Bioresource 7-echnology 38 ( 1991 ) 115-119

Lipid Production by Phaeodactylum tricornuturn

V. Veloso," A. Reis, b L. Gouveia," H. L. Fernandes, b J. A. Empis," & J. M. Novais"
"Laboratdrio de Engenharia Bioqufmica/IST, Avenida Rovisco Pais, 1096 Lisboa Codex, Portugal
t'Departamento de Energias Renovfiveis/LNETI, Estrada do Paqo do Lumiar, 1699 Lisboa Codex, Portugal

Abstract (EPA), are included. PUFAS were claimed as pre-

cursors for prostaglandins (Cohen et al., 1988a),
7he results of a three month investigation on the playing an important role as agents to prevent
outdoor mass culture of Phaeodactylum tri- blood platelet aggregation (Cohen et al., 1988a;
cornutum Bohlin are reported. Experiments" car- Seto et al., 1984). In this work, lipid production by
ried out from January to March 1989 gave an Phaeodac~lum tricornutum is considered in par-
average output rate of 4"O g m - 2 day -I in ash-free ticular in what affects the yields of PUFAS.
dry weight (AFD W) basis for semicontinuous mode
(days 0--74) and 2"0 g m - : day -1 for batch mode
(days 74-84). Eicosapentaenoic acid (20.'5 ~o3 METHODS
EPA) was the predominant fatty acid synthesized
by this diatom during semicontinuous growth con- Organism and growth media
stituting 3"9% of AFD W, yielding 0.15 g m : day-i. Phaeodactylum tricornutum Bohlin S/PHAEO-
7he content of EPA was decreased to 0"7% of I(TFX-1) was obtained from the Solar Energy
AFD W during batch mode, 16:0 and 16:1 becom- Research Institute (SER1) Culture Collection
ing the main fatty acids. The relationship between (Golden, Colorado, USA) and was cultivated in
optical density and output productivity rate was water collected from the Tagus Estuary, enriched
studied. In the range studied, the lower the con- with some components (KNO3, K:HPO 4,
centration the higher the productivity and the lower Na2SiO3, trace elements) added according to the
the lipid production rate. Several flocculant systems GMP medium composition (SER1, 1986). NaCI
were tried and calcium hydroxide, Ca(OH):, was was added in order to obtain a salinity of 30 g
found to be best at concentrations from 30 to 100 iitre-~.
mg litre - ~, in an apparently random way. The addi- Growth conditions
tion of chitosan did not improve the efficiency of Phaeodactylum tricornutum was cultiwlted in
flocculation. 2-2 m 2 PVC ponds in a polyethylene sheet
covered greenhouse. The depth of the culture was
Key words." Mass culture, microalgae, diatom, 10 cm. Agitation was obtained by a paddle-wheel
Phaeodactvlurn tricornutum, lipid composition, that maintained a water velocity of 25 cm s-L
EPA, N-deficiency, flocculation. Harvesting and dilution was performed daily in
order to maintain the desired optical density
values.
INTRODUCTION
Description of experiment
Microalgae are biological systems that transform This experiment was carried out for 84 days from
CO 2 into organic matter utilizing solar radiation January to March 1989. The culture was main-
as the energy source. Much work has been done in tained in a semicontinuous mode for the first 44
the outdoor mass culture of several species of days. From day 1 to day 9, ponds were run as
microalgae in order to study biomass production. duplicates (%T (540 nm) = 15%). From day 10 to
A wide range of chemicals can be expected from day 13, pond 2 was gradually diluted up to a %T
algae, namely lipids, proteins and pigments. (540 n m ) = 2 5 % (days 13-21). From day 22 to
Among these, polyunsaturated fatty acids day 27, ponds 1 and 2 were diluted to 20% and
(PUFAS), mainly eicosapentaenoic acid 20:5 033 40% of transmittance, respectively (days 28-33).
115
Biores(mrce Technolo~, 0960-8524/91/S03.50 1991 Elsevier Science Publishers I_.td, England. Printed in
Great Britain
116 V. Veloso, A. Reis, L. Gouveia, H. L. Fernandes, J. A. Empis, J. M. Novais

Growth parameters decreased from day 34 to day 1951 ). Total lipids were assayed by the Bligh and
41 in pond 1, and from day 41 to day 45 in pond Dyer (1959) procedure.
2, due to a protozoan contaminant (not identified).
On day 45, a chemical application based on Fatty acids analysis
previous works (Richmond, 1986)was success- Fatty acid methyl esters (FAME) were prepared
fully tested (addition of acetic acid to reduce the by transesterification of freeze-dried samples as
pH to 4-5 for 30 min followed by neutralization described by Lepage and Roy (1986) with the
with NH4OH until pH = 9.7) in pond 1, and pond modifications introduced by Cohen (Cohen et al.,
2 was discarded. From day 46 to day 56, the 1988b). Analyses were performed in a Hewlett-
recovery of the culture took place. On day 57, a Packard (Avondale, PA) 5880A gas-liquid
new semicontinuous culture was started from the chromatograph equipped with FID. FAME
supernatant (pond 1). At day 63, two flow analyses were carried out with a 1/8 in x 2 m
deflectors were installed on pond 3, which was stainless steel column packed with 20% DEGS on
inocculated from pond 1. From day 72 to day 74, 80/100 chromosorb W held isothermally (178C).
CO 2 was added to both ponds in order to adjust Injector and flame ionization detector were held
pH values to the range 7.0-8-0. Finally, cultures at 230 and 250C, respectively. Carrier gas was
were operated in batch mode (no addition of flesh N 2 (20 cm 3 rain l). Fatty acid identification was
medium) for the last 10 days to test starvation done by comparison of retention times with
effects. known standards (SIGMA) and cod liver oil
(Ackman & Burgher, 1964).
Harvesting Thin-layer chromatography
The removal of algae from the effluent collected Thin-layer chromatography was performed on
daily from ponds was performed with vigorous commercial 2 0 x 2 0 cmX0-25 mm silica gel
stirring followed by the addition of the flocculant. plates (Merck, Darmstadt). Chromatography was
The mixture was left to settle and the supernatant carried out by the ascending method using the fol-
was discharged. Wet biomass was filtered and lowing solvent mixtures:
dried at different temperatures.
(a) chloroform/acetone/methanol/acetic acid/
water ( 100:40: 20 : 20 : 10 v/v/v/v/v) (Kates,
Analytical methods 1972).
To determine algal growth, the following para- (b) chloroform/methanol/water (65:25:4 v/v/
meters were measured: optical density (540 nm), v) (Negishi etal., 1964).
ash-free dry weight (AFDW) and chlorophyll by Spots were visualized by exposure to iodine
methanol extraction (4 min at 70C). A correla- vapours (total lipids), UV light (254 and 356 nm)
tion between AFDW and optical density was and a naphthol/H2SO 4 solution (Kates, 1972)
obtained: A F D W (g/litre) = (0-78 _+0.10) OD. (glycolipids). Identification of spots was done by
A F D W was measured by placing sintered glass co-elution with known standards (Sigma, St.
fiber filters during 1 h at 550C. Chlorophyll a Louis, Mo).
determination usually involves rather tedious
extraction procedures and the use of multi-
RESULTS AND DISCUSSION
parameter equations which correlate observations
at various wavelengths. A simple method, similar
Figure 1 depicts the history of the 84-day growth
to the one used by Vonshak, was found by us to
experiments in ponds 1, 2 and later 1, 3 (see text).
predict [Chl,] with the empirically derived equa-
As can be seen, pond 2 was diluted more than
tion:
pond 1 at day 13, and resisted predation better
[Chla] (mg/litre)= 10"80D (665 nm) than the culture in pond 1, as would be expected.
After treatment, the culture from pond 1 was
which can be applied to Phaeodactylum tricor- recovered and used as inocuhim in pond 3.
nutum. Nitrate and phosphate concentrations in Cultures in ponds 1 and 3 from day 63 to day
the medium were analysed by the Cawse (1967) 84, should not differ except from the inclusion of
and the perchloric acid method (Standard hemicylindrical baffles at both ends of pond 3 (at
Methods, 1976), respectively. Protein was half width). It is apparent from Fig. 1 that a culture
measured by the Lowry method (Lowry et al., in pond 3 did not do so well as in pond 1, and this
 Lipid production by Phaeodactylum tricornutum 117

0.St 3
i //
c~ 2.5

2
0 0.3
1.5
t~
0.2
1
,,=,
u. 0.1 0.5 o
i
O [ ~ J ~ 0
0 20 40 60 80 100 20 40 60 80 100
TIME(DAYS) TIME (DAYS)

" POND 1 + POND 2 ~ - - POND 3 POND 1 * POND 2 c~ POND 3

Fig. 1. Biomass evolution of l'haeodaco'lum tricornumm Fig. 2. Evolution of IChl,,1/AF'[)W ratios f r o m outdoors
grown outdoors. Air maximal and minimal temperature cultures o f Phaeodactylumtricornttlum.
ranges (C) were, respectively: January (days 0-28) 32-13,
10-1.5; February (days 29-56) 35-18, 10-4; March (days
57-84) 3t~-26, 15-5"5.
7-

POND
we take to mean that the extra turbulence is bene-
~ 1 ~ 2 ~_i_~ 3
ficial to algal growth, as well as the absence of
added shading effects. ')., 5
Figure 2 shows the evolution of the ratio [Chl,/
E
AFDWI for the culture depicted in Fig. 1, only 4
during days 2 0 - 4 4 and 5 7 - 8 4 . As stated in the u.I

introduction, we found this ratio to be extremely nr" 3

I'--
sensitivc to the health of the culture, its value :2)

diminishing both by lower chlorophyll and for 2

higher organic charge. T h e value of the [(Chl,/ O

A F D W ratio was found to be a simple and 1

characteristic indicator of culture state. A value of
less than 1% means impending population crash, 0
1 2 3 4 5 6 7 8 9 10 11 12
be it due to a predator (e.g. days 3 5 - 4 5 ) or to a
lack of nutrients (e.g. days 78-84), and the inclu- WEEK
sion of CO~ in the nutrient mixture immediately Fig. 3. Productivities (AFDW basis l versus time of experi-
ments.
shows an increase: of this ratio (e.g. day 7 2 - 7 4 )
(Fig. 2). Weekly productivity can be seen in F'ig. 3
to be fairly constant at 3 - 6 g m -2 day-~, and a r . . . . 7
4
mean value of 4 g m - : day-~ was found, after
excluding the effect of the p r o t o z o a n contaminant 3.5 ~ --~-~! ~ _ ~<~ -~ _ _
which is obvious during weeks 6, 7 and 8. T h e ! [ i i~ .....

influence of cell concentration u p o n the various

"1:) 2.5 ' , I ~
lipidic fractions can be summarized as in Fig. 4,
where daily p r o d u c t i o n is shown as a function of
uJ
AFDW. and a c o m p r o m i s e value of A F D W of
()" 18 g litre f is apparent.
Chitosan addition can be seen (Fig. 5) to have
D 0.5
had an adverse effect at relatively low Ca(OH): 0

concentration and was not tried at higher ones. 0

o,14 o.18 0.22 0.29
This effect was not found in other Chrysophyceae
and Chlorophyta algae, for which chitosan/
AFDW (g litre'1)
Ca(OH.,, was a very efficient flocculant and can COMPOSITION
m LIPID but FA ~ FA non EPA ~ EPA
possibly be attributed to the nature of the diatoms PROTEIN ~ OTHER
frustule.
Fig. 4 . Effect of cell concentration on chemical composi-
Inclusion of baffles proves to be disadvantage- tions (with special emphasis on lipidic fraction) and output
ous to the growth of the culture, both as measured rate IAFDW basis).
118 V. Veloso, A. Reis, L. Gouveia, H. L. Fernandes, J. A. Ernpis, J. M. Novais

100 8TAR~TION
1 0 DAYI! [ 4 DAYB ~ 10 DAY8

8oI( 40

35

30
>. 617 O~
t~
O O 25
z <:
W
>" 20
_. 40 I-
1.1.

LU CHITOSAN ..I
20 0 mg litre "~ -4-- 2.5 mg litre "~
O
5 mg litre "1 ~ - 10 mg litre "~ I--
g

O I J i J I i - ~ i i i i
14:0 : 16:1 16:2 18:0 18;1 18:2 18:4 2 0 : 3 2 0 : 5 22:3 22:6
0 50 100 150 200 250 300 350 400 450 500
FATTY ACID
Ca(OH)2CONCENTRATION (rag litre-~) Fig. 7. Effect of starvation on fatty acid profile.
Fig. 5. Effect of Ca(OH)2 and/or chitosan addition on
Phaeodactylum tricormaum harvesting efficiency.

dactylum tricornutum is a promising candidate for

an outdoor culture in a temperate to cool climate,
even if silica is an important part of the total
40 ~ . m l n ~ AVERAGE ~.MAX
weight.
35 For this outdoor culture we developed a semi-
CO
a automatic computer monitoring system, which
30 will later be expanded to perform as a means of
<
>- culture control.
I- 25

20
,.J ACKNOWLEDGEMENTS
6- 15
O
Work was partially supported by Sociedade
10
Nacional de Sab6es Ltd, grants n FCI
5 i 310.86.258 from JNICT and n 3.3/P198 from
FLAD. Special thanks are due to Mrs Ana Maria
0 Ferrfio (LNETI/DCEAI) for the help with the
14:0 16:0 1(5:1 16:2 18:0 18:1 1 8 : 2 1 8 : 4 2 0 : 3 2 0 : 5 2 2 : 3 2 2 : 6
FATTY ACID analysis techniques and to Mr Irineu Batista
Fig. 6. Main fatty acid profile during experiment. (INIP) for supplying chitosan samples.

by A F D W and by the ratio [(Chla)/AFDW]. EPA REFERENCES

is the predominant fatty acid, and is produced at a
rate of 0-15 g m - 2 day- 1. The fatty acid mixture is Ackman, R. G. & Burgher, R. D. (1964). Cod liver oil: com-
ponent fatty acid as determined by gas-liquid chroma-
very rich in UFA (unsaturated fatty acid) (Fig. 6), tography. J. Fish. Res. Bd. Canada, 21 (2), 319-26.
even after starvation (Fig. 7). Lipids of Phaeod- Bligh, E. G. & Dyer, W. J. (1959). A rapid method of total
actylum tricornutum, as shown by TLC, consisted lipid extraction and purification. Can. J. Biochem.
Physiol., 37, 911-7.
of monogalactosyldiglyceride, digalactosyldigly- Cawse, E A. (1967). The determination of nitrate in soil
ceride, sulfoquinovosyldiglyceride, phosphatidyl- solution by ultraviolet spectrophotometry. Analyst, 92,
glyceride, phosphatidylcholine, phosphatidyl- 311.
Cohen, Z., Vonshak, A., Boussiba, S. & Richmond, A.
inositol, phosphatidylethanolamine, phosphatidic (1988a). The effect of temperature and cell concentration
acid, triglyceride and minor unidentified ones. As on the fatty acid composition of outdoor cultures of Porp-
a concluding remark it can be stated that Phaeo- hyridium cruentum. In Algal Biotechnology, ed. T. Stadler,
 Lipid production by P h a e o d a c t y l u m tricornutum 1 19

J. Mollion, M.-C. Verdus, Y. Karamanos, H. Morvan & D. Negishi, T., McKillican, M. E. & Lepage, M. (1964). The use
Christiaen. Elsevier Applied Science Publishers, London, of Saren wrap protect chromatoplates during their
pp. 412-9. exposure to iodine vapours. J. Lipid Res., 5 (3), 486.
Cohen, Z., Vonshak, A. & Richmond, A. (1988b). Effect of Richmond, A, (1986). Outdoor Mass Cultures of Microalgae.
environmental conditions on fatty acid composition of the In Handbook of Microalgae Mass Culture. ed. A.
red alga l'otphyridiurn cruenturn: correlation to growth Richmond. CRC Press, Boca Raton, pp. 312-29.
rate. J. Phycol., 24,328-32. SERI (1986). Microalgae Culture Collection 1~)85-1~)8~5,
Kates, M. (1972). Techniques in Lipidology. North Holland/ Solar Energy Research Institute, Golden, Colorado.
American Elsevier Publishing Co. Inc., Amsterdam/New Seto. A., Wang, H. C. & Hessehine, C. W. (1984). Culture
York. conditions affect eicosapentaenoic acid content of
Lepage, G. & Roy, C. R. (1986). Direct transesterification of Chlorella minutissima. JAOCS, 61 (5), 892-4.
all classes of lipids in a one-step reaction. J. Lipid Res., 27, Standard Methods (1976). Standard Methods for the t:xami-
114-9. nation of Water and Wastewater. American Public Health
Lowry, O. H., Rosebrough, N. H., Farr, A. L. & Randall, R. J. Association, American Water Works Association, Spring-
( 1951). Protein measurements with folinphenol reagent. J. field, pp. 448-52.
Biol. ('hem., 193,265.