Bioassay-Guided Isolation of Anti-Inflammatory Constituents of the Subaerial Parts of Cyperus articulatus (Cyperaceae)

Based on data from a previous ethnobotanical study in northern Angola, phytochemical investigations into the methanolic rhizomes and roots extract of Cyperus articulatus, monitored by in vitro assays, resulted in the recovery of 12 sesquiterpenes, 3 stilbenes, 2 phenolic acids, 1 monoterpene, and 1 flavonoid. Among them, 14 compounds were isolated for the first time from this species. Their inhibitory potential against nitric oxide (NO) production, as well as inducible nitric oxide synthase (iNOS) and cyclooxygenase-2 (COX-2) expression, was evaluated in LPS-treated J774A.1 murine macrophages. Especially, both stilbene dimer trans-scirpusin B and trimer cyperusphenol B showed promising inhibitory activity against the production of the inflammatory mediator, NO, in a concentration-dependent manner (10–1 µM). The obtained data are the first results confirming the anti-inflammatory potential of C. articulatus and support its indigenous use as a traditional remedy against inflammation-related disorders.


Introduction
The sedge plant Cyperus articulatus (L.) (Cyperaceae), commonly known in the Amazonia region as Piri-Piri, is native to the subtropical areas of the world and spreads throughout African countries such as Cameroon, Nigeria, and Angola to the Amazonia region and is common in wet habitats [1]. The decoction of its horizontal rhizomes and roots is used to gain traditional remedies to treat a number of diseases such as epilepsy, migraine, cough, fever, gastrointestinal disorders, malaria, and tuberculosis [2,3]. The rhizome essential oils are acclaimed for their multiple beneficial effects as anticonvulsant [4,5], antimalarial [6], antimicrobial [7][8][9], antiproliferative [10], and antioxidant [11]. The chemical composition of the apolar extracts from rhizomes and the roots of C. articulatus has been exhaustively investigated by GC-MS, and this has revealed the identification of terpenoids. The biological activities described are attributed mostly to sesquiterpenes, which represent the major components present [12][13][14]. According to a previous ethnobotanical survey in the province of Uíge in northern Angola, 6 informants out of 62 recommended the underground plant material of C. articulatus for anti-inflammatory indications, such as stomach pains, backache, foot infection, and toothache [15]. Pharmacological investigations of its methanol extract in murine macrophages showed promising inhibitory effects against NO production (96.24-36.09% inhibition) and COX-2 expression (85.37-42.45% inhibition) for six different concentrations, between 100-1.25 µg mL −1 [16]. Reports on the inflammation-related use of the rhizomes and roots opened a new research focus for C. articulatus. The increased

Results
The anti-inflammatory effect of the methanolic rhizomes and roots extract of C. articulatus was confirmed in a previous ethnopharmacological report [16]. Due to the promising pharmacological activity of the crude extract, the isolation and identification of its active principles are described in the present study. In preliminary screening, the obtained PE, DEE, PRC, EtOAc, 1-BuOH, and H 2 O fractions were analyzed for their inhibition of NO 2 − release in LPS-stimulated J774A.1 murine macrophages at non-antiproliferative concentrations of 10, 5, and 1 µg mL −1 (Figure 1). The DEE fraction exhibited the highest anti-inflammatory activity and was selected, along with the also active PE fraction, for further classical activity-guided isolation. The latter contains the essential oil of the rhizomes, which was acclaimed for multiple beneficial effects, as described above; however, anti-inflammatory reports on the essential oil from C. articulatus are missing in the literature. The antiproliferative results indicated a significant activity for the PE fraction at concentrations above 12.5 µg mL −1 [21] ( Figure S1). Separation of the highly active DEE fraction by NP-MPLC, with its HPLC-UV chromatogram shown in Figure 2B, yielded subfractions A-R. The HPLC-UV chromatograms of subfractions A-R are presented in the Supplementary Information (Figure S2). Their inhibitory effects on NO production and COX-2 and iNOS expression in LPS-stimulated macrophages were analyzed at lower concentrations (5, 1, and 0.1 µg mL −1 ) to evaluate their concentration-dependent activity (results shown in Table 1). The potential antiproliferative activities of the obtained subfractions were determined using the MTT assay ( Figure S3). The criterion for including a subfraction in one of the subsequent studies was initially a requirement of at least 40% inhibition among one of the three pro-inflammatory parameters, as the positive controls used yielded an inhibition of approximately 44% at a concentration of 1 µM, respectively. Considering the available amount and purity of the samples, subfractions E, F, H, M, N, and P were finally selected for detailed phytochemical and pharmacological investigations. Subfraction E showed promising inhibitory effects for NO production (31.33% at 5 µg mL −1 ), iNOS expression (33.67% at 1 µg mL −1 ) and COX-2 expression (43.00% at 5 µg mL −1 ). Subfraction F was of interest due to an antiproliferative effect of 15.33% on the J774A.1 murine macrophage at a concentration of 5 µg mL −1 . Subfractions H and N revealed 32.33% and 57.67% inhibition for NO production in the macrophages stimulated by LPS; subfraction M exhibited a 35.0% inhibition during the iNOS expression inhibition assay, and subfraction P suppressed COX-2 expression by 40.33% at respective concentrations of 5 µg mL −1 vs. LPS alone. The major constituents of the selected subfractions, and the second most active PE fraction, were isolated as described in 4.4. The HPLC-DAD and GC analysis of the PE fraction displayed a rather complex composition with similar molecular ion peaks and an identical, expected polarity for the main compounds 1-6 at m/z 219.1-223.0 in the positive mode according to HPLC-ESI-MS. Thus, the compounds of the PE fraction were isolated without any intermediate pharmacological testing carried out on the subfractions, as described in Section 4.4.1.
Molecules 2022, 27, 5937 3 of 15 Figure 1. Inhibitory effects of the methanolic extract (Extract) and its subfractions (Petroleum ether, Diethyl ether, Precipitate, Ethyl acetate, 1-Butanol, and Water) obtained from the methanolic rhizomes and roots extract of C. articulatus on LPS-stimulated J774A.1 murine macrophages, which were incubated with subfractions at concentrations of 10, 5, and 1 µ g mL −1 , followed by measurement of the NO production vs. LPS ± SEM (n = 3). L-NAME was used as positive control (PC) at a concentration of 1 µ M.
Separation of the highly active DEE fraction by NP-MPLC, with its HPLC-UV chromatogram shown in Figure 2B, yielded subfractions A-R. The HPLC-UV chromatograms of subfractions A-R are presented in the Supplementary Information ( Figure S2). Their inhibitory effects on NO production and COX-2 and iNOS expression in LPS-stimulated macrophages were analyzed at lower concentrations (5, 1, and 0.1 µ g mL −1 ) to evaluate their concentration-dependent activity (results shown in Table 1). The potential antiproliferative activities of the obtained subfractions were determined using the MTT assay ( Figure S3). The criterion for including a subfraction in one of the subsequent studies was initially a requirement of at least 40% inhibition among one of the three pro-inflammatory parameters, as the positive controls used yielded an inhibition of approximately 44% at a concentration of 1 µ M, respectively. Considering the available amount and purity of the samples, subfractions E, F, H, M, N, and P were finally selected for detailed phytochemical and pharmacological investigations. Subfraction E showed promising inhibitory effects for NO production (31.33% at 5 µ g mL −1 ), iNOS expression (33.67% at 1 µ g mL −1 ) and COX-2 expression (43.00% at 5 µ g mL −1 ). Subfraction F was of interest due to an antiproliferative effect of 15.33% on the J774A.1 murine macrophage at a concentration of 5 µ g mL −1 . Subfractions H and N revealed 32.33% and 57.67% inhibition for NO production in the macrophages stimulated by LPS; subfraction M exhibited a 35.0% inhibition during the iNOS expression inhibition assay, and subfraction P suppressed COX-2 expression by 40.33% at respective concentrations of 5 µ g mL −1 vs. LPS alone. The major constituents of the selected subfractions, and the second most active PE fraction, were isolated as described in 4.4. The HPLC-DAD and GC analysis of the PE fraction displayed a rather complex composition with similar molecular ion peaks and an identical, expected polarity for the main compounds 1-6 at m/z 219.1-223.0 in the positive mode according to HPLC-ESI-MS. Thus, the compounds of the PE fraction were isolated without any intermediate pharmacological testing carried out on the subfractions, as described in Section 4.4.1.

Discussion
The COX-2 activity of fraction E, obtained at a concentration of 5 µg mL −1 , was lost during the isolation processing of compounds 12-14. In contrast to data from the literature, compound 12 showed a strong in vitro inhibition of NO production in the used assay at concentrations of 10 and 5 µM. The inhibitory potential of compound 13 was influenced by high antiproliferative effects that were revealed in the MTT assay [48]. Compound 14 suppressed iNOS expression at concentrations of 10 and 5 µM by 40.32% and 32.25%, respectively [49]. The antiproliferative properties of fraction F were not detectable for its isolated major component, sesquiterpene 15. The in vitro anti-inflammatory activity of 4R/4S-4hydroxy-1,10-seco-muurol-5-ene-1,10-dione is reported here for the first time, indicating moderate activities against NO production (40.27%) and iNOS expression (35.48%) at the highest test concentration of 10 µM. Although trans-sobrerol (18) was isolated from fraction N and can be found in the subaerial parts of other Cyperus species, such as Cyperus longus (L.) [50] or Cyperus rotundus L. [51], the use of extraction solvents may cause the formation of compound 18 from α-pinene oxide [52]. This may be a reason for the loss of activity during the isolation process, which occurred for subfraction N, and the observed cytotoxicity of its isolated compound 18 up to the lowest test concentration of 1 µM (Table S2). In fraction M, the isolated stilbene dimer 17 (trans-scirpusin B, a dimer of piceatannol) was the dominant constituent, and it strongly attenuated the NO production in LPS-stimulated murine macrophages at all tested concentrations (46.50-31.65% inhibition vs. LPS alone). Additionally, COX-2 expression was reduced by 36.03% at a concentration of 10 µM. Apart from compound 17, stilbene trimer 19 (cyperusphenol B, a trimer of piceatannol) was obtained as a highly active constituent from subfraction P. Treatment with compound 19 inhibited the NO production to a similar extent (as the tested stilbene dimer) at all tested concentrations (49.63-30.33% inhibition vs. LPS alone). The discrepancy in activity from the stilbene oligomers to the respective positive control was minimized at a test concentration of 1 µM. According to HPLC-DAD analysis, at 205 nm, stilbene oligomers 17 and 19 represent major components of the DEE fraction (10.2% and 8.4% of total integral, respectively) and are most likely the key contributors to the determined in vitro anti-inflammatory activity of the DEE fraction. Stilbene monomer 16, isolated from subfraction H, revealed lower inhibitory effects on all tested pro-inflammatory parameters compared to stilbene dimer 17 and trimer 19. Piceatannol (16) can be considered a well-known anti-inflammatory agent, showing a stronger in vitro inhibition of NO and PGE 2 production (as well as NF-κB activation in LPS-stimulated macrophages) than resveratrol [53]. Anti-inflammatory reports for compounds 17 and 19 are missing in the literature and are less abundant for other stilbene oligomers [54]. α-Viniferin, a trimer of resveratrol, has been reported to show inhibitory effects on NO and COX-2 production in LPS-activated RAW 264.7 cells (IC 50 < 10 µM) [55]. A previous report assessed the effect of some stilbene oligomers to inhibit NO production in LPS-activated murine microglia BV-2 cells, indicating the partially higher inhibition potential of stilbene dimers and trimers over resveratrol [56]. As it appears from the results of the performed assays, the pharmacological activities of the DEE fraction and its compounds isolated are mostly mediated by the identified stilbene oligomers.
Compounds isolated from the DEE fraction exhibited the highest potency for inhibiting NO production. Thus, inhibition assays for NO production in LPS-stimulated J774A.1 murine macrophages were only carried out at three different concentrations (20, 10, and 5 µM) to evaluate the anti-inflammatory activities of the 11 identified sesquiterpenes of the second most active PE fraction. Unfortunately, the sesquiterpenes were almost completely inactive and showed superior antiproliferative effects at the tested concentrations (Table S1). As compounds 7 and 8 are isomers of humulene dioxide, only compound 7, available in a larger amount, was included in the assay and was identified as a very slight NO production inhibitor (31.50% at 20 µM vs. LPS alone). However, the anti-inflammatory effects of various essential oils, with a predominance of sesquiterpenes, have been reported in the literature. On the one hand, caryophyllene oxide (29.5%) (2), humulene epoxide II (4.2%) (3), and mustakone (1.2%) (4) represent the main constituents of the essential oil obtained from the leaves of Allophylus edulis Radlk., acclaimed for its multiple anti-inflammatory capacities in mouse paw oedema, cold allodynia, and hyperalgesia [57]. The essential oil of Algerian Cyperus rotundus L. is rich in sesquiterpenes, such as humulene epoxide II (21.3%) (3) and caryophyllene oxide (13.3%) (2), and reduces inflammation by inhibiting 5-lipoxygenase (5-LO) and leukotriene A4 hydrolase (LTA4H) [58]. On the other hand, NO production in BV-2 microglial cells was suppressed only slightly by kobusone (IC 50 > 85 µM) (5) [59], and no inhibition effect on inflammatory cytokine (IL-6) release in murine macrophages could be noted [20].
All in all, the inhibitory activity of the PE fraction of the methanolic extract of C. articulatus on NO release in LPS-stimulated J774A.1 murine macrophages got lost during the fractionation process. The overall anti-inflammatory activity of the PE fraction could result from the potential additive or synergistic effects of sesquiterpenes or the effects of non-identified constituents. The beneficial interactions between multiple agents might be a pure summation effect, causing diminished biological activity in the isolated sesquiterpenes on their own, as observed in the NO production inhibition assay [60]. The data presented identified trans-scirpusin B (17) and cyperusphenol B (19) as the most active constituents in the methanol extract from the rhizomes and roots of C. articulatus. Interestingly, both compounds responsible for the observed biological activity were identified by HPLC-DAD and HPLC-ESI-MS and also in the PRC fraction of the methanolic extract. In comparison to the PE and DEE fractions, its inhibition of NO production was slightly lower in the preliminary screening model. According to HPLC-DAD analysis at 205 nm, it is likely that the dominant lipophilic area at the end of its chromatogram attenuated the activity of this fraction. Moreover, the HPLC-UV chromatogram of the PRC fraction was almost identical to that of the DEE fraction of the methanolic extract of the rhizomes and roots of Scirpoides holoschoenus L., except for the lipophilic region at the end ( Figure S5). In comparison to the published data [47], the HPLC-DAD and HPLC-ESI-MS analysis indicated the presence of two stilbene dimers: scirpusin A (21) and cassigarol E (26), five stilbene trimers: cyperusphenol D (20), passiflorinol A (22), cyperusphenol A (24), mesocyperusphenol A (25), and passiflorinol B (23), and one ferulic acid derivative: smiglaside C (27) in the PRC fraction of the methanolic extract of C. articulatus, in addition to 2R/2S dihydroluteolin (14), piceatannol (16), trans-scirpusin B (17), and cyperusphenol B (19), which were already isolated and tested. Finally, the PRC fraction, containing many potentially active stilbene oligomers, might be interesting for further investigations as well.

Solvent and Reagents
VWR International (Darmstadt, Germany) supplied the reagents and solvents of an analytical grade. Solvents used for HPLC analysis were purchased from Merck (Darmstadt, Germany). A Sartorius Arium 611 UV water purification system (Sartorius AG, Göttingen, Germany) provided ultrapure water for the HPLC analysis. NMR deuterated solvents: acetone-d 6 , chloroform (chloroform-d), and methanol (MeOH-d 4 ), were obtained from Euriso-top SAS (Saint-Aubin Cedex, France).

General Experimental Methods
Fractionation and isolation steps were carried out using both column chromatography systems, Sephadex LH-20 (Pharmacia Biotech, Uppsala, Sweden) and silica gel 60 (Merck KGaA, Darmstadt, Germany), as well as a Reveleris X2 MPLC system (Büchi, Flawil, Switzerland) with commercial prepacked columns. TLC was performed on precoated silica gel 60 F 254 plates (Merck, Darmstadt, Germany, 1 mm thickness). Chromatographic experiments were performed using a HP1100 system (Agilent, Waldbronn, Germany), equipped with an auto-sampler, a DAD, and column thermostat. For HPLC-ESI-MS purposes, the system was hyphenated to an Esquire 3000 plus ion trap (Bruker Daltonics, Bremen, Germany) utilizing electrospray ionization (ESI). LC parameters for the petroleum ether fraction of the methanolic extract: stationary phase: YMC-Pack Pro-C 18  Semi-preparative HPLC experiments, as final isolation steps, were taken on a 1260 Infinity II (Agilent, Waldbronn, Germany) and a Dionex UltiMate 3000 (Thermo Fisher scientific Inc., Waltham, MA, USA), both equipped with an auto-sampler, a DAD, a column thermostat, and a fraction collector.

Plant Material
The fresh rhizomes and roots of Cyperus articulatus were purchased from the village Kilomosso, in the province of Uíge, in northern Angola in February 2018 (S 7 • 38 31.0 , E 15 • 00 21.9 ). The plant material was dried immediately after collection in a drying cabinet at 40 • C (HTD 100 Bench Top, LinTek) and stored in bags until extraction. Voucher specimens are deposited at the Herbarium Dresdense (DR), Institute of Botany, TU Dresden (voucher no. DR050752, DR051629).

Extraction and Isolation
The dried and milled rhizomes and roots of C. articulatus (239 g) were extracted with 125 mL methanol in the ultrasonic bath for 10 min. Subsequently, the filtrate was evaporated to dryness using a vacuum rotavapor with a waterbath at 40 • C. To assure exhaustive extraction, the procedure was repeated 10 times until colorlessness of the organic solvent to yield a total of 47.5 g crude methanolic extract. An amount of 45 g extract was suspended in 1 L HPLC water and was successively extracted four to six times with 500 mL portions, each of petroleum ether (PE), diethyl ether (DEE), ethyl acetate (EtOAc), and water saturated 1-butanol (1-BuOH) until exhaustion (yield after solvent removal: PE: 8.78 g, DEE: 2.20 g, EtOAc: 0.82 g, 1-BuOH: 3.45 g, H 2 O: 25.87 g). During the extraction with diethyl ether, an off-white precipitate was formed, which was separated and treated as additional fraction (PRC: 4.24 g). American Type Culture Collection (ATCC, Manassas, VA, USA) supplied the J774A.1 murine monocyte macrophage cell line used for assays. Cells were grown adherent to Petri dishes with Dulbecco's modified Eagle's medium (DMEM), which was supplemented with 10% fetal calf serum (FCS), 25 mM HEPES, 2 mM glutamine, 100 u/mL penicillin and 100 mg mL −1 streptomycin at a humidified air with 5% CO 2 and 37 • C in an incubator to enable the cells to attach and reach the exponential phase of growth. Macrophages J774A.1 were plated into 96-well plates (5 × 10 4 /well) and adhered. The spent medium was removed and replaced with fresh DMEM alone or with serial dilutions of the fractions (100-12.5 µg mL −1 ), subfractions (5-0.1 µg mL −1 ), and isolated compounds 1-19 (20-1 µM) for 24 h. Relative cell viability was determined by 3-(4,5-dimethyltiazol-2yl)-2,5-phenyl-2H-tetrazolium bromide (MTT) assay [61]. Briefly, 25 µL of MTT at a concentration of 5 mg mL −1 was added, and the cells were incubated for additional 3 h. Once the cells lysed, the dark blue crystals were solubilized with 100 µL of a solution, consisting of 50% (v/v) N,N-dimethylformamide, 20% (mL L −1 ) sodium dodecyl sulfate (SDS), which was adjusted to a pH of 4.5. The amount of MTT reduction was measured by detecting the optical density (OD) in a microplate spectrophotometer (Titertek Multiskan MCC/340; Dasit, Cornaredo, Milan, Italy) at a wavelength of 550 nm and a reference wavelength of 620 nm.

J774A.1 Murine Macrophage Cell Line
Macrophages J774A.1 were plated into 96-well plates (5 × 10 4 /well) and adhered. The spent medium was removed and replaced with fresh DMEM alone or with serial dilutions of the extracts (10-1 µg mL −1 ) for 1 h, and ultimately, the cells were co-exposed to a final concentration of LPS (1 µg mL −1 ) for an additional 24 h to detect the production of NO. The same protocol was used to assess the production of NO, and the expression of iNOS and COX-2 after treatment with fractions A-R and the isolated compounds 1-19 based on three concentrations, 5-0.1 µg mL −1 or 20-1 µM, respectively.

Measurement of NO Release
After 24 h of LPS stimulation, NO levels were measured by Griess reaction in the culture medium of J774A.1 macrophages as nitrite (NO 2 − ), the index of NO released by cells [62]. Briefly, the macrophages were exposed to samples and LPS; subsequently, 100 µL of cell culture medium were mixed with 100 µL of Griess reagent (equal volumes of 1% (w:v), sulphanilamide in 5% (v/v), phosphoric acid and 0.1% (w:v), and N-(1-napthyl)ethylenediamine dihydrochloride in water) and incubated at room temperature for 10 min. The absorbance was detected in a microplate reader (Titertek Multiskan MCC/340) at a wavelength of 550 nm. The amount of NO 2 − measured in the samples is expressed as µM concentration, which was calculated via a sodium NO 2 − standard curve.

Measurement of iNOS and COX-2 Expression
After the J774A.1 macrophages were exposed to samples and LPS, they were processed as follows: collected, washed with phosphate-buffered saline (PBS), and first incubated in fixing solution (containing PBS, 2% FBS and 4% formaldehyde) for 20 min at 4 • C and then incubated in fix perm solution (containing PBS, 2% FBS and 4% formaldehyde and 0.1% Triton X) for another 30 min at 4 • C. Anti-iNOS/Anti-COX-2 (BD Laboratories, Milan, Italy) were added for 30 min. The secondary antibody was added in fix perm solution and finally, the cells were evaluated using a fluorescence-activated cell sorting (FACSscan; Becton Dickinson, NJ, Franklin Lakes, USA) and elaborated using Cell Quest software [63].

Data Analysis
The data are expressed as mean values in % inhibition ± standard error of the mean (s.e.m), referred to as the LPS stimulated control group. At least three independent experiments, at least in triplicate, were performed for each sample and concentration.

Conclusions
The bioassay-guided isolation of the PE and DEE fractions of the methanolic rhizomes and roots extract of C. articulatus afforded the recovery of 12 sesquiterpenens (1-11, 15), 2 phenolic acids (12, 13), 1 flavonoid (14), 3 stilbenes (16, 17, 19), and 1 monoterpene (18). Among these constituents, compounds 1, 5, 7-10, and 12-19 were isolated for the first time from this species. In summary, the stilbene dimer trans-scirpusin B (17) and trimer cyperusphenol B (19) showed promising inhibition over the production of the inflammatory mediator, NO, in LPS-stimulated J774A.1 murine macrophages and can be assumed to be the major active principle in the methanolic rhizomes and roots extract of C. articulatus, according to the in vitro assays performed. The moderate inhibitory activity of compounds 12, 13, 15, and 16 against NO release, compounds 14, 15, and 16 against iNOS expression, and compounds 16 and 17 against COX-2 expression contribute to the anti-inflammatory potential of the extract. The data obtained through bioactivityguided isolation, monitored by in vitro assays, are the first results confirming the antiinflammatory potential of C. articulatus [16]. Further bio-guided studies on the PRC fraction, containing at least seven stilbene oligomers, might be needed to identify more of the active principles in C. articulatus. The mode of action and the in vivo activity of the active constituents: trans-scirpusin B (17) and cyperusphenol B (19), should be investigated to prove their suitability as anti-inflammatory candidates. In conclusion, the phytochemical and ethnopharmacological investigations of the rhizomes and roots extracts of C. articulatus support its use as a traditional remedy to treat inflammation-related disorders in northern Angola and other countries.
Supplementary Materials: The following supporting information can be downloaded at: https:// www.mdpi.com/article/10.3390/molecules27185937/s1, Figure S1: Cell viability of the obtained PE, DEE, PRC, EtOAc, 1-BuOH and H 2 O fractions.; Figure S2: HPLC-UV chromatogram (205 nm) of DEE subfractions A-R after separation by NP-MPLC.; Figure S3: Cell viability of DEE subfractions A-R of the methanolic extract of the subaerial parts of C. articulatus.; Figure S4: GC chromatogram and peak assignment of the PE fraction of the methanolic extract of the subaerial parts of C. articulatus.; Figure S5: Comparison of the chromatograms (205 nm) of the HPLC-UV analysis of extract subfractions of S. holoschoenus and C. articulatus.; Table S1: Antiproliferative activity and NO production inhibition of compounds 1-11 obtained from the PE fraction of the subaerial parts of C. articulatus.; Table S2: Antiproliferative activity of compounds 12-19 obtained from the DEE fraction of the subaerial parts of C. articulatus.