 2°C for 1 hour. Sodium borohydride was
added as a 1% aqueous solution by injection under the
surface of the solution. To ensure complete reduction,
a 15 fold molar excess of NaBH4 was added relative to
PLP, with small additions to minimize foaming. The pH
of the solution was maintained between 7.5 and 7.7
during the reduction by the periodic addition of 6 N
sodium hydroxide.

After storage overnight at 2°C, the protein
solution was transferred to several dialysis bags of
1 1/8 inch diameter (Fisher Scientific Co.) and dialyzed

in a five-gallon stainless steel container against

184

distilled water at 2°C. Dialysis continued for 1 week,
with six changes of the cold distilled water. The
dialyzed protein was freeze dried in a shallow tray in
a Virtis Model FFD 42WS Freeze-Dryer. Drying was per-
formed at a platen temperature of 110°F to an absolute
pressure of 5 um Hg.

High Performance Liquid Chromatggraphic
Determination ofVitaminSE

Materials

 

l. Extraction buffer. Potassium acetate
(39.26 g) was dissolved in about 1800 ml distilled water,
adjusted to pH 4.5 with glacial acetic acid, and diluted
to 2 1, providing a 0.2 M acetate buffer.

2. Trichloroacetic acid, 20% (w/v), was prepared
by dissolving 20 g crystalline TCA in distilled water
and diluting to 100 ml. TCA solutions were prepared
fresh daily prior to analysis.

3. HPLC mobile phase buffer, 0.033 M potassium
phosphate, pH 2.2, was prepared by adding 4.6 ml concen-
trated phosphoric acid to about 1800 ml distilled water,
adjusting the pH with 6 N potassium hydroxide, and dilut-
ing to 2 1. Prior to use, the buffer was filtered
through a 0.45 pm filter (Gelman Instrument Co.) under
vacuum in order to remove small particulate matter. This

step provided satisfactory degassing of the solution.

185

4. Potato acid phosphatase (2 U/mg) was obtained
from Sigma Chemical Co. A 1.0 mg/ml solution was pre-
pared (2 U/ml) by dissolving 50 mg of the dry enzyme
preparation in 50 m1 of cold 0.2 M potassium acetate,
pH 4.5. This solution was prepared immediately prior
to use.

5. Vitamin B6 standard stock mixture was pre-
pared at a concentration of 10 ug free base/ml with
respect to PM, PN, and PL by dissolving 14.2 mg pyri-
doxamine dihydrochloride (ICN Pharmaceuticals, Inc.),
12.2 mg pyridoxal hydrochloride (ICN Pharmaceuticals,
Inc.), and 12.2 mg pyridoxine hydrochloride (Sigma
Chemical Co.) in l l 0.2 M potassium acetate, pH 4.5.
Working standards were prepared over the range of 0.25
to 2.5 pg free base/ml by transferring appropriate
volumes of the stock to 25 ml volumetric flasks, adding
8.33 ml 20% TCA, and diluting to 25 ml with 0.2 M
potassium acetate, pH 4.5.

6. Recovery standard mixtures were prepared at
a concentration of 30 ug free base/ml for PM and PN,
and 41.1 ug/ml for PLP (equivalent to 30 ug PL/ml).
Pyridoxal phosphate monohydrate (22.2 mg; ICN Pharma-
ceuticals, Inc.), Pyridoxine hydrochloride (18.2 mg),
and pyridoxamine dihydrochloride (21.5 mg) were dissolved

in 500 ml 0.2 M potassium acetate, pH 4.5.

186

Sample Extraction

 

Finely ground model system (about 3 g) was
accurately weighed and transferred into a 25 ml volumetric
flask. Fifteen ml of 0.2 M potassium acetate, pH 4.5,
was added and the sample thoroughly agitated to ensure
complete wetting. For recovery samples, 1.0 ml of the
recovery mixture was added at this point. To promote
the extraction, all flasks were sonicated 30 minutes
in a 65 watt ultrasonic cleaning bath (Fisher Scientific
Co.) at room temperature. After sonication, the contents
of each flask were diluted to 25 ml with the pH 4.5 acetate
buffer. Insoluble solids were removed by centrifugation
of the samples at 6000 rpm (4340 x g max) in a Beckman
Model J-21C centrifuge at 25°C for 20 minutes (JA-20
rotor).

Ten ml of each of the supernatants was transferred
into a clean 40 ml screw cap centrifuge tube. Phosphory-
lated forms of the vitamin were hydrolyzed with potato
acid phosphatase (Takanashi, EE.El" 1970). Two ml of
the 1.0 mg/ml enzyme solution was added to each 10 ml
aliquot of the sample supernatants, followed by incu-
bation in a 37°C water bath for 1 hour. Preliminary
studies showed that this enzyme treatment quantitatively
hydrolyzed PLP in the extracts. The reaction was termi-
nated and protein precipitated by the addition of 4 m1

of 20% TCA, followed by incubation in a 50°C water bath

187

for 15 minutes. After cooling to room temperature, the
samples were recentrifuged at 6000 rpm for 20 minutes
at 25°C. Aliquots (50 pl) of each supernatant were taken

for direct injection into the chromatograph.

Chromatographic Analysis

 

Chromatographic separations were performed using

a pBondapak C ‘reverse phase column (4mm ID x 30 cm;

18
Waters Associates). The mobile phase was pumped for at
least 1 hour prior to analysis to maximize the stability
of the system. Chromatography of the standards and
samples was carried out at a flow rate of 2 ml/minute.
The absorbance of the column effluents was monitored at
280 nm, 0.02 AUFS. The fluorescence of the effluents
was monitored using an Aminco Fluoro Monitor equipped
with a Germicidal lamp (General Electric Corp.), a
295 nm interference excitation filter (American Instru-
ment Co.), and a Corning 7-51 (370 nm narrow pass)
emission filter. Fluorescence spectra of the B6 vitamers
in the mobile phase buffer exhibited excitation and
emission maxima at approximately 290 and 395 nm,
respectively.

A standard curve was prepared by injecting
50 pl of each calibration standard mixture and plotting
peak height, either relative fluorescence or absorbance,

against micrograms for each vitamer. The concentration

188

of the 36 vitamers in the sample extracts was determined
by comparison of peak height with the appropriate standard
curve.

Following analysis, the HPLC column was washed
with approximately 20 ml distilled water, then 30 ml
methanol. The column was stored in a 33% methanol:67%
water mixture.

High Performance Liquid Chromatographic
Determination of e-PyridoxyleL-lySine

Model System Protein Extraction
and HydrOIysis

 

Model systems (5 g) were finely powdered and
suspended in 15 ml 0.05 M KCl-HCl, pH 2.5. The mixtures
were sonicated 30 minutes and centrifuged at 6000 rpm in
a Beckman Model J-21C centrifuge with a JA-21 rotor.

The supernatants were dialyzed in 0.39 inch diameter
tubing (Fisher Scientific Co.) against 4 1 of 0.05 M
KCl-HCl, pH 2.5, with three changes of the buffer over
a 4-day period. After dialysis, the protein concentra-
tions were determined by the procedure of Lowry, 2353;.
(1951). Solutions of bovine serum albumin and phos-
phopyridoxyl-bovine serum albumin were prepared in the
same pH 2.5 buffer at a concentration of 10 mg/ml.

A volume equivalent to 10 mg of protein of each
sample was pipetted into duplicate 10 ml glass ampules,
adjusted to 6 N in constant boiling HCl, and diluted to

5 m1 total volume. The ampules were frozen in a dry

189

ice-ethanol mixture, followed by slow thawing under
vacuum to degas the solutions. The ampules were then
refrozen and sealed under vacuum.

.The proteins were hydrolyzed by incubation of
the ampules in a 105°C oven for 24 hours. This procedure
also hydrolyzed the phosphate esters of the bound 36
compounds. Upon cooling, the hydrolysates were quanti-
tatively transferred to 100 ml round bottom flasks and
evaporated to dryness in a rotary evaporator at approxi-
mately 50°C. Small portions of water were added to each
flask several times, followed by redrying, to ensure
complete removal of the HCl. Each hydrolysate was dis-
solved in 10 ml 0.2 M potassium acetate, pH 4.5, and
filtered through a 5 pm membrane (Type HA, Millipore
Corporation) prior to chromatography.

Chromatographic Analysis of
e-nyidoxyllysine

 

 

The chromatographic conditions for the determi-
nation of pyridoxyllysine were identical to those
described for the HPLC determination of vitamin B6 in
model system extracts. The fluorescence of the column
effluents was monitored with an Aminco Fluoro Monitor
using a Germicidal lamp (General Electric Corp.), a
295 nm interference excitation filter (American Instru-
ment Co.), and a Corning 7-51 (370 nm narrow pass)

emission filter.

190

Crystalline e-pyridoxyl-L-lysine was dissolved
in 0.2 M potassium acetate, pH 4.5, and its concentration
was determined using the molar absorptivity at 323 nm of

1cm.1 (Dempsey and Christensen, 1962). Injections

5800 M‘
of 5 to 40 p1 portions of the standard solution provided
a standard curve range of 0.087 to 0.689 pg e-pyri-
doxyllysine. A standard curve was prepared by plotting
peak height against pg bound pyridoxal, determined using
the gravimetric factor of 0.563 pg bound PL/pg pyri-
doxyllysine.

High Performance Liquid Chromatographic
Determination of 4-Pyridoxic Acid

 

 

 

Materials

 

1. HPLC mobile phase buffer was prepared by
adding 100 m1 anhydrous methanol (Burdick and Jackson)
and 4.6 ml concentrated phosphoric acid to approximately
1800 ml distilled water. The pH was adjusted to 2.2
with 6 N potassium hydroxide, and the solution was
diluted to 2 l with distilled water. This provided
a final concentration of 0.033 M phosphate and 5% (v/v)
methanol. The mobile phase buffer was filtered through
a 0.45 pm filter (Gelman Instrument Co.) prior to use.

2. 4-Pyridoxic acid standards were prepared as
follows. Crystalline 4-pyridoxic acid (4-PA) was
obtained from Sigma Chemical Co. A working stock

solution was prepared immediately before use by

191

dissolving 20 mg 4-PA in l 1 0.2 M potassium acetate,

pH 4.5. Calibration standards were prepared over the
range of 0.05 to 6.0 pg/ml. To 25 ml volumetric flasks
containing 5 ml of the pH 4.5 acetate buffer, appropriate
volumes of the 4-PA stock were added, followed by 8.33 ml
20% TCA. The contents were diluted to 25 ml with the
acetate buffer.

Chromatographic Analysis of
Z-Pyridoxic Acid

 

 

The extraction and enzymatic hydrolysis of the
model systems were performed as described for the HPLC
analysis of vitamin 86' Therefore, any 4-pyridoxic acid
phosphate formed in PLP-fortified systems would be
determined as 4-PA.

All extracts and standard solutions were held
in an ice bath until chromatographed. Chromatographic
conditions were identical to those employed for the HPLC
analysis of vitamin B6, with the exception of 5% methanol
added to the mobile phase to shorten the retention time
of 4-PA. The fluorescence of the column effluents was
monitored using an Aminco Fluoro Monitor with a Corning
7-60 (355 nm narrow pass) excitation.fi1ter and a Wratten
47B (436 nm narrow pass) emission filter. The concen-
tration of 4-PA in model system extracts was determined
from a standard curve of fluorescence peak height against

pg 4—PA injected.

192

Discontinuous Polyacrylamide Gel
Electrophoresis

 

 

Gel electrophoresis of B-lactoglobulin was per-
formed by a modification of the discontinuous method of
Melachouris (1969). The running gel solution was pre-
pared by dissolving 22.5 g Cyanogum 41 (E-C Apparatus
Corp.) in 0.380 M Tris-HCl buffer, pH 8.9, and diluting
to 250 ml, giving a total gel concentration of 9%. To
this gel solution, 0.25 ml N, N, N',N'-tetramethyl-
ethylenediamine was added. Polymerization of the gel
was initiated by the addition of 2 ml of fresh 10%
aqueous ammonium persulfate solution to 190 ml of the
gel solution. The mixture was quickly poured into a
vertical slab gel apparatus (E-C Apparatus Corp.) in its
horizontal position, and the sample slot former positioned.
Thirty minutes were allowed for gel polymerization. No
spacer gel was employed.

Excess gel was cut away from the sample slot
former. The electrophoresis apparatus was then placed
in its vertical position. Electrophoresis chamber
buffer, 0.046 M Tris-glycine, pH 8.3, was gently loaded
into the upper and lower buffer chambers of the apparatus.
After filling the buffer chambers, the sample slot former
was carefully removed.

The protein samples were dissolved at a concen-
tration of 5 mg/ml in 0.062 M Tris-HCl, pH 6.7. Sucrose

was added to the sample solution to provide a greater

193

density than the chamber buffer, thus facilitating
layering of the protein solutions in the sample slots.
Bromphenol blue was added to the protein solutions to
serve as a tracking dye.

Electrophoresis was performed at approximately
17°C at a constant voltage of 130 V for two and one-half
hours (Heathkit Power Supply). During this time, the
tracking dye migrated 10 cm.

Following electrophoresis, the gels were care-
fully removed from the apparatus and fixed overnight by
soaking in 15% (w/v) TCA. The gels were stained with a
solution of 0.1% Coomassie brilliant blue in 13.75% TCA
for 2 hours. Destaining was accomplished by several

washes with 15% TCA.

194

Darin-d from Journal of Food Science.
0 1977 by Institute of Food Technologists.

J. F. GREGORYandJ. R. KIRK
Dept. of Food Science a Human Nutrition
Michigan State University, East Lansing, All 48824

IMPROVED CHROMATOGRAPHIC SEPARATION AND FLUOROMETRIC DETERMINATION
OF VITAMIN B. COMPOUNDS IN FOODS

 

 

ABSTRACT

Several fluorometric promdures have been previously developed for the
determination of vitamin B. compounds in foods. These have mainly
been based on the separation of pyridoxal (PAL), pyridoxine (PIN),
and pyridoxamine (PAM) by column chromatography on Dowex AG
SON-X8, chemical conversion of PIN and PAM to PAL, and formation
(1 the 4-pyridoxic acid lactone fluorophore by reaction with KCN.
M procedures were limited by a cumbersome chromatographic pro-
adure utilizing boiling potassium acetate buffers, and interfering com-
pounds in the PAL fraction often precluded accurate measurement. In
this study, Maillard browning pigments resulting from the autoclave
extraction procedure were shown to interfere by reacting with KCN to
form highly fluorescent products. To alleviate the interference problem,
an alternative ion exchange chromatographic method was developed
which eliminated the KCN-reactive browning pigments from the PAL
fraction. After the ion exchange clean up steps, PAL was determined
directly by reaction with KCN, whereas PIN and PAM were first con-
verted to PAL by reaction with MnO, and sodium glyoxylate, respec-
tively. The effectiveness of the procedure was demonstrated on casein
it the presence or absence of glucose, and on selected food samples.

 

INTRODUCTIW

THE DEVELOPMENT of an adequate chemical procedure for
the determination of biologically active forms of vitamin B. in
foods has been a complex and enigmatic problem. Basic stud-
ies by Bonavita (I960), Toepfer et al. ( 1961), and Polansky et
a). (1964) have demonstrated the feasibility of fluorometric
measurement of pyridoxal (PAL), pyridoxamine (PAM), and
pyrodixine (PIN) by conversion to PAL, and reaction with
KCN, forming the fluorophore, 4opyridoxic acid lactone. Vari-
ous fluorometric methods have been applied to vitamin B.
compounds in biological materials (Fujita et al., 1955; Con-
tractor and Shane, I968; Loo and Badger, I969;Takanashi et
al., l970a, b; FiedIerova and Davidek, I974; Chin, I975). The
recent results of Chin (1975) have suggested that interfering
compounds may be present in the PAL fraction after column
chromatographic separation of the B. analogs by the pro-
cedure of Toepfer and Lehman (1961). This possibility has
not been previously investigated.

In this study, certain Maillard browning pigments formed
during the standard autoclave extraction procedure were
shown to react with KCN to form highly fluorescent products.
Since these pigments appeared in the PAL chromatographic
fraction using the ion exchange method of Toepfer and Leh-
mann (I961), erroneously high PAL values would be obtained
with all fluorometric assay procedures based on autoclave ex-
traction and this chromatographic separation.

To alleviate the problem, various ion exchange chromato-
yaphic procedures were ex‘amined to develop a method which
would eliminate these interfering compounds from the PAL
fraction. The fluorometric procedure of Chin (I975) was mod-
ified to be compatible with this chromatographic method, pro-
viding a more accuratechemical assay for B. compounds in
foods.

MATERIALS & METHODS

1. Vitamin B. stock standards were individually prepared at 10 ug
free base per ml in IN HCI (Toepfer and Polansky, 1970) using
pyridoxal-NC) and pyridoxamine-2 HCL (ICN Pharmaceuticals,
Inc.) and pyridoxine-NC) (Sigma Chemical Co.).

Calibration standards were prepared over the range of 0.02 —0.10
1‘ free base per ml by dilution of the stock standards with the
appropriate buffer.

2. Extraction and hydrolysis of the samples were performed by the
procedure of Toepfer and Polansky (1970). Finely ground sample
(0.5—2.0g) was weighed into a 250 ml Erlenmeyer flask. To plant
products, 180 ml 0.44N HCI was added and autoclaved for 2 hr at
121°C; animal products required [80 ml 0.055N HCl and auto-
claving for 5 hr at 121°C. Recovery standards were prepared by
adding 1 ml of stock solution (10 ug/ml) of each form of vitamin
B, to an aliquot of the sample prior to autoclaving. After cooling
to ambient temperature, the pH of the samples was adjusted to 4.5
with 6N KOH. Each sample was diluted to a final volume of 250
ml with distilled water and filtered through Whatman No. 42 filter
paper. These filtrates were applied to the ion exchange columns.

3. Chromatographic procedures—All ion exchange fractionation: of
the food extracts were performed on Dowex AG SOW-XB,
100—200 mesh, K*-form (Bio-Rad Laboratories). Separation of
PAL was done with a 0.7 x 20 cm column packed with 5.5 ml of
resin, while PAH and PIN were eluted in a single fraction from a
separate 1.8 x 40 cm column packed with 30 ml of resin. The
surfaces of the resin beds of the large columns were protected by
discs of Whatman No. 1 filter paper. The elution procedures for
PAL, PIN and PAM are shown in Table I. The PAL fraction was
eluted into a 25 ml volumetric flask, and diluted to 25 ml with the
pH 6.30 phosphate buffer. The PAM and PIN fraction was eluted
from the larger column into a 200 ml volumetric flask. All separa~
tions were performed at room temperature. After three to four
separations, the column resins were regenerated accordiru to the
procedure of Toepfer and Polansky (I970).

4. Conversion of the B. analogs to 4-pyridoxic acid lactone was
based on the procedure of Chin (1975). Reagents: 0.5N sodium
glyoxylate (Sigma Chemical Co.), manganese dioxide prepared ac-
cording to Polansky et a1. (1964), and 1M KCN (Mallinckrodt
Chemical Works).

Pyrfloxfl. Two ml of PAL calibration standard or column eluate
was pipetted into each of duplicate 15 ml screw cap test tubes labeled
A and B (A - sample, B - blank). Each was diluted by the addition of 8
ml 0.4“ K, "PO, pH 7.5. For conversion to 4-pyridoxic acid lactone,
0.4 ml 1M KCN was added to A, white 0.4 ml distilled water was added
to B. with incubation 2 hr at 50°C.

he. Duplicate 2-ml aliquots of PAH calibration standard
or eluate was pipetted into IS ml screw cap test tubes, labeled as above.
Each was diluted by the addition of 8 ml 0.4“ K, HPO. pH 7.5. For
conversion to pyridoxal, 0.4 ml 0.5M sodium glyoxylate was added to
A. 0.4 ml distilled water was added to B. Following the addition of
glyoxylate or water the samples were incubated for [5 min in a boilim
water bath. After cooling the tubes to room temperature in running
water, 0.4 ml 1H KCN was added to each tube followed by incubation
2 hr at 50°C to convert PAL to 4-pyridoxic acid lactone.

Pyriloxtln. A 20-ml aliquot of sample eluate or calibration standard
was transferred to separate 125 ml Erlenmeyer flasks and the pH of

Val. 42, No. 4 (1977) - JOURNAL OF FOOD SCIENCE — 1073

195

Table I—Ion exchange column elation procedures

 

Buffer type

 

Potnei urn acetate

Potassi um phosphate

 

PAL Column
(1) Equilibration
(2) Sample applied (10—15 ml) _
(3) Wash -
(4) PAL elution

(5) Wash before "equilibration

PIN eel! PAM Column
(1) Ecullibration
I2) Sample applied (SO-100 ml) —
(3) Wash
(4) PIN and PAM elution

10 ml 0.2M pH 4.50 —

18 ml 0.1M pH 5.70
18 ml 0.15M pH 6.30
(dilute collected
traction to 26 ml
using this buffer)

35 ml 0.5M DH 8.00 -

100 ml 0.2M pH 4.50 —

100 ml 0.15M OH 6.20 -
200 ml 0.5M pH 8.00 —

 

each flask adjusted to 5.5 with IN HCI. Oxidation of PIN to PAL was
carried out by adding 0.1g MnO, and gently mixing at room tempera-
ture for 30 min. Addition of MnO, was omitted for the PIN blanks.
The pH of the samples was then adjusted to 7.2 with 0.4M K, HPO,
(about 14 ml), followed by filtration using quantitative technique

Table 2- Typical calibration curses for fluorometric determinations.
Free base concentration 0.02—0. 10 m/ml

 

Clone
(Floor. urinal we...

 

 

 

 

 

 

 

 

 

 

 

Intercept
I. analog (Floor. units) “I’M" coefficient
PAL 233 673.8 +9977
PIN 2.3 512.5 £989
PAM 11.1 572.5 +9984
no II
‘Wk 1"
I“
ll) “I. II ‘ [on 13.0) ”V“
w “Cl"fl!‘
1 (Le) up;
I” w
7] I ”I
.3" ‘ I‘m “.4. 7“ PI‘
“3",“ n ‘ "m ' m I
“(m 0 I“ C
”I WM! I 0
Ilium “I“ ° l
“mu?" II A

 

 

 

Fig. f-Continuous flow analysis system for determination of vitamin
B, as 4~pynaoxic acid lactone (modified from Chin, 1975). Figures in
parentheses represent flow rate in nil/min.

7074 — JOURNAL OF FOOD SCIENCE — Vol. 42, No. 4 (7977)

through Whatman No. 42 filter paper directly into a 100 ml volumetric

flask. The filtrates were diluted to l00 ml with distilled water. Ten ml
aliquots were then transferred to screw cap test tubes. 0.4 ml 1M KCN
added, and the tubes were held at 50°C for 2 hr to convert the PAL to
4-pyridoxic acid lactone.

5. Fluorometric procedures were based on those of Chin (1975) for
automated analysis, with the following modifications:

a. The higher ionic strength buffers employed necessitated
the use of wash buffers of similar composition and ionic
strength. The following buffers were used as wash solu-
tions between samples: PAL: l part 0.l5M phosphate,
pH 6.30; 4 parts 0.4M phosphate, pH 7.4. PIN: I part
part 0.5M acetate, pH 8.00; 4 parts 0.4M phosphate, pH
7.4.

b. Fluorometric apertures were set at a reference value of 2
and a sample value of 4.

c. The flow pattern through the Technicon AutoAnalyzer
was improved by the addition of 2 ml of 30% Brij 35
(Fisher Scientific Co.) per liter of 0.4M Na, (‘0,.

d. Due to the differences in sample buffers and, hence,
buffer capacity, slight adjustments in fluorometer cali-
bration were required between determinations of the
three forms of the vitamin. Typical calibration curve
data appear in Table 2.

6. Studies on interfering compounds.

Various food samples and a model system (described below) were
extracted, chromatographed according to Toepfer and Lehmann
(I961), and examined for the presence of interfering compounds. The
PAL ion exchange column effluents for sample extracts and PAL stand-
ards were evaporated to dryness under vacuum at 50°C in a rotary
evaporator and reconstituted with water to give a tenfold concentra-
tion. Aliquots (2 ml) of this concentrate were treated with 0.1 ml 1M
KCN (0.) ml distilled water for blanks) and held at 50°C for 2 hr.
These samples were examined qualitatively for fluorescing compounds
by thin-layer chromatography. Seventy~fivc microliters of each sample
were spotted on a silica gel G precoated plate (Analtech, Inc.) and
developed with a 50:50 mixture (v/v) of methanol and water. Fluores-
cence was observed upon excitation at 366 nm. Differences in the
pattern of fluorescing compounds between KCN-treated and blank sam-
ples were indicative of the presence of KCN-reactive interfering sub-
stances.

For evaluation of the effectiveness of various column chromatog-
raphy procedures in removing interfering compounds from the PAL
fraction, extracts from autoclaved model food systems were employed.
The model systems consisted of 5g of “vitamin free" test casein
(General Biochemicals. Inc.), either with or Without 5g of Dglucose
(‘Baker Analyzed Reagent', LT. Baker Chemical Co.) to induce Maillard
browning during autoclaving. The ratio of apparent ug of PA L/g casein

196

lee J—Efflciemy of ion exchange calm meal of interfering eonpouna from PAL fraction“,

DETERMINING VITAMIN B. COMPOUNDS IN FOODS . . .

 

 

 

W us “Us Casein
Preeedun Casein Only Casein + Glucose Ratio"
1. Boiling acetate buffers 1.20 s 0.08 2.11 s 0.08 0.57 a 0.05
(Toepfer and Lahmann.1001l
2. Room temperature acetate butters
PAL eluant - 0.17M acetate pH 0.26
Proceeding mesh: 0.1M acetate pH 5.70
Volume
100 ml 0.91 t 0.08 1.20 2 0.04 0.72 t 0.07
150ml 1m10.00 1.1330." 08810.14
m0ml 05010.04 086:0.01 0380.06
3. Room temperature, mined. 1.00 a 0.21 1.44 s 0.00 0.15 s 0.10
Wash-100 ml 0.01M phosphate pH 5.70
PAL fluent-100 ml 0.17M acetate pH 0.25
4. Room temperature phosphate buffers 0.81 t 0.10 0.78 2 one 1.04 1 0.23

Wash-1m ml 0.1M phosphate pH 5.70
PAL eluant-tw ml 0.16M phosphate pH e30

 

‘ Thlrty ml reeln bed columns. duplicate separations and PAL determinations. Values are mean a standard deviation.
5 A retlo of one would lndleate complete removal of lntaderlng compounds.

in the casein plus glucose system to ug PAL/g casein in the casein
without glucose system was used as a quantitative indicator of column
cleanup efficiency

RESULTS t. DISCUSSION

STUDIES undertaken to investigate the possibility of inter-
fering compounds in the PAL fraction following ion exchange
cinematography confirmed the presence of non-PAL, KCN-
reactive substances. Qualitative differences were observed in
the thin-layer chromatographic pattern of KCN treated PAL
fractions of food extracts and their corresponding blanks. As
this qualitative difference was also observed with model sys-
tems of casein plus glucose, but not casein alone (Fig. 2, plate
I), it was felt that the interfering compounds could be
Maillard browning pigments formed during the autoclave ex-
traction procedure. Analysis of these compounds by gel filtra-
tion chromatography and ultraviolet spectrophotometry sug-
gested a strong similarity to the Maillard browning pigments
previously characterized (Clark and Tannenbaum. I970).
Therefore, the ratio of PAL/g casein values of model systems
with or without glucose was employed as a quantitative indica-
tor of cleanup efficiency in the development of an ion ex-

change chromatographic procedure to eliminate these interfer-
ing compounds. After evaluating several ion exchange systems,
a room temperature potassium phosphate wash and elution
procedure was developed which provided adequate cleanup of
the PAL fraction. As shown in Table 3, interference was found
with all other procedures tested, the most noted example
being the boiling acetate elution method. The absence of any
qualitative differences in the thin-layer chromatographic pat-
terns exhibited by the casein and casein plus glucose samples
shown on Figure 2, plate 2, was further evidence for removal
of interfering compounds from the PAL fraction by the phos-
phate elution system. Similar thin-layer chromatographic re-
sults were observed with all food extracts tested.

The results of Chin (I975) for the determinations of PIN
and PAM suggested little, if any, problem with interfering
compounds. Studies utilizing casein and casein plus glucose
model systems confirmed this observation. Values of apparent
PIN or PAM/g casein were not affected by the presence of
glucose in the sampling indicating that browning pigments did
not interfere with these determinations. Such specificity may
be attributed to the selective motions for converting PIN and
PAM to PAL, in addition to blank determinations which, in

 

 

Plate 1 Plate 2

"’ L

.. 4» 1.0 u- .

i- 6 @ 4» 0.8 «l- ..
a O a ] Fig. 2-1'Mn layer MW of PAL
0 l" 0'6 " O I' calm dimes. Exelmion-mnm.n'sml
,_ 0 Oil 4 w analysis. Extent ofhetehingproporrionalto
‘ flwreeeence. Designetion:A -KCN veered;
I. .L 01 .. 4. 8 - Blank; PICL - 4-pyridoxicacid lactone.
Here 1, (selling acetate and: and elation.
0 O O . C 4* 0.0 4L J» "I" 2, room WNW. ”Wm. M

anddution.

 

 

 

 

 

 

ma 5...? Wm mama
(1m

6m»? PN.

am

Vol. 42, No. 4 (1977) — JOURNAL OF FOOD SCIENCE — 1075

197

Table 4-Recovery of 8. stendrds from ion exchange columnshb

 

 

.a m up Applied Recovery (S)
PAL 5.0 95.2 s 3.1
PIN 10.0 103.2 2 1.9
PAM 10.0 “3.4 s 4.3

 

‘ Vitamin 5. solutions were prepared in 0.2M pH 4.50 potassium
acetate buffer.
5 Mean and standard deviation of four separations

Table 5- Vitamin 8, content of ales-ted food samplesl

 

 

 

ug free lime/g
84mph PAL PIN PAM
Corn flakes 7.3 2 0.9 17.8 a 2.0 8.6 s 3.1
Infant formula 1.8 s 0.6 0.6 s 0.5 11.0 t 2.4
(powdered)
Cured ham 1.7 s 0.5 2.5 s 1.4 9.2 s 1.4

‘ Mean and standard deviation of triplicate determinations

Table 6—Recovery a! PAL, PAH, and PIN added to food sen'iplesP-b

 

 

 

Percentage recovery
Sample PAL PIN PAM
Corn flakes 78.3 s 5.8 “1.0 t 5.7 105.0 s 13.2
infant formula 72.9 s 3.6 123.3 s 5.5 110.0 a 14.1
(poweredl
Cured ham 54.4 a 4.4 100.0 s 10.0 115.0 s 21.2

 

‘ Mean and standard deviation of triplicate determinations
10 up of each analog added prior to autoclaving.

contrast to the PAL procedure, provide a measure of extrane-
ous KCN-teactive compounds.

Whereas the elution of PIN and PAM was readily accom-
plished from the PAL column using the room temperature
phosphate system, several problems in the determination of
PIN precluded the use of a single column for separation and
purification of the other 5. analogs. First, the oxidation of
PIN to PAL by Mn01 was found to be inhibited by the phos-
phate buffer. This could possibly be alleviated by using anoth-
er oxidizing agent. Second, attempts to reequilibrate the
column with 0.15M potassium acetate, pH 6.30, caused excep-
tionally high losses of PIN from the column. Therefore, the
two-column procedure was adopted. For convenience, the
PAL elution procedure was scaled down to permit the use of a
smaller column. The cleanup efficiency of the column was not
affected by this change. The recovery data (Table 4) indicate
quantitative and reproducible elution of the three 8,, analogs
from the Dowex AG SOW-XS columns. Studies concerning the
oxidation of PIN to PAL in a phosphate buffer are in progress
and will be completed in the near future.

Selected food samples were assayed for total 3‘ com-
pounds by the fluorometric procedure (Table 5). For all forms
of the vitamin in these individual samples, the chemically de-
termined values were considerably higher than published re-
sults based on microbiological assay (Polansky and Toepfer,
1969). By direct comparison of chemical and microbiological
procedures, (‘hin (1975) observed similar large differences in

1076 - JOURNAL OF FOOD SCIENCE — Vol. 42, No. 4 (1977)

apparent vitamin B. in foods. Using a chemical assay based on
totally different sample cleanup and fluorophore formation
procedures, Kraut and Imhoff (1967) also obtained results
which were significantly greater than published microbiolog-
ical values. These independent results suggest that either the
chemical assay may be measuring forms of the vitamin which
are not available microbiologically or that factors in the food
extract may inhibit microbial growth, thereby affecting the
accuracy of the microbiological procedure. Further research
will be required to resolve this difference.

The recovery of 86 compounds added to food samples
prior to autoclaving was observed to be somewhat variable
(Table 6), with net recovery and precision being dependent of
the food sample. The variability could not be satisfactorily
explained, although certain heat-induced reactions of the
added B. vitamins with various food components may be in-
volved. Similar results were observed by Chin (1°75). Because
of this variability, duplicate recovery samples should be run
with each product assayed to assure accuracy.

The precision of this method is slightly less than that of
other recently published semi-automated methods of vitamin
analysis for B complex vitamins, which is probably due to the
complexity of the analysis. However, the sensitivity and specif-
icity of the procedure have been demonstrated. Therefore, this
assay method provides a satisfactory means for the chemical
complexity of the analysis. However, the sensitivity and speci-
ficity of the procedure have been demonstrated. Therefore,
this assay method provides a satisfactory means for the
chemical determination of vitamin B. in foods and biological
materials.

REFERENCES

Bonavita, V. 1960. The reaction of pyridoxal phosphate with cyanide
and its analytical use. Arch. Biochem. Biophys. 55: 366.

Chin, Y.P. 1975. Chromatographic separation and fluorometric deter-
mination of pyridoxal. pyridoxamine, and pyridoxine in food sys-
tem. M.S. thesis. Mlchlgan State University. East Lansing. Mich.

Clerk. A.V. and Tannenbaum. 8.11. 1970. Isolation and characterisation
of protein-carbonyl browning systems. Isolation. purification. and
properties. J. Agr. Food Chem. 15: 591.

Contractor, 8.1". and Shane, B. 1965. Estimation of vitamin 5. corn-
pounds in human blood and urine. Clin. Chirn. Acts 21: 71.

Pledlarova. V. and Davidek, J. 1974. Fluorometric determination of
pyridoxal in dried milk. 2. Lebenam. Untera-P‘orach. 155: 277.

Nita, A.. Funita. D. and Fujino, K. 1955. Fluorometric determination
of vitamin B . 3. Fractional determination of pyridoxal and
4-pyridoxlc act‘s. J. Vltarninol. 1: 279.

Hennessy, D..I.. Steinberg. A.M., Wilson. 6.5. and Keaveney. W.P. 1960.
Fluorometric detenninatlon of added pyridoxine in enriched white
flour and bread baked from it. J. Assoc. Offic. Agr. Chemists 45:

766.

Kraut, H. and Imhoff, U. 1967. Porschungber. [Andes Nordrheirn-
Weatfallen. I533.

Loo, Y.H. and Badger. L. 1959. Spectroflurometrlc my of vitamin B.
analogs in brain tissue. J. Neurochem. 16: 501.

Naaultawa, K.. Nakama, A. and Monaka. H. 1971. Differential determi-
nation of pyridoxal-5'-phosphate and pyridoxamine-b'ohonhate in
biological materials. Vitamins 44: 527.

Polansky. M.S.. Camarra. 1LT. and Toepfer, 3.7!. 1964. Pyridoxine
determined fluorometrically as pyridoxal cyanide compound. J.
Amoc. Offic. Agr. Chemists 47: 527.

Polansky, 54.8. and Toepfer, E.\V. 1969. Vitamin 5. components in
some meats, fish, dairy products and commercial infant formulas. J.
An. Food Chem. 17: 1394.

Takanashi, S., Matsunaga, I. and Tarnura. 2. 1970s. Fluorometric de-
termination of pyridoxal-5'-phosphate. Biochem. Biophys. Acta
804: 765.

Takanashi. S., Matsunaga. 1. and Tamura. Z. 1970b. Fluorometric de-
termination of pyridoxal and its 5'-phosphate in biological materi-
als. J. Vitamlnol. 16: 132.

Toepfer, ELW. and Lehmann, J. 1961. Procedure for chromatographic
separation and microbiological amay of pyridoxine, pyridoxal, and
pyridoxamine In food extracts. J. Aeaoc. Offic. Agr. Chemists 44:
426

Toepfer, E.W.. Polansky. M.M. and flewston. 8.14. 1961. Fluorometric
determination of pyridoxamine by conversion to pyridoxal cyanide
compound. Anal. Biochem. 2: 463.

Toepfer, E.W. and M.M. Polansky, 1970. Microbiological any of vita-

min B. and its components. J. Assoc. Offic. Agr. Chemists 53: 546.

Ma received 6/29/76: revised 11/30/76; accepted 12/5l76.

Presented at the 36th Annual Meeting of the Institute of Food
Technologists, Anaheim. CA, June 6-9, 1976.
Michigan Agricultural Experiment Station Journal Article Number
7575.

 

SELECTED BIBLIOGRAPHY

SELECTED BIBLIOGRAPHY

Abbott, E. H., and A. E. Martell. 1970. Pyridoxine and
pyridoxal analogs. XIII. A nuclear magnetic
resonance study of the condensation of poly-
functional amino acids with pyridoxal. J. Am.
Chem. Soc. 9;:1754.

Ahrens, H., and W. Korytnyk. 1970. Thin-layer chroma-
tography and thin-layer electrophoresis of vitamin
Bs. In: Methods in Enzymology, Vol. 18A.
"Vitamins and Coenzymes." Academic Press, New
York, p. 489.

Anderson, B. B., P. A. Newmark, M. Rawlins, and R. Green.
1974. Plasma binding of vitamin 36 compounds.
Nature 250:502.

Anderson, B. M., C. D. Anderson, and J. E. Churchich.
1966. Inhibition of glutamic dehydrogenase by
pyridoxal 5'-phosphate. Biochem. Ԥ:2893.

Anderson, J. A., H. W. Chang, and C. J. Grandjean. 1971.
Nature of the binding site of pyridoxal 5'-
phosphate to serum albumin. Biochem. 12:2408.

Anderson, R. H., D. L. Maxwell, A. E. Mulley, and C. W.
Fritsch. 1976. Effects of processing and
storage on micronutrients in breakfast cereals.
Food Technol. 32:110.

Arrio-Dupont, M. 1970. Etude par fluorescence de bases
de Schiff du pyridoxal, comparaison avec la
L-aspartate—aminotransferase. Photochem.
Photobiol. 1g:297.

Arx, A. E. von, and R. Neher. 1963. Eine multidimen-
sionale technik zur chromatographischen identifi-
zierung von aminosauren. J. Chromatog. 1g:329.

Atkin, L., A. S. Schultz, W. L. Williams, and C. N.
Frey. 1943. Yeast microbiological methods for
determination of vitamins. Pyridoxine. Ind.
Eng Chem. Anal. Ed. 1§:141.

198

_. '1

199

Bain, J. A., and H. L. Williams. 1960. Inhibition in
the nervous system and gamma-aminobutyric acid.
Proceedings of an International Symposium held
in 1959. Pergamon Press, New York. p. 275.

Baker, H., and H. Sobotka. 1962. Microbiological assay
methods for vitamins. In: Advances in Clinical
Chemistry. Ed. H. Sobotka and C. Stewart.
Academic Press, New York.

Barton-Wright, E. C. 1971. The microbiological assay
of the vitamin B6 complex (pyridoxine, pyridoxal
and pyridoxamine) with Kloeckera brevis.

Analyst 96:314.

 

Bayoumi, R. A., and S. B. Rosalki. 1976. Evaluation of
methods of coenzyme activation of erythrocyte
enzymes for detection of deficiency of vitamins
B1, B2, and B6. Clin. Chem. 22:327.

Beaton, G. H., and M. C. Cheney. 1965. Vitamin B
requirement of the male albino rat. J. Nu r.
87:125.

Bellamy, W. D., and I. C. Gunsalus. 1944. Tyrosine
decarboxylation of Streptococci: growth require-
ments for active ceIl production. J. Bacteriol.
48:191.

 

Benson, E. M., J. Peters, M. A. Edwards, and S. A. Stor-
vick. 1964. Problems in the microbiological
determination of vitamin B6 in blood. Fed. Proc.
23:138.

Bergel, F., and K. R. Harrap. 1961. Interaction between
carbonyl groups and biologically essential sub-
stituents. Part III. The formation of a thia-
zolidine derivative in aqueous solution from
pyridoxal phosphate and L-cysteine. J. Chem.
Soc. 4051.

Bernhart, F. W., E. D'Amato, and R. M. Tomarelli. 1960.
The vitamin 36 activity of heat-sterilized milk.
Arch. Biochem. BiOphys. §§;267.

Birch, T. W., and P. Gyorgy. 1936. Vitamin B5 and
methods for preparation in a concentrated state.
Biochem. J. 30:304.

Bliss, C. 1., and C. White. 1967. Statistical methods
in biological assay of the vitamins. In: The
Vitamins, Vol. VI. Ed. P. Gyorgy and W. N.
Pearson. Academic Press, New York. p. 21.

200

Bonavita, V. 1960. The reaction of pyridoxal phosphate
with cyanide and its analytical use. Arch.
Biochem. BiOphys. 33:366.

Brin, M. 1970. A simplified Toepfer-Lehmann assay for
the three 36 vitamers. In: Methods in Enzy-
mology, Vol. 18A. Academic Press, New York.

p. 519.

Brin, M., and V. F. Thiele. 1967. Relationships between
vitamin B5-vitamer content and the activities of
transaminase enzymes in rat tissues at varying
intake levels of vitamin 85. J. Nutr. 33:213.

Buell, M. V., and R. E. Hansen. 1960. Reaction of
pyridoxal-S-phosphate with amino-thiols.
J. Am. Chem. Soc. 33:6042.

Bunting, W. R. 1965. The stability of pyridoxine added
to cereals. Cereal Chem. 13:569.

Cabantchik, I. Z., M. Balshin, W. Breuer, and A. Roth-
stein. 1975. Pyridoxal phosphate. An anionic
probe for protein amino groups on the outer and
inner surfaces of intact human red blood cells.
J. Biol. Chem. 332:5130.

Calbiochem. 1975. GOT reagents technical bulletin.
Doc. No. L03023. Calbiochem, La Jolla, Cal.

Cennamo, C. 1964. Metal ion independent non-enzymatic
transamination between pyridoxal and amino acid
esters. Biochim. Biophys. Acta 33:323.

Chen, L. H., and A. L. Marlatt. 1975. Effects of
dietary levels of vitamin 85 on glutamic-pyruvic
transaminase activity in rat tissues. J. Nutr.
323:401.

Chin, Y. P. 1975. Chromatographic Separation and Fluoro-
metric Determination of Pyridoxal, Pyridoxamine,
and Pyridoxine in Food System. M.S. Thesis.
Michigan State University, East Lansing.

Christensen, H. N. 1958. Three Schiff base types formed
by amino acids, peptides and proteins with pyri-
doxal and pyridoxal-S-phosphate. J. Am. Chem.
Soc. §Q}99.

Churchich, J. E. 1965. Energy transfer in protein pyri-
doxamine-S-phosphate complexes. Biochem. 1:1405.

201

Clark, A. V., and S. R. Tannenbaum. 1970. Isolation and
characterization of pigments from protein-
carbonyl browning systems. Isolation, purifi-
cation, and properties. J. Agr. Food Chem.
18:891.

Clarke, M. F., and M. Lechycka. 1943. The biological
assay of pyridoxine (vitamin Be). J. Nutr.
25:571.

Coates, M. E., S. K. Kon, and E. E. Shepheard. 1950.
The use of chicks for the biological assay of
members of the vitamin B complex. 1. Tests with
pure substances. Br. J. Nutr. 1:203.

Columbini, C. E., and E. E. McCoy. 1970. Rapid thin-
layer electrophoretic separation and estimation
of all vitamin B6 compounds and some 5-hydroxy-
indoles. Anal. Biochem. 31:451.

Conger, T. W., and C. A. Elvehjem. 1941. The biological
estimation of pyridoxine (vitamin B6). J. Biol.
Chem. 138:555.

Contractor, S. F., and B. Shane. 1968. Estimation of
vitamin 85 compounds in human blood and urine.
Clin. Chim. Acta .El‘7l°

Cort, W. M., B. Borenstein, J. H. Harley, M. Osadca, and
J. Scheiner. 1976. Nutrient stability of
fortified cereal products. Food Technol. 32:52.

Cortijo, M., J. Llor, J. S. Jiminez, and F. Garcia-Blanco.
1976. Studies on the pyridoxal phosphate site
in glycogen phospharylase b. Eur. J. Biochem.
65:521.

Coursin, D. B. 1954. Convulsive seizures in infants
with pyridoxine-deficient diet. J. Am. Med.
Assoc. 154:406.

Cunningham, E., and E. E. Snell. 1945. The vitamin B6
group. VI. The comparative stability of pyri-
doxine, pyridoxamine, and pyridoxal. J. Biol.
Chem. 133:491.

Daoud, H. N. 1973. Factors affecting the microbiological
assay of vitamin B . Ph.D. Thesis. University
of California, Dav1s. Davis, Cal.

202

Daoud, H. N., B. S. Luh, and M. W. Miller. 1977. Effect
of blanching, EDTA and NaHSO3 on color and
vitamin 85 retention in canned garbonzo beans.

J. Food Sci. 135375.

Davies, M. K., M. E. Gregory, and K. M. Henry. 1959.
The effect of heat on the vitamin B6 of milk.
II. A comparison of biological and microbio-
logical tests of evaporated milk. J. Dairy Res.
26:215.

Dempsey, W. B., and H. N. Christensen. 1962. The spe-
cific binding of pyridoxal 5'-phosphate to bovine
plasma albumin. J. Biol. Chem. 237:1113.

Dempsey, W. B., and E. E. Snell. 1963. Pyridoxamine-
pyruvate transaminase. II. Characteristics of
the enzyme. Biochem. 3:1414.

Dennison, D., J. Kirk, P. Kokoczka, and D. Heldman. 1977.
Storage stability of thiamin and riboflavin in
a dehydrated food system. J. Food Proc. Preserv.
1:43.

Diding, N. A. 1955. Biochemical mutants of micro-
organisms in vitamin assay of vitamin B6.
Svensk Farm. Tidskr. 33:321.

Dixon, H. B. F. 1966. Transamination between pyridoxal
and proteins. In: Pyridoxal Catalysis: Enzymes
and Model Systems. Ed. E. E. Snell, A. E. Bauern-
stein, E. S. Severin, and Y. M. Torchinsky.
Interscience Publishers, New York. p. 99.

Ellman, G. L. 1959. Tissue sulfhydryl groups. Arch.
Biochem. Biophys. 33:70.

Everson, G. J., J. Chang, S. Leonard, B. S. Luh, and
M. Simone. 1964. Aseptic canning of foods.
III. Pyridoxine retention as influenced by
processing method, storage time and temperature,
and type of container. Food Technol. 13:87.

Feeney, R. E., G. Blankenhorn, and H. B. F. Dixon. 1975.
Carbonyl-amine reactions in protein chemistry.
Adv. Protein Chem. 33:135.

Fiedlerova, V., and J. Davidek. 1974. Fluorometric
determination of pyridoxal in dried milk.
Z. Lebensm. Unters.-Forsch. 155:277.

203

Fields, R. 1972. The rapid determination of amino
groups with TNBS. In: Methods in Enzymology,
Vol. 258. ”Enzyme Structure.” Academic Press,
New York. p. 464.

Fischer, E. H., A. B. Kent, E. R. Snyder, and E. G. Krebs.
1958. The reaction of sodium borohydride with
muscle phosphorylase. J. Am. Chem. Soc. 33:2906.

Fredericq, E. 1953. Reversible dissociation of insulin.
Nature 171:570.

Fujita, A., D. Fujita, and K. Fujino. 1955. Fluorometric
determination of vitamin 35. II. Determination
of pyridoxamine. J. Vitaminol. 1:272.

Fujita, A., D. Fujita, and K. Fujino. 1955. Fluorometric
determination of vitamin B5. III. Fractional
determination of pyridoxal and 4-pyridoxic acid.
J. Vitaminol. ‘1:279.

Fujita, A., K. Masuura, and K. Fujino. 1955. Fluoro-
metric determination of vitamin B5. I. Determi-
nation of pyridoxine. J. Vitaminol. 1:267.

Gale, E. F., and H. M. Epps. 1944. Studies on bacterial
amino acid decarboxylases. Distribution and
preparation of codecarboxylase. Biochem. J.
38:250.

Garabedian, A. M. Der, and M. A. Der Garabedian. 1976.
Transamination between w-amino acids, a, w-diamino
acids and pyridoxal phosphate: formation of
aldimines. FEBS Letters 13:87.

Gregory, J. F., and J. R. Kirk. 1977a. Improved chroma-
tographic separation and fluorometric determi-
nation of vitamin Be compounds in foods. J.
Food Sci. 13:1073.

Gregory, J. F., and J. R. Kirk. 1977b. Determination of
urinary 4-pyridoxic acid using high performance
liquid chromatography. Am. J. Clin. Nutr.
(Submitted for publication).

Gregory, M. E. 1959. The effect of heat on the vitamin
B5 of milk. I. Microbiological tests. J.
Dairy Res. 33:203.

Gregory, M. E., and L. A. Mabbitt. 1961. The occurrence
of pyridoxamine phosphate in milk. J. Dairy
Res. 33:293.

204

Grillo, M. A. 1968. The effect of pyridoxal phosphate
on yeast hexokinase. Enzymologia 31:7.

Gruenwedel, D. W., and R. K. Patnaik. 1971. Release of
hydrogen sulfide and methyl mercaptan from
sulfur-containing amino acids. J. Agr. Food
Chem. 13:775.

Gunsalus, E. C., and W. D. Bellamy. 1944. The function
of pyridoxine and pyridoxine derivatives in the
decarboxylation of tyrosine. J. Biol. Chem.
133:557.

Gyorgy, P. 1938. Crystallization of vitamin B6. J. Am.
Chem. Soc. 33:983.

Hamilton, G. A., and A. Revesv. 1966. Oxidation by
molecular oxygen. IV. A possible model reaction
for some amine oxidases. J. Am. Chem. Soc.
33:2069.

Harding, R. S., I. C. Plough, and T. E. Friedmann.
1959. The effect of storage on the vitamin B5
content of a packaged army ration, with a note
on the human requirement for the vitamin. J.
Nutr. 33:323.

Harris, D. L. 1952. Interaction of thiamine and pyri-
doxine in Neurospora. I. Studies of the pyri-
doxineless mutants. Arch. Biochem. Biophys.
41:294.

Harris, S. A., D. Heyl, and K. Folkers. 1944a. The
structure and synthesis of pyridoxamine and
pyridoxal. J. Biol. Chem. 154:315.

Harris, S. A., D. Heyl, and K. Folkers. l944b. The
vitamin B5 group. II. The structure and syn-
thesis of pyridoxamine and pyridoxal. J. Am.
Chem. Soc. 33:2088.

Haskell, B. E., and E. E. Snell. 1970. Microbiological
determination of the vitamin B5 group. In:
Methods in Enzymology, Vol. 18A. "Vitamins and
Coenzymes." Academic Press, New York. p. 512.

Haskell, B. E., and U. Wallnofer. 1967. D-alanine
interference in microbiological assays in human
blood. Anal. Biochem. 13:569.

205

Hassinen, J. B., G. T. Durbin, and F. W. Bernhart. 1954.
The vitamin B5 content of milk products.
J. Nutr. £949.

Hedin, P. A. 1963. The separation of pyridoxine, pyri-
doxal, and pyridoxamine by a sulfonic acid ion
exchange resin. J. Agr. Food Chem. 11:343.

Hellstrom, V. 1960. The influence of sunlight on vitamin
Bs in milk. Int. Z. Vitaminforsch. 33:323.

Hennessy, D. J., A. M. Steinberg, G. S. Wilson, and W. P.
Keaveney. 1960. Fluorometric determination of
added pyridoxine in enriched white flour and
white bread baked from it. J. Assoc. Offic.

Agr. Chem. 13:765.

Heyl, D., S. A. Harris, and K. Folkers. 1948a. The
chemistry of vitamin Ba. VI. Pyridoxylamino
acids. J. Am. Chem. Soc. 33:3429.

Heyl, D., E. Luz, S. A. Harris, and K. Folkers. 1948b.
Chemistry of vitamin B5. VII. Pyridoxylidene-
and pyridoxylamines. J. Am. Chem. Soc. 13:3669.

Heyl, D., E. Luz, S. A. Harris, and K. Folkers. 1951a.
Phosphates of the vitamin Be group. I. The
structure of codecarboxylase. J. Am. Chem. Soc.
73:3430.

Heyl, D., E. Luz, S. A. Harris, and K. Folkers. 1951b.
Phosphates of the vitamin B5 group. III.
Pyridoxamine phosphate. J. Am. Chem. Soc.
13:3436.

Heyl, D., E. Luz, S. A. Harris, and K. Folkers. 1952.
Chemistry of vitamin B5. VIII. Additional
pyridoxylideneamines and pyridoxylamines.

J. Am. Chem. Soc. 31:414.

Hilak, M. C., B. J. M. Harmsen, J. J. M. Joordens, and
G. A. J. Van Os. 1975. Binding of pyridoxal
5'-phosphate to bovine serum albumin and albumin
fragments obtained after proteolytic hydrolysis.
Localization and nature of the primary PLP bind-
ing site. Int. J. Peptide Protein Res. ‘Z:4ll.

Hill, J. M. 1972. The oxidation of Schiff bases of
pyridoxamine with a-oxo acids by manganous ions
and peroxidase. Biochem. J. 128:701.

206

Hochberg, M., D. Melnick, L. Siegel, and B. L. Oser.
1943. Destruction of vitamin B (pyridoxine)
by light. J. Biol. Chem. 148: 53.

Hodson, A. Z. 1956. Vitamin B6 in sterilized milk and
other milk products. J. Agr. Food Chem. 1:876.

Hopkins, R. H., and R. J. Pennington. 1947. The assay
of the vitamin B6 complex. Biochem. J. 11:110.

Horvath, C., W. Melander, and I. Molnar. 1977. Liquid
chromatography of ionogenic substances with non-
polar stationary phases. Anal. Chem. 13:142.

Ichiba, A., and K. Michi. 1938. Crystalline vitamin B6-
Sci. Papers Inst. Phys. Chem. Res. 31:623.

Ikawa, M., E. E. Snell. 1954. Oxidative deamination of
amino acids by pyridoxal and metal salts.
J. Am. Chem. Soc. 13:4900.

Johansson, 8., S. Lindstedt, and H.-G. Tiselius. 1968.
Metabolism of 3H8 pyridoxine in mice. Biochem.
7:2327.

Johnson, G. S., and W. C. Deal. Inactivation of tetra-
meric rabbit muscle pyruvate kinase by specific
binding of 2 to 4 moles of pyridoxal 5'-phosphate.
J. Biol. Chem. 313:238.

Jones, A., and S. Morris. 1959. A 24-hour plate assay
technique for the vitamin Ba complex of yeast.
Analyst 33:608.

Kent, A. B., E. G. Krebs, and E. H. Fischer. 1958.
Properties of crystalline phosphorylase b.
J. Biol. Chem. 232:549.

Keresztesy, J. C., and J. R. Stevens. 1938. Vitamin B5.
Proc. Soc. Exptl. Biol. Med. 33:64.

Kierska, D., and C. Maslinski. 1971. Conditions for the
formation of complexes between pyridoxal phos-
phate and some imidazole derivatives. Biochem.
Pharmacol. 33:1951.

Kirk, J., D. Dennison, P. Kokoczka, and D. Heldman. 1977.
Degradation of ascorbic acid in a dehydrated food
system. J. Food Sci. 13:1274.

207

Kirksey, A., R. L. Pang, and W. J. Lin. 1975. Effects
of different levels of pyridoxine fed during
pregnancy superimposed upon growth in the rat.
J. Nutr. 133:607.

Kishi, H., and K. Folkers. 1976. Improved and effective
assays of the glutamic oxaloacetic transaminase
by the coenzyme-apoenzyme system (CAS) principle.
J. Nutr. Sci. Vitaminol. 33:225.

Klosterman, H. J. 1974. Vitamin 36 antagonists of
natural origin. J. Agr. Food Chem. 33:13.

Klostermeyer, H., and R. E. Humbel. 1966. The chemistry
and biochemistry of insulin. Angew. Chem.
Internat. Edit. 3:807.

Korytkyk, W. 1970. Gas chromatography of vitamin B5.
In: Methods in Enzymology, Vol. 18A. ”Vitamins
and Coenzymes." Academic Press, New York. p. 500.

Kraut, H., and U. Imhoff. 1967. Forschungsber. Landes
Nordrheim-Westfalen 1833.

Kuhn, R., and G. Wendt. 1938. Uber das antidermatitische
vitamin der Hefe. Ger. Chem. Ges. 11:780.

Labuza, T. P. 1975. Final Report. Contract NAS-12560.
National Aeronautics and Space Administration.
Food and Nutrition Office. Houston, Texas.
Aug. 1974 to Aug. 1975. pp. 126-215.

Lea, C. H., and R. S. Hannan. 1949. Studies of the
reaction between proteins and reducing sugars in
the "dry” state. I. The effect of water, of pH
and of temperature on the primary reaction
between casein and glucose. Biochim. Biophys.
Acta 3:313.

Lee, D. J. W., A. R. Robblee, and R. Blair. 1976. The
use of growth and liver aspartate aminotrans-
ferase to assess the effect of source of non-
essential nitrogen on pyridoxine depletion,
repletion and requirements of chicks. Br.
Poult. Sci. 11:43.

Leklem, J. E. 1977. Vitamin B-6 enrichment of wheat
flour: stability and bioavailability. Report
to the Food and Nutrition Board: Committee on
Food Protection of the National Academy of
Science. Technology of Fortification of Cereal-
Grain Products. May 16-17.

208

Lepkovsky, S. J. 1938. The isolation of factor one in
crystalline form. J. Biol. Chem. 124:125.

Levine, J. B., and J. B. Hill. 1966. Automated fluoro-
metric determinations of serum glutamic oxalacetic
transaminase and serum glutamic pyruvic trans-
aminase. In: Automation in Analytical Chemistry-
Proceedings of the 1965 Technicon Symposium.

Ed. L. T. Skeggs. Mediad, Inc., New York.

Linkswiler, H., C. A. Bauman, and E. E. Snell. 1951.
Effect of aureomycin on the response of rats to
various forms of vitamin B5. J. Nutr. 13:565.

Loncin, M., J. J. Bimbenet, and J. Lenges. 1968.
Influence of the activity of water on the
spoilage of food stuffs. J. Food Technol.
3:131.

Loo, Y. H., and L. Badger. 1969. Spectrofluorometric
assay of vitamin B5 analogues in brain tissue.
J. Neurochem. 13:801.

Lowry, O. H., N. J. Rosebrough, A. L. Farr, and R. J.
Randall. 1951. Protein measurement with the
Folin phenol reagent. J. Biol. Chem. 193:265.

Lucas, N., H. K. King, and S. J. Brown. 1962. Sub-
strate attachment in enzymes. The interaction
of pyridoxal phosphate with amino acids. Bio-
chem. J. 31:118.

Lushbough, C. H., J. M. Weichmann, and B. S. Schweigert.
1959. The retention of vitamin B5 in meat during
cooking. J. Nutr. 31:451.

Lyon, J. B., J. A. Bain, and Williams. 1962. The dis-
tribution of vitamin B5 in the tissues of two
inbred strains of mice fed complete and vitamin
BG-deficient rations. J. Biol. Chem. ‘331:1989.

MacArthur, M. J., and Lehmann. 1959. Chromatographic
separation and fluorometric measurement of
vitamin B5 components in aqueous solutions.

J. Assoc. Offic. Agr. Chem. 13:619.

Mackay, D. 1963. The reaction of thiol amino acids
with pyridoxal 5'-phosphate in the presence and
absence of L-glutamic acid. Biochim. Biophys.
Acta 13:445.

209

Mackay, D., and D. M. Shepherd. 1962. Inhibition of
histidine decarboxylases and interaction of the
inhibitors with pyridoxal 5'-phosphate. Biochim.
Biophys. Acta 33:553.

Masukawa, K., A. Nakama, H. Monaka, T. Kondo, and K.
Okumara. 1971. Vitamin B5 derivatives (XIV).
Differential determination of pyridoxal, pyri-
doxamine, pyridoxal 5'-phosphate, and pyridoxamine
5'-phosphate in biological materials. Vitamins
44:168.

Matsuo, Y. 1957a. Formation of Schiff bases of pyridoxal
phosphate. Reaction with metal ions. J. Am.
Chem. Soc. 13:2011.

Matsuo, Y. 1957b. Pyridoxal catalysis of non-enzymatic
transamination in ethanol solution. J. Am.
Chem. Soc. 13:2016.

Mauron, J. 1960. The concept of amino acid availability
and its bearing on protein evaluation. Intern.
Conf. Protein Needs, Washington, D.C., Aug. 21-22.

McCormick, D. B., and E. E. Snell. 1961. Pyridoxal
phosphokinases. II. Effects of inhibitors.
J. Biol. Chem. 236:2085.

McDonald, R., J. R. Kirk, and J. R. Brunner. 1975.
Peptide contamination from cellulose dialysis
membrane. J. Agr. Food Chem. 31:435.

Melachouris, N. 1969. Discontinuous gel electrophoresis
of whey proteins, casein and clotting enzymes.
J. Dairy Sci. 33:456.

Metzler, D. E. 1957. Equilibria between pyridoxal and
amino acids and their imines. J. Am. Chem.
Soc. 13:485.

Metzler, D. E., M. Ikawa, and E. E. Snell. 1954a. A
general mechanism for vitamin B5-cata1yzed
reactions. J. Am. Chem. Soc. 13:648.

Metzler, D. E., J. B. Longenecker, and E. E. Snell.
1954b. The reversible catalytic cleavage of
hydroxyamino acids by pyridoxal and metal salts.
J. Am. Chem. Soc. 13:639.

210

Metzler, D. E., and E. E. Snell. 1952. Some trans-
amination reactions involving vitamin B5.
J. Am. Chem. Soc. 11:979.

Miller, C. F., D. G. Guadagni, and S. Kon. 1973. Vitamin
retention in bean products: cooked, canned and
instant bean powders. J. Food Sci. 33:493.

Moore, W. J. 1962. Physical Chemistry, 3rd ed.
Prentice-Hall, Inc. Englewood Cliffs, N.J.
p. 253.

Morrison, A. L., and R. F. Long. 1958. The photolysis
of pyridoxal phosphate. J. Chem. Soc. 211.

Nelson, E. W., H. Lane, and J. J. Cerda. 1976. Compara-
tive human intestinal bioavailability of vitamin
B5 from a synthetic and a natural source.
J. Nutr. 133:1433.

Neter, J., and W. Wasserman. 1974. Applied Linear Sta-
tistical Models. Richard D. Irwin, Inc.,
Homewood, Ill.

Ohishi, N., and S. Fukui. 1968. Further study on the
reaction products of pyridoxal and pyridoxal
5'-phosphate with cyanide. Arch. Biochem.
Biophys. 133:606.

O'Leary, M. H. 1971. A proposed structure for the 330 nm
chromophore of glutamate decarboxylase and other
pyridoxal 5'-phosphate dependent enzymes.

Biochim. Biophys. Acta 313:484.

Oncley, J. L., E. Ellenbogen, D. Gitlin, F. R. N. Gurd.
1952. Protein-protein interactions. J. Phys.
Chem. 33:85.

Ott, W. H. 1946. Antipyridoxine activity of 2,4-dimethy1-
3-hydroxy-5-hydroxymethlypyridine in the chick.
Proc. Soc. Exptl. Biol. Med. 31:125.

Page, R. C., E. P. Benditt, and C. R. Kirkwood. 1968.
Schiff base formation by the lysyl and hydroxy-
1ysy1 side chains of collagen. Biochem. Biophys.
Res. Comm. 33:752.

Palnitkar, M. P., and D. R. Heldman. 1971. Equilibrium
moisture characteristics of freeze-dried beef
components. J. Food Sci. 33:1015.

Parrish, W. P., H. W. Loy, and O. L. Kline. 1955. A
study of the yeast method for vitamin B5. J.
Assoc. Offic. Agr. Chem. 33:506.

211

Parrish, W. P., H. W. Loy, and O. L. Kline. 1956.
Further studies on the yeast method for vitamin
86' J. Assoc. Offic. Agr. Chem. 33:157.

Peterson, E. A., and H. A. Sober. 1954. Preparation of
crystalline phosphorylated derivatives of vitamin
B6. J. Am. Chem. Soc. 13:169.

Pharmacia Fine Chemicals. 1974. Sephadex Gel Filtration
in Theory and Practice. Pharmacia Fine Chemicals
A B, Uppsala, Sweden.

Polansky, M. M., R. T. Camarra, and E. W. Toepfer. 1964.
Pyridoxine determined fluorometrically as pyri-
doxal cyanide compound. J. Assoc. Offic. Agr.
Chem. 11:827.

Polansky, M. M., and E. W. Toepfer. 1969. Vitamin 85
components in some meats, fish, dairy products,
and commercial infant formulas. J. Agr. Food
Chem. 11:1394.

Raab, C. A., B. S. Luh, and B. S. Schweigert. 1973.
Effects of heat processing on the retention of
vitamin 86 in Lima beans. J. Food Sci. 33:544.

Rabinowitz, J. C., and E. E. Snell. 1947. Extraction
procedures for the microbiological determination
of vitamin B6. Ind. Eng. Chem. Anal. Ed. 13:277.

Rabinowitz, J. C., and E. E. Snell. 1948. The vitamin
B5 group. XIV. Distribution of pyridoxal, pyri-
doxamine, and pyridoxine in some natural products.
J. Biol. Chem. 113:1157.

Rabinowitz, J. C., and E. E. Snell. 1949. The vitamin
B6 group. XV. Urinary excretion of pyridoxal,
pyridoxamine, and 4-pyridoxic acid in human
subjects. Proc. Soc. Exptl. Biol. Med. 13:235.

Rabinowitz, J. C., and E. E. Snell. 1953. The microbio-
logical activity of pyridoxylamines. J. Am.
Chem. Soc. 13:998.

Raica, N., and H. E. Sauberlich. 1964. Blood cell
transaminase activity in human vitamin B5
deficiency. Am. J. Clin. Nutr. 13:67.

Raines, R. C., and B. E. Haskell. 1968. Peptide stimu-
lation of Streptococcus faecium, a test organism
for pyridoxal and pyrideamine. Anal. Biochem.
23:413.

 

212

Register, U. D., U. J. Lewis, W. R. Ruegamer, and C. A.
Elvehjem. 1950. Studies on the nutritional
adequacy of army combat rations. J. Nutr.
40:281.

Reiber, H. 1972. Photochemical reactions of vitamin B5
compounds, isolation and properties of products.
Biochim. Biophys. Acta 279:310.

Reiber, H. 1976. Vitamin B5-catalyzed B-elimination of
serine and O-phosphoserine. Biochim. Biophys.
Acta 444:734.

Richardson, L. R., S. Wilkes, and S. J. Ritchey. 1961.
Comparative vitamin B5 activity of frozen,
irradiated and heat-processed foods. J. Nutr.
73:363.

Rippa, M., L. Spanio, and S. Pontremoli. 1967. A spe-
cific interaction of pyridoxal 5'-phosphate and
6-phosphog1uconate dehydrogenase. Arch. Biochem.
Biophys. 113:48.

Riquelme, P., W. E. Brown, and F. Marcus. 1975. Modifi-
cation of bovine pancreatic ribonuclease A with
pyridoxal 5'-phosphate. Isolation and identifi-
cation of derivatives. Int. J. Peptide Protein
Res. 1:379.

Ronchi, S., M. C. Zapponi, and G. Ferri. 1969. Inhibi-
tion by pyridoxal-phosphate of glyceraldehyde-3-
phosphate dehydrogenase. Eur. J. Biochem. 3:325.

Rotilio, G., A. M. Wolf, A. Finazzi Agro, and B. Mondovi.
1970. Oxidation of a lysine-copper-pyridoxal
phosphate complex. Ital. J. Biochem. 13:28.

Rubin, S. H., J. Scheiner, and E. Hirschberg. 1947.
The availability of vitamin B5 in yeast and
liver for growth of Saccharomyces carlsbergensis.
J. Biol. Chem. 131:599.

 

Sarma, P. S., E. E. Snell, and C. A. Elvehjem. 1946.
The vitamin B5 group. VIII. Biological assay
of pyridoxal, pyridoxamine, and pyridoxine. J.
Biol. Chem. 133:55.

Sarma, P. S., E. E. Snell, and C. A. Elvehjem. 1947.
The bioassay of vitamin 36 in natural materials.
J. Nutr. 33:121.

213

Sauberlich, H. E. 1961. Effect of vitamin B5 on the
growth of rats fed diets limiting in an essential
amino acid and on the utilization of isomers of
tryptophan, methionine, and valine. J. Nutr.
74:289.

Sauberlich, H. E. 1967. Vitamin B5. In: The Vitamins,
Vol. VII. Ed. P. Gyorgy and W. N. Pearson.
Academic Press, New York. p. 169.

Sauberlich, H. E. 1968. Vitamin B5 group. VIII. Active
compounds and antagonists. In: The Vitamins,
Vol. II. Ed. W. H. Sebrell and R. S. Harris.
Academic Press, New York. p. 33.

Scheller, F., and H. Will. 1973. A polarographic study
on pyridoxal-5'-phosphate binding in glycogen
phosphorylase b. FEBS Letters 33:47.

Schnickels, R. A., H. C. Warmbier, and T. P. Labuza.
1976. Effect of protein substitution on non-
enzymatic browning in an intermediate moisture
food system. J. Agr. Food Chem. 31:120.

Schonbeck, N. D., M. Skalski, and J. A. Shafer. 1975.
Reactions of pyridoxal 5'-phosphate, 6-amino-
caproic acid, cysteine, and penicillamine.
Models for reactions of Schiff base linkages
in pyridoxal 5'-phosphate-requiring enzymes.
J. Biol. Chem. 333:5343.

Severin, E. S., N. N. Gulyaev, E. N. Khurs, and R. M.
Khomutov. 1969. The synthesis and properties
of phosphopyridoxyl amino acids. Biochem.
Biophys. Res. Comm. 33:318.

Shaltiel, S., and M. Cortijo. 1970. The mode of binding
of pyridoxal 5'-phosphate in glycogen phosphoryl-
ase. Biochem. Biophys. Res. Comm. 11:594.

Shane, B., and E. E. Snell. 1972. The nature of "pyri-
doxine oxidase" from rabbit liver preparations.
Arch. Biochem. Biophys. 153:333.

Shapiro, S., M. Enser, E. Pugh, and B. L. Horecker. 1968.
The effect of pyridoxal phosphate on rabbit
muscle aldolase. Arch. Biochem. Biophys.

133:554.

Sheppard, A. J., and A. R. Prosser. 1970. Gas chro-
matography of vitamin B5. In: Methods in
Enzymology, Vol. 18A. Academic Press, New York.
p. 494.

214

Smith, M. A., and L. S. Dietrich. 1971. Preparative
thin-layer chromatography for the separation of
the various forms of vitamin B5 in tissues.
Vitamin B5 content of the chick embryo liver
during the midperiod of development. Biochim.
Biophys. Acta 333:262.

Snell, E. E. 1942. Effect of heat sterilization on
growth-promoting activity for Streptococcus
lactis R. Proc. Soc. Exptl. Biol. Med} 31:356.

 

Snell, E. E. l944a. The vitamin B6 group. I. Formation
of additional members from pyridoxine and evidence
concerning their structure. J. Am. Chem. Soc.
33:2082.

Snell, E. E. l944b. The vitamin activities of "pyridoxal"
and "pyridoxamine." J. Biol. Chem. 154:313.

Snell, E. E., B. M. Guirrard, and R. J. Williams. 1942.
Occurrence in natural products of a physiologi-
cally active metabolite of pyridoxine. J. Biol.
Chem. 113:519.

Snell, E. E., and J. C. Rabinowitz. 1948. The micro-
biological activity of pyridoxylamino acids.
J. Am. Chem. Soc. 13:3433.

Soda, K., T. Yorifuji, H. Misono, and M. Moriguchi. 1969.
Spectrophotometric determination of pyridoxal
and pyridoxal 5'-phosphate with 3-methy1-2-benzo-
thiazolone hydrochloride, and their selective
assay. Biochem. J. 111:629.

Sood, V. C., and D. R. Heldman. 1974. Analysis of a
vapor pressure manometer for measurement of

water activity in nonfat dry milk. J. Food Sci.
33:1011.

Srivastava, S. K., and E. Beutler. 1973. A new fluoro-
metric method for the determination of pyridoxal
5'-phosphate. Biochim. Biophys. Acta 304:765.

Srncova, V., and J. Davidek. 1972. Reaction of pyridoxal
and pyridoxal-S-phosphate with proteins. Reaction
of pyridoxal with milk serum proteins. J. Food
Sci. 31:310.

Stokes, J. L., L. L. Larson, C. R. Woodward, and J. W.
Foster. 1943. A Neurospora assay for pyridoxine.
J. Biol. Chem. 150:17.

 

215

Storvick, C. A., E. M. Benson, M. A. Edwards, and M. J.
Woodring. 1964. Chemical and microbiological
determination of vitamin B5. Methods of Bio—

(chemical Analysis 13:183.

Storvick, C. A., and J. M. Peters. 1964. Methods for
the determination of vitamin B5 in biological
materials. Vitamins and Hormones 33:833.

Takanashi, S., I. Matsunaga, and Z. Tamura. 1970.
Fluorometric determination of pyridoxal and
its 5'-phosphate in biological materials.
J. Vitaminol. 13:132.

Takanashi, S., and Z. Tamura. 1970., Preliminary studies
for fluorometric determination of pyridoxal and
its 5'-phosphate. J. Vitaminol. 13:129.

Tappan, D. V., U. J. Lewis, A. H. Methfessel, and C. A.
Elvehjem. 1953. Studies concerning the pyri-
doxine requirements of rats receiving highly
processed rations. J. Nutr. 31:479.

Taylor, A. A. 1961. Determination of moisture equilibria
in foods. Food Technol. 13:536.

Technicon Instruments Corp. 1969. Hemoglobin. Technicon
AutoAnalyzer Method File N-186 I/II. Technicon
Instrument Corp. Tarrytown, N.Y.

Thiele, V. F., and M. Brin. 1966. Chromatographic
separation and microbiological assay of vitamin
B5 in tissues from normal and vitamin B5-
deficient rats. J. Nutr. 33:347.

Tiselius, H.-G. 1972. A chromatographic separation of
the different forms of vitamin B5. Clin. Chim.
Acta 13:319.

Toepfer, E. W., and J. Lehmann. 1961. Procedure for
chromatographic separation and microbiological
assay of pyridoxine, pyridoxal, and pyridoxamine
in food extracts. J. Assoc. Offic. Agr. Chem.
44:426.

Toepfer, E. W., M. J. MacArthur, and J. Lehmann. 1960.
Chromatographic separation of vitamin B6 compo-
nents in food extracts. J. Assoc. Offic. Agr.
Chem. 13:57.

216

Toepfer, E. W., and M. M. Polansky. 1970. Microbiological
assay of vitamin 85 and its components. J. Assoc.
Offic. Anal. Chem. ‘33:546.

Toepfer, E. W., M. M. Polansky, and E. M. Hewston. 1961.
Fluorometric determination of pyridoxamine by
conversion to pyridoxal cyanide compound. Anal.
Biochem. ‘3:463.

Toepfer, E. W., M. M. Polansky, L. R. Richardson, and
S. Wilkes. 1963. Comparison of vitamin B5
values of selected food samples by bioassay and
microbiological assay. J. Agr. Food Chem.
11:523.

Tomarelli, R. M., E. R. Spence, and F. W. Bernhart. 1955.
Biological availability of vitamin B5 of heated
milk. J. Agr. Food Chem. ‘3:339.

Tsuji, H., J. Okada, K. Iwami, K. Yasumoto, and H. Mit-
suda. 1977. Availability as vitamin B6 and
small intestinal absorption of pyridoxine-B-D-
glucoside in rats. Vitamins 31:153.

Wada, H., and E. E. Snell. 1961. The enzymatic oxi-
dation of pyridoxine and pyridoxamine phosphates.
J. Biol. Shem. 236:2089.

Waibel, W. E., W. W. Cravens, and E. E. Snell. 1952.
The effect of diet on the comparative activities
of pyridoxal, pyridoxamine, and pyridoxine for
chicks. J. Nutr. 13:531.

Waletzko, P., and T. P. Labuza. 1976. Accelerated
shelf-life testing of an intermediate moisture
food in air and in an oxygen-free atmosphere.
J. Food Sci. 11:1338.

Warmbier, H. C., R. A. Schnickels, and T. P. Labuza.
1976. Effect of glycerol on nonenzymatic brown-
ing in a solid intermediate moisture food system.
J. Food Sci. 11:528.

Warren, R. M., and T. P. Labuza. 1977. Comparison of
chemically measured available lysine with rela-
tive nutritive value measured by a Tetrahymena
bioassay during early stages of nonenzymatic
browning. J. Food Sci. 13:429.

Wendt, G., and F. W. Bernhart. 1960. The structure of
a sulfur-containing compound with vitamin B6
activity. Arch. Biochem. Biophys. 33:270.

217

Williams, A. K. 1974. Vitamin B6: gas-liquid chroma-
tography of pyridoxol, pyridoxal, and pyridox-
amine. J. Agr. Food Chem. 33:107.

Williams, A. K., and P. D. Cole. 1975. Vitamin B5. Ion
exchange chromatography of pyridoxal, pyridoxol,
and pyridoxamine. J. Agr. Food Chem. 33:915.

Woodring, M. J., and C. A. Storvick. 1960. Vitamin B5
in milk: review of literature. J. Assoc. Offic.
Agr. Chem. 13:63.

Yasumoto, K., K. Iwami, H. Tsuji, J. Okada, and H. Mitsuda.
1976. Bound forms of vitamin B5 in cereals and
seeds. Vitamins 33:327.

Yasumoto, K., K. Tadera, H. Tsuji, and H. Mitsuda. 1975.
Semi-automated system for analysis of vitamin B5
complex by ion-exchange column chromatography.

J. Nutr. Sci. Vitaminol. 31:117.

Yasumoto, K., H. Tsuji, K. Iwami, and H. Mitsuda. 1977.
Isolation from rice bran of a bound form of
vitamin 35 and its identification as 5'-O-
(B-D-glucopyranosyl) pyridoxine. Agr. Biol.
Chem. 11:1061.

Yen, J. T., A. H. Jensen, and D. H. Baker. 1976.
Assessment of the concentration of biologically
available vitamin B in corn and soybean meal.
J. Animal Sci. 13: 66.