-:
94-
_ _°,
Non-Aqueous
Reverse Phase Purification
of Carotenes
on a Small Particle Preparative Packing
Kevin C. O'Connor and George J. Vella
Chemical Products Division
,Introduction
Carotenoids are a class of naturally occurring hydrocarbons consisting
mostly of C40 molecules varying in the degree of conjugated double bonds. They
are ubiquitous in nature and can be found in fungi, bacteria, photosynthetic and
non-photosynthetic plant tissue such as algae and seaweed. Some birds, fish,
amphibians and reptiles owe their vivid colors to carotenoids. Carotenes and
xanthophylls are subclasses of carotenoids; the xanthophylls are oxidized forms of
the carotenes and both can be dark red to yellow in color. Figure 1 shows a few
of the carotene structures and their absorbance maxima.
Due to the conjugated nature of these molecules considerable interest in the
carotenoids has initiated research endeavors in elucidating the role they play in
photosynthetic processes, vitamin A activity in vision and more recently betacarotene's role as an anticarcinogenic agent 1. Carotenes are also used as food
additives, antioxidants or are available as food supplements. Hence the need for
large amounts of pure carotenoids is of paramount importance.
Both analytical and preparative methods for the separation of the classes of
carotenoids have been reported in the literature 2-7. These methods have utilized
thin layer chromatography, as well as normal phase and reverse phase column
chromatography. Each method has certain advantages and disadvantages and the
conditions of elution depends on the sample matrix and the particular carotenoids
to be separated or isolated.
Described here is a non-aqueous reverse phase method which separates
configurational carotene isomers extracted from a carrot homogenate, and also
permits resolution between cis and trans isomers of beta-carotene. In addition this
methodology is amenable to scale-up so' that larger quantities of purified
compounds can be isolated.
I
Figure 1
Structures
and absorption
maxima
Beta- Carotene
425 450 477
Alpha-Carotene
420 442 472
_ nm )* of carotenes
s.
V
Gamma-Carotene
437 462 492
Delta-Carotene
428 458 490
Zeta-Carotene
Lycopene
380 400 425
448 473 504
* in hexane
2
i ,
Methods
Sample Preparation
Raw Carrots
Blend w/MeOH
& Filter
/
Carrot Pulp
MeOH ( Discard )
Wash with THF
Until Colorless
Pulp ( Discard )
THF Concentrate
Evaporate to Dryness
Crude Carotenes
Dissolve in 25%
Toluene
75% Mobile Phase
HPLC Analysis
Optimization
of Analytical
Separation
The separation of the carotenoids was first optimized on a high resolution
analytical column (3.9 x 300mm) packed with 4gin Nova-Pak C18 material
(Figure 2 A,B). Identification of the various carotenoids in the crude mixture
was confirmed by comparing relative retention times to known standards and by
their characteristic UV/'Vis absorbance spectra which were obtained by using an
on-line Waters 994 photodiode array detector. Spectra were collected between
200nm and 600nm and were critical in the identification of several of these
carotenoids.
Scale Up
The optimized conditions on the analytical column were transferred to a scaling
column ( 8 x 100mm, cartridge column ) packed with a small particle preparative
packing, 6gm Prep Nova-Pak HR C18 (Figure 2C). A loading study was carried
out to determine the maximum amount of sample this column could tolerate
under these conditions without compromising resolution of the components of
interest.
Once established, the flow rate and maximum load on the scaling
column can now be used to determine the conditions to be employed for any
larger volume column packed with the same particle size and of the same length.
The sample load and flow rate are increased proportionally to the cross-section of
the packed bed8 as shown in Equations 1 and 2.
Figure 2"
Scale-up of Crude Carrot Extract from
Analytical to Preparative Particle Sizes
Column: Walers Novo-PakTM CI8, 411m,3.9mm x 300mm
Mobile Phase: CH3CN/CH3OH/THF
58/35/7
Flow Rale: 1.Oral/rain
Tnmperalure'
30.0 C
AU /
cis.zelo-caroleno
0.44
40Onto
_
"co
0.4
\
ly ,pone
o4
04
0 4t
_
I
I
I
alphacarolono
/
/
Irons-zelol
hans belacaroh:ne
A
ammaXX I
"_ao_mnaeX
\_
deha<arotone
\
I
I|
I A/ H
c.bo,o
co,o,o.o
A_JU_Ld
carotene/
I|
!
I
I ........
0.4
0.4
2A
phyli_one
o,°'-_
0d-I
0
I
5
I
!0
I
15
I
20
A_ /L
I
25
_
..................
_
30
Colu,,,n: 8mm x lOOmm RodiaI-Pak" Carlridge packed with Prep Nova.Pak TM CI8,
61tin
Mobile Phase: CH3CN/CH3OH/IIIF
58/35/7
Flow Rale: l.Oml/min
TemperatuJ_: 3qOC
AU
oO203041
O0
0
I
I0
I
20
Minulos
!
30
40
2C
Scaling Equations
Mass (prep) / Mass (scaling) -- r2 (prep) / r2 (scaling)
( 1)
Flow (prep) / Flow (scaling) = r2 (prep) / r2 (scaling)
(2 )
r = column radius
If both column diameter and column length are changing, the flow rate and
sample load can be changed proportionally to column volume, as shown in
Equations 3 and 4 below:
Flow (prep)
Flow (scaling)
= Volume (prep)
= Length (prep) • r2 (prep)
Volume (scaling)
Mass (prep)
Mass (scaling)
=
(3)
Length (scaling) • r2 (scaling)
Length (prep) ° r2 (prep)
(4)
Length (scaling) o r2 (scaling)
This results in identical chromatographic
performance ( retention times, peak
widths and resolution ) at the preparative scale as that of the scaling column. The
flow rate and the maximum sample load on the scaling column were found to be
1.0ml/min ( 4.0cm/min linear velocity ) and 1.2mg ( 1501.tl) respectively.
Employing equations 1-4 the flow rates and the loads were calculated for three
column cartridge configurations, packed with 6p.m Prep Nova-Pak HR C18
material, which were used in a stackable cartridge holder ( Table 1 ).
Chromatography
the preparative
is shown in Figure 3 A,B,C.
Fractions were collected during
separation using 25 x 300 cartridge configuration ( Fig 3A ).
6
Figure 3:
Preparative Separation of Crude Carrot
Extract using Waters PrepPak® Cartridge Columns
Column: Walers Nova-PakTM CI8, 61.1m25mm x I00, 200 and 300mm Carlddges
Mobile Phase: CH3CN/CH3OH/IIIF
58/35/7
Sample: Crude Carrot Exlract
Temperalure: 25.0 C
Pump: Walers 600E Muhisolvenl Delivery System
Delection: UV @ 400nm
Injeclor: Walers U6K wilh a lOml loop
'1
25mm x 300mm
@ 30ml/min,
4.5ml
loaded
\
0.4
t,n 3
0 4-
3A
Fm!
A°
t
__
O.4
0.4
--
O.4_L_
o4-°'_-°^u'_-1
25mm-Od-
I
x 200mm
I
_
I
@ 20ml/min,
_
!
l
--
3.Oral Ioa_
I
h
IFfn8 !
"
I
\4-7
I
'
3B
I
0.A040404(/,t
Au - I
()
_
25ram x' l O0mm
I
@ l Oral/rain,
_I .... 1.5ml Ioaded_I
I0
20
minutes
30
I
dO
------T
...........
SO
:_C
I
Table 1.
Number of Cartridges
Column Volume
Flow Rate
1.0ml/min
Sample Load Run Time
1 x ( 8 xl00mm )*
5.0mls
1 x (25 xlOOmm )
50.Omls
2 x ( 25 xl00mm )
100.0mls
20.0ml/min
24.0mgs (30001Ltl) 35rain
3 x ( 25 xl00mm)
150.0mls
30.0ml/min
36.0mgs (45001Ltl) 35rain
lO.Oml/min
* Scaling cartridge column ( 8 x 100mm )
1.2mgs (150_tl)
35rain
12.0mgs (15001.tl) 35rain
Results and Discussion
Aliquots ( 50-200gi ) of the fractions collected during the preparative
separation were analyzed using the high resolution 4l.t Nova-Pak column.
Samples were injected in the form they were collected i.e. without evaporation or
alteration ( Figures 4-_' ). Fraction 8 was identified as phytoene based on its
strong absorbance a 786nm3. This spectrum is shown in Figure 4. Fractions 1
and 2 have been tentatively identified as cis-zeta-carotene and trans-zeta-carotene
based on their UV/vis spectral maxima and retention relative to the other acyclic
carotenes 5 ( Figures 5 and 6. ).
Fraction 3 ( Figure 7 ) was identified as alpha-carotene by its retention and
spectral characteristics which were identical to those of a standard. These three
fractions together with phytoene represent pure components. Fractions 4,5,6 and
7 contained approximately 80%, 90%, 65% and 30% trans-beta-carotene as the
spectral characteristics were very similar to those of a trans-beta-carotene
standard. These fractions also exhibited absorptions between 310 and 390nm
which had three maxima at 33 lnm, 347nm and 366nm in addition to absorptions
attributable to cis-beta-carotene. A comparison of the three maxima matched
those of phytofluene 3,5.
The amount of phytofluene and cis-beta-carotene appears to vary from
fractions 4 to 7 suggesting that more than one isomer of each may be present.
Several isomers are known to exist for both phytofluene and cis-beta-carotene5,6.
Therefore it is suggested that fraction 4 contained predominantly trans-betacarotene with a small amount of phytofluene and cis-beta-carotene. Fraction 5
appears to be most enriched with trans-beta-carotene
( >90% ) with a small amount of phytofluene. ( Figure 8 ) Fractions 6 and 7
contained decreasing amounts of trans-beta-carotene and what may be other
isomers of cis-beta-carotene and phytofluene. Yet the fact that similar spectra
are observed from noncontiguous fractions suggest that these other isomers are
being resolved from each other however, they coelute with other carotenes found
in this mixture.
9
Figure 4
Analysis of Fraction 8 Containing
Phytoene.
Column: Waters Nova.Fak TM C18, 4Ltm, 3.9ram x 300mm
Mobiie Phase: CH3CN/CH3OH/THF
58/35/7
FiowRate: 1.0ml/min
Temperature: 30.0 C
Detectzon: UV @ 286nm
,._d
0.035
0.030
n..025
0.020
0.015
0.010
0.005
0.000
=
0
5
I0
Fraction 8 ( 46.9-49.2
15
20
/V_nufe$
25
30
35
UV/Vis @ 26.0rain )
AU
0.6
0.9
/A
0.2
0.4,_
0.0
,-
..ii
200
250 300 350
400
i
450 560
5._0 600
am
uv/vis spectrum obtained at _eak
maxima during analytical separation
10
Figure 5"
AU
0.140.12
Analysis of Fraction 1 Containing cis-ZetaCarotene.
Coiumn: Waters Nova-Fak"rMC18, 4u.m,3.9ram x 300ram
Mobile Phase: CH3CN/CH3OH/THF 58/35/7
Flow Rate: 1,0ml/min
Temaerature: 30.0 C
Detection: UV/Vis @ 400nm
0.10
2.._3 0.06
0.040.020.000
I
I
|
I
I
5
10
15
20
25
Min_es
Fraction 1 ( 32.4-33.8 UV/Vis @ 18.9rain )
AU
0.10.
0.05
0.0
200 2.so 30o sso 460 4._o 560 550 6oo
uv/vis spectrumobta,ned at peak
maxima during analvrXcaJ separation
nrn
Figure 6"
Column: Waters Nova-_akTM C18, 4L_m.3.gmm x 300mm
Mobile Phase: CH3CN/CH3OH/TH.: 55/35/7
Flow Rate: i.0ml/min
Temaerature: 30.0 C
• ,-,__
.....
0.C23
Analysis of Fraction 2 Containing trans-ZetaCarotene.
-
Detection:
UV/Vis
@' 40Onto
0.025 0.020
0.015
0.010
0.005.
0.000 .L
mw
0
10
Minutes
15
20
25
Fractian2 ( 34.6-35.4 UV/Vis @ 19.5rain )
AU
0.8.
0.6
0.4
t
0.2.
0.0.
200 2.so 560 s._o 460 _.so 500 550 600
nm
uv/vis spectrumo_ta,ned at peak
max,maduring analytical seDarat_on
Figure 7:
Analysis of Fraction 3 Containing
Carotene.
Alpha-
Coiumn: Wavers Nova-PakTM C18, 4_m, 3.9ram x 300ram
Mobile Phase: CH3CN/CH3OH/THF
58/35/7
AU |
Temperature: 30.0 C
Detection: UV/Vis @ 450nm
Flow Rate: 1.0ml/min
0.0601
0.050
0.040 0.030
0.020
0.010
0.000
k,,.
0
I
I
I
5
10
15
I
20
Minulei
I
i
I
25
25
25
Fraction 3 ( 36.9-38.4 UV/Vis @ 21.3rain )
i
AU
0.15-
0.10-
0.05 -
0.00-
200
I
I
t
I
250
300
350
400
1
450
I
I
500
600
600
rlm
'dV/Vis spectrum obtained a_ peak
maxima
during
analytical
seoorallon
.°
Figure 8"
Analysis of Fraction 5 Containing
Carotene.
,
(
trans-Beta-
Column: Waters Nova-PakTM C18, 4urn 3.9ram x 300ram
Mobile Phase: CH3CN/CH3OH/THF
58/35/7
Flow Rote: !.0ml/min
Temperature: 30.0 C
DetecTion: UV/Vis @ 450nm
0.060-
0.050-
O.04O -
0.030-
0.020-
as
0.000-
•
0
_.
5I
I I0
Fraction 5 ( 40.9-41.7,
;
I
15
Minutes
u_
2b
1
25
I
30
I
35
UV/Vis 22.Tmin )
Aut
0.12
0.I0-
0.080.060.04-
0.020.00200
_
2.S0
360
350
400
430
500
550
600
rim
UV/VJs
maxima
spectrum obtained
at peak
during anatyticai
separation
lh
Conclusion
The difficulty in performing a separation of this type is that all these
compounds are closely related and cannot be separated using one simple
chromatographic method. The separation must be developed to isolate
components of interest. An analytical method for the separation and
identification of carotenes in carrot homogenate has been presented. This
methodology can be used to scale to larger preparative cartridges and can be used
to analyze fractions collected from the preparative separation. The large scale
purification of the carotenoids using a 6gm preparative reversed phase packing
material permitted the isolation of five different compounds as pure components.
Nine different carotenes were tentatively identified by their spectral
characteristics and relative retention times. Positive identification of these would
require the aid of additional structural analysis such as NMR or mass
spectroscopy 5. Alternative mobile or stationary phases may afford sufficiently
different selectivities, thereby resolving compounds which coelute or to separate
components not addressed here or from other sources.
References
1. Burton, G.W. and Ingold, K.U., beta-Carotene: An unusual type of lipid
antioxidant. Science 224, 569-575, (1984)
2. Neils, H.J. and De Leenheer, A.P., Isocratic nonaqueous reverse phase liquid
chromatography of carotenoids. Analytical Chemistry. 55,270-275, (1983)
3. Tan, B. and Andersson, A., Analytical and preparative chromatography
tomato paste carotenoids. J. Food Science, 53(3), 954-959, (1988)
of
4. Bushway, R2., Separation of carotenoids in fruits and vegetables by high
performance liquid chromatography. J. Liquid Chrom., 8(81, 1527-1547, (1985)
5. Davies, B.H., Carotenoids In " Chemistry and Biochemistry of Plant
Pigments," T.W. Edwin (Ed.)
Vol. 2, pg. 38 Academic Press, Orlando FL.
(1976) and references therein
6. Quackenbush F.W., Reverse phase HPLC separation of cis- and transcarotenoids and its application to beta-carotene in food materials. J Liquid
Chrom., 10(4), 643-653, (1987)
7. Scalia, S. and Francis, G.W., Preparative scale reversed-phase
for simultaneous separation of carotenoids and carotenoid esters.
Chromatographia,
28(3,4)_, 129-132, ( 1989 )
8. Neue, U. Private Communication
HPLC method
(1990)
16
P.57