Does warm-water extraction of benthic diatoms yield extracellular polymeric substances or does it extract intracellular chrysolaminaran?

In this issue, Chiovitti et al. (2004) argue that a 1-h extraction of diatoms with water at 30^oC results in the liberation of the intracellular storage carbohydrate chrysolaminaran and does not yield extracellular polymeric substances (EPS). This conclusion challenges work published over the last 5 years in which we have used this method for the extraction of EPS from benthic epipelic diatoms isolated from intertidal mudflats. We argue here that Chiovitti et al. (2004) have not proved their contention that our work is in error, nor have they provided conclusive evidence that they have extracted chrysolaminaran alone.

Benthic epipelic diatoms secrete copious amounts of EPS, which are of considerable ecological importance (e.g. Paterson & Black, 1999). Because of this, the composition of these substances and the metabolic control of their formation have been intensively researched by a number of research groups. One approach has been to use isolated strains of benthic diatoms in controlled laboratory conditions, and extraction methods have ranged from centrifugation of diatom cultures and subsequent analysis of EPS in the supernatant (Smith & Underwood, 2000) to extensive sequential extraction procedures (Wustman et al., 1997). The extraction method clearly influences the nature and amount of material obtained. In our experience, centrifugation alone did not recover all the EPS from diatom cultures (Staats et al., 1999; de Brouwer et al., 2002). Therefore, we developed a protocol to extract EPS by treatment with water at 30°C for 1 h (Staats et al., 1999). Although this method was intensively checked for possible contamination of the EPS with internal sugars (Staats et al., 1999; de Brouwer et al., 2002), we could not fully exclude the possibility that intracellular products were also contained in this fraction. Therefore, we routinely separated the high molecular weight polymers from low molecular weight material by overnight precipitation in cold ( − 20°C) ethanol (75% final concentration). The high molecular weight carbohydrates that are isolated using this protocol have been termed ‘bound’ or ‘attached’ EPS.

Chiovitti et al. (2004) have examined the composition and molecular arrangement of the glucose-rich carbohydrates that were extracted in 1 h using MilliQ water at 30 and 60°C. They showed that the carbohydrate pool was dominated by 1,3-linked glucose moieties characteristic of 1,3-β-D-glucans, and concluded that this carbohydrate was the intracellular storage glucan chrysolaminaran, which has a molecular size distribution of 4 – 13 kDalton. The possibility that other glucose-rich polymers in the EPS may have this common molecular arrangement was, thereby, excluded. Glucose-rich polymers larger than 13 kDalton have been identified from the warm-water extract of the diatom Navicula subinflata (Bhosle et al., 1995) as well as from natural microphytobenthos (de Brouwer & Stal, 2001). This shows that, although the carbohydrates in these studies consisted predominantly of glucose, they could not be attributed to chrysolaminaran. This conclusion is supported by our own unpublished observations that bound EPS isolated by precipitation in cold ethanol has a molecular weight > 100 kDalton. The internal storage glucan chrysolaminaran, does not precipitate in cold ethanol (Staats et al., 2000) and cannot, therefore, form part of this bound EPS. Since Chiovitti et al. (2004) did not apply size fractionation, which was an essential part of our method, they could not discriminate between laminaran and glucans of different molecular weight.

In addition, Chiovitti et al. (2004) studied different strains of diatoms and cultivated them under different conditions. Our extraction method for bound EPS was specifically developed for benthic epipelic diatoms isolated from intertidal mudflats. In these environments, diatoms are subjected to strong and rapid fluctuations in temperature up to a maximum of ca. 38°C (Blanchard et al., 1997). We have cultivated our strains at temperatures up to 35°C and were able to measure photosynthesis, CO₂-fixation and EPS-secretion at this temperature. Therefore, we are confident that the temperature used for the EPS extraction does not compromise our strains. Similarly, benthic diatoms experience osmotic stress on a regular basis, due to rainfall or evaporation, and contain osmoprotectants to deal with these high or low salinity events (Van Bergeijk et al., 2003). We do not know how the strains used by Chiovitti et al. (2004) reacted to their extraction conditions, but it is possible that these species were compromised during the warm water treatment. This may apply especially to the pelagic species studied that live under more constant environmental conditions than benthic diatoms in intertidal areas. We suggest that these authors should have used our strains and our complete EPS extraction method (and not just a part of it) before challenging our conclusions.

In order to relate the presence of 1,3-ß-D-glucan in the warm-water extract to the possible release of chrysolaminaran, Chiovitti et al. (2004) used immunocytochemistry and dye labelling experiments with DIBAC₄(3). It was shown that extraction of diatoms in MilliQ water at 30°C increased the fraction of DIBAC₄(3)-labelled cells to 100%, whereas no labelling was present in the original diatom cultures. In similar experiments, de Brouwer et al., (2002) found that the fraction of DiBAC₄(3)-labelled cells in a culture of Cylindrotheca closterium increased from 12% in the original culture to 44% after extraction with distilled water at 30°C. These two sets of results are not contradictory, even though Chiovitti et al. (2004) found a higher percentage of DiBAC4(3)-labelled cells. Cell labelling with DiBAC₄(3) provides information on the permeability of cell membranes but does not indicate the release of chrysolaminaran into the warm-water extract. Neither does it discriminate between viable and dead cells (Laflamme et al., 2004). The increase in DiBAC₄(3)-labelled cells after warm-water extraction may well have been due to osmotic shock. It is known that benthic diatoms permeabilize their membranes and release dimethyl sulfonopropionate (DMSP) into the medium upon osmotic shock. However, such cells remain viable and resynthesize the osmoprotectant DMSP upon salinity upshock (Van Bergeijk et al., 2003). Since Chiovitti et al. (2004) also classified most of their diatoms as alive and labelled after warm-water extraction, cell membranes were probably permeabilized but not severely compromised. This is in agreement with Staats et al., (1999), who showed that water extraction at 30°C did not result in loss of cellular protein content, although this did occur when stronger extraction methods (e.g. at 70°C) were applied.

Chiovitti et al. (2004) used a monoclonal antibody raised against 1,3-β-D-glucan to confirm its localization inside the vacuoles of diatom cells. While this is an elegant method, the observation that chrysolaminaran was present inside the cells is not surprising since this glucan is the common storage product in diatoms (e.g. Darley, 1977). Nevertheless, these results do not provide evidence for the release of chrysolaminaran upon extraction with warm water. For this purpose, immunolabelling experiments should have been conducted with warm-water extracted diatom cells to demonstrate that immunolabelling had indeed decreased within the cells. In addition, Chiovitti et al. (2004) report that immunolabelling only occurred inside the cells while extracellular immunolabelling above background levels was not apparent. However, careful observation of the images (especially for Thalassiosira pseudonana) showed immunolabel outside the cells, but not in the corresponding negative controls. This may be a critical observation, since it suggests the presence of extracellular 1,3-ß-glucan. This was not discussed by Chiovitti et al. (2004). Furthermore, cells were prepared for TEM by freeze substitution, which minimized the risk of displacing cell contents or disruption of the cell ultrastructure. The possibility of modifying or removing extracellular components during sample preparation was however not discussed.

In summary, we conclude that the results presented by Chiovitti et al. (2004) do not substantiate their conclusions that ‘bound’ or ‘attached’ EPS fractions should be re-interpreted as intracellular chrysolaminaran. Their results cannot be extended to ‘bound’ or ‘attached’ EPS because they used different methods and applied them to different diatom species. Furthermore, neither the labelling of cells with DIBAC₄(3) nor the immunolabelling experiments demonstrated that chrysolaminaran alone was dissolved in the water extract.