Planta (1992)187:185-191
P l a n t a 9 Springer-Verlag1992
Abscisic acid biosynthesis in roots I. The identification of potential abscisic acid precursors, and other carotenoids Andrew D. Parry and Roger Horgan Department of Biological Sciences, The University College of Wales, Aberystwyth,Dyfed SY23 3DA, UK Received 25 September; accepted 7 November 1991
Abstract. The pathway of water-stress-induced abscisic acid (ABA) biosynthesis in etiolated and light-grown leaves has been elucidated (see A.D. Parry and R. Horgan, 1991, Physiol. Plant. 82, 320-326). Roots also have the ability to synthesise ABA in response to stress and it was therefore of interest to examine root extracts for the presence of carotenoids, including those known to be ABA precursors in leaves. All-trans- and 9'-cisneoxanthin, all-trans- and 9-cis-violaxanthin, antheraxanthin (all potential ABA precursors), lutein and p-carotene were identified on the basis of absorbance spectra, reactions with dilute acid, retention times upon high-performance liquid chromatography and by comparison with leaf carotenoids that had been analysed by mass spectrometry. The source of the extracted carotenoids was proved to be root tissue, and not contaminating compost or leaf material. The levels of total carotenoids in roots varied between 0.03-0.07% of the levels in light-grown leaves (Arabidopsis thaliana (L.) Heynh, Ni~ cotiana plumbaginifolia Viv., Phaseolus vulgaris L. and Pisum sativum L.) up to 0.27% (Lycopersicon esculentum Mill.). The relative carotenoid composition was very different from that found in leaves, and varied much more between species. All-trans-neoxanthin and violaxanthin were the major carotenoids present (64-91% of the total), but while Lycopersicon contained 67-80% alltrans-neoxanthin, Phaseolus, Pisum and Zea mays L. contained 61-79% all-trans-violaxanthin. Carotenoid metabolism also varied between species, with most of the carotenoids in older roots of Phaseolus being esterified. Roots and leaves of the ABA-deficient aba mutant of Arabidopsis had reduced epoxy-xanthophyll levels compared to the wild-type. Key words: Abscisic acid biosynthesis (precursors) - Carotenoid - Lycopersicon (ABA biosynthesis) - Neoxanthin (isomers) - Root (abscisic acid) - Violaxanthin (isomers) Abbreviations: ABA= abscisic acid; r.p.HPLC= reversed-phase high performanceliquid chromatography
Introduction Abscisic acid (ABA) has been identified and quantified in root extracts, and exudates from roots, on numerous occasions and in many species (Audus 1983; Davies and Zhang 1991). Although ABA can be translocated from the shoot to the roots within the phloem (Zeevaart and Boyer 1984; Wolf et al. 1990), experiments involving phloem-blocking and using detached roots have demonstrated that roots do possess the ability to synthesise ABA (Barr 1973; Walton et al. 1976; Cornish and Zeevaart 1985; Lachno and Baker 1986; Zhang and Davies 1987). The importance of root-derived ABA in plant/soil water relations, acting as a chemical messenger, has been recently reviewed (Davies and Zhang 1991). Within the roots, ABA may also influence both hydraulic conductivity (Glinka 1980) and growth (Sharp 1990). After years of uncertainty it is now known that, at least in water-stressed leaves, ABA biosynthesis relies on the asymmetric cleavage of the xanthophyU 9"-cisneoxanthin (see Parry and Horgan 1991a). Experiments utilising ~80 2 have indicated indirectly that the pathway in roots is the same as that in leaves (Creelman et al. 1987), and the use of cell-free systems have demonstrated that roots possess the enzymes necessary for the conversion of xanthoxin to ABA (Sindhu and Walton 1987), i.e. the post-cleavage part of the biosynthetic pathway. With the exception of carrot and sweet potato the carotenoid content of roots has received little attention (Goodwin 1980), although all-trans-violaxanthin was identified as the major carotenoid in root caps of Zea mays (Maudinas and Lematre 1979). Prior to investigating ABA biosynthesis in roots it was therefore necessary to examine the occurrence and distribution of carotenoids, especially those known to be potential ABA precursors, in roots of a variety of species. The accompanying paper (Parry et al. 1992) discusses the metabolism of specific xanthophylls in relation to stress-induced ABA biosynthesis in roots of soil-grown and hydroponicallygrown plants of Lycopersicon esculentum.
186
A.D. Parry and R. Horgan: Abscisic acid biosynthesis in roots. I
Materials and methods
Results
Plant material. Seeds of Lycopersicon esculentum Mill. cv. Ailsa Craig were a gift from Dr. I.B. Taylor, University of Nottingham School of Agriculture, Sutton Bonington, Leicestershire, UK, while seeds of Arabidopsis thaliana (L.) Heynh. cv. Lansberg erecta and line A26 (homozygous for the aba1 mutant allele) were a gift from Dr. M. Koornneef, Department of Genetics, Wageningen Agricultural University, The Netherlands. Seeds of Arabidopsis were germinated on 1% agar (containing half-strength Murashige and Skoog (1962) salts) in Petri dishes. Germination required pre-treatment at 4 ~ C for 4 d followed by exposure to red light for 8 h at 20~ C, after which the petri dishes were placed in a growth cabinet (22 ~ C, 16 h of light daily). After 10 d the seedlings were transferred to Levington Universal compost and kept at a high-humidity. Light-grown plants of Lycopersicon, Nicotiana plumbaginifolia Viv. (seeds from Dr. P.J. King, Friedrich-Miescher Institut, Basel, Switzerland), Phaseolus vuloaris L. cv. Canadian Wonder and Pisum sativum L. cv. Meteor were grown as described previously, as were etiolated seedlings of Lycopersicon, Phaseolus and Zea rnays. L. cv. Golden Bantam (Parry et al 1990; Parry and Horgan 1991a). Unless otherwise stated all seed was obtained from Booker Seeds, Sleaford, Lincs., UK. Roots were carefully washed free of compost, blotted dry and extracted immediately or frozen in liquid nitrogen. All frozen sampies, light-grown and etiolated leaves and roots, were stored at - 70~ C.
Identification o f carotenoids in roots. The procedure a d o p t e d for the extraction o f carotenoids f r o m roots was based on one developed and validated previously (Parry et al. 1990). All parts o f the procedure were carried out in the d a r k or dim light and exposure to extremes o f p H or high temperatures was avoided, to prevent the breakd o w n or rearrangement o f the carotenoids. This was j u d g e d to have been successful as the results obtained were entirely reproducible, and no b r e a k d o w n p r o d u c t s (furanoids or apo-carotenoids) were detected. Extraction was complete within 15 min a n d a single r . p . H P L C step was sufficient to fractionate all o f the carotenoids o f interest. The data used for the identification o f the carotenoids extracted are presented in Tables 1 and 2. A n extract o f Lyeopersicon esculentum roots (5 g F W ) was concentrated and the total analysed by r . p . H P L C (50 ~tl injection). Fractions were collected (real-time m o n i t o r i n g o f the eluate was at 440 nm), reduced to dryness u n d e r v a c u u m at 30 ~ C, and the residues dissolved in 50 ~tl o f ethanol. The low concentrations o f carotenoids necessitated the use o f a micro-cuvette for s p e c t r o p h o t o m e t r y . The absorbance spectra obtained are c o m p a r e d with published values in Table 1. F u r t h e r evidence for the identity o f these c o m p o u n d s comes f r o m their reactions with acid. Acid catalyses the specific isomerisation o f 5,6-epoxide g r o u p s to 5,8-furanoid groups, resulting in a h y p s o c h r o m i c shift o f approx. 20 n m for m o n o epoxides such as neoxanthin (Fig. 1) and 40 n m for di-epoxides such as violaxanthin (Fig. l; Davies 1976). The carotenoids identified in roots o f Lyeopersicon were neoxanthin, violaxanthin, antheraxanthin, lutein and 13-carotene. I f an a p o - c a r o t e n o i d A B A biosynthetic p a t h w a y operates in roots, and X a n t h o x i n is the immediate postcleavage intermediate as in leaves (Parry and H o r g a n 1991a), then the X a n t h o x i n precursor m u s t be a 9-cis isomer o f a xanthophyll such as neoxanthin or violaxanthin (Sindhu and W a l t o n 1987; Parry et al. 1988; 1990; P a r r y and H o r g a n 1991b). Carotenoids containing eis double b o n d s a b s o r b at lower wavelengths than the c o r r e s p o n d i n g all-trans isomers (approx. 2-5 n m per cis double bond), and so it is possible to tentatively assign geometric configurations to m a n y carotenoids on the basis o f their absorbance spectra ( W e e d o n 1971; Davies 1976; M o l n a r et al. 1986). In addition, the location o f the cis double bond(s) determines the intensity (Q) o f the so-called "cis" peak, a subsidiary peak in the nearultraviolet (where Q is the ratio o f the absorbance at )~max to the m a x i m u m absorbance in the cis peak; M o l n a r et al. 1986). Both all-trans and 9-cis isomers have small eis peaks and consequently large Q values. The spectra for the two principal isomers o f neoxanthin isolated are consistent with them being all-trans-neoxanthin (Q = 20.0) and 9"-cis-neoxanthin (Q = 7.2). A n u n k n o w n carotenoid eluting just after 9"-eis-neoxanthin could have been a third neoxanthin isomer (retention time (Rt) 23.1 min, )~m,x ethanol 463, 436, 412, Q = 1.9), possibly the lY-cis isomer, occurring at levels o f only 4% o f the
Determination and identification of earotenoids. Root samples (3-10 g FW) were extracted in approx. 25 ml of cold ( - 2 0 ~ C) ethanol (containing 0.1% v/v re-distilled tri-ethylamine [TEA]) using a Polytron blender (Kinematica, Lucerne, Switzerland). The extracts were then centrifuged at 1500 g in a bench-top centrifuge (IEC Centra-4X; Damon/IEC (UK), Dunstable, Beds.) for 5 rain, and the supernatant decanted. The pellets were washed with a further 10 ml of ethanol (+ 0.1% TEA), the extracts recentrifuged, and the combined supernatants reduced to dryness, under vacuum at 30~ C. The extracts were re-dissolved in ethanol (0.2-0.5 ml) and an aliquot (50 ~tl) analysed by reversed-phase high-performance liquid chromatography (r.p.HPLC), The procedure for light-grown and etiolated leaves was as described in Parry et al 1990. A Waters 600E system linked to a Waters 990 photodiode-array detector (Millipore Corporation, Waters Chromatography Division, Milford, Mass., USA) was used for HPLC. For r.p.HPLC an ODS-Spherisorb column (Phase Separations, Queensferry, Clwyd, UK; 250 mm long, 4.5 mm i.d.) was utilised. The column was eluted at 0.7 ml 9min- ~with a linear gradient of 10-50% Y in X over 25 rain, followed by a further 15 min at 100% Y, where X is 85% methanol (+ 0.1% TEA; v/v) and Y is 6: 4 dichloromethane :methanol. The eluate was monitored between 250-500 nm and spectra recorded at 3-s intervals. Absorbance spectra were recorded in ethanol using a Beckmann spectrophotometer (Beckmann Instruments Inc., Fullerton, Cal., USA) and micro-cuvette (50 Ixl), or in running solvent following r.p.HPLC using the photodiode-array detector. Tests for the presence of 5,6-epoxide groups were carried out by taking spectra in ethanol before and after the addition of 1.0 M hydrochloric acid (< 10 Ixl). Authentic zeaxanthin was a gift from Dr. B.H. Davies, Department of Biochemistry, The University College of Wales, Aberystwyth. Quantitation of carotenoids was based on their absorbance at )~.... and the use of suitable calibration curves. Saponification. This was performed on a small scale after Gregory et al. (1987). Approx. 5 ml of 10% potassium hydroxide in methanol (w/v) was added to 100 ~tl of ethanolic carotenoid extract, the mixture purged with nitrogen and heated for 5 rain at 50~ C in the dark. The carotenoids were recovered by partitioning twice against diethyl ether. The ether was washed with water to remove traces of alkali before being reduced to dryness under vacuum at 30~ C. The residue was dissolved in ethanol and analysed by r.p.HPLC.
A.D. Parry and R. ttorgan: Abscisic acid biosynthesis in roots. I Table 1. Absorbance maxima (in ethanol)
of carotenoids extracted from roots of Lycopersicon compared with published values, and their hypsochromic shifts following the addition of acid
187
Carotenoid
Extracted ~'ma~
All-trans-neoxanthin 9"-cis-neoxanthin All-trans-violaxanthin 9-cis-violaxanthin Antheraxanthin Lutein B-Carotene
471,443, 468, 441, 470; 442, 465, 438, 469, 443, 474, 446, 477, 450,
Published Lm~ 420 417 419 416 422 423 (428)
470, 467, 469, 464, 472, 474, 475,
441, 438, 440, 436, 444, 445, 449,
422 415 417 415 422 422 (427)
Referencea
Shift plus acid (nm)
1 t 2 2 2 2 2
20-21 18-19 41-43 37-39 18-22 0-1 0
a: 1, Cholnoky et al. 1969; 2, Davies 1976 Table 2. Absorbance maxima (in eluate) and r.p.HPLC retention times of carotenoids extracted from leaves and roots of Lycopersicon
Carotenoid
Root L~x
All-trans-neoxanthin 9"-cis-neoxanthin All-trans-violaxantlfin 9-cis-violaxanthin Antheraxanthin Lutein l-Carotene
Off afl-fra ns-violoxanthin 9-cis-
all-frans-neoxonfhin
H
O
~
OH
9'-cis-ne
an~heroxonfhin
H
O
~ lufein
.6
~ 8-coroIene
Fig. 1. Structures of the carotenoids extracted from roots
combined levels of the other neoxanthin isomers. Similarly the spectra for the two violaxanthin isomers indicated they were all-trans-violaxanthin (Q= 13.4) and 9-cis-violaxanthin (Q = 5.7). Antheraxanthin and lutein were both present as the all-trans isomers (Q = 7.2 and 26.0 respectively). In previous studies (Parry et al. 1990), carotenoids extracted from etiolated and light-grown leaves were conclusively identified by mass spectrometry and from the quasi-equilibria resulting from either iodinecatatysed or chlorophyU-sensitised photoisomerisation. The low levels of carotenoid present in roots precluded the use of either of these methods. However, Table 2 compares the absorbance spectra recorded using the photodiode-array detector and the R,s on r.p.HPLC of carotenoids from roots and leaves of Lycopersicon. This established that the carotenoids tentatively identified from their absorbance spectra in ethanol and reactions with dilute acid were the same as the ones extracted from leaves, and therefore confirmed their identities.
473, 467, 473, 467, 475, 477, 483,
443, 439, 443, 439, 447, 449, 457,
419 415 4t8 415 (418) (422) (432)
Rt (min)
~.~
Leaf
21.9 22.7 25.2 27.3 28.4 31.3 39.9
472, 468, 472, 466, 474, 478, 485,
Rt (min) 444, 439, 443, 439, 446, 450, 457,
419 415 418 415 (420) (422) (430)
21.8 22.7 25.2 27.3 28.4 31.2 39.8
Quantification of carotenoids. Using the procedure described above, carotenoids were extracted from roots of a variety of species. The levels of the xanthophylls, the oxygenated carotenoids, are shown in Table 3. For comparison the levels of carotenoids in light-grown and etiolated leaves of some of the same species are shown in Tables 4 and 5. Qualitatively there was no difference between the distribution of carotenoids in roots of the various species examined, or between the roots and either etiolated or light-grown leaves. Neoxanthin, violaxanthin, antheraxanthin and lutein were the only xanthophylls identified, and [3-carotene the only carotene. Prior to quantification the identity of each carotenoid was confirmed from absorbance spectra recorded during r.p.HPLC and Rt. The levels of total carotenoids in the roots varied between 0.03-0.07% of the levels in light-grown leaves, for Arabidopsis, Nicotiana, Phaseolus and Pisum, up to 0.27 % for Lycopersicon. The percentage of total carotenoids in roots compared with etiolated leaves was between 0.32% (Lycopersicon) and 1.7% (Zea). As the levels of carotenoid in the root extracts were so low it was possible that their presence was due to contamination by fragments of shoot material or residual carotenoids in the compost. The latter possibility was investigated by extracting approx. 5 g of both Levington Universal and John Innes No.2 composts using the same procedure as for roots. In both cases no trace of carotenoid was detected, with the exception of lutein. However, the amount of lutein was equivalent, on a fresh-weight basis, to only 2-4% of that in roots. As the amount of compost extracted with the root samples was probably less than 5 % of the root fresh weight, the contribution of lutein from the compost would have been negligible. Contamination of a 5' g FW root sample with only 1-10 mg of leaf tissue (i.e. 0.02-0.2 %) would be sufficient to give the observed levels of carotenoids. However a
188
A.D. Parry and R. Horgan: Abscisic acid biosynthesis in roots. I
Table 3, Xanthophyll composition in roots of a variety of plants Age (weeks) Arabidopsis thaliana Lycopersicon esculentum
3 3 13
Nicotiana plumbaginifolia Phaseolus vulgaris
10 2 13
Pisum sativum Zea mays
2 2
Xanthophyll (nmol - (g FW)-1, (% of total xanthophylls)) t-neo
c-neo
t-viola
c-viola
anthera
lutein
total
0.085 (39) 0.287 (80) 0.797 (67) 0.099 (46) 0.042 (11) 0.002 (1) 0.012 (6) 0.024 (2)
0.032 (15) 0.009 (2) 0.042 (3) 0.013 (6) 0.005 (1) 0.002 (t) 0.003 (1) 0.090 (9)
0.055 (25) 0.041 (t 1) 0.287 (24) 0.077 (37) 0.227 (73) 0.089 (61) 0.165 (79) 0.746 (78) ,
0.006 (3) 0.003 (1) 0.023 (2) 0.006 (3) 0.019 (6) 0.022 (15) 0.015 (7) 0.017 (2)
0.005 (2) 0.002 ( < 1) 0,005 (< l) 0.005 (2) 0.010 (3) 0.005 (3) 0.003 (1) 0.041 (4)
0.033 (15) 0.021 (6) 0.035 (3) 0.010 (5) 0.019 (6) 0.029 (19) 0.011 (5) 0.037 (4)
0.218 0.360 1.190 0.210 0.322 0.149 0.194 0.956
"t- = alt-trans; c- = 9/9"-cis; neo = neoxanthin; viola = violaxanthin; an thera = antheraxanthin Table 4, Xanthophyll composition in light-grown leaves of a variety of plants Age (weeks) Arabidopsis thaliana Lycopersicon esculentum Nicotiana plumbaginifolia Phaseolus vulgaris Pisum sativum
3 13 10 13 2
Xanthophyll (nmol " (g FW)-1, (% of total xanthophylls)) t-neo
c-neo
t-viola
c-viola
anthera
lutein
total
1.3 (< 1) 1.5 (< 1) 0.6 (< 1) 0.7 (< 1) 3.1 (1)
52.1 (17) 85.3 (19) 63.7 (16) 84.1 (19) 73.3 (16)
72.7 (25) 91.1 (21) 123.9 (31) 94.5 (22) 155.0 (35)
0.4 (< 1) 0.6 (< 1) n.d. (-) 0.3 (< 1) 0.9 ( < 1)
4.7 (2) 4.2 (1) 2.6 (1) 10.2 (2) 3.6 (1)
164.6 (55) 254.6 (58) 205.4 (52) 246.7 (56) 225.7 (48)
296 438 396 436 462
"t- = all-trans; c- = 9/9'-cis; neo = neoxanthin; viola = violaxanthin; anthera = antheraxanthin; n.d. = not detected detailed c o m p a r i s o n o f xanthophyll profiles o f roots and leaves reveals large differences. The relative xanthophyll composition in all of the light-grown leaves examined was very consistent (Table 4), the three major xanthophylls being 9"-cis-neoxanthin (16-19 % of the total carotenoid), all-trans-violaxanthin (21-35%) and lutein (48-58%). All-trans-neo xanthin, 9-cis-violaxanthin and antheraxanthin were all m i n o r c o m p o n e n t s ( < 1-2%). The levels o f l)-carotene present were between 28-33 % o f those o f the total xanthophylls (data not shown). This is consistent with findings reported in the literature (Goodwin 1980; G o o d w i n and Britton 1988). In etiolated leaves the situation was slightly different, and the carotenoid distribution m o r e varied (Table 5). Lutein and all-transviolaxanthin remained m a j o r carotenoids but the relative levels o f 9"-cis-neoxanthin were lower. In etiolated leaves o f Phaseolus the relative levels of antheraxanthin were elevated. In addition to the absolute levels being m u c h lower in roots, the relative distribution o f carotenoids was m u c h
m o r e varied between species, and very different f r o m that found in leaves. F o r example all-trans-neoxanthin varied between being present as 1-80 % of the total root xanthophylls, and all-trans-violaxanthin between 11-79%. The all-trans isomers o f neoxanthin and violaxanthin were the m a j o r carotenoids found in the roots examined here (the combined levels making up 64-91% o f the total xanthophyll levels). The levels of 13-carotene varied between 3% (Lycopersicon) and 12% (Pisum) o f the total xanthophyll levels (data not shown). The species listed in Table 3 appear to fall into three catorgaries: firstly, Lycopersicon is unusual in having a very large percentage of all-trans-neoxanthin (67-80%); secondly, there are those species whose roots contain approximately equal amounts o f all-trans-neoxanthin and all-trans-violaxanthin, such as Arabidopsis and Nicotiana; and thirdly, those such as Phaseolus, Pisum and Zea which contain predominantly all-trans-violaxanthin (61-79%). The data for Zea are consistent with those previously published (Maudinas and Lematre 1979; Feldman et al. 1985). In
189
A.D. Parry and R. Horgan: Abscisic acid biosynthesis in roots. I Table 5. Xanthophyll composition in etiolated leaves of a variety of plants
Lycopersicon esculentum Phaseolus vulgaris Zea mays
Age (weeks)
Xanthophyll (nmol 9(g FW)- 1, (% of total xanthophylls)) t-neo c-neo t-viola c-viola anthera
lutein
total
1
4.8 (4) 1.6 (3) n.d. (-)
61.0 (54) 23.3 (38) 31.2 (57)
113
1 2
7.0 (6) 4.1 (7) 1.7 (3)
37.7 (33) 25.4 (41) 21.1 (38)
1.8 (2) 0.6 (< 1) 0.3 (< 1)
0.8 (1) 6.8 (11) 0.4 (1)
62 55
at- = all-trans; c- = 9/9'-cis; neo = neoxanthin; viola = violaxanthin; anthera = antheraxanthin; n.d. = not detected
Table 6. Properties of suspected violaxanthin esters extracted from 13-week-old Phaseolus roots Isomer
All-trans9-cis-
a Recorded in ethanol
Non-saponified ~,m~a
Rtb
% of total carotenoid
469, 440, 417 466, 437, 412
39.8 38.2
28 57
Saponified ~'max
Rt
% of total carotenoid
468, 438, 415 466, 436, 413
24.9 27.0
33 58
Shift plus acid (nm) 39-42 37-43
b r.p.HPLC
leaves, especially light-grown ones, the predominant neoxanthin isomer is 9"-cis (Tables 4, 5), whilst in the roots o f the majority of species shown in Table 3 the all-trans isomer is predominant. It therefore seems unlikely that the carotenoids identified in the root extracts are derived from contaminating photosynthetic tissue. Measurements were made o f carotenoid levels in roots o f different aged plants of Lycopersicon and Phaseolus. The total levels in roots o f Lycopersicon were 3.3-fold higher in 13-week-old roots compared with those o f 3-week-old ones. The composition was similar except that the relative amounts o f all-trans-neoxanthin were lower and those of all-trans-violaxanthin higher in the older roots. In contrast the total levels o f carotenoid in roots of 13-week-old Phaseolus plants appeared to be only half those at 2 weeks. Interestingly the levels of all-trans-neoxanthin and all-trans-violaxanthin were much lower whilst those o f 9-cis-violaxanthin increased with age. Analysis by r.p.HPLC of extracts from the older (13-week) roots of Phaseolus revealed the presence of two large peaks eluting just prior to I]-carotene. Their spectra (Table 6) were very similar to those o f the xanthophylls but their Rts were characteristic o f compounds much less polar. Such compounds had been observed previously in extracts of orange fruit peel (data not shown). In oranges purchased early in the season the predominant carotenoids were all-trans- and 9-cis-violaxanthin, but in fruit purchased much later the normal levels of violaxanthin isomers were not observed. In their place large peaks of non-polar compounds possessing similar spectra were found. These were presumed to be violaxanthin esters. A carotenoid extract from 13-weekold Phaseolus roots was saponified to investigate the presence of esterified carotenoids. The procedure chosen was particularly mild, to avoid possible xanthophyll rearrangement and breakdown (Davies 1976; Gregory et al.
1987). The carotenoids were removed by partitioning from the saponified extract and analysed by r.p.HPLC. The two peaks predominant in the non-saponified extract disappeared following saponification, and were replaced by comparable peaks o f all-trans- and 9-cis-violaxanthin (identified by Rt, absorbance spectra and reaction with dilute acid; Table 6). In saponified extracts of 13-weekold Phaseolus roots estimates of total xanthophyll levels ranged from 0.5 to 2.3 nmol- (g FW) -1, that is 1.7-7.0fold higher than in roots o f 2-week-old plants. The violaxanthin esters were not further characterised, nor were they detected in young Phaseolus roots or roots of any other species examined. Arabidopsis mutants deficient in A B A . During the course
of this study it was revealed that the cause o f the ABAdeficiencies in leaves of the aba mutants of Arabidopsis was a deficiency of neoxanthin and violaxanthin (Duckham et al. 1991 ; Rock and Zeevaart 1991). The mutation is believed to affect the conversion of zeaxanthin, which accumulates in the mutant leaves, to antheraxanthin. It was therefore of interest to measure the xanthophyll levels in the roots of such a mutant (Table 7). The xanthophyll composition of wild-type and mutant leaves were consistent with previous findings (Duckham et al. 1991; Rock and Zeevaart 1991), with the epoxy-xanthophylls being reduced from 45 % o f the total xanthophylls in the wild-type to approx. 8 % in the mutant. This difference was more pronounced in the roots, with epoxyxanthophylls constituting 85% of the total xanthophylls in the wild-type but only 8 % in the mutant roots. Discussion The imposition of water-stress leads to the accumulation of ABA in leaves and roots, as the consequence o f an increased rate of biosynthesis (Zeevaart and Creelman
190
A.D. Parry and R. Horgan: Abscisic acid biosynthesis in roots. I
Table 7. Xanthophyll composition in leaves and roots of normal and mutant (aba, A26) Arabidopsis thaliana Xanthophyll (nmol 9(g F W ) - l, (% of total xanthophylls)) t-neo
c-neo
t-viola
c-viola
anthera
lutein + zea a
total
1.3 (
P l a n t a 9 Springer-Verlag1992
Abscisic acid biosynthesis in roots I. The identification of potential abscisic acid precursors, and other carotenoids Andrew D. Parry and Roger Horgan Department of Biological Sciences, The University College of Wales, Aberystwyth,Dyfed SY23 3DA, UK Received 25 September; accepted 7 November 1991
Abstract. The pathway of water-stress-induced abscisic acid (ABA) biosynthesis in etiolated and light-grown leaves has been elucidated (see A.D. Parry and R. Horgan, 1991, Physiol. Plant. 82, 320-326). Roots also have the ability to synthesise ABA in response to stress and it was therefore of interest to examine root extracts for the presence of carotenoids, including those known to be ABA precursors in leaves. All-trans- and 9'-cisneoxanthin, all-trans- and 9-cis-violaxanthin, antheraxanthin (all potential ABA precursors), lutein and p-carotene were identified on the basis of absorbance spectra, reactions with dilute acid, retention times upon high-performance liquid chromatography and by comparison with leaf carotenoids that had been analysed by mass spectrometry. The source of the extracted carotenoids was proved to be root tissue, and not contaminating compost or leaf material. The levels of total carotenoids in roots varied between 0.03-0.07% of the levels in light-grown leaves (Arabidopsis thaliana (L.) Heynh, Ni~ cotiana plumbaginifolia Viv., Phaseolus vulgaris L. and Pisum sativum L.) up to 0.27% (Lycopersicon esculentum Mill.). The relative carotenoid composition was very different from that found in leaves, and varied much more between species. All-trans-neoxanthin and violaxanthin were the major carotenoids present (64-91% of the total), but while Lycopersicon contained 67-80% alltrans-neoxanthin, Phaseolus, Pisum and Zea mays L. contained 61-79% all-trans-violaxanthin. Carotenoid metabolism also varied between species, with most of the carotenoids in older roots of Phaseolus being esterified. Roots and leaves of the ABA-deficient aba mutant of Arabidopsis had reduced epoxy-xanthophyll levels compared to the wild-type. Key words: Abscisic acid biosynthesis (precursors) - Carotenoid - Lycopersicon (ABA biosynthesis) - Neoxanthin (isomers) - Root (abscisic acid) - Violaxanthin (isomers) Abbreviations: ABA= abscisic acid; r.p.HPLC= reversed-phase high performanceliquid chromatography
Introduction Abscisic acid (ABA) has been identified and quantified in root extracts, and exudates from roots, on numerous occasions and in many species (Audus 1983; Davies and Zhang 1991). Although ABA can be translocated from the shoot to the roots within the phloem (Zeevaart and Boyer 1984; Wolf et al. 1990), experiments involving phloem-blocking and using detached roots have demonstrated that roots do possess the ability to synthesise ABA (Barr 1973; Walton et al. 1976; Cornish and Zeevaart 1985; Lachno and Baker 1986; Zhang and Davies 1987). The importance of root-derived ABA in plant/soil water relations, acting as a chemical messenger, has been recently reviewed (Davies and Zhang 1991). Within the roots, ABA may also influence both hydraulic conductivity (Glinka 1980) and growth (Sharp 1990). After years of uncertainty it is now known that, at least in water-stressed leaves, ABA biosynthesis relies on the asymmetric cleavage of the xanthophyU 9"-cisneoxanthin (see Parry and Horgan 1991a). Experiments utilising ~80 2 have indicated indirectly that the pathway in roots is the same as that in leaves (Creelman et al. 1987), and the use of cell-free systems have demonstrated that roots possess the enzymes necessary for the conversion of xanthoxin to ABA (Sindhu and Walton 1987), i.e. the post-cleavage part of the biosynthetic pathway. With the exception of carrot and sweet potato the carotenoid content of roots has received little attention (Goodwin 1980), although all-trans-violaxanthin was identified as the major carotenoid in root caps of Zea mays (Maudinas and Lematre 1979). Prior to investigating ABA biosynthesis in roots it was therefore necessary to examine the occurrence and distribution of carotenoids, especially those known to be potential ABA precursors, in roots of a variety of species. The accompanying paper (Parry et al. 1992) discusses the metabolism of specific xanthophylls in relation to stress-induced ABA biosynthesis in roots of soil-grown and hydroponicallygrown plants of Lycopersicon esculentum.
186
A.D. Parry and R. Horgan: Abscisic acid biosynthesis in roots. I
Materials and methods
Results
Plant material. Seeds of Lycopersicon esculentum Mill. cv. Ailsa Craig were a gift from Dr. I.B. Taylor, University of Nottingham School of Agriculture, Sutton Bonington, Leicestershire, UK, while seeds of Arabidopsis thaliana (L.) Heynh. cv. Lansberg erecta and line A26 (homozygous for the aba1 mutant allele) were a gift from Dr. M. Koornneef, Department of Genetics, Wageningen Agricultural University, The Netherlands. Seeds of Arabidopsis were germinated on 1% agar (containing half-strength Murashige and Skoog (1962) salts) in Petri dishes. Germination required pre-treatment at 4 ~ C for 4 d followed by exposure to red light for 8 h at 20~ C, after which the petri dishes were placed in a growth cabinet (22 ~ C, 16 h of light daily). After 10 d the seedlings were transferred to Levington Universal compost and kept at a high-humidity. Light-grown plants of Lycopersicon, Nicotiana plumbaginifolia Viv. (seeds from Dr. P.J. King, Friedrich-Miescher Institut, Basel, Switzerland), Phaseolus vuloaris L. cv. Canadian Wonder and Pisum sativum L. cv. Meteor were grown as described previously, as were etiolated seedlings of Lycopersicon, Phaseolus and Zea rnays. L. cv. Golden Bantam (Parry et al 1990; Parry and Horgan 1991a). Unless otherwise stated all seed was obtained from Booker Seeds, Sleaford, Lincs., UK. Roots were carefully washed free of compost, blotted dry and extracted immediately or frozen in liquid nitrogen. All frozen sampies, light-grown and etiolated leaves and roots, were stored at - 70~ C.
Identification o f carotenoids in roots. The procedure a d o p t e d for the extraction o f carotenoids f r o m roots was based on one developed and validated previously (Parry et al. 1990). All parts o f the procedure were carried out in the d a r k or dim light and exposure to extremes o f p H or high temperatures was avoided, to prevent the breakd o w n or rearrangement o f the carotenoids. This was j u d g e d to have been successful as the results obtained were entirely reproducible, and no b r e a k d o w n p r o d u c t s (furanoids or apo-carotenoids) were detected. Extraction was complete within 15 min a n d a single r . p . H P L C step was sufficient to fractionate all o f the carotenoids o f interest. The data used for the identification o f the carotenoids extracted are presented in Tables 1 and 2. A n extract o f Lyeopersicon esculentum roots (5 g F W ) was concentrated and the total analysed by r . p . H P L C (50 ~tl injection). Fractions were collected (real-time m o n i t o r i n g o f the eluate was at 440 nm), reduced to dryness u n d e r v a c u u m at 30 ~ C, and the residues dissolved in 50 ~tl o f ethanol. The low concentrations o f carotenoids necessitated the use o f a micro-cuvette for s p e c t r o p h o t o m e t r y . The absorbance spectra obtained are c o m p a r e d with published values in Table 1. F u r t h e r evidence for the identity o f these c o m p o u n d s comes f r o m their reactions with acid. Acid catalyses the specific isomerisation o f 5,6-epoxide g r o u p s to 5,8-furanoid groups, resulting in a h y p s o c h r o m i c shift o f approx. 20 n m for m o n o epoxides such as neoxanthin (Fig. 1) and 40 n m for di-epoxides such as violaxanthin (Fig. l; Davies 1976). The carotenoids identified in roots o f Lyeopersicon were neoxanthin, violaxanthin, antheraxanthin, lutein and 13-carotene. I f an a p o - c a r o t e n o i d A B A biosynthetic p a t h w a y operates in roots, and X a n t h o x i n is the immediate postcleavage intermediate as in leaves (Parry and H o r g a n 1991a), then the X a n t h o x i n precursor m u s t be a 9-cis isomer o f a xanthophyll such as neoxanthin or violaxanthin (Sindhu and W a l t o n 1987; Parry et al. 1988; 1990; P a r r y and H o r g a n 1991b). Carotenoids containing eis double b o n d s a b s o r b at lower wavelengths than the c o r r e s p o n d i n g all-trans isomers (approx. 2-5 n m per cis double bond), and so it is possible to tentatively assign geometric configurations to m a n y carotenoids on the basis o f their absorbance spectra ( W e e d o n 1971; Davies 1976; M o l n a r et al. 1986). In addition, the location o f the cis double bond(s) determines the intensity (Q) o f the so-called "cis" peak, a subsidiary peak in the nearultraviolet (where Q is the ratio o f the absorbance at )~max to the m a x i m u m absorbance in the cis peak; M o l n a r et al. 1986). Both all-trans and 9-cis isomers have small eis peaks and consequently large Q values. The spectra for the two principal isomers o f neoxanthin isolated are consistent with them being all-trans-neoxanthin (Q = 20.0) and 9"-cis-neoxanthin (Q = 7.2). A n u n k n o w n carotenoid eluting just after 9"-eis-neoxanthin could have been a third neoxanthin isomer (retention time (Rt) 23.1 min, )~m,x ethanol 463, 436, 412, Q = 1.9), possibly the lY-cis isomer, occurring at levels o f only 4% o f the
Determination and identification of earotenoids. Root samples (3-10 g FW) were extracted in approx. 25 ml of cold ( - 2 0 ~ C) ethanol (containing 0.1% v/v re-distilled tri-ethylamine [TEA]) using a Polytron blender (Kinematica, Lucerne, Switzerland). The extracts were then centrifuged at 1500 g in a bench-top centrifuge (IEC Centra-4X; Damon/IEC (UK), Dunstable, Beds.) for 5 rain, and the supernatant decanted. The pellets were washed with a further 10 ml of ethanol (+ 0.1% TEA), the extracts recentrifuged, and the combined supernatants reduced to dryness, under vacuum at 30~ C. The extracts were re-dissolved in ethanol (0.2-0.5 ml) and an aliquot (50 ~tl) analysed by reversed-phase high-performance liquid chromatography (r.p.HPLC), The procedure for light-grown and etiolated leaves was as described in Parry et al 1990. A Waters 600E system linked to a Waters 990 photodiode-array detector (Millipore Corporation, Waters Chromatography Division, Milford, Mass., USA) was used for HPLC. For r.p.HPLC an ODS-Spherisorb column (Phase Separations, Queensferry, Clwyd, UK; 250 mm long, 4.5 mm i.d.) was utilised. The column was eluted at 0.7 ml 9min- ~with a linear gradient of 10-50% Y in X over 25 rain, followed by a further 15 min at 100% Y, where X is 85% methanol (+ 0.1% TEA; v/v) and Y is 6: 4 dichloromethane :methanol. The eluate was monitored between 250-500 nm and spectra recorded at 3-s intervals. Absorbance spectra were recorded in ethanol using a Beckmann spectrophotometer (Beckmann Instruments Inc., Fullerton, Cal., USA) and micro-cuvette (50 Ixl), or in running solvent following r.p.HPLC using the photodiode-array detector. Tests for the presence of 5,6-epoxide groups were carried out by taking spectra in ethanol before and after the addition of 1.0 M hydrochloric acid (< 10 Ixl). Authentic zeaxanthin was a gift from Dr. B.H. Davies, Department of Biochemistry, The University College of Wales, Aberystwyth. Quantitation of carotenoids was based on their absorbance at )~.... and the use of suitable calibration curves. Saponification. This was performed on a small scale after Gregory et al. (1987). Approx. 5 ml of 10% potassium hydroxide in methanol (w/v) was added to 100 ~tl of ethanolic carotenoid extract, the mixture purged with nitrogen and heated for 5 rain at 50~ C in the dark. The carotenoids were recovered by partitioning twice against diethyl ether. The ether was washed with water to remove traces of alkali before being reduced to dryness under vacuum at 30~ C. The residue was dissolved in ethanol and analysed by r.p.HPLC.
A.D. Parry and R. ttorgan: Abscisic acid biosynthesis in roots. I Table 1. Absorbance maxima (in ethanol)
of carotenoids extracted from roots of Lycopersicon compared with published values, and their hypsochromic shifts following the addition of acid
187
Carotenoid
Extracted ~'ma~
All-trans-neoxanthin 9"-cis-neoxanthin All-trans-violaxanthin 9-cis-violaxanthin Antheraxanthin Lutein B-Carotene
471,443, 468, 441, 470; 442, 465, 438, 469, 443, 474, 446, 477, 450,
Published Lm~ 420 417 419 416 422 423 (428)
470, 467, 469, 464, 472, 474, 475,
441, 438, 440, 436, 444, 445, 449,
422 415 417 415 422 422 (427)
Referencea
Shift plus acid (nm)
1 t 2 2 2 2 2
20-21 18-19 41-43 37-39 18-22 0-1 0
a: 1, Cholnoky et al. 1969; 2, Davies 1976 Table 2. Absorbance maxima (in eluate) and r.p.HPLC retention times of carotenoids extracted from leaves and roots of Lycopersicon
Carotenoid
Root L~x
All-trans-neoxanthin 9"-cis-neoxanthin All-trans-violaxantlfin 9-cis-violaxanthin Antheraxanthin Lutein l-Carotene
Off afl-fra ns-violoxanthin 9-cis-
all-frans-neoxonfhin
H
O
~
OH
9'-cis-ne
an~heroxonfhin
H
O
~ lufein
.6
~ 8-coroIene
Fig. 1. Structures of the carotenoids extracted from roots
combined levels of the other neoxanthin isomers. Similarly the spectra for the two violaxanthin isomers indicated they were all-trans-violaxanthin (Q= 13.4) and 9-cis-violaxanthin (Q = 5.7). Antheraxanthin and lutein were both present as the all-trans isomers (Q = 7.2 and 26.0 respectively). In previous studies (Parry et al. 1990), carotenoids extracted from etiolated and light-grown leaves were conclusively identified by mass spectrometry and from the quasi-equilibria resulting from either iodinecatatysed or chlorophyU-sensitised photoisomerisation. The low levels of carotenoid present in roots precluded the use of either of these methods. However, Table 2 compares the absorbance spectra recorded using the photodiode-array detector and the R,s on r.p.HPLC of carotenoids from roots and leaves of Lycopersicon. This established that the carotenoids tentatively identified from their absorbance spectra in ethanol and reactions with dilute acid were the same as the ones extracted from leaves, and therefore confirmed their identities.
473, 467, 473, 467, 475, 477, 483,
443, 439, 443, 439, 447, 449, 457,
419 415 4t8 415 (418) (422) (432)
Rt (min)
~.~
Leaf
21.9 22.7 25.2 27.3 28.4 31.3 39.9
472, 468, 472, 466, 474, 478, 485,
Rt (min) 444, 439, 443, 439, 446, 450, 457,
419 415 418 415 (420) (422) (430)
21.8 22.7 25.2 27.3 28.4 31.2 39.8
Quantification of carotenoids. Using the procedure described above, carotenoids were extracted from roots of a variety of species. The levels of the xanthophylls, the oxygenated carotenoids, are shown in Table 3. For comparison the levels of carotenoids in light-grown and etiolated leaves of some of the same species are shown in Tables 4 and 5. Qualitatively there was no difference between the distribution of carotenoids in roots of the various species examined, or between the roots and either etiolated or light-grown leaves. Neoxanthin, violaxanthin, antheraxanthin and lutein were the only xanthophylls identified, and [3-carotene the only carotene. Prior to quantification the identity of each carotenoid was confirmed from absorbance spectra recorded during r.p.HPLC and Rt. The levels of total carotenoids in the roots varied between 0.03-0.07% of the levels in light-grown leaves, for Arabidopsis, Nicotiana, Phaseolus and Pisum, up to 0.27 % for Lycopersicon. The percentage of total carotenoids in roots compared with etiolated leaves was between 0.32% (Lycopersicon) and 1.7% (Zea). As the levels of carotenoid in the root extracts were so low it was possible that their presence was due to contamination by fragments of shoot material or residual carotenoids in the compost. The latter possibility was investigated by extracting approx. 5 g of both Levington Universal and John Innes No.2 composts using the same procedure as for roots. In both cases no trace of carotenoid was detected, with the exception of lutein. However, the amount of lutein was equivalent, on a fresh-weight basis, to only 2-4% of that in roots. As the amount of compost extracted with the root samples was probably less than 5 % of the root fresh weight, the contribution of lutein from the compost would have been negligible. Contamination of a 5' g FW root sample with only 1-10 mg of leaf tissue (i.e. 0.02-0.2 %) would be sufficient to give the observed levels of carotenoids. However a
188
A.D. Parry and R. Horgan: Abscisic acid biosynthesis in roots. I
Table 3, Xanthophyll composition in roots of a variety of plants Age (weeks) Arabidopsis thaliana Lycopersicon esculentum
3 3 13
Nicotiana plumbaginifolia Phaseolus vulgaris
10 2 13
Pisum sativum Zea mays
2 2
Xanthophyll (nmol - (g FW)-1, (% of total xanthophylls)) t-neo
c-neo
t-viola
c-viola
anthera
lutein
total
0.085 (39) 0.287 (80) 0.797 (67) 0.099 (46) 0.042 (11) 0.002 (1) 0.012 (6) 0.024 (2)
0.032 (15) 0.009 (2) 0.042 (3) 0.013 (6) 0.005 (1) 0.002 (t) 0.003 (1) 0.090 (9)
0.055 (25) 0.041 (t 1) 0.287 (24) 0.077 (37) 0.227 (73) 0.089 (61) 0.165 (79) 0.746 (78) ,
0.006 (3) 0.003 (1) 0.023 (2) 0.006 (3) 0.019 (6) 0.022 (15) 0.015 (7) 0.017 (2)
0.005 (2) 0.002 ( < 1) 0,005 (< l) 0.005 (2) 0.010 (3) 0.005 (3) 0.003 (1) 0.041 (4)
0.033 (15) 0.021 (6) 0.035 (3) 0.010 (5) 0.019 (6) 0.029 (19) 0.011 (5) 0.037 (4)
0.218 0.360 1.190 0.210 0.322 0.149 0.194 0.956
"t- = alt-trans; c- = 9/9"-cis; neo = neoxanthin; viola = violaxanthin; an thera = antheraxanthin Table 4, Xanthophyll composition in light-grown leaves of a variety of plants Age (weeks) Arabidopsis thaliana Lycopersicon esculentum Nicotiana plumbaginifolia Phaseolus vulgaris Pisum sativum
3 13 10 13 2
Xanthophyll (nmol " (g FW)-1, (% of total xanthophylls)) t-neo
c-neo
t-viola
c-viola
anthera
lutein
total
1.3 (< 1) 1.5 (< 1) 0.6 (< 1) 0.7 (< 1) 3.1 (1)
52.1 (17) 85.3 (19) 63.7 (16) 84.1 (19) 73.3 (16)
72.7 (25) 91.1 (21) 123.9 (31) 94.5 (22) 155.0 (35)
0.4 (< 1) 0.6 (< 1) n.d. (-) 0.3 (< 1) 0.9 ( < 1)
4.7 (2) 4.2 (1) 2.6 (1) 10.2 (2) 3.6 (1)
164.6 (55) 254.6 (58) 205.4 (52) 246.7 (56) 225.7 (48)
296 438 396 436 462
"t- = all-trans; c- = 9/9'-cis; neo = neoxanthin; viola = violaxanthin; anthera = antheraxanthin; n.d. = not detected detailed c o m p a r i s o n o f xanthophyll profiles o f roots and leaves reveals large differences. The relative xanthophyll composition in all of the light-grown leaves examined was very consistent (Table 4), the three major xanthophylls being 9"-cis-neoxanthin (16-19 % of the total carotenoid), all-trans-violaxanthin (21-35%) and lutein (48-58%). All-trans-neo xanthin, 9-cis-violaxanthin and antheraxanthin were all m i n o r c o m p o n e n t s ( < 1-2%). The levels o f l)-carotene present were between 28-33 % o f those o f the total xanthophylls (data not shown). This is consistent with findings reported in the literature (Goodwin 1980; G o o d w i n and Britton 1988). In etiolated leaves the situation was slightly different, and the carotenoid distribution m o r e varied (Table 5). Lutein and all-transviolaxanthin remained m a j o r carotenoids but the relative levels o f 9"-cis-neoxanthin were lower. In etiolated leaves o f Phaseolus the relative levels of antheraxanthin were elevated. In addition to the absolute levels being m u c h lower in roots, the relative distribution o f carotenoids was m u c h
m o r e varied between species, and very different f r o m that found in leaves. F o r example all-trans-neoxanthin varied between being present as 1-80 % of the total root xanthophylls, and all-trans-violaxanthin between 11-79%. The all-trans isomers o f neoxanthin and violaxanthin were the m a j o r carotenoids found in the roots examined here (the combined levels making up 64-91% o f the total xanthophyll levels). The levels of 13-carotene varied between 3% (Lycopersicon) and 12% (Pisum) o f the total xanthophyll levels (data not shown). The species listed in Table 3 appear to fall into three catorgaries: firstly, Lycopersicon is unusual in having a very large percentage of all-trans-neoxanthin (67-80%); secondly, there are those species whose roots contain approximately equal amounts o f all-trans-neoxanthin and all-trans-violaxanthin, such as Arabidopsis and Nicotiana; and thirdly, those such as Phaseolus, Pisum and Zea which contain predominantly all-trans-violaxanthin (61-79%). The data for Zea are consistent with those previously published (Maudinas and Lematre 1979; Feldman et al. 1985). In
189
A.D. Parry and R. Horgan: Abscisic acid biosynthesis in roots. I Table 5. Xanthophyll composition in etiolated leaves of a variety of plants
Lycopersicon esculentum Phaseolus vulgaris Zea mays
Age (weeks)
Xanthophyll (nmol 9(g FW)- 1, (% of total xanthophylls)) t-neo c-neo t-viola c-viola anthera
lutein
total
1
4.8 (4) 1.6 (3) n.d. (-)
61.0 (54) 23.3 (38) 31.2 (57)
113
1 2
7.0 (6) 4.1 (7) 1.7 (3)
37.7 (33) 25.4 (41) 21.1 (38)
1.8 (2) 0.6 (< 1) 0.3 (< 1)
0.8 (1) 6.8 (11) 0.4 (1)
62 55
at- = all-trans; c- = 9/9'-cis; neo = neoxanthin; viola = violaxanthin; anthera = antheraxanthin; n.d. = not detected
Table 6. Properties of suspected violaxanthin esters extracted from 13-week-old Phaseolus roots Isomer
All-trans9-cis-
a Recorded in ethanol
Non-saponified ~,m~a
Rtb
% of total carotenoid
469, 440, 417 466, 437, 412
39.8 38.2
28 57
Saponified ~'max
Rt
% of total carotenoid
468, 438, 415 466, 436, 413
24.9 27.0
33 58
Shift plus acid (nm) 39-42 37-43
b r.p.HPLC
leaves, especially light-grown ones, the predominant neoxanthin isomer is 9"-cis (Tables 4, 5), whilst in the roots o f the majority of species shown in Table 3 the all-trans isomer is predominant. It therefore seems unlikely that the carotenoids identified in the root extracts are derived from contaminating photosynthetic tissue. Measurements were made o f carotenoid levels in roots o f different aged plants of Lycopersicon and Phaseolus. The total levels in roots o f Lycopersicon were 3.3-fold higher in 13-week-old roots compared with those o f 3-week-old ones. The composition was similar except that the relative amounts o f all-trans-neoxanthin were lower and those of all-trans-violaxanthin higher in the older roots. In contrast the total levels o f carotenoid in roots of 13-week-old Phaseolus plants appeared to be only half those at 2 weeks. Interestingly the levels of all-trans-neoxanthin and all-trans-violaxanthin were much lower whilst those o f 9-cis-violaxanthin increased with age. Analysis by r.p.HPLC of extracts from the older (13-week) roots of Phaseolus revealed the presence of two large peaks eluting just prior to I]-carotene. Their spectra (Table 6) were very similar to those o f the xanthophylls but their Rts were characteristic o f compounds much less polar. Such compounds had been observed previously in extracts of orange fruit peel (data not shown). In oranges purchased early in the season the predominant carotenoids were all-trans- and 9-cis-violaxanthin, but in fruit purchased much later the normal levels of violaxanthin isomers were not observed. In their place large peaks of non-polar compounds possessing similar spectra were found. These were presumed to be violaxanthin esters. A carotenoid extract from 13-weekold Phaseolus roots was saponified to investigate the presence of esterified carotenoids. The procedure chosen was particularly mild, to avoid possible xanthophyll rearrangement and breakdown (Davies 1976; Gregory et al.
1987). The carotenoids were removed by partitioning from the saponified extract and analysed by r.p.HPLC. The two peaks predominant in the non-saponified extract disappeared following saponification, and were replaced by comparable peaks o f all-trans- and 9-cis-violaxanthin (identified by Rt, absorbance spectra and reaction with dilute acid; Table 6). In saponified extracts of 13-weekold Phaseolus roots estimates of total xanthophyll levels ranged from 0.5 to 2.3 nmol- (g FW) -1, that is 1.7-7.0fold higher than in roots o f 2-week-old plants. The violaxanthin esters were not further characterised, nor were they detected in young Phaseolus roots or roots of any other species examined. Arabidopsis mutants deficient in A B A . During the course
of this study it was revealed that the cause o f the ABAdeficiencies in leaves of the aba mutants of Arabidopsis was a deficiency of neoxanthin and violaxanthin (Duckham et al. 1991 ; Rock and Zeevaart 1991). The mutation is believed to affect the conversion of zeaxanthin, which accumulates in the mutant leaves, to antheraxanthin. It was therefore of interest to measure the xanthophyll levels in the roots of such a mutant (Table 7). The xanthophyll composition of wild-type and mutant leaves were consistent with previous findings (Duckham et al. 1991; Rock and Zeevaart 1991), with the epoxy-xanthophylls being reduced from 45 % o f the total xanthophylls in the wild-type to approx. 8 % in the mutant. This difference was more pronounced in the roots, with epoxyxanthophylls constituting 85% of the total xanthophylls in the wild-type but only 8 % in the mutant roots. Discussion The imposition of water-stress leads to the accumulation of ABA in leaves and roots, as the consequence o f an increased rate of biosynthesis (Zeevaart and Creelman
190
A.D. Parry and R. Horgan: Abscisic acid biosynthesis in roots. I
Table 7. Xanthophyll composition in leaves and roots of normal and mutant (aba, A26) Arabidopsis thaliana Xanthophyll (nmol 9(g F W ) - l, (% of total xanthophylls)) t-neo
c-neo
t-viola
c-viola
anthera
lutein + zea a
total
1.3 (
