OUP user menu

Temporal variation of the microbial community associated with the mediterranean sponge Aplysina aerophoba

Anja B. Friedrich, Isabell Fischer, Peter Proksch, Jörg Hacker, Ute Hentschel
DOI: http://dx.doi.org/10.1111/j.1574-6941.2001.tb00888.x 105-113 First published online: 1 December 2001


Sponges of the Aplysinidae family contain large amounts of bacteria that are embedded within the sponge tissue matrix. In order to determine the stability and specificity of the Aplysina–microbe association, sponges were maintained in recirculating seawater aquariums for 11 days. One aquarium was left untreated, a second one contained 0.45 μm filtered seawater (starvation conditions) and the third one contained 0.45 μm filtered seawater plus antibiotics (antibiotics exposure). Changes in the microbial community were monitored using group-specific, 16S rRNA targeted oligonucleotide probes, by denaturing gradient gel electrophoresis and by electron microscopic observations. Furthermore, the changes in the natural product profile were monitored using high-performance liquid chromatography. The measured parameters showed that a large fraction of the sponge-associated microbial community could not be cleared under the given experimental conditions. Based on these cumulative results we postulate that a large fraction of sponge-associated bacteria resides permanently in the Aplysina aerophoba mesohyl pointing to a highly integrated interaction between the host sponge and associated microorganisms.

  • Aplysina
  • Sponge
  • Sponge-associated microorganism
  • Symbiosis
  • Denaturing gradient gel electrophoresis
  • Fluorescence in situ hybridization
  • Natural product

1 Introduction

Sponges (Porifera) are primitive metazoans that were probably the starting point for the metazoan explosion during the Precambrium about 650 million years ago [1]. They are widely distributed on today's tropical reefs but are also found in lower abundances with increasing latitudes. Sponges have received a lot of attention as they are a rich source of natural products with antiviral, antitumor, antimicrobial or generally cytotoxic properties that are of considerable biotechnological interest [2,3]. Functionally sponges share many features with unicellular protozoa particularly with respect to nutrition, cellular organization, gas exchange, reproduction and response to environmental stimuli [4,5]. Instead of organs or tissues, sponges possess amoeboid, omnipotent and phagocytotically active cells that move freely through the sponge matrix, termed the mesohyl. Nevertheless, sponges are true metazoans that can reach a considerable size (1 m or more in height) particularly in tropical waters. Sponges contain genes and proteins (cell adhesion and receptor molecules, signal transduction and others) that are highly homologous to vertebrate analogs [6]. This provides molecular evidence for the functional similarities between sponges and higher eukaryotes.

Sponges are also known to be sometimes associated with large amounts of bacteria that can amount to 40% of the biomass of the sponge [7,8]. Various microorganisms have evolved to reside in sponges, including cyanobacteria [9], diverse heterotrophic bacteria [10,11], unicellular algae [12,13] and zoochlorellae [14]. Since sponges are filter feeders, a certain amount of transient bacteria are trapped within the vascular system or attached to the sponge surface. Considering that sponge–bacterial associations are (i) evolutionarily ancient, (ii) widely distributed and (iii) in some cases specific to their host [1517], the question is what functions the bacteria may have. Putative symbionts may provide benefits to their hosts, such as a source of nutrition by translocation of metabolites [18], access to bacterial-specific traits such as autotrophy, nitrogen fixation and nitrification [19], protection against UV radiation [20] and secondary metabolite production [2123]. However, only once has a symbiotic activity been linked to a specific bacterium [15]. Experimental evidence is difficult to come by because the sponge microflora is a diverse consortium and there appears to be no physical compartment that separates putative symbionts from commensal microflora. Moreover, most permanently sponge-associated bacteria cannot be cultured outside their host rendering investigations regarding their function difficult.

Sponges of the Aplysinidae family contain high concentrations of brominated alkaloids with cytotoxic activities and repellent properties against predators [24,25] and are also a rich source of microorganisms with antimicrobial activities [26]. Aplysina sponges also harbor large amounts of microorganisms that are embedded within the mesohyl. The microbial population is physically separated from the surrounding seawater by contiguous host membranes. Microorganisms are removed from the seawater passing through the canal system and transferred into the mesohyl interior. In order to establish themselves in this environment, bacteria must be resistant to digestion by sponge archaeocytes. Indeed, transmission electron microscopy (TEM) observations document that the physically abundant bacteria contain various additional membranes, slime capsules and enlarged periplasms which presumably serve as barriers to prevent digestion [17,27].

The purpose of this study was to identify bacteria that are permanently associated and therefore putatively symbiotic with Aplysina aerophoba. Sponges were maintained in seawater aquariums under starvation conditions and upon exposure to antibiotics with the aim to clear the sponges from transient microorganisms that are only present by coincidence. The temporal variations of the microbial consortium were investigated by various morphological, microscopic and molecular techniques. The results are expected to provide insights into quantities and types of microorganisms that reside permanently in the A. aerophoba mesohyl.

2 Materials and methods

2.1 Sponge collection

Specimens of the mediterranean sponge A. aerophoba (class Demospongiae, order Verongida, family Aplysinidae) were collected by scuba diving at a depth of 5–15 m off Banyuls-sur-Mer (France) in April 1999 and in May 2000. The sponges, arbitrarily numbered 1, 2, 4, 5, 7 and 8, were collected on one dive within a 50-m2 radius on a rocky reef and the sponges 3, 6 and 9 were collected from the same location 4 days later (Table 1). Physical and chemical parameters of the sampling sites were obtained using standard methods according to the manufacturer's suggestions (B. Lange GmbH, Berlin, Germany) (Table 1). Sponge specimens were placed into sampling bags under water to avoid contact with air and transported to the laboratory.

View this table:

Environmental parameters of the seawater at the collection site and the maintenance aquariums

2.2 Aquarium experimental design

Three 12-1 aquariums were set up for a time period of 11 days at an ambient temperature of 12°C. Seawater pumps (Aqua Clear 1000 Powerhead, Hagen-Aquaristic, Holm, Germany) were installed for recirculating water movement and for oxygenation. One aquarium contained freshly collected seawater, the second aquarium was filled with 0.45 μm filtered seawater and the third aquarium contained 0.45 μm filtered seawater, the β-lactam antibiotic ampicillin and the aminoglycoside antibiotic gentamicin, each added to a final concentration of 100 μg ml−1. To compensate for the potential loss of antibiotics during the experiment, the aquariums were supplemented with 50 μg ml−1 of each antibiotic at days 4 and 7. At specified time intervals the aquarium seawater was plated onto ZoBell medium to confirm that the antibiotics were effective. The aquariums were partially covered with plastic plates to reduce microbial contamination from the air. At the beginning of the experiment, each aquarium contained two sponges and a third one was added to each aquarium after 4 days.

The sponges were marked with colored strings for identification. The health status of the sponges was assessed by frequent visual inspection as actively pumping sponges generate sufficient water movement through their chimneys to cause swirls on the water surface. At time intervals of 0, 4, 7 and 11 days, a tissue sample of each specimen was taken with a cork borer (11 mm in diameter) under water. This procedure had no apparent effects on the sponges as they resumed pumping activity shortly thereafter. The tissue sample was placed into sterile seawater and the reddish surface tissue was removed from the yellow core mesohyl tissue with a scalpel blade. The mesohyl tissue was rinsed two more times in sterile seawater, cut into pieces and processed as described below.

2.3 Determination of total/culturable bacteria

The sponge tissue was weighed, minced with a razor blade and homogenized using a Dounce homogenizer with a constant number of strokes. The tissue homogenates were diluted with sterile seawater from 10−1 to 10−5. A 100-μl aliquot of each dilution was plated in triplicate onto ZoBell agar (Difco 2216) and colonies were counted after 1 week incubation at room temperature. To determine the total number of bacteria, a defined volume of the tissue homogenate was fixed with an equal volume of 8% paraformaldehyde/phosphate-buffered saline (PBS) and stored at 4°C until use. The tissue homogenate was centrifuged for 5 min at 13 000 rpm, washed two times with PBS and fixed in an ethanol series (50%, 70%, 100%). DAPI (4,6-diamidino-2-phenylindole) was added to a final concentration of 5 μg ml−1 (Sigma). The DAPI-stained homogenate was diluted from 10−1 to 10−6 with sterile PBS. A volume of 1 ml was filtered onto a black, 25-mm-diameter, 0.2-μm polycarbonate membrane (Millipore) which was supported by a 0.45-μm cellulose nitrate filter (Schleicher and Schuell, Dassel, Germany). Vacuum (<10 cm Hg) was applied carefully with a hand pump. The filters were washed two times with sterile seawater. They were then mounted with Citifluor (Citifluor, London, UK) onto a microscope slide. Bacterial numbers were determined following microscopical inspection (Axiolab Microscope, Zeiss, Germany). Three independent samples were processed per sponge. For each sample an average bacterial number from 10 different counting fields was determined.

2.4 TEM

A. aerophoba samples were fixed in 2.5% glutaraldehyde/PBS. The samples were rinsed 3×20 min in 1×PBS, fixed overnight in 2% osmium tetroxide, rinsed again and cut into small pieces. After overnight incubation in 0.5% uranyl acetate, the pieces were dehydrated in an ethanol series (30%, 50%, 70%, 100%) and incubated 3×20 min in 1×propylene oxide. Following overnight incubation in 1:1 (v/v) propylene oxide/Epon 812 (Serva) they were embedded in Epon 812 at a temperature of 60°C. The embedded samples were sectioned with a ultramicrotome (OM U3, C. Reichert, Austria) and examined by TEM (Zeiss EM 10, Zeiss, Germany).

2.5 Fluorescence in situ hybridization (FISH)

A tissue core was punched out from the center of the sponge with an ethanol-sterilized cork borer (11 mm in diameter). Tissues which had been exposed to the surface of the sponge were removed with an ethanol-sterilized scalpel blade. The remainder was frozen in liquid nitrogen. For hybridization, sponge pieces were sectioned into 2–4-μm-thick slices with a cryomicrotome (Mikrom HM 500 OM, Walldorf, Germany), placed on microscope slides and air-dried. The sections were fixed with Carnoy solution (60% ethanol, 30% trichloromethane, 10% acetic acid) at room temperature for 45 min. After drying, the sections were dehydrated in an aqueous ethanol series (50%, 80%, 96%) for 3 min each, dried again and stored at −20°C until use. In situ hybridizations were performed as described previously [28].

Oligonucleotide probes were used at a concentration of 3 ng of labelled probe μl−1 hybridization buffer (MWG, Biotech, Ebersberg, Germany) (Table 2). The probes EUB338, EUB338-II and EUB338-III were used as a 1:1:1 mix [29]. The probes BET42a and GAM42a were used with the unlabelled competitors GAM42a and BET42a, respectively (Table 2). The slides were rinsed, air-dried and mounted in Citifluor. Examination was done with an Axiolab microscope equipped with Zeiss filter sets 10 and 15. Color micrographs were taken with a low-light charge-coupled device camera (Intas, Germany). Digital image processing was performed using the software Adobe Photoshop. The signal abundances were visually estimated after triplicate inspection in comparison to a reference chart (Fig. 3).

View this table:

Oligonucleotide probes used for in situ hybridization


FISH of A. aerophoba tissue cryosections with the 16S rRNA oligonucleotide probes EUB-I, -II, -III (a), NHGC (b), PLA46 (c), SRB385 (d), BACT (e) and HGC69a (f). The chart was used as an internal reference for the visual estimation of bacterial abundances (Table 3). Magnification ×1729.

2.6 Denaturing gradient gel electrophoresis (DGGE)

Sponge tissue cores were punched out from the center of the sponge with the aid of a cork borer. Tissues which had been exposed to the sponge surface were removed with an ethanol-sterilized scalpel blade. The remainder was rinsed three times in sterile seawater and frozen in liquid nitrogen. For DNA extraction, the tissue was ground using a mortar and pestle while being submerged in liquid nitrogen. Genomic DNA was extracted using the Fast DNA Spin Kit for soil (Q-Biogene, Heidelberg, Germany) according to the manufacturer's instructions and stored at 4°C. The universal primers 341F with the GC-clamp [30] spanning Escherichia coli positions 341–357 and 518R spanning E. coli positions 518–534 were used for 16S rDNA amplification [31]. PCR was performed using a Mastercycler Gradient (Eppendorf, Hamburg, Germany) as follows: one initial denaturation step for 2 min at 95°C; 30 cycles of 1 min at 95°C, 1 min at 54°C, 1.5 min at 72°C; and one final elongation step for 10 min at 72°C. The PCR mix consisted of 45 μl 1×reaction buffer, 1 μl of each primer (100 pmol final concentration), 1 μl of deoxyribonucleoside triphosphates (10 μmol), 2.5 U MasterTaq DNA polymerase (Eppendorf, Hamburg, Germany) and 1 μl DNA template (corresponding to 20–500 ng of DNA). DGGE was performed using a Bio-Rad DCode Universal Mutation Detection System (Bio-Rad Laboratories GmbH, Munich, Germany) on a 10% (w/v) polyacrylamide gel in 0.5×TAE and using a 0–100% denaturing gradient. 100% denaturant corresponded to 7M urea and 40% (v/v) formamide. Electrophoresis was performed for 6 h at 150 V and 60°C. Gels were stained for 30 min in aqueous ethidium bromide solution (0.5 μg ml−1) and photographed with a GelDoc system (Gel Doc 2000, Bio-Rad).

Several experimental parameters were initially tested to establish maximum amplification efficiency and to check the reproducibility of the DNA banding pattern. A regular PCR protocol with an annealing temperature at 54°C resulted in more bands than touchdown PCR and was therefore chosen as the standard protocol. A comparison of DNAs extracted immediately after collection and DNAs extracted from liquid nitrogen-frozen tissues showed the same banding pattern. Therefore, DNA was routinely extracted from liquid nitrogen-frozen material. A comparison of different gradients (0–100%, 30–70%) revealed that optimal resolution was obtained at 0–100% denaturation. The weak banding of some sponges (Fig. 4a, lane 1; Fig. 4b, lanes 8, 9) is likely due to poor PCR amplification as the PCR gels repeatedly displayed weaker amplification products (data not shown). Each gel was run three times to confirm the reproducibility of the overall pattern.


DGGE profiles of PCR-amplified 16S rDNA fragments from individual A. aerophoba sponges. Each number corresponds to a sponge and was repeatedly sampled at the given time intervals. Three sponges were maintained each in regular seawater (a), in 0.45 μm filtered seawater (starvation conditions) (b) and in 0.45 μm filtered seawater supplemented with 100 μg ml−1 ampicillin and gentamicin (antibiotics exposure) (c).

2.7 High-performance liquid chromatography (HPLC) analysis

Liquid nitrogen-frozen tissue was lyophilized and extracted with 1 ml methanol g−1 dry weight for 1 h at room temperature. Salicylic acid (10 mg ml−1) was used as an internal standard. For HPLC analysis, 20 μl of a 1:10 dilution was injected into an HPLC system coupled to a photodiode array detector (Gynkotek, Munich, Germany). Routine detection was performed at 254 nm in aqueous methanol. The separation column (125×4 mm) was prefilled with Eurospher C18 (Knauer, Berlin, Germany). Quantification of brominated alkaloids was performed using previously isolated pure compounds as external standards.

3 Results and discussion

The aim of this study was investigate the interactions between the sponge A. aerophoba and its potential symbionts. In defining an interaction, several aspects need to be taken into consideration [32,33]. First, the partners of the interaction need to be identified. Methods centering around the 16S rDNA gene now make it possible to phylogenetically describe complex consortia of unculturable bacteria [3436]. Secondly, the stability of the interaction needs to be considered. For example, an unanswered question was whether high amounts of bacteria pass transiently through the animal or whether a defined, permanent mesohyl population exists that is unaffected by the animals’ filtration activity. A recent study by Webster et al. [37] was the first to show that the microbial community of the sponge Rhopaloeides odorabile could be perturbed upon exposure to Cu2+ as an environmental stress. Thirdly, the most challenging questions concern the mechanisms of interaction between bacterial symbionts and their sponge hosts. This study is one of the first contributions that investigate temporal variations of sponge-associated microorganisms.

Counting of bacterial numbers in A. aerophoba tissue extracts revealed 6.4±4.6×108 bacteria g−1 sponge tissue (n=9). To our knowledge this is the first report to show that bacterial concentrations in sponges exceed those of seawater by two to three orders of magnitude [38]. Only 0.15% of the bacterial population was culturable as determined by counts on ZoBell medium (1±1.4×106 bacteria g−1 sponge tissue). These numbers are consistent with estimates that >99% of the microorganisms in the environment cannot be cultivated on laboratory media [34,39]. The bacterial concentration of seawater was determined as an internal control and fell within the range of typical seawater (data not shown). The total bacterial numbers remained unchanged over the time course of the experiment irrespective of the maintenance conditions. The culturable bacterial units (CFUs) increased by one to two orders of magnitude following maintenance in regular seawater (Fig. 1a) and decreased following starvation (1×104 g−1 mesohyl tissue) or antibiotics exposure (5×103 g−1 mesohyl tissue), respectively (Fig. 1b,c). Because the culturable bacterial fraction was two to five orders of magnitude lower than the total bacterial fraction these changes were not reflected in the DAPI counts.


Determination of total (DAPI-stained) and culturable (CFUs) bacterial numbers in A. aerophoba tissues. Sponges were maintained in regular seawater (a), 0.45 μm filtered seawater (starvation condition) (b) and 0.45 μm filtered seawater supplemented with 100 μg ml−1 ampicillin and gentamicin (antibiotics exposure) (c). The closed symbols represent the DAPI counts, the crosses represent CFUs.

These quantitative data are supported by TEM observations. Visual inspection revealed that the bacterial composition and density within the mesohyl tissues did not change (Fig. 2a,b,c,d). The sponge archaeocytes are highly vacuolized which indicates an active metabolism. Bacterial cell division was observed infrequently. Interestingly, the same bacterial morphotypes that have been observed in A. cavernicola from Marseille and Elba [7,17] were also present in A. aerophoba from Banyuls-sur-Mer. These physically abundant bacteria contain various additional membranes (type C), slime capsules (type D) and enlarged periplasms (type E). The phenotypic characteristics of type E (enlarged periplasm, putative nuclear membrane) are consistent with bacteria of the genus Pirellula[40,42].


Transmission electron micrographs of A. aerophoba mesohyl tissues. Samples were taken following maintenance in regular seawater at day=0 (a) and day=11 (b) and following exposure to antibiotics at day=0 (c) and day=11 (d). Scale bar: 5 μm.

The application of group-specific, 16S rDNA-targeted oligonucleotide probes to A. aerophoba tissue sections revealed an overall profile whose composition is very similar to the microbial profile of A. cavernicola, [17]. Among the Eubacteria the δ-Proteobacteria were most abundant, followed by the Bacteroides cluster, the γ-Proteobacteria and the high-GC Gram-positive bacteria. With FISH probes specific to the Archaea (ARCH915), the α- (ALF1b) and β- (BET42a) Proteobacteria and Cytophaga-Flavobacterium cluster (CF319), signals could not or only rarely be obtained. Following maintenance of the sponges under the respective aquarium conditions, the abundances of the different phylogenetic groups remained largely unchanged (Table 3).

View this table:

FISH of A. aerophoba tissues

The application of a planctomycete-specific FISH probe (PLA46) revealed an abundant signal in A. aerophoba tissues that rivals the amounts of the δ-Proteobacteria. The probe also resulted in high signal abundances in the sponge R. odorabile from the Great Barrier Reef [41] and in all Aplysina species investigated so far (Hentschel et al., in preparation). Planctomycetes are naturally ampicillin-resistant because of their peptidoglycan-free cell walls [40,42]. From an ecological perspective it might be proposed that the constant exposure to bioactive secondary metabolites in the mesohyl may select for bacteria that are naturally resistant to antibiotics. Furthermore, the identification of ampicillin-resistant bacteria provides a phylogenetic explanation for the observed lack of bacterial clearance upon ampicillin exposure. It is tempting to speculate that the unusual membranes of planctomycetes may be responsible for resistance to phagocytotic digestion by sponge cells.

With the DGGE analysis, a second, independent molecular approach was employed to describe microbial community changes in sponges over time and with respect to maintenance conditions (Fig. 4). Because the same specimen can be sampled repeatedly without visible injury or reduction in pumping activity, it is possible to monitor the changes in DGGE banding pattern of an individual sponge. The DGGE analysis from aquarium-maintained Aplysina sponges resulted in complex profiles with an average of 15–20 bands per lane. When the banding pattern of freshly collected sponges was compared, it is noticeable that sponges collected on the same day shared more similarities than compared to the third one, which was added 4 days later. When the sponges were maintained in regular seawater, the DGGE pattern remained largely unchanged (Fig. 4a). When maintained in 0.45 μm filtered seawater, there was a noticeable appearance of additional bands at t=11 days which coincided with the bacterial bloom observed in this tank (Fig. 4b). When maintained in the presence of antibiotics, some bands disappeared and existing ones appeared more faded at day 11 (Fig. 4c). This indicates that the respective microorganisms have been cleared from the mesohyl tissue.

Several hypotheses exist as to whether the natural products characteristic for Aplysina sponges are produced by the animal or possibly by symbiotic bacteria. Evidence for the former hypothesis is provided by the observation that brominated alkaloids are localized in host cells (spherulous cells) of A. fistularis[43]. On the other hand, the chemical structure of several produced metabolites points to enzymes of microbial or fungal origin [44,45]. It is also conceivable that bacteria are involved in the synthesis of precursors which are then transferred and stored in host vesicles. To investigate whether a correlation exists between bacterial abundances and natural product composition, the concentrations of six selected metabolites were followed over time (Fig. 5). The metabolite concentrations did not change irrespective of the maintenance conditions. Since the microbial abundances and diversity did not change significantly either, conclusions regarding the possible involvement of bacteria in secondary metabolism cannot be drawn. The lack of change, particularly of the wound-activated metabolite dienone, suggests that the animals were not stressed during the experiment, or that the secondary metabolites are not suitable indicators for stress. Betancourt et al. [46] determined that the antimicrobial activities of A. fistularis in the environment were remarkably stable leading to the conclusion that these compounds were produced continuously.


HPLC analysis of six brominated alkaloids characteristic of A. aerophoba. Three sponges were maintained each in regular seawater (a), in 0.45 μm filtered seawater (starvation conditions) (b) and in 0.45 μm filtered seawater aquarium supplemented with 100 μg ml−1 ampicillin and gentamicin (antibiotics exposure) (c). The bars represent isofistularin-3 (white), aplysinamysin-1 (red), aerophobin-2 (yellow), aerophobin-1 (blue), aeroplysinin-1 (turquoise) and dienone (black). The total amount of brominated compounds in freshly collected A. aerophoba (set at 100%) is equal to 23.7±2.9 mg g−1 dry sponge weight (n=21). The individual contributions of metabolites in freshly collected animals were isofistularin-3 (48.9±3.9%), aplysinamysin-1 (15.7±8.2%), aerophobin-2 (20.3±8.8%), aerophobin-1 (8.1±1.8%), aeroplysinin-1 (5.8%±3.3) and dienone (1.1±1.3%) (n=21).

In summary, the sponge mesohyl of A. aerophoba provides for a specialized environmental niche that contains high numbers of bacteria exceeding those of seawater by two or three orders of magnitude. A large fraction of the microbial community could not be cleared by starvation of the sponges or antibiotics exposure, at least not over the time course of the experiment. These cumulative results indicate the existence of a highly stable population within sponge tissues. Since sponge bacterial interactions date back to the beginning of the metazoan evolution, a significant subset of bacteria may be associated with the sponge possibly over evolutionarily long periods of time. Future studies aim at a detailed phylogenetic identification of the microorganisms associated with the sponge A. aerophoba as well as to provide insights into the mechanisms underlying the bacteria–sponge interactions.


We thank C. Gernert for establishing the DGGE analysis, C. Pabel for help in establishing the DAPI assay and M. Steinert for many interesting discussions and for critically reading the manuscript (all at Universität Würzburg). We also owe thanks to G. Brauers (Universität Düsseldorf) for support in the HPLC analysis. We gratefully acknowledge the marine operations personnel at the Laboratoire Arago (Banyuls-sur-Mer, France) for their expert help during sponge collection and P. Lebaron for sharing his laboratory. This work was supported by grants to U.H. (BMB+F 03F0235A and BMB+F 03F0239A) and to P.P. (SFB 251).


  1. [1]
  2. [2]
  3. [3]
  4. [4]
  5. [5]
  6. [6]
  7. [7]
  8. [8]
  9. [9]
  10. [10]
  11. [11]
  12. [12]
  13. [13]
  14. [14]
  15. [15]
  16. [16]
  17. [17]
  18. [18]
  19. [19]
  20. [20]
  21. [21]
  22. [22]
  23. [23]
  24. [24]
  25. [25]
  26. [26]
  27. [27]
  28. [28]
  29. [29]
  30. [30]
  31. [31]
  32. [32]
  33. [33]
  34. [34]
  35. [35]
  36. [36]
  37. [37]
  38. [38]
  39. [39]
  40. [40]
  41. [41]
  42. [42]
  43. [43]
  44. [44]
  45. [45]
  46. [46]
  47. [47]
  48. [48]
  49. [49]
View Abstract