Sterol chemotaxonomy of marine pelagophyte algae

Article abstract of DOI:10.1002/cbdv.200800316

Several marine algae of the class Pelagophyceae produce the unusual marine sterol 24-propylidenecholesterol, mainly as the (24E)-isomer. The (24Z)-isomer had previously been considered as a specific biomarker for Aureococcus anophagefferens, the 'brown ti

Full text of DOI:10.1002/cbdv.200800316

CHEMISTRY & BIODIVERSITY – Vol. 6 (2009)
1111
Sterol Chemotaxonomy of Marine Pelagophyte Algae
by Jose´-Luis Giner*a), Hui Zhaoa), Gregory L. Boyera), Michael F. Satchwella), and
Robert A. Andersenb)
a) Department of Chemistry, State University of New York-ESF, Syracuse, NY 13210, USA
(phone: þ1-315-470-6895; fax: þ1-315-470-6856 ; e-mail: jlginer@syr.edu)
b) Bigelow Laboratory for Ocean Sciences, West Boothbay Harbor, ME 04575, USA
Dedicated to Professor Carl Djerassi on the occasion of his 85th birthday
Several marine algae of the class Pelagophyceae produce the unusual marine sterol 24-
propylidenecholesterol, mainly as the (24E)-isomer. The (24Z)-isomer had previously been considered
as a specific biomarker for Aureococcus anophagefferens, the ꢀbrown tideꢁ alga of the Northeast coast of
the USA. To test this hypothesis and to generate chemotaxonomic information, the sterol compositions
of 42 strains of pelagophyte algae including 17 strains of Aureococcus anophagefferens were determined
by GC analysis. A more comprehensive sterol analysis by HPLC and 1H-NMR was obtained for 17
selected pelagophyte strains. All strains analyzed contained 24-propylidenecholesterol. In all strains
belonging to the order Sarcinochrysidales, this sterol was found only as the (E)-isomer, while all strains in
the order Pelagomonadales contained the (Z)-isomer, either alone or together with the (E)-isomer. The
occurrence of D22 and 24a-sterols was limited to the Sarcinochrysidales. The first occurrence of D22-24-
propylcholesterol in an alga, CCMP 1410, was reported. Traces of the rare sterol 26,26-dimethyl-24-
methylenecholesterol were detected in Aureococcus anophagefferens, and the (25R)-configuration was
proposed, based on biosynthetic considerations. Traces of a novel sterol, 24-propylidenecholesta-5,25-
dien-3b-ol, were detected in several species.
Introduction. – The C30 marine sterol, 24-propylidenecholesterol, is generated in an
unusual and mechanistically interesting biosynthetic reaction that involves a proto-
nated cyclopropane [1][2]. The first isolation of this sterol was from the scallop
Placopecten magellanicus, where it was found as a mixture of the (24E)- and (24Z)-
isomers (1 and 2, resp.) [3–5]. Its presence in this and other marine invertebrates is
thought to reflect dietary input. The (E)-isomer 1 was subsequently detected in two
marine algae, Chrysoreinhardia sp. [6], and Nematochrysopsis roscoffensis [7]. Later, a
mixture of 1 and 2 was found in Aureococcus anophagefferens [8], the species
responsible for the ꢀbrown tidesꢁ that have decimated the scallop fisheries of the
Northeast coast of the USA [9]. Another harmful alga, Aureoumbra lagunensis, the
cause of the ꢀTexas brown tideꢁ along the coast of the Gulf of Mexico, was found to
contain 1 [10].
All of the algae that make 24-propylidenecholesterol belong to a single class, the
Pelagophyceae, which is comprised of the orders Pelagomonadales and Sarcinochry-
sidales [11–14]. These small unicellular marine algae were formerly grouped in the
Chrysophyceae, and include species that grow on the surface of substrates in shallow
waters (Sarcinochrysis marina and Chrysoreinhardia sp.), algae that form harmful
ꢂ 2009 Verlag Helvetica Chimica Acta AG, Zꢃrich

1112
CHEMISTRY & BIODIVERSITY – Vol. 6 (2009)
blooms in coastal waters (A. anophagefferens and A. lagunensis), and open-ocean
picoplanktonic species that are important to global carbon fixation (Pelagomonas
calceolata and Pelagococcus subviridis) [15–17]. Sterols have long been recognized as
useful characteristics for distinguishing algal groups [18][19], and it was proposed that
all pelagophyte algae contain 24-propylidenecholesterol [8]. It was also proposed that
the (Z)-isomer 2 was specific to Aureococcus anophagefferens and could be regarded as
a biomarker for this species [8]. To test these hypotheses, a survey of the sterol
compositions of 42 strains of Pelagophyte algae including 17 strains of Aureococcus
anophagefferens was carried out.
Results and Discussion. – 1. GC Analysis. Quantitative gas-chromatographic
analysis was carried out for the sterols of 17 strains of Aureococcus anophagefferens
(Table 1), twelve strains of other members of the Pelagomonadales (Table 2), and 13
strains from the Sarcinochrysidales (Table 3). The sterol identities were assigned by
comparison of their retention times with those of known standards.
To assist in the interpretation of the large amount of information, the method to
display the data graphically was developed based on biosynthetic parameters (Fig. 1).
In the biosynthetic modification of sterol side chains [2], successive methylation
reactions of S-adenosylmethionine (SAM)-dependent sterol methyltransferases con-
vert C27 sterols to C28 sterols, C28 sterols to C29 sterols, and C29 sterols to C30 sterols
(Scheme 1). These sterols can then undergo subsequent modification by redox enzymes
whereby the side-chain C¼C bond is first hydrogenated by a sterol side-chain
reductase, and a new C¼C bond is introduced at the 22-position by a D22 sterol
desaturase. Two parameters were calculated for the total sterols of each organism, the
methylation degree (MD) and reduction-oxidation degree (ROD). These parameters
were defined according to the equations:
Methylation degree (MD)¼1ꢀ% C28 sterolsþ2ꢀ% C29 sterolsþ3ꢀ% C30 sterols
Reduction–oxidation degree (ROD)¼1ꢀ% sterols with saturated side chains
þ2ꢀ% sterols with D22-side chains
1.1. Aureococcus anophagefferens. For the brown tide alga, Aureococcus anopha-
gefferens, the major sterols found by GC analysis were 24-methylenecholesterol (19),
24-methylcholesterol (20/21), (E)-24-propylidenecholesterol (1), and (Z)-24-propyli-

CHEMISTRY & BIODIVERSITY – Vol. 6 (2009)
1113

1114
CHEMISTRY & BIODIVERSITY – Vol. 6 (2009)

CHEMISTRY & BIODIVERSITY – Vol. 6 (2009)
1115

1116
CHEMISTRY & BIODIVERSITY – Vol. 6 (2009)
denecholesterol (2; Table 1). Except for CCMP 1853, for which 2 was not detected, the
(E)/(Z)-ratio of 24-propylidenecholesterol (1 vs. 2) varied from 1.4 to 4.3, with an
average of 3.2. The methylation degree (MD) calculated for these strains ranged from
1.3 to 2.4, with a cluster centered at ca. 2.1 and another centered at ca. 1.6 (Fig. 1). C28
and C30 sterols were predominant, with no C27 sterols and only small amounts of C29
sterols typically being present. Unusually high amounts of C29 sterols were found in
CCMP 1789 (36.4%), CCMP 1850 (16.4%), and CCMP 1853 (14.0%). The ROD was
low in this species, ranging between 0 and 0.6. This variation was largely due to the
amount of 24-methylcholesterol (20/21) present. However, because the retention times
of 24-methylcholesterol (20/21) and 24-methylenecholesterol (19) were close, the ratios
of these two sterols could not be accurately determined by GC.
The observed variation in sterol composition between the different strains within
this species could be due to genetic factors, and such traits as a high ratio of 1 vs. 2 or a
low ROD might be characteristic for particular strains. Alternatively, the variation in
MD or ROD could reflect the availability of the cofactors for the biosynthetic enzymes
(SAM and NADPH), which might be influenced by the metabolic state of the cells. It
was unknown to what extent the sterol composition changes during the growth cycle of
a culture. To test this, a culture of CCMP 1708 was sampled, and the sterols were

CHEMISTRY & BIODIVERSITY – Vol. 6 (2009)
1117
Scheme 1. Sterol Side-Chain Biosynthesis
analyzed at different stages during its growth (Table 4). Again, the proportions of 24-
methylenecholesterol (19) and 24-methylcholesterol (20/21) were difficult to evaluate.
Despite this, the sterol compositions at different time points proved to be fairly
constant. However, it is worth noting that this strain showed significantly larger
amounts of 24-propylidenecholesterol in the growth study (Table 4) than in the GC
analysis (Table 1), and that this strain would, therefore, be relocated into the upper
cluster of Aureococcus strains in the biosynthetic graph (Fig. 1). Every strain that was
re-examined by HPLC (see below), also mapped into the upper cluster.
1.2. Pelagomonas and Pelagococcus. Besides Aureococcus, two other genera of the
order Pelagomonadales were analyzed (Table 2). Both of these, Pelagomonas and
Pelagococcus, are picoplanktonic organisms that grow in the open ocean. The former is
a flagellate occurring in tropical and warm temperate waters, and the latter a coccoid
species found in colder waters. Pelagomonas calceolata is the basis for the order
Pelagomonadales and the class Pelagophyceae [11].
The predominant sterol of Pelagomonas calceolata was shown to be 24-methyl-
enecholesterol (19), which accounted for 48–80% of total sterols. The calculated
biosynthetic parameters clustered very closely for this species with a MD of ca. 1.7 and
a ROD of ca. 0 (Fig. 1). Like Aureococcus, C29 sterols were relatively minor

1118
CHEMISTRY & BIODIVERSITY – Vol. 6 (2009)
^
Fig. 1. Biosynthetic mapping of the sterol compositions of pelagophyte algae based on GC analysis.
:
&
*
Aureococcus anophagefferens, : Aureoumbra lagunensis, ꢀ: unidentified coccoid species, : Pelago-
~
~
&
^
:
monas calceolata,
Pelagomonas cf. calceolata, : Pelagococcus subviridis, : Pelagococcus sp.,
*
Chrysoreinhardia sp., : Sarcinochrysis marina. *: MD and ROD were re-evaluated by HPLC and NMR
analysis (see Tables 1–3).
components. However, in contrast to Aureococcus, the (E)-isomer of 24-propylide-
necholesterol (1) was absent, and only the (Z)-isomer (2) was found. The strain CCMP
1954 was substantially different from the other strains because of the presence of sterols
containing a D22 C¼C bond. This was regarded suspicious because of the typical
ROD¼0 found in this species. Also, although D22-24-methylcholesterol (22/23) was
present, none of its biosynthetic precursor, 24-methylcholesterol (20/21), was detected
in this or any of the other strains of this species (see Scheme 1). Contamination during
culture or analysis was, therefore, considered likely. In the HPLC analysis (see below),
D22-24-methylcholesterol (22/23) was not detected.
The alga CCMP 1145 (Pelagomonas cf. calceolata) was an anomaly. Unlike the
other strains of Pelagomonas calceolata, it is a coccoid species, but it is clearly closely
related as shown by an identical 18S rRNA gene sequence as CCMP 1214 [20]. The

CHEMISTRY & BIODIVERSITY – Vol. 6 (2009)
1119
Table 4. Sterol Composition of a Culture of Aureococcus anophagefferens (CCMP 1708) at Different
Points in the Growth
Sterolsa)
RRTb)
Percentage of Total Sterols
Day 9
Day 11
Day 12
Day 14
Day 16
24MC
24Me
24EP
24ZP
1.14
1.15
1.44
1.50
28.6
5.6
48.9
16.9
26.8
6.2
51.8
15.2
28.9
5.6
50.1
15.4
31.6
6.1
48.5
13.8
33.3
9.3
45.2
12.2
MDc)
2.32
0.06
2.9
1.85ꢀ106
2.34
0.06
3.4
4.20ꢀ106
2.31
0.06
3.3
4.80ꢀ106
2.25
0.06
3.5
4.43ꢀ106
2.15
0.09
3.7
2.18ꢀ106
RODc)
(E)/(Z)c)
Cells/ml
a) Abbreviations of sterols: 24MC¼24-methylenecholesterol (19); 24Me¼24-methylcholesterol (20/
21); 24EP¼(E)-24-propylidenecholesterol (1); 24ZP¼(Z)-24-propylidenecholesterol (2). b) RRT¼
c
GC Retention time to relative cholesterol. ) MD¼Methylation degree (see text). ROD¼Reduction–
oxidation degree (see text). (E)/(Z)¼Ratio of the (24E)- vs. (24Z)-isomers of 24-propylidenecholes-
terol (1 vs. 2).
sterol profile it displayed, however, was very different, since its major sterols, 24-
methyl- and 24-ethylcholesterol (14/15 and 20/21, resp.), bear saturated side chains.
Also in contrast to most strains of P. calceolata, the (E)-isomer 1 of 24-propylide-
necholesterol was present, and predominated over the (Z)-isomer 2. Based on the
graph of biosynthetic parameters (Fig. 1), and its ratio of 1 and 2 (Table 2), this alga
was found to be very similar to CCMP 1429 (Pelagococcus subviridis).
Three strains of the genus Pelagococcus, CCMP 1429 (Pelagococcus subviridis), and
two unnamed species (CCMP 1395 and CCMP 1252), were also analyzed by GC
(Table 2). These three strains contained both isomers of 24-propylidenecholesterol, 1
and 2, at ratios between 1.5 and 3.8. The sterol composition of Pelagococcus subviridis
(CCMP 1429) was quite similar to that of the Pelagococcus species CCMP 1395, which
also contained more than 80% sterols with saturated side chains, in this case
predominantly as 24-ethylcholesterol (14/15). However, the Pelagococcus species
CCMP 1252 contained mainly unreduced sterols and more closely resembled A.
anophagefferens.
1.3. Sarcinochrysidales. A total of 13 strains were analyzed from the order
Sarcinochrysidales, representing four different species including the ꢀTexas brown tideꢁ
alga, Aureoumbra lagunensis, Chrysoreinhardia sp., Sarcinochrysis marina, and an
unidentified coccoid CCMP 1410 (Table 3). The (Z)-24-propylidenecholesterol (2)
was not detected in any of these strains. Unlike the Pelagomonadales, all members of
the Sarcinochrysidales contained D22-sterols.
The nine strains of the ꢀTexas brown tideꢁ, Aureoumbra lagunensis, showed very
similar sterol compositions. The MD ranged between 1.9 and 2.4, and the ROD from 0.7
to 1.3 (Fig. 1). The major sterols were D22-24-ethylcholesterol (16/17; average 30.4%),
(E)-24-propylidenecholesterol (1; average 28.7%), cholesterol (25; average 13.8%),
24-propylcholesterol (4/5; average 11.7%), and ethylcholesterol (14/15; average

1120
CHEMISTRY & BIODIVERSITY – Vol. 6 (2009)
11.5%). This species is unusual among the pelagophytes because of its relatively large
content of cholesterol (25) and 24-propylcholesterol (4/5). We previously suggested 4/5
could be a useful biomarker for the ꢀTexas brown tideꢁ [10][21], and this study confirms
its presence in all strains of A. lagunensis.
Four other strains from the Sarcinochrysidales were analyzed. Two of these
represent the type species for this order, Sarcinochrysis marina, which forms colonies
on rocks in the splash zone (superlitoral) in temperate waters. In the literature, the
sterol composition of this species is mentioned twice in reference to prior studies
[22][23], however, the citations referred to in those publications do not deal with this
species, and it is likely to have been confused with CCMP 292, which was extensively
investigated chemically before it was securely classified biologically. CCMP 292, which,
like Sarcinochrysis marina, also forms colonies on substrates in shallow water, was
previously referred to either as an unidentified chrysophyte [6][22], Chrysoderma
mucosa [1], or Pulvinaria sp. [8], and has now been identified as Chrysoreinhardia sp.
Our analysis showed the sterol compositions of Chrysoreinhardia sp. and Sarcinochrysis
marina to be quite similar (Table 3). Like Aureoumbra, both algae contained sterols
predominantly with saturated side chains and D22 C¼C bonds. However, the C30 sterols
were largely unmodified, indicating that the side chain reductase apparently does not
function well with the 24-propylidene group, probably due to its steric bulk. The sterol
composition of CCMP 1410 (an unidentified coccoid strain collected from the Sargasso
Sea at a depth of 84 m) very closely resembled that of Aureoumbra lagunensis.
2. Analysis by HPLC and NMR. Although GC analysis is sensitive and requires
only small amounts, it often does not provide adequate information for molecular
identification, even when coupled to MS (mass spectrometry) detection, especially for
stereoisomers or in situations where rare and unusual sterols might be present [24].
Also, at the high temperatures required for volatilization, selective decomposition of
the less stable sterols may occur. To achieve greater confidence in our data, a detailed
analysis was undertaken of selected strains using reversed-phase (RP) HPLC to
separate the sterol mixtures, and 600-MHz NMR to provide definite identities of the
components.
Several algal strains were chosen for reinvestigation, including some obvious
outliers that were evident from the graph of biosynthetic descriptors, e.g., CCMP 1954
and CCMP 1848 (Fig. 1). Larger sample sizes (8-l cultures vs. 200 ml for GC analysis)
and the sensitivity of the NMR instrument allowed the detection of even minor sterols.
In total, 17 were examined, and 23 sterols were identified including a new C30 sterol, 3.
All of the sterols had the D5-sterol nucleus.
2.1. Aureococcus anophagefferens. Seven strains of A. anophagefferens were
reinvestigated (Table 5). The results were generally consistent with the GC analysis,
although there were some differences. All strains of A. anophagefferens contained both
(E)- and (Z)-24-propylidenecholesterol (1 and 2, resp.) as a high percentage of the
total sterols (48–67%). The (E)/(Z) ratios (1/2) were somewhat higher for the HPLC
analysis than for GC analysis (Table 1). The strain CCMP 1853, for which 2 was not
detected by GC analysis, was found to have similar amounts of 2 as the other A.
anophagefferens strains.
A factor complicating the ROD calculation from the GC data was that 24-
methylenecholesterol (19) and 24-methylcholesterol (20/21) had very close retention

CHEMISTRY & BIODIVERSITY – Vol. 6 (2009)
1121

1122
CHEMISTRY & BIODIVERSITY – Vol. 6 (2009)
times. In some cases, only one of the two sterols was detected, perhaps because of poor
resolution (Table 1). HPLC led to a complete separation of these two sterols and
showed that both sterols were present in all of the A. anophagefferens strains analyzed,
although the proportions were highly variable (Table 5). Nevertheless, good general
agreement was found in the relative amounts of the two sterols as determined by the
two methods. All strains, with the exception of CCMP 1785, that had higher amount of
one sterol when measured by GC, also had more of that sterol when measured by
HPLC. The high variability in the ratios of 24-methylcholesterol (20/21) to 24-
methylenecholesterol (19) may be explained by differences in the activities of the
biosynthetic enzymes among strains. In some strains, the alkylation of the 24-methylene
side chain were faster, while reduction of D24(28) C¼C bond was favored in other strains.
Whether this is a genetic characteristic of the strains, or it depends on the metabolic
conditions, remains to be determined.
Relatively high levels of C29 sterols were observed for CCMP 1789, 1850, and 1853
by GC analysis (Table 1). By HPLC analysis, however, all strains had low amounts of
C29 sterols (8, 10, 11, and 12), ranging from 2.9–4.4%. While the reason for the high
levels of C29 sterols as determined by GC is unknown, HPLC analysis is more reliable,
and low levels of C29 sterols should, therefore, be considered typical of this species.
1
The high sensitivity of 600-MHz H-NMR analysis enabled us to identify several
minor sterols in A. anophagefferens. From CCMP 1847 and CCMP 1853, two minor
compounds were found to co-elute with (E)-24-propylidenecholesterol (1). Attempts
to isolate them were carried out using different HPLC mobile phases and preparative
argentic TLC. However, none of these methods was successful. Based on the retention
1
time and by comparison with published H-NMR data, the two compounds were
identified as the two epimers 8 and 9 of 24-isopropenylcholesterol [25][26]. A
cyclopropa-sterol was isolated from the (24S)-24-methylcholesterol (21) fraction of
CCMP 1847, and was shown to be the (24S,28S)-isomer of 24,28-methylenestigmast-5-
1
en-3b-ol (10) based on H-NMR data [19][27][28]. Another minor sterol co-eluted
with (Z)-24-propylidenecholesterol (2) in all seven of the strains investigated. This
could not be separated using different HPLC mobile phases, but was purified by
repeated preparative argentic TLC of the acetate esters. After saponification, the
minor sterol was identified as 26,26-dimethyl-24-methylenecholesterol (11) by
1
comparison with the published H-NMR data [22]. These four minor sterols, 8–11,
are C30 sterols that are believed to be by-products in the biosynthesis of 24-
propylidenecholesterol (Scheme 2). The cyclopropa-sterol 10 is product of simple
deprotonation of the hypothetical nonclassical carbocationic intermediate 26 [1][2].
Although the configuration of sterol 11 is presently unknown, based on least motion
pathways for the two possible biosynthetic alkyl shifts from 27 to either 28 or 29
(Scheme 2), we postulate that the configuration of C(25) is (R).
Another minor sterol was found in all Aureococcus strains analyzed by HPLC. It co-
eluted with clerosterol (18), and was isolated by changing the HPLC mobile phase to
MeCN/MeOH/AcOEt 11:4 :4 (v/v/v). Mass spectrometry showed an Mþ ion peak at
1
m/z 424, suggesting the formula C30H48O. The H-NMR spectrum of this compound
showed signals of four olefinic H-atoms that were consistent with a D5-sterol-triene
bearing three olefinic H-atoms in the side chain. In CDCl3, two broad singlets in a
pattern typical of an olefinic ¼CH2 group were found at 4.976 and 4.867 ppm. The third

CHEMISTRY & BIODIVERSITY – Vol. 6 (2009)
1123
Scheme 2. Mechanistic Details of the Biosynthetic Formation of C30 Sterols in Pelagophytes
olefinic signal was a triplet at 5.508 ppm, typical of a trisubstituted olefin. Structure
determination by 2D-NMR (Fig. 2) was performed in C6D6, because of overlapping
signals in the Me region of the CDCl3 spectrum. In C6D6, the chemical shifts of these
three H-atoms changed to 5.015, 5.225, and 5.608 ppm, respectively. Correlations were
Fig. 2. Selected 2D-NMR correlations for the novel C30 sterol D25-24-propylidenecholesterol (3): a) COSY
correlations; b) HMBC correlations.

1124
CHEMISTRY & BIODIVERSITY – Vol. 6 (2009)
observed in the COSY spectrum for the olefinic methylene H-atoms and a Me singlet at
1.925 ppm, indicating an isopropenyl group. Both the triplet olefinic H-atom peak and a
Me triplet at 1.007 ppm had a correlation with a CH2 multiplet at 2.167 ppm. In the
HMBC spectrum, the Me triplet (1.007 ppm) had correlations with an olefinic C-atom
signal at 129.6 and a signal at 22.5 ppm. The multiplet at 2.167 ppm also had correlations
with the olefinic C-atom signal at 129.6 and another olefinic C-atom signal at
140.6 ppm, indicating the multiplet came from two allylic H-atoms. The olefinic C-atom
signal at 140.6 ppm correlated to the Me singlet at 1.925 ppm, which also correlated to
two other olefinic C-atom signals at 144.1 and 111.2 ppm. All of the information was
consistent with D25-24-propylidenecholesterol (3). An attempt to determine the
configuration of the D24(28) C¼C bond by ROESY was unsuccessful. This novel
compound was isolated as a minor component in all of the Pelagophyte species
analyzed by HPLC. Because it is so ubiquitous, it is believed to originate from the
decomposition of 24-propylidenecholesterol (1/2), perhaps via an allylic hydroperoxide
intermediate. In our experience, 24-propylidenecholesterol (1/2), like other trisub-
stituted olefinic sterols, is rather susceptible to decomposition. Precautions were,
therefore, taken throughout this study to prevent autoxidation: exposure to light was
minimized, especially in the presence of pigments, and tetramethylethylene was added
to the solvents as a singlet-oxygen scavenger. It is unknown whether the two isomers of
24-propylidenecholesterol display different stabilities, but we suspect the (Z)-isomer 2
is more labile.
2.2. Pelagomonas. Of the eight strains of Pelagomonas calceolata analyzed by GC,
four were reinvestigated by HPLC; strain CCMP 1214 was only analyzed by HPLC.
The results (Table 5) were generally consistent with the GC data (Table 2). The
previous detection of D22-sterols in CCMP 1954, which raised suspicions of contam-
ination, was not confirmed by HPLC analysis, and this strain looked very similar to the
others of the species. All strains displayed RODs close to zero. All of the strains except
CCMP 1214 contained almost exclusively only the (Z)-isomers of 24-propylidenecho-
lesterol and of 24-ethylidenecholesterol (2 and 13, resp.). To rule out the possibility to
contamination, the HPLC/NMR analysis of CCMP 1214 was repeated with a fresh
culture, giving similar results. The unusual (E)/(Z) ratio of CCMP 1214 compared to
other strains of the same species is further complicated by the fact that CCMP 1214 is
the authentic strain for this species [11]. The reasons that most strains of this species
favor the (Z)-isomers 2 and 13 are unknown. We have previously shown that
isofucosterol (13) is the biosynthetic precursor of (E)-24-propylidenecholesterol (1) in
Chrysoreinhardia [1], and it would be interesting to determine the precursor of the (Z)-
isomer 2 in Pelagomonas calceolata.
The HPLC results confirmed the differences that GC analysis indicated between
Pelagomonas cf. calceolata (CCMP 1145) and the different strains of Pelagomonas
calceolata. A second HPLC analysis with a fresh culture of CCMP 1145 also gave
similar results. As discussed previously, despite an 18S rRNA sequence identical to
Pelagomonas calceolata, the sterol composition of CCMP 1145 closely resembles that of
Pelagococcus subviridis (CCMP 1429). This may be related to the fact that CCMP 1145,
unlike the other Pelagomonas calceolata strains which are flagellate [11], is a
nonflagellate, coccoid cell surrounded by a thick, three-layered cell wall. Similarly,
Pelagococcus subviridis is a coccoid cell with a trilayered wall [29], and it is possible

CHEMISTRY & BIODIVERSITY – Vol. 6 (2009)
1125
that similarities in sterol composition may be partly related to the life form. Since the
same species may have different life forms (e.g., alternating between a flagellate
swimming stage and a nonmotile coccoid stage), the sterol composition could
potentially change with the life form, which may account for the anomalous
chemotaxonomic affinities of CCMP 1145.
2.3. Pelagococcus. Detailed analysis of Pelagococcus subviridis (CCMP 1429) and
two unnamed Pelagococcus species (CCMP 1252 and CCMP 1395) by HPLC and NMR
(Table 5) gave generally similar results to those obtained by GC analysis (Table 2). As
discussed previously, the sterol composition of CCMP 1395 was similar to that of the
type species, Pelagococcus subviridis (CCMP 1429), and like P. subviridis, has a similar
trilayered cell wall (unpublished results). Of all of the pelagophyte algae investigated,
only CCMP 1395 contained significant amounts of sterol esters. After saponification,
the composition of these, however, did not differ much from that of the free sterols. The
configuration of the 24-alkyl-sterols in this genus was shown by 1H-NMR to be
exclusively 24b (21 and 15). The D22-sterols that had previously been detected by GC
analysis were not confirmed by NMR. All Pelagococcus strains showed both isomers of
24-propylidenecholesterol, and the (E)/(Z)-ratios (1/2) were similar to Aureococcus,
but the quantities were considerably lower, especially for CCMP 1395.
The sterol composition of CCMP 1252 was quite different from the others because
of the lack of side chain reductase activity in this strain as inferred from the lack of
reduced sterol side chains (Scheme 1). In contrast, CCMP 1395 and CCMP 1429
contained mainly sterols with saturated side chains. In the GC analysis, CCMP 1252 had
contained a small amount of sterol with a saturated side chain (i.e., 14/15), but, upon
HPLC analysis, no reduced sterols were evident. The HPLC analysis also showed more
24-methylenecholesterol (19) than before. From the GC analysis, CCMP 1252 most
closely resembled Aureococcus, but this similarity was not strong by HPLC analysis.
Instead, CCMP 1252 was found to strongly resemble the authentic strain for
Pelagomonas calceolata (i.e., CCMP 1214).
2.4. CCMP 1410. The algal strain CCMP 1410 is an unidentified coccoid species
belonging to the Sarcinochrysidales. GC Analysis had shown this species had a sterol
profile similar to that of Aureoumbra lagunensis (Table 3). HPLC and NMR analyses
displayed a complex mixture (Table 6). In total, 14 sterols were isolated. The
predominant sterol of this alga was (E)-24-propylidenecholesterol (1); none of the (Z)-
isomer 2 was observed. The other major sterols were D22-(24S)-24-ethylcholesterol
(stigmasterol, 16) and cholesterol (25). Phylogenetic studies based on 18S rRNA and
rbcL had shown this alga to be closely related to Aureoumbra lagunensis, and also
related to Chrysoreinhardia sp. and Nematochrysopsis roscoffensis (Fig. 3)
[13][14][30]. Comparison of the published sterol compositions of these algae showed
that the sterol pattern of CCMP 1410 more closely resembles that of N. roscoffensis and
A. lagunensis, than that of Chrysoreinhardia (Table 6).
In a previous study, 24a-sterols, but no 24b-sterols were isolated from Aureoumbra
lagunensis [10]. Because of this, it was suggested that 24a-sterols were a characteristic
of chemotaxonomic value for the Sarcinochrysidales, and it was predicted that, if an
algal source were found for D22-24-propylcholesterol (6/7), it would contain the 24a-
configuration (6) [21]. While the 24a-sterols predominated in CCMP 1410, and the
24a-configuration was the only configuration of the C29 sterols 14 and 16, 24b-sterols

1126
CHEMISTRY & BIODIVERSITY – Vol. 6 (2009)
Table 6. Sterol Composition of CCMP 1410 Compared to Other Algae in the Sarcinochrysidales
Sterolsa)
RRTb)
Percentage of Total Sterols
1410c)
1681d)
292e)
Nrf)
24-Methylenecholesterol (19)
(24S)-D22-24-Methylcholesterol (22)
(24R)-D22-24-Methylcholesterol (23)
(24S)-D25-24-Ethylcholesterol (18)
D25-24-Propylidenecholesterol (3)
Cholesterol (25)
(E)-24-Propylidenecholesterol (1)
(24S)-D22-24-Ethylcholesterol (16)
(24S)-24-Methylcholesterol (21)
(24R)-D22-24-Propylcholesterol (7)
(24S)-D22-24-Propylcholesterol (6)
(24R)-24-Ethylcholesterol (14)
(24S)-24-Propylcholesterol (5)
(24R)-24-Propylcholesterol (4)
0.86
0.90
0.93
0.94
0.94
1.00
1.05
1.07
1.12
1.17
1.17
1.21
1.30
1.30
0.5
1.3
4.7
0.3
0.3
11.5
38.9
27.2
2.9
0.4
0.7
0.4
0.9
–
1.1
–
14.1
35.7
22.2
0.7h)
–
0.9
0.2
9.8
0.8
–
0.8
46.5
4.3
9.5
–
4g)
–
2
12
40
25
4g)h)
–
19.2
–
–
6.2
1.7
3.4
10.2
6
0.6h)
9h)
5.2
a) MD and ROD values for CCMP 1410 are given in Table 3. b) RRT¼GC Retention time relative to
c
e
cholesterol. ) HPLC Analysis; total sterol: 2.3 mg. d) Aureoumbra lagunensis; data from [10]. ) Chrys-
oreinhardia sp.; data from [18]. ) Nr¼Nematochrysopsis roscoffensis; data from [7]. g) Identified only
f
as either 24-methylenecholesterol or 24-methylcholesterol [6]. h) The configuration at C(24) was not
determined.
Fig. 3. Distribution of sterol characteristics within the class Pelagophyceae, shown with the single most
parsimonious tree (phylogram) derived from a branch-and-bound cladistic analysis using concatenated
18S rRNA and rbcL genes (tree length, 1931; 687, parsimony-informative characters). The outgroup (not
shown) was four species of the closely related Dictyochophyceae.

CHEMISTRY & BIODIVERSITY – Vol. 6 (2009)
1127
were also present, and dominated the C28 sterols 21 and 23. The C30 sterols included
both stereoisomers of 24-propylcholesterol (i.e., 4 and 5) and D22-24-propylcholesterol
(i.e., 6 and 7). These were identified using 1H-NMR of reference compounds obtained
through chemical synthesis [21]. In the C30 sterols, the 24a-configuration (i.e., 4 and 6)
was found to predominate by a factor of 2 :1. This is the first time D22-24-
propylcholesterol has been isolated from an alga.
Conclusions. – All pelagophyte algae analyzed contained one or both of the two
isomers of 24-propylidenecholesterol (i.e., 1 and 2). While this sterol has been detected
in other marine organisms such as shellfish, its presence there is thought to derive from
dietary sources. It has not been detected in other algae, and the ability to produce this
sterol appears to be limited to the Pelagophyceae, where it can be regarded as a
defining characteristic of this class.
The sterol compositions correlated well with the biological classification based on
18S rRNA and rbcL analyses (Fig. 3), and are, therefore, of high chemotaxonomic
value. All species from order Pelagomonadales contained the (Z)-24-propylidenecho-
lesterol (2), either alone or together with the (E)-isomer 1. On the other hand, 2 was
absent in all species of the order Sarcinochrysidales. Other differences between these
two orders are that D22-sterols and 24a-sterols were found in all Sarcinochrysidales
investigated, but were absent in all algae of the Pelagomonadales.
Within the Pelagomonadales, Pelagomonas calceolata contained only sterols with
nonreduced side chains (ROD¼0), and, with the exception of CCMP 1214, did not
contain the (E)-isomer 1 of 24-propylidenecholesterol. In all other algae in this order,
both isomers were present, with the (E)-isomer 1 predominating. Aureococcus
anophagefferens had both isomers, with an average (E)/(Z)-ratio (1/2) of 4.0 based on
HPLC analysis. A similar ratio was found in Pelagococcus subviridis, but this species
had much lower amounts of the 24-propylidenecholesterols than Aureococcus, where
they represented greater than 50% of the total sterols. Another difference between
these two species was that Aureococcus had low amounts (< 5%) of C29 sterols, while,
in P. subviridis, these represented nearly 50% of the total sterols. Two strains that
belong to genus Pelagococcus also had relatively low amounts of the 24-propylide-
necholesterols. CCMP 1395 closely resembled P. subviridis, but contained significant
amounts of sterol esters. CCMP 1252 had low amounts of the 24-propylidenecholes-
terols, but differed from the other Pelagococcus strains, because it did not contain
sterols with reduced side chains (ROD¼0) and had a lower MD. Unlike the other
Pelagococcus algae, CCMP 1145 is a nonflagellate, coccoid cell, much like Pelagococcus
subviridis. The fact that its sterol composition more closely resembles the latter species
than examples from its own genus suggests that the sterol composition may depend
upon the life form of the alga.
Because of the actions of oxidoreductases on the sterol side chain, the sterol
compositions in the order Sarcinochrysidales are much more complex than in the
Pelagomonadales. The species Aureoumbra lagunensis showed high amounts of (24S)-
D22-24-ethylcholesterol (stigmasterol, 16), (24R)-24-ethylcholesterol (sitosterol, 14),
cholesterol (1), and 24-propylcholesterol (4/5). These represent characteristic sterols of
this species. However, because of the prevalence of the first three sterols in terrestrial
plants and animals, only 24-propylcholesterol (4/5) would be a useful biomarker.

1128
CHEMISTRY & BIODIVERSITY – Vol. 6 (2009)
CCMP 1410, an unidentified coccoid species, displayed a nearly identical sterol
composition to Aureoumbra, as did the previously reported analysis of Nematochry-
sopsis roscoffensis [7]. The sterol composition of Sarcinochrysis marina was reported
for the first time and was found to be similar to that of Chrysoreinhardia sp. [6][22].
Compared to Aureoumbra, these did not contain nearly as much (24S)-D22-24-
ethylcholesterol (stigmasterol, 16), cholesterol (25), or 24-propylcholesterol (4/5).
An algal source for D22-24-propylcholesterol (6/7) was identified for the first time in
the unidentified alga CCMP 1410. The rare C30 sterol 26,26-dimethyl-24-methylene-
cholesterol (11) was isolated from Aureococcus anophagefferens, and the (25R)-
configuration was proposed based on biosynthetic considerations. Traces of a new
sterol, D25-24-propylidenecholesterol (3) were detected and identified. This is thought
to be an autoxidation product of 24-propylidenecholesterol (1/2).
While (Z)-24-propylidenecholesterol (2) can no longer be regarded as a specific
biomarker for the ꢀbrown tideꢁ alga Aureococcus anophagefferens, this species was
shown to have a typical (E)/(Z)-ratio (1/2) of ca. 4. Similar mixtures were only found in
Pelagococcus and a Pelagomonas species (CCMP 1145). However, because these
species do not occur in coastal waters, mixtures of these two isomers at ratios in the
range of 3–5 represent a useful biomarker for the coastal harmful algal blooms caused
by A. anophagefferens.
Financial support was provided by New York Sea Grant, and the National Oceanographic and
Atmospheric Administrationꢀs Coastal Ocean Program. R. A. A. was supported by New York Sea Grant
R/XG-5 and National Science Foundation grants 0206590, 0444418, and 0629654.
Experimental Part
General. All solvents were obtained commercially and were further purified by distillation. As a trap
for singlet O2, 2,3-dimethylbut-2-ene was added to all solvents, except the HPLC solvents, at 0.5% (v/v).
All extraction and purification steps were carried out under low light conditions, especially before the
pigments had been removed, to minimize autoxidation. Anal. TLC: aluminum-backed plates coated with
a 0.25-mm layer of silica gel 60 F254. Prep. TLC: glass-backed plates coated with a 0.25-mm layer of silica
gel 60 F254. Argentic TLC plates were prepared by wetting prep. and anal. TLC plates with sat. AgNO3/
MeOH, allowing the solvent to evaporate in the dark, and then repeating the process a second time.
HPLC: Waters 6000A pump, Waters 410 differential refractometer, and two Altex Ultrasphere ODS
columns (5 mm, 10ꢀ250 mm) in series, using either MeOH or MeCN/AcOEt/MeOH 11:4 :4 as the
mobile phase at a flow rate of 3 ml/min. GC: Hewlett-Packard 6890 gas chromatograph with a HP-5
column (cross-linked 5% methyl(phenyl)siloxane, film thickness 0.25 mm; 30 mꢀ0.25 mm), a splitless
injector at 2508, He carrier gas at 1.2 ml/min (133 kPa), a temp. gradient of 150 to 2608 at 108/min and a
FID detector at 3008. NMR Spectra: Bruker Avance 600 MHz instrument; CDCl3 was the solvent unless
otherwise specified; referenced to residual CHCl3 signals (1H: 7.26 ppm, 13C: 77.0 ppm); only selected
1H-NMR peaks (Me groups and isolated signals) are reported. MS: Hewlett-Packard 5989B mass
spectrometer operating at 70 eV.
Acetylation of sterols were preformed by treatment with Ac2O/pyridine 1:4 at 608 for 2 h. The
mixture was evaporated with a stream of N2. The crude product was purified by prep. TLC with hexane/
AcOEt 9 :1. Saponification of sterol esters was carried out by refluxing with 10% KOH/EtOH for 1 h.
The solvent was partially removed with a stream of N2, and the residue was extracted with Et2O and H2O.
After evaporation of the org. layer with a stream of N2, the residue was purified by prep. TLC (hexane/
AcOEt 2 :1).

CHEMISTRY & BIODIVERSITY – Vol. 6 (2009)
1129
Algal culture. Cultures of pelagophyte algae for GC analysis were cultured at 228 in L1 medium using
fluorescent light at ca. 100 mE mꢁ2 sꢁ1. The 200-ml cultures were harvested by filtration on glass fiber
discs, and the cells were shipped frozen with liq. N2.
The culture of Aureococcus anophagefferens (CCMP 1708) for the growth curve experiment (see
Table 4) was grown at 188 in a 8-l glass carboy containing ASP artificial seawater and modified f/2
seawater nutrients [31] using a 14 :10 h light/dark cycle at 50–75 mE·mꢁ2 ·sꢁ1. Cell growth was measured
using in vivo chlorophyll-a fluorescence and by visual cell counts under phase contrast. Volumes
containing 3ꢀ108 cells were collected at different times on the growth curve by filtration on glass fiber
discs.
Cultures of algae for analysis by HPLC were grown at 18 or 218 with aeration in 12-l glass carboys as
described above. Each culture was inoculated (10%) with cells in exponential growth phase. Cell growth
was measured using in vivo chlorophyll-a fluorescence and by visual cell counts under phase contrast.
Cells were harvested in late exponential or stationary phase by use of a continuous flow centrifuge.
Strains 1429, 1507, 1509, 1510, and 1756 were axenic. The remaining strains contained uncharac-
terized bacterial species, but no other eukaryotes.
Sterol Analysis. GC Analysis. Algae were extracted twice at r.t. with 1 ml portions of AcOEt using
mechanical shaking (Tony Microtube Mixer) for 45 min each time. The org. extracts were combined and
concentrated with a stream of N2. The contaminating pigments were removed by chromatography on a
pipet column containing Florisil (1 g; 60–100 mesh), eluting with hexane/AcOEt 19 :1 (5 ml), followed
by hexane/AcOEt 4 :1 (5 ml). The hexane/AcOEt 4 :1 fractions were concentrated to dryness with a
stream of N2, and the residue was trimethylsilylated by reacting with N,O-bis(trimethylsilyl)trifluoro-
acetamide (BSTFA) at 608 for 1 h to give a mixture of sterol TMS ethers. The samples were analyzed by
GC, and the sterols were identified by comparison of their retention times with standards. The identities
of all sterols were confirmed by GC/MS through comparison of their mass spectra with standards. The
relative proportions were determined from integration of the GC FID traces. The absolute amounts were
estimated by comparison with a known amount of an external standard (cholesterol).
HPLC Analysis. The algal pellets were extracted with AcOEt (4 ml) in 20-ml glass vials with
vigorous shaking by hand for 5 min. Additional extractions were repeated until no more sterols were
detected by anal. TLC. The solvents were evaporated with a stream of N2, and the pigments were
removed by Florisil chromatography (see above). The sterols were obtained by prep. TLC with hexane/
AcOEt 2 :1. Only strain CCMP 1395 had significant amounts of sterol esters which were saponified to
give free sterols. The sterols were fractionated by reversed-phase (RP) HPLC with MeOH. The isolated
sterol fractions were evaporated with a stream of N2 and characterized by 1H-NMR. The relative
proportions of sterols were determined from the HPLC integrals. The absolute amount of sterols in each
sample was estimated from the NMR spectrum by comparison of the signals of HꢁC(3) and HꢁC(6)
with that of a known concentration of stigmasterol using the residual CHCl3 solvent peak as a reference.
24-Propylidenecholesta-5,25-dien-3b-ol (3). 1H-NMR (CDCl3): 5.508 (t, J¼7.3, 1 H); 5.353 (m, 1 H);
4.976 (s, 1 H); 4.867 (s, 1 H); 3.524 (m, 1 H); 2.123 (quint., J¼7.5, 2 H); 1.879 (s, 3 H); 1.018 (s, 3 H); 1.018
1
(t, J¼7.4, 3 H); 1.015 (d, J¼6.5, 3 H); 0.696 (s, 3 H). H-NMR (C6D6): 5.608 (t, J¼7.3, 1 H); 5.350 (m,
1 H); 5.224 (s, 1 H); 5.015 (s, 1 H); 3.394 (m, 1 H); 2.167 (quint., J¼7.3, 2 H); 1.925 (s, 3 H); 1.081 (d, J¼
6.6, 3 H); 1.007 (t, J¼7.6, 3 H); 0.952 (s, 3 H); 0.670 (s, 3 H). EI-MS: 424 (2, Mþ, C30H48O), 409 (1),
391(1), 381(2), 314 (2), 271 (12), 159 (10), 145 (17), 105 (51), 95 (98), 93 (62), 91(64), 81(94), 79 (74),
67(96), 57 (57), 55 (100). For 2D-NMR results, see the text.
Phylogenetic Analysis. The phylogenetic analysis was conducted using a combined matrix of 18S
rRNA and rbcL gene sequences. There were ten ingroup taxa belonging to the Pelagophyceae and four
outgroup taxa (not shown in the figure) belonging to the Dictyochophyceae (Dictyocha speculum,
Apedinella radians, Pseudopedinella elastica, Rhizochromulina cf. marina). There were 3336 total
characters (unordered, equal weights), gaps were treated as missing characters, and there were 687
parsimony-informative characters. The analysis was carried out using PAUP* Version 4.0b10, and a
branch-and-bound exhaustive search was conducted. A single most parsimonious tree was recovered
(Fig. 3), and it had a length of 1931, a consistency index of 0.673, and a retention index of 0.674.

1130
CHEMISTRY & BIODIVERSITY – Vol. 6 (2009)
REFERENCES
[1] J.-L. Giner, C. Djerassi, J. Am. Chem. Soc. 1991, 113, 1386.
[2] J.-L. Giner, Chem. Rev. 1993, 93, 1735.
[3] D. R. Idler, L. M. Safe, E. F. MacDonald, Steroids 1971, 18, 545.
[4] D. R. Idler, M. W. Khalil, J. D. Gilbert, C. J. W. Brooks, Steroids 1976, 27, 155.
[5] M. Kobayashi, H. Mitsuhashi, Steroids 1975, 26, 605.
[6] M. Rohmer, W. C. M. C. Kokke, W. Fenical, C. Djerassi, Steroids 1980, 35, 219.
[7] D. Raederstorff, M. Rohmer, Phytochemistry 1984, 23, 2835.
[8] J.-L. Giner, G. L. Boyer, Phytochemistry 1998, 48, 475.
[9] C. J. Gobler, D. J. Lonsdale, G. L. Boyer, Estuaries Coasts 2005, 28, 726.
[10] J.-L. Giner, X. Li, G. L. Boyer, Phytochemistry 2001, 57, 787.
[11] R. A. Andersen, G. W. Saunders, M. P. Paskind, J. P. Sexton, J. Phycol. 1993, 29, 701.
[12] G. W. Saunders, D. Potter, M. P. Paskind, R. A. Andersen, Proc. Natl. Acad. Sci. U.S.A. 1995, 92, 244.
[13] N. Daugbjerg, R. A. Andersen, J. Phycol. 1997, 33, 1031.
[14] G. W. Saunders, D. Potter, R. A. Andersen, J. Phycol. 1997, 33, 310.
[15] N. Simon, R. G. Barlow, D. Marie, F. Partensky, D. Vaulot, J. Phycol. 1994, 30, 922.
[16] R. A. Andersen, R. R. Bidigare, M. D. Keller, M. Latasa, Deep Sea Res., Part II 1996, 43, 517.
[17] A. Z. Worden, J. K. Nolan, B. Palenik, Limnol. Oceanogr. 2004, 49, 168.
[18] G. W. Patterson, in ꢀPhysiology and Biochemistry of Sterolsꢁ, Eds. G. W. Patterson, W. D. Nes,
AOCS, Champaign, IL, 1999, p. 118.
[19] J. K. Volkman, Org. Geochem. 1986, 9, 83.
[20] D. Potter, T. C. Lajeunesse, G. W. Saunders, R. A. Andersen, Biodivers. Conserv. 1997, 6, 99.
[21] J.-L. Giner, X. Li, Tetrahedron 2000, 56, 9575.
[22] W. C. M. C. Kokke, J. N. Shoolery, W. Fenical, C. Djerassi, J. Org. Chem. 1984, 49, 3742.
[23] C. Billard, J. C. Dauguet, D. Maume, M. Bert, Bot. Mar. 1990, 33, 225.
[24] N. Gerst, B. Ruan, J. Pang, W. K. Wilson, G. J. Schroepfer Jr., J. Lipid Res. 1997, 38, 1685.
[25] W. C. M. C. Kokke, C. S. Pak, W. Fenical, C. Djerassi, Helv. Chim. Acta 1979, 62, 1310.
[26] T. Kikuchi, S. Kadota, T. Shima, Tetrahedron Lett. 1985, 26, 3817.
[27] J. L. Giner, M. P. Zimmerman, C. Djerassi, J. Org. Chem. 1988, 53, 5895.
[28] J.-L. Giner, S. P. Gunasekera, S. A. Pomponi, Steroids 1999, 64, 820.
[29] M. Vesk, S. W. Jeffrey, J. Phycol. 1987, 23, 322.
[30] J. C. Bailey, R. A. Andersen, J. Phycol. 1999, 35, 570.
[31] R. R. L. Guillard, ꢀCulture of phytoplankton for feeding marine invertebratesꢁ, in ꢀCulture of
Marine Invertebrate Animalsꢁ, Eds. W. L. Smith, M. H. Chanley, Plenum Press, New York, USA,
1975, pp. 29–60.
Received October 22, 2008