Tree Physiology 24, 1129 –1136

© 2004 Heron Publishing—Victoria, Canada

A method for routine measurements of total sugar and starch content

in woody plant tissues
PAK S. CHOW1 and SIMON M. LANDHÄUSSER1,2
1
Centre for Enhanced Forest Management, Department of Renewable Resources, University of Alberta, Edmonton, AB, T6G 2E3, Canada
2
Corresponding author (simon.landhausser@ualberta.ca)

Downloaded from https://academic.oup.com/treephys/article/24/10/1129/1646984 by guest on 13 April 2023

Received December 15, 2003; accepted February 15, 2004; published online August 2, 2004

Summary Several extraction and measurement methods cur- first, a hot ethanol solution (Ebell 1969, MacRae et al. 1974,
rently employed in the determination of total sugar and starch Rose et al. 1991); and the second, a methanol:chloroform:
contents in plant tissues were investigated with the view to water solution (Dickson 1979, Rose et al. 1991). Both methods
streamline the process of total sugar and starch determination. are thought to be effective in removing soluble sugars from
Depending on the type and source of tissue, total sugar and plant tissues, but their efficacies on sugar and subsequent
starch contents estimated from samples extracted with 80% hot starch determinations have not been compared.
ethanol were significantly greater than from samples extracted Plant extracts contain diverse mixtures of sugars. The pres-
with a methanol:chloroform:water solution. The residual etha- ence of glucose, fructose, galactose, sucrose, maltose, melibi-
nol did not interfere with the sugar and starch determination, ose, raffinose and stachyose have been reported from tissues of
rendering the removal of ethanol from samples unnecessary. beech (Dietrichs and Schaich 1964), aspen (Wildman and Par-
The use of phenol–sulfuric acid with a phenol concentration kinson 1979), poplar (Fege and Brown 1984, Sauter 1988) and
of 2% provided a relatively simple and reliable colorimetric birch (Sauter and Ambrosius 1986). Although these sugars can
method to quantify the total soluble-sugar concentration. Per- be separately determined by high-performance liquid chroma-
forming parallel sugar assays with and without phenol was tography (HPLC) (Casterline et al. 1999) and gas chromatog-
more useful for accounting for the interfering effects of other raphy (GC) (Carlsson et al. 1992), the process is expensive es-
substances present in plant tissue than using chloroform. For pecially when only the total amount of glucose equivalents is
starch determination, an enzyme mixture of 1000 U α-amylase of interest (e.g., reserve and carbon allocation studies). The en-
and 5 U amyloglucosidase digested starch in plant tissue sam- zymatic method based on NADPH absorption (Blunden and
ples more rapidly and completely than previously recom- Wilson 1985, Hendrix 1993) requires a specific enzyme for
mended enzyme doses. Dilute sulfuric acid (0.005 N) was less each of the sugars, making it a relatively expensive and lengthy
suitable for starch digestion than enzymatic hydrolysis because process. Methods that measure reducing sugars (Ashwell 1957,
the acid also broke down structural carbohydrates, resulting in Miller 1959) exclude many of the oligosaccharides (e.g., su-
overestimates of starch content. After the enzymatic digestion crose), unless they are first hydrolyzed. Colorimetically, the
of starch, the glucose hydrolyzate obtained was measured with anthrone method (Scott and Melvin 1953) can detect all types
a peroxidase–glucose oxidase/o-dianisidine reagent; absorb- of sugars; however, because of the large difference in absorp-
ance being read at 525 nm after the addition of sulfuric acid. tion coefficients among the different types of sugars, it is
With the help of this series of studies, we developed a refined and known to produce errors in the analysis of sugar mixtures
shortened method suitable for the rapid measurement of total (Ashwell 1957). A simpler colorimetric method for sugar de-
sugar and starch contents in woody plant tissues. termination uses phenol–sulfuric acid (Dubois et al. 1956). Al-
though Buysse and Merckx (1993) optimized the phenol–sul-
Keywords: acid hydrolysis, analysis, colorimetric, enzymatic furic acid method for a sugar mixture of glucose, fructose and
digestion, ethanol extraction, MCW extraction, phenol–sul- sucrose by adjusting the amount of phenol, an optimization of
furic acid. this method for a mixture that includes all dominant sugars
present in plant extracts has not been reported.
Alcohol-soluble substances present in plant extracts, in-
cluding chlorophylls, lipids and proteins, react with the con-
Introduction centrated sulfuric acid in the sugar assay and therefore signifi-
Many physiological studies of growth and reserve allocation cantly interfere with the absorbance reading (Ashwell 1957).
in plants require the separate measurements of sugars and To remove these substances, the use of chloroform is consid-
starch in tissues. Generally, water-soluble sugars are extracted ered easier and more effective than hexane, activated char-
from the tissue sample and starch content is determined in the coal or ion-exchange columns (Haslemore and Roughan 1976,
residue. The process is time-consuming and costly. Two meth- Dickson 1979, Fege and Brown 1984). As an alternative to the
ods are frequently used for the extraction of soluble sugars: the removal of these interfering substances, Ashwell (1957) sug-
1130 CHOW AND LANDHÄUSSER

gested that running parallel assays for each sample with and For each tissue sample, five subsamples (50 mg) were ex-
without a color developer could also provide a correction of tracted three times with 5 ml of 80% ethanol, by boiling the
the absorbance readings. This approach has yet to be tested in samples in glass tubes capped with glass marbles in a 95 °C
sugar assays using phenol–sulfuric acid. water bath for 10 min each. After each extraction, the tubes
To estimate starch content in tissue samples, most methods were centrifuged at 2500 rpm for 5 min, and the supernatants
hydrolyze the starch to glucose, which is subsequently as- of the three extractions combined for sugar analysis. A fourth
sayed. Among these methods are perchloric acid hydrolysis ethanol extraction yielded less than 0.5% of the total sugar in
(MacRae et al. 1974, Rose et al. 1991), which is considered the first three extracts. The residues remaining in the tubes
dangerous because of the instability of perchloric acid; sulfu- were stored wet at –20 °C for starch analysis. Whether resi-
ric acid hydrolysis (Smith et al. 1964), which is considered the dues were oven-dried or wet had no effect on the subsequent
simplest and most rapid; and enzyme digestion with amylase starch analysis (see below).
and amyloglucosidase (Smith 1969, Haissig and Dickson 1979, A second set of five subsamples was extracted three times

Downloaded from https://academic.oup.com/treephys/article/24/10/1129/1646984 by guest on 13 April 2023

Rose et al. 1991, Hendrix 1993), which is considered the most with 5 ml of MCW solution. Sample tubes were loosely cap-
accurate but the most labor intensive. Grotelueschen and Smith ped, placed in a sonic bath for 5 s, and left at room temperature
(1967), Greub and Wedin (1969) and Kozloski et al. (1999) for 10 min. Samples were then centrifuged at 2500 rpm for
compared the sulfuric acid and enzymatic methods on legume 10 min and the supernatants of the three extractions were com-
roots, grasses and cereals; however, they determined the total bined (Rose et al. 1991). An earlier study had shown that three
nonstructural carbohydrate content without separating the sol- extractions with MCW removed about 98% of the water-alco-
uble sugars from starch. In addition, the authors used sulfuric hol-soluble compounds from leaf material (Dickson 1979). A
acid at concentrations of 0.2 N or higher, which could break 5-ml sample was taken from each combined extract, mixed
down structural carbohydrates. So far, few data are available with 3 ml of deionized water (dH2O) and separated into two
on the use of more dilute acid for the exclusive estimation of phases by centrifuging at 2500 rpm for 5 min. The chloroform
starch in woody plant tissues. Rose et al. (1991) described phase was discarded and the methanol:water phase was ana-
methods for digesting starch with enzyme mixtures of α-amy- lyzed for sugar. The residues were oven-dried at 50 °C over-
lase and amyloglucosidase, but the completeness of digestion night to remove the residual solvent, and stored at –20 °C for
was not tested in their studies. starch analysis.
Currently, the removal of the ethanol after extraction is For both extraction methods, sugar concentrations in ex-
among the most time-consuming steps in the determination of tracts were determined by the phenol–sulfuric acid method
sugar and starch in tissue samples. Buysse and Merckx (1993) without removing the aqueous ethanol or methanol solvent.
showed that, in pure sugar solutions, absorbance readings in-
creased for glucose but decreased for fructose and sucrose, Optimization of phenol dosage and absorbance wavelength
with increasing ethanol content when analyzed by the phe- for sugar mixtures
nol–sulfuric acid method. However, the effects of residual eth- Oligosaccharides are hydrolyzed by concentrated sulfuric acid
anol on the analysis of total sugar in plant extracts and on the during the phenol–sulfuric assay and form monomers, namely
subsequent analysis of starch have not been investigated. glucose, fructose and galactose (Sturgeon 1990). Therefore, in
The overall objective of this paper was to review selected the analysis of mono- and oligosaccharides in a plant extract,
methodologies for total sugar and starch analyses in plant tis- the intermediate product after acid hydrolysis is mostly a mix-
sues, and to shorten the existing procedures to establish a pro- ture of glucose, fructose and galactose, which are thus the
tocol suitable for rapid routine measurements of total sugar principal compounds measured in the sugar assay. Thus opti-
and starch reserves in plant tissues. mizing the phenol concentration and absorbance wavelength
for these three monomers should increase the accuracy of total
sugar estimation in plant extracts based on a single measure-
Materials and methods ment.
For the optimization, five 0.5 ml solutions of glucose, fruc-
Plant materials and preparation
tose, or galactose (50 µg ml –1) were mixed with 1 ml of a phe-
Leaf, stem and root samples of aspen (Populus tremuloides nol solution at one of five concentrations (1, 1.5, 2, 3 and 4%
Michx.), black spruce (Picea mariana (Mill.) BSP.) and lodge- phenol) followed by the rapid addition of 2.5 ml of concen-
pole pine (Pinus contorta Loudon) seedlings and saplings trated sulfuric acid (H2SO4). After 10 min of color develop-
were oven-dried at 68 °C for 2–3 days, ground in a Wiley mill ment in the dark and an additional 30 min of cooling in a water
to pass 40-mesh and stored in airtight containers at room tem- bath at 22 °C, absorbance was measured at wavelengths rang-
perature, in the dark, until analysis. Tissue samples analyzed in ing from 465 to 505 nm in 5 nm increments. Each sugar and
the different studies varied by type and collection. phenol combination was replicated five times.
After optimization of the phenol concentration and the ab-
Comparison of sugar extraction methods sorption wavelength, the absorption coefficient of the sugar
Two methods of soluble sugar extraction, hot ethanol and mixture (1:1:1, glucose, fructose, galactose; GFG) was com-
methanol:chloroform:water (MCW, 12:5:3, v/v/v), were com- pared with the absorption coefficients of the single sugars: glu-
pared on dried samples of five different woody plant tissues cose, fructose, galactose, sucrose, maltose, melibiose, raffi-
(leaves of aspen, spruce and pine, aspen roots and pine shoots). nose and stachyose, to verify that the GFG mixture can be used

TREE PHYSIOLOGY VOLUME 24, 2004

 MEASURING TOTAL SUGAR AND STARCH IN WOODY TISSUES 1131

as the calibration standard for the colorimetric analysis of the A = a s [sugar]c + A I (2)
single sugars. Because the oligosaccharides increase in molec-
ular weight after hydrolysis to monomers, the concentrations where [sugar]c is the corrected sugar concentration and AI is
of all sugar solutions were standardized based on their mono- the absorbance of interfering substances. In the parallel assay
mer equivalent. without phenol, the absorbance (A′) measured is given by:
The sugar assay was also verified on prepared pure sugar
mixtures (Table 1) with known concentrations, from 50 to A′ = a ′s [sugar]c + A′I (3)
200 µg ml –1. The composition of the pure sugar mixtures was
derived from information on plant extracts found in the pub- where a ′s is the absorption coefficient of the GFG standard
lished literature (Wildman and Parkinson 1979, Sauter and without phenol and A′I is the absorbance of interfering sub-
Ambrosius 1986, Sauter 1988). Sugar concentrations were de- stances without phenol. Assuming that the interfering sub-
termined against the GFG mixture standard. stances react with the concentrated H2SO4 similarly in both

Downloaded from https://academic.oup.com/treephys/article/24/10/1129/1646984 by guest on 13 April 2023

assays and thus have the same interfering absorbance, i.e., AI =
Interfering substances in the ethanol extract A′I , then, subtracting Equation 3 from Equation 2 gives:
Two methods that account for the interfering effects on sugar
determination of compounds (primarily pigments and pro- A – A′ = ( a s – a ′s )[sugar] c (4)
teins) found in plant extracts were compared. One method
physically removes the interfering substances with chloro- or
form, whereas the other method corrects the absorbance read-
ings based on parallel assays. A – A′
[sugar] c = (5)
Raw ethanol extracts were made from three samples each of a s – a ′s
four plant tissue types (leaves of aspen, pine and spruce, and
pine roots). To clean the plant extracts with chloroform by The corrected sugar concentration of each extract solution was
fractionation, 1 ml of chloroform was mixed with 3 ml of ex- calculated according to Equation 5 after the absorption coeffi-
tract to form a monophasic liquid. After shaking with 5 ml of cients of the GFG standard were determined from standard
dH2O and centrifuging at 2500 rpm for 5 min, two phases sep- curves.
arated. The chloroform phase was discarded and the aqueous The parallel assay method was verified on mixtures contain-
phase was assayed for sugar. The experiment was repeated by ing a GFG standard mixed with a pine leaf extract that contains
increasing the chloroform to extract ratio from 1:3 to 1:2.5 and significant amounts of interfering substances. Four mixtures
1:2. For comparison, sugar assays were carried out on the raw were made, with 0.5 ml of a leaf extract combined with 0, 2.5,
extracts before chloroform treatment. 5.0 or 7.5 ml of 100 µg ml –1 GFG sugar solution producing
In the second method, the raw plant extracts were subjected samples with decreasing interfering substances relative to
to parallel sugar assays: with and without phenol (the phenol sugar content.
solution was replaced in the control by an equivalent volume
of dH2O). In the assay with phenol, the absorbance (A) mea- Acid hydrolysis versus enzymatic digestion for starch
sured is given by: estimation
To explore the possibility of using a dilute acid for hydrolyzing
A = a s [sugar]u (1) the starch content of woody plant tissues instead of digestive
enzymes, eight pure polysaccharides (cotton cellulose, citrus
where a s is the absorption coefficient of the GFG standard and pectin, starch from arrowroot, corn and potato, and hemicellu-
[sugar]u is the sugar concentration uncorrected for interfer- lose including arabic, damar and guaiac gum) and four de-sug-
ence. However, when taking the interference into account: ared tissue samples (shoots and roots of aspen and spruce)
were tested.
The method of acid hydrolysis was adapted from Grotel-
Table 1. Sugar composition (SC) as a percentage of total sugar. Sugar ueschen and Smith (1967); however, our sulfuric acid concen-
compositions found in plant extracts are based on published values for tration was 40 times lower (0.005 N) than that used by the
various tissues (Wildman and Parkinson 1979, Sauter and Ambrosius previous authors. Plant tissue (20 mg), starch (5 mg), and cel-
1986, Sauter 1988). lulose, hemicellulose and pectin (50 mg) samples were re-
Sugar % SC found Pure sugar mixtures (%) fluxed for 0.5 and 1 h with 5 ml of 0.005 N H2SO4 in a 95 °C
in plant extracts water bath. After hydrolysis, the contents were vacuum fil-
Mix 1 Mix 2 Mix 3 tered through Whatman No. 40 filter paper. Filtrates were ana-
Glucose 7.2–36.7 30 15 5 lyzed for glucose, by the phenol–sulfuric acid method and
Fructose 6.3–52.8 50 25 5 concentrations determined against a glucose standard. Starch
Galactose 0 –23.7 20 10 0 content of the samples was estimated by multiplying the glu-
Sucrose 0 – 40.2 0 25 35 cose content by the glucose equivalent of 0.9.
Maltose 0–36.6 0 10 25 A replicate set of samples was digested with an enzyme
Raffinose 0 –10.3 0 5 10 mixture containing α-amylase and amyloglucosidase (Rose et
Stachyose 1.6–20.7 0 10 20
al. 1991). Samples were weighted as described above. Each

TREE PHYSIOLOGY ONLINE at http://heronpublishing.com

1132 CHOW AND LANDHÄUSSER

sample was heated with 2 ml of 0.1 N sodium hydroxide cose hydrolysate was corrected for all samples. After diges-
(NaOH) in a 50 °C water bath for 30 min with intermittent tion, the resulting solution of each sample was split into four
mixing. After neutralizing with 2.5 ml of 0.1 N acetic acid, subsamples for glucose measurement. The first subsample was
0.5 ml of a digestive enzyme mixture containing 400 U ml –1 of analyzed with the addition of 2 ml of PGO-color solution
α-amylase (from Bacillus licheniformis, ICN-190151, ICN alone (as described above) (Haissig and Dickson 1979, Rose et
Biomedicals, Aurora, OH) and 2 U ml –1 of amyloglucosidase al. 1991), whereas the second subsample was analyzed with
(from Aspergillus niger, Sigma A-1602, Sigma Chemicals, St 2 ml of PGO-color solution followed by the addition of 0.4 ml
Louis, MO) in 0.05M sodium acetate buffer (pH 5.1) was of 75% H2SO4 (Ebell 1969, Rose et al. 1991). To determine the
added. The combined solution was incubated for 48 h in a amount of interference from sample impurities, the third sub-
50 °C water bath. To determine the amount of glucose hydro- sample was analyzed with 2 ml of dH2O alone and the fourth
lysate, the digest was centrifuged at 2500 rpm for 10 min and subsample with 2 ml of dH2O plus 0.4 ml of 75% H2SO4.
the supernatant diluted with 0.05 M sodium acetate buffer, as Absorbances of all four subsamples were read at wavelengths

Downloaded from https://academic.oup.com/treephys/article/24/10/1129/1646984 by guest on 13 April 2023

required. To 0.2 ml of the diluted sample solution, 2 ml of from 400 to 575nm in 25nm increments. The experiment was
a peroxidase–glucose oxidase/o-dianisidine reagent (PGO- repeated five times.
color solution, prepared by dissolving one capsule of PGO en-
zymes (Sigma P-7119) in 100 ml of dH2O and mixed with The effects of residual ethanol on sugar and starch analyses
1.6 ml of o-dianisidine solution (50 mg of o-dianisidine dihy- To determine the effect of residual ethanol on the sugar assay, a
drochloride (Sigma D-3252) in 20 ml of dH2O) was added and sample of spruce leaf tissue was extracted with ethanol, and
mixed. The absorbance was read at 450 nm, after leaving the equal aliquots of the extract were transferred to 25 test tubes.
mixture in darkness at room temperature for 45 min. The After drying in a 50 °C oven overnight, ethanol of five concen-
amount of glucose was calculated against a glucose standard trations (0, 20, 40, 60 and 80%) was added to redissolve the
prepared in the sodium acetate buffer solution. A preliminary residue in replicates of five. A GFG standard (100 µg ml –1)
test showed that using phenol–sulfuric acid for the glucose as- was also prepared in ethanol of the same five concentrations.
say after enzyme digestion of starch was not possible, because All samples were assayed for sugar, and absorbances were
the concentrated H2SO4 hydrolyzed the enzyme and therefore read and corrected for interference.
gave unreliable results. The effect of the residual ethanol on the starch analysis was
studied on six tissue samples (leaves and roots of aspen and
Efficiency of enzymatic digestion of starch spruce, and leaves and shoots of pine). From each sample,
To examine the amount of enzyme and time required for com- 10 subsamples were extracted with ethanol and residues were
plete hydrolysis of starch in plant tissues, tissue samples from randomly divided into two sets of five. One set was oven-dried
aspen and spruce roots, aspen shoots, spruce and pine leaves, at 50 °C overnight to remove the residual ethanol and water,
as well as purified potato starch (Sigma S-2630) were digested and the second set was kept frozen at –20 °C overnight without
with either 400 U ml –1 of α-amylase + 2 U ml –1 of amyloglu- drying. Both sets of residue were analyzed enzymatically for
cosidase (Rose et al. 1991) or 2000 U ml –1 of α-amylase + 10 U starch as described above.
ml –1 of amyloglucosidase. The high dosage was calculated
from the guideline set out by Haissig and Dickson (1979), Statistical analysis
based on a 50 mg sample weight with a starch content of 30%. In all studies, sugar and starch contents were evaluated by
Tissue samples (50 mg) and potato starch samples (5 mg) were analysis of variance of paired comparisons (SAS Institute,
digested for either 24 or 48 h. Cary, NC).

Measurement of glucose hydrolysate from enzymatic

digestion of starch Results and discussion
During starch solubilization, colored impurities are readily no-
ticeable when sample material is heated with sodium hydrox- Comparison of sugar extraction methods
ide, resulting in a brown solution. The brown color is main- Overall, samples extracted with methanol:chloroform:water
tained throughout the enzyme digestion process and interferes (MCW) solution gave total nonstructural carbohydrate (TNC)
with the colorimetric measurement of glucose hydrolysates estimates 6 to 29% lower than samples extracted with ethanol.
obtained after the digestion (Haissig and Dickson 1979). To Estimates of sugar content in the five different plant tissue
overcome this interference, parallel assays were performed to samples were 7 to 16% lower when extracted with MCW than
test at which wavelength an absorbance correction can be when extracted with 80% hot ethanol (P < 0.001). Similarly,
made. starch content estimated in MCW-extracted samples was 2 to
After ethanol extraction, the residues of five different tis- 96% lower than in samples extracted with ethanol (P = 0.019)
sues (leaves of aspen, spruce and pine, aspen roots and pine (Table 2). Some of the differences could be explained by the
shoots) were digested with 0.5 ml of an enzyme mixture con- extraction process. During the MCW extraction, sugars are re-
taining 2000 U ml –1 of α-amylase and 10 U ml –1 of amyloglu- moved from the tissues by brief sonication without any further
cosidase for 20 h at 50 °C. Samples of potato starch and an en- agitation, whereas samples extracted with ethanol are agitated
zyme blank were also included. Because the enzyme mixture constantly by boiling and thus sugars might be removed more
contained a small amount of glucose, the total amount of glu- efficiently. Further, starch grains are gelatinized by boiling in

TREE PHYSIOLOGY VOLUME 24, 2004

 MEASURING TOTAL SUGAR AND STARCH IN WOODY TISSUES 1133

ethanol, so they dissolve better in the NaOH during starch di-

gestion (Rose et al. 1991). Although not tested, the large range
of differences between the two extraction methods, especially
in starch content, might indicate that the success of the extrac-
tion technique can vary with tissue type (e.g., leaf versus root)
(Table 2).

Optimization of phenol concentration and absorbance

wavelength for sugar mixtures
Absorption peaks of glucose, fructose and galactose were
observed at 490 nm, which is similar to the findings reported
by Dubois et al. (1956). Our study indicates that a phenol con-

Downloaded from https://academic.oup.com/treephys/article/24/10/1129/1646984 by guest on 13 April 2023

centration of 2% produced the smallest difference in the absor-
bance readings among the three sugars (Figure 1). As a result,
measuring absorbance at 490 nm with the use of 2% phenol in
the sugar assay gave the closest match to the absorption coeffi-
cients of the three basic monomers without compromising the
sensitivity of the assay (Figure 1C). Figure 1. Absorbance of glucose, fructose and galactose (50 µg ml –1)
Based on a phenol concentration of 2% and an absorption at various wavelengths and phenol concentrations in the phenol–sul-
wavelength of 490 nm, the standard curves of the GFG stan- furic acid sugar assay. Data represent means and SE of 5 replicates.
dard and the sugars (glucose, fructose, galactose, sucrose, mal-
tose, melibiose, raffinose and stachyose) were linear up to
200 µg ml –1. Although there were differences in absorption phenol, the sugar concentrations in the extracts estimated after
coefficients among the sugars, which ranged from 0.00628 cm2 the interference correction were 11 to 36% lower (P < 0.001)
µg –1 for fructose to 0.00716 cm2 µg –1 for glucose, the ab- than those without correction, suggesting that the chloroform
sorption coefficient of the GFG standard (0.00679 cm2 µg –1) fractionation removed only a portion of the interfering sub-
was close to the mean absorption of the eight single sugars stances from the plant tissue extracts. In addition, another
(0.00668 cm2 µg –1). Considering that fructose and galactose study indicated that the compounds removed by chloroform
are equivalent to glucose in molecular weight, this sugar assay (primarily chlorophylls and carotenoids) absorbed mainly at
using GFG as a reference standard can approximate the glu- wavelengths below 420 nm (data not shown). These sub-
cose equivalent of the tested sugars. stances would interfere only slightly with sugar absorption at
The sugar assay quantified 96 to 100% of the total sugar 490 nm. As increasing the amount of chloroform did not re-
contained in the pure sugar mixtures, indicating that it can sat- duce the estimated quantity of sugar, interfering substances
isfactorily estimate the glucose equivalent of sugar mixtures. that remained were likely water-soluble compounds such as
However, when working with plant tissues that might contain proteins and other non-carbohydrate reducing agents (Stur-
sugars other than those tested in this study, absorbances of geon 1990).
those particular sugars need to be tested against the GFG stan- When extracts treated with and without chloroform were
dard. analyzed in parallel assays, the sugar estimates after interfer-
ence corrections were not significantly different (P = 0.448).
Interfering substances in the ethanol extract Using the parallel-assay method on the four GFG/pine leaf ex-
The sugar concentrations were 2 to 7% lower in extracts tract mixtures gave sugar recoveries between 98 and 99%.
treated with chloroform than in extracts without the chloro- This indicates that the correction of interference by means of
form treatment (P < 0.001); however, increasing the amount of parallel sugar assays can give accurate estimates of sugar in
chloroform resulted in no further reduction in the sugar esti- plant extracts.
mates (P = 0.968). By running parallel assays with and without The amount of interfering substances was found to be clos-
ely related to the sugar content of the plant tissue. The ratio of
the corrected to the uncorrected sugar content measured in
leaf, shoot and root samples of aspen, spruce and pine were
Table 2. Sugar and starch contents (% dry mass) of plant tissue sam- found to be constant within each plant-tissue type (Table 3).
ples extracted with either 80% hot ethanol (ETOH) or methanol:chlo- Consequently, we believe that for each type of plant tissue the
roform:water (MCW) solution. Values are means of 5 subsamples.
sugar content can be calculated from Equation 1 and corrected
Sample Sugar (%) Starch (%) using the ratio obtained from a group of representative sam-
ples. This saves time by eliminating the need for parallel sugar
ETOH MCW ETOH MCW
assays of all samples.
Aspen leaf 24.7 22.9 0.28 0.01
Aspen root 10.4 9.3 11.82 11.57 Comparison of acid hydrolysis and enzymatic digestion
Pine leaf 7.3 6.1 3.17 1.38 methods for starch estimation
Pine shoot 7.5 6.8 0.72 0.26 After half an hour, 0.005 N sulfuric acid hydrolyzed the three
Spruce leaf 11.8 10.6 4.15 2.13
sources of starch (arrowroot, corn and potato) by 67 to 86%,

TREE PHYSIOLOGY ONLINE at http://heronpublishing.com

1134 CHOW AND LANDHÄUSSER

Table 3. Mean ratio ± SD of corrected to uncorrected sugar contents Table 4. Starch content (% dry mass) of plant tissue samples estimated
measured in ethanol extracts of various plant tissues from n different by enzyme digestion (Enzyme) and dilute sulfuric acid (H2SO4 ) hy-
trees. drolysis (n = 2). Abbreviations: OS = old shoot; and NS = new shoot.

Species Tissue Ratio n Sample Starch estimated (%)

Aspen Leaf 0.88 ± 0.04 29 Enzyme H2SO4 hydrolysis (0.005 N)
Shoot 0.91 ± 0.02 39 (48 h)
Root 0.89 ± 0.03 28 0.5 h 1h

Pine Leaf 0.74 ± 0.03 39 Aspen OS 0.40 3.08 4.47

Shoot 0.82 ± 0.04 39 Aspen root 2.78 4.46 6.55
Root 0.81 ± 0.03 19 Spruce NS 0.62 8.10 10.25
Spruce root 0.94 3.68 4.31
Spruce Leaf 0.75 ± 0.04 38

Downloaded from https://academic.oup.com/treephys/article/24/10/1129/1646984 by guest on 13 April 2023

Shoot 0.83 ± 0.01 18
Root 0.80 ± 0.04 20
ples that had been digested for 48 h with the low enzyme dose
(P < 0.001). However, extending the digestion time to 48 h at
arabic gum and citrus pectin by 74 and 53%, respectively, and the high dose did not result in higher glucose yields (P =
cotton cellulose and guaiac gum both by 1%. Damar gum was 0.060). This indicates that the high enzyme dose is preferable,
not hydrolyzed. Thus, even at low concentrations, sulfuric acid particularly for rapid assays. A separate test showed that, at the
can partially hydrolyze structural carbohydrates in plant tis- high enzyme dose, the digestion period could be further short-
sues, giving rise to an overestimate of starch content. However, ened to 20 h without compromising the starch estimation (P =
the enzyme mixture digested the three sources of starch by 84 0.765).
to 89% without digesting cellulose or hemicellulose, although
Measurement of glucose hydrolysate from enzymatic
it broke down 19% of the citrus pectin. As woody tissues con-
digestion of starch
tain little pectin (Kramer and Kozlowski 1979), starch estima-
tion by enzyme digestion can be considered acceptable. Colored impurities formed during the enzymatic digestion of
Sulfuric acid hydrolysis gave generally higher estimates of starch had a very broad absorption spectrum with an absorp-
starch content than the enzyme method for all plant tissue sam- tion maximum at or below 400 nm (Figure 2A). Absorption
ples (P = 0.004), suggesting that sulfuric acid hydrolyzed curves of sample solutions analyzed with the addition of dH2O
some structural carbohydrates to glucose (Table 4). showed that the degree of interference at 525 nm was less than
half that at 450 nm (Figure 2A). The addition of sulfuric acid
Efficiency of enzymatic digestion of starch to the sample solution lowered the absorption at 525 nm by be-
Independent of enzyme dosage (P = 0.942) and digestion time tween 0 to 50%, depending on the type of plant tissue (Fig-
(P = 0.285), there was, overall, no significant difference in the ure 2B). Samples analyzed with PGO-color solution alone
amount of purified starch digested. However, in plant tissues had an absorption maximum at 450 nm where interference
incubated with the low enzyme dosage, starch estimates after is still high (Figure 2C). In contrast, samples analyzed with
48 h of digestion were 3 to 16% higher (P = 0.013) than those PGO-color solution and sulfuric acid had absorption peaks at
digested for only 24 h (Table 5), suggesting that the low dos- 525 nm where interference was significantly less (Figure 2D).
age may have been insufficient for complete hydrolysis. Hais- Combining these effects, the interfering substances will have a
sig and Dickson (1979) reported that without enough enzyme, smaller impact on the measurement of glucose hydrolyzate
the amount of glucose yielded by starch digestion increases when absorbance is measured at 525 nm after the addition of
with digestion time, but full digestion is not achieved. In our sulfuric acid.
tests, samples digested for 24 h with the high enzyme dose After making corrections for interference in the plant sam-
yielded starch estimates that were 5 to 28% higher than sam- ples by subtracting the absorbance readings obtained without

Table 5. Mean starch content ± SD (% dry mass) of plant tissue samples estimated following digestion with low (400 U ml –1 α-amylase + 2 U ml –1
amyloglucosidase) and high (2000 U ml –1 α-amylase + 10 U ml –1 amyloglucosidase) enzyme dose for 24 or 48 h (n = 5).

Sample Low enzyme dose High enzyme dose

24 h 48 h 24 h 48 h
Aspen shoot 2.96 ± 0.05 3.05 ± 0.09 3.90 ± 0.06 3.87 ± 0.07
Aspen root 1 5.30 ± 0.23 5.49 ± 0.12 6.20 ± 0.27 6.33 ± 0.12
Aspen root 2 7.40 ± 0.05 8.58 ± 0.27 9.05 ± 0.17 9.79 ± 0.37
Pine leaf 19.9 ± 0.2 20.8 ± 0.1 22.2 ± 0.2 22.6 ± 0.5
Spruce leaf 23.4 ± 0.4 25.0 ± 0.5 26.5 ± 0.5 27.4 ± 0.5
Spruce root 7.87 ± 0.56 9.17 ± 0.59 10.1 ± 0.3 11.7 ± 0.4
Potato starch 99.4 ± 3.7 97.3 ± 3.2 99.6 ± 3.1 96.8 ± 7.9

TREE PHYSIOLOGY VOLUME 24, 2004

 MEASURING TOTAL SUGAR AND STARCH IN WOODY TISSUES 1135

starch contents of oven-dried samples and undried samples

containing residual ethanol (P = 0.619).

Reproducibility and application

Combining the above findings and amendments of existing
procedures provides a method that yielded consistent results.
Throughout this study, we used a benchmark tissue sample of
aspen root to assess the precision and reproducibility of the
method. Sugar and starch estimations of the benchmark sam-
ple remained consistent throughout the 61 tests over a period
of 10 months during which different batches of chemicals
were used. The sugar and starch contents in the benchmark

Downloaded from https://academic.oup.com/treephys/article/24/10/1129/1646984 by guest on 13 April 2023

sample were 15.09 ± 0.57% (SD) and 3.38 ± 0.17% (SD), re-
spectively.
Overall, our method requires small samples (50 mg) and
only common laboratory equipment. It is practical for all tis-
sues of woody plants, including leaves, buds, stems, bark and
Figure 2. Absorption curves of sample solutions after digestion with roots, is sensitive enough to detect a sugar or starch concentra-
α-amylase and amyloglucosidase. Tissue materials were extracted tion of 0.04%, and can handle samples with up to 40% sugar
with hot ethanol before digestion. Data are means of 5 replicates. (A)
and 30% starch content. For samples with higher anticipated
Analyzed with dH2 O. (B) Analyzed with dH2O + H2SO4. (C) Ana-
lyzed with peroxidase–glucose oxidase/o-dianisidine reagent (PGO- carbohydrate contents, sample size should be reduced to half.
color solution). (D) Analyzed with PGO-color solution + H2SO4. Starting with a ground tissue sample, both sugar and starch as-
says can be completed in 27 h, including incubation time. In
routine measurements, one operator working with one set of
equipment can finish a complete analysis of 60 ground
the PGO-color solution from those with the PGO-color solu- samples in four working days, including preparation and clean-
tion, there was no detectable difference in starch content when up work.
estimated at 450 or 525 nm (P = 0.267).
The color developed with PGO-color solution darkens over
time and absorbance readings increased by about 3% within Conclusion
the first hour. However, when sulfuric acid was added, the en-
zymatic action of the PGO-color solution was stopped, and the Based on the above studies, we have developed a rapid method
color of the solution turned from orange to pink and darkened for the measurement of total sugar and starch in woody plant
only by about 0.1% per hour. Therefore, we believe that the ad- tissues. Briefly, 50 mg of dried ground tissue sample is ex-
dition of sulfuric acid allows for the more accurate estimation tracted with 80% hot ethanol. The ethanol extract is analyzed
of glucose hydrolysate concentrations. for sugar using phenol–sulfuric acid against a GFG standard,
with 2% phenol and absorbance read at 490 nm. Interference is
corrected by running a parallel sugar assay without phenol.
Effects of residual ethanol on sugar and starch analyses The residue is analyzed for starch by enzymatic digestion with
Increasing ethanol content in plant extracts and GFG standard a mixture of 1000 U of α-amylase and 5 U of amyloglucosi-
solution samples resulted in decreased absorbance readings dase for 20 h, followed by the colorimetric measurement of
and consequently underestimation of sugar concentrations. At glucose hydrolyzate with a peroxidase–glucose oxidase/o-
ethanol concentrations of 20 to 80%, absorbance readings de- dianisidine reagent. Absorbance is read at 525 nm after the ad-
creased from 3 to 13%. There was no difference in the effect of dition of H2SO4.
ethanol on absorbance readings between the GFG solution and
the plant extract (P = 0.466). Acknowledgment
Generally, sample extracts are diluted with water (1:10) be-
We thank W. Liu for the collection and preparation of the plant tissue
fore the sugar assay and, as a result, only 8% or less of ethanol samples. We also thank P.V. Blenis and K. Stadt for their help in re-
will be present in the sample solution. By interpolating our re- viewing this manuscript.
sults, this concentration will result in a 1% reduction in ab-
sorbance. However, if the sugar content of the tissue sample is References
expected to be less than 1% of dry mass (no dilution), residual
ethanol concentrations up to 80% are possible. In that case, the Ashwell, G. 1957. Colorimetric analysis of sugars. Methods En-
zymol. 3:73–105.
expected reduction in absorbance associated with the presence
Blunden, C.A. and M.F. Wilson. 1985. A specific method for the de-
of alcohol will be 13% (e.g., a tissue sugar content of 1% will termination of soluble sugars in plant extracts using enzymatic
measure as 0.87%). analysis and its application to the sugar content of developing pear
Ethanol that remained in the residues after ethanol extrac- fruit buds. Anal. Biochem. 151:403–408.
tion had no significant effect on the subsequent measurement Buysse, J. and R. Merckx. 1993. An improved colorimetric method to
of starch content. No significant difference was found in the quantify sugar content of plant tissue. J. Exp. Bot. 44:1627–1629.

TREE PHYSIOLOGY ONLINE at http://heronpublishing.com

1136 CHOW AND LANDHÄUSSER

Carlsson, N.G., H. Karlsson and A.S. Sandberg. 1992. Determination Kozloski, G.V., J.B.T. Rocha, H.M.N. Ribeiro-Filho and J. Perottoni.
of oligosaccharides in foods, diets, and intestinal contents by high- 1999. Comparison of acid and amyloglucosidase hydrolysis for es-
temperature gas chromatography and gas chromatography/mass timation of non-structural polysaccharides in feed samples. J. Sci.
spectrometry. J. Agric. Food Chem. 40:2404–2412. Food Agric. 79:1112–1116.
Casterline, Jr., J.L., C.J. Oles and Y. Ku. 1999. Measurement of sugars Kramer, P.J. and T.T. Kozlowski. 1979. Physiology of woody plants.
and starches in foods by a modification of the AOAC total dietary Academic Press, New York, pp 258–281.
fiber method. J. AOAC Int. 82:759 –765. MacRae, J.C., D. Smith and R.M. McCready. 1974. Starch estimation
Dickson, R.E. 1979. Analytical procedures for the sequential extrac- in leaf tissue—a comparison of results using six methods. J. Sci.
tion of 14 C-labeled constituents from leaves, bark and wood of cot- Food Agric. 25:1465–1469.
tonwood plants. Physiol. Plant. 45:480 – 488. Miller, G.L. 1959. Use of dinitrosalicylic acid reagent for determina-
Dietrichs, H.H. and E. Schaich. 1964. Type, proportion and distribu- tion of reducing sugar. Anal. Chem. 31:426– 428.
tion of low-molecular carbohydrates in Fagus sylvatica. Forstwiss. Rose, R., C.L. Rose, S.K. Omi, K.R. Forry, D.M. Durall and W.L. Bigg.
Centralbl. 83:212–222. 1991. Starch determination by perchloric acid vs enzymes: evaluat-

Downloaded from https://academic.oup.com/treephys/article/24/10/1129/1646984 by guest on 13 April 2023

Dubois, M., K.A. Gilles, J.K. Hamilton, P.A. Rebers and F. Smith. ing the accuracy and precision of six colorimetric methods. J.
1956. Colorimetric method for determination of sugars and related Agric. Food Chem. 39:2–11.
substances. Anal. Chem. 28:350-356. Sauter, J.J. 1988. Temperature-induced changes in starch and sugars
Ebell, L.F. 1969. Specific total starch determinations in conifer tissues in the stem of Populus × canadensis “Robusta”. J. Plant Physiol.
with glucose oxidase. Phytochemistry 8:25-36. 132:608–612.
Fege, A.S. and G.N. Brown. 1984. Carbohydrate distribution in dormant Sauter, J.J. and T. Ambrosius. 1986. Changes in the partitioning of
Populus shoots and hardwood cuttings. For. Sci. 30:999 –1010. carbohydrates in the wood during bud break in Betula pendula
Greub, L.J. and W.F. Wedin. 1969. Determination of total available Roth. J. Plant Physiol. 124:31– 43.
carbohydrates in forage legume roots by extraction with takadias- Scott, T.A. and E.H. Melvin. 1953. Determination of dextran with
tase, amyloglucosidase or sulfuric acid. Crop Sci. 9:595–598. anthrone. Anal. Chem. 25:1656–1661.
Grotelueschen, R.D. and D. Smith. 1967. Determination and identifi- Smith, D. 1969. Removing and analyzing total nonstructural carbohy-
cation of nonstructural carbohydrates removed from grass and le- drates from plant tissue. Wisconsin Agric. Exp. Sta. Res. Report R-
gume tissue by various sulfuric acid concentrations, takadiastase, 2107. College of Agric. and Life Sci., Univ. Wisconsin, Madison,
and water. J. Agric. Food Chem. 15:1048–1051. 12 p.
Haissig, B.E. and R.E. Dickson. 1979. Starch measurement in plant Smith, D., G.M. Paulsen and C.A. Raguse. 1964. Extraction of total
tissue using enzymatic hydrolysis. Physiol. Plant. 47:151–157. available carbohydrates from grass and legume tissue. Plant Phys-
Haslemore, R.M. and P.G. Roughan. 1976. Rapid chemical analysis iol. 39:960 –962.
of some plant constituents. J. Sci. Food Agric. 27:1171–1178. Sturgeon, R.J. 1990. Monosaccharides. In Methods in Plant Biochem-
Hendrix, D.L. 1993. Rapid extraction and analysis of nonstructural istry. Vol. 2. Carbohydrates. Ed. P.M. Dey. Academic Press, Lon-
carbohydrates in plant tissues. Crop Sci. 33:1306–1311. don, pp 1–37.
Wildman, H.G. and D. Parkinson. 1979. Seasonal changes in wa-
ter-soluble carbohydrates of Populus tremuloides leaves. Can. J.
Bot. 59:862–869.

TREE PHYSIOLOGY VOLUME 24, 2004