Analytica Chimica Acta 776 (2013) 3140

Contents lists available at SciVerse ScienceDirect

Analytica Chimica Acta

journal homepage: www.elsevier.com/locate/aca

Simultaneous assay of pigments, carbohydrates, proteins and lipids

in microalgae
Yimin Chen, Seetharaman Vaidyanathan
ChELSI Institute, Department of Chemical and Biological Engineering, The University of Shefeld, Shefeld S1 3JD, UK

h i g h l i g h t s

g r a p h i c a l

a b s t r a c t

Simultaneous assay of major biochemical components with a unied

method.
New formulae for the assay of pigments with higher sensitivities.
Standardised pretreatment of samples for the assay of carbohydrates
and proteins.
Conservation of sample, time, chemicals, cost and energy using the unied
assay.

a r t i c l e

i n f o

Article history:
Received 11 January 2013
Received in revised form 18 February 2013
Accepted 3 March 2013
Available online 19 March 2013
Keywords:
Biochemical composition
Pigments
Carbohydrates
Proteins
Lipids

a b s t r a c t
Biochemical compositional analysis of microbial biomass is a useful tool that can provide insight into the
behaviour of an organism and its adaptational response to changes in its environment. To some extent,
it reects the physiological and metabolic status of the organism. Conventional methods to estimate
biochemical composition often employ different sample pretreatment strategies and analytical steps for
analysing each major component, such as total proteins, carbohydrates, and lipids, making it labour-,
time- and sample-intensive. Such analyses when carried out individually can also result in uncertainties
of estimates as different pre-treatment or extraction conditions are employed for each of the component
estimations and these are not necessarily standardised for the organism, resulting in observations that
are not easy to compare within the experimental set-up or between laboratories. We recently reported a
method to estimate total lipids in microalgae (Chen, Vaidyanathan, Anal. Chim. Acta, 724, 6772). Here,
we propose a unied method for the simultaneous estimation of the principal biological components,
proteins, carbohydrates, lipids, chlorophyll and carotenoids, in a single microalgae culture sample that
incorporates the earlier published lipid assay. The proposed methodology adopts an alternative strategy
for pigment assay that has a high sensitivity. The unied assay is shown to conserve sample (by 79%), time
(67%), chemicals (34%) and energy (58%) when compared to the corresponding assay for each component,
carried out individually on different samples. The method can also be applied to other microorganisms,
especially those with recalcitrant cell walls.
2013 Elsevier B.V. All rights reserved.

1. Introduction
Biochemical composition of cells that includes pigments, carbohydrates, proteins and lipids can provide a valuable indication of
the physiology and metabolic status of an organism. The response of

Corresponding author. Tel.: +44 0114 222 7526; fax: +44 0114 222 7501.
E-mail address: s.vaidyanathan@shefeld.ac.uk (S. Vaidyanathan).
0003-2670/$ see front matter 2013 Elsevier B.V. All rights reserved.
http://dx.doi.org/10.1016/j.aca.2013.03.005

an organism to its living conditions, such as light, temperature and

pH can be gleaned by monitoring changes in these gross indicators
with respect to growth and survival. In photosynthetic organisms,
changes in pigment levels can indicate how the organism adapts
its strategy to capture light energy and convert it to biochemical energy. Carbohydrates and lipids are often the major energy
stores, changes in which can be indicative of how the organism perceives its environment and plans to adapt. Changes in total protein
can reect the rate of metabolic activity in actively growing cells.

32

Y. Chen, S. Vaidyanathan / Analytica Chimica Acta 776 (2013) 3140

The cellular distribution of these components varies in response to

environmental stress [13]. Several studies have reported compositional analyses to indirectly monitor the environment quality or
used to stimulate the accumulation of a desired component, for e.g.,
lipids [4,5].
In cases where the biochemical composition of the cells is
desired, measurement of these intracellular components usually
requires several analytical steps and can be time and labour intensive with the requirement of sufcient sample volumes for analysis.
It is highly desirable to have a unied methodology that can be
used to estimate the intracellular composition of gross metabolic
indicators, namely, total pigments, lipids, carbohydrates and proteins. This will not only save on time and effort, but will also
produce a standardised methodology to monitor changes in the
distribution of the physiological indicators from the same sample.
Analysis of total cellular pigments has commonly been carried out by direct extraction in an organic solvent or in a mixture
of solvents, followed by spectrophotometry. A set of formulae
corresponding to the absorbance of the pigments in the solvent employed is subsequently applied to estimate the levels
of different pigments, including chlorophyll a (Chl a), chlorophyll b (Chl b) and total carotenoids (Tc ). Different extraction
solvents and formulae have been reviewed and reported [68].
This method of pigments assay strongly relies on the extraction
efciency, i.e., it depends on the amount of pigments extracted
by the solvent. Underestimation may occur due to incomplete
extraction of pigments from an organism, especially those with
thick cell walls [9]. Even if the extraction is complete, overlap in
absorbance of different pigments at the wavelengths used can cause
interferences. This is an issue that is usually not given due consideration.
A number of methods have been developed for the measurement of carbohydrates, proteins and lipids. However, there is still
ambiguity with respect to pre-treatment of cells prior to the estimations. For example, the phenol-sulphuric acid [10], the orcinol
[11] and the resorcinol [12] methods can all be applied to quantify carbohydrates. Another method that uses anthrone is a widely
employed methodology for the assay of carbohydrates [13,14]. This
simple colorimetric method is relatively robust to interferences
from the other cellular components [13]. However, several studies
that report the use of this method do not explicitly clarify the pretreatment of the sample prior to the estimation step [4,15]. There
are also several methods for the assay of proteins like Folin phenol method [16] and Bradford method [17] but the same issue of
pretreatment as is noted with carbohydrates still exists [4,15]. In
addition, some methods involve a series of repeated procedures for
the protein extraction making it labour intensive [18].
Recently, we reported a reproducible methodology for the
estimation of lipids in microalgae [19]. We have developed this
methodology to incorporate the estimation of pigments, carbohydrates and proteins. Here we report this unied methodology
that enables all the four components to be estimated from a
single sample. Two different microalgal species were tested, Nannochloropsis salina and Dunaliella salina. The former has smaller
cells but thicker cell walls, while the latter has relatively larger
cells and a fragile cell wall. Cells were lysed by glass bead-beating
and saponied in an alkaline solution. The saponication procedure
further promoted the lysis of cells and the release of compounds
making this method more widely applicable for different organisms including those with thick cell walls. We investigated the
spectral characteristics of the saponied pigments and developed
formulae for their quantication. The effects of different treatments and carbohydrate classes on the carbohydrate assay were
also studied, as was the effect of saponication time on the proteins
assay.

2. Experimental
All the solvents and chemicals used were from SigmaAldrich,
unless specied otherwise.
2.1. Microalgae cultivation and harvest
N. salina and D. salina were selected as the model microalgae.
They were obtained from the Culture Collection of Algae and Protozoa (CCAP, UK) and grown in the f/2 medium (recipe from CCAP)
[20]. Stock cultures were grown in 1 L Erlenmeyer ask placed on a
shaking platform at 25 C with 12 h light/dark cycles. For biochemical compositional analysis by using the proposed method, 1.5 mL of
the microalgae culture at different stages of growth (OD of 0.051.5)
were harvested by centrifugation (3000 g for 10 min) and the cell
pellets were stored frozen at 20 C until analysis.
2.2. Reference pigment assay
Cellular pigments were determined using a spectrophotometric
method after extraction with 80% acetone [8,21]. Briey, a 2 mL
aliquot of culture was harvested, washed twice with distilled water
and centrifuged. The pellet was resuspended in 0.4 mL of 0.1 M
phosphate buffer (pH = 7) and disrupted by bead-beating using a
cell disruptor (DISRUPTOR GENIE , USA) for 10 min. 1.6 mL of pure
acetone was added to form an 80% acetone solution, vigorously
vortexed for 2 min and incubated in the dark for 15 min. After centrifugation, the supernatant was decanted to read the absorbance in
a spectrophotometer (Ultrospec 2100 pro, Thomas Scientic, USA).
Subsequently, the amount of pigments was calculated using the
formulae (1)(3) [8]:
Ca = 12.21A663 2.81A646

(1)

Cb = 20.13A646 5.03A663

(2)

Ct =

1000A470 3.27Ca 104Cb

198

(3)

where Ca is the chlorophyll a, Cb is the chlorophyll b, and Ct is the

total carotenoids (g mL1 ).
2.3. Reference carbohydrate assay
Cellular carbohydrates were estimated using the anthrone
method [13] after hot alkaline extraction [22]. Briey, N. salina pellets (obtained from 1.5 mL of culture) were resuspended in 0.2 mL
dH2 O and then heated in 0.4 mL 40% (w/v) KOH at 90 C for 1 h. After
cooling down, 1.2 mL cold absolute ethanol was added and stored
in a fridge at 20 C overnight. The sample was spun down and
the supernatant was discarded. The pellet was then resuspended
in 1.5 mL H2 O. The sample was then reacted with anthrone reagent
and a high concentration of sulphuric acid [13]. An aliquot (0.2 mL)
of sample was mixed and vortexed with 0.4 mL of pre-chilled 75%
H2 SO4 solution in a test tube. To this mixture 0.8 mL of the anthrone
reagent (2 g L1 in 75% H2 SO4 , freshly prepared) was added and subsequently boiled at 100 C for 15 min. After cooling, the absorbance
was read at 578 nm by using a spectrophotometer. It is worth noting that normally in the anthrone method, the wavelength used
is 630 nm. However, sometimes the absorbance was too high for
the tested carbohydrates at this wavelength. Therefore, the shoulder of the peak at wavelength 578 nm was used throughout [23,24].
Blank absorbance of the sample was read by reacting 0.2 mL sample
with 1.2 mL75% H2 SO4 without the anthrone reagent. The amount
of carbohydrate was estimated using a standard curve created using
d-glucose. d-Glucose, sucrose and starch were investigated as standards to study the inuence of the saponication in the method
proposed by us here.

Y. Chen, S. Vaidyanathan / Analytica Chimica Acta 776 (2013) 3140

33

Scheme 1. Proposed combined assay for total pigments, carbohydrates, proteins and lipids. Chl a/b: chlorophyll a/b; Car: carbohydrates; Pro: proteins; Polys: polysaccharides;
Tc : total carotenoids.

2.4. Reference protein assay

3. Results and discussion

The extraction of proteins was performed using alkali [18]. An

aliquot (2 mL) of sample was centrifuged, 1 mL 0.5 N NaOH was
added to the pellet and extracted at 80 C for 10 min with occasional stirring. After cooling and centrifugation, the supernatant
was transferred to a new tube. The alkali extraction was repeated
3 times. The nal repeat was heated at 100 C for 10 min for complete extraction of residual proteins. All the three extractions were
pooled and mixed well before analysis.
The protein content in the extract was estimated using a rapid
and sensitive microbiuret method [25]. An aliquot (0.05 mL) of
copper sulphate (0.21% CuSO4 5H2 O in 30% NaOH) was added to
0.1 mL of the samples and colour formation monitored at 310 nm.
This method was not affected by the presence of high concentrations of deoxyribonucleic acid (DNA). Bovine serum albumin
(BSA) from SigmaAldrich was used as the standard for calibration.

Biochemical compositional analysis is typically carried out in

separate samples. This requires that for each time point several
samples be harvested in parallel from the same culture for the
assay of each component. The sample volume required depends
on the number of assays as well as the harvesting time points.
However, the sample volume taken out from the culture should
be kept to a minimum to avoid the disruption of the organisms
culture conditions, especially in a small scale experimental set-up.
In comparison, simultaneous measurement of several biochemical
components in a single sample would drastically reduce the sample
volume needed and save on time and labour involved. We therefore
investigated the unied procedure proposed above and assessed it
for the estimation of each of the major components in two model
microalgae.

2.5. Proposed combined assay for total pigments, carbohydrates,

proteins and lipids

3.1.1. Spectra of pigments

The two phases resulting from alkaline saponication and
organic extraction in the proposed procedure (Scheme 1) showed
different colours, reecting the components extracted into the
respective phase. The top aqueous phase exhibited a green colour
while the bottom organic phase was yellow. The spectra of these
two phases in the visible region can be seen in Fig. 1.
The aqueous spectrum of N. salina (Fig. 1a) has a strong peak
at 416 nm and a smaller one at 644 nm. These two peaks can be
attributed to chlorophyll a as the strain of Nannochloropsis is known
to have only this type of chlorophyll [26,27]. Interestingly, an additional shoulder at 453 nm was observed in the aqueous spectrum
of D. salina which can be attributed to chlorophyll b as this species
has both chlorophyll a and b [28]. The saponication resulted in the
conversion of chlorophylls a/b to the magnesium-rhodochlorins a/b
(Mg-Rchln a/b) by hydrolytic cleavage of both the phytyl ester bond
and the isocyclic ring [9]. This made the extracted chlorophyll more
hydrophilic and resulted in its partitioning in the aqueous phase
when the organic solvent was added. The structure of Mg-Rchln
a/b is shown in Fig. 1a. It is known that Mg-Rchln a exhibits two
absorbance peaks, one at 416 nm and the other at 641.2 nm, while
Mg-Rchln b has prominent absorbances at 452.6 nm and 623 nm in
alkaline methanol [9]. However, the peak of Mg-Rchln b at 623 nm
is not seen in the aqueous spectrum of D. salina as it is too small to
observe and overlaps with the peak of Mg-Rchln a at 644 nm.
The spectra of the bottom organic phase of N. salina and D.salina
can be seen in Fig. 1b. Both the species show absorbance maxima
at 430 nm, 450 nm and 480 nm. The absorbance can be attributed
to intact carotenoids (such as -carotene and lutein) that are stable
under the alkaline conditions used. Yuan and Chen [29] employed
high performance liquid chromatography (HPLC) to study the differences in pigments from Haematococcus pluvialis before and
after saponication. They found that all of the chlorophylls were

As shown in Scheme 1, pigments, carbohydrates and proteins

were extracted and assayed at different stages during the assay of
lipids we have previously reported [19]. Briey, harvested pellets
were resuspended in 20 L phosphate buffer (0.05 M, pH 7.4), to
which 0.48 mL of R1 (25% methanol in 1 N NaOH) was added and
disrupted by bead-beating for 10 min. The phosphate buffer and R1
were both pre-boiled to remove any entrained air and prevent oxidation. To the lysate another aliquot (1 mL) of R1 was added and
vortexed. Two aliquots (0.2 mL) (one to act as blank and the other as
sample) were taken to measure the total carbohydrates or the dissolved carbohydrates in the supernatant following centrifugation,
using the anthrone method detailed above.
The remaining sample was saponied by heating at 100 C for
30 min and cooled down to room temperature. Two aliquots (0.1 mL
each, one to act as blank and the other as sample) were transferred to eppendorf tubes and centrifuged; the supernatant was
used for estimating proteins using the microbiuret method detailed
above.
Another aliquot (0.5 mL) of sample was pipetted to an
eppendorf tube containing 0.75 mL of a solvent mixture R2 (chloroform/methanol, 2:1, v/v) and vortexed for 2 min. The mixture was
spun at 12,000 g for 2 min to get two phases. The top aqueous
phase was used for the assay of chlorophyll, while the lower organic
phase was used for the assay of total carotenoids and lipids. The
absorbance of the lower phase at 430 nm, 450 nm and 480 nm (for
total carotenoids) was co-read with that at 260 nm (for lipids) after
the organic phase was reacted with R3 (1 M triethanolamine:1 N
acetic acid, 9:1, v/v) required for the lipid assay. All the absorbances
were normalised to the absorbance at 750 nm to eliminate errors
caused by baseline drifts.

3.1. Estimation of pigments

34

Y. Chen, S. Vaidyanathan / Analytica Chimica Acta 776 (2013) 3140

Fig. 1. Spectra of saponied chlorophylls (a) in aqueous phase and total carotenoids (b) in organic phase. The inset in (a) shows the structure of Mg-Rchln a/b, where R is
methyl group in Mg-Rchln a and formyl group in Mg-Rchln b.

removed from the organic solvent by saponication. In comparison,

astaxanthin ester was hydrolyzed and converted to astaxanthin.
This and most other carotenoids, including -carotene, adonirubin,
canthaxanthin and lutein, were found to be retained in the organic
phase, intact.
It is noteworthy that pre-heating or blowing nitrogen is necessary to remove any entrained oxygen from the alkaline solution
before saponication, as the presence of oxygen is known to lead to
the destruction and structural transformation of some carotenoids
[29,30]. Porra [9] found that the addition of a reducing agent
to the alkaline solution was not necessary because the heating
during saponication eliminated allomerization of chlorophylls.
Pre-heating the solvent and/or blowing nitrogen before saponication and manipulations under darkness should be sufcient to
take care of this concern.
3.1.2. Determination of chlorophylls
Natural chlorophylls generally show two noticeable peaks in the
visible region, a major peak at about 400 nm and a minor peak in
the red region, at around 650 nm. Although the major peak has a
higher sensitivity, the wavelength of the minor peak has been commonly adopted for the determination of chlorophylls in all previous
reports, including that of Porras [9] (where the chlorophylls were
converted to Mg-Rchlns). The reason for this option is to evade
the signicant interference of carotenoids which mainly absorb in
the vicinity of the blue light (400500 nm). Nevertheless, we have

been able to utilise the major peaks of Mg-Rchln a/b at 416 nm and
453 nm, as a result of the extraction procedure adopted here, where
the carotenoids are partitioned in the organic phase and hence do
not interfere. The choice of the stronger peaks at 416 and 453 nm
enables higher sensitivities to be achieved for the determination of
chlorophylls, as shown in Fig. 2, with little or no interference from
carotenoids.
Serial dilutions of N. salina and D. salina samples were saponied
in R1 and used to establish a correlation between the absorbance at
416 nm and 453 nm and chlorophylls determined by the reference
method. As can be seen from Fig. 2a, a strongly linear relationship can be observed. The slopes of the correlation for chlorophyll
a from N. salina at 416 nm and 453 nm were directly taken as the
extinction coefcients of chlorophyll a at 416 nm and 453 nm (E416a
and E453a ) because N. salina is known to have only chlorophyll a. It
needs to be pointed out that, in D. salina, the absorbance of chlorophyll a at 416 nm contains the contribution from chlorophyll b
and the absorbance of chlorophyll b at 453 nm contains the contribution from chlorophyll a. Consequently, E416a and E453a derived
from the N. salina spectra were used to calculate the absorbance of
chlorophyll a in D. salina at 416 nm and 453 nm, based on the reference method estimations. The calculated chlorophyll a absorbance
was then subtracted from the observed absorbance of D. salina at
the two wavelengths (416 nm and 453 nm), in order to remove
the interference from chlorophyll a and calculate the extinction
coefcients of chlorophyll b at 416 nm and 453 nm (E416b and E453b ).

Fig. 2. Spectral sensitivity to chlorophylls. (a) Absorbance of chlorophyll a and b in N. salina and D. salina at different wavelengths as a function of chlorophyll concentration;
(b) relationship of the predicted chlorophylls using the present coefcient equations and the chlorophylls estimated using the 80% acetone method [8,21]; (c) test limitation
of 0.025 g mL1 for both chlorophylls is encircled in the inset.

Y. Chen, S. Vaidyanathan / Analytica Chimica Acta 776 (2013) 3140

35

Table 1
Comparison of extinction coefcients at corresponding wavelengths among different methods. ad are the corresponding extinction coefcients (g mL1 ).
Method

a/b

d/c

Present method
Porras method [9]
80% acetone [8]
Chloroform [8]
Methanol [8]
Diethyl-ether [8]
Dimethyl-formamide [8]
Dimethyl-sulphoxide [8]

0.1571
0.0611
0.0869
0.0946
0.0793
0.101
0.0904
0.0881

0.0212
0.0063
0.0121
0.0069
0.0215
0.0048
0.0117
0.0138

0.0064
0.0162
0.0217
0.0264
0.0329
0.0194
0.0197
0.0213

0.1711
0.0218
0.0527
0.063
0.0459
0.062
0.0506
0.0488

7.41
9.69
7.18
13.65
3.69
21.26
7.74
6.37

26.73
1.35
2.43
2.39
1.4
3.2
2.57
2.29

The absorbance of chlorophylls at 416 nm and 453 nm can be

represented using formulae (4) and (5):
A416 = E416a Ca + E416b Cb

(4)

A453 = E453a Ca + E453b Cb

(5)

where Ca and Cb are the concentrations of chlorophylls a and b

(in g mL1 ), and E416a , E453a , E416b and E453b are the apparent
extinction coefcients determined as detailed above. The relevant
concentrations can then be calculated from the following expressions:
Ca = 6.40 A416 0.79 A453

(6)

Cb = 5.87 A453 0.24 A416

(7)

For calculating chlorophyll a concentration, when it is the only

type of chlorophyll present in the organism, the following formula
can be applied,
Ca = 0.1571 A416

(8)

The generalised Eqs. (6)(8) can be applied to species containing

chlorophyll a and/or chlorophyll b, which is the case in most species.
The empirical constants are based on the organisms and conditions
employed in the study. For species containing other chlorophylls
in addition to chlorophylls a and b, the equations may have to be
reassessed. It should be pointed out that the formulae (6)(8) are in
terms of the chlorophyll concentrations in the aqueous phase after
extraction by organic solvent R2. To eventually calculate the chlorophyll concentrations in the real samples, a dilution effect should be
considered as methanol in the organic solvent can go into the aqueous phase leading to the increase in the volume of aqueous phase.
This dilution factor was found to be around 1.4 by measuring the
volumes of the two phases after mixing and centrifugation. This
was used in the calculations to arrive at the chlorophyll concentrations in D. salina and N. salina. The results were compared with
the 80% acetone method [8,21], as shown in Fig. 2b. The results
show a strong correlation, suggesting that the proposed method is
applicable for chlorophyll assay.
In order to compare the performance of the method proposed
here with that of others in published literature, formulae (4) and
(5) were generalised as below ((9) and (10)):
A1 = a Ca + b Cb

(9)

B2 = c Ca + d Cb

(10)

where 1 is the major peak wavelength of chlorophyll a; 2 is

the major peak wavelength of chlorophyll b; ad are the corresponding extinction coefcients of chlorophyll a at 1, chlorophyll
b at 1, chlorophyll a at 2 and chlorophyll b at 2, respectively.
Table 1 shows a comparison of the apparent extinction coefcients
calculated in the proposed method with that reported by Porras
[9] and Wellburn [8], where six other organic solvents were compared. It must be noted that the resolution used for all the methods
compared was in the 14 nm range. It can be seen that both a
and d values in the proposed method were much higher than that

reported for the other methods, especially the value of d which

was between 3 and 8 folds higher. It indicates that the method
proposed in this investigation has the highest sensitivity as a and
d are the extinction coefcients of chlorophyll a and b at their
major peaks. Meanwhile, b and c, respectively represent the interference of chlorophyll b at 1 when calculating chlorophyll a and
the interference of chlorophyll a at 2 when calculating chlorophyll b. Therefore, the ratio of a/b and d/c can be used to represent
the extinction coefcient that accounts for interference. Higher the
ratio, lower the interference. As can be seen from Table 1, the ratio
of a/b for the method proposed by us in this investigation is around
the average for all the methods, while d/c is the highest among the
methods compared.
3.1.3. Determination of total carotenoids
The structural transformation of chlorophylls caused by saponication leads them to separate from the family of fat-soluble
pigments. As a result, the total pigments are segregated into
two parts: water-soluble chlorophylls in the aqueous phase (light
green) and fat-soluble carotenoids in the organic phase (yellow).
Because three prominent carotenoid peaks were present in the
organic phase (Fig. 1b), we investigated the feasibility of using these
peaks for the total carotenoid estimation. A linear relationship of
the carotenoids measurement can be observed (Fig. 3a) between
the carotenoid concentration measured by the 80% acetone method
[8] and the absorbance at the three wavelengths (430 nm, 450 nm
and 480 nm) for N. salina. Different sensitivities were noted at each
wavelength, and the observations were similar for D. salina (not
shown). The formulae for N. salina and D. salina at these three wavelengths are compared to explore if there is one universal formula
which can be used for carotenoids measurement independent of
species, as shown in Table 2. Although the correlation of the sensitivities among these three wavelengths is similar between these
two microalgal species (A450 > A480 > A430 ), the slopes are not the
same between these two species at each wavelength. This might
be due to the differing composition of carotenoids, resulting in
different contributions to the absorbance. Therefore, it would not
be prudent to apply the formulae generated from one species to
another species.
However, the difference in absorbance A450430 and A480430
which is also proportional to the concentration of total carotenoids
(Fig. 3b) can be seen to be less inuenced between the two species.
The value of , as given in the last column of Table 2, is the difference of two slopes divided by their average, indicating the varied
degree of those two tted lines obtained from the two species at one
wavelength. Lower was found in the difference curves justifying
the corresponding lines and formulae. Especially for the A480430
curve, it only produces a smallest of 0.02. It appears that the
variation in the composition of carotenoids could be minimised
by subtracting the absorbance at these wavelengths. The resulting
formulae for these two microalgae species were found to be similar (Table 2). Average formulae of y = 0.0425x and y = 0.0295x were
eventually adopted for total carotenoids determination by using
the absorbance difference of A450430 and A480430 , respectively. The

36

Y. Chen, S. Vaidyanathan / Analytica Chimica Acta 776 (2013) 3140

Fig. 3. Correlation of absorbance at three wavelengths (a), and spectral difference at the given wavelengths (b) with total carotenoid concentration (Tc ), estimated by the
reference method, in serial dilutions of N. salina.

detection limit of the present method for the assay of chlorophyll

a, chlorophyll b and total carotenoids were found to be 0.025, 0.025
and 0.006 g mL1 , respectively.
3.2. Estimation of carbohydrates
3.2.1. Effects of different pre-treatments
The most common methods used for the assay of carbohydrates
include the phenol-sulphuric acid [31,32] and the anthronesulphuric acid methods [33]. Phenol-sulphuric acid method was
rst developed by Dubois et al. in 1950s [31,32]. Simple sugars,
oligosaccharides, polysaccharides, and their derivatives form an
orange yellow colour when reacted with 80% phenol and concentrated sulphuric acid in a water bath at 2530 C for 1020 min.
The absorbance of the characteristic yellow orange colour is measured at 490 nm for hexoses and 480 nm for pentoses and uronic
acids. Anthrone-sulphuric acid method for the measurement of
carbohydrate material was rst described by Dreywood [33] in
1946. Carbohydrates are treated (no additional heating) with 0.2%
solution of anthrone in concentrated sulphuric acid whose nal
concentration in the test solution should always be greater than
50% to avoid the production of a milky suspension. The absorbance
of the green colour produced is measured at 630 nm. These two
methods have been extensively employed for the colorimetric estimation of a variety of carbohydrates [3437]. However, the original
forms of these two methods had some common issues. For example,
their sensitivities are not high enough for the measurement of low
levels of carbohydrates; different hexose and pentoses show different responses. These problems were solved for the later method
by Yemm and Willis [38] by increasing the heating temperature
to 100 C and pre-cooling the anthrone reagent. By doing this, the
sensitivity was greatly improved and no evidence of serious interaction of hexoses and pentoses were found. This high temperature
condition was therefore adopted by the modern versions [22,39]. In

Schneegurts method [22], before reaction with anthrone the carbohydrates were extracted by a hot KOH solution followed by cold
precipitation of carbohydrates before colour development. Nevertheless, recent estimation approaches appear to gloss over the
extraction procedure [4,15,39]. In these studies, there is no reference to sample pretreatment such as physical lysis, except washing
the cell pellets before conducting the assay. For instance, Laurens
et al. [39] employed the phenol-sulphuric acid method to estimate
the carbohydrates in the water-resuspended biomass. It appears
that the sample with intact cells can be directly applied to the carbohydrate assay given the digestion by the concentrated sulphuric
acid. However, it is unclear whether a separate extraction approach
before colour formation would have made a difference or not. Evidently, the success of the method depends on the sample material
and the effectiveness of the employed methodology in reproducibly
extracting the carbohydrates. Therefore, it is necessary to clarify the
kind of treatment required for a given sample. Scheme 2 illustrates
the six treatments that we examined in this investigation, including the hot alkaline extraction method adopted by Schneegurt [22].
All the samples were subsequently subjected to the carbohydrate
assay using the anthrone-sulphuric acid method.
The results of the pre-treatment are compared in Fig. 4a. It
can be observed that the original culture sample without any pretreatment yielded the maximum absorbance. This either indicates
that there are extracellular carbohydrates in the medium that
are detected, in addition to intracellular carbohydrates, or that
the method is inuenced by media components giving an overestimation. After a wash step, the medium was substituted by
distilled water and the measured carbohydrate content decreased.
Lysis of cells with the help of glass beads led to a higher carbohydrate content than water-washed cell pellets. This illustrates
that more carbohydrates were released by the physical grinding
of glass beads. The saponication step appears to reduce the carbohydrate level to similar levels observed in the washed pellets.

Table 2
Formulae for total carotenoids at different wavelength, y: absorbance; x: concentration (g mL1 ). The value of is the difference of two slopes divided by their average, i.e.,
= (S1 S2 )/0.5*(S1 + S2 ), S1 and S2 are the two comparing slopes.
Wavelength (nm)

A430
A450
A480
A450430
Average
A480430
Average

N. salina

D. salina
2

Formula

Formula

y = 0.096x
y = 0.1364x
y = 0.1258x
y = 0.0404x

0.9835
0.9831
0.9818
0.9806

y = 0.1189x
y = 0.1634x
y = 0.1481x
y = 0.0445x

0.9562
0.9585
0.9603
0.9636

0.213
0.18
0.163
0.096

y = 0.0298x

0.9729

y = 0.0292x

0.9713

0.02

y = 0.0425x
y = 0.0295x

Y. Chen, S. Vaidyanathan / Analytica Chimica Acta 776 (2013) 3140

37

Scheme 2. Pre-treatment methods examined prior to anthrone sulphuric acid estimation of carbohydrates. Sample numbers indicated correspond to the ones used in Fig. 4a.

This might be attributable to the degradation of carbohydrates

during the saponication. The addition of organic solvent R2 separated the sample into three phases (aqueous, organic and solid)
and the aqueous phase contained only the water-soluble carbohydrates yielding the lowest value. In addition, it can be observed
that the carbohydrate content using Schneegurts method [22] is
lower than that in the lysed sample. It is possible that some carbohydrates were lost in the precipitation step employed in the
Schneegurt method.
To further understand how these pre-treatments inuence the
carbohydrate assay, we measured the water-soluble and -insoluble
carbohydrate fractions in the samples with three different treatments: cell lysis in distilled water, cell lysis in R1, and saponication
in R1 but without organic extraction. The sample was centrifuged
after treatment and both the pellet and supernatant were analysed
for carbohydrates. Fig. 4b shows the results observed. As can be
seen, lysis in the alkaline solution (R1) enhanced both the total and
the supernatant carbohydrates and showed a lower level in the
pellet. It appears that alkali promoted the release of more carbohydrate, perhaps from complexes like glycoprotein or glycolipid by

breaking the bonds between carbohydrate and other compounds

[40,41]. The saponication procedure decreased the total carbohydrate a little while dramatically extracted more carbohydrate from
the precipitate to the supernatant.
Therefore, the wash step and the cell lysis are required to effectively release cellular carbohydrates. Lysis in the alkaline solution
(R1) further released the carbohydrate. Saponication led to a
decrease in the total carbohydrate, perhaps due to the degradation of monosaccharides like glucose [42]. The aqueous phase after
saponication and stratication showed lower carbohydrate levels
possibly due to the fact that some polysaccharides such as starch
can be extracted into the organic solvent, and therefore lost from
the aqueous phase. Moreover, it needs to be pointed out that, for
precise measurements, two blanks should be included for each
sample. One is a general blank that contains anthrone but no algae,
and the other has the algal sample but does not contain anthrone.
The second blank is required because we found that the pigments
contained in the algae have some absorbance (depending on the
pigment concentration) at the wavelength used and this interferes
with the estimation of carbohydrates.

Fig. 4. Effects of different pre-treatment on the assay of carbohydrates (triplicate determinations). (a) Total carbohydrate in the same samples by different treatment (for
details refer to Scheme 2). (b) Carbohydrate proportion in supernatant and pellet from samples subjected to water-washing (H2 O), lysis in R1 reagent (R1) and saponication
(R1 + Sap.), respectively.

38

Y. Chen, S. Vaidyanathan / Analytica Chimica Acta 776 (2013) 3140

3.3. Effect of saponication time on the protein assay

Fig. 5. Effects of saponication on the assay of carbohydrate (triplicate determinations).

3.2.2. Effects of carbohydrate type on the assay

As discussed above, the saponication step decreased the
detectable amount of carbohydrates (Fig. 4a). We therefore examined the saponication step for three different carbohydrate
types, a monosaccharide (glucose), a disaccharide (sucrose), and a
polysaccharide (starch), all at the same concentration of 0.09 g L1 .
We looked at the carbohydrate levels estimated with or without
the saponication step by using the reference carbohydrate assay
based on the anthrone-sulphuric acid method. In the unsaponied
samples glucose and starch were estimated precisely, while sucrose
appears to be slightly over-estimated (Fig. 5). In the saponied
samples, the glucose estimation fails, perhaps due to its oxidation to gluconic acid in the hot alkali [42]. Nevertheless, it has
been known that the alkaline conditions promoted the release of
oligosaccharides from glycoproteins [40,41]. The concentrations of
sucrose and starch were similar to those without saponication.
When we spiked an amount of starch (0.1 g L1 ) in an algal sample
and subjected it to saponication, we found the recovery of starch
was 101.2 1.6%. The addition of organic solvent R2 decreased
the amount of starch tested indicating that starch was possibly
extracted into the organic phase. Therefore, the lysed sample in R1
(methanol/NaOH) can be used to measure the total carbohydrate
because the physical lysis maximises the release of carbohydrates.
Saponied sample can be used for the assay of polysaccharides. The
extraction of organic solvent separates carbohydrates into watersoluble and water-insoluble fractions. Estimation on these two
fractions sometimes might be needed to provide more information
about the composition of the cellular carbohydrates.

All of the complicated metabolic reactions are associated with

the catalysis of various enzymes. The level of proteins, to some
extent, indicates the physiological status and therefore the assay
of proteins in organisms is usually required. Rausch [18] developed
a method of protein extraction in which the sample was heated in
0.5 N NaOH twice at 80 C for 10 min and once at 100 C for another
10 min. However, this strategy is both time and sample consuming as it requires repeated sample centrifugation and transfer. The
repeated heating in fresh NaOH solution has a dilution effect on
the tested sample so that the amount of samples should not be too
small. One step for protein extraction would ideally be desired to
simplify the assay. In the proposed method, samples are saponied
in 1 N NaOH in 25% Methanol (R1) instead of NaOH solution. As the
inuence of the reagent on the efciency of protein extraction and
estimation is not known, we investigated the effect of saponication time on the protein assay. The results are exhibited in Fig. 6a. It
can be seen that the protein standard (BSA) was hardly affected by
the saponication procedure for 30 min. The protein content of N.
salina reached its maximum value at 10 min and kept stable within
30 min. In comparison, the protein content of D. salina increased
gradually with respect to time and appears to plateau at 25 min.
An algal sample spiked with BSA (0.056 g L1 ) was subjected to
saponication and the recovery of BSA was found to be 101.3 2.2%.
The addition of organic solvent R2 had no signicant inuence on
the protein assay but still slightly reduced the estimation by 4%. As
discussed above, the saponication treatment used for lipid assay
does not appear to have any signicant inuence on the protein
assay. It promotes the release of cellular proteins and can fully
extract all the proteins. Saponied samples before organic extraction can be used for the protein assay. This is also veried by a good
correlation between the proteins levels estimated by the method
and that taken for analysis (Fig. 6b).
3.4. Biochemical composition in N. salina and D. salina
A practical experiment was conducted to apply the present
method to analyse the biochemical composition of two species of
microalgae (N. salina and D. salina). Cells were harvested to measure dry weight of biomass based on the pre-determined standard
curves of OD against dry weight and stored for biochemical analysis. The time proles for dry cell weight, dissolved carbohydrates,
proteins and lipids are shown in Fig. 7a; carbohydrates, proteins
and lipids as a percent of dry biomass is plotted in Fig. 7b; time
prole of pigments is shown in Fig. 7c.
In the ten days of cultivation, the amount of biomass as well
as the biochemical components all increased with respect to time

Fig. 6. (a) Effect of saponication time on the assay of proteins (triplicate determinations); (b) correlation between known and estimated protein concentrations.

Y. Chen, S. Vaidyanathan / Analytica Chimica Acta 776 (2013) 3140

39

Table 3
Comparison of consumptions for analysing ten samples in terms of time, material and cost between the proposed method and the individual methods. P represents the GBP
pounds. Cost estimation was on the basis of the UK rates of chemicals and electricity, as on 5th November, 2012.
Assays

Sample (mL)

Time (h)

Chemicals (P)

Energy (P)

Method

Pigments
Carbohydrates
Proteins
Lipids
Sub-total
Simultaneous
Conservation %

2.00
1.50
2.00
1.50
7.00
1.50
78.57

3.00
3.00
3.00
3.00
12.00
4.00
66.67

0.59
0.43
0.32
0.39
1.73
1.14
34.15

0.09
0.08
0.14
0.18
0.49
0.21
58.18

80% acetone [8]

Anthrone-sulphuric acid [13]
Rausch [18] and Itzhaki [25]
Chen [19]

(Fig. 7a). D. salina contained more pigments than N. salina (both

chlorophylls and carotenoids) possibly enhancing light energy harvest by this species (Fig. 7c). The protein content in D. salina was
higher than that in N. salina (Fig. 7a and b) indicating that the physiological reactions in D. salina might be more active or complicated
than that in N. salina. The carbohydrate content in D. salina was
also higher than that in N. salina, but the lipids was not (Fig. 7a and
b). The initial lipid content was higher than other growing stages
in these two species due to the inoculum from stationary phase in
which lipid is usually accumulated. It demonstrates that under optimum conditions without deciency of any nutrients, microalgae do
not show a tendency to accumulate lipids. These two microalgae
were only cultured up to the logarithmic phase and no signicant
lipid accumulation was observed, presumably because their growing cycles have not stepped into the stationary phase when nitrogen
becomes decient. By comparing these two species, it appears that
N. salina tends to store the energy and carbon in the form of lipids
instead of carbohydrates. Overall, the biochemical composition of
the two species, estimated with the unied methodology, for the

The proposed method

cultures harvested towards the end of the cultivation (240 h for N.

salina and 260 h for D. salina), was found to be: carbohydrates28%,
proteins35% and lipids28%, others9%, for N. salina and 30%, 55%,
11% and 5% respectively for D. salina. These values were found to
be in the range reported in the literature [43,44].
3.5. Cost and energy savings with the unied methodology
To better understand the advantages of the proposed unied
assay, the cost and energy savings for analysing ten microalgal
samples were compared with the scenario where the components
were estimated individually (Table 3). Indicative cost of chemicals
involved was estimated using the SigmaAldrich catalogue prices
as on 5th November, 2012. The energy consumption contribution
is primarily for cell lysis by bead-beating, centrifugation and heating, the cost for which was evaluated on the basis of electricity
cost 0.137 pence/kW h. It can be seen that only 21% of the sample
volume was required for the unied assay, saving on 79% of the
sample. The fewer requirements for samples enable this method
to be a better choice when the scale of experiment is limited. It
also saved on time consumption making the compositional analysis faster, especially when the number of samples is large. The
conservation on chemical cost was the least at 34% as most of the
chemicals are still needed for the simultaneous assay except the
acetone used for pigments extraction. The proposed method does
not need any additional solvent for pigments assay. Reduction in
the cost of energy was considerable up to nearly 60%. Because the
energy consumption was evaluated based on the time consumed
for major devices, it also includes the using frequencies of these
devices, although the wear and tear of the devices is not included
as it could not be measured.
4. Conclusions

Fig. 7. Growth curve and biochemical composition in N. salina (left) and D. salina
(right). (a) Yields of biomass (dry weight), carbohydrate (Car), protein (Pro) and
lipid (Lip); (b) percentage content of carbohydrate, protein and lipid in the basis of
biomass; (c) yields of chlorophyll a (Chl a), chlorophyll b (Chl b) and total carotenoids
(Tc ).

We have successfully developed and demonstrated a unied

procedure for the simultaneous assay of microalgae biochemical composition (namely, pigments, lipids, carbohydrates and
proteins). We have extended an earlier protocol for lipid assay
in microalgae, published by us, that now enables a unied
methodology to be adopted for monitoring the major biochemical compositions in microalgae. The coefcient of variation (CV)
of the triplicate determinations was typically <5%. Total carbohydrates should be assayed after cell lysis in alkaline solution
but before saponication, as saponication results in the degradation of monosaccharides. Saponied samples can be used for
the estimation of polysaccharides and proteins but before the
addition of the organic solvent to extract lipids, because some of
carbohydrates and proteins can be co-extracted with lipids. The
chloroform/methanol (2:1) solvent straties the sample into two
phases. The top aqueous phase can be used for chlorophyll assay,
while the lower organic phase can be used for total carotenoids
and lipids assay. In fact, saponication converts chlorophylls a/b to
magnesium-rhodochlorins a/b (Mg-Rchln a/b) by hydrolytic cleavage of both the phytyl ester bond and of the isocyclic ring. These

40

Y. Chen, S. Vaidyanathan / Analytica Chimica Acta 776 (2013) 3140

Mg-Rchlns a/b are hydrophilic and consequently dissolve in the top

aqueous phase; meanwhile the total carotenoids are dissolved in
the lower chloroform phase. We have also developed a new set of
formulae for the pigment assay that gives higher sensitivities than
existing methods.
Acknowledgements
We gratefully acknowledge Chinese Scholarship Council (CSC)
scheme and EPSRC UK (ChELSI) for the funding support that made
this work possible. We are also grateful to Dr. Jim Gilmour for the
provision of D. salina cultures that we based our early work on.
References
[1] R.J. Stevenson, Y. Pan, The Diatoms: Applications for the Environmental and
Earth Sciences, vol. 1, 1999, 4.
[2] D. Redalje, E. Laws, J. Exp. Mar. Biol. Ecol. 68 (1983) 59.
[3] M.A. Torres, M.P. Barros, S.C.G. Campos, E. Pinto, S. Rajamani, R.T. Sayre, P.
Colepicolo, Ecotoxicol. Environ. Safe. 71 (2008) 1.
[4] S.L. Pahl, D.M. Lewis, F. Chen, K.D. King, J. Appl. Phycol. 22 (2010) 165.
[5] B. Allard, M. Danger, L. Ten-Hage, G. Lacroix, Aquatic Sciences-Research Across
Boundaries (2011), 1.
[6] D. Skran, G. Tohit, S. Rdvan, Turk. J. Bot. 22 (1998) 13.
[7] R.J. Ritchie, Photosynth. Res. 89 (2006) 27.
[8] R. Wellburn, J. Plant Physiol. 144 (1994) 307.
[9] R.J. Porra, Biochim. Biophys. Acta (BBA): Bioenergetics 1015 (1990) 493.
[10] T. Masuko, A. Minami, N. Iwasaki, T. Majima, S.I. Nishimura, Y.C. Lee, Anal.
Biochem. 339 (2005) 69.
[11] M. Irwin, A. Leaver, Nature 177 (1956) 1126.
[12] M. Monsigny, C. Petit, A.C. Roche, Anal. Biochem. 175 (1988) 525.
[13] P. Gerhardt, R.G.E. Murray, W.A. Wood, N.R. Krieg (Eds.), Methods for General
and Molecular Bacteriology, ASM, Washington, DC, 1994.
[14] A. Leyva, A. Quintana, M. Snchez, E.N. Rodrguez, J. Cremata, J.C. Snchez,
Biologicals 36 (2008) 134.
[15] Y. Sun, C. Wang, Biotechnol. Bioprocess Eng. 14 (2009) 225.

[16] O. Classics Lowry, N. Rosebrough, A. Farr, R. Randall, J. Biol. Chem. 193 (1951)
265.
[17] T. Zor, Z. Selinger, Anal. Biochem. 236 (1996) 302.
[18] T. Rausch, Hydrobiologia 78 (1981) 237.
[19] Y. Chen, S. Vaidyanathan, Anal. Chim. Acta 724 (2012) 67.
[20] C. F/2 medium. Accessed from http://www.ccap.ac.uk/media/documents/f2.pdf
[21] J. Kirk, Planta 78 (1967) 200.
[22] M.A. Schneegurt, D.M. Sherman, S. Nayar, L.A. Sherman, J. Bacteriol. 176 (1994)
1586.
[23] X.-P. Liu, T.E. Grams, R. Matyssek, H. Rennenberg, Plant Physiol. Biochem. 43
(2005) 147.
[24] W. Zwerschke, B. Mannhardt, P. Massimi, S. Nauenburg, D. Pim, W. Nickel, L.
Banks, A.J. Reuser, P. Jansen-Drr, J. Biol. Chem. 275 (2000) 9534.
[25] R. Itzhaki, D. Gill, Anal. Biochem. 9 (1964) 401.
[26] L.M. Lubin, O. Montero, I. Moreno-Garrido, I.E. Huertas, C. Sobrino, M.
Gonzlez-del Valle, G. Pars, J. Appl. Phycol. 12 (2000) 249.
[27] J.S. Brown, Plant Physiol. 83 (1987) 434.
[28] W. Chen, N. Jin, Y. Shi, Y. Su, B. Fei, W. Li, D. Qiao, Y. Cao, Photosynthetica 48
(2010) 355.
[29] J.P. Yuan, F. Chen, J. Agric. Food Chem. 46 (1998) 3371.
[30] G. Lietz, C. Henry, Food Chem. 60 (1997) 109.
[31] M. Dubois, K. Gilles, J. Hamilton, P. Rebers, F. Smith, Nature 168 (1951)
167.
[32] M. Dubois, K.A. Gilles, J.K. Hamilton, P. Rebers, F. Smith, Anal. Chem. 28 (1956)
350.
[33] R. Dreywood, Ind. Eng. Chem. Anal. Ed. 18 (1946) 499.
[34] J.F. Robyt, S. Bemis, Anal. Biochem. 19 (1967) 56.
[35] D. Liu, P. Wong, B. Dutka, Water Res. 7 (1973) 741.
[36] F.J. Viles Jr., L. Silverman, Anal. Chem. 21 (1949) 950.
[37] S. Seifter, S. Dayton, B. Novic, E. Muntwyler, Arch. Biochem. 25 (1950) 191.
[38] E. Yemm, A. Willis, Biochem. J. 57 (1954) 508.
[39] L.M.L. Laurens, T.A. Dempster, H.D.T. Jones, E.J. Wolfrum, S. Van Wychen, J.S.P.
McAllister, M. Rencenberger, K.J. Parchert, L.M. Gloe, Anal. Chem. 84 (2012)
1879.
[40] S.I. Ogata, K.O. Lloyd, Anal. Biochem. 119 (1982) 351.
[41] T. Nakajima, C.E. Ballou, J. Biol. Chem. 249 (1974) 7679.
[42] C. Degani, M. Halmann, J. Am. Chem. Soc. 90 (1968) 1313.
[43] M. Rebolloso-Fuentes, A. Navarro-Perez, F. Garcia-Camacho, J. Ramos-Miras, J.
Guil-Guerrero, J. Agric. Food Chem. 49 (2001) 2966.
[44] A. Singh, S.I. Olsen, Appl. Energy 88 (2011) 3548.