All methods below are described in Tatters et al. 2012.

Culture, Media, and Sampling
Pseudo-nitzschia fraudulenta was isolated from public nearshore water collected at 34.08 N, 119.05 W in Ventura County, California in March, 2010. The culture was grown at 16 degrees C on a 12-h light:12-h dark cycle in modified f/2 enriched seawater growth media (Guillard 1975) under 90 photons per square meter per second of cool white fluorescent illumination. Silicate (Si(OH)4) final concentrations in the medium were 10.6 uM for the Si(OH)4-limited treatment and 106.1 uM for the nutrient replete treatment. Dissolved CO2 concentrations were controlled by gentle bubbling with commercially prepared air/CO2 mixtures (Praxair Gas).

Cell Counts and Growth Rates
Steady state semi-continuous culture methodology was employed to maintain cultures in exponential state and specific growth rates were calculated as the equation:
u = (lnNb - lnNa) / (tb - ta)
where Na and Nb are the average cell density at times ta (directly after a dilution) and tb (directly before the next day's dilution). Growth rate as determined by bi-weekly microscopic cell counts and in vivo fluorescence determined the dilution rate of each bottle. Cells of P. fraudulenta preserved in acidified Lugol's solution were vortexed and enumerated by direct counts using an Accu-Scope 3032 inverted microscope according to the Utermohl method (1931). A minimum of 300 cells were counted to guarantee a 95% confidence interval with +/- 11.5% accuracy (Guillard 1973). Cultures were acclimated for a period of three months to the respective experimental conditions prior to splitting into triplicates, which were then further acclimated for one month (5 to 17 cell divisions) prior to final sampling.

Domoic acidy by high performance liquid chromatography
High performance liquid chromatography with ultraviolet detection (HPLC-UV) of domoic acid (DA) was performed using a SCL-10ADVP controlled system (Shimadzu). The UV detector was programmed for sample and reference wavelengths of 242 and 280 nm respectively and the system was operated by EZ START software version 7.4 SP1 (Shimadzu). Cellular concentrations of domoic acid were determined according to Mafra et al., (2009) with slight modifications (Mafra et al 2009). Culture subsamples of 10 to 20 ml were carefully measured and cells were collected by gentle filtration on 25 mm GF/F filters (Whatman). The filters were stored in the dark at -20 degrees C. Filters were subsequently subject to sonication in 10% aqueous methanol for 2 min. at 40W in a water bath. The cell extracts were then clarified by centrifugation at 3000 x g for 10 min at 4 degrees C. The pellet was discarded and the clarified extracts were transferred to 300 ul polyspring inserts (National Scientific) placed inside clean 2.0 ml Target DPTM vials (National Scientific). Prior to analysis, all samples were treated with 0.15% trifluoroacetic acid (TFA).

The chromatographic separation was carried out on a reversed phase Luna C18 (2) column (3 um, 2x100 mm, Phenomenex) at 25 degrees C with a mobile phase system consisting of water with 0.1% TFA (A), and acetonitrile (MeCN) with 0.1% TFA (B). The elution gradient began with a 10–35% B transition over 10 min, then was held at 35% B until 15 min, followed by a subsequent decrease to 10% B at 16 min, and held at 10% B. The flow rate was 0.2 ml per min and the injection volume 5 to 10 ul. Quantification of domoic acid was determined using certified reference material CRM-domoic acid-e obtained from the National Resource Council, Canada at a range of concentrations. Calibration curves of CRM-domoic acid-e were determined by linear regressions (r squared values=0.99) for each sample treatment. Interpolation from the standard curves was used to calculate the amount of compound injected from the peak areas of each sample under the same experimental conditions. Domoic acid per cell concentrations were determined by the ratio of reconstituted volume (300 ul) to the volume injected (5 to 10 ul) and dividing by the total cell count in the original sample. Reported values represent means of the results (n= 3). Domoic acid production rates were calculated by multiplying the growth rate by toxin per cell.

Carbonate Buffer System analysis
Dissolved inorganic carbon analysis was performed using a CM5230 CO2 coulometer (UIC). pH was determined on freshly collected samples using a calibrated Orion 5-star plus pH meter using an NBS buffer system with three-point calibration. The pCO2 in the experimental media was calculated from these two parameters using CO2SYS software. For clarity, pCO2 treatments in the cultures are referred to in the data and text (Tatters et al. 2012) using rounded-off values of 200 ppm, pre-industrial atmospheric levels; 360 ppm, modern atmospheric levels: and 765 ppm, projected year 2100 levels.

Carbonate buffer system parameters:

Treatment	pH	DIC (umol/kg)	Calculated pCO2 (ppm)
190 ppm	8.43	1965	198
380 ppm	8.23	2107	357
750 ppm	7.95	2249	764

Chlorophyll a, POC, PON, and Biogenic Silica
For chlorophyll a (Chl a) measurements, samples were filtered in duplicate onto 25 mm GF/F filters. Five ml of 90% acetone was later added and each vial was allowed to extract overnight in the dark at -20 degrees C. After 24 hours, Chl a was determined using a Turner Designs 10-AU fluorometer. For particulate organic carbon (POC) and nitrogen (PON), sample volumes of 20 ml were collected onto pre-combusted (450 degrees C for 5 h) GF/F glass fiber filters, stored at -20 degrees C, and dried at 55 degrees C before analysis. Molar POC and PON content was analyzed using a 4010 Costech Elemental Combustion System calibrated with methionine and atropine as reference materials. For measurement of biogenic silica, sample volumes of 20 ml were collected onto 0.6 uM polycarbonate filters until analysis. Cellular biogenic silica (BSi) was measured according to Brzezinski and Nelson (1995).

P_fraudulenta_exp.csv