LIFE CYLE AND SEXUALITY OF THE FRESHWATER RAPHIDOPHYTE
GONYOSTOMUM SEMEN (RAPHIDOPHYCEAE)1
Rosa Isabel Figueroa
Instituto Oceanogra ´fico Vigo, Cabo Estai-Canido, 36200 Vigo, Spain
Limnology, Department of Ecology, Lund University, Ecology Building, 22362 Lund, Sweden
Previously unknown aspects in the life cycle of
the freshwater flagellate Gonyostomum semen (Eh-
renb.) (Raphidophyceae) are described here. This
species forms intense blooms in many northern
temperate lakes, and has increased in abundance
and frequency in northern Europe during the past
decades. The proposed life cycle is based on obser-
vations of life cycle stages and transitions in cul-
individually isolated and monitored by time-lapse
photography. The most common processes under-
taken by the isolated cells were: division, fusion
followed by division, asexual cyst formation, and
sexual cyst formation. Motile cells divided by two
different processes. One lasted between 6 and 24h
and formed two cells with vegetative cell size and
with or without the same shape. The second division
process lasted between 10 and 20min and formed
two identical cells, half the size of the mother cell.
Planozygotes formed by the fusion of hologametes
subsequently underwent division into two cells.
Asexual cyst-like stages were spherical, devoid
of a thick wall and red spot, and germinated in
24–48h. Heterogamete pairs were isogamous, and
formed an angle of 0–901 between each other. Plano-
zygote and sexual cyst formation were identified
within strains established from one vegetative cell.
The identity of these strains, which was studied by
an amplified fragment length polymorphism analy-
sis, was correlated with the viability of the planozy-
gote. Resting cyst germination was described using
cysts collected in the field. The size and morphol-
ogy of these cysts were comparable with those
formed sexually in culture. The excystment rate
was higher at 241 C than at 19 or 161 C, although the
cell liberated during germination (germling) was
only viable at 161 C. The placement of G. semen
within the Raphidophyceae family was confirmed
by sequence analysis of a segment of the 18S rib-
Key index words:
nyostomum semen life cycle; raphydophyceans; sex-
AFLP; cysts; encystment; Go-
Abbreviation: AFLP, amplified fragment length po-
Gonyostomum semen is the most common freshwater
member of the raphidophyceans, and like several of its
marine counterparts it is considered a nuisance alga. G.
semen often forms intensive blooms (Pithart et al. 1997,
Wille ´n 2003) and may dominate the phytoplankton
community by as much as 98% for extended periods
(Le Cohu et al. 1989). This alga adversely affects lakes
used for recreation, as it discharges mucilaginous
strands upon contact, thereby covering bathers with a
slimy layer causing itching and other allergic reactions
(Cronberg et al. 1988). The distribution of G. semen is
widespread, and has been described in various loca-
tions in Europe, Asia, Africa, North America, and South
America (Eloranta and Ra ¨ike 1995). In the last decades,
there has been an increase in the abundance and oc-
currence of this flagellate in Nordic countries including
Sweden, Finland, and Norway (Hongve et al. 1987,
Cronberg et al. 1988, Lepisto ¨ et al. 1994), while before
the 1980s, the occurrence of these blooms was rare.
G. semen is most common in brown-water lakes, i.e.
lakes with moderate to high humic content (50–60mg
Pt?L?1or approximately 10mg DOC?L?1) and slight
acidity, which are typically found in forested areas
(Cronberg et al. 1988). However, principal component
analyses show no direct connection with pH, but rather
with humic content and nutrients, especially phospho-
rus (Cronberg et al. 1988, Eloranta and Ra ¨ike 1995).
Although peat processing and general eutrophication
of forest lakes have been suggested as a probable cause
(Eloranta and Ra ¨ike 1995), the recent spreading of this
species remains largely unexplained.
G. semen can form extremely dense blooms, reach-
ing a biomass of over 1800mg chl a?L?1(Pithart et al.
1997). However, an explanation behind the bloom
formation and competitive advantage of G. semen
over other algae is lacking. Vertical migration has
been suggested as one important adaptation allowing
1Received 30 August 2005. Accepted 26 April 2006.
2Author for correspondence: e-mail email@example.com.
J. Phycol. 42, 859–871 (2006)
r 2006 by the Phycological Society of America
G. semen to maximize both nutrient uptake and
photosynthesis (Eloranta and Ra ¨ike 1995). Another
important characteristic of G. semen is that it has an
overwintering benthic stage or resting cyst (Drouet and
Cohen 1935). The resting cyst is considered as an ad-
aptation to survive during unfavorable environmental
conditions (Fryxell 1983), including predation (Hans-
son 1996, Hansson 2000). A life cycle with an alterna-
tion between a benthic and a planktonic stage is a trait
that G. semen has in common with a number of other
freshwater and marine harmful algal bloom formers,
including other raphidophytes, dinoflagellates, and
cyanobacteria. In these species, the cyst stage has
been shown to play an important role in bloom initi-
ation (Anderson and Wall 1978) and seasonal succes-
sion (Anderson and Rengefors 2006). Moreover, in the
dinoflagellates, resting cysts are known to form follow-
ing sexual reproduction, making the cyst stage impor-
tant in connection to gene recombination and genetic
diversity of populations (Pfiester and Anderson 1987).
Despite the importance of the life cycle in the ecol-
ogy of phytoplankton, it has to date been poorly in-
vestigated in G. semen. Previous works have reported
aspects of asexual reproduction (Drouet and Cohen
1935) and of isolated steps of the sexual cycle (Cronb-
erg 2005). Nevertheless, a detailed description has
never been given of sexual reproduction, in which sin-
gle cells are followed through the process, nor has
there been a study of mating type in cultures, two as-
pects of great importance in phytoplankton life cycles.
Marine raphidiophytes, a small taxon whose members
are almost all classified as ‘‘harmful’’ species, have in con-
trast been studied in more detail. Despite previous re-
ports of sexual cyst formation (Nakamura et al. 1990) in
these marine relatives, the cyst stage was later claimed to
be haploid (Yamaguchi and Imai 1994, Itakura et al.
1996). Chattonella antiqua (Hada) and Chattonella marina
(Subrahmanyan) Hara et Chihara have diplontic life his-
tories, in which vegetative cells are diploid and cysts are
haploid (Yamaguchi and Imai 1994). A similar life cycle is
also suggested for Heterosigma akasiwo (Hada) Hada (Imai
et al. 1993a), which, like G. semen (Salonen and Rosenb-
erg 2000), exhibits diel vertical migrations (Watanabe
et al.1983,YamochiandAbe1984,Nagasakiet al.1996).
Here, we investigated the life cycle of G. semen by
following the development and fate of the morpholog-
ical types that were commonly observed in culture. The
individual monitoring of these cells allowed us toreport
undescribed aspects in the life cycle of raphidiophytes,
such as two kinds of vegetative division, fusion of holo-
and heterogametes, and both asexual and sexual cyst
formation. Part of the small subunit (SSU) of the rib-
osomal DNA was sequenced in order to provide a mo-
lecular identification of G. semen, and to determine its
phylogenetic relationship with other raphidophytes.
MATERIAL AND METHODS
Cultures. In these experiments, cultures of G. semen were
established through the isolation of several individual motile
cells from a water sample of Lake Dagstorpsjo ¨n in August
2002. The cultures were grown in modified WC culture me-
dium (MWC) (Guillard and Lorenzen 1972) and maintained
at 151 C at 30mmol photons?m?2?s?1on a 12:12 light:dark
(L:D) photo-cycle. Two strains from single-cell isolations,
named G2 and G3, were established by the isolation of indi-
vidual cells from this culture.
Life cycle study. Observations and size determination were
carried out on vegetative cells, cysts formed in laboratory,
and natural cysts. For monitoring of individual cells, different
life cycle stages of G. semen (Table 1) were isolated from six
different treatments: (i) MWC-replete medium, (ii) medium
without nitrate (N) added, (iii) medium without phosphate
(P) added, (iv) medium without N nor P added, (v) medium
without N and placed 72h in darkness, and (vi) medium
without P and placed 72h in darkness. Both cultures were
exposed to each of the treatments. Duplicate sterile polysty-
rene petri dishes (Iwaki, Tokyo, Japan, 35mm diameter)
were inoculated with exponentially growing cells (4000–
6000cells?mL?1) to a final concentration of 700cells?mL?1
in a total volume of 10mL of each treatment medium. Fifty
cells or pairs of cells with different morphology were indi-
vidually isolated, photographed, and separately transferred
to wells of tissue culture plates (Iwaki, 6.4mm diameter)
filled with MWC medium. Culture plates were incubated un-
der the same conditions as those previously described for
culture maintenance. The isolated cells were checked at least
once daily and were photographed with a CANON EOS dig-
ital camera. They were also measured at ?630 magnification
TABLE1. Morphologies most commonly observed in clonal
cultures of Gonyostomum semen after 5 days of incubation in
Morphology and numbers
of cells isolated from each
Division type 1
Division type 2
Fusing pairs of hologametes
Fusing pairs of heterogametes
aClassification based on results from this study.
MWC, modified WC culture medium.
ROSA I. FIGUEROA AND KARIN RENGEFORS
using an inverted microscope (Axiovert Zeiss 135, Ober-
kochen, Germany) and an Image IPplus analyzer (Media
cybernetics, Berkshire, UK).
Cyst germination study. For natural cyst observations, sed-
iment samples were collected from Lake Bokesjo ¨n, Southern
Sweden, in May 2004, using a gravity core sampler. Small
particles were removed with a 20mm sieve and the remaining
sediment was stored in airtight plastic jars in darkness at
41 C. Twenty to 25 cysts were individually isolated to sterile
polystyrene petri dishes (Iwaki, 6.4mm diameter), and sub-
sequently incubated at three different temperatures (16, 19,
and 241 C), and checked for excystment every 24h for 15
days. The excystment was defined as the complete emer-
gence of the protoplast from the cyst even if the germling
remained non-motile (Anderson and Wall 1978). The num-
ber of germinated cysts divided by the total number of cysts
was used to determine germination success. The develop-
ment of the germinated cells was checked at least daily, and
the cells were photographed and measured at ? 630 as
Nuclear staining. Individual cells were fixed for 10–15min
in 2% glutaraldehyde in 0.01M PBS buffer, pH 7.4 (Sigma,
St. Louis, MO, USA). Subsequently, they were washed with
several drops of PBS buffer, stained with 10mM Hoechst
33342 (Sigma) or 1:100 Sybr green (Molecular Probes, New
Brunswick, NJ, USA) in 0.01M PBS pH 7.4 for 30min,
washed again, and observed with an epifluorescence micro-
scope (LEICA DMLA, Wetzlar, Germany) at 497nm. Photo-
graphs were taken using a digital camera (CANON EOS,
Genetic analysis. DNA fingerprint analysis was performed
on the G2 and G3 strains in order to check whether
phenotypic differences were reflected in the respective gen-
otypic profiles. Amplified fragment length polymorphism
(AFLP) was utilized to distinguish the strains, as this tech-
nique has been successfully implemented to detect clonal
differences in dinoflagellates (Figueroa et al. 2006).
DNA extraction. Cells in the exponential growth phase
from the G2 and G3 strains were harvested by centrifuga-
tion (3000g for 10min at 41 C) and the total DNA was
extracted, according to Bolch et al. (1999), and stored
at ?801 C.
AFLPanalysis. Modified protocols were followed for AFLP
analyses based on Vos et al. (1995). Ten microliters of
the extracted DNA (10ng/mL) from each sample was first
digested with 2.5 units of EcoRI (Amersham Pharmacia
Biotech, Piscataway, NJ, USA) and TruI (Fermentas, Vilnius,
Lithuania) in a total volume of 20mL for 1h at 371 C. Ligation
to the ends of the DNA fragments was made by adding T-4
ligase (USB corporation, Cleveland, OH, USA) (0.5U per
sample) and adaptors, at a concentration of 0.01mM for the
E-adaptor and at 0.1mM for the M-adaptor and incubating
3h at 371 C. Duplicate DNA templates were made for all cul-
tures. The preamplification reaction was performed using a
DNA Thermal Cycler (Perkin Elmer Applied Biosystems
9600, Foster City, CA, USA), and carried out using 20 cycles
(941 C, 30s; 561 C, 30s; 721 C, 60s). Following the pream-
plification step, the product was diluted (10?) with water and
2.5mL was used for selective amplification. Six primer com-
binations were used: 1: EcoRI-TCG and MseI-CGC; 2: EcoRI-
TCT and MseI-CGA; 3: EcoRI-CGG and MseI-CGG; 4:
EcoRI-TGA and MseI:CGG; 5: EcoRI-TCG and MseI-CGA;
and 6: EcoRI-TGA and MseI-CGA. The reaction mix con-
tained 10mL preamplified product, 1.8mL of water, 0.4U of
Taq DNA polymerase, 4mL 1mM dNTPs, 0.06mL E-primer
(100mM), 0.06mL M-primer (100mM), 2mL MgCl2(25mM),
and 2mL PCR buffer (10?). Amplification by touchdown
PCR was performed with an initial denaturation at 941 C
for 2min and a first cycle at 941 C for 30s, 651 C for 30s, and
721 C for 60s. During the next 12 cycles, the annealing tem-
perature was reduced by 0.71 C per cycle down to 561 C,
whereas the last 23 cycles were the same as described for
The selective amplification was stopped by adding 10mL of
formamide dye (100% formamide, 10mM EDTA, 0.1% xylene
cyanol ff, 0.1% bromophenol blue) to the samples that were
stored at þ41 C overnight before running on the gel. After
3min denaturation at 951 C, 3.5mL was loaded onto a 6%
polyacrylamide gel. The fragments were separated at 30 W
for 1–2h, and detected by fluorescein-labeled E-primers in a
FlourImager (Vistra Fluorescens, Molecular Dynamics Inc.,
Sunnyvale, CA, USA). The results were stored as TIFF files for
Phylogenetic analyses. Strain GSBO2 isolated from a single
cell collected in Lake Bokesjo ¨n, Southern Sweden, was used.
DNAwas extractedas for AFLP. The SSU of ribosomal DNAwas
amplified and sequenced using two different universal primers:
4616 (forward) 50-AACCTGGTTGATCCTGCCAG-30and 4618
reactions were set up in 25mL volumes, with 25ng DNA tem-
plate, 0.5U of Taq DNA polymerase (AmpliTaq, Applieds
Biosystems, Stockholm, Sweden), 3mL MgCl2 (25mM),
1.25mL dNTP (10mM, Applied Biosystems), and 1mL of
each primer. The following amplification protocol was fol-
lowed: 5min at 941 C, 35 cycles at 941 C for 1min, 551 C for
5min, 721 C for 2min, and a final extension step at 721 C for
7min. The PCR product was cleaned with the PCR-M Clean
Up System (Viogene, Sunnyvale, CA, USA) and sequenced
using the BigDye Terminatior v1.1 Cycle Sequencing Ready
Reaction Kit (Applied Biosystems) using the primers 4616F,
4618R, Euk29F, and Euk517R (Coyne et al. 2005). The partial
SSU sequence was submitted to GenBank (accession no.
For the phylogenetic analyses, the sequences were edited
with BioEdit (v7.04.1, Hall 1999) and aligned using MEGA3
(Kumar et al. 2004). Additional raphidophyte and outgroup
sequences were downloaded from GenBank for phylogeny
constructions. The phylogeny was estimated using Neighbor-
Joining as implemented in MEGA3 under the Tajima–Nei
substitution model with a g-distributed rate of variation across
sites. Branch support for the phylogram was estimated by 1000
Based on our results, we have proposed a life cycle
for G. semen, which is summarized in Figure 1. The
physiological development of each depicted process is
based on the individual isolation and monitoring of a
total of more than 50 cells.
Vegetative cycle. Vegetative cells of G. semen with
normal morphology and size (Table 2) have two
flagella attached at an apical pit, with one directed
forward and one backward when swimming (Fig. 2a).
These cells had a central nucleus that occupied
approximately 1/7 of the cell volume (Fig. 2b).
Vegetative cells in culture underwent three main
processes: division, fusion, or asexual cyst formation.
Division occurred by two different mechanisms. Divi-
sion type 1 (Figs. 1 and 3) lasted between 6 and 24h,
and involved the elongation of the dividing cell
proceeding until two cells with vegetative size were
formed. The newly formed cells had the same or dif-
ferent lengths. This process took place in two planes,
with the formation of a lobe on one side of the caudus
LIFE CYLE AND SEXUALITY OF GONYOSTOMUM SEMEN
at an early stage of the division (Fig. 3, a and b). This
lobe later formed the two ‘‘tails’’ of the cells. In this
example, the cells formed had a different size and re-
mained attached by a thin cytoplasmic bridge during
several hours before splitting (Fig. 3c). At other times,
the process proceeded in the same plane (Fig. 3, d–f).
In these cases, cell enlargement (without or with sym-
metry, Fig. 3, d and e, respectively) was not followed by
the formation of the two cell tails, which only appeared
at the end of the division process (Fig. 3f). The division
of the nucleus may have occurred in very early stages
of division, since cells in initial stages of division
(Fig. 3g) already had two distinct nuclei (Fig. 3h). In
addition, some large (63–78mm in length) vegetative
cells in culture contained two nuclei (Fig. 3i).
In division type 2 (Figs. 1 and 4, a–f) cell constric-
tion was evident at the posterior and anterior part of
the cell (Fig. 4a). Cells with a morphology similar to
Figure 4a had one large nucleus that occupied most of
the cell, and in which short and thick chromosomes
(30–50) were observed (Fig. 4b). The mid-furrow
(Fig. 4c) and the nuclear division (Fig. 4d) progressed
until the formation of two identical cells half the size of
the mother cell (Fig. 4, e–f). The main characteristic
FIG. 1. Life cycle of Gonyostomum semen in culture and DNA contents assuming that vegetative cells are 2C in the G1 phase (see text
for explanation). The scissors indicate the steps not allowed in the clonal culture G3.
TABLE2. Cell dimensions and distinctive features of the different Gonyostomum semen life cycle stages in culture.
Stage Dimensions Distinctive featuresDuration
Vegetative cell 36–92mm long No fixed shape, yellowish-green in color, with
two flagella in an apical pit
Old (1–3 days):
Fusing pairs Isogamy. 0–901 angle of coupling10min to 3h
PlanozygoteFlat apex. Could present a red spot 48–72h
Double wall, reservoir granules, one or two
Vegetative morphology. The red spot is left in
the cyst shell.
28–38mm Temporary cystAlmost spherical, dark color, and no granules or
ROSA I. FIGUEROA AND KARIN RENGEFORS
that differentiated this process from the division type 1
was that there was no cell enlargement, and conse-
quently, the cells formed were always identical to each
other and half the size of the mother cell. Another im-
portant difference was the nuclear development,
where nuclear division occurred before cell enlarge-
ment in type 1, whereas nuclear and cytoplasmic divi-
sions were simultaneous processes in type 2. Division
type 1 was commonly observed in the morning. Type 2
division, on the other hand, was more frequent before
the onset of the dark period, and was also enhanced by
stressful conditions, such as nutrient limitation (mainly
N reduction), or long periods of darkness.
We also observed that some cells lost cytoplasmic
volume through a cytoplasmic furrow (Fig. 4g). Cells in
the early stages of this process had one nucleus located
in the constriction area (Fig. 4h), although it ended
with the formation of only one nucleated cell and a
Hologamete fusion. Cell fusion was a rather fre-
quent process in exponentially growing cultures. Fu-
sions occurred between isogamous or anisogamous
cell pairs and followed two pathways: at an angle or
longitudinal. Fusions at an angle (Fig. 5) were more
common and began with the attachment between one
cell’s apical area and the upper lateral area of the
other cell, forming an angle of 01–901 between the
cells (Fig. 5a). At this stage, nuclei were approaching
each other (Fig. 5b). The fusion progressed as shown
in Figure 5, c–e, and a larger and longer cell than the
originals was formed (Fig. 5f). After the complete
fusion, only two flagella were observed in these
zygotes. This cell contained one nucleus of 22–26mm
FIG. 2. Vegetative cell of Gonyostomum semen. Scale bars,
10mm. (a) Vegetative cell with the two flagella pointing in oppo-
site direction. (b) Nucleus of a vegetative cell after staining with
FIG. 3. Time-lapse photogra-
phy of cells in vegetative division
type 1. Scale bars, 10mm. (a)
First division stage showing the
formation of a lobe (arrow) be-
tween the emerging cells. (b) Di-
vision stage in (a) after 24h
development. (c) Final stage of
the division and formation of
two cells of different sizes. The
arrow indicates the cytoplasmic
bridge that is the last nexus be-
tween both cells. (d) Asymmetric
division type 1 without lobe for-
mation. (e) Symmetric division
type 1 without lobe formation.
(f) Final stage of the division (e)
and formation of two cells with
the same size. (g) Culture cell in
division type 1. (h) Nuclear
staining of the cell in (g). (i) Cul-
ture vegetative cell showing the
presence of two nuclei.
LIFE CYLE AND SEXUALITY OF GONYOSTOMUM SEMEN
diameter (Fig. 5g), and the subsequent division result-
ed in two cells with similar morphology (Fig. 5h).
Longitudinal fusion began by the attachment of
both lateral sides of the cells, which progressively fused
(Fig. 6, a–c). After 24h, the cell formed by this fusion
process (Fig. 6d) divided (Fig. 6e), and formed two
similar cells with smaller size and a marked constriction
between the main body and the tail area (Fig. 6f).
These cells were viable and progressively acquired
a normal vegetative morphology. Fusing stages were
FIG. 4. Time-lapse photography of a vegetative divisions type 2. Scale bars, 10mm. (a) Initial stage of division and formation of
cytoplasmic constrictions on both sides of the cell (arrow heads). (b) Nuclear staining of a cell in similar stage of division than (a).
(c) Progression of the division (a) and formation of a mid-cytoplasmic furrow. (d) Nuclear division in a cell with a morphology similar
to (c). (e–f) Final stages of division (a). (g–h) Culture cell undergoing cystoplasmic reduction by the formation of a cytoplasmic furrow
(arrow) and the presence of one nucleus in the constricted area (arrows) in a early reduction stage (h).
FIG. 5. Time-lapse photography of fusion with angle of hologametes. Scale bars, 10mm. (a) Initial stage of fusion. (b) Nuclear
arrangement in a cell with similar morphology than (Fig. 6a). (c–f) Progression of the fusion (a) and new cell formation. (g) Nuclear
staining of a cell formed by a fusion with angle. (h) One of the two cells formed by the division of the cell (f).
ROSA I. FIGUEROA AND KARIN RENGEFORS
always viable in the G2 strain but never in the G3
strain. This pattern of viability was observed in all
studied cases, which involved the monitoring of at least
10 cell pairs in the process of fusion from each strain.
Intermediate morphologies of the fusing process as
shown in these time-lapse photographs were routinely
observed in cultures.
Asexual cyst formation. In stressed cultures, the
formation of asexual cyst-like stages was observed.
Under long periods of darkness and/or N and/or P
depletion, vegetative cells acquired a circular shape
and settled on the bottom of the culture plates
(Fig. 7a). These cysts were distinct from sexual cysts
(see paragraph below), because they were larger, and
lacked a double wall and red spot. Furthermore, an
empty wall was not left after excystment (Fig. 7b),
which occured 24–72h after their isolation in fresh
MWC medium. Figure 7c depicts a massive tempo-
rary cyst formation at the bottom of a culture previ-
ously exposed to N-limited medium and 72h of
Heterogamete fusion, planozygote, and resting cyst for-
mation. Fusing gamete pairs were observed in old
cultures and in medium with N?or P?depletion. As
seen with the fusion of vegetative cells, the zygotes
were unviable in clone G3, but encysted successfully
in strain G2. Heterogametes were isogamous (both
gametes had the same size), and paler and smaller
than vegetative cells (Table 2), and sometimes con-
tained a reddish brown vacuole. Fusing gamete pairs
formed an angle of 01–901 between each other. Fig-
ure 8a shows a fusing pair with a 01 angle between the
gametes. During the fusing process (6–12h), the cells
first attached the apices, which led to the formation of
typical two-tailed cells with a flat apex (Fig. 8b). Two-
tailed cells can be considered syncytiums, as they
contained two nuclei (Fig. 8c) and four flagella aris-
ing from the apical pit (Fig. 8d). In Figure 8e, a fus-
ing pair with a 901 angle between the gametes is
shown. Although the fusing process was different
from that in the previous case (compare Fig. 8f with
8, a and b), it eventually led to the formation of a two-
tailed cell (Fig. 8g). Later, the tails fused (Fig. 8h) and
zygote formation was completed (Fig. 8i). One-tailed
zygotes contained one elongated nucleus in the
central position (Fig. 8j) and only two flagella. Old
planozygotes usually contained one brown-reddish
spot (Fig. 8k), and the nucleus was positioned at the
FIG. 6. Time-lapse photogra-
phy of an equatorial fusion of
hologametes. Scale bars, 10mm.
(a) Initial stage of fusion. (b–d)
Progression of the fusion and
new cell formation. (e) Cell in di-
vision type 2. (f) One of the two
cells formed by the division (e).
FIG. 7. Formation
cystment of temporary cysts.
Scale bars, 10mm. (a) Temporary
cyst. (b) Germling from tempo-
rary cyst. (c) Temporary cysts in
the bottom of a culture.
LIFE CYLE AND SEXUALITY OF GONYOSTOMUM SEMEN
tail of the cell (Fig. 8l). Between 2 and 4 days from the
isolation of fusing pairs, double walls and round cysts
(29–39mm diameter) were formed in clone G2 (Fig.
8m). However, these cysts degraded 6–8 days after
their formation in the culture plates of gamete pair
isolation. No indications existed that planozygotes
could omit the resting stage and undergo division
directly. Sexual cysts contained one nucleus (9.5–
13mm in diameter, Fig. 8, n and o) located toward
one end of the cyst. In some cysts, the nucleus was
elongated and located in a more central position (Fig.
8p). Some culture cysts had two red spots (Fig. 8q).
This feature was explained by the existence of red
spots in both fusing gametes and often, although not
in all cases, co-occurred with the presence of two nu-
clei (Fig. 8r). In these cysts, the distance between the
two nuclei was variable. Some of them were adjacent
and almost fused (Fig. 8s) and were localized at a
more central position (Fig. 8t). In the followed cases,
fusing gametes with none or only one red spot
formed one-spot cysts.
Resting cyst germination. To describe the morphol-
ogy of the germling from resting cysts, we studied the
germination of natural cysts. The size and morphol-
ogy of those cysts were comparable with those ob-
tained in our cultures (Table 3). Cysts were round,
green, and with a thick wall. They typically had a
reddish-brown spot, although two spots were also
observed (Fig. 9, a and b) as in the culture cysts. Cysts
germinated in greater proportions at high tempera-
tures, but the germlings were only highly viable at
161 C. At 241 C, the germination percentage reached
61%, but no single viable germling was observed. At
191 C, 43% of the cysts germinated, although germ-
ling viability remained low (11%). Only 36% of the
cysts germinated at 161 C, but the post-excystment
viability was as high as 60%.
Excystment began by the appearance of a small hole
in the double wall of the cyst (Fig. 9c). Later, the germ-
ling exited through this opening, always leaving be-
hind the red spot in the empty cyst (Fig. 9d). Zero to
24h after excystment, the germling had attained veg-
etative cell morphology (Table 2 and Fig. 9e). On one
occasion, we observed the germination of two cells.
Newly excysted cells had one nucleus with the same
characteristics as observed in vegetative stages (Fig. 9f).
The germling divided into two cells after 72–120h.
Genetic analyses. To determine whether G. semen
strain identity impacted the efficiency of the life-cycle
processes studied, an AFLP analysis was carried out
on the cultures G2 and G3. In this analysis, 3 of the 6
primers combinations used showed clear differences
between the strains (primer combinations 2, 3, and 6,
Fig. 10). As several differences were observed with
the selective amplification 6, only the most obvious
ones are indicated by arrows in Figure 10.
The partial SSU (436bp) phylogeny clustered G. se-
men with the other freshwater and marine rap-
hidophytes (Fig. 11). This clade had high bootstrap
support (94%). G. semen was most closely related to
Vacuolaria virescens, the only other freshwater species
in the analyses (100% bootstrap).
Knowledge of the life cycle stages and processes is
essential to understanding the ecology and environ-
mental role of a species in a certain area (Litaker et al.
2002). Here, we present detailed photographic evi-
dence of cell morphology and transitions between life
stages of the full life cycle of the nuisance alga G. semen.
Our main results can be summarized as follows: G.
semen ordinarily multiplies with a longitudinal cell di-
vision (division type 1). Close to onset of darkness, and
mainly under stressful conditions, a depauperating di-
vision (division type 2) also takes place. G. semen has a
sexual cycle, which is induced by N?or P?depletion.
Under these conditions, small, pale gametes are
formed, which fuse in pairs and form a two-tailed cell
that becomes rounded, acquires one to two brown-
reddish pigmented spots inside, and transforms into a
round resting cyst. Vegetative stages can undergo asex-
ual cyst formation and fusion, but the zygote formed
by this process divides but never encysts. Our data are
the first to our knowledge that reported step by step
the formation of resting sexual cysts in a rap-
hidophycean flagellate, from the fusion of gametes to
the zygote encystment. These results show interesting
similarities to and differences from marine members of
Asexual reproduction. Asexual reproduction of G.
semen vegetative stages occurred through two types
of pathways that we have termed as division type 1
and 2. Division type 1 was the main way of division in
exponentially growing cultures, and resulted in cells
with vegetative size and morphology. In contrast, di-
vision type 2 was a depauperating kind of division, as
two cells half the size of the mother cell were formed.
Presumably, type 2 division occurred as a response to
stress, as it was more frequent near the dark period,
after placing the cultures in darkness for 72h, or in
N?-limited medium. These results agree with previ-
ous work on marine raphidophytes showing that ir-
radiance was important for nuclear DNA replication.
For example, C. antiqua and C. marina generally di-
vide once during one period of darkness (Ono 1988),
and several divisions occurred at night in cultures of
Heterosigma akashiwo (Honjo and Tabata 1985). Previ-
ously, only division type 2 has been described for the
asexual division of this alga (Cronberg 2005). This
fact may reflect stressful conditions, given that no
acclimated cultures but rather lakes samples were
The two types of divisions described for G. semen
differed in their nuclear development. Nuclear divi-
sion occurred before cell division in division type 1,
whereas both processes were simultaneous in division
type 2. In the latter process, the chromosomic struc-
ture was clearer (Fig. 4b), which might be explained by
a smaller DNA amount. The characteristics of type 1
ROSA I. FIGUEROA AND KARIN RENGEFORS
FIG. 8. Time-lapse photography of heterogamete fusion and zygote formation in the strain G3. Scale bars, 10mm. (a) Isolated fusing
gamete pair forming a 01 angle between the gametes. (b and c) Planozygote formation and nuclear staining. (d) Different planozygote
showing four flagella arising from the apical pit (arrows). (e) Isolated fusing gamete pair forming 901 angle between the gametes. (f–i)
Planozygote formation. (j) Nuclear staining of the planozygote (i). (k and l) Planozygote with a red spot showing the nucleus near the tail
area. (m–p) One-spot cysts and nuclear staining showing the nucleus usually opposite the red spot and in peripheral position (n and o),
though on occasion almost central (p). (q) Two-spot resting cyst. (r-t). Nuclear staining of two-spot resting cysts showing different
arrangement of the nuclei.
LIFE CYLE AND SEXUALITY OF GONYOSTOMUM SEMEN
division suggest that this is the mitotic and normal type
of vegetative cell division. Type 2 division, on the other
hand, may be a haploidization process, based on cell
size, morphology, nuclear development (Fig. 1), and in
the fact that the resulting cells are smaller than vege-
tative ones and probably contain less DNA.
Sexual reproduction. In this study, we were able to
prove sexual reproduction in G. semen. We provided
the step-by-step evidence that small and pale cells
fused and that the resulting two-tailed planozygote
eventually lost motility and formed resting cysts.
Nutrient stress has been reported as a trigger for
sexuality and resting stage formation in many dino-
flagellate species (von Stosch 1973, Chapman et al.
1982, Sako et al. 1987, Kelley and Pfiester 1990). Ac-
cordingly, we observed the formation of fusing gam-
ete pairs under nutrient-deficient conditions, mainly
in medium without N. In contrast to results for many
dinoflagellate species, in G. semen, we observed that
gametes capable of resting cyst formation were dis-
tinctly different from vegetative cells. Heterogametes
were characteristically smaller, paler, and differently
shaped than vegetative cells (heterogamy) (Fig. 8, a
and e). ‘‘Small cells’’ were also identified as gametes
in C. antiqua (Nakamura et al. 1990). In H. akashiwo,
pre-encystment cells also differed in shape, size, and
color from vegetative cells (Imai 1989, 1990, Imai
et al. 1993b). In the present study, we also verified
that cells of G. semen with vegetative, rather than
gamete morphology (hologametes) underwent proc-
esses of fusion in culture. The fusion of hologametes,
is a novel finding in the study for raphidophytes. This
process differed from the fusion of heterogametes in
both the morphology of the cells involved and the
fate followed by the fused cell. This kind of fusion did
not form two-tailed zygotes, and the resulting cell
underwent division. This fact suggests that although
vegetative cells may have surface recognition mole-
cules as those in gametes, the fusion cannot produce
resting cysts. Interestingly, we never observed fusion
of holo- with heterogametes.
We documented an important phenotypic differ-
ence in the life cycle processes between the strains of
G. semen used. Fusion of vegetative-like cells and sexual
cyst formation were unviable processes in one of the
strains, while zygote division and encystment were suc-
cessfully achieved in the other one. The AFLP analysis
on the cultures G2 and G3 showed different finger-
print patterns, suggesting that there were genotypic
differences between the two strains. In Heterosigma
strains, RAPD variability was related to phenotypic
variation between natural and cultured samples (Han
et al. 2002). Based on this evidence, we inferred that
genetic factors likely affected the sexual behavior of
the G. semen strains.
Another result is that sexuality might be under the
control of at least two loci, as cell mating and cytoplas-
mic fusion were processes achieved independently of
the viability of the zygote. Consequently, only asexual
reproduction was possible in the strain G3, albeit that
cell mating and fusion processes were also observed. In
this strain, outcrossing would be necessary to complete
the cycle proposed in Fig. 1. The existence of a sterile
TABLE3. Size comparison between natural and culture cysts
of Gonyostomum semen.
Origin of cysts
Median ? standard
deviation (mm)Range (mm)
31.7 ? 1.89
33.6 ? 3.41
FIG. 9. Resting cyst germina-
tion. Scale bars, 10mm except in
(f), which is 5mm. (a–b) Resting
cysts from natural samples with
one and two red spots, respec-
tively. (c) In an initial stage of
germination, an opening (arrow)
is formed in the double wall of
the resting cyst. (d) Resting cyst
germination. (e) Newly excysted
cell (0–24h). (f) Nucleus from
the germling in Fig. 10e after
staining with Hoescht.
ROSA I. FIGUEROA AND KARIN RENGEFORS
strain is also plausible. The same sexual pattern was
observed in four other new clonal strains tested (un-
published data), which in all cases were capable of
gamete mating and cyst formation. In conclusion, our
results suggest that in most strains of G. semen, sexual
fusion and cyst formation can be completed within
strains established from a diploid, vegetative stage.
Nonetheless, the sexual cysts formed during our
experiment degraded with time. This phenomenon
was also observed by Imai (1989), who reported that
some unknown factors affected the formation of viable
cysts of C. marina in culture. As seen in our results,
encystment was successfully achieved in C. marina but
the cysts had a very low incidence of germination. One
explanation for this result might be that culture con-
ditions were not adequate for cyst maturation. C.
antiqua and C. marina complete cyst formation after
sinking to the sea bottom (Imai and Itoh 1988). For
these species, the combination of factors such as nutri-
ent depletion, adherence to solid surfaces, and low
light intensity was assumed to be essential for cyst for-
mation and maturation (Imai 1989). In fact, for mat-
uration and then germination to occur, Chattonella cysts
need a low storage temperature (below 111 C) for
more than 4 months (Imai et al. 1991).
We observed that uninucleate sexual cysts are pro-
duced in G. semen cultures, and that binucleate cysts
might be produced by a premature encystment of the
fused cell (syncytium), as is suggested by the existence
of two nuclei in different steps of fusion (Fig. 8, r–t)
that are located centrally as in the young planozygote
stage (Fig. 8i). Imai (1989) also observed that most
cysts formed in C. marina cultures were uninucleate,
although some cysts were binucleate. Imai proposed
that, as described for some chrysophycean flagellates
(Sandgren 1983), three types of cysts might be pro-
duced in cultures: uninucleate asexual, binucleate
asexual (autogamic), and binucleate sexual (zygotic).
In our experiments, however, we found no evidence of
asexual resting cyst formation.
Resting cysts of H. akashiwo, C. antiqua, C. marina,
and G. semen have been suggested to be haploid, based
on the small size of both the cyst diameter and nucleus.
This fact has been attributed to their presumed ha-
ploid DNA content, which was similar to that observed
in small cells formed in N-limited medium and more
or less half in content of that observed in normal veg-
etative cells (Imai et al. 1993a). The sexually formed
cysts of G. semen in our cultures were also smaller than
the vegetative cells. However, the cell size was not re-
lated to an asexual encystment, but was due to the
FIG. 10. Amplified fragment length polymorphism analysis
showing the pattern of the clonal strains G2 and G3. SA 2, SA3,
and SA6 are the primer combinations used. Each two lines cor-
respond to one duplicate analysis. Arrows indicate when bands
are missing in one of the strains.
FIG. 11. Neighbor-joining tree based on a partial 18S rDNA
sequences of raphidophyte and xanthophytes. Note the well-
supported clade including Gonyostomum semen with other marine
rapahidophytes and the freshwater species Vacuolaria virecens.
Emiliania huxley was included as an outgroup. Bootstrap values
are based upon 100 replicates. Scale shows substitutions per sites.
*Heterosigma carterae is synonymous to Heterosigma akashiwo (see
text for explanation).
LIFE CYLE AND SEXUALITY OF GONYOSTOMUM SEMEN
small size of the gametes along with the contraction
of planozygotes during encystment. For example, in
Fig. 8m, we show a cyst of 28.4mm of diameter, even
though it was formed by a planozygote of 37mm in
length (5h after gamete fusion). Sexual cyst size fitted
the normal range found for natural cysts, which
ranged from 27 to 37mm.
Among the marine raphidiophytes studied, a dor-
mancy period is not always required before cyst acti-
vation can occur. For example, Heterosigma resting cells
do not require a mandatory period of dormancy (Han
et al. 2002). In contrast, Chattonella cysts undergo a
mandatory dormancy that in nature extends from
summer to the following spring (Imai and Itoh 1987,
Imai et al. 1991). G. semen cysts may also have over-
wintering benthic stages that require a mandatory pe-
riod of dormancy, because in preliminary studies,
natural cysts stored in darkness at 41 C failed to excyst
during a 3-month period (unpublished data). The
highest viability was found at 161 C, which corresponds
with late spring/early summer temperatures in Swed-
ish lakes. Temperatures between 20 and 251 C pro-
moted the germination of cysts out of dormancy,
although they have a negative effect on the viability
of germlings. Temperature is also a crucial factor in the
germination of natural cysts of Chattonella and Hetero-
sigma. Cysts from H. akashiwo did not germinate below
101 C, and reached maximum germination values be-
tween 15 and 251 C (Imai and Itakura 1999). Chatton-
ella cysts also have an optimum temperature range for
germination and viability between 20 and 251 C, which
corresponds with Japanese seawater temperatures
during summer (Imai et al. 1991).
Sexual versus asexual reproduction. Previous studies
on the life cycles of raphidophytes (Imai 2003)
showed interesting and contradictory results regard-
ing sexuality. Initial studies on marine species point-
ed to some evidence of sexuality. For example,
Nakamura et al. (1990) observed sexual fusion and
zygote encystment in C. antiqua under culture condi-
tions, which suggested that cysts were diploid and
products of haploid fusion. Similarly, Imai (1989) re-
ported that both sexual and asexual processes could
be responsible for cyst formation in C. marina. How-
ever, Imai (1989) also suggested that asexual cyst
formation seemed to be most important in nature, as
natural cysts contained a nucleus with a size similar
to the one observed in vegetative stages cultured in
N-limited medium. Based on fluorometric analysis
of nuclear DNA, Yamaguchi and Imai (1994) later
concluded that C. antiqua and C. marina cysts had
a haploid DNA content, and this suggested the same
life cycle pattern as observed for the encystment of
H. akashiwo (Itakura et al. 1996). In these species, it
was proposed that DNA diploidization occurred after
excystment without fertilization (autodiploidy), as the
DNA content of cells of C. antiqua and C. marina
established from germinated cysts was 2C and 4C
(2 and 4 DNA contents, assuming that vegetative cells
are 2C in G1 phase). In contrast, we have verified
that the cell and nucleus size of G. semen planozygotes
are attained before encystment, and that the germ-
ling acquires vegetative morphology within 24h after
germination. In contrast to depictions by Cronberg
(2005), only one diploid cell germinated in most
cases, although germination of two cells was also
observed on one occasion.
Apart from resting cyst formation, we have ob-
served that G. semen cells formed asexual, short-lived
cysts when they were exposed for 72h to dark condi-
tions. These temporary resting states allow the cell to
withstand short-term environmental fluctuations, as
has been reported for the marine raphidophycean
H. akashiwo (Imai and Itakura 1999). In G. semen, this
round non-motile form can re-establish a vegetative,
motile existence in 24–72h when conditions become
favorable again (replete medium and light).
Phylogenetic analyses. Additionally, we confirmed
hidophycean family by nucleotide sequence analyses
of part of the 18S (SSU) rDNA. The phylogenetic tree
showed that G. semen clustered in a well-supported
clade with the marine genera Heterosigma and Chat-
tonella. G. semen also formed a distinct group with the
other freshwater taxon, V . virescens Cienkowski. Pot-
ter et al. (1997) have previously shown that V . viresc-
ens is a sister taxon to Chattonella subsalsa Biecheler
and Heterosigma carterae (Hulburt) Taylor (synony-
mous of H. akashiwo (Throndsen 1996)). We also
noted that the marine species Fibrocapsa japonica
Toriumi et Takano was not part of the G. semen,
Heterosigma, and Chattonella clade. Its position re-
ceived low bootstrap support and thus remains
We thank Polina Polykarpou, M. Svensson, I. Ramilo, and A.
Ferna ´ndez-Villamarı ´n for technical assistance. We also thank
Ramiro Logares for help with the phylogenetic analyses. The
research was supported by a CSIC grant for training of re-
search staff to R. I. F., by the Swedish Research Council (VR)
to K. R., and by the SEED project (GOCE-CT-2005-003875).
Anderson, D. M. & Rengefors, K. 2006. Community assembly and
seasonal succession of marine dinoflagellates in a temperate
estuary—the importance of life cycle events. Limnol. Oceanogr.
Anderson, D. M. & Wall, D. 1978. Potential importance of benthic
cysts of Gonyaulax tamarensis and G. excavata in initiating toxing
dinoflagellate blooms. J. Phycol. 14:224–34.
Bolch, C. J. S., Blackburn, S. I., Hallegraeff, G. M. & Vaillancourt,
R. E. 1999. Genetic variation among strains of the toxic dino-
flagellate Gymnodinium catenatum (Dinophyceae). J. Phycol.
Chapman, D. V., Dodge, J. D. & Heaney, S. I. 1982. Cyst formation
in the freshwater dinoflagellate Ceratium hirundinella (Din-
ophyceae). J. Phycol. 18:121–9.
Coyne, K. J., Handy, S. M., Demir, E., Whereat, E. B., Hutchins, D.
A., Doblin, M. A. & Cary, S. C. 2005. Improved quantitative
real-time PCR assays for enumeration of harmful algal species
in field samples using an exogenous DNA reference standard.
Limnol. Oceanogr. Methods 3:381–91.
Cronberg, G. 2005. The life cycle of Gonyostomum semen (Rap-
hidophyceae). Phycologia 44:285–93.
ROSA I. FIGUEROA AND KARIN RENGEFORS
Cronberg, G., Lindmark, G. & Bjo ¨r, S. 1988. Mass development of Download full-text
the flagellate Gonyostomum semen (Raphidophyta) in Swedish
forest lakes—an effect of acidification? Hydrobiologia 161:
Drouet, F. & Cohen, A. 1935. The morphology of Gonyostomum
semen from Woods Hole, Massachusetts. Biol. Bull. Mar. Biol
Lab., Woods Hole 68:422–39.
Eloranta, P. & Ra ¨ike, A. 1995. Light as a factor affecting the vertical
distribution of Gonyostomum semen (Ehr.) Diesing (Rap-
hidophyceae) in lakes. Aqua Fennica 25:15–22.
Figueroa, R. I., Rengefors, K. & Bravo, I. 2006. Effects of parental
factors and meiosis on sexual offspring of Gymnodinium nolleri
(Dinophyceae). J. Phycol. 42:350–62.
Fryxell, G. A. 1983. Survival Strategies of the Algae. Cambridge Uni-
versity Press, New York, 154 pp.
Guillard, R. R. L. & Lorenzen, C. J. 1972. Yellow-green algae with
chlorophyllide c. J. Phycol. 8:10–24.
Hall, T. A. 1999. BioEdit: a user-friendly biological sequence align-
ment editor and analysis program for Windows 95/98/NT.
Nucleic Acids Symp. Ser. 41:95–8.
Hansson, L.-A. 1996. Behavioural reponse in plants: adjustment in
algal recruitment induced by herbivores. Proc. Roy. Soc. London
Hansson, L.-A. 2000. Synergistic effects of food chain dynamics and
induced behavioral responses in aquatic ecosystems. Ecology
Han, M.-S., Kim, Y.-P. & Cattolico, R. A. 2002. Heterosigma akashiwo
(Raphidophyceae) resting cell formation in batch culture:
strain identity versus physiological response. J. Phycol.
Honjo, T. & Tabata, K. 1985. Growth and dynamics of Olisthodiscus
luteus in outdoor tanks with flowing coastal water and in small
vessels. Limnol. Oceanogr. 30:653–64.
Hongve, D., L?vstad, O. E. & Bj?rndalen, K. 1987. Gonyostomum
semen—a new nuisance to bathers in Norwegian lakes. Verh.
Internat. Verein. Limnol. 23:430–34.
Imai, I. 1989. Cyst formation of the noxious red tide flagellate
Chattonella marina (Raphidophyceae) in culture. Mar. Biol.
Imai, I. 1990. Physiology, morphology, and ecology of cysts of
Chattonella (Raphidophyceae), causative flagellates of noxious
red tides in the Inland Sea of Japan. Bull. Nansei Natl. Fish. Res.
Inst. 23:63–166 (in Japanese with English abstract).
Imai, I. 2003. Life history of raphidophycean flagellates. In Ok-
aichi, T. [Ed.] Red Tides. Terra Scientific Publishing Company,
Tokyo, pp. 111–24.
Imai, I. & Itakura, S. 1999. Importance of cysts in the population
dynamics of the red tide flagellate Heterosigma akashiwo (Rap-
hidophyceae). Mar. Biol. 133:755–62.
Imai, I., Itakura, S. & Itoh, K. 1991. Life cycle strategies of the red
tide causing flagellates Chattonella (Raphidophyceae) in the
Seto Inland sea. Mar. Pollut. Bull. 23:165–70.
Imai, I., Itakura, S. & Itoh, K. 1993a. Cysts of the red tide flagellate
Heterosigma akashiwo, raphidophyceae, found in bottom sedim-
ents of northern Hiroshima Bay, Japan. 1993. Nippon Suisan
Imai, I., Itakura, S. & Ouchi, A. 1993b. Occurrence of Chattonella
red tide and cyst dynamics in sediments in northern Hiroshi-
ma Bay, the Seto Inland Sea, Japan. Nippon Suisan Gakk. 59:
1–6 (in Japanese with English abstract).
Imai, I. & Itoh, K. 1987. Annual life cycle of Chattonella spp.,
causative flagellates of noxious red tides in the Inland Sea of
Japan. Mar. Biol. 94:287–92.
Imai, I. & Itoh, K. 1988. Cysts of Chattonella antiqua and C. marina
(Raphidophyceae) in sediments of the Inland Sea of Japan.
Bull. Plankton Soc. Japan 35:35–44.
Itakura, S., Nagasaki, K., Yamaguchi, M. & Imai, I. 1996. Cyst
formation in the red tide flagellate Heterosigma akashiwo
(Raphidophyceae). J. Plankton Res. 18:1975–9.
Kelley, I. & Pfiester, L. A. 1990. Sexual reproduction in the
freshwater dinoflagellate Gloeodinium montanum. J. Phycol. 26:
Kumar, S., Tamura, K. & Nei, M. 2004. MEGA 3: Integrated soft-
ware for Molecular Evolutionary Genetics Ana ´lysis and
sequence alignment. Briefings Bioinform. 5:150–63.
Le Cohu, P. R., Guitard, J., Comoy, N. & Brabet, J. 1989. Gonyosto-
mum semen a potential nuisance in large French reservoirs? The
case of the Pareloup lake. Arch. Hydrobiol. 117:225–36.
Lepisto ¨, L., Antikainen, S. & Kivinen, J. 1994. The occurrence of
Gonyostomum semen (Ehr.) Diesing in Finnish lakes. Hydrobiol-
Litaker, R. W., Vandersea, M. W., Kliber, S. R., Noga, E. J. & Tester,
P. A. 2002. Reply to comment on the life cycle and toxicity of
Pfiesteria piscicida revised. J. Phycol. 38:1268–72.
Nagasaki, K., Itakura, S., Imai, I., Nakagiri, S. & Yamaguchi, M.
1996. The disintegration process of a Heterosigma akashiwo
(Raphidophyceae) red tide in northern Hiroshima Bay,
Japan, during the summer of 1994. In Ysumoto, T., Oshima,
Y. & Fukuyo, Y. [Eds.] Harmful and Toxic Algal Blooms. Inter-
governmental Oceanographic Commission of UNESCO,
Paris, pp. 251–54.
Nakamura, Y., Umemori, T., Watanabe, M., Kulis, D. M. & And-
erson, D. M. 1990. Encystment of Chattonella antiqua in labo-
ratory cultures. J. Oceanogr. Soc. Japan 46:35–43.
Ono, C. 1988. Cell cycles and growth rates of red tide organisms in
Harima Nada aea, eastern part of Seto Inland Sea. Japan. Bull.
Akashiwo Res. Inst. Kagawa Pref. 3:1–67 (in Japanese with
Pfiester, L. & Anderson, D. M. 1987. Dinoflagellate reproduction.
In Taylor, F. J. R. [Ed.] Biology of Dinoflagellates. Blackwell
Scientific Publications, Oxford, pp. 611–48.
Pithart, D., Pechar, L. & Mattsson, G. 1997. Summer blooms of
raphidophyte Gonyostomum semen and its diurnal vertical
migration in a floodplain pool. Arch. Hydrobiol. Suppl. Algolog.
Potter, D., Saunders, G. W. & Andersen, R. A. 1997. Phylogenetic
relationships of the Raphidophyceae and Xanthophyceae as
inferred from nucleotide sequences of the 18S ribosomal
RNA gene. Am. J. Bot. 84:966–72.
Sandgren, C. D. 1983. Survival strategies of chrysophycean flagel-
lates: reproduction and the formation of resistant resting cysts.
In Fryxell, G. A. [Ed.] Survival Strategies of Algae. Cambridge
University Press, Cambridge, pp. 23–48.
Sako, Y., Ishida, Y., Nishijima, T. & Hata, Y. 1987. Sexual re-
production and cyst formation in the freshwater dino-
flagellate Peridinium penardii. Nippon Suisan Gakk. 53:473–8.
Salonen, K. & Rosenberg, M. 2000. Advantages from diel vertical
migration can explain the dominance of Gonyostomum semen
(Raphidophyceae) in a small, steeply-stratified humic lake.
J. Plankton Res. 22:1481–853.
Throndsen, J. 1996. Note on the taxonomy of Heterosigma akashiwo
(Rapidophyceae). Phycologia 35:367.
von Stosch, H. A. 1973. Observations on vegetative reproduction
and sexual life cycles of two freshwater dinoflagellates, Gym-
nodinium pseudopalustre Schiller and Woloszynskia apiculata sp.
nov. Br. Phycol. J. 8:105–34.
Vos, P., Hogers, R., Bleeker, M., Reijans, M., van de Lee, T.,
Hornes, M., Frijters, A., Pot, J., Peleman, J. & Kuiper, M.
1995. AFLP: a new technique for DNA fingerprinting. Nucleic
Acids Res. 23:4407–14.
Watanabe, M. M., Nakamura, Y. & Kohata, K. 1983. Diurnal
vertical migration and dark uptake of nitrate and phosphate
of the red tide flagellates, Heterosigma akashiwo HADA and
Chattonella antiqua (Hada) Ono (Raphidophyceae). Jpn. J.
Wille ´n, E. 2003. Dominance patterns of planktonic algae in
Swedish forest lakes. Hydrobiologia 502:315–24.
Yamaguchi, M. & Imai, I. 1994. A microfluorometric analysis of
nuclear DNA at different stages in the life history of Chattonella
antiqua and Chattonella marina (Raphidophyceae). Phycologia
Yamochi, S. & Abe, T. 1984. Mechanisms to initiate a Heterosigma
akashiwo red tide in Osaka Bay. II. Diel vertical migration.
Mar. Biol. 83:255–61.
LIFE CYLE AND SEXUALITY OF GONYOSTOMUM SEMEN