METABOLISM AND NUTRITION
A Method for Estimating the Relative Degree of Saponification
of Xanthophyll Sources and Feedstuffs
D. L. Fletcher1
Department of Poultry Science, University of Georgia, Athens 30602-2772
ABSTRACT Saponification of xanthophyll esters in var-
ious feed sources has been shown to improve pigmenta-
tion efficiency in broiler skin and egg yolks. Three trials
were conducted to evaluate a rapid liquid chromatogra-
phy procedure for estimating the relative degree of xan-
thophyll saponification using samples of yellow corn,
corn gluten meal, alfalfa, and 6 commercially available
marigold meal concentrates. In each trial, samples were
extracted using a modification of the 1984 Association of
Official Analytical Chemists hot saponification procedure
with and without the addition of KOH. A comparison of
the chromatography results was used to estimate percent
Key words: saponification, xanthophyll, pigmentation
INTRODUCTION
Numerous factors have been identified that affect
broiler skin and egg yolk pigmentation. Factors such as
xanthophyll source, xanthophyll concentration, type of
bird, storage of feed, age of feedstuffs, and disease condi-
tion of the bird have been extensively reviewed (Marusich
and Bauernfeind, 1981; Fletcher, 1989). The chemical
structure, nomenclature, and biochemistry of the carot-
enoids as a general class of compounds are extensively
reviewed in the works by Bauernfeind (1981) and Good-
win (1980, 1984).
The influence of the chemical state of the xanthophylls,
bound or esterified, in the native state as opposed to being
nonbound, or free, in various products has received little
attention, except concerning extraction procedure
(Quackenbush and Miller, 1972) and a few contradictory
reports concerning egg yolk pigmentation. Coon and
Couch (1976) reported that saponification of marigold
meal could increase biological availability approximately
30% in laying hens. Philip et al. (1976) reported that lutein
fatty acid esters were better utilized by laying hens than
was crystalline lutein. Papa et al. (1985) reported that
saponification had no effect on egg yolk color. Lai et al.
2006 Poultry Science Association, Inc.
Received November 11, 2005.
Accepted January 8, 2006.
1
Corresponding author: fletcher@uga.edu
866
saponification of the original sample by dividing the non-
saponified extraction values by the saponified extraction
values. A comparison of the percent saponified xantho-
phylls for each product (mg/kg) was: yellow corn, 101;
corn gluten meal, 78; alfalfa, 97.9; and marigold concen-
trates A through F, 99.8, 4.6, 99.0, 95.6, 96.8, and 6.6,
respectively. These results indicate that a modification
of the 1984 Association of Official Analytical Chemists
procedure and liquid column chromatography can be
used to quickly verify saponification and can be used
to estimate the relative degree of saponification of an
unknown xanthophyll source.
2006 Poultry Science 85:866–869
(1996) reported that saponification of paprika did not
affect the color of egg yolks compared with a nonsaponi-
fied paprika oleoresin. Galobart et al. (2004) reported that
saponification of both marigold and paprika extracts im-
proved pigmenting efficiency.
Little research has been reported on the effects of ester
vs. free form of xanthophyll on broiler skin pigmentation.
However, the poultry industry prefers the use of saponi-
fied marigold products based on their judgment of supe-
rior performance. Using oral intubation and blood serum
analyses, Middendorf et al. (1980) concluded that saponi-
fication of marigold meal had no significant effect on
serum xanthophyll availability. Fletcher et al. (1986) di-
rectly compared saponified and nonsaponified marigold
extracts using both white and yellow corn basal diets
with broilers. The results clearly indicated that saponifi-
cation of the marigold extract concentrate resulted in a
significant improvement in broiler skin color.
The Association of Official Analytical Chemists (AOAC
International, 2000) extraction procedure, using either the
cold (overnight) or hot saponification procedure, could
not be expected to differentiate products that might or
might not have been subjected to saponification. The tra-
ditional method commonly used in the poultry industry
to determine whether a product was saponified is a thin
layer chromatography procedure distributed by commer-
cial sources. The procedure involves the spotting of silica
gel G plates with a xanthophyll extract dissolved in hex-
ane (AOAC extraction) and developing the plate in a
 ESTIMATING XANTHOPHYLL SAPONIFICATION
hexane:acetone (70:30) solution. The more polar xantho-
phyll esters develop more rapidly, and the less polar, free
xanthophylls develop more slowly. By comparing the
relative migration of the more polar xanthophyll esters,
one can observe the intensity of spots to qualitatively
determine whether the product has indeed been saponi-
fied. Quantification depends on scraping the plates, resus-
pending the xanthophylls, and subjecting them to tradi-
tional column or HPLC techniques. In addition, free xan-
thophylls result in 2 spots, representing a mixture of free
monohydroxy xanthophylls and mixed xanthophyll
diesters.
During the earlier research reported by Fletcher et al.
(1986), several methods were used to qualitatively esti-
mate whether marigold concentrates had been saponified
during production. These methods used both the thin
layer chromatography procedure and a liquid column
chromatography procedure, both of which are relatively
simple, rapid, and give a visual confirmation of the rela-
tive degree of saponification. The purpose of this project
was to evaluate a modification of the AOAC Official
Method 970.64, carotenes and xanthophylls in dried plant
materials and mixed feeds (AOAC International, 2000) by
comparing chromatographic results of various feedstuffs
with and without saponification during sample prepara-
tion. Emphasis was placed on quantitatively comparing
the results to determine the degree to which the xantho-
phylls in a given source are free or esterified, primarily
to evaluate if marigold products were indeed saponified
during processing or if saponified and nonsaponified
products might have been blended.
MATERIALS AND METHODS
Yellow corn, corn gluten meal, and alfalfa were ob-
tained from commercial feed sources. Six different com-
mercial marigold products, identified A through F, were
obtained either directly from the manufacturer or were
obtained from commercial feed mills. All samples were
placed in plastic bags and stored at −20°C in the dark
prior to analysis.
Carotenoids [both carotenes (CAR) and xanthophylls]
were extracted using the AOAC hot saponification proce-
dure (AOAC International, 2000). As recommended in
the procedure, the size of the sample was based on esti-
mated total xanthophyll (TX) content, and the addition
of water during extraction was based on the moisture
content of the sample. To determine the optimum sample
size, each product was subjected to a preliminary AOAC
extraction and chromatography.
A second set of extractions was conducted, identical to
those previously described, except for the deletion of
KOH. Depending on the product being extracted, either
2 or 4 mL of methanol was added in place of the 2 or 4
mL of 40% methanolic KOH. The samples were then held
at 56°C for 1 h in the dark prior to chromatography.
Total CAR, monohydroxy pigments (MHP), dihydroxy
pigments (DHP), residual pigments (RP), and TX (MHP
+ DHP + RP) were also determined as described by the
867
AOAC International procedure (2000) using adsorbent 1
(silica gel G and diatomaceous earth, 1:1) and eluting the
fractions as follows: CAR with hexane:acetone at 96:4;
MHP with hexane:acetone at 90:10; DHP with hexane:ace-
tone at 80:20; and RP with hexane:acetone:methanol at
80:10:10. The RP eluent is part of the AOAC TX procedure
using adsorbent II, but was used in this case to recover
the slow-eluting xanthophylls that would otherwise re-
main on the column using adsorbent I; this is why the
RP fraction in used in the calculation for TX. The highly
polar polyoxy pigments were retained on the column. To
reduce the influence of moisture affecting the elutions,
adsorbent I was dried at 100°C for 20 min and held in a
desiccator prior to column packing and immediate use.
During initial test trials (in which the sample sizes were
also estimated), it was found that excessive amounts of
fraction “bleeding” were observed for some of the nonsa-
ponified extracts. Apparently, the xanthophyll esters or
nonsaponified species are not homogenous and do not
elute in a clean manner. To help reduce this problem,
greater eluent volumes were used for both the saponified
and nonsaponified extracts than recommended in the
AOAC procedure. It was found during these initial trials
that 30 mL of CAR, 80 mL of MHP, 50 mL of DHP, and
75 mL of RP eluents brought to total volumes of 100 mL
prior to spectrophotometer readings were most reproduc-
ible. The absorbance for the CAR fraction was determined
at 436 nm, and for the MHP, DHP, and RP fractions,
absorbance was determined at 474 nm, immediately fol-
lowing elution and bringing to standard volume.
Duplicate columns were run on each extraction and
averaged. The entire experiment, extraction through chro-
matography, was replicated 3 times. Data were summa-
rized using the means procedures in the SAS/STAT soft-
ware program (SAS Institute, 1988).
RESULTS AND DISCUSSION
The results for the preliminary tests for the 9 different
xanthophyll feedstuffs and concentrates resulted in the
following samples sizes: yellow corn, 4.00 g; corn gluten
meal and alfalfa, 2.50 g; marigold concentrate A, 0.20 g;
marigold concentrate D, 0.10 g; and marigold concen-
trates B, C, E, and F, 0.05 g. As per the AOAC procedure,
1 mL of water was added to all samples except for the
yellow corn.
The fractionation results for CAR, MHP, DHP, RP, TX,
and total carotenoids are presented in Table 1. The values
for the saponified extraction of all products were similar
to those obtained in a previous study (Fletcher et al.,
1986). Total xanthophyll ranged from 15 mg/kg for the
yellow corn up to 43,237 mg/kg for 2 of the marigold
concentrates. Saponification had little effect on the frac-
tion profiles except for marigold products B and F. In the
nonsaponified samples, the esterified xanthophylls eluted
with the less polar CAR and MHP fractions as previously
described by Quackenbush (1973). However, following
saponification, the free, more-polar xanthophylls eluted
primarily with the DHP fraction.
868 FLETCHER
Table 1. Mean results of 3 experimental trials for the saponified (Sap) and nonsaponified (Nonsap) extractions
for total carotene (CAR), monohydroxy (MHP), dihydroxy (DHP), residual pigment (RP), total xanthophyll
(TX = MHP + DHP + RP), and total carotenoids (total = CAR + TX) of the 3 xanthophyll-containing feedstuffs
(yellow corn, corn gluten meal, and alfalfa) and the 6 marigold concentrates, A through F

Fraction
Extraction
Product treatment CAR MHP DHP RP TX Total

(mg/kg)
Yellow corn Nonsap 4 2 9 4 15 19
Sap 4 2 12 2 15 20
Corn gluten Nonsap 33 20 42 5 67 100
Sap 12 13 56 4 73 84
Alfalfa Nonsap 36 72 52 18 142 178
Sap 12 5 41 6 51 62
Marigold A Nonsap 212 254 10,329 227 10,811 11,023
Sap 186 269 10,389 227 10,885 11,070
Marigold B Nonsap 52,825 1,568 714 185 2,467 55,292
Sap 913 907 41,673 657 43,237 44,150
Marigold C Nonsap 1,144 1,551 19,844 967 22,362 23,507
Sap 907 654 20,943 834 22,432 23,339
Marigold D Nonsap 1,252 1,120 17,285 298 18,703 19,956
Sap 408 537 16,925 234 17,696 18,104
Marigold E Nonsap 767 994 12,028 1,250 14,272 15,039
Sap 281 393 11,606 480 12,479 12,760
Marigold F Nonsap 47,267 2,516 522 264 3,302 50,569
Sap 415 1,363 38,782 663 40,808 41,223

The results for the CAR, MHP, DHP, RP, and TX are
reported as a percentage of total carotenoids (TX + CAR)
in Table 2. The differences between the saponified and
nonsaponified fractions for marigold products B and F,
noted previously, are better illustrated as a percentage of
the total eluted pigments. In both marigold B and F, the
nonsaponified treatment resulted in a 95.5 and 93.5% CAR
fraction and a 1.3 and 1.0% DHP fraction, respectively.
However, following saponification, the CAR fractions
were 2.1 and 1.0%, and the DHP fractions were 94.4 and
94.1%, respectively. This pattern was also reflected in the
TX fractions.
Percent saponification was determined by dividing the
nonsaponified extraction TX by the saponified extraction
TX for both the absolute and percentage values (Table 3).
Because the addition of KOH during the hot saponifica-
tion procedure resulted in differences in total extracted
pigments, the percent extraction was determined by di-
viding the total carotenoids from the saponified extraction
by the total carotenoids from the nonsaponified extrac-
tion; this is also presented in Table 3.
The results indicate that substantial differences exist
among the 9 products as to their response to saponifica-
tion during extraction of the carotenoids. Yellow corn

Table 2. Percentage of total carotene (CAR), monohydroxy pigment (MHP), dihydroxy pigment (DHP), residual
pigment (RP), and total xanthophyll (TX) as a percentage of total carotenoid content for the saponified (Sap)
and nonsaponified (Nonsap) extracts of the 3 xanthophyll-containing feedstuffs (yellow corn, corn gluten meal,
and alfalfa) and the 6 marigold concentrates, A through F

Fraction
Extraction
Product treatment CAR MHP DHP RP TX
(%)
Yellow corn Nonsap 21.2 8.5 49.5 20.9 78.8
Sap 22.0 8.4 61.4 8.3 78.0
Corn gluten Nonsap 32.7 20.0 42.3 5.0 67.3
Sap 13.7 15.3 66.2 4.8 86.3
Alfalfa Nonsap 20.4 32.4 35.2 11.9 79.6
Sap 18.7 7.2 65.2 8.9 81.3
Marigold A Nonsap 1.9 2.4 93.7 2.1 98.1
Sap 1.7 2.4 93.8 2.1 98.3
Marigold B Nonsap 95.5 2.8 1.3 0.3 4.5
Sap 2.1 2.1 94.4 1.5 97.9
Marigold C Nonsap 4.9 6.6 84.4 4.1 95.1
Sap 3.9 2.8 89.7 3.6 96.1
Marigold D Nonsap 6.6 5.5 86.5 1.5 93.4
Sap 2.3 3.0 93.5 1.3 97.7
Marigold E Nonsap 5.3 6.5 80.6 7.6 94.7
Sap 2.2 3.1 91.0 3.7 97.8
Marigold F Nonsap 93.5 5.0 1.0 0.5 6.5
Sap 1.0 3.3 94.1 1.6 99.0
 ESTIMATING XANTHOPHYLL SAPONIFICATION 869
Table 3. Percentage of saponification calculated as total xanthophyll It should be pointed out that the hot saponification
(TX) and percentage of total xanthophyll (%TX) of total carotenoids
extracted and percentage of extraction of saponified extraction of nonsa- extraction (as opposed to the cold overnight extraction)
ponified extraction of the 3 xanthophyll-containing feedstuffs (yellow was initially developed for products such as marigold
corn, corn gluten meal, and alfalfa) and the 6 marigold concentrates, meals that have high native esterified xanthophyll levels
A through F
(Quackenbush and Miller, 1972). These results also have
Saponification1 (%) impact on the use of high performance liquid chromatog-
Product TX TX% Extraction2 (%) raphy relative to the preparation of the sample (saponifi-
cation of the sample) prior to elution.
Yellow corn 100 101 104
Corn gluten 92 78 85
Alfalfa 278 98 35 REFERENCES
Marigold A 99 100 100
Marigold B 6 5 80 AOAC International. 2000. Official Methods of Analyses. 17th
Marigold C 100 99 99 ed. W. Horwitz, ed. AOAC Int., Gaithersburg, MD.
Marigold D 106 96 91 Bauernfeind, J. C. 1981. Carotenoids as Colorants and Vitamin
Marigold E 114 97 85
Marigold F 8 7 82
A precursors. J. C. Bauernfeind, ed. Acad. Press, Inc., New
York, NY.
Percentage of saponification = (nonsaponified / saponified) × 100.
1
Coon, C. N., and J. R. Couch. 1976. Effect of storage and fatty
2
Percentage of extraction = [saponified total carotenoids (mg/g) / acid esters on the utilization of xanthophyll from marigold
nonsaponified total carotenoids (mg/g)] × 100. meal by laying hens. Poult. Sci. 55:841–847.
Fletcher, D. L. 1989. Factors influencing pigmentation in poultry.
Crit. Rev. Poult. Biol. 2:149–170.
Fletcher, D. L., C. M. Papa, and F. X. Tirado. 1986. The effect of
and corn gluten meal resulted in similar chromatography
saponification on the broiler coloring capability of marigold
profiles for both extraction treatments. The similarity in extracts. Poult. Sci. 65:1708–1714.
results between the 2 treatments indicates that the extract- Galobart, J., R. Sala, X. Ricon-Carruyo, E. G. Manzanilla, B. Vila,
able carotenoids are predominately free and saponifica- and J. Gasa. 2004. Egg yolk color as affected by saponification
tion during extraction has little effect. These results are of different natural pigmenting sources. J. Appl. Poult. Res.
13:328–334.
similar to those reported by Quackenbush (1973).
Goodwin, T. W. 1980. The Biochemistry of the Carotenoids,
The results for the alfalfa sample are interesting. Appar- Volume I, Plants. 2nd ed. Chapman Hall, New York, NY.
ently, the severity of the hot saponification procedure Goodwin, T. W. 1984. The Biochemistry of the Carotenoids,
was sufficient to cause a reduction in total carotenoids Volume II, Animals. 2nd ed. Chapman Hall, New York, NY.
from 178 to 62 mg/kg (Table 1) and a reduction in extrac- Lai, S. M., J. I. Gray, and C. J. Flegal. 1996. Deposition of carot-
enoids in eggs from hens fed diets containing saponified and
tion efficiency to 35% (Table 3). This may reflect the unsta-
unsaponified oleoresin paprika. J. Sci. Food Agric. 72:166–
ble nature of carotenoids in alfalfa and would explain 170.
why alfalfa xanthophyll analyses are highly variable and Marusich, W. L., and J. C. Bauernfeind. 1981. Oxycarotenoids
alfalfa generally is regarded as having poor carotenoid in poultry feeds. In Carotenoids as Colorants and Vitamin A
stability during storage (Middendorf et al., 1980; Fletcher, Precursors. J. C. Bauernfeind, ed. Acad. Press, Inc., New York.
Middendorf, D. F., G. R. Childs, and W. W. Cravens. 1980.
1989). Because the xanthophylls in alfalfa appear to be
Variations in the biological availability of xanthophyll within
more labile to hot KOH than the other samples tested, and among generic sources. Poult. Sci. 59:1460–1470.
the percent extraction [saponified total carotenoids (mg/ Papa, C. M., D. L. Fletcher, and H. R. Halloran. 1985. Utilization
g) ÷ nonsaponified total carotenoids (mg/g) × 100] can and yolk coloring capability of xanthophylls from synthetic
also represent the severity of the saponification process and high xanthophyll concentrates. Poult. Sci. 64:1464–1469.
Philip, T., C. W. Weber, and J. W. Berry. 1976. Utilization of
on extraction efficiency. Thus, the greater the difference in
lutein and lutein fatty acid esters by laying hens. J. Food Sci.
TX content between the nonsaponified and the saponified 41:23–25.
samples, the greater the negative effect of the KOH on Quackenbush, F. W. 1973. Use of heat to saponify xanthophyll
that sample. esters and speed analysis for carotenoids if feed materials.
Results for marigold products A through F ranged from Collaborative study. J. AOAC 56:748–753.
lows of 5.7 and 8.1% saponification up to 100%. Because Quackenbush, F. W., and S. L. Miller. 1972. Composition and
analysis of the carotenoids in marigold petals. J. AOAC
only one of the products tested was known to be nonsa- 55:617–621.
ponified prior to the experiment, it was surprising to find SAS Institute. 1988. SAS/STAT User’s Guide for Personal
2 products with such low relative saponification values. Computers, Release 6.03. SAS Inst., Inc., Cary, NC.