Diatom cultures, sample preparation, and EPS extraction
P. tricornutum (UTEX 646) was selected for culturing in autoclaved f/2 and f/2-Si media (salinity of 26) at a temperature of 19 + 1 oC with a light cycling of 14 h : 10 h under a saturating irradiance of 100 umol quanta m-2 s-1. In order to deplete the diatom of Si supply, cultures were transferred into f/2-Si medium over at least six generations by harvesting cells (2694 g, 30 min) and resuspending them in fresh f/2-Si medium. Sterile polycarbonate bottles were also used to prevent Si supply from glassware. The growth status of P. tricornutum was monitored by changes in optical density at 750 nm. Cells, frustules, and EPS were collected when P. tricornutum reached the stationary phase.

Laboratory cultures of P. tricornutum were centrifuged (2694 x g, 30 min) and filtered (0.2 um) to collect the whole cells. The frustules were repeatedly treated by using a hydrogen peroxide (30%, room temperature) treatment until bubbles were no longer generated, followed by concentrated nitric acid (HNO3) digestion (85oC, 1 h) to remove organic matter adopted from Robinson et al. (2004). 

The resulting organic carbon (C), nitrogen (N), and sulfur (S) contents of the cleaned frustules were measured using a Perkin Elmer CHNS 2400 analyzer to ensure the removal of organic materials using cysteine as a standard according to Guo and Santschi (1997).

EPS extraction was followed the procedures described in Xu et al. (2011b), which minimize cell rupture and molecular alterations and maximize extraction efficiency. EPS here is referring to those biopolymers that are attached on the diatom frustules. Hereafter, EPS Si+ and EPS Si2 denote the EPS extracted from diatoms cultured under Si-replete (f/2 medium) and Si-depleted (f/2-Si medium) conditions, respectively. Briefly, laboratory cultures were centrifuged (2694 x g, 30 min) and filtered (0.2 um) when diatoms reached stationary phase. The diatom cells were soaked with 0.5 mol L-1 sodium chloride (NaCl) solution for 10 min and followed by centrifugation at 2000 x g for 15 min to remove the medium and weakly bound organic material on the cells. The pellet from previous step was resuspended in a new 100 mL 0.5 mol L-1 NaCl solution and stirred gently overnight at 4oC. The resuspended particle solution was ultracentrifuged at 12,000 x g (30 min, 4uC), and the supernatant was then filtered through a 0.2 um polycarbonate membrane. The filtrate was desalted and collected with a 1 kDa cutoff cross-flow ultrafiltration and diafiltration membrane and then freeze-dried for later use.

Characterization of exopolymeric substances

After partitioning EPS collected from lab cultures into aliquots for freeze-drying, subsamples were analyzed for individual components. Concentration of total carbohydrate (TCHO) concentration was determined by the TPTZ (2,4,6-tripyridyl-s-triazine) method using glucose as the standard, and uronic acids were measured by the meta-hydroxyphenyl method using glucuronic acid as the standard (Hung and Santschi 2001). Protein content was determined using a modified Lowry protein assay, using bovine serum albumin (BSA) as the standard (Pierce, Thermo Scientific). C, N, and S contents were determined as described above. Iron was measured using an atomic absorption spectrometer (Varian) after overnight digestion with 12 mol L^-1 HNO3 at 85^oC (Von Loon 1985). To evaluate the protein size distribution pattern in EPS Si+ and EPS Si2, sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS-PAGE) was carried out according to Sambrook et al. (1989) using standard molecular weight markers (Dual Xtra Standards, Bio-Rad).

Fourier transform infrared spectroscopy (FTIR) was used to characterize samples using a Varian 3100 model interfaced with a single reflection horizontal attenuated total reflectance (ATR) accessory (PIKE Technologies). A diamond plate was used as the internal reflection element. A freeze-dried EPS sample was mounted at the surface of the diamond. Absorbance spectra from 800 to 2000 cm21 were collected and integrated using Varian Resolution Pro 4.0 software. ATR-FTIR spectroscopy provides a noninvasive way to quickly gain information about the contents of major secondary structures of biopolymers (Xu et al. 2011b; Jiang et al. 2012). Major infrared (IR) peaks were assigned according to Xu et al. (2011b) and Jiang et al. (2012). Characteristic bands found in the IR spectra of proteins and polypeptides include the amide I (1652–1648 cm^-1) and amide II (1550–1548 cm^-1) band. The absorption associated with the amide I band leads to stretching vibrations of the C=O bond of the amide, and absorption associated with the amide II band leads primarily to bending vibrations of the N-H and C-N bond. The symmetric stretching peak due to deprotonated carboxyl groups is observed at 1400 cm^-1 along with the CH₂ bending mode at 1455 cm^-1. In the 800–1200 cm^-1 regions, responses from C-O, C-O-C, P-O-P, C-O-P, and ring vibrations of the main polysaccharide functional groups are present in polysaccharide mixtures. The peaks at 1241 and 1113 cm^-1 correspond to P-O stretching in phosphate groups.

Isoelectric focusing of radionuclide-labeled EPS

EPS Si+ and EPS Si2 were incubated separately with 234Th, 233Pa, 210Pb, 210Po, and 7Be, respectively, for  subsequent IEF electrophoresis separation to determine the pHIEF of selected radionuclide binding ligands (Alvarado Quiroz et al. 2006). Briefly, radiolabeled biopolymers and 140 mL of rehydration solution were loaded onto an immobilized pH gradient strip (General Electric Healthcare Immobiline Drystrip, pH 3–10, 11 cm) and were reswelled overnight. Afterward, the strip was loaded into the device for isoelectric focusing for 17.5 h. The strip was then cut into 11 1 cm pieces and followed by 1% SDS extraction overnight. Five radionuclide activities of each fraction were subsequently analyzed. Due to the limited amount of each strip fraction, selected chemical compositions (TCHO, proteins, and Fe) of individual fraction were characterized as described above.