Introduction

Macroalgae are an important source of a wide range of compounds, endowed with relevant biological activity and potential applications in many different areas (^{Cardozo et al., 2007}; ^{Machado et al., 2010}; ^{Blunt et al., 2011}). It is believed that these chemical compounds serve as defense mechanisms (^{Paul et al., 2011}).

Red algae of Laurencia complex (Ceramiales, Rhodophyta), specifically genus Laurencia, represent one of the most studied marine organisms from environment (^{Suzuki et al., 2009}; ^{Wang et al., 2013}). More than 700 halogenated compounds have already been reported for Laurencia genus (^{Kamada and Vairappan, 2012}; ^{Wang et al., 2013}). Previous studies have focused on the discovery of chemical compounds with pharmaceutical or ecological potential in Laurencia complex (^{Blunt et al., 2011}). In spite of the reports regarding antibacterial, antiviral and cytotoxic activities (^{Machado et al., 2010}), there are only four studies on the antileishmanial action of Laurenciacrude extracts (^{Freile-Pelegrin et al., 2008}; ^{Bianco et al., 2013}) and their sesquiterpenes (^{Santos et al., 2010}; ^{Machado et al., 2011}). Only few reports have addressed inter or intraspecific variation in secondary metabolite composition and none of them have considered the impact of chemical variability on bioactivity (^{Yuan et al., 2009}; ^{Stengel et al., 2011}; ^{Sudatti et al., 2011}; ^{Oliveira et al., 2013}).

Currently, Laurencia complex is comprised by six genera: Laurencia J.V. Lamouroux, Osmundea, Chondrophycus(J.Tokida & Y.Saito) Garbary & J.T.Harper, LaurenciellaV.Cassano, Gil-Rodríguez, Sentíes, Díaz-Larrea, M.C.Oliveira & M.T.Fujii, Palisada K.W. Nam and Yuzurua (Nam) Martin-Lescanne (^{Fujii et al., 2011}; ^{Cassano et al., 2012a}). The genus Laurencia is represented by 419 species (and infraspecific) names in the database at present (^{Guiry and Guiry, 2014}), of which 130 have been flagged as being currently taxonomically accepted. In Brazil, five of the six genera of the complex can be found: Laurencia itself, Laurenciella, Osmundea, Palisada, and Yuzurua, which are widely distributed along the coast in intertidal and shallow subtidal zones, except for the last, whose occurrence is restricted to the northeast.

The aim of this study was to evaluate the geographical variability of metabolite production by Laurencia complex, and the impact of chemical diversity on bioactivity. Chemical profiles and Principal Component Analysis were employed in the study. The samples were submitted to antileishmanial assay against promastigote and intracellular amastigote forms of Leishmania amazonensis, the etiological agent of leishmaniasis, which is a public health problem in many countries (^{Croft and Coombs, 2003}; ^{Chappuis et al., 2007}). The most active samples were tested against amastigote forms and their cytotoxicity was evaluated in macrophages in order to estimate antiparasitic selectivity.

Materials and methods

Results

Chemical profiles of twelve crude extracts from four different species of Laurencia complex collected in different sites of the Brazilian coast were obtained by HPLC/DAD (Fig. 1) and compared using Principal Component Analysis (PCA).

Figure 1 HPLC fingerprint chromatograms (λ=200 nm) of Laurencia complex. 

During collection, two populations were distinguished at Ubú beach (LdU) and three at Manguinhos beach (LdM). While LdU1 displayed dense red to brown-purple tufts, LdU2 exhibited brown-purple patterns (Box 2). Differently, the samples LdM1, LdM2 and LdM3 were greenish or brown-purple with orange tips. Additionally, it remarkable differences were observed on Figure 1 – HPLC fingerprint chromatograms (λ=200 nm) of Laurencia complex.

the chemical profiles (Fig. 1). Chemical profiles of LdAR, LdM1, LdM2, LdM3 and LdMa displayed greater chemical diversity than LaFl, LdU1, LdU2, PfFl, PpFl, PpM and PpU.

In this study, the first eight minutes of the chemical fingerprint were removed for PCA, since they comprise a set of overlapping peaks of more polar compounds that do not have useful discriminant information.The two first components of PCA accounted for 96.1% (PC1 = 53.3% and PC2 = 42.8%) of the total chromatographic variance. The resulting PCA score plot discriminated the samples (Fig. 2A) according to the strong influence of two main substances, with a strong weight in CP1 and CP2, as depicted by the plot of PCA loadings against the retention time (Fig. 2B and 2C). Basically, PC1 discriminated the samples according to the presence of a compound with retention time of 32.74 min while PC2 according to a compound eluting in 30.78 min. The identification of these compounds was accomplished by comparison of their retention times with purified chamigranes (-)-elatol (1) and obtusol (2), respectively.

Figure 2 Principal Component Analysis of HPLC fingerprints at 200 nm obtained from twelve populations of Laurencia complex. Components 1 and 2 account for 96.1% of the total chromatographic variation among the populations. Factor scores (A) and loads plotted against retention time (B and C) are given. 

The PCA score plot grouped the sample of L. dendroidea from Macaé (LdMa) alone at the lower left plot quadrant, characterized mostly by the presence of (-)-elatol (1) confirmed after the analysis of ¹H NMR spectra. Conversely, the samples of L. dendroidea from Manguinhos (LdM1, LdM2 and LdM3) were characterized by an increased signal of obtusol (2) and no signal of (-)-elatol and were grouped at the upper right quadrant. Similarly, NMR data of these extracts also revealed the presence of obtusol. Containing both (-)-elatol and obtusol and placed alone at the upper left quadrant was the L. dendroidea from Angra dos Reis (LdAr). Finally, at the lower right quadrant were the samples characterized mostly by flat chromatograms, with the absence of (-)-elatol and obtusol, which corresponds to the species P. flagellifera from Flexeira (PfFl1) and P. perforata from Flexeira, Manguinhos and Ubú (PpFl, PpM and PpU), along with L. dendroidea from Ubú (LdU1 and LdU2) and L. aldingensis from Flexeira (LaFl). Fig. 3 shows the comparison of the ¹H NMR spectra of purified (-)-elatol and obtusol with the crude extracts LdMa, LdM1, LdM2, LdM3 and the LdAR characterized by the presence of these compounds. Additionally, Table 1 shows the relative area of the pure compounds obtained by HPLC/DAD analysis.

Figure 3 Comparison of ¹H NMR spectra (200 MHz, CDCl₃) of (-)-elatol, obtusol and crude extracts LdAR, LdMa, LdM1, LdM2 and LdM3. 

In spite of the chemical differences, all the tested samples displayed antileishmanial action (Table 2). Comparing the results against promastigote form with the groups showed on PCA, important differences can be observed on bioactivity even within a cluster. L. dendroidea from Angra dos Reis (LdAR) was the most active (IC₅₀ 17.9 ± 1.3 µg/ml). Although eight samples were identified as L. dendroidea their bioactivity was remarkably different with IC₅₀ values ranging from 17.9 ± 1.3 µg/ml (LdAR) to 97.2 ± 2.0 µg/ml (LdU1). It is important to mention that isolated compounds (-)-elatol and obtusol are known to possess antileishmanial action with IC₅₀ 9.7 ± 1.2 µg/ ml and 6.2 ± 0.5 µg/ml against promastigote forms (^{Machado et al., 2011}). Despite the lack of chamigranes, P. flagellifera from Fleixeira (PfFl) was, along with LdU2, the third most active crude extract. Nevertheless, in the anti-amastigote assay, more relevant in terms of chemotherapy, LaFl was the most active sample (IC₅₀ 12.5 ± 1.4 µg/ml) followed by LdU2 (IC₅₀ 16.8 ± 2.8 µg/ml) and LdM3 (IC₅₀ 19.2 ± 4.6 µg/ml). Moreover, none of the tested samples displayed significant cytotoxic activity against macrophages with IC₅₀ greater than 198 ± 6.1 µg/ml. Since none of the samples significantly activated the production of NO by infected macrophages, their antileishmanial activity is likely to be direct on the parasites.

Discussion

Chemical profiling has been proved to be useful in the study of chemical variability of natural products (^{Yuan et al., 2009}; ^{Nylund et al., 2011}; ^{Payo et al., 2011}; ^{Hou et al., 2012}). These approaches are relevant because of the existence of inter-and intrapopulation chemical differences between marine organisms, described in several studies (^{Miller et al., 2001}; ^{Sacristan-Soriano et al., 2011}). Basically, the sources of chemical variation can be either genetic or environmental.

The genetic influence on chemical differences was proposed by ^{Oliveira et al. (2013)} for elatol content on L. dendroidea. Additionally, ^{Howard et al. (1980)}, using unialgal culture, observed that the chemical compounds produced by Chondrophycus undulatus (as Laurencia undulata), L. gardneri, L. nippponica, L. intermedia, L. okamurai, and Osmundea spectabilis (as L. spectabilis), were qualitatively and quantitatively identical to those produced in natural populations suggesting that the secondary metabolites of Laurencia complex populations reflect genetic differences. Crossing experiments were used to study the chemical diversity in Laurencia nipponica revealing that genetic variation is a determining factor for the synthesis of halogenated metabolites (^{Masuda et al., 1997}).

The outcome of genetic differences, proposed in previous studies, is the existence of chemical types, once chemical diversity is observed not only between species but also among populations from different sites, as observed for Laurencia nipponica (^{Masuda et al., 1997}), L. obtuse (^{Caccamese et al., 1981}), L. pacifica (^{Fenical, 1976}) and recently for L. nangii (^{Kamada and Vairappan, 2012}).

In the present study, four groups of L. dendroidea, were recognized by PCA analysis. Therefore, four chemical types were distinguished, determined by collection site. Previous reports, with samples from another regions, revealed that some Brazilian populations of L. dendroidea (as L. scoparia) also produce a different array of chamigrane and bisabolene sesquiterpenes (^{Davyt et al., 2001}; ²⁰⁰⁶), corroborating the existence of a specific chemical profile for each collection site.

Environmental factors can modulate the levels of secondary metabolites in marine organisms. In this regard, ^{Kuwano et al. (1998)}reported that an increase in salinity led to higher levels of sesquiterpenes in L. okamurae. In another study, temperature and salinity were reported to influence the amount of sesquiterpenes and severe conditions decreasing the level of elatol in L. dendroidea (^{Sudatti et al., 2011}).

Environmental conditions are also considered to be the cause of morphological plasticity in algae (^{Oliveira et al., 2013}). It is known that Laurencia dendroidea display morphological differences that may potentially complicate the identification of species (^{Cassano et al., 2012b}). In the present study, intrapopulation and interpopulation morphological differences were observed (Box 2). However, chromatographic profiles indicated that intrapopulation morphological differences do not represent important chemical variations. For instance, morphologically different samples LdM1, LdM2 and LdM3 displayed similar chemical profiles, and the same pattern is observed for LdU1 and LdU2. Conversely, despite the quantitative difference on obtusol content (Table 1), antileishmanial action of LdM1, LdM2 and LdM3 are very similar, while LdU1 (IC₅₀ 97.2 ± 2.0 μg/ml) is about three times less active than LdU2 (IC₅₀ 30.1 ± 1.5 μg/ml). These results suggest that microscale factors acting at Úbu beach (LdU1 and LdU2) result, not only on morphological differences, but also on antileishmanial action. Once chemical profiles are very similar, the pharmacological dissimilarities may be explained as quantitative variation on active metabolites produced.

Although three different collection sites were compared, all Palisadaspp. samples displayed similar chromatograms, with the absence of sesquiterpenes (-)-elatol and obtusol, as reported previously (^{Fujii et al., 2011}). Nevertheless, samples PfFl, PpFl, PpM and PpU displayed antileishmanial action comparable to Laurencia spp., samples containing chamigranes (LdAR, LdMa, LdM1, LdM2 and LdM3). This result indicates the existence of additional active compounds that may be eluting in the first eight minutes of the chromatograms, or the presence of additional compounds that were not detected by DAD. In both cases, it is clear that compounds other than (-)-elatol and obtusol are active against Leishmania amazonensis.

The antileishmanial action also differed among the samples that contained (-)-elatol (1) and obtusol (2). The most active extract against promastigote form was LdAR, which contained both sesquiterpenes. The presence of (-)-elatol and obtusol on LdAR could explain the antileishmanial action observed. Samples with only one of these two compounds (Table 1) were less potent and displayed similar antileishmanial action, for instance IC₅₀ for LdMa was 40.6 ± 1.6 while for LdM2 was 41.9 ± 1.6 µg/ml.

In conclusion, the present results showed a geographic variation among the metabolites produced by Laurencia complex from Brazilian coast, especially within L. dendroidea species, with small scale factors affecting the levels of leishmanicidal compounds and morphological features at Ubú beach. The antileishmanial action was not restricted to the samples containing chamigranes, suggesting the existence of other classes of active compounds. This fact reinforces that Laurencia complex can be a source of new antileishmanial compounds other than chamigrane-type sesquiterpenes.