An efficient and scalable extraction and quantification method for algal derived biofuel Authors: Egan J. Lohman, Robert D. Gardner, Luke Halverson, Richard E. Macur, Brent M. Peyton, & Robin Gerlach NOTICE: this is the author’s version of a work that was accepted for publication in Journal of Microbiological Methods. Changes resulting from the publishing process, such as peer review, editing, corrections, structural formatting, and other quality control mechanisms may not be reflected in this document. Changes may have been made to this work since it was submitted for publication. A definitive version was subsequently published in Journal of Microbiological Methods, 94, 3, September 2013. DOI# 10.1016/j.mimet.2013.06.007. Lohman EJ, Gardner RD, Halverson L, Macur RE, Peyton BM, Gerlach R, "An efficient and scalable extraction and quantification method for algal derived biofuel," Journal of Microbiological Methods, September 2013 94(3): 235–244. Made available through Montana State University’s ScholarWorks scholarworks.montana.edu An efficient and scalable extraction and quantification method for algal derived biofuel Egan J. Lohman, Robert D. Gardner, Luke H Peyt Montan g a A B S Keywor Triacylg tty acid Micro nd cts such a sp nce slight t th ent a met nd eir charac cu nes which esp ed; inform algae candidates as a biodiesel source. Traditional methods of analyz tal conditions. Here we detail a new method to extract lipids from live cultures of microa ds, mono-, di- or tri-acylglycerides) and carbo ed into FAMEs and directly compared to tota of total FAMEs derived from extractable lip nd phospholipids, ste-rols, etc.). This approach ive abundance, is responsible for each FAM om Phaeo l Chlo ris. The m hly re me course e-res to obtain 2 h. Con current percen politica resourc fuels (D 2008; biofuel market sources mass r 2008). food” (Ferrel in nd om eat Es) uel the current distribution s. Biodiesel has similar nd a higher combustion um I N T, they hold great potential for truly displacing traditional due to their high biomass productivity, relatively small land equirements and high lipid yield (Chisti, 2007; Hu et al., efficiency than gasoline (Demirbas, 2007). However, if biodiesel is to replace a large part of petroleismukes et al., 2008; Greenwell et al., 2010; Groom et al., Hill et al., 2006; Sheehan, 1998). Although algae-derived s command a relatively small percentage of this emerging since no modifications have to be made to infrastructure or diesel combustion engine combustion properties to petroleum diesel at-age being imported from foreign sources (EIA, 2012). The l, economic and environmental controversies over these es have resulted in increased interest in advanced alternative mono, di, and tri-acylglycerides (MAG, DAG, TAG), derived fr plant or animal sources, react with methanol in the presence of h and base or acid, to produce fatty acid methyl esters (FAM (Laurens et al., 2012a). Biodiesel is considered a “drop-in” fFurther, algal fuels do not directly cont debate because they can grow in non-a l and Sarisky-Reed, 2010). in the United States with a significant Biodiesel is the end result of a transesterification reaction which lipids such as phospholipids, free fatty acids (FFA), asumption rates of traditional fossil fuels ly exceed 18 million barrels per dayds: lycerol (TAG) - Microalgae - Biodiesel - Fa dactylum tricornutum, the mode ethod is shown to be robust, hig of culturing, thus providing tim ing analytical results is less than R O D U C T I O N methyl ester (FAME) Nile Red fluorescence - GC–FID/MS which uses microwave energy as a reliable, single-step cell disruption technique lgae. After extractable lipid characterization (including lipid type (free fatty aci n chain length determination) by GC–FID, the same lipid extracts are transesterifi l biodiesel potential by GC–MS. This approach provides insight into the fraction ids compared to FAMEs derived from the residual fraction (i.e. membrane bou can also indicate which extractable lipid compound, based on chain length and relat E. This method was tested on three species of microalgae; the marine diat rophyte Chlamydomonas reinhardtii, and the freshwater green alga Chlorella vulga producible, and fast, allowing for multiple samples to be analyzed throughout the ti olved information regard-ing lipid quantity and quality. Total time from harvestinging these precursor molecules are time intensive and prone to high degrees of variation between species and experimen ation that is fundamental for identifying preferred microon, Robin Gerlach ⁎ a State University, Department of Chemical and Biological Engineerin T R A C T algae are capable of synthesizing a multitude of compou s omega-3-fatty acids. However, accurate analysis of the variations in saturation and carbon chain length can affec hod that allows for fast and reliable extraction of lipids a terization using gas chromato-graphic analysis with a fo range of biologically synthesized compounds is likely rend in the “fuel vs. rable environments alverson, Richard E. Macur, Brent M. nd Center for Biofilm Engineering, Bozeman, MT 59717, United States s including biofuel precursors and other high value produ ecific compounds produced by microalgae is important si e quality, and thus the value, of the end product. We pres similar compounds from a range of algae, followed by th s on biodiesel-relevant compounds. This method determi onsible for each fatty acid methyl ester (FAME) produchydro-carbon based fuels, the compositional makeup of biodiesel will be-come highly important since biodiesel can have poorer performance than petroleum-based diesel (Hunter-Cevera et al., 2012). For exam-ple, (i) the viscosity of fully saturated hydrocarbons increases signifi-cantly at low temperatures and can lead to operational issues and (ii) short chain fatty acid methyl esters tend to be more susceptible to oxidation which can lead to corrosion, resulting in reduced engine durability (Xue et al., 2011). Therefore a comprehensive and diverse Gardner et al., 2011, 2013). Hence, Nile Red-based measurements may underestimate the total biofuel potential of algal species that produce high fractions of FFA, MAG, DAG or membrane lipids. Additionally, the Nile Red assay is qualitative (possibly semi-quantitative) and cannot be compared between species. In contrast, gravimetric analysis can overestimate total biofuel potential by including non-fuel components (e.g. chlorophyll) in the total weight (Laurens et al., 2012b). Gas chro- matographic analysis of FAMEs has proven to be a reliable method for quantifying total biodiesel potential (Eustance et al., in press; Laurens et al., 2012b), but fails to identify the biological precursor molecules. Here, a new method is described which allows for high-throughput harvesting and extraction of live cultures by utilizingmicrowave energy for a one-step cell disruption-and-extraction of lipid precursors (FFA, MAG, DAG, TAG). Microwave energy has recently been shown to be an effective cell disruption technique, generating comparative or better lipid yields than more traditional methods such as sonication or bead beating (Lee et al., 2010; Patil et al., 2012; Prabakaran and Ravindran, 2011). Additionally, extraction of lipids from live culture significantly reduces process time and decreases the potential for degradation of intracellular lipid compounds. A portion of the lipid extract is analyzed by GC coupled with flame ionization detection (GC–FID) to identify the lipid class and carbon chain length of the fatty acid(s). In parallel, a portion of the lipid extract is transesterified and analyzed by GC–MS to identify FAMEs derived from extractable lipid. Additionally, total biodiesel potential is determined by direct in situ transesterification of live cultures. Comparisons can bemade between extractable lipid precur- sors, FAMEs derived from extractable lipids and total biodiesel potential by contrasting the carbon chain length and saturation of each molecule respectively. This approach was demonstrated for three different fre- quently usedmicroalgae species; themodel Chlorophyte Chlamydomonas reinhardtii, the marine diatom Phaeodactylum tricornutum and the fresh- water green alga Chlorella vulgaris. These three organisms have been extensively studied (e.g. Gardner et al., 2012; Merchant et al., 2007; Stephenson et al., 2010; Valenzuela et al., 2012) and represent a diverse set of microalgae which are commonly used for biofuel investigations (Gardner et al., 2013; Liu et al., 2011; Spiekermann et al., 2003). The method described herein is relatively simple, fast, utilizes fairly standard equipment, and results in a comprehensive lipid profile within 2 h of harvesting. 2. Materials and methods 2.1. Microalgae strains 2.1.1. Phaeodactylum tricornutum P. tricornutum strain Pt-1 (CCMP 2561) (Pt-1) was acquired from the Provasoli-Guillard National Center for Culture of Marine Phytoplankton (East Boothbay, Maine, USA) and accessions of P. tricornutum have pre- viously been described (Martino et al., 2007). Pt-1 was cultured on ASP2 medium (Provasoli et al., 1957) amended with 50 mM Tris buffer (Sigma-Aldrich, USA), pKa 7.8, to maintain a stable pH throughout culturing. Cultures were continuously sparged with ambient air andfeedstock of lipid precursors along with the derived fatty acid methyl esters will be necessary to achieve the desired properties of the biodiesel fuel (Knothe, 2005). Traditionally, microalgae species have been screened for strains that produce a high abundance of TAGusingfluorometric techniques such as the Nile Red assay (Cooksey et al., 1987) and total lipid extracts have been quantified gravimetrically (Bligh and Dyer, 1959) or by gas chro- matography after transesterification (e.g. Guckert and Cooksey, 1990). However, these methods are limited by their inherent inability to iden- tify all, or only those, compounds which may be utilized for fuel. Nile Red, for instance, roughly correlates with the amount of TAG precursors per cell (Chen et al., 2009; Cooksey et al., 1987; Elsey et al., 2007;amended with 25 mM of sodium bicarbonate (final concentration)just prior to nitrate depletion in the medium to induce TAG accumula- tion as demonstrated by Gardner et al. (2012). 2.1.2. Chlorella vulgaris C. vulgaris, UTEX 395 (University of Texas at Austin) was cultured on Bold's basal medium (Nichols and Bold, 1965) with pH adjusted to 8.7 prior to autoclaving. Cultures were sparged with 5% CO2 in air during the light hours. 2.1.3. Chlamydomonas reinhardtii C. reinhardtii CC124, obtained from the Chlamydomonas Center (Minnesota State University, Minneapolis MN), was kindly provided by John Peters, Department of Chemistry and Biochemistry at Montana State University, and was cultured on Sager's minimal medium (Harris, 1989). Cultures were sparged with 5% CO2 (v/v) until just prior to ammonium depletion, at which time the gas sparge was switched to ambient concentrations of CO2 (0.04%; v/v) and 50 mM of sodium bicarbonate (final concentration) were added to induce TAG accumula- tion (Gardner et al., 2013). 2.2. Culturing conditions All organisms were checked for bacterial contamination by inocu- lation into respective medium supplemented with 0.05% yeast extract and 0.05% glucose and incubation in the dark. Experiments were conducted in triplicate batch cultures using 70 × 500 mm glass tubes containing 1.2 L medium submersed in a water bath to control temperature. Rubber stoppers, containing ports for aeration and sampling, were used to seal the tubes. Temperature was maintained at 24 °C ± 1 °C. Light (400 μmol photons m−2 s−1) was maintained on a 14:10 light–dark cycle using a light bank containing T5 tubes. Aeration (400 mL min−1) was supplied by humidified compressed air (supplemented with 5% CO2 (v/v) for C. reinhardtii and C. vulgaris) and controlled using individual rotameters for each bioreactor (Cole-Parmer, USA). ACS grade sodium bicarbonate was used in all experiments involving bicarbonate addition (Sigma-Aldrich, St. Louis, MO). 2.3. Culture analysis Cell concentrations were determined using an optical hemacy- tometer with a minimum of 400 cells counted per sample for statisti- cal reliability. Cell dry weights (CDWs) for C. reinhardtii and C. vulgaris were determined by harvesting 30 mL of culture into a tared 50 mL Falcon tube (Fisher Scientific, Palatine, IL) followed by centrifugation at 4800 ×g and 4 °C for 10 min (Thermo Scientific, Sorvall Legend XTR, Waltham, MA). The concentrated biomass was rinsed with deionized H2O (diH2O), 18 MΩ, to remove media salts and excess bicarbonate, before centrifuging again. Remaining algae pellets were frozen and lyophilized (Labconco lyophilizer, Kansas City, MO) for 48 h. CDWs were calculated by subtracting the weight of the biomass-free Falcon tube from the weight of the Falcon tube with freeze-dried biomass. CDWs for P. tricornutum were determined by filtering 10 mL of culture using 1 μm pore size glass fiber filters (Fisher Scientific) to collect the biomass. The biomass was washed with diH2O to remove media salts and excess bicarbonate. Algal cells were dried at 70 °C until the weight of the filter (plus biomass) remained constant. CDWswere calculated by subtracting the dry weight of the clean filter from the oven-dried weight of the filter with biomass. This method was employed for the diatom due to the concern that the fragility of the organism's silicon-based frustules may result in cell disruption during the centrifugation and lyophilization steps utilized for the Chlorophytes, thereby underestimating overall CDWs. This technique was not employed for the Chlorophytes because it has been shown by us and others (Laurens et al., 2012a) that the freeze drying method is more reliable for green microalgae. 2.4. Analysis of media components Medium pH was measured using a standard bench top pH meter. Ammonium and nitrate concentrations were measured using Nessler reagent (Hach, Loveland, CO) and ion chromatography (Dionex, Sunnyvale, CA), respectively, using previously described protocols (Gardner et al., 2013). 2.5. Harvesting At the end of each experiment, cultures were transferred to a grad- uated cylinder and total harvest volume was recorded. Two aliquots of 45 mL were dispensed into 50-mL Falcon tubes (Fisher Scientific, Palatine, IL) and centrifuged (Thermo Scientific, Sorvall Legend XTR, Waltham, MA) at 5800 ×g and 4 °C for 10 min. Concentrated biomass was then transferred to Pyrex test tubes with Teflon lined screw caps (Kimble-Chase, Vineland, NJ) for microwave extraction of lipid precur- sors (Section 2.6.1) and direct in situ transesterification for total biodie- sel potential (Section 2.7), respectively. The remaining ~1 L of culture was centrifuged (Thermo Scientific, Sorvall Legend XTR, Waltham, MA) at 5800 ×g and 4 °C for 10 min. The supernatant was discarded and the remaining concentrated biomass was combined in a 50 mL Falcon tube, washed with diH2O, re-centrifuged, and the supernatant discarded. The washed biomass pellets were frozen at −80 °C, lyophilized, and stored at −20 °C for subsequent cell disruption/lipid extraction method comparisons (Section 2.6). 2.6. Cell disruption and extraction of lipids Six different cell disruption and lipid extraction techniques were tested in biological triplicate to compare their extraction yields against yields from microwave extraction on live cultures. Unless otherwise stated, all methods were performed using approximately 30 mg of dried biomass, and the final extract was supplemented with 10 μL/mL of a 10 mg/mL internal standardmix (isooctane in chloroform) to monitor instrument performance. All mixtures involving solvent and heat were carefully sealed in Pyrex test tubes with Teflon lined screw caps (Kimble-Chase, Vineland, NJ). Final extracts were analyzed for lipid precursors by gas chromatography–flame ionization detection (GC–FID) (Section 2.9.2). 2.6.1. Microwave extraction of lipid from wet biomass 45 mL of culture were (45 milliliters) transferred into a 50 mL Fal- con tube and centrifuged at 4800 ×g and 4 °C for 10 min. The super- natant was discarded and the pellets were re-suspended in 3 mL diH2O to facilitate transfer from Falcon tubes to screw-cap glass tubes for subsequent lipid extraction. Three mL of chloroform were added to the algal slurry in the test tubes, and the mixture was ho- mogenized by vortexing for 10 s. Caps were securely fastened and samples were microwaved (Sharp Carousel, Mahwah, NJ) at 1000 W on power level 1 (~100 °C and 2450 MHz) for 1 min and then allowed to cool for 5 min. This process of vortexing for 10 s, microwaving for 1 min and cooling for 5 min was repeated until op- timal extraction efficacy was established for each culture (Supple- mental data). A total of 3 min microwave exposure was required for samples of P. tricornutum and C. reinhardtii. Samples of C. vulgaris re- quired a total of 10 min microwave exposure to completely disrupt cells and release lipids into the organic solvent. Transmitted light mi- croscopy was used to verify complete disruption of cells, and epifluorescence microscopy was used to verify all extractable lipid (stained with Nile Red) had been released into the organic solvent (Nikon Eclipse E800).It should be noted that extreme caution should be taken when applying microwave energy to an organic solvent under high heat and pressure. To address this inherent safety concern this method has since been modified by adding only 5 mL diH2O to the algal pellet before transferring the slurry to screw cap test tubes, and test tubes were further secured by placing them in a microwaveable Tupperware container. Samples were then microwaved at 1000 W power level 4 (2450 MHz) for a continuous 5 min and allowed to cool. Three mL of chloroform were added to each test tube when they had reached room temperature. Test tubes were vortexed for 10 s and heated at 90 °C for 5 min to improve partitioning of the extracted lipids into the organic phase. Equivalent results were obtained with both versions of the method. The latter (i.e. water only-based extraction) being inherently safer, it is the recommended approach. In either case, samples were then centrifuged at 1200 ×g for 2 min to enhance phase separation. One mL of the organic phase was removed from the bottom of the test tube using a glass syringe and transferred to a 2 mL GC vial (Fisher Scientific, Palatine, IL) for GC–FID analysis and subsequent transesterification (Sections 2.9.2 and 2.8, respectively). 2.6.2. Microwave extraction of lipids from dry biomass Microwave extraction of lipids from dry biomass was conducted as described in Section 2.6.1 except ~30 mg of dried algae biomass were used instead of live culture. Samples of P. tricornutum and C. vulgaris were extracted using the optimized microwave extraction technique with just diH2O and no organic solvent during the microwave step. Samples of C. reinhardtii were collected before the optimized method was established; therefore they were extracted with diH2O and solvent together in the microwave extraction step. 2.6.3. Sonication extraction of lipids from dry biomass Dried biomass was combined with 3 mL of chloroform in a dispos- able 15 mL boro-silicate test tube. Samples were sonicated with a Branson S-450D sonifier equipped with a microtip probe set to 80 W (Branson, Danbury, CT) for six 20 s pulses, with each pulse followed by a 1 min cool down period in ice water. Total sonication time was 2 min. Samples were then centrifuged at 1200 ×g for 2 min to pellet the residual biomass. One mL of the organic phase was removed from the test tube using a glass syringe and transferred to a 2 mL GC vial for GC–FID analysis. 2.6.4. Bead beating extraction of lipids from dry biomass Dried biomass was combined with 1 mL of chloroform in a 1.5 mL stainless steel bead beating microvial with silicone cap (BioSpec Products, Bartlesville, OK). Three types of beads (0.6 g of 0.1 mm zirconium/silica beads, 0.4 g of 1.0 mm glass beads, two 2.7 mm glass beads) were added to each vial before capping. A FastPrep bead beater (Bio101/Thermo Savant, Carlsbad, CA) was used to agitate the vials for six 20 s pulses at power level 6.5 followed by a 1 min cool down period between pulses. Total bead beating exposure time was 2 min. The mixture of solvent, residual biomass and beads was then transferred to a disposable 15 mL boro-silicate test tube, and the steel microvial was rinsed twice with 1 mL of chloroform, which was also added to the test tube, bringing the total solvent volume to 3 ml. Samples were then centrifuged at 1200 ×g for 2 min to pellet the residual biomass. One mL of the organic phase was removed from the bottom of the test tube using a glass syringe and transferred to a 2 mL GC vial for GC–FID analysis. 2.6.5. Heat induced extraction of lipids from dry biomass Dried biomass was combined with 3 mL of chloroform and 3 mL diH2O in a test tube. Samples were homogenized by vortexing for 10 s. Caps were securely fastened to test tubes and samples were heated in an oven at 95–100 °C for 1 h with vortexing every 10–15 min. Samples were then centrifuged at 1200 ×g for 2 min to enhance phase separa- tion. One mL of the organic phase was removed from the bottom of the test tube using a glass syringe and transferred to a 2 mL GC vial for GC– FID analysis. 2.6.6. Bligh and Dyer extraction of lipid from dry biomass The modified Bligh and Dyer method (as described by Guckert et al. (1988), without phosphate buffer) was used without any additional physical cell disruption steps. Briefly ~50 mg of dry biomass from each of the three organisms in triplicate was transferred to Pyrex test tubes with Teflon lined screw caps. 1.5 mL of chloroform and 3 mL of methanol were added to each test tube and sampleswere homogenized by vortexing for 10 s. Capswere securely fastened, sampleswere placed horizontally on a shaker table and allowed tomix continuously for 18 h. 1.5 mL diH2O and 1.5 mL chloroform were then added, and samples were vortexed for 10 s before being centrifuged at 1200 ×g for 2 min to enhance phase separation. One mL of the chloroform phase was removed from the bottom of the test tube using a glass syringe and transferred to a 2 mL GC vial for GC–FID analysis. 2.7. Direct in situ transesterification for FAME analysis Direct in situ transesterification of live cultures was conducted using a previously described protocol (Griffiths et al., 2010) with modifica- tions. FAME composition was analyzed using gas chromatography– mass spectroscopy detection (GC–MS) (Section 2.9.3). Briefly, 45 mL of culture were transferred into a 50 mL Falcon tube and centrifuged at 4800 ×g and 4 °C for 10 min. The supernatant was discarded and the pellets were re-suspended in 3 mL diH2O to facilitate transfer from Falcon tubes to screw-cap glass tubes for subsequent transesterification. The tubes were centrifuged at 1200 ×g to remove as much water from the remaining biomass as possible. One mL of tol- uene and 2 mL sodium methoxide (Fisher Scientific, Pittsburgh PA) were added to each test tube along with 10 μL of a 10 mg/mL standard mix (C11:0 and C17:0 TAG) to monitor transesterification efficiency of the TAG into FAMEs. Samples were heated in an oven for 30 min at 90 °C and vortexed every 10 min. Samples were allowed to cool to room temperature before 2 mL of 14% boron tri-fluoride in methanol (Sigma-Aldrich, St. Louis, MO) were added and samples were heated again for an additional 30 min. Samples were again allowed to cool be- fore 10 μL of a 10 mg/mL of C23:0 FAMEwere added to assess the com- pleteness of partitioning FAMEs into the organic phase. Additionally, 0.8 mL of hexane and 0.8 mL of a saturated salt water solution (NaCl in diH2O) were added. Samples were heated for 10 min to facilitate FAME partitioning into the organic phase, vortexed for 10 s and centrifuged at 1200 ×g for 2 min to enhance phase separation. One mL of the organic phase was removed from the top layer using a glass syringe and transferred to a 2 mL GC vial for GC–MS analysis. 2.8. Transesterification of extractable lipids from wet biomass Lipids extracted fromwet biomass (described in Section 2.6.1) were transesterified for FAME analysis. Briefly, 0.5 mL of the extract was removed from the GC vial and transferred to a screw-cap glass tube. One half mL of a 5% HCl in methanol mixture was added to each test tube along with 10 μL of a 10 mg/mL standard mix (C11:0 and C17:0 TAG) to assess transesterification efficiency. Samples were capped with a Teflon lined screw cap, heated for 1 h at 90 °C and vortexed every 10 min. Samples were allowed to cool to room temperature before 0.5 mL of a 15% salt water solution (NaCl in diH2O) was added along with 10 μL of 10 mg/mL C23:0 FAME to assess completeness of FAME partitioning into the organic phase. Samples were heated for an additional 10 min to help partition FAMEs into the organic phase, vortexed for 10 s and centrifuged at 1200 ×g for 2 min to enhance phase separation. One mL of the organic phase was removed from the bottom organic layer using a glass syringe and transferred to a 2 mL GC vial for GC–MS analysis.2.9. Lipid analysis 2.9.1. Nile Red Cellular TAG accumulation was monitored throughout the experi- ments using the Nile Red (9-diethylamino-5H-benzo(α)phenoxazine- 5-one) (Sigma-Aldrich, St. Louis, MO) fluorescence method (Cooksey et al., 1987), which has become a generally accepted screening meth- od for analyzing TAG in algal cultures both in academia and industry (e.g. Chen et al., 2009; Cooksey et al., 1987; da Silva et al., 2009; Elsey et al., 2007). Furthermore, Nile Red fluorescence has been correlated to TAG content using gas chromatography (GC) analysis on extracted lipid from C. reinhardtii, r2 = 0.998 (Gardner et al., 2013). Briefly, 1 mL aliquots were removed from cultures and assayed directly with Nile Red (4 μL from 250 μL/mL in acetone) or by diluting 1:5 with diH2O or salt water, for freshwater or marine cultures respectively, before staining with Nile Red (20 μL from 250 μL/mL in acetone). To maintain linearity of the Nile Red assay, dilution was required when population counts exceeded 1 × 107 cells/mL. Total Nile Red fluores- cence was quantified on a microplate reader (Bio-Tek, USA) utilizing a monochromator set to 480/580 nm excitation/emission wavelengths. A baseline sensitivity setting of 50 was experimentally determined to maximize the signal-to-noise ratio, while accommodating fluorescent level changes over 10,000 units. To minimize fluorescence spillover, black walled 96 well plates were loaded with 200 μL of sample. Unstained samples were used for background medium and cellular autofluorescence correction. 2.9.2. GC–FID GC–FID (Agilent 6890N, Santa Clara, CA) analysis was conducted using 1-μL splitless injections onto a 15 m (fused silica) RTX biodiesel column (Restek, Bellefonte, PA). The initial column temperature was held at 100 °C for 1 min, before being increased to 370 °C at a rate of 10 °C min−1. The injector temperature was held constant at 320 °C. Helium was used as the carrier gas and column flow was held at 1.3 mL min−1 for 22 min, increased to 1.5 mL min−1, held for 2 min, increased to 1.7 mL min−1 and held for 12 min. All flow rate increases were set to 0.2 mL min−2. Calibration curves were constructed for each of the following standards: C10:0, C12:0, C14:0, C16:0, C18:0, C20:0 FFA; C12:0, C14:0, C16:0, C18:0 MAG; C12:0, C14:0, C16:0, C18:0 DAG; and C11:0, C12:0, C14:0, C16:0, C17:0, C18:0, C20:0, C22:0 TAG (Sigma-Aldrich, St. Louis, MO) for quantification (r2 N 0.99). This GC method allows for an estimate of the amounts of FFA, MAG, DAG and TAG in a single analysis (Fig. 1a). 2.9.3. GC–MS GC–MS analysis was performed according to a previously published protocol (Bigelow et al., 2011). Briefly, 1 μL splitless injections were performed via an autosampler into a GC–MS (Agilent 6890N GC and Agilent 5973NetworkedMS) equippedwith a 30 m × 0.25 mmAgilent HP-5MS column (0.25 μm film thickness). The injector temperature was 250 °C and the detector temperature was 280 °C. The initial col- umn temperature was 80 °C and was increased to 110 °C at a rate of 8.0 °C min−1, immediately followed by a ramp at 14.0 °C min−1 to a final temperature of 310 °Cwhichwas held for 3 min before run termi- nation. Heliumwas used as the carrier gas and column flowwas held at 0.5 mL min−1. Quantities of FAMEs were determined by quantifying each response peak with the nearest eluting calibration standard based on retention time, using MSD ChemStation software (Ver. D.02.00.275), with additional analyses performed using a custom pro- gram described in Section 2.9.4. A 28-component fatty acid methyl ester standard prepared in methylene chloride (“NLEA FAME mix”; Restek, Bellefonte, PA) was used for GC–MS retention time identifica- tion and response curve generation (r2 N 0.99). It should be noted that C18:1, C18:2 and C18:3 co-elute on the columnused and have sim- ilar fragmentation patterns. They were therefore quantitated together as C18:1–3. 2.9.4. Lipid quantification and analysis Lipid quantification was performed using a custom program developed specifically for this purpose. Chromatogram data from the GC–FID and GC–MS were exported from the instrument software (Chemstation Ver. B.02.01-SR1) as a Microsoft Excel spreadsheet. The spreadsheet was then imported into a web-based software application written in Microsoft's ASP.NET framework using C# and MVC as the coding language and paradigm, respectively. Data were stored with unique identifiers in a Microsoft SQL database. Six point calibration Fig. 1. GC–FID chromatograms: (a) Standard mix of 22 analytical grade lipid standards (0.1 mg/mL); from left to right — C10:0, C12:0, C14:0, C16:0, C18:0 FFA, C12:0 MAG, C20:0 FFA, C14:0, C16:0, C18:0 MAG, C12:0, C14:0 DAG, C11:0 TAG, C16:0 DAG, C12:0 TAG, C18:0 DAG, C14:0, C16:0, C17:0, C18:0, C20:0, C22:0 TAG. (b) Olive oil (1 mg/mL) was used as a control (black line) against (c) a sample of the same concen- tration which was microwaved for 10 min (dashed line) to assess the potential for thermal degradation of TAG. Lipid extracted viamicrowave energy from (d) C. vulgaris, (e) C. reinhardtii and (f) P. tricornutum. Chromatograms are offset to allow for alignment with standards.curves were constructed using the LINEST function which is built into the ASP.NET framework library. Chromatogram peaks were quantified based on the closest standard, as determined by retention time of the sample peak and standard peak, respectively. All chromatograms were manually integrated and inspected prior to export to ensure accuracy and reliability. Manual calculations were performed periodically using Excel to verify software results. 2.9.5. Control samples Olive oil controls were analyzed by GC–FID to verify no thermal degradation of TAG was occurring when samples were exposed to microwave energy. Olive oil is predominantly comprised of C16 and C18 TAG compounds (Kiritsakis et al., 1998). A stock solution of olive oil dissolved in chloroform (1 mg mL−1) was prepared, and aliquots (n = 3) were run directly on the GC–FID as a control (Fig. 1b). To verify thermal degradation was not occurring, 3 mL of the stock solu- tion were transferred to a Pyrex test tube with Teflon lined screw caps and an additional 3 mL of diH2Owere added. Samples (n = 3)were ho- mogenized by vortexing for 10 s, caps were securely fastened, samples microwaved (Sharp Carousel, Mahwah, NJ) at 1000 W on power level 1 (~100 °C and 2450 MHz) for 1 min and then allowed to cool for 5 min. This process of vortexing for 10 s, microwaving for 1 min and cooling for 5 min was repeated 10 times, for a total of 10 min exposure to microwave energy. Samples were then centrifuged at 1200 ×g for 2 min to enhance phase separation. One mL of the organic phase was removed from thebottomof the test tube using a glass syringe and trans- ferred to a 2 mLGC vial (Fisher Scientific, Palatine, IL) for GC–FID analysis (Fig. 1c). Microorganisms are known to produce a wider range of mono-, di- and triacylglycerides than higher plants with carbon chain length varying by more than six carbon atoms and a range of saturation states (Brown et al., 2009). However, it is uncommon to find a glycerolipid synthesized by any organism with individual fatty acids deviating by more than two carbon atoms, albeit saturation states can differ significantly (Brown et al., 2009). Therefore, the challenge of separating and quantifying all potentially possible glycerolipids was addressed by developing calibration ranges based on the resi- dence time of each of the twenty even-chain length standards as described in Section 2.9.2 and further comparing these results to mass spectroscopy analysis of FAMEs derived from the extractable2.9.6. Statistical analysis A “two one-sided tests” (TOST) procedure (Richter and Richter, 2002) was performed on the control samples described in Section 2.9.5 to determine statistical equivalency between treated and untreated sam- ples. Equivalence was determined for each response peak based on peak area using n = 3 samples for each group. The TOST procedure requires an acceptance parameter “θ” to represent the hypothesized mean differ- ence boundary. Samplemeans are considered equivalent if |μ1 − μ2| ≤ θ with a 95% confidence interval. Statistical parameters were set at: θ = 0.5, α = 0.05. 3. Results and discussion 3.1. Chromatograph calibration, quantification and data compilation The quality of the results obtained using the methods described in this paper relies heavily on proper calibration, chromatogram quantification, quality control and quality assurance. Fig. 1 depicts GC–FID chromatograms for a surrogate mix of calibration standards (Fig. 1a), two olive oil standards (to check for the potential of ther- mal degradation during microwave treatment; Fig. 1b, c), and sam- ples from each of the three organisms used in this study (Fig. 1d, e, f). The surrogatemix of standards is comprised of twenty two analyt- ical grade lipid standards; six free fatty acids (C10:0–C20:0), four monoacylglycerides (C12:0–C18:0), four diacylglycerides (C12:0– C18:0), and eight triacylglycerides (C11:0–C22:0). All of these stan- dards are fully saturated and of even chain length, except for the C11:0 and C17:0 TAG. Six different calibration standard concentra- tions (0.005, 0.01, 0.05, 0.1, 0.25 and 0.5 mg/mL) were used to create the calibration curves for each compound listed above. Regression coefficients for each linear calibration curve were r2 N 0.99. Excellent peak shape and resolution were achieved for each of the standards in the mix (Fig. 1a). The algae extract chromatograms exhibit more noise, presumably due to a range of different lipids (chain length and saturation) being produced. As an example, consider the TAG range for C. vulgaris (Fig. 1d). The increase in baseline response during retention times of approximately 25–30 min is likely the re- sult of the presence of a number of TAG with different combinations of fatty acids of different chain lengths per molecule of TAG. In com- parison, the chromatogram of olive oil has a mostly flat baseline and exhibits only three groups of peaks in the TAG region. Olive oil consists predominantly of TAG with fatty acid chains consisting of 55–83% oleic acid (C18:1), 3.5–21% linoleic acid (C18:2) and 7.5–20% palmitic acid (C16:0) (Kiritsakis et al., 1998). The three groups of peaks in Fig. 1b and c for the olive oil standards are TAG with various combinations of these three fatty acids at the sn− 1, sn− 2 and sn− 3 positions of the glycerol backbone. The peak at 27.15 min (longest retention time (RT)) is a TAG with oleic acid at each position (sn− 1, sn− 2 and sn− 3). The left-most TAG peak (ret. time 25.76) is predominantly palmitic acid, but the shift in retention time compared to the C16:0 TAG stan- dard suggests oleic acid or linoleic acid must also be present at one position.lipids (Section 2.8). The details are discussed in detail in Section 3.3. the timecourse of culturing and implement adjustments as necessary. Furthermore, extensive processing times carry a greater risk of the sam- ple composition changing between collection and analysis. For example, it has been shown that TAG can be degraded during dark hours in some algal cultures (Bigelow et al., 2011; Gardner et al., 2012) making long processing times undesirable. Microwave extraction of lipid precursors from live culture required less than 2 h from time-of-harvest to generation of GC data and was shown to yield as much, if not more, lipid than the other techniques tested on these three organisms (Fig. 2). Further, relatively small for the sample and control are shown in Fig. 1 (b and c). Statistical equivalence was determined for each response peak as described in3.2. Comparison of various cell disruption and lipid extraction techniques To test the validity of microwave energy as a suitable and rapid method for cell disruption and lipid extraction, it was necessary to com- pare the results of the microwave technique to those from more tradi- tional methods (Fig. 2). Historically, the Bligh and Dyer (1959) gravimetric method for quantifying lipid has been considered the stan-The described GC–FID analysis of extractable lipid provides infor- mation regarding the lipid class and chain length (cf. Fig. 1), however, it is not capable of chromatographically separating a number of the acylglycerides based on the degree of fatty acid chain saturation. This limitation was addressed to a large extent by transesterifying the lipid extract and comparing the FAME profile to the relative abundance of the lipid classes, based on chain length and saturation. Fig. 2. Average and standard deviation of extractable lipid % (w/w) using various cell disruption techniques on P. tricornutum, C. vulgaris and C. reinhardtii: Microwave extraction of wet biomass (Wet MW), bead beating (Beads), microwave extraction ondry biomass (DryMW), heat induced extraction (Oven), sonicationprobe (Sonication), a modified Bligh and Dyer method and direct in situ transesterification for total biodiesel potential (Direct trans.) (n = 3).dard for lipid extraction (Gouveia and Oliveira, 2009; Laurens et al., 2012b; Singh and Kumar, 1992; Teixeira, 2012). However, it has been shown that the Bligh and Dyer method can produce significantly lower estimates of lipid content, and this underestimation tends to increase significantly with increasing lipid content of the sample (Iverson et al., 2001). More recently, researchers have demonstrated that cell lysis bymechanical means such as sonication and bead beating aids in lipid extraction (Lee et al., 1998, 2010; Zheng et al., 2011). However, these techniques require significant processing time due to limitations of the instruments and extensive preparation of sample and materials. For example, biomass must first be frozen and lyophilized, requiring up to 48 h or more from the time of harvest until processing can commence, and often only one sample can be processed at a time. These factors can limit one's ability to acquire lipid profiles throughout Table 1 Comparison of mean and standard deviation of final culture growth and TAG accumulation, m Organism Dry weight (g L−1; DCW)a Final cell density (×106 cells mL−1) P. tricornutum 0.18 ± 0.06 8.65 ± 0.49 C. vulgaris 0.88 ± 0.05 200 ± 10 C. reinhardtii 0.9 ± 0.02 7.55 ± 0.3 a Dry cell weight (DCW) determined gravimetrically with lyophilized biomass. b Calculated by fluorescence signal/cell density × 10,000 (scaling factor).Section 2.9.6 with 95% confidence. 3.3. Comparison of extractable lipid-derived FAMEs to total biodiesel potential Another significant benefit of this method is the ability to deter- mine which fraction of total biodiesel potential was derived from extractable lipid precursors (e.g. FFA, MAG, DAG, TAG) as opposed to residual precursors (e.g. membrane lipids). Additionally, the lipid class itself can be determined by the carbon chain length of the extractable lipid-derived FAMEs. This advancement was achieved due to the relative simplicity of extracting lipids from live cultures using microwave energy, transesterifying the extract and quantifying both the extract and transesterified extract by GC–FID and GC–MS, respectively. This approach results in two chromatograms for each sample whose comparison allows for an estimate of the lipid class and carbon chain length of the extractable lipid precursors as well as the carbon chain length and saturation of the extractable lipid- derived FAMEs. The comparison of the results of extractable lipid- derived FAMEs with direct in situ transesterified FAMEs (Griffiths et al., 2010, which was modified here to allow for the processing of small quantities of live cultures), makes it possible to assess which fraction onitored by Nile Red fluorescence, of P. tricornutum, C. vulgaris and C. reinhardtii (n = 3). Final total Nile Red fluorescence (×103 units) Final Nile Red specific fluorescence (units cell−1)b 9.71 ± 0.27 1.12 ± 0.03 4.06 ± 0.6 0.02 ± 0.003 3.37 ± 0.16 0.45 ± 0.003volumes of culture (usually b50 mL) are necessary for each sample point, resulting in the ability to potentially collect multiple data points throughout even lab scale experiments. Results from these techniques were also compared against results from direct in situ transesterification of live culture. Exhaustive direct in situ transesterification converts free and bound fatty acids into FAMEs (Laurens et al., 2012b), thereby generating larger amounts of FAMEs than the other techniques due to the additional conversion of fatty acids from non-extractable cellular sources (i.e. membrane bound lipids, chlorophyll, etc.). It is important to assess potential losses or degradation of lipids due to the extraction technique. There was a concern that high heat gener- ated during the microwave disruption process may result in thermal degradation of lipid compounds such as TAG. Hence, samples of olive oil were subjected to the microwave extraction process and compared to the untreated olive oil sample. Briefly, 3 mL of a 1 mg/mL (olive oil/ chloroform) mixture was microwaved for 10 min according to the protocol described in Section 2.9.5. One mL of this mixture was re- moved and analyzed byGC–FID against a control of the same concentra- tion which had not been exposed to microwave heat. Chromatograms these C16 FAMEs derived from extractable lipid potentially originated as C16 TAG. Additionally, Fig. 3 shows that all unsaturated C18 FAMEs were derived from extractable lipid. The sum of these unsaturated C18 FAMEs was 3.9 ± 0.35% (w/w) (Table 4) and total C18 TAG was 3.4 ± 0.04% (w/w) (Table 3), indicating that the majority of unsaturated C18 FAMEs originated as C18 TAG. Furthermore, under these conditions P. tricornutum produced 4.52 ± 0.07% (w/w) eicosapentaenoic acid (EPA) (Table 4), which is a valuable omega-3 fatty acid marketed for its suggested beneficial effects on human health (Alonso et al., 1996). Over half of the total EPA produced was derived from extractable lipid (Table 4), all of which originated as C20 TAG (Table 3). Of the 1.99 ± nd total biodiesel potential from direct in situ transesterification of P. tricornutum, C. vulgaris )a Tri-glyceride (%)a Sum of extracted (%)b Total biodiesel potential (%)c 27.39 ± 0.32 31.51 ± 0.33 51.19 ± 0.75 17.35 ± 0.95 21.56 ± 0.06 32.96 ± 0.93 3.46 ± 0.28 8.53 ± 0.18 14.45 ± 0.95of total biodiesel potential was likely derived from extractable lipid. The remaining difference between direct in situ transesterified FAMEs and extractable lipid-derived FAMEs is considered to be the residual fraction (e.g. membrane lipids). Further, the originating lipid class of the extractable lipid-derived FAMEs can be determined with relatively high confidence by comparing the carbon chain length of the molecule to the extractable lipid precursors quantified by GC–FID. This approach makes it possible to gain a comprehensive lipid assessment of a microalgae culture. It should be noted, that while themethods described here are capable of detecting a multitude of lipid compounds, e.g. cyclo- propane fatty acids or amide-linked fatty acids (sphingolipids), by mass-spectroscopy, such compounds were not detected in significant abundance in any of the cultures. If these compounds are present and quantifiable in samples, theywould be grouped into the “Other” category in the reported values. Additionally, two different transesterification techniques were used to derive total FAMEs versus extractable FAMEs. The direct in situ transesterification (Section 2.7) was used for total cellular fatty acid conversion to FAMEs. This method uses both a base (sodium methoxide) and an acid (boron tri-fluoride in methanol) to fully transesterify all cellular fatty acid compounds into FAMEs. The sequential addition of a basic and acidic catalyst has been shown to im- prove the efficiency of direct transesterification when water is present, as is the case when using live cultures (Griffiths et al., 2010). The basic catalyst, in the presence of water, promotes the saponification of glycerolipids via alkaline hydrolysis, thereby cleaving the ester bonds be- tween fatty acids and glycerol. The subsequent addition of acid catalyst increases the rate of esterification of liberated free fatty acids. These com- bined steps reduce the detrimental effects of water and result in faster and more efficient conversion of lipids to FAMEs (Griffiths et al., 2010). Extractable lipid compounds were transesterified using only acid (HCl) as a catalyst (Section 2.8). This was determined to be sufficient because there was no significant amount of water present in the extract containing the extractable lipid. Both techniques were compared using the precursor lipid extract and no yield differences were found. Since the acid only method is faster and uses less material, it was determined to be more suitable. 3.3.1. Phaeodactylum tricornutum P. tricornutum was cultured as previously described by Gardner et al. (2012) and harvested at peak TAG accumulation, as monitored by Nile Red fluorescence (Table 1). Lipids were extracted by the micro- wave extraction method described in Section 2.6.1. Total FAMEs were determined as described in Section 2.7, and FAMEs derived Table 2 Comparison of mean and standard deviation of lipid extracted with microwave energy, a and C. reinhardtii (n = 3). All values expressed as % (lipid/CDW). Organism Free fatty acid (%)a Mono-glyceride (%)a Di-glyceride (% P. tricornutum 1.4 ± 0.2 0.65 ± 0.02 2.08 ± 0.22 C. vulgaris 1.88 ± 0.54 0.06 ± 0.01 2.26 ± 0.36 C. reinhardtii 2.56 ± 0.15 0.07 ± 0.01 2.43 ± 0.03 a From microwave extraction of wet biomass. b Sum of neutral lipid precursors — free fatty acids, mono, di, tri-glycerides. c Total FAMEs from direct in situ transesterification.from extractable lipid were determined as described in Section 2.8. Total FAME was 51.2% (w/w) and the sum of all extractable lipids was 31.5% (w/w) with TAG contributing the majority at 27.4% (w/ w) or 87% of the extract (Table 2). Fig. 3 shows the complete FAME profile for P. tricornutum. Results from direct in situ transesterification indicate that under these growth conditions this organism preferen- tially synthesized C16:1 and C16:0 lipid compounds. Total C16:1 and C16:0 FAMEs were 24.8% (w/w) and 14.7% (w/w) respectively (Table 4). The fractions of total C16 FAMEs from extractable lipids were 54.5% for C16:1 and 51.8% for C16:0 (Table 4). The sum of all C16 FAMEs derived from extractable lipid was 21.1 ± 0.3% (w/w) (Table 4) and total C16 TAG was 21.4 ± 0.8% (w/w) (Table 3), indicating that0.02% (w/w) C14:0 FAMEs produced by this culture, 58% were derived from extractable lipid, of which 57% originated as C14 TAG and 43% originated as C14 MAG. 3.3.2. Chlorella vulgaris Cultures of C. vulgaris were sparged continuously with air amended with 5% CO2. Once peak TAG accumulation was reached, as monitored by Nile Red fluorescence (Table 1), the cultures were harvested for lipid extraction (Section 2.6.1), total FAMEs determina- tion (Section 2.7) and an estimate of FAMEs derived from extractable lipids (Section 2.8). Fig. 4 shows the complete FAME profile for C. vulgaris. Interestingly, this freshwater greenmicroalga preferentially synthesized unsaturated C18 fatty acids under the culture conditions employed here, compared to P. tricornutumwhich predominantly syn- thesized unsaturated C16 fatty acids. Total FAME content for C. vulgaris was 33% (w/w) and the sum of all extractable lipids was 21.6% (w/w) with TAG contributing the majority at 17.4% (w/w) or 80.6% of the total extract (Table 2). Twenty two percent (w/w) of total FAME was unsaturated C18 fatty acids, of which 66%were derived from extractable lipid (Table 4). The sum of these unsaturated C18 FAMEs derived from extractable lipids was 14.7 ± 0.2% (w/w) (Table 4) and total C18Fig. 3. Average and standard deviation of the FAME profile for P. tricornutum comparing direct in situ transesterification of all fatty acids (Direct) to FAMEs derived from only extractable lipid precursors (Extractable). The residual fraction is the difference between direct and extractable transesterification (n = 3). Table 3 Comparisons of mean and standard deviation of lipid class and carbon chain length percen (n = 3). All values expressed as % (lipid/CDW). Organism C12 C14 C1 Free fatty acid (%)a P. tricornutum N/D 0.25 ± 0.02 1 ± C. vulgaris 0.11 ± 0.03 0.23 ± 0.04 0.5 C. reinhardtii N/D N/D 1.1 Mono-glyceride (%)a P. tricornutum N/D 0.56 ± 0.04 0.0 C. vulgaris N/D N/D N/ C. reinhardtii N/D N/D 0.0 Di-glyceride (%)a P. tricornutum 0.39 ± 0.08 0.02 ± 0.01 1.0 C. vulgaris 0.24 ± 0.03 0.07 ± 0.01 0.4 C. reinhardtii 0.24 ± 0 0.38 ± 0.02 0.4 Tri-glyceride (%)a P. tricornutum N/D 0.73 ± 0.06 21 C. vulgaris N/D 0.1 ± 0.1 4.2 C. reinhardtii N/D 0.01 ± 0.01 2.0 N/D — not detected. a From microwave extraction of wet biomass. Table 4 Comparisons of mean and standard deviation of percent composition of FAMEs derived fro being reported as the residual fraction of total FAME for P. tricornutum, C. vulgaris and C. re Organism C14:0 C16:1 C16:0 C18:1–3b C18:0 Direct trans-esterification % lipid content (w/w) P. tricornutum 1.99 ± 0.02 24.81 ± 0.4 14.66 ± 0.59 3.62 ± 0.32 0.39 ± 0 C. vulgaris 0.04 ± 0 3.07 ± 0.07 6.88 ± 0.17 22.18 ± 0.72 0.67 ± 0 C. reinhardtii 0.04 ± 0 2.01 ± 0.1 3.35 ± 0.22 8.64 ± 0.53 0.33 ± 0 % Extractable lipid fraction (w/w)a P. tricornutum 1.16 ± 0.04 13.52 ± 0.22 7.59 ± 0.12 3.89 ± 0.52 0.25 ± 0 C. vulgaris N/D 2.57 ± 0.16 4.04 ± 1.08 14.72 ± 0.19 0.65 ± 0 C. reinhardtii 0.03 ± 0 1.99 ± 0.31 1.7 ± 0.14 4.28 ± 0.27 0.03 ± 0 % Residual lipid fraction (w/w) P. tricornutum 0.83 ± 0.06 11.29 ± 0.62 7.07 ± 0.7 −0.27 ± 0.2 0.14 ± 0 C. vulgaris 0.04 ± 0 0.49 ± 0.05 2.84 ± 0.42 7.46 ± 1.1 0.02 ± 0 C. reinhardtii 0.02 ± 0 0.02 ± 0.71 1.65 ± 0.83 4.36 ± 0.21 0.3 ± 0.0 N/D — not detected. a From microwave extraction of wet biomass, followed by transesterification of extracted b C18:1, C18:2, and C18:3 taken together. c Sum of other compounds detected. Fig. 4. Average and standard deviation of the FAME profile for C. vulgaris comparing direct in situ transesterification of all fatty acids (Direct) to FAMEs derived from only extractable lipid precursors (Extractable). The residual fraction is the difference between direct and extractable transesterification (n = 3).tages of extractable lipid constituents from P. tricornutum, C. vulgaris and C. reinhardtii 6 C18 C20 C22 0.19 0.13 ± 0 0.01 ± 0.01 N/D 7 ± 0.18 0.92 ± 0.28 0.05 ± 0.02 N/D 3 ± 0.05 1.43 ± 0.11 N/D N/D 4 ± 0.03 0.06 ± 0.01 N/D N/D D 0.06 ± 0.01 N/D N/D 2 ± 0.01 0.06 ± 0.03 N/D N/D 6 ± 0.04 0.61 ± 0.19 N/D N/D 8 ± 0.09 1.47 ± 0.24 N/D N/D 5 ± 0.01 1.36 ± 0.04 N/D N/D .38 ± 0.81 3.4 ± 0.04 1.87 ± 0.47 N/D 9 ± 0.5 12.81 ± 0.35 0.15 ± 0.07 N/D 5 ± 0.12 1.4 ± 0.17 N/D N/D m in situ transesterification, FAMEs derived from extractable lipid, and the differenceTAGwas 12.8 ± 0.35% (w/w) (Table 3). This indicates that themajority of unsaturated C18 FAMEs were derived from extractable lipid and originated as C18 TAG, with the remainder originating as C18 DAG (1.5% w/w) and C18 FFA (1% w/w) (Table 3). Furthermore, 10% (w/w) of total FAMEs for this culture of C. vulgaris were C16 fatty acid methyl ester, with 6.6% (w/w) derived from extractable lipid. Table 3 shows that 65% of the C16 FAMEs derived from extractable lipidwere originally C16 TAG with the remainder originating as C16 DAG and C16 FFA. 3.3.3. Chlamydomonas reinhardtii The model Chlorophyte, C. reinhardtii, was cultured as previously reported by Gardner et al. (2013) until peak TAG accumulation was achieved as monitored by Nile Red fluorescence (Table 1). The cul- tures were harvested and lipid was extracted using the microwave extraction method described in Section 2.6.1. Total FAMEs were determined as described in Section 2.7, and FAMEs derived from extractable lipid were determined as described in Section 2.8. Total FAME content was determined to be 14.5% (w/w) and the sum of all ex- tractable lipids was 8.5% (w/w) with TAG contributing the majority at 3.5% (w/w) or 41.2% of the extract and DAG accounting for 2.4% (w/w) or 28.2% of the extract (Table 2). C. reinhardtii is not known as an organism capable of producing a significant amount of TAG, and therefore is not typically regarded as a contender for industrial produc- tion of algal biofuel (Hu et al., 2008). However, it is widely studied and inhardtii (n = 3). All values expressed as % (lipid/CDW). C20:5 C20:1 C20:0 C22:6 C24:0 Otherc .01 4.52 ± 0.07 0.37 ± 0.01 0.12 ± 0 0.34 ± 0.01 0.25 ± 0 0.13 ± 0.02 .02 N/D 0.05 ± 0.03 0.04 ± 0 N/D 0.05 ± 0 0.03 ± 0.02 .03 N/D 0.07 ± 0.04 0.05 ± 0.03 N/D N/D 0.02 ± 0.0 .02 2.45 ± 0.12 0.23 ± 0.01 N/D 0.16 ± 0.02 N/D N/D .05 N/D N/D N/D N/D N/D N/D N/D 0.04 ± 0.02 0.03 ± 0.02 N/D N/D N/D .03 2.07 ± 0.19 0.14 ± 0.01 0.12 ± 0 0.18 ± 0.02 0.25 ± 0 0.13 ± 0.02 .02 N/D 0.05 ± 0.03 0.04 ± 0 N/D 0.05 ± 0 0.03 ± 0.02 3 N/D 0.04 ± 0.06 0.01 ± 0.01 N/D N/D 0.02 ± 0.0 lipid. one of the firstmicroalgae to have its genome fully sequenced. Thus, it is still a reasonable candidate for future research regarding lipid profile and metabolite analysis. Fig. 5 shows the complete FAME profile for C. reinhardtii. Similar to C. vulgaris, C. reinhardtii predominantly synthesized unsaturated C18 lipid compounds. Total unsaturated C18 FAMEs were 8.6% (w/w), of which ~50% were derived from extractable lipid. Unlike C. vulgaris, in this culture of C. reinhardtii, not all of the C18 FAMEs derived from extractable lipid came from C18 TAG. Instead, ~33% were originally C18 FFA, ~33% C18 DAG and ~33% C18 TAG (Table 3). All unsaturated C16 FAME compounds were derived from extractable lipid. However, the original source of these compounds is ambiguous. Total unsaturated C16 FAMEs derived from extractable lipid were 1.99 ± 0.31% (w/w), total C16 FFA were 1.0 ± 0.19% (w/w), total C16 DAG were 0.45 ± 0.01% (w/w) and total C16 TAG were 2.05 ± 0.12% (w/w). Unambiguously discerning which fraction of these extractable lipid precursors was responsible for the unsaturated C16 FAMEs is not possible with this method and will require additional method development. Fig. 5. Average and standard deviation of the FAME profile for C. reinhardtii comparing direct in situ transesterification of all fatty acids (Direct) to FAMEs derived from only extractable lipid precursors (Extractable). The residual fraction is the difference between direct and extractable transesterification (n = 3).4. Conclusion We describe a new rapid method for the efficient extraction of lipids from live cultures of microalgae for the purpose of quantifying the lipid content and characterizing the lipid carbon chain length and saturation. Additionally, we developed methods to determine the fraction of total biodiesel potential (as FAME) derived either from extractable lipid or from the residual fraction. Furthermore, the method can frequently determine from which lipid class (FFA, MAG, DAG, TAG) a particular FAME was derived. Since properties of biodiesel depend heavily on the length and degree of saturation of the FAME, the methods described here can assist in ‘fine-tuning’ the final product, be it biodiesel or a higher value product, through the choice of certain algae or growth conditions. Specifically, here we demonstrate the usefulness of this method by demonstrating differences in lipid profiles for three frequently studied microalgae. It is demonstrated that the marine diatom P. tricornutum preferentially produced unsaturated C16 lipid com- pounds, whereas the two freshwater green microalgae, C. vulgaris and C. reinhardtii, predominantly synthesized unsaturated C18 com- pounds. From an industrial standpoint, it will be beneficial to choose the appropriate algae and culture conditions to produce the largest possible amount of biodiesel with the appropriate FAME profile orto screen for algal products with higher value than biodiesel. For in- stance P. tricornutum was shown to produce an omega-3 fatty acid, specifically eicosapentaenoic acid (C20:5), which has been suggested to benefit human health. This method has determined that these compounds were primarily derived from triacylglyceride storage molecules. In order to apply the methods described here to other algae, plants or different culturing conditions, verification of extraction and transesterification efficacy will have to be conducted. Such checks are fairly quick and are described in detail for the algae and culturing conditions employed here. Financial disclosure This material is based upon work supported by the National Science Foundation (NSF) under CHE-1230632. A portion of this re- search was supported by the U.S. Department of Energy (DOE) Office of Energy Efficiency and Renewable Energy (EERE) Biomass Program under Contract Nos. DE-EE0003136 and DE-EE0005993. Support for EL and RDG was also provided by the NSF IGERT Program in Geobiological Systems (DGE 0654336). Instrumental support was provided through the Center for Biofilm Engineering (CBE) at Montana State University (MSU) as well as by the Environmental and Biofilm Mass Spectrometry Facility (EBMSF) funded through DURIP Contract Number: W911NF0510255 and the MSU Thermal Biology Institute from the NASA Exobiology Program Project NAG5-8807. References Alonso, D.L., Segura, C.I., Grima, E.M., 1996. First insights into improvement of eicosapentanoic acid content in Phaeodactylum tricornutum (Bacillariophyceae) by induced mutagenesis. J. Phycol. 32, 339–345. Bigelow, N.W., Hardin, W.R., Barker, J.P., Ryken, S. a, Macrae, A.C., Cattolico, R.A., 2011. A comprehensive GC–MS sub-microscale assay for fatty acids and its applications. J. Am. Oil Chem. Soc. 88, 1329–1338. Bligh, E.G., Dyer, W.J., 1959. A rapid method of total lipid extraction and purification. Can. J. Biochem. Physiol. 37, 911–917. Brown, A.P., Slabas, A.R., Rafferty, J.B., 2009. Fatty acid biosynthesis in plants — metabolic pathways, structure and organization. In: Govindjee, Wada, H., Murata, N. (Eds.), Lipids in Photosynthesis: Essential and Regulatory Functions. Advances in Photosyn- thesis and Respiration, pp. 11–34. Chen, W., Zhang, C., Song, L., Sommerfeld, M., Hu, Q., 2009. A high throughput Nile red method for quantitative measurement of neutral lipids in microalgae. J. Microbiol. Methods 77, 41–47. Chisti, Y., 2007. Biodiesel from microalgae. Biotechnol. Adv. 25, 294–306. Cooksey, K.E., Guckert, J.B., Williams, S.A., Callis, P.R., 1987. Fluorometric determination of the neutral lipid content of microalgal cells using Nile Red. J. Microbiol. Methods 6, 333–345. Da Silva, T.L., Reis, A., Medeiros, R., Oliveira, A.C., Gouveia, L., 2009. Oil production towards biofuel from autotrophic microalgae semicontinuous cultivations monitorized by flow cytometry. Appl. Biochem. Biotechnol. 159, 568–578. Demirbas, A., 2007. Importance of biodiesel as transportation fuel. Energy Policy 35, 4661–4670. Dismukes, G.C., Carrieri, D., Bennette, N., Ananyev, G.M., Posewitz, M.C., 2008. Aquatic phototrophs: efficient alternatives to land-based crops for biofuels. Curr. Opin. Biotechnol. 19, 235–240. EIA, 2012. How Dependent Are We on Foreign Oil? [WWW Document]. U.S. Energy Information Administration. (URL http://www.eia.gov/energy_in_brief/foreign_oil_ dependence.cfm). Elsey, D., Jameson, D., Raleigh, B., Cooney, M.J., 2007. Fluorescent measurement of microalgal neutral lipids. J. Microbiol. Methods 68, 639–642. Eustance, E., Gardner, R.D., Moll, K., Menicucci, J., Gerlach, R., Peyton, B.M., 2013. Growth, nitrogen utilization and biodiesel potential for two chlorophytes grown on ammonium or nitrate. J. Appl. Phycol. http://dx.doi.org/10.1007/s10811-013- 0008-5 (in press). Ferrell, J., Sarisky-Reed, V., 2010. National Algal Biofuels Technology Roadmap. U.S. Department of Energy's Office of Fuels Development. Gardner, R., Peters, P., Peyton, B., Cooksey, K.E., 2011. Medium pH and nitrate concentra- tion effects on accumulation of triacylglycerol in two members of the chlorophyta. J. Appl. Phycol. 23, 1005–1016. Gardner, R.D., Cooksey, K.E., Mus, F., Macur, R., Moll, K., Eustance, E., Carlson, R.P., Gerlach, R., Fields, M.W., Peyton, B.M., 2012. Use of sodium bicarbonate to stimu- late triacylglycerol accumulation in the chlorophyte Scenedesmus sp. and the dia- tom Phaeodactylum tricornutum. J. Appl. Phycol. 1311–1320. Gardner, R.D., Lohman, E., Gerlach, R., Cooksey, K.E., Peyton, B.M., 2013. Comparison of CO(2) and bicarbonate as inorganic carbon sources for triacylglycerol and starch accumulation in Chlamydomonas reinhardtii. Biotechnol. Bioeng. 110, 87–96. Gouveia, L., Oliveira, A.C., 2009. Microalgae as a raw material for biofuels production. J. Ind. Microbiol. Biotechnol. 36, 269–274. Greenwell, H.C., Laurens, L.M.L., Shields, R.J., Lovitt, R.W., Flynn, K.J., 2010. Placing microalgae on the biofuels priority list: a review of the technological challenges. J. R. Soc. Interface 7, 703–726. Griffiths, M.J., Van Hille, R.P., Harrison, S.T.L., 2010. Selection of direct transesterification as the preferred method for assay of fatty acid content of microalgae. Lipids 45, 1053–1060. Groom, M.J., Gray, E.M., Townsend, P. a, 2008. Biofuels and biodiversity: principles for creating better policies for biofuel production. Conserv. Biol. 22, 602–609. Guckert, J.B., Cooksey, K.E., 1990. Triglyceride accumulation and fatty acid profile changes in Chlorella (Chlorophyta) during high pH-induced cell inhibition. J. Phycol. 26, 72–79. Guckert, J.B., Cooksey, K.E., Jackson, L., 1988. Lipid solvent systems are not equivalent for analysis of lipid classes in the microeukaryotic green-alga, Chlorella. J. Microbiol. Methods 8, 139–149. Harris, E.H., 1989. The Chlamydomonas Sourcebook: A Comprehensive Guide to Biology and Laboratory Use. Academic Press, San Diego. Hill, J., Nelson, E., Tilman, D., Polasky, S., Tiffany, D., 2006. Environmental, economic, and energetic costs and benefits of biodiesel and ethanol biofuels. Proc. Natl. Acad. Sci. U. S. A. 103, 11206–11210. Hu, Q., Sommerfeld, M., Jarvis, E., Ghirardi, M., Posewitz, M., Seibert, M., Darzins, A., 2008. Microalgal triacylglycerols as feedstocks for biofuel production: perspectives and advances. Plant J. 54, 621–639. Hunter-Cevera, J.C., Boussiba, S., Cuello, J.L., Duke, C.S., Efroymson, R.A., Golden, S.S., Holmgren, J., Johnson, D.L., Jones, M.E., Smith, V.H., Steger, C., Stephanapoulos, G.N., Walker, L.P., Williams, E., Zimba, P.V., 2012. Sustainable development of algal biofuels committee on the sustainable development of algal biofuels. Board on Agriculture and Natural Resources. Board on Energy and Environmental Systems. Iverson, S.J., Lang, S.L., Cooper, M.H., 2001. Comparison of the Bligh and Dyer and Folch methods for total lipid determination in a broad range of marine tissue. Lipids 36, 1283–1287. alkyl esters. Fuel Process. Technol. 86, 1059–1070. Laurens, L., Quinn, M., Van Wychen, S., Templeton, D.W., Wolfrum, E.J., 2012a. Accurate and reliable quantification of total microalgal fuel potential as fatty acid methyl esters by in situ transesterification. Anal. Bioanal. Chem. 403, 167–178. Laurens, L., Dempster, T. a, Jones, H.D.T., Wolfrum, E.J., VanWychen, S., McAllister, J.S.P., Lee, J.-Y., Yoo, C., Jun, S.-Y., Ahn, C.-Y., Oh, H.-M., 2010. Comparison of several methods for effective lipid extraction frommicroalgae. Bioresour. Technol. 101, S75–S77 (Suppl.). Liu, J., Huang, J., Sun, Z., Zhong, Y., Jiang, Y., Chen, F., 2011. Differential lipid and fatty acid profiles of photoautotrophic and heterotrophic Chlorella zofingiensis: assessment of algal oils for biodiesel production. Bioresour. Technol. 102, 106–110. Martino, A. De, Meichenin, A., Shi, J., Pan, K., Bowler, C., 2007. Genetic and phenotypic characterization of Phaeodactylum tricornutum (Bacillariophyceae) accessions 1. J. Phycol. 43, 992–1009. Merchant, S.S., Prochnik, S.E., Vallon, O., Harris, E.H., Karpowicz, S.J., et al., 2007. The Chlamydomonas genome reveals the evolution of key animal and plant functions. Science (New York, N.Y.) 318, 245–250. Nichols, H., Bold, H., 1965. Trichosarcina polymorpha gen. et sp. nov. J. Phycol. 1, 34–38. Patil, P.D., Reddy, H., Muppaneni, T., Mannarswamy, A., Schuab, T., Holguin, F.O., Lammers, P., Nirmalakhandan, N., Cooke, P., Deng, S.G., 2012. Power dissipation in microwave-enhanced in situ transesterification of algal biomass to biodiesel. Green Chem. 14, 809–818. Prabakaran, P., Ravindran, a D., 2011. A comparative study on effective cell disruption methods for lipid extraction from microalgae. Lett. Appl. Microbiol. 53, 150–154. Provasoli, L., McLaughlin, J., Droop, M., 1957. The development of artificial media for marine algae. Arch. Microbiol. 25. Richter, S.J., Richter, C., 2002. A method for determining equivalence in industrial appli- cations. Qual. Eng. 14, 375–380. Sheehan, J., 1998. A Look Back at the U.S. Department of Energy's Aquatic Species Program: Biodiesel from Algae. U.S. Department of Energy's Office of Fuels Development. Singh, Y., Kumar, H.D., 1992. Lipid and hydrocarbon production by Botryococcus spp. under nitrogen limitation and anaerobiosis. 8, 121–124. Spiekermann, P., Lerchl, J., Beckmann, C., Kilian, O., Kroth, P.G., Boland, W., Za, U., Heinz, E., 2003. New insight into Phaeodactylum tricornutum fatty acid metabolism. Cloning and Functional Characterization of Plastidial and Microsomal Δ12-Fatty Acid Desaturases 1 [w], 131, pp. 1648–1660. Stephenson, A.L., Dennis, J.S., Howe, C.J., Scott, S.A., Smith, A.G., 2010. Influence of nitrogen-limitation regime on the production by Chlorella vulgaris of lipids for bio- diesel feedstocks. 1, 47–58. Valenzuela, J., Mazurie, A., Carlson, R.P., Gerlach, R., Cooksey, K.E., Peyton, B.M., Fields, M.W., 2012. Potential role of multiple carbon fixation pathways during lipid accumulation in Phaeodactylum tricornutum. Biotechnol. Biofuels 5, 40. Xue, J., Grift, T.E., Hansen, A.C., 2011. Effect of biodiesel on engine performances and Rencenberger, M., Parchert, K.J., Gloe, L.M., 2012b. Algal biomass constituent analysis: method uncertainties and investigation of the underlying measuring chemistries. Anal. Chem. 84, 1879–1887. Lee, S.J., Yoon, B., Oh, H., 1998. Rapid method for the determination of lipid from the green alga Botryococcus braunii. 12, 553–556.emissions. Renewable Sustainable Energy Rev. 15, 1098–1116. Zheng, H., Yin, J., Gao, Z., Huang, H., Ji, X., Dou, C., 2011. Disruption of Chlorella vulgaris cells for the release of biodiesel-producing lipids: a comparison of grinding, ultrasonication, bead milling, enzymatic lysis, and microwaves. Appl. Biochem. Biotechnol. 164, 1215–1224.Teixeira, R.E., 2012. Energy-efficient extraction of fuel and chemical feedstocks from algae. Green Chem. 14, 419.Kiritsakis, A.K., Lenart, E.B., Willet, W.C., Hernandez, R.J., 1998. Olive Oil: from the Tree to the Table, 2nd ed. Food & Nutrition Press, Trumbull, CT. Knothe, G., 2005. Dependence of biodiesel fuel properties on the structure of fatty acid