Key words cyanobacteria -
Hapalosiphon
- hapalindoles - immune modulation - lymphocytes
Abbreviations
CFSE:
carboxyfluorescein diacetate succinimidyl ester
CPT:
camptothecin
CsA:
cyclosporine A
PBMC:
peripheral blood mononuclear cells
Introduction
Cyanobacteria are an intriguing source for structurally diverse and biologically active
natural products. Especially the genera Microcystis, Nostoc, and Lyngbya or Moorea are chemically well-characterized [1 ], [2 ], [3 ], [4 ], [5 ], [6 ], [7 ]. Several strains of the family Hapalosiphonaceae produce indole alkaloids [8 ]. The major classes of these indole alkaloids include hapalindoles, ambiguines, fischerindoles,
and welwitindolinones, of which more than 80 variants have been described in the literature
[9 ], [10 ], [11 ], [12 ], [13 ], [14 ], [15 ], [16 ], [17 ], [18 ], [19 ], [20 ], [21 ], [22 ], [23 ], [24 ]. They share several common motifs: an indole or oxindole core, a cyclohexane fused
to that core, an isonitrile or isothiocyanate functional group, a chlorine substituent,
or an additional ring [25 ]. The largest group of these alkaloids are the hapalindoles. So far 30 different
hapalindoles have been reported [9 ], [10 ], [15 ], [18 ], [21 ], [26 ]. The first hapalindoles, hapalindoles A and B, were discovered in 1984 by Moore
et al. from Hapalosiphon fontinalis [9 ]. Since then, diverse hapalindoles have been isolated from various strains of the
genera Hapalosiphon, Fischerella , and Westiellopsis [9 ], [17 ], [21 ], [26 ].
Hapalindoles have a broad spectrum of biological activities, e. g., activity against
bacteria [17 ], [19 ], [21 ], [27 ], fungi [19 ], [27 ], and algae [27 ]. Studies by Doan et al. suggested inhibition of RNA polymerase and consequently
the disturbance of the protein biosynthesis as a possible mode of action for the antibacterial
activity [27 ], [28 ]. Furthermore, cytotoxic activity against normal mammalian and cancer cell lines
has been reported for various hapalindoles [21 ], [27 ]. However, no mode of action for this activity has been described yet. Finally, they
have been shown to be toxic to insects [18 ], [29 ] and vertebrates [23 ]. The insecticidal activity could be explained by sodium channel modulating activity
of the indole alkaloids [30 ]. Interestingly, inhibition of sodium channels did not lead to any cytotoxicity in
neuroblastoma cell lines. Thus, compared to insects, a different mode of action must
underlie the cytotoxicity on mammalian cells. Although various bioactivities can be
attributed to hapalindoles, modulation of the activity of human immune cells has not
yet been described. Immunomodulatory drugs are used as modifiers of the immune system
to either enhance the immune response against infectious diseases, tumors, and immunodeficiency,
or to suppress the immune reaction in organ transplants or to treat autoimmune responses.
Screening of a cyanobacteria extract collection (575 extracts) derived from strains
from all cyanobacteria orders for inhibition of T cell proliferation resulted in 35
extracts with an activity at 1 µg/mL or lower. One of the active extracts was derived
from a Hapalosiphon sp. strain. Bioassay-guided fractionation of the extract led to the isolation of
3 known hapalindoles, hapalindole A (1 ) [9 ], D (2 ), and M (3 ) [10 ], as well as 2 formamide-bearing hapalindoles that are reported here for the first
time from a natural source, namely hapalindole A-formamide (4 ) [31 ] and hapalindole J-formamide (5 ) [32 ]. Here, we report the isolation, structure elucidation, and the investigation of
the immunomodulatory properties of the hapalindole derivatives isolated from Hapalosiphon sp. CBT1235.
Results and Discussion
Our initial screening of a cyanobacteria extract collection for immunomodulating activity
on human immune cells resulted in 35 extracts with an activity at a concentration
of 1 µg/mL or lower. Nine out of the 11 most potent extracts were derived from cyanobacteria
of the genus Nostoc . Dereplication of the natural products in these Nostoc strain extracts showed the presence of the well-known cytotoxic compound cryptophycin-1
in most of these strains [33 ], [34 ], [35 ], [36 ], so they were excluded from further investigation. Chromatographic evaluation of
the remaining extracts by HPLC-DAD/MS showed that few of them featured prominent peaks.
In the HPLC-DAD chromatogram of an Hapalosiphon sp. extract, several prominent peaks were observed; thus, this strain was selected
for follow-up work. T cell proliferation was inhibited by this extract with an IC50 of 0.11 µg/mL ([Fig. 1 a ]). At the same time, a trend to an increased amount of apoptotic cells at an extract
concentration of 10 µg/mL could be observed (statistically non-significant, [Fig. 1 b ]). No induction of necrosis has been observed (data not shown). Microfractionation
of the extract and subsequent bioassays showed that indeed the major compounds observable
in the chromatogram were responsible for the bioactivity and led to the significant
retardation of T cell proliferation (Fig. 1S a – c in the Supporting Information).
Fig. 1 Effects of Hapalosiphon sp. CBT1235 extract on T cell proliferation and apoptosis
induction. Primary human lymphocytes were cultured in the presence of medium (NC)
or were stimulated with anti-human CD3 and anti-human CD28 mAb (PC; 100 ng/mL). Activated
T cells were further incubated with cyclosporine A (CsA; 5 µg/mL), camptothecin (CPT;
30 µg/mL), or different concentrations of the Hapalosiphon sp. CBT1235 extract. Cell division analysis was carried out using CFSE staining and
flow cytometry. Levels of apoptosis were determined using flow-cytometric analysis
of annexin V-stained cells. a Effects of Hapalosiphon sp. CBT1235 extract on lymphocyte proliferation. b Induction of apoptosis by Hapalosiphon sp. CBT1235 extract. Data of 4 (a ) or 3 (b ) independent experiments are presented as mean ± SD in relation to stimulated T cells
(PC = 100%). Asterisks indicate significant differences from PC controls (**p < 0.01,
***p < 0.001).
Five compounds present in the active micro-fraction were isolated from the extract
by semi-preparative HPLC (1 –5 ). 1 H NMR spectra of the pure compounds displayed typical signals of hapalindoles. The
molecular formulas were deduced from HRESIMS data. The structures of 1 –5 were elucidated based on 1D and 2D NMR data. NMR and MS spectral data of compounds
1 –3 matched the published data for hapalindole A (1 ), hapalindole D (2 ), and hapalindole M (3 ), respectively (structures of all isolated compounds, see [Fig. 2 ]). NOESY experiments as well as the determination of specific rotation values confirmed
the absolute configuration of compounds 1 –3 as originally described in detail by Moore et al. [10 ]. Hapalindole A (1 ) was the main compound isolated from Hapalosiphon sp. CBT1235.
Fig. 2 Chemical structures of hapalindoles 1 –5 isolated from Hapalosiphon sp. CBT 1235.
Compounds 4 and 5 were hapalindole derivatives not yet described as natural products. Compound 4 was isolated as a brown oil. HRESIMS showed an [M + H]+ ion with m/z 357.1723, corresponding to the molecular formula C21 H25 N2 OCl (calcd. 357.1728, Δ 1.54 ppm). The 1 H-spectrum of 4 was almost identical with the respective spectrum of 1 but with 2 additional peaks in the low-field region (8.04 ppm and 8.57 ppm) and a
deshielded H-11 (4.84 ppm) compared to 1 (4.37 ppm), suggesting C-11 to be substituted with a formamide moiety ([Table 1 ]). Key HMBC and COSY correlations confirmed the planar structure ([Fig. 3 ]). Evaluation of the NOESY spectrum showed the relative configuration of 4 to be the same as hapalindole A with the formamide group being attached axially to
C-11 (strong correlations of the N-formamide proton with H-10, H-13 and H-15) [9 ]. Comparing the specific rotation values with already published data confirmed our
assignment of the absolute configuration of 4 [37 ]. Therefore, compound 4 was identified as the new natural product hapalindole A-formamide.
Table 1 1 H and 13 C NMR assignments for 4 and 5 (600 MHz for 1 H, 150 MHz for 13 C, DMSO-d6 ). 13 C chemical shifts for 5 were extracted from the 2D spectra.
Position
4
5
δH (J in Hz)
δC
δH (J in Hz)
δC
1
10.81, s
indole N
10.69, s
indole N
2
7.24, t (1.8)
120.5
7.13, m
119.7
3
110.6
111.8
4
137.1
138.0
5
6.81, d (7.0)
112.3
6.79, br d (7.02)
111.9
6
7.02, t (7.6)
121.9
6.99, t (7.4)
121.5
7
7.14, d (7.9)
108.7
7.10, d (8.0)
108.1
8
133.3
133.4
9
124.0
124.4
10
3.29
36.7
3.30
36.3
11
4.84, m
54.6
4.59, s
51.6
12
44.6
39.1
13 ax
4.61, dd (12.4, 4.0)
65.3
1.74, td (13.2, 3.5)
30.2
13 eq
1.27, m
14 ax
1.26, m
31.3
0.85, m
19.3
14 eq
1.96, td (13.0, 3.4)
1.60, br d (10.5)
15
2.11, br td (13.1, 3.8)
44.4
1.87, m
43.4
16
37.4
37.3
17
1.46, s
24.3
1.43, s
24.5
18
1.04, s
31.9
1.06, s
31.5
19
0.85, s
19.6
0.74, s
26.4
20
5.82, dd (17.1, 11.0)
144.2
5.87, m
147.7
21
5.08, m
114.0
4.87, m
109.9
22
8.57, br d (9.9)
N
8.23, br d (9.16)
N
23
8.04, d (1.37)
159.7
8.00, br d (1.45)
159.6
Fig. 3 1 H-1 H COSY (bold connections) and selected HMBC correlations (arrows) of 4 and 5 .
The main difference with 5 compared to 1 and 4 was the absence of the chlorine. Compound 5 was also isolated as a brown oil. HRESIMS showed an [M + H]+ ion with m/z 323.2114, corresponding to the molecular formula C21 H26 N2 O (calcd. 323.2118, Δ 1.08 ppm). Evaluation of the NMR spectra confirmed the presence
of a formamide group at C-11 of the hapalindole backbone, as well ([Fig. 3 ] and [Table 1 ]). The absolute configuration of 5 matches the one of hapalindole J, the nonchlorinated variant of hapalindole A (1 ), and has been confirmed by comparing the NOESY correlations and the specific rotation
values with those originally published by Moore et al. [10 ]. Additional NOESY correlations for the axial formamide could be observed ([Fig. 4 ]). Compound 5 was thus confirmed to be the new natural product hapalindole J-formamide.
Fig. 4 Selected NOESY correlations (arrows) of compounds 4 (R = Cl) and 5 (R = H).
Coupling constants between H-22 (NH) and H-23 were found to be 1.37 Hz and 1.45 Hz
for hapalindole A-formamide (4 ) and J-formamide (5 ), respectively, indicating a cis conformation of the amide bond in both compounds [31 ].
Hapalindole formamides have been described as intermediates during the total synthesis
of hapalindoles, as derivatization products with formic acid, and they can also form
during storage in d -chloroform [13 ], [31 ], [32 ], [38 ]. Here, we report for the first time the isolation of 2 hapalindole formamides as
natural products from Hapalosiphon sp. CBT1235. 4 and 5 could readily be detected by HPLC-MS in a fresh Hapalosiphon sp. CBT1235 extract that has not been in contact with formic acid or other acids
(Fig. 2S , Supporting Information), ruling out the possibility that the isolated formamides
are processing or isolation artifacts.
All isolated hapalindoles showed significant effects in the T cell proliferation assay
([Fig. 5 ]). Hapalindole A (1 ) displayed the highest antiproliferative activity with an IC50 of 1.56 µM. Hapalindoles D (2 ) and M (3 ) showed a weaker activity and suppressed proliferation with an IC50 value of 27.15 µM for hapalindole D. Hapalindoles A-formamide (4 ) and J-formamide (5 ) again showed lower activity ([Fig. 5 ]). Except for hapalindole J-formamide, all hapalindoles induced T cell apoptosis
at highest concentrations ([Fig. 6 ]).
Fig. 5 Effects of different hapalindoles on T cell proliferation. Primary human lymphocytes
were cultured in the presence of medium (NC) or were stimulated with anti-human CD3
and anti-human CD28 mAb (PC; 100 ng/mL). Activated T cells were further incubated
with cyclosporine A (CsA; 5 µg/mL) or different concentrations of 1 –5 . Cell division analysis was carried out using CFSE staining and flow cytometry. Data
of 3 independent experiments are presented as mean ± SD in relation to stimulated
T cells (PC = 100%). Asterisks indicate significant differences from PC controls (**p < 0.01,
***p < 0.001).
Fig. 6 Levels of T cell apoptosis after treatment with different hapalindoles. Primary human
lymphocytes were cultured in the presence of medium (NC) or were stimulated with anti-human
CD3 and anti-human CD28 mAb (PC; 100 ng/mL). Activated T cells were further incubated
with camptothecin (CPT; 30 µg/mL) or different concentrations of 1 –5 . Levels of apoptosis were determined using flow-cytometric analysis of annexin V-stained
cells. Data of 3 independent experiments are presented as mean ± SD in relation to
stimulated T cells (PC = 100%). Asterisks indicate significant differences from PC
controls (**p < 0.01, ***p < 0.001).
Although relevant amounts of T cell apoptosis were detected, retardation of T cell
proliferation was measurable at up to 10-fold lower concentrations. The most potent
compound, hapalindole A, was effective at a concentration of 3.0 µM without showing
cytotoxic effects. Therefore, potentially a therapeutic range for an application as
an anti-inflammatory remedy is given. The isonitrile functional group seems to be
crucial for the antiproliferative bioactivity of the hapalindoles, as the formamide
derivatives possess a weaker activity. This is in agreement with previous findings,
where a reduced antifungal and antibacterial activity of the hapalindole formamides
compared to their isonitrile and isothiocyanate counterparts has been reported [31 ]. Our results, therefore, strengthen the key role of this functional group in regard
to the bioactivity of members of the hapalindole family. Our work shows that antiproliferative
effects on human T cells are more pronounced than toxicity on human T cells, adding
this activity to the wide range of reported bioactivities of the hapalindoles. The
mode of action in human immune cells remains to be investigated. Moreover, concerning
the fact that hapalindoles have been described to be neurotoxic metabolites [30 ], detailed studies which further discriminate toxicity and immunomodulatory effects
need to be carried out in order to estimate whether or not a development as immunomodulatory
drugs would be possible.
Materials and Methods
General experimental procedures
HRESIMS data were obtained using a Q Exactive Plus mass spectrometer (Thermo Fisher
Scientific) coupled to an UltiMate 3000 HPLC system (Thermo Fisher Scientific). Semi-preparative
HPLC was conducted on an UltiMate 3000 HPLC system (Thermo Fisher Scientific). NMR
spectra were either recorded at 600 MHz (1 H frequency) on a Bruker AV-III spectrometer using a cryogenically cooled 5 mm TCI-triple
resonance probe equipped with 1-axis self-shielded gradients at 300 K or at 400 MHz
(1 H frequency) on an Agilent DD2 spectrometer. Spectra were referenced indirectly. If
1D spectra were not separately recorded, 13 C chemical shifts were extracted from the 2D spectra (compounds 1 –3, and 5 ).
Cyanobacterial material
Hapalosiphon sp. CBT1235 was taxonomically identified as Hapalosiphon sp. on the basis of its morphology. The strain has kindly been provided by Algenol
Biotech Inc. (USA) as ABCC 804 and is deposited in the culture collection of the Cyano
Biotech GmbH (Germany) under the accession number CBT 1235. The strain was cultivated
in BG11 medium [39 ] at 28 °C under continuous light (60 – 80 µmol m2 · s−1 ) in 20 L scale photobioreactors and harvested semi-continuously over a period of
several weeks.
Extraction, bioassay-guided fractionation, and isolation of compounds 1 –5
Cyanobacterial cells were harvested and lyophilized. Seven grams dry biomass were
suspended in 100 mL 50% methanol in water (v/v), treated with an ultrasonication rod
(Bandelin), and extracted on a shaker for 30 min at room temperature. After centrifugation
(20 min, 10 800 g ), the biomass was extracted using 100 mL 80% methanol (v/v). The solutions were combined
and dried under reduced pressure, yielding 0.4 g of biomass extract. For micro-fractionation,
4 mg of extract were suspended in 0.1 mL acetonitrile and separated into 23 fractions
by HPLC using a C18 column (250 × 4.6 mm, 5 µm, 100 Å, Luna, Phenomenex) and 5 – 100% acetonitrile-water
as the mobile phase at 1 mL/min in 23 min. All fractions were tested for inhibitory
activity on T cell proliferation.
For preparative isolation of the active compounds, the remaining extract was dissolved
in acetonitrile and fractionated by semi-preparative HPLC using a phenyl-hexyl column
(250 × 10 mm, 5 µm, 100 Å, Luna, Phenomenex) and 60 – 80% acetonitrile-water as the
mobile phase at 4.7 mL/min in 25 min to afford 11 fractions. Fraction 1 (tR 6.8 min) was further purified by an additional round of semi-preparative HPLC using
a phenyl-hexyl column (250 × 100 mm, 5 µm, 100 Å, Luna, Phenomenex) and 40 – 47% acetonitrile-water
as the mobile phase at 9.5 mL/min in 20 min to afford hapalindole A-formamide (4 , 5.0 mg, tR 16.5 min) and hapalindole J-formamide (5 , 2.0 mg, tR 15.1 min). Fraction 6 (tR 15.2 min) was further purified by semi-preparative HPLC using a pentafluorophenyl
column (250 × 100 mm, 5 µm, 100 Å, Luna, Phenomenex) and 45 – 63% acetonitrile-water
as mobile phase at 9.5 mL/min in 30 min yielding hapalindole A (1 , 16.1 mg, tR 18.4 min). No further purification was needed for fraction 8 (tR 19.2 min) and fraction 10 (tR 22.4 min), which corresponded to hapalindole D (2 , 5.3 mg) and hapalindole M (3 , 6.4 mg), respectively.
Hapalindole A (
1
): [α ]D
23 − 64.2° (CH2 Cl2 , c 1.2); HRESIMS (positive ion mode): m/z 339.1616 [M + H]+ ; NMR spectra, see Fig. 3S, 4S .
Hapalindole D (
2
): [α ]D
23 + 45.2° (CH2 Cl2 , c 0.31); HRESIMS (positive ion mode): m/z 337.1727 [M + H]+ ; NMR spectra, see Fig. 5S, 6S .
Hapalindole M (
3
): [α ]D
23 − 6.7° (CH2 Cl2 , c 0.45); HRESIMS (positive ion mode): m/z 337.1727 [M + H]+ ; NMR spectra, see Fig. 7S, 8S .
Hapalindole A-formamide (
4
): brown oil; [α ]D
23 − 56.2° (CH2 Cl2 , c 0.45); UV (MeOH) λmax (log ε ): 223 (4.38), 282 (3.60), 292 (3.52) nm; HRESIMS (positive ion mode): m/z 357.1723 [M + H]+ (calcd. for C21 H25 N2 OCl, 357.1728); NMR data, see [Table 1 ]; NMR spectra Fig. 9S–14S .
Hapalindole J-formamide (
5
): brown oil; [α ]D
23 + 12.2° (CHCl3 , c 0.9); UV (MeOH) λmax (log ε ): 224 (4.24), 282 (3.50), 292 (3.41) nm; HRESIMS (positive ion mode): m/z 323.2114 [M + H]+ (calcd. for C21 H26 N2 O, 323.2118); NMR data, see [Table 1 ]; NMR spectra Fig. 15S–19S .
Ethics statement
Written informed consent was obtained from patients prior to blood donation for research
purposes. All experiments conducted on human material were approved by the ethics
committee of the University Freiburg (55/14; February 11th, 2014).
Preparation and cultivation of human immunocompetent cells
Human PBMC were isolated from the blood of adult donors obtained from the Blood Transfusion
Centre (University Medical Center Freiburg). Venous blood was centrifuged on a LymphoPrep
gradient (density: 1.077 g/cm3 , 20 min, 500 g, 20 °C; Progen). Afterwards, cells were washed twice with medium and
cell viability and concentration was determined using the trypan blue exclusion test.
Cells were cultured in Roswell Park Memorial Institute (RPMI) 1640 medium (Invitrogen)
supplemented with 10% heat-inactivated fetal bovine serum (GE Healthcare), 2 mM L-glutamine,
100 U/mL penicillin and 100 U/mL streptomycin (Invitrogen) at 37 °C in a humidified
incubator with a 5% CO2 /95% air atmosphere. PBMC were additionally stimulated with anti-human CD3 (clone
OKT3) and anti-human CD28 (clone 28.6) mAb (100 ng/mL; both from eBioscience). Incubation
was carried out as indicated in the figure captions in the presence of medium alone,
camptothecin (CPT; 30 µg/mL; Tocris), cyclosporine A (CsA, 5 µg/mL, Sandimmun 50 mg/mL,
Novartis), or different concentrations of the Hapalosiphon sp. CBT1235 extract (0.01, 0.1, 1, 10 µg/mL), fractions (semi-quantitative using
dilutions 1 : 250, 1 : 500, 1 : 1000, 1 : 2000) or various hapalindoles (0.1, 0.3,
1, 3, 10, 30 µg/mL).
Cell division tracking using CFSE
PBMC were harvested, washed twice in cold PBS (Invitrogen) and resuspended in PBS
at a concentration of 5 × 106 cells/mL. CFSE (5 mM; Sigma) was diluted 1/1000 and incubated for 10 min at 37 °C.
The staining reaction was stopped by washing twice with complete medium. PBMC were
activated as described above and cultured with the Hapalosiphon sp. CBT1235 extract, fractions, hapalindoles, or DMSO as a solvent control for 72 h.
Cell division progress was analyzed from 3 independent experiments with a BD FACSCalibur
flow cytometer using BD CellQuest Pro Software.
Determination of apoptosis and necrosis using annexin V and propidium iodide staining
Cells were cultured as described, and levels of apoptosis and necrosis were determined
using annexin V-FITC apoptosis detection kit (eBioscience) according to the manufacturerʼs
instructions. After annexin V and propidium iodide staining, cells were analyzed by
flow cytometry. CPT and Triton-X 100 (0.5%; Sigma-Aldrich) were used as positive controls
for apoptosis and necrosis, respectively.
Data analysis
For statistical analysis, data were processed with Microsoft Excel and SPSS software
(IBM, Version 22.0). Data were adjusted in relation to untreated control cells (= 100% ± SD)
and values are presented as mean ± SD. Statistical significance was determined by
one-way ANOVA followed by Dunnettʼs post hoc pairwise comparisons. P values < 0.05 were considered as statistically significant
(* p < 0.05, ** p < 0.01, *** p < 0.001). IC50 and EC50 values were determined using GraphPad Prism 6.
