Screening, Down-Selection, Characterization, and Genetic Tool

SCREENING, DOWN-SELECTION, CHARACTERIZATION,

AND GENETIC TOOL DEVELOPMENT IN

HIGH-PRODUCTIVITY MICROALGAE

by Lukas Royce Dahlin A thesis submitted to the Faculty and the Board of Trustees of the Colorado School of

Mines in partial fulfillment of the requirements for the degree of Doctor of Philosophy (Applied

Chemistry).

Golden, Colorado

Date:

Signed: Lukas R. Dahlin

Signed: Dr. Matthew C. Posewitz Thesis Advisor

Signed: Dr. Michael T. Guarnieri Thesis Advisor

Golden, Colorado

Date:

Signed: Dr. Thomas Gennett Professor and Head Department of Chemistry

ABSTRACT

Microalgae are a promising source of renewable biomass, and biocatalysts which can be utilized for the renewable production of fuel and chemical intermediates. During the time the work of this thesis was pursued microalgal productivity was recognized as a limiting factor to achieve competitive pricing with fossil fuel alternatives. To address this, a microalgal culture collection was screened, down-selecting for promising strains showing robust growth under relevant outdoor conditions.

After isolating several promising strains, they were further characterized, ultimately leading to the identification of strains for robust winter and summer cultivation. Winter strains were grown outdoors in 1000 L raceway ponds in February of 2016 to confirm and characterize real world growth metrics. Two promising summer isolates were identified, Picochlorum renovo and Scenedesmus sp. These two strains both display high productivity, however, have remarkably different life cycles. Picochlorum renovo has a compact genome, grows rapidly, undergoes cell division during both day and night, and once nitrogen deprived photosynthesis ceases. In contrast, Scenedesmus sp. has a relatively large genome, regulates cell division to occur during dark periods, and has a remarkably high photosynthetic rate when nitrogen deprived.

To further enhance growth and enable basic science inquiries genetic tools were developed for Picochlorum renovo. Tools necessary for the overexpression of proteins from either the nuclear or chloroplast genomes were developed.

iii

TABLE OF CONTENTS

ABSTRACT ...... iii

LIST OF FIGURES ...... vii

LIST OF TABLES ...... xi

ACKNOWLEDGEMENTS ...... xii

CHAPTER 1 INTRODUCTION ...... 1 1.1 Motivation ...... 1 1.2 Background and Previous Work ...... 3 1.3 Outline of Chapters ...... 6 1.4 References ...... 7

CHAPTER 2 DOWN-SELECTION AND OUTDOOR EVALUATION OF NOVEL HALOTOLERANT ALGAL STRAINS FOR WINTER CULTIVATION ...... 11 2.1 Abstract ...... 11

2.2 Introduction ...... 12 2.3 Results ...... 14

2.3.1 Indoor Screening ...... 14 2.3.2 Outdoor Deployment ...... 15 2.4 Discussion ...... 21

2.4.1 Indoor Screening ...... 21 2.4.2 Outdoor Deployment ...... 23 2.5 Methods...... 26

2.5.1 Culture Collection Screening ...... 26 2.5.2 18S Sequencing ...... 28

2.5.3 Outdoor Growth ...... 28

2.5.4 Compositional Analysis ...... 29 2.6 Acknowledgements ...... 30 2.7 References ...... 31

CHAPTER 3 RECENT ADVANCES IN ALGAL GENETIC TOOL DEVELOPMENT ...... 35 3.1 Abstract ...... 35 3.2 Introduction ...... 36 3.3 Random Gene Integration and Homologous Recombination ...... 37 3.4 Multi-copy Gene Expression and Expression Control...... 39 3.5 In vivo genome editing ...... 41 3.6 Technical Hurdles and Considerations ...... 43 3.7 Concluding Remarks ...... 44 3.8 References ...... 46

CHAPTER 4 DEVELOPMENT OF A HIGH-PRODUCTIVITY, HALOPHILIC, THERMOTOLERANT MICROALGA, PICOCHLORUM RENOVO ...... 50 4.1 Abstract ...... 50 4.2 Introduction ...... 51 4.3 Results ...... 52

4.3.1 Down-Selection, Physiology, and Compositional Analysis ...... 52

4.3.2 Genomic Analysis and Speciation ...... 55

4.3.3 Transcriptome Response to Salinity...... 59

4.3.4 Nuclear and Chloroplast Engineering ...... 60 4.4 Discussion ...... 63 4.5 Conclusions ...... 67 4.6 Methods...... 68

4.6.1 Strain Screening and Characterization of Algal Growth...... 68

4.6.2 Compositional Analysis ...... 70

4.6.3 Genome Sequencing, Assembly and Annotation ...... 70

4.6.4 Transcriptome Response to Salinity...... 71

4.6.5 Nuclear Engineering...... 73

4.6.6 Chloroplast Engineering...... 74

4.6.7 Fluorescent Plate Reader Analysis ...... 75

4.6.8 Statistics and Reproducibility ...... 76

4.6.9 Microscopy ...... 76 4.7 Acknowledgements ...... 77

4.8 Data Availability ...... 77 4.9 References ...... 78

CHAPTER 5 CHARACTERIZATION OF A NOVEL SCENEDESMUS...... 83 5.1 Abstract ...... 83 5.2 Introduction ...... 83 5.3 Results ...... 84

5.3.1 Strain Down-Selection, Physiology, and Compositional Analysis ...... 84

5.3.2 Preliminary Genetic Engineering ...... 86

5.3.3 Effect of Prolonged Darkness on Biomass...... 91 5.4 Discussion ...... 92 5.5 Methods...... 93 5.6 References ...... 94

CHAPTER 6 CONCLUSIONS AND FUTURE DIRECTIONS ...... 96 6.1 Conclusions and Future Directions ...... 96

APPENDIX A FURTHER WINTER STRAIN CHARACTERIZATION ...... 100 A.1 Temperature, Salinity, and N:P Optima ...... 100

APPENDIX B FURTHER P. RENOVO DEVELOPMENT ...... 103 B.1 Nitrogen: Phosphorus Ratios ...... 103

B.2 Additional Genetic Engineering Notes ...... 105

APPENDIX C COPYRIGHT LICENSES...... 108

LIST OF FIGURES

Figure 2.1 (Left) Ash-Free Dry Weight accumulation and (right) areal productivities for strains cultivated outdoors during the winter month of February 2016 (1st – 19th). Maximum productivity is calculated from the steepest portion of growth curve, while harvest-yield productivity measures all biomass accumulated up to day 18. Average and standard deviation error bars are from 3 biological replicates...... 16

Figure 2.2 FAME and carbohydrate analysis of strains grown outdoors during the winter month of February 2016 (1st – 19th). Average and standard deviation error bars are from 3 biological replicates...... 17

Figure 2.3 Nitrogen (as ammonium) depletion of strains grown outdoors during the winter month of February 2016 (1st – 19th) (Top Left). Phosphate depletion of strains grown outdoors under winter conditions (Bottom Left). Representative change in temperature of ponds throughout the experiment (Top Right). Representative change in PAR of ponds throughout the experiment (Bottom Right). Average and standard deviation error bars are from 3 biological replicates...... 18

Figure 2.4 Carbohydrate and FAME profile of strains grown outdoors during the winter month of February 2016 (1st – 19th), x axis lists the strain and time in days. Average and standard deviation error bars are from 3 biological replicates...... 19

Figure 3.1 Schematic overview of microalgal genetic tools developed to date, and target organelles. Random integration and homologous recombination have been demonstrated in the chloroplast (green), mitochondria (orange), and nucleus (blue). Episomal plasmids and genome editing tools, including ZFNs, TALE(N)s, and CRISPR/Cas9, have been applied for nuclear targeting. Transit peptides further allow for indirect engineering of organelles outside the nucleus, including mitochondria, chloroplast, and/or endoplasmic reticulum (red) ...... 37

Figure 4.1 Representative culture collection growth screening data. The rapid growth and high optical density phenotype of P. renovo is highlighted in black. Nannochloropsis salina CCMP 1776 is bolded in red for reference...... 53

vii

P. renovo growth rate as a function of temperature. Average and standard deviation are from n=3 biological replicates. (c) Growth curve comparison of P. renovo at 8.75 and 35 g/L salinity. Average and standard deviation are from n=2 biological replicates. (d) Endpoint biomass titer following 6 days of growth at varying salinities for P. renovo. Average and standard deviation are from n=2 replicates are reported...... 54

Figure 4.3 Overview of P. renovo productivity and associated biomass composition, as a function of time. Alternating black and yellow bars depict the light-dark growth cycle. (a) Growth curves as a function of ash-free dry weight (g per L) and cell density (cells per mL). Areal productivity is shown for hour 6 to 30, representing one light-dark cycle (day). (b) The biomass content of lipid (as FAMEs), carbohydrate, protein, and the fraction of biomass not identified. (c) Carbohydrate speciation via hydrolysis of biomass. (d) Fatty acid speciation via fatty acid methyl ester analysis, representative of the lipid fraction of the biomass. All data points are an average of n=3 biological replicates; error bars depict the standard deviation of the replicates...... 55

Figure 4.4 Division of P. renovo highlighting autosporulation and mother cell wall (arrow), a defining trait of the Picochlorum genus. Red coloring indicates chlorophyll autofluorescence...... 57

Figure 4.5 Chloroplast and mitochondria genome maps of P. renovo...... 58

Figure 4.6 Gene ontology analysis, visualized with REVIGO. Top panels represent gene ontology terms upregulated at high salinity, with both color and circle size indicating log fold change, as indicated in the color scale. Bottom panels represent gene ontology terms downregulated at high salinity, with both color and circle size indicating log fold change, as indicated in the color scale...... 59

Figure 4.7 Overview of P. renovo nuclear transformation. (a) Construct design showing genetic elements and primers used to generate DNA for electroporation (49 and 11) and subsequent PCR confirmation of transformants. (b) PCR verification of 12 clones utilizing primers shown in panel A. (c) Dot plot of fluorescent plate reader data of wild type and mCherry transformants, normalized to chlorophyll autofluorescence. Data is from 3 biological replicates. (d) Confocal microscopy images of wild type and transformant microalgae expressing mCherry. Green coloring represents chlorophyll autofluorescence, red coloring represents mCherry fluorescence, 10 µm scale bar...... 61

viii

Figure 4.8 Overview of P. renovo chloroplast transformation. (a) Construct design showing genetic elements utilized and homology directed integration into the chloroplast genome, along with the primers used for subsequent PCR confirmation of transformants. (b) PCR verification of 3 clones utilizing primers shown in panel A. (c) Dot plot of fluorescent plate reader data of wild type and sfGFP transformants, normalized to chlorophyll autofluorescence. Data is from 3 biological replicates. (d) Epifluorescent microscopy images of wild type and sfGFP transformant microalgae. Red coloring represents chlorophyll autofluorescence, green coloring represents sfGFP fluorescence, 10 µm scale bar...... 62

Figure 5.1 Representative culture collection growth screening data. The high biomass accumulation capacity phenotype of Scenedesmus sp.(46B- D3) following 6 days of growth is highlighted in black. Nannochloropsis salina CCMP 1776 is bolded in red for reference...... 86

Figure 5.2 Detailed time course growth analysis of Scenedesmus sp. (46B-D3). Alternating black and yellow bar represents the lighting cycle. Top: Growth curves in terms of ash free dry weight (AFDW) and cells per mL. Middle: Growth curves in terms of lipids (as fatty-acid methyl esters, g/L) and in terms lipids (% AFDW). Bottom: Growth curves in term of carbohydrates (g/L) and carbohydrates (% AFDW). Data points represent the average and standard deviation of n=2 biological replicates...... 87

Figure 5.3 Temperature optima of Scenedesmus sp. (46B-D3), data points represent the average and standard deviation of n=3 biological replicates...... 88

Figure 5.4 Salinity tolerance screening of Scenedesmus sp. (46B-D3), data points represent the average and standard deviation of n=2 biological replicates, following 6 days of growth...... 88

Figure 5.5 Representative growth curves of Scenedesmus sp. (46B-D3) with and without addition of ammonium chloride to a final concentration of 5mM at hour 100, following entry into stationary growth phase...... 89

Figure 5.6 Nitrogen: Phosphorus ratio screening of Scenedesmus sp. (46B-D3), data points represent the average and standard deviation of n=2 biological replicates, following 6 days of growth...... 89

Figure 5.7 Lipid analysis via fatty acid methyl esters as a function of time, all data points are an average of n=2 biological replicates; error bars depict the standard deviation of the replicates ...... 90

Figure 5.8 Carbohydrate analysis via acid hydrolysis of the biomass, as a function of time, all data points are an average of n=2 biological replicates; error bars depict the standard deviation of the replicates. .... 90

Figure 5.9 FAME and carbohydrate analysis during prolonged darkness, as a function of time, data points are an average of n=2 biological replicates; error bars depict the standard deviation of the replicates. .... 91

Figure A.1 Optimal growth temperature for 14-F2 and 4-C12. Data point represent the average and standard deviation from n=3 biological replicates...... 100

Figure A.2 Optimal salinity data for 14-F2 and 4-C12, salinity in g/L is listed on the x axis, with strain ID. Data point represent the average and standard deviation from n=2 biological replicates, after growth in winter conditions for 14 days ...... 101

Figure A.3 Optimal nitrogen to phosphorus ratio data for 14-F2, N:P ratio is listed on the x axis. Data point represent the average and standard deviation from n=2 biological replicates, after growth in winter conditions for 13 days...... 102

Figure B.1 Growth analysis of P. renovo at varying nitrogen to phosphorus ratios. Top: endpoint biomass titer after 6 days of growth. Bottom: optical density growth curves...... 104

LIST OF TABLES

Table 2.1 Down-selection criteria and data for strains selected for outdoor cultivation...... 14

Table 4.1 Whole genome alignment analysis of Picochlorum spp...... 57

ACKNOWLEDGEMENTS

I would like to thank my advisors Dr. Michael Guarnieri and Dr. Matthew

Posewitz for accepting me into their respective groups and offering invaluable guidance throughout graduate school. I would also like to thank my thesis committee, Dr. Kim

Williams, Dr. Brian Trewyn, and Dr. Nanette Boyle for their support. Dr. Brian Vogler,

Dr. Calvin Henard, and Dr. Jeffrey Linger deserve thanks for insightful discussion regarding genetic engineering in eukaryotic and prokaryotic systems. Thanks to Nick

Sweeney at NREL for help in the design and upkeep of various bioreactors utilized in my work. Thank you to Stefanie Van Wychen, Dr. Henri Gerken, Dr. John McGowen, Dr.

Alida Gerritsen, Yuliya Kunde, Dr. Blake Hovde, and Dr. Shawn Starkenburg for their work on the resultant publications from this thesis. This work could not have been accomplished without the continued financial support from the United States Department of Energy, thank you to Daniel Fishman for his recognition of the continued importance of this work.

I owe a great deal of gratitude for those that encouraged me to go to graduate school and pursue a PhD, especially Dr. Steven Sternberg from the University of

Minnesota Duluth, and Dr. Sigmund Degitz, Dr. Michael Hornung, Joe Korte, Jon

Haselman, and Dr. Tony Schroeder from the US Environmental Protection Agency, Mid-

Continent Ecology Division. Special thanks to my family and friends for their unwavering support.

xii

CHAPTER 1

INTRODUCTION

1.1 Motivation

The U.S. Energy Information Administration predicts a high demand (~35 quads) for liquid fuels to satisfy U.S. energy needs in the future 1. Historically these liquid fuels have been sourced from fossil fuel resources, which are declining due to human consumption. Concurrent with this consumption, atmospheric CO2 concentrations are increasing, leading to rising global temperatures due to the greenhouse effect. As such, human society is becoming increasingly reliant on a finite resource, while causing detrimental environmental and economic effects due to this reliance. Numerous problems are expected due to a changing climate, which in part include ocean acidification, reduced crop yields, freshwater scarcity, species extinction, coastal flooding, and increased heatwaves and wildfires 2. These problems are further exacerbated by unsustainable farming practices that have led to topsoil erosion and degradation of agricultural lands 3–5. These farming practices have ultimately led to fertilizer runoff causing blooms of toxic algae and dead zones in the ocean 6,7. Combined, these problems will have distinct detrimental economic costs on human civilization. As such, a solution is needed to address this multitude of problems. Microalgae are a potential option to ameliorate these problems, as they can be grown as a food or liquid fuel source and can be utilized for wastewater treatment.

Microalgae are amongst the most efficient organisms on Earth at converting solar energy,

8 CO2, and water into useful biomass and oxygen . This is in part due to their small size allowing rapid diffusion of growth substrates, and simple structure, for example lacking the roots and recalcitrant biopolymers such as lignin in higher plants 9. These organisms can be sustainably

1 farmed at large scales and the subsequent biomass can be used to produce food, fuels, and numerous bioproducts through fermentative processes. As such, microalgae can address the above-mentioned problems by offsetting petroleum-based products and producing a food source while concurrently pulling CO2 out of the atmosphere, reducing the greenhouse effect. Indeed microalgae are estimated to perform half of global photosynthesis, providing food for 70% of the world’s biomass 10. Microalgae have demonstrated a 10-fold increase in areal biomass yields over traditional crops such as corn grain and switch grass 9. When extrapolated to biodiesel yields, and accounting for differences in biomass composition, it has been concluded that microalgae (at 30% oil by dry weight) are capable of a 10-fold increase in biodiesel yields when compared to oil palm 11. Further, microalgae are capable of utilizing saline water for growth, as such their production does not compete with terrestrial agriculture and does not need to utilize dwindling freshwater resources. Indeed microalgal systems use far less water than traditional oilseed crops, with abundant land available for cultivation 8. Despite the promise of microalgae to ameliorate multiple problems, widespread cultivation has yet to be realized. This is in part due to low productivities compared to theoretical and lab scale yields, previous investments in freshwater strains, and a lack of genetic engineering tools available to enhance growth or direct production of valuable products. This thesis was pursued with the goal of advancing the biofuel potential of microalgae, however lessons learned will likely be translatable to other uses of microalgae, such as a food source. Regarding biofuel production, technoeconomic analysis has shown a need to achieve 25 g/m2/day biomass productivity to achieve cost competitive biofuels

12. Current state of technology metrics have demonstrated year-round biomass productivities of

~12 g/m2/day. Thus, to realize the potential of microalgae, a doubling in biomass productivity is necessary 8,13.

1.2 Background and Previous Work

Starting in the 1940’s, significant research on the use of microalgae as an agricultural crop has been performed 14. This preliminary work focused on microalgae as a food source.

Starting in 1978, concurrent with a decline in U.S. oil production, research into the use of microalgae as a source of biofuels was pursued through the Department of Energy funded

Aquatic Species Program (ASP) 8. This research began with an extensive bioprocessing effort, identifying several promising isolates. Work then progressed to understanding lipid production in microalgae, with the observation that nutrient stress induces lipid production through a cessation in cell division. However, despite induction of lipids under stressful conditions researchers concluded that the concurrent decrease in cell division led to an overall decrease in lipid productivity when cells were stressed. In order to address this disparity researchers invested heavily in the genetic engineering of microalgae, with the ultimate hope of engineering algae to concurrently maintain high productivity, under nutrient replete conditions, with concurrent lipid accumulation. As such, they developed tools for manipulation of two diatom species, as much research was done on silica limitation in diatoms to induce lipid production, and the enzyme

ACCase was identified as a putative rate limiting step in lipid synthesis. Ultimately researchers were able to overexpress ACCase in two diatom species, but no increase in lipids were observed.

Of particular note in terms of genetic enhancement was the assertion that reduction of the photosynthetic antenna would solve multiple problems by reducing wasted photons and reducing photooxidative damage, ultimately increasing overall culture productivity by efficient photon usage.

Despite significant advances in the understanding of utilizing algae for biofuels, the aquatic species program funding ended in 1996. Ultimately ASP researchers concluded there

3 were no fundamental engineering or economic issues (e.g. net energy input vs. output) that limit the feasibility of microalgae as a source of biomass for biofuel production. One key conclusion of the ASP was the need for the use of open raceway ponds for biofuels to be cost competitive, as such significant work focused on the use of these ponds for outdoor cultivation. These open raceway ponds have operated on the scale of a few meters squared, up to multiple (~20) hectares in size. Ponds are run at depths of approximately 5-20 centimeters, and paddle wheels provide mixing. If CO2 is available and cost effective it can be injected into ponds via microbubbles, to increase productivity while concurrently controlling pond pH 15. Harvesting of algal biomass occurs via gravity settling, followed by hollow fiber membranes, and finally centrifugation to concentrate biomass from ~0.5 g/L to 200 g/L. These relatively simple raceway ponds are widely regarded as the only economically viable way to produce algal biomass for conversion to low value products such as biofuels 16.

To process algal biomass into biofuels and bioproducts researchers have devised two methodologies for conversion. First, hydrothermal liquefaction uses high temperatures and pressures, combined with natural gas and hydrogen addition to catalytically convert wet algal biomass to biofuels (gasoline, diesel, jet). This process has the distinct advantage that the whole algal cell (proteins, carbohydrates, lipids) can be converted to fuels, as such a high lipid state is not necessary. Further, the waste streams generated are rich in nutrients and can be fed back into algal ponds to recover nutrients and reduce operating expenditures. However, significant hurdles remain, particularly concerning the quality of the biocrude oil produced, which is high in nitrogen and oxygen, and has a high viscosity 17–19. Another hurdle with hydrothermal liquefaction is the limited products obtained, biofuels are primarily produced, with little leeway available for the production of other bio-based products. An alternative approach for algal

4 processing developed by Dong et al. is termed combined algal processing. This approach utilizes an acid pretreatment of wet algal biomass to release monomeric sugars from the biomass into the aqueous phase. After neutralization this sugar rich broth can then be utilized via established fermentative technologies to produce an array of high value bioproducts (e.g. succinic acid).

After fermentation and extraction of bioproducts, the residual biomass and broth is sent to an extraction process wherein hexane is used to extract lipids. These lipids can then be converted to biodiesel using established biodiesel manufacturing technology. The high value of the fermentation produced product(s) can be utilized to offset the low value of biofuels, making this a potentially more economic approach to biomass processing. However, for high yields of bioproducts and biofuels the starting biomass must be high in fermentable sugars (carbohydrates) and lipids, approximately 40% of each 20.

Concurrent with the work of the ASP, other researchers pursued the alga

Chlamydomonas reinhardtii, ultimately developing it as a model algal species 21–23. As such, much of the work regarding microalgae biology in the last ~35 years has been done in

Chlamydomonas reinhardtii, due to its genetic tractability, availability of mutants, and robust research community built around this organism. Chlamydomonas has proven invaluable for basic science inquiries into phototroph biology, contributing to the understanding of a wide range of processes including, photosynthesis, carbon concentrating mechanisms, carbon fixation, lipid biosynthesis, and starch biosynthesis 24–27. These processes were elucidated in part due to the genetic tractability of Chlamydomonas, a necessary tool in today’s age for basic and applied sciences. Unfortunately, Chlamydomonas has proven to be difficult to culture outdoors, and relatively unproductive when compared to other strains. As such researchers sought to identify more productive, genetically tractable organisms, these efforts led to the discovery of various

5 species of Nannochloropsis and Chlorella which have higher productivities, and established genetic engineering and omics knowledge bases 28–31. The diatom Phaeodactylum tricornutum has also been evaluated, along with numerous other less studied microalga 32–34. Despite the higher productivities of these algae, biofuels that are cost competitive with fossil fuels have yet to be achieved. This is most starkly demonstrated by the fact numerous companies have formed and failed over the last decade with the goal of commercializing algal biofuels. These companies are now defunct or have transitioned to higher value products (e.g. nutraceuticals).

Even though significant bioprospecting and screening efforts have occurred researchers have yet to evaluate the majority of available species found in nature. Indeed, published estimates of microalgal diversity range from 200,000 to several million species, thus nature has provided a diverse set of microalgae to examine for exploitation by humans 10. It is logical to hypothesize that evolution has selected for algae that would be more suitable for intensive farming, given the large diversity. As such, screening of diverse strains is one approach to identify microalgae that have characteristics beneficial for intensive farming and downstream processing to biofuels and bioproducts. To test this hypothesis a diverse culture collection was established, by Elliott et al. which opened the door for screening and down-selection of valuable strains, providing the basis for this thesis 35.

1.3 Outline of Chapters

The current chapter (1) serves to introduce the reader into the general concept and utility of microalgal cultivation, prior work done in this field, and how this thesis advances the field. In general, this thesis involved the screening of ~100 strains of microalgae under conditions

(seawater based media, temperature, light cycling) that simulated winter and summer growth conditions in raceway ponds in order to identify isolates that are promising for seasonal biomass

6 production. I then went on to develop genetic tools allowing for overexpression of proteins from both the nucleus and chloroplast for one of the top candidate strains, an isolate from the genus

Picochlorum.

Chapter 2 describes the culture collection screening for isolation of strains suitable for growth under winter conditions. Importantly, these strains were grown outdoors in Mesa Arizona at the AzCATI test bed site. Results regarding the productivity of the outdoor grown cultures, and associated biomass composition are presented. Chapter 3 is a review paper I wrote with Dr.

Michael Guarnieri wherein genetic tools utilized in algae are discussed. This chapter serves to introduce Chapter 4, wherein the results of a promising Picochlorum isolate are presented.

Biomass composition, general growth characteristics, and genetic tools for nuclear and chloroplast engineering are presented. Chapter 5 presents another promising strain, this one from the genus Scenedesmus which has an interesting growth phenotype, especially in contrast to the

Picochlorum. At the time of writing chapters 2, 3, and 4 are all published in the scientific literature, I reformatted them to fit the thesis formatting requirements, as such I would direct any readers to the original published manuscripts. There were some interesting findings that did not make it into the published manuscripts, I have included a series of appendices at the end of this thesis that briefly goes over these findings, as they may be helpful and interesting for other researchers.

1.4 References

1. U.S. Energy Information Administration - EIA - Independent Statistics and Analysis. Annual Energy Outlook 2019 Available at: https://www.eia.gov/outlooks/aeo/data/browser/#/?id=1-AEO2019®ion=0- 0&cases=highprice&start=2017&end=2050&f=A&linechart=~highprice-d111618a.31-1- AEO2019~highprice-d111618a.30-1-AEO2019~highprice-d111618a.32-1- AEO2019~highprice-d111618a.33-1-AEO2019~hig. (Accessed: 16th July 2019)

2. Field, C.B., V.R. Barros, D.J. Dokken, K.J. Mach, M.D. Mastrandrea, T.E. Bilir, M. Chatterjee, K.L. Ebi, Y.O. Estrada, R.C. Genova, B. Girma, E.S. Kissel, A. N. L. & S. MacCracken, P.R. Mastrandrea, and L. L. W. Climate Change 2014 Part A: Global and Sectoral Aspects. Climate Change 2014: Impacts, Adaptation, and Vulnerability. Part A: Global and Sectoral Aspects. Contribution of Working Group II to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change (2014). 3. Montgomery, D. R. Soil erosion and agricultural sustainability. (2007). 4. Pimentel, D. SOIL EROSION: A FOOD AND ENVIRONMENTAL THREAT. doi:10.1007/s10668-005-1262-8 5. Udom, B. E., Omovbude, S. & Abam, P. O. Topsoil removal and cultivation effects on structural and hydraulic properties. CATENA 165, 100–105 (2018). 6. Diaz, R. J. & Rosenberg, R. Spreading Dead Zones and Consequences for Marine Ecosystems. Science (80-. ). 321, 926–929 (2008). 7. Anderson, D. M., Glibert, P. M. & Burkholder, J. M. Harmful Algal Blooms and Eutrophication Nutrient Sources, Composition, and Consequences. 25, (2002). 8. Sheehan, J., Dunahay, T., Benemann, J. & Roessler, P. Look back at the U. S. Department of Energy’s Aquatic Species Program: Biodiesel from Algae; Close-Out Report. (1978). 9. Dismukes, G. C., Carrieri, D., Bennette, N., Ananyev, G. M. & Posewitz, M. C. Aquatic phototrophs: efficient alternatives to land-based crops for biofuels. Curr. Opin. Biotechnol. 19, 235–240 (2008). 10. Singh, J. & Saxena, R. C. An Introduction to Microalgae: Diversity and Significance. Handb. Mar. Microalgae 11–24 (2015). doi:10.1016/B978-0-12-800776-1.00002-9 11. Chisti, Y. Biodiesel from microalgae. Biotechnol. Adv. 25, 294–306 (2007). 12. Quinn, J. C. & Davis, R. The potentials and challenges of algae based biofuels: A review of the techno-economic, life cycle, and resource assessment modeling. Bioresour. Technol. 184, 444–452 (2015). 13. McGowen, J. et al. The Algae Testbed Public-Private Partnership (ATP3) framework; establishment of a national network of testbed sites to support sustainable algae production. Algal Res. 25, 168–177 (2017). 14. Algal Culture From Laboratory to Pilot Plant. (1953). 15. McGowen, J. et al. The Algae Testbed Public-Private Partnership (ATP3) framework; establishment of a national network of testbed sites to support sustainable algae production. Algal Res. 25, 168–177 (2017). 16. Davis, R. et al. Process Design and Economics for the Production of Algal Biomass: Algal Biomass Production in Open Pond Systems and Processing Through Dewatering for Downstream Conversion. (2016). 17. Biller, P., Sharma, B. K., Kunwar, B. & Ross, A. B. Hydroprocessing of bio-crude from continuous hydrothermal liquefaction of microalgae. Fuel 159, 197–205 (2015).

18. Jones, S. et al. Process Design and Economics for the Conversion of Algal Biomass to Hydrocarbons: Whole Algae Hydrothermal Liquefaction and Upgrading. (2014). 19. Garcia Alba, L., Torri, C., Fabbri, D., Kersten, S. R. A. & (Wim) Brilman, D. W. F. Microalgae growth on the aqueous phase from Hydrothermal Liquefaction of the same microalgae. Chem. Eng. J. 228, 214–223 (2013). 20. Dong, T. et al. Combined algal processing: A novel integrated biorefinery process to produce algal biofuels and bioproducts. Algal Res. (2015). doi:10.1016/j.algal.2015.12.021 21. Kindle, K. L., Schnell, R. A., Fernández, E. & Lefebvre, P. A. Stable nuclear transformation of Chlamydomonas using the Chlamydomonas gene for nitrate reductase. J. Cell Biol. 109, 2589–2601 (1989). 22. Boynton, J. et al. Chloroplast transformation in Chlamydomonas with high velocity microprojectiles. Science 240, 1538–41 (1988). 23. Rochaix, J.-D. & van Dillewijn, J. Transformation of the green alga Chlamydomonas reinhardii with yeast DNA. Nature 296, 70–72 (1982). 24. Goodenough, U. Historical perspective on Chlamydomonas as a model for basic research: 1950-1970. Plant J. 82, 365–369 (2015). 25. Harris, E. H. The Chlamydomonas Sourcebook : a Comprehensive Guide to Biology and Laboratory Use. (Elsevier Science, 1989). 26. Merchant, S. S., Kropat, J., Liu, B., Shaw, J. & Warakanont, J. TAG, You’re it! Chlamydomonas as a reference organism for understanding algal triacylglycerol accumulation. Curr. Opin. Biotechnol. 23, 352–363 (2012). 27. Moroney, J. V & Ynalvez, R. A. Proposed carbon dioxide concentrating mechanism in Chlamydomonas reinhardtii. Eukaryot. Cell 6, 1251–9 (2007). 28. Kilian, O., Benemann, C. S. E., Niyogi, K. K. & Vick, B. High-efficiency homologous recombination in the oil-producing alga Nannochloropsis sp. Proc. Natl. Acad. Sci. 108, 21265–21269 (2011). 29. Radakovits, R. et al. Draft genome sequence and genetic transformation of the oleaginous alga Nannochloropsis gaditana. Nat. Commun. 3, 686 (2012). 30. Guarnieri, M. T. et al. Examination of Triacylglycerol Biosynthetic Pathways via De Novo Transcriptomic and Proteomic Analyses in an Unsequenced Microalga. PLoS One 6, e25851 (2011). 31. Lammers, P. J. et al. Review of the cultivation program within the National Alliance for Advanced Biofuels and Bioproducts. Algal Res. 22, 166–186 (2017). 32. Breuer, G., Lamers, P. P., Martens, D. E., Draaisma, R. B. & Wijffels, R. H. The impact of nitrogen starvation on the dynamics of triacylglycerol accumulation in nine microalgae strains. Bioresour. Technol. 124, 217–226 (2012). 33. Radakovits, R., Jinkerson, R. E., Darzins, A. & Posewitz, M. C. Genetic engineering of

algae for enhanced biofuel production. Eukaryot. Cell 9, 486–501 (2010). 34. Daboussi, F. et al. Genome engineering empowers the diatom Phaeodactylum tricornutum for biotechnology. Nat. Commun. 5, 3831 (2014). 35. Elliott, L. G. et al. Establishment of a bioenergy-focused microalgal culture collection. Algal Res. 1, 102–113 (2012).

CHAPTER 2

DOWN-SELECTION AND OUTDOOR EVALUATION OF NOVEL, HALOTOLERANT ALGAL STRAINS FOR WINTER CULTIVATION

Reproduced with permission from Frontiers in Plant Science. Copyright © 2019 Frontiers Media Lukas R. Dahlin1, Stefanie Van Wychen2, Henri G. Gerken3, John McGowen3, Philip T. Pienkos2, Matthew C. Posewitz1, and Michael T. Guarnieri2

2.1 Abstract

Algae offer promising feedstocks for the production of renewable fuel and chemical intermediates. However, poor outdoor winter cultivation capacity currently limits deployment potential. In this study, 300 distinct algal strains were screened in saline medium to determine their cultivation suitability during winter conditions in Mesa, Arizona. Three strains, from the genera Micractinium, Chlorella and Scenedesmus, were chosen following laboratory evaluations and grown outdoors in 1000 L raceway ponds during the winter. Strains were down-selected based on doubling time, lipid and carbohydrate amount, final biomass accumulation capacity, cell size and phylogenetic diversity. Algal biomass productivity and compositional analysis for lipids and carbohydrates show successful outdoor deployment and cultivation under winter conditions for these strains. Outdoor harvest-yield biomass productivities ranged from 2.9 to 4.0 g/m2/day over an 18-day winter cultivation trial, with maximum productivities ranging from 4.0 to 6.5 g/m2/day, the highest productivities reported to date for algal winter strains grown in saline media in open raceway ponds. Peak fatty acid levels ranged from 9% to 26% percent of biomass,

1 Department of Chemistry, Colorado School of Mines, Golden, CO, USA 2 National Bioenergy Center, National Renewable Energy Laboratory, Golden, CO, USA 3 Arizona Center for Algae Technology and Innovation, Arizona State University, Mesa, AZ, USA

11 and peak carbohydrate levels ranged from 13% to 34% depending on the strain. Changes in the lipid and carbohydrate profile throughout outdoor growth are reported. This study demonstrates that algal strain screening under simulated outdoor environmental conditions in the laboratory enables identification of strains with robust biomass productivity and biofuel precursor composition. The strains isolated here represent promising winter deployment candidates for seasonal algal biomass production when using crop rotation strategies.

2.2 Introduction

Algae are a promising source of renewable biomass that can be converted to biofuels and bioproducts. Of the various terrestrial feedstocks currently grown at large scales, algae have the potential to achieve orders of magnitude higher productivities and can be grown on land not currently suitable for food production 1. Algae can also mitigate the problems associated with climate change by reducing the need to burn additional fossil fuels. In contrast to fossil fuels, algae utilize CO2 already in the atmosphere and can potentially be integrated with power plants to consume CO2 emissions from the generation of electricity or other point source CO2 emissions. Furthermore, many algae are able to use saline or brackish water resources which should reduce costs and avoid further demands on limited freshwater resources. Lastly, algae can be grown throughout the calendar year in many locations. However, despite the apparent promise of algal biomass, algae have yet to be fully utilized at industrial scales, due in part to the poor correlation of productivities measured in the laboratory to outdoor performance and ultimately lack of cost competitiveness with traditional petroleum-based fuel resources at current prices 2,3.

To realize the potential of algae, significant productivity improvements are required 3,4.

Cost and sustainability improvements can be made through the use of halotolerant algae grown

12 in saltwater. This is due to the global abundance of saltwater relative to limited freshwater resources, which is a critical consideration for contemporary agriculture and sustainability.

Further improvements can be made if a seasonal crop rotation strategy is employed. There is typically a substantial decrease in algal productivities during the winter months at most locations, and the identification and use of superior winter strains will improve overall annual biomass yields if strains better suited to outdoor growth in colder conditions are deployed during this season 2–7. Unfortunately, it is often difficult to identify strains in controlled laboratory environments that can be successfully cultured outdoors. As noted by Griffiths et al., 8, outdoor productivities are often diminished relative to laboratory yields due to the more extreme dynamics and levels of outdoor temperatures and light intensities, environmental conditions that are often not replicated during strain selection indoors. Therefore, we designed a growth system that better reflected the outdoor environmental parameters of light and temperature and then screened uncharacterized algal strains for isolates that perform well under real-world fluctuating temperature and light to identify strains with improved biomass productivities relative to previously tested isolates.

Herein, we targeted identification of halotolerant algal strains suitable for outdoor cultivation during winter for the production of biofuels and bioproducts. To achieve this, a previously uncharacterized algal culture collection was screened under simulated outdoor winter conditions using a seawater-based medium. As described by Elliott et al., this collection comprising novel strains was established via the isolation of algae from diverse water samples through the southwestern United States 9. Utilizing a custom-built photobioreactor, we screened this algal collection using real-world winter pond temperature and photosynthetically active radiation (PAR) inputs derived from 1000 L outdoor algal ponds located in Mesa, Arizona. After

13 screening in simulated outdoor winter conditions, three diverse, high-potential, down-selected strains were deployed outdoors to verify and quantify their suitability for large scale biomass and biofuel production.

2.3 Results

2.3.1 Indoor Screening

Initial down-selection involved screening for the ability of 300 strains to grow in saline media, as the algae were collected from diverse water sources with varied salinity. Of these 300,

107 showed growth in saline media, and were thus subjected to more detailed growth analysis as described in the methods. Of the 107 halotolerant strains screened only 69 grew under the simulated winter conditions. Table 2.1 shows the primary metrics used to determine further down-selection of the 107 strains, including biomass accumulation, FAME and carbohydrate content, and doubling time. Of the strains deployed outdoors, strain 14-F2 showed the highest net biomass accumulation (2.32 g/L) and shortest doubling time (~40 hours, including the dark period). Strain 4-C12 showed the highest FAME percentage (~37 %). Strain 46B-D3 had both the largest cell size (~12 µm) and carbohydrate percentage (~30 %), along with a relatively rapid

Table 2.1 Down-selection criteria and data for strains selected for outdoor cultivation.

14 auto settling capacity and carotenogenesis (data not shown). When choosing which strains to deploy, outdoor biomass productivity and composition were used as primary selection criteria, however qualitative observations of cell size, morphology, and auto settling capacity were also utilized to ensure a diverse set of organisms were cultivated outdoors. The diversity is ultimately reflected in the genus level speciation of the strains grown outdoors, 14-F2 (Micractinium reisseri), 4-C12 (Chlorella vulgaris), 46B-D3 (Scenedesmus rubescens); along with their respective cell sizes, 4.2 ± 0.5, 5.7 ± 0.7, and 11.7 ± 1.6 µm.

We attempted to use Nannochloropsis salina (CCMP 1776) as a baseline control in all down-selection trials. Nannochloropsis sp. are widely regarded as industrially relevant for outdoor growth due to their relatively high productivities, high lipid content, and stability during outdoor cultivation 6–8,10–13. Furthermore, the United States Department of Energy recently defined the winter strain state of technology based upon Nannochloropsis oceanica (KA32) productivity 14. Notably, Nannochloropsis salina and Nannochloropsis oceanica failed to grow under the simulated winter growth conditions employed here, likely due to the extreme cold temperatures employed. This observation further underscores the need for improved winter cultivation strains.

2.3.2 Outdoor Deployment

Due to the highly controlled nature of indoor laboratory screening, which shields algae from many outdoor environmental stresses, the top winter strains were grown outside in Mesa

Arizona at the AzCATI testbed site to determine whether the down-selected strains can be successfully cultivated outdoors during the winter growing season. Briefly, strains were cultivated in 1000 L raceway ponds during February 2016, which led to a cold start to cultivation

15 as discussed below, providing a well-suited winter experiment, which tested the extreme cold tolerance of the strains.

Overall harvest-yield biomass productivities were greatest for both 14-F2 and 4-C12, with productivities of 4.0 and 3.8 g/m2/day respectively. Strain 46B-D3 showed the lowest productivity of 2.9 g/m2/day (Figure 2.1). Maximum productivities showed a similar trend, however were naturally greater, as this excluded the ~6-day long lag phase (Figure 2.1). It is interesting to note that indoor biomass titers were as high as 2.32 g/L, whereas outdoor titers only reached 0.295 g/L (Table 2.1, Figure 2.1). This represents approximately an order of magnitude change in biomass titers from indoor to outdoor cultivation, underscoring the need to assess strains under real world outdoor conditions. Qualitative microscopic examination of outdoor cultures indicated negligible amounts of contaminating organisms in comparison to the

16 alga of interest. It is important to note that the strains were grown in relatively cold temperatures, mostly varying between 10-20°C, with a cold minimum (0°C) early in cultivation

(Figure 2.3). Additionally, the lower winter light duration and intensities tend to decrease growth rates and lipid content 15. Actual light intensities measured during the experiment are shown in

Figure 2.3. PAR peaked at approximately 1475 µmol photons/m2s, with measurable light for approximately 10.5 hours each day and 13.5 hours of darkness. Ammonium and phosphate consumption correlated with algal growth. Ammonium concentration in the media was not detectable at day 9 for strains 4-C12 and 14-F2, and at day 12 for 46B-D3, with phosphate concentrations showing a similar trend (Figure 2.3). Significant biomass accumulation was

17 observed after nutrients were depleted, indicating active carbon assimilation and photosynthesis

(Figure 2.1). As seen in Figure 2.2, the percentages of FAMEs and carbohydrates begins to increase at the approximate time that nutrients are depleted in the media; however, carbon fixation continues after this as AFDW continues to increase. Putative chlorophyll degradation was observed for all strains grown outdoors, as evident by a decrease in 680 nm/750 nm absorption ratios (data not shown).

Perhaps most notable, considering the growth conditions, is the demonstration of a relatively high lipid state in two of the strains grown outdoors, 14-F2 and 4-C12, which showed peak FAME percentages of 22.5 and 25.8, and titers of 0.066 and 0.072 g/L, respectively (Figure

2.2). As shown in Figure 2.4, lipids contained almost exclusively 16 and 18 long carbon chains with varying degrees of unsaturation depending on the strain and time point. The change in fatty acid chain lengths for all strains was quite dramatic from inoculation to endpoint harvest, with a substantial decrease in 16:2 and 18:2 chain lengths after inoculation. Concurrently, a decrease in their respective saturation was observed. The increase in lipids for strains 14-F2 and 4-C12 was

19 primarily due to accumulation of 18:1, 18:2, and 16:0 fatty acids, as shown in Figure 2.4. Strain

46B-D3 did not display an increase in lipid accumulation under the conditions tested.

Total carbohydrate accumulation is shown in Figure 2.2, which demonstrates that the down-selected strains are capable of accumulating high levels of total carbohydrates during winter months. Strains 14-F2 and 46B-D3 have the highest carbohydrate titers at ~0.07 g/L at day 18, with 46B-D3 having the highest carbohydrate percentage at ~34%. Figure 2.4 shows the sugar monomer content of hydrolyzed biomass. Glucose is heavily represented; however, measurable amounts of galactose, rhamnose, ribose and mannose are also present depending on the strain and time point. Notably, ~40-60% of the carbohydrate fraction or 13% of the biomass in 46B-D3 is mannose. The most dramatic changes in sugar monomer profiles for all strains were observed at 7 days after inoculation into the outdoor raceway ponds when the strains were beginning to exit from their initial lag phase and entering maximum productivity. After this time, the carbohydrate profiles of each of the strains trended back toward the initial inoculation profile.

All strains showed depletion in the amount of glucose from day 0 to 7. At day 7, strains 14-F2 and 4-C12 were largely composed of galactose in regard to sugar monomers; at the same time strain 46B-D3 was composed largely of mannose and glucose in regard to sugar monomers

(Figure 2.4).

Another important observation to note is the relatively high FAME and carbohydrate percentages of the seed culture used to inoculate the raceway ponds, likely representing an early stationary phase culture (Figure 2.2). This is putatively due to the seed cultures being grown in flat panel photobioreactors with a decreased light path, allowing greater light saturation, and in turn, accumulation of energy storage molecules. Figure 2.2 clearly shows these energy storage molecules are utilized for cell growth and maintenance upon inoculation into the ponds with

20 fresh media; as the percentages of both FAME and carbohydrates decrease from the time of inoculation to day 7. In fact, total carbohydrates (g/L) decreased for all strains from day 0 to 7.

The cold temperatures the cultures experienced from days 0-6, likely required remobilization and utilization of stored carbon to survive (Figure 2.3).

2.4 Discussion

2.4.1 Indoor Screening

Both lipids and carbohydrates provide potential precursors for biofuels and bioproducts.

Lipids can be directly converted to hydrocarbon-based fuels, whereas carbohydrates can be converted to fuel substitutes or valuable coproducts via fermentation in a CAP (combined algal processing) approach, and thus add additional value to algal biomass 16. It is clear that many algae can attain a high percentage of storage carbon; however, this metric alone does not necessarily predict overall lipid productivity on a volume or areal basis. Furthermore, it is well documented that outdoor growth in ponds does not necessarily correlate with the extremely high lipid storage state that is often observed in a laboratory setting or by the use of photobioreactors

(PBRs) 7,17. Therefore, both doubling time and overall biomass accumulation along with storage carbon percentage were used as selection criteria when deciding which strains would be deployed for outdoor testing.

Although the above criteria (as shown in Table 2.1) were the most heavily factored attributes used in the down-selection process, other assets including organismal diversity, cell size, auto settling capacity, and the ability of strains to accumulate high value co-products, were also considered. With respect to diversity, it was hypothesized that moving forward exclusively with highly similar isolates could lead to common outdoor growth outcomes. For example, it is

21 reported that many Chlorella isolates rapidly drop in productivity, ultimately requiring fresh inoculum in outdoor trials due to problems associated with predation and contaminating organisms, such as Vampirovibrio chlorellavorus 18,19. As such, genus level diversity was targeted in the final down-selection. Cell size and morphology is suggested to be a mechanism that algae use to avoid predation; therefore cells with unusually large cell size, or the tendency to clump or stack may be preferable for outdoor cultivation despite slightly lower productivities measured in the laboratory 5,20,21. Algal morphological alterations have been observed in response to predation/predator infochemicals. Thus, sterile lab conditions may not correlate to the actual morphology observed during pond deployment, as previously reported for

Scenedesmaceae and Micractinium 22–24, hence the logic in choosing these genera for the outdoor deployment reported here. The ability to avoid predation by this mechanism is simply conferred by the algal cell’s ability to become too large for many predator classes to consume. Another important metric is settling time, as this allows for greatly reduced volumes of water containing algae that need to be processed during dewatering. Gravity sedimentation is a relatively cheap primary harvesting technique that can concentrate algal biomass to 1.5% TSS (total suspended solids), from an initial TSS of ~0.04% 25. Deploying strains that inherently settle out of solution quickly could bypass the need for additional bio or chemical based flocculation, which contributes to the operational expenses of an algal farm, ultimately reducing final production costs 26,27. The ability of strains to accumulate carotenoids or high value polyunsaturated fatty acids was also considered as these are generally considered high value co-products, which could be used to increase the cost competitiveness of algal biofuels 27. A number of strains were noted to produce the high value omega 3 fatty acids, EPA (eicosapentaenoic acid) and DHA

(docosahexaenoic acid) unfortunately, these strains displayed relatively low productivities under

22 the tested conditions and were therefore not pursued past initial screening.

2.4.2 Outdoor deployment

The data presented here validated the indoor screening and selection methodology, with the down-selected strains showing growth and composition suitable for biofuel and bioproducts production when considering the cold temperatures of the ponds (Figures 2.1-2.3). These data represent the highest productivities reported to date for algal winter strains grown in saline media in open raceway ponds. This is in comparison to the productivities reported by McGowen et al, wherein similar cultivation conditions were utilized to grow N. oceanica (KA32) 28. As discussed above, storage carbon accumulation and chlorophyll degradation correlate with nutrient depletion; however, carbon fixation continued, as reflected by AFDW increase. This suggests the cells are initiating a nutrient deprivation response once the nutrients in the media are depleted, despite the fact there are sequestered nutrients in the cells. This is supported by Li et al. 29, wherein the authors postulated that nitrogen containing compounds such as chlorophyll are remobilized as a nitrogen source for cell growth once nitrogen in the media is depleted. These authors observed chlorophyll depletion despite continued biomass accumulation 29 and reported a doubling of biomass after nitrogen depletion in the media, which is similar to the results obtained in this study.

4-C12 and 14-F2 show a clear increase in the percentages of specific fatty acid chains lengths, likely due to nitrogen limitation (Figure 2.4), an essential attribute for successful biofuel production. There is clearly a distinct FAME profile when the strains are grown in the indoor flat panel photobioreactors, which were used to generate the starting seed culture (t=0 in Figure 2.4) as this initial profile is not recapitulated in the outdoor raceway ponds, likely due to variations in

23 pond temperature and light intensity in comparison to the photobioreactors. From day 0 to day 7 there is a marked increase in unsaturated lipids, as C16:2 appears to be replaced by C16:3 or

C16:4 depending on the strain; simultaneously C18:2 is replaced by C18:3. This observation of an increase in unsaturation at cold temperatures is well documented in the literature for a variety of organisms 30–32.

Although carbohydrates are not as easily utilized for biofuels relative to lipids, they represent an important fraction of biomass that can be fermented to fuels and/or higher value products. The dramatic changes in glucose content is likely due to the utilization of glucose for cell division, maintenance, and coping with the relatively cold temperatures from day 0 to 7

(Figure 2.3). This is supported by the increased glucose percentages after day 7, likely due to the accumulation of carbohydrates for energy storage. Galactose and mannose may contribute to the cell wall composition as suggested by Blumreisinger et al. and Takeda 33,34; the portions of rhamanose (14-F2) and ribose (4-C12 and 14-F2) likely represent other structural components of the cell, or for the case of ribose, Calvin-Benson cycle intermediates and nucleic acids. A mannose polysaccharide may be used as an energy store and cell wall constituent in strain 46B-

D3, given the comparatively high mannose percentage 35. Determining the exact nature of the mannose quantified will be addressed in future research as non-starch based polysaccharides have a variety of industrial uses 36–38.

As noted above, relatively high storage carbon percentages were observed in seed cultures used to inoculate the raceway ponds, likely representing an early stationary phase culture. Inoculation of cells with relatively high stored energy reserves could prove useful as the cells will have the energy needed to quickly adapt to the new conditions present in the pond. The physiological status (growth phase and biomass composition i.e. degree of lipid unsaturation) of

24 seed cultures represents an area of future research optimization as cells that experience a reduced initial transfer shock as the algae are transitioned from a photobioreactor to a pond, will likely have increased productivities.

The winter productivity data reported here are unique in that metrics for biomass, along with the associated lipid and carbohydrate composition are reported under real world deployment conditions. Furthermore, algal ponds were used for growth, which is significant as techno- economic analysis has suggested that for algal biofuels to be cost competitive with traditional petroleum sources, algal ponds should be employed due to the low capital expenditure relative to photobioreactors 3,39. Additionally, there is an economic benefit in using low-cost saline media, as well as positive attributes associated with the sustainability of saline water when compared to freshwater resources. The data herein provide a rare data set of algal composition and productivity during the winter growth season that researchers can utilize in techno-economic analysis.

Our data demonstrate that algal strain screening under simulated outdoor environmental conditions enables identification of strains with robust biomass productivity and biofuel precursor composition. This can be especially useful in future efforts as algae could be cultivated in a variety of geographic locations representing unique temperature and light regimes, along with other environmental variables. Importantly, we have identified a series of halotolerant strains with the highest productivities reported to date for algal winter strains grown in saline media in open raceway ponds, with biomass compositions useful for biofuels and bioproducts.

These strains represent promising candidates for further development and deployment. Future studies will focus on optimization of culture conditions, down-selection of superior summer strains, and generation of facile genetic toolkits for these promising halotolerant algae.

2.5 Methods

2.5.1 Culture Collection Screening

Preliminary screening efforts utilized a custom-built photobioreactor capable of simulating the variations in PAR and temperature common in outdoor raceway ponds. This was achieved using a 73.4-liter polycarbonate aquarium (109.9 x 29.2 x 29.2 cm), connected to a programmable water chiller/heater circulating pump (VWR, Model AP15R-30-V11B). White

LED (light emitting diode) lights (Lithonia, Model

IBH24000LMSD080MDMVOLTGZ1040K70CRICS93WWH) illuminated both of the longer sides of the water bath, and were connected to a programmable dimmer (Pulsar, Model P02C1-

375HS_SERIAL), this provided uniform lighting to all culture tubes. Carbon dioxide was mixed with air to 2% CO2 and delivered via rotameters to 36 individual tubes at 100 mL/min, which also provided aeration-mediated mixing to the 100 mL cultures. Winter water temperature cycled from 6°C to 14°C in a sinusoidal fashion daily which represents the average temperatures measured in 1000 L elevated raceway ponds at the Arizona Center for Algae Technology and

Innovation (AzCATI) testbed site located at Arizona State University in Mesa Arizona from

December 17th 2014 to January 28th 2015. At the same time, PAR cycled from 0 to 965 µmol m-2 s-1, over 9 h, followed by 15 h of darkness; this closely represented the PAR conditions at outdoors ponds at the time and location described above.

To initiate the screening process, 300 strains from diverse environments were first evaluated for their ability to grow in seawater-based media (defined below). Of the 300, 107 strains were found to be halotolerant (growth in seawater-based media) and were further screened in duplicate under simulated winter conditions. In each screening run, approximately 35

26 unique strains were each grown in single chambers, from a starting optical density (750 nm) of

0.2. To compare the most promising isolates, the top performing strains from each winter run were run for a 3rd time to produce triplicate data. This represented 33 promising winter isolates chosen by the down-selection criteria that included: doubling time, lipid and carbohydrate amount, final biomass accumulation capacity, cell size and phylogenetic diversity. During all runs, optical density (750 nm) was measured over a time course (measurements every other day), from which doubling time could be calculated, followed by endpoint compositional analysis, which included analysis of moisture, ash, lipid content as converted fatty-acid-methyl-esters

(FAME), and carbohydrates. The reported doubling time, calculated during linear growth phase, includes time in which the lights were off simulating night time (15 hours a day), and thus is a representation of a real-world winter doubling time including light and temperature variation rather than the shortest doubling time possible under ideal conditions. The cell size reported was calculated by averaging 20 linear phase cells measured with a Nikon Eclipse Ci microscope, utilizing NID Elements v. 4.40.00 software and calibrated with an improved Neubauer hemocytometer; average and the respective standard deviation is reported in the results.

The media used in screening was a modified f/2 formulation 13. To seawater (Gulf of

Maine, Bigelow Laboratory) the following were added to the indicated final concentration

-3 followed by addition of 12 M HCl to attain pH 8.0: NH4Cl (5.0 x 10 M), NaH2PO4∙H2O (0.313

-3 -4 -5 x 10 M), Na2SiO3∙9H2O (1.06 x 10 M), FeCl3∙6H2O (1.17 x 10 M), Na2EDTA∙2H2O (1.17 x

-5 -8 -8 - 10 M), CuSO4∙5H2O (3.93 x 10 M), Na2MoO4∙2H2O (2.60 x 10 M), ZnSO4∙7H2O (7.65 x 10

8 -8 -7 -7 M), CoCl2∙6H2O (4.20 x 10 M), MnCl2∙4H2O, (9.10 x 10 M), thiamine HCl (2.96 x 10 M), biotin (2.05 x 10-9 M), cyanocobalamin (3.69 x 10-10 M), Tris base (24.76 x 10-3 M).

2.5.2 18S sequencing

Phylogenetic sequencing was conducted using universal eukaryotic primers to amplify the 18S rRNA gene. Reverse primer, 1391RE 5- GGGCGGTGTGTACAARGRG-3 and forward primer 360FE 5-CGGAGARGGMGCMTGAGA-3 were used for all amplifications 40. Colony

PCR was conducted wherein a single colony was picked from a solid agar matrix, placed in 20

µL of Y-PER lysis buffer (Sigma), boiled for 5 min, and diluted by addition of 150 µL of H2O.

PCR was carried out with 2 µL of the cell lysate, 25 µL of Q5® Hot Start High-Fidelity 2X

Master Mix (New England Biolabs), 18 µL of H2O, and 2.5 µL each of the forward and reverse primers (10 µM) 41. The thermocycler protocol is as follows: 94°C initial denature for 30 s, followed by 30 cycles of 94°C denaturation for 30 s, 63°C annealing for 15 s, 72°C elongation for 1 min and 20 s, a 2 min final elongation at 72°C was utilized. The 18S amplicon was sequenced (Genewiz) and the NCBI BLAST algorithm 42 used as an initial determination of speciation of the novel strains. The above methods were also used to verify the speciation in outdoor growth, DNA was extracted using a MasterPure™ Yeast DNA Purification kit (Lucigen) from day 0 seed cultures, lyophilized biomass samples, ultimately confirming correct speciation via 18S sequences. 18S sequences were deposited into the Sequence Read Archive, submission

ID: SRP136934.

2.5.3 Outdoor growth

Following down-selection via indoor screening, 3 promising winter isolates 14-F2

(Micractinium reisseri), 4-C12 (Chlorella vulgaris), and 46B-D3 (Scenedesmus rubescens) were deployed outdoors utilizing the AzCATI testbed site located at Arizona State University in Mesa

Arizona during February of 2016. Cultures were scaled up indoors using bubble columns and 15

L flat panel reactors as described in McGowen, et al. 2017 28, and finally inoculated into 1000 L raceway ponds at a 25 cm depth, with an area of 4.2 m2. Media consisted of 35 g/L salinity achieved using Crystal Sea (Marine Enterprises International), 2.14 mM ammonium chloride and

0.134 mM phosphate (NaH2PO4∙H2O). Trace metals were added to the following final

-5 -5 concentrations (standard f/2 media), FeCl3∙6H2O (1.17 x 10 M), Na2EDTA∙2H2O (1.17 x 10

-8 -8 -8 M), CuSO4∙5H2O (3.93 x 10 M), Na2MoO4∙2H2O (2.60 x 10 M), ZnSO4∙7H2O (7.65 x 10

-8 -7 M), CoCl2∙6H2O (4.20 x 10 M), MnCl2∙4H2O, (9.10 x 10 M). Inoculation density was targeted at 0.05 g/L of algal biomass, and sodium carbonate was added as necessary to control the pH drop due to consumption of NH3 from the added NH4Cl. Pond pH was adjusted to 7.9 via a pH controller connected to a solenoid delivering CO2 gas, and PAR was measured continuously with a LiCor LI-190R quantum pyranometer 28. Freshwater was added prior to sampling to account for evaporative loses. Biomass samples for analysis were taken in the morning (07:30 – 08:00), except for inoculation biomass which was sampled at (15:00). Qualitative microscopic examination of outdoor cultures was conducted periodically to assess contamination. In order to compare strains, two productivity metrics were used. The first was productivity calculated during linear phase growth, termed maximum productivity here. The second, harvest-yield productivity, is a measure of the net productivity over the course of the entire cultivation trial, which includes the lag, linear, and stationary culturing phases. The end of the harvest-yield was chosen when cultures were not adding significant biomass during the day, and nutrients had not been measurable for multiple days in the media.

2.5.4 Compositional Analysis

Compositional analysis for ash (Van Wychen and Laurens 2013), FAME 44,45, and carbohydrate (Templeton et al. 2012; Van-Wychen and Laurens 2013) were conducted as

29 previously reported. Briefly, residual moisture after lyophilization was determined by heating of pre-weighed algal biomass at 40°C for at least 18 hours, followed by ash analysis by ramping at

20°C/min to 575°C, and held for 180 minutes. FAME analysis was conducted by transesterification of 10 mg of lyophilized algal biomass in 200 µL chloroform:methanol (2:1 v/v) and 300 µL 0.6M HCl:methanol (2.1% v/v) at 85°C for 1 hour. FAMEs were extracted with

1.0 mL hexane and analyzed via a gas chromatograph with flame ionization detection (Agilent

7890A equipped with a split/splitless inlet and an Agilent J&W GC Column DB-Wax 30m x

0.25mm x 0.25 um). Tridecanoic acid methyl ester (C13:0ME) was used as an internal standard.

Carbohydrates were determined by hydrolysis of 25 mg lyophilized biomass with 250 µL of 72%

(w/w) sulfuric acid at 30°C for 1 hour, followed by dilution with 7 mL of water and autoclaving for one hour at 121°C. Samples were filtered through a 0.2 um nylon filter, and subjected to high pressure anion exchange liquid chromatography with pulsed amperometric detection on a PA-20 column (Thermo Scientific – Dionex ICS-5000+ Analytical HPIC System). For outdoor cultivation, ash free dry weight (AFDW) was tracked throughout outdoor growth using the methodology described in McGowen et al. 2017 28, which involved filtering the algal culture through a pre-weighed glass microfiber filter followed by drying, ashing and re-weighing. Three independent trials (biological triplicates) were conducted for all analyses; average and standard deviations were obtained for replicates.

2.6 Acknowledgements

The authors would like to thank Nick Sweeney (National Renewable Energy Laboratory,

NREL) for contributions to culture and facilities maintenance, Jeff Wolfe (NREL) for assistance with FAME analyses, and Theresa Rosov and Thomas Dempster (Arizona Center for Algae

Technology and Innovation) for their contributions to outdoor cultivation.

2.7 References

1. Chisti, Y. Biodiesel from microalgae. Biotechnol. Adv. 25, 294–306 (2007). 2. Sheehan, J., Dunahay, T., Benemann, J. & Roessler, P. Look back at the U. S. Department of Energy’s Aquatic Species Program: Biodiesel from Algae; Close-Out Report. (1978). 3. Quinn, J. C. & Davis, R. The potentials and challenges of algae based biofuels: A review of the techno-economic, life cycle, and resource assessment modeling. Bioresour. Technol. 184, 444–452 (2015). 4. Davis, R. et al. Process Design and Economics for the Production of Algal Biomass: Algal Biomass Production in Open Pond Systems and Processing Through Dewatering for Downstream Conversion. (2016). 5. White, R. L. & Ryan, R. A. Long-Term Cultivation of Algae in Open-Raceway Ponds: Lessons from the Field. Ind. Biotechnol. 11, 213–220 (2015). 6. Quinn, J. C. et al. Nannochloropsis production metrics in a scalable outdoor photobioreactor for commercial applications. Bioresour. Technol. 117, 164–171 (2012). 7. Lammers, P. J. et al. Review of the cultivation program within the National Alliance for Advanced Biofuels and Bioproducts. Algal Res. 22, 166–186 (2017). 8. Griffiths, M. J. & Harrison, S. T. L. Lipid productivity as a key characteristic for choosing algal species for biodiesel production. J. Appl. Phycol. 21, 493–507 (2009). 9. Elliott, L. G. et al. Establishment of a bioenergy-focused microalgal culture collection. Algal Res. 1, 102–113 (2012). 10. Neofotis, P. et al. Characterization and classification of highly productive microalgae strains discovered for biofuel and bioproduct generation. Algal Res. 15, 164–178 (2016). 11. Rodolfi, L. et al. Microalgae for oil: Strain selection, induction of lipid synthesis and outdoor mass cultivation in a low-cost photobioreactor. Biotechnol. Bioeng. 102, 100–112 (2009). 12. Bondioli, P. et al. Oil production by the marine microalgae Nannochloropsis sp. F&M- M24 and Tetraselmis suecica F&M-M33. Bioresour. Technol. 114, 567–572 (2012). 13. Guillard, R. R. L. & Ryther, J. H. STUDIES OF MARINE PLANKTONIC DIATOMS: I. CYCLOTELLA NANA HUSTEDT, AND DETONULA CONFERVACEA (CLEVE) GRAN. Can. J. Microbiol. 8, 229–239 (1962). 14. Office of Energy Efficiency and Renewable Energy, D. of E. Productivity Enhanced Algae and Tool-Kits. (2017). 15. Khotimchenko, S. V. & Yakovleva, I. M. Lipid composition of the red alga Tichocarpus crinitus exposed to different levels of photon irradiance. Phytochemistry 66, 73–79 (2005). 16. Dong, T. et al. Combined algal processing: A novel integrated biorefinery process to

produce algal biofuels and bioproducts. Algal Res. (2015). doi:10.1016/j.algal.2015.12.021 17. Unkefer, C. J. et al. Review of the algal biology program within the National Alliance for Advanced Biofuels and Bioproducts. Algal Res. (2016). doi:10.1016/j.algal.2016.06.002 18. Ganuza, E., Sellers, C. E., Bennett, B. W., Lyons, E. M. & Carney, L. T. A novel treatment protects Chlorella at commercial scale from the predatory bacterium Vampirovibrio chlorellavorus. Front. Microbiol. 7, 1–13 (2016). 19. Coder, D. M. & Starr, M. P. Antagonistic association of the chlorellavorus bacterium (“Bdellovibrio” chlorellavorus) withChlorella vulgaris. Curr. Microbiol. 1, 59–64 (1978). 20. National Alliance for Advanced Biofuels and Bio-products. Final Report Section II. 197 (2014). 21. Porter, K. G. Selective Grazing and Differential Digestion of Algae by Zooplankton. Nature 244, 179–180 (1973). 22. Wu, X., Zhang, J., Qin, B., Cui, G. & Yang, Z. Grazer density-dependent response of induced colony formation of Scenedesmus obliquus to grazing-associated infochemicals. Biochem. Syst. Ecol. 50, 286–292 (2013). 23. Yang, Z. et al. Increased Growth of Chlorella pyrenoidosa (Chlorophyta) in Response to Substances from the Rotifer Brachionus calyciflorus. J. Freshw. Ecol. 234, 545–552 (2008). 24. Luo, W., Pflugmacher, S., Pröschold, T., Walz, N. & Krienitz, L. Genotype versus Phenotype Variability in Chlorella and Micractinium (Chlorophyta, Trebouxiophyceae). Protist 157, 315–333 (2006). 25. Uduman, N., Qi, Y., Danquah, M. K., Forde, G. M. & Hoadley, A. Dewatering of microalgal cultures: A major bottleneck to algae-based fuels. J. Renew. Sustain. Energy 2, 012701 (2010). 26. Chen, C. Y., Yeh, K. L., Aisyah, R., Lee, D. J. & Chang, J. S. Cultivation, photobioreactor design and harvesting of microalgae for biodiesel production: A critical review. Bioresour. Technol. 102, 71–81 (2011). 27. Brennan, L. & Owende, P. Biofuels from microalgae-A review of technologies for production, processing, and extractions of biofuels and co-products. Renew. Sustain. Energy Rev. 14, 557–577 (2010). 28. McGowen, J. et al. The Algae Testbed Public-Private Partnership (ATP3) framework; establishment of a national network of testbed sites to support sustainable algae production. Algal Res. 25, 168–177 (2017). 29. Li, Y., Horsman, M., Wang, B., Wu, N. & Lan, C. Q. Effects of nitrogen sources on cell growth and lipid accumulation of green alga Neochloris oleoabundans. Appl. Microbiol. Biotechnol. 81, 629–636 (2008).

30. Neidleman, S. Effects of Temperature on Lipid Unsaturation. Biotechnol. Genet. Eng. Rev. 5, 245–268 (1987). 31. Upchurch, R. G. Fatty acid unsaturation, mobilization, and regulation in the response of plants to stress. (2008). doi:10.1007/s10529-008-9639-z 32. Sheng, J. et al. Effects of temperature shifts on growth rate and lipid characteristics of Synechocystis sp. PCC6803 in a bench-top photobioreactor. Bioresour. Technol. 102, 11218–11225 (2011). 33. Blumreisinger, M., Meindl, D. & Loos, E. Cell wall composition of chlorococcal algae. Phytochemistry 22, 1603–1604 (1983). 34. Takeda, H. Sugar Composition of the Cell Wall and the Taxonomy of Chlorella.pdf. J. Phycol. 27, 224–232 (1991). 35. Northcote, D. H., Goulding, K. J. & Horne, R. W. The Chemical Composition and Structure of the Cell Wall of Chlorella pyrenoidosa. Biochem. J. 70, 391–7 (1958). 36. Barati, R. & Liang, J.-T. A review of fracturing fluid systems used for hydraulic fracturing of oil and gas wells. J. Appl. Polym. Sci. 131, n/a-n/a (2014). 37. Abd Alla, S. G., Sen, M. & El-Naggar, A. W. M. Swelling and mechanical properties of superabsorbent hydrogels based on Tara gum/acrylic acid synthesized by gamma radiation. Carbohydr. Polym. 89, 478–485 (2012). 38. Brummer, Y., Cui, W. & Wang, Q. Extraction, purification and physicochemical characterization of fenugreek gum. Food Hydrocoll. 17, 229–236 (2003). 39. Davis, R., Aden, A. & Pienkos, P. T. Techno-economic analysis of autotrophic microalgae for fuel production. Appl. Energy 88, 3524–3531 (2011). 40. Dawson, S. C. & Pace, N. R. Novel kingdom-level eukaryotic diversity in anoxic environments. (2002). 41. Packeiser, H., Lim, C., Balagurunathan, B., Wu, J. & Zhao, H. An Extremely Simple and Effective Colony PCR Procedure for Bacteria, Yeasts, and Microalgae. Appl. Biochem. Biotechnol. 169, 695–700 (2013). 42. NCBI Resource Coordinators. Database Resources of the National Center for Biotechnology Information. Nucleic Acids Res. 45, D12–D17 (2017). 43. Van Wychen, Stefanie, Laurens, L. M. L. Determination of Total Solids and Ash in Algal Biomass Laboratory Analytical Procedure (LAP) Issue Date : December 2 , 2013 Determination of Total Solids and Ash in Algal Biomass Laboratory Analytical Procedure. (2013). 44. Laurens, L. M. L., Quinn, M., Van Wychen, S., Templeton, D. W. & Wolfrum, E. J. Accurate and reliable quantification of total microalgal fuel potential as fatty acid methyl esters by in situ transesterification. Anal. Bioanal. Chem. 403, 167–178 (2012).

45. Van Wychen, S. & Laurens, L. M. L. Determination of Total Lipids as Fatty Acid Methyl Esters (FAME) by in situ Transesterification. Contract 303, 275–3000 (2013). 46. Templeton, D. W., Quinn, M., Van Wychen, S., Hyman, D. & Laurens, L. M. L. Separation and quantification of microalgal carbohydrates. J. Chromatogr. A 1270, 225– 234 (2012). 47. Van-Wychen, Stefanie, Laurens, L. M. . [NREL] Determination of Total Carbohydrates in Algal Biomass. (2013).

CHAPTER 3

RECENT ADVANCES IN ALGAL GENETIC TOOL DEVELOPMENT

3.1 Abstract

The goal of achieving cost-effective biofuels and bioproducts derived from algal biomass will require improvements along the entire value chain, including identification of robust, high- productivity strains and development of advanced genetic tools. Though there have been modest advances in development of genetic systems for the model alga Chlamydomonas reinhardtii, progress in development of algal genetic tools, especially as applied to non-model algae, has generally lagged behind that of more commonly utilized laboratory and industrial microbes. This is in part due to the complex organellar structure of algae, including robust cell walls and intricate compartmentalization of target loci, as well as prevalent gene silencing mechanisms, which hinder facile utilization of conventional genetic engineering tools and methodologies.

Recent progress in global tool development has opened the door for implementation of strain- engineering strategies in industrially-relevant algal strains. These developments will facilitate the use of microalgae in a variety of applications relating to biofuels and bioproducts. Here, we review recent advances in algal genetic tool development and applications in eukaryotic microalgae.

1 Department of Chemistry, Colorado School of Mines, Golden, CO, USA 2 National Bioenergy Center, National Renewable Energy Laboratory, Golden, CO, USA

3.2 Introduction Algal cell factories offer promising platforms for the production of an array of biofuels and bioproducts, including pharmaceuticals, industrial platform chemicals and materials, and nutraceuticals 1,2. Though some wild-type microalgae have natural properties amenable to direct industrial deployment (e.g. astaxanthin production in Haematococcus sp.), genetic engineering has played, and will continue to play, a critical role in development of algal biocatalysts. Fine- tuned, coordinated gene expression and carbon flux maximization, facilitated by targeted genetic enhancements, will be required to develop robust algal biocatalysts suitable for outdoor cultivation, producing viable titers, rates, and yields of biofuels and bioproducts. To this end, advances in algal strain engineering have already led to the development of platforms targeting the production of therapeutic proteins and carotenoid-based nutraceuticals 3–6, with multiple industrial entities actively pursuing additional high-value products and fuels from algae.

Additionally, advances in genomic and functional post-genomic applications have led to the identification of promising strain-engineering targets for the development of algal biofuel platforms 7–11.

However, the absence of genetic toolboxes in many promising strains has largely limited progress in the development of deployable algal biocatalysts. Recent progress in genetic tool development has now opened the door for implementation of strain-engineering strategies in these promising, multi-use feedstocks (Figure 3.1). Herein, we highlight some key advances in eukaryotic microalgal genetic tool development, with a focus on non-model algae, and discuss inherent challenges to algal genetic engineering.

3.3 Random Gene Integration and Homologous Recombination

Historically, algal transformation has been achieved via the employment of conventional integrative plasmids, which are chromosomally incorporated through non-homologous end joining (NHEJ). This process inserts target DNA (typically in the form of, or coupled to, a selective marker) via direct ligation of double strand break ends, which facilitates chromosomal integration of DNA. Numerous examples of NHEJ random integration have been demonstrated in the model alga, Chlamydomonas reinhardtii, notably contributing to elucidation of mechanisms governing carbon concentration and partitioning 12–14, among a myriad of other key findings (for a complete review of genetic tool development in C. reinhardtii, please refer to 15 and 16. Random integration in Phaeodactylum tricornutum has been used for the industrially- relevant applications of tuning fatty acid chain length by overexpression of specific thioesterases, as well as engineering increased omega-3 polyunsaturated fatty acids 17,18. Another recent advance is transformation of Chaetoceros gracilis via random gene integration, which

37 demonstrated expression of both selectable markers and fluorescent proteins. However the authors note that a method for directed genome editing, discussed below, is needed via newer technologies 19. Indeed, despite the large role random integration approaches have played in the advancement of algal biology, they lack almost all ability to fine-tune expression levels or target specific integration loci, limiting their application in complex, targeted algal metabolic engineering endeavors. Additionally, complications may arise due to concatenation, incomplete or altered insertion of target cassettes, and/or deletion of indirect targets.

Through the use of homologous recombination (HR), an alternative mechanism for repair of breaks in double-strand DNA that proceeds via direct exchange between two similar or identical DNA molecules, some of the aforementioned problems may be resolved, enabling precision genome engineering. Unsurprisingly, homologous recombination in algae was first observed in C. reinhardtii 20. It has since been employed in other, more industrially-relevant algal species, such as Porphyridium sp., Nannochloropsis sp., Ostreococcus tauri, and

Dunaliella tertiolecta 21–24, successfully demonstrating both targeted transgene integration and expression, as well as targeted gene knockout. In both Nannochloropsis and Ostreococcus, the nuclear genome was targeted for alterations in nitrogen regulatory mechanisms to show the presence of homologous recombination mechanisms, whereas in Dunaliella and Porphyridium the chloroplast was targeted for expression of industrially relevant enzymes and herbicide resistance, respectively. Interestingly the chloroplast is more readily transformed via homologous recombination than NHEJ 24. HR within the mitochondria is also possible, though to date has only been reported in C. reinhardtii, where it has been modified to express a fluorescent protein and bleomycin resistance 25–27. Genomic interrogation to identify the absence of Ku heterodimer repair machinery, essential for NHEJ, may implicate the potential for HR in other non-model

38 algae, and offers a target for knockout in promising strains where NHEJ is prevalent, as described previously 16.

3.4 Multi-copy Gene Expression and Expression Control

Though random integration and targeted HR can offer effective means to integrate transgenes, gene copy number is often limited to a single copy (barring multi-insertion, duplication, and/or concatenation events). Replicative plasmids offer a means to express multiple copies of a gene, often accompanied by extreme phenotypic differential. Though replicative plasmids are routinely employed in industrial microbes, the lack of identified replication origins, origins of transfer, and/or autonomous replicating sequences in algae have effectively limited their utilization in basic and applied algal biology pursuits.

Eukaryotic episomes, extrachromosomal sequences that are stably maintained and replicated independently of nuclear DNA, are analogous to prokaryotic replicative plasmids.

These constructs represent an ideal vector for protein expression, given their potential for high copy number and plastid replication capacity, avoiding limitations typically associated with nuclear expression. Recently Karas, et al. 28 demonstrated the ability to express nuclear episomes in the diatoms Phaeodactylum tricornutum and Thalassiosira pseudonana. This process employed an ‘episome rescue’ technique in which episomes harboring varied algal DNA sequences were iteratively extracted and reintroduced into E. coli and P. tricornutum; ultimately, the group identified a heterologous yeast sequence on the plasmid backbone that confers episomal replication capacity in algae in the absence of antibiotic selection. The authors further demonstrated the ability to transfer these episomes to the diatom via conjugation with E.coli, representing the first demonstration of transkingdom conjugational transfer in Stramenopiles, and yielding an extremely efficient means of transformation.

The potential for this technology is immense for elucidation and confirmation of an array of proposed algal regulatory mechanisms. For example, acetyl-CoA carboxylase (ACCase) and diacylglyceride acyl transferase (DGAT) have been identified as obvious strain engineering targets, due to their rate-limiting, integral roles in fatty acid and triacylglyceride biosynthesis, respectively 8. Episomal deployment offers a means to examine overexpression of these enzymes, potentially elucidating alternative upstream and downstream bottlenecks in their respective pathways and global cellular metabolism via parallel omic analyses. Additionally, post-transcriptional regulation has been observed by our group and others to be widespread in an array of microalgal species 7,29–31. Multi-copy overexpression of genes of interest offers a means to decipher the degree of post-transcriptional regulation and/or gene silencing governing phenotypes of interest (such as high-lipid states induced by overexpression of the aforementioned targets).

An alternative means to target multi-copy and multi-gene insertion is via transcriptional fusion, also termed CHYSEL (cis-acting hydrolase element). This technique fuses the foot-and- mouth-disease-virus (FMDV) 2A peptide in between two genes to be transformed, all contained in one open reading frame. Equal expression of the two genes of interest is achieved due to the failure of a peptide bond to form in the growing FMDV polypeptide; thus, two fully function proteins are formed, in equal ratios. Rasala and coworkers 32–34 have used this approach extensively for various purposes, including expression of xylanase and bleomycin resistance inside the nucleus, while simultaneously targeting the xylanase for secretion. The authors have also leveraged this technology to express numerous fluorescent proteins in C. reinhardtii, followed by fusion of bleomycin resistance to native α-tubulin and a fluorescent protein, which in turn were localized to the cytoskeleton and flagella. Finally, two reporter genes were

40 expressed under the control of one selection agent and both reporter genes were successfully targeted for expression in separate organelles.

Control of gene expression is also a critical factor in induction of a desired phenotype. A classical means of controlling gene expression is through selection and/or manipulation of the promoter sequence. This can be achieved by choosing constitutive promoters with varying strengths; the chloroplast psbA promoter is among the strongest -transgene promoters, and is routinely utilized for constitutive transgene expression. Alternatively, inducible promoters can be employed, enabling temporal control of gene expression. For example Iwai and colleagues 35 recently used a phosphate starvation inducible promoter from C. reinhardtii to drive DGTT4 expression in Nannochloropsis sp. Light- and macronutrient-responsive promoters are also commonly utilized for inducible gene expression in microalgae 36. Such promoters can often be identified by differential gene expression under light- and/or nutrient-limited conditions, as assessed by comparative omic analyses.

3.5 In vivo genome editing

A series of powerful new genetic engineering technologies, generally referred to as

“genome editing” tools, have recently emerged at the forefront of molecular biology. These techniques leverage DNA recognition/binding proteins coupled to endonucleases for targeted

DNA insertion or deletion. One such technology, zinc finger nucleases (ZFNs), is created via fusion of zinc finger DNA-binding domains and a DNA-cleavage domain. At present, there has been only one published report of ZFN deployment in microalgae; Sizova and colleagues designed a series of ZFNs to successfully target the C. reinhardtii channelrhodopsin gene, COP3

37. However, as noted by Carroll 38, TALENs and CRISPR/Cas systems (defined below) seem to be replacing ZFNs in other model organisms, except in a few highly specialized cases.

Similar to ZFNs, TALENs (Transcription activator-like effector nucleases) are synthetic restriction enzymes containing a DNA recognition domain and a DNA cleavage domain. The recognition domain utilizes a TAL effector, a protein capable of binding to specific DNA sequencing by altering two key amino acids in the TAL protein 38. The DNA cleavage domain is typically non-specific, such as the FokI endonuclease, but can be modified for specificity and activity. Once targeted DNA has been cleaved, genome editing can occur through HR or NHEJ.

There are a few examples of successful TALEN use in algae, first described by Daboussi, et al.

39, wherein the authors demonstrated targeted mutagenesis in the UDP-glucose pyrophosphorylase gene (a key carbohydrate biosynthetic component) of P. tricornutum, ultimately leading to a 45-fold increase in triacylglycerol accumulation. Knockout of the urease gene was also achieved via TALENs in P. tricornutum, which was confirmed by lack of growth on urea media, PCR, Southern blot, and Western blot analysis 40. An alternative application of this technology is through the use of TALEs (transcription activator-like effector, in the absence of nuclease activity). TALEs serve as a transcriptional activator when targeted to the promoter region of a gene. This application has recently been deployed in C. reinhardtii for enhanced expression of two arylsulfatases, as well as HLA3, a protein involved in inorganic carbon uptake

41,42.

CRISPR/Cas9 (clustered regularly interspaced short palindromic repeats/CRISPR associated protein 9) is a prokaryotic immune system, offering protection from viruses by integrating viral DNA into the genome, followed by destruction of viral DNA upon subsequent infection. This has been exploited for genetic engineering purposes, and offers a third genome- editing tool that allows for alteration or addition of specific genes to target loci of the genome.

This is possible through the use of guide RNAs that “guide” the Cas9 protein to specific loci, at

42 which time Cas9 causes a double strand break, allowing for NHEJ or HR and insertion of DNA.

(for an extensive review of CRISPR/Cas9 technology, please refer to 43. In algae, Jiang and coworkers showed that expression of Cas9 targeting four different genes, led to distinct modification of those genes within 24 hours of transformation in C. reinhardtii 44. However transformation efficiency was extremely low, giving rise to a single viable colony from sixteen different transformations, encompassing transformation of >109 cells. The authors hypothesized the low transformation efficiency was due to Cas9 toxicity, leading to transient expression.

Mussgnug concluded “expression of a Cas9-YFP fusion protein in C. reinhardtii was indeed possible and furthermore demonstrated specific Cas9 nuclease activity” 15.

3.6 Technical Hurdles and Considerations

The emergence of the above genetic tools holds great promise for the implementation of hypothesis-driven strain engineering strategies in microalgae. However, microalgal cellular organization and architecture present a series of inherent hurdles for genetic transformation.

Algal cell wall composition can vary greatly, but regularly includes complex arrangements of carbohydrates, polymers, proteins, and lipids 45,46. Additionally, the photosynthetic nature of these organisms is conferred, in part, by the chloroplast, a specialized organelle housing the machinery required for an array of functions including photosynthesis, the Calvin cycle, and importantly from a biofuels perspective, fatty acid biosynthesis. These structures contain their own genomic DNA as well as a unique multi-membrane structure, generally consisting of outer and inner membranes and a thylakoid system. Combined, the algal cell wall and chloroplast membranes can present a robust, recalcitrant physical barrier to introduction of foreign DNA.

This often necessitates biolistic particle bombardment as a means of mechanical disruption, though electroporation and conjugation (as discussed above), among other techniques, have been

43 successfully applied in some species. In addition to robust exterior physical barriers, algal chromosomal DNA is packaged in an extremely compact nucleosomal structure with tight histone association, often making for poor accessibility of target regions. This structure has been proposed to be a cause of chromatin-mediated gene silencing, another major algal transformation hurdle 47. Indeed, gene-silencing mechanisms are prevalent in algal systems, often resulting in transient and/or unstable gene expression. Related, design of integration cassettes has been shown to be highly sensitive to selection of untranslated region (UTR) and promoter sequences, selection markers, and codon optimization 36. Failure to utilize native and/or self-recognizing sequences can lead to effective gene silencing and/or transient expression.

Lastly, there are a series of practical considerations unique to algal transformation. These include the requirement for light and CO2 for photoautotrophic cultivation. Following mechanical or electrical disruption of algal cells, transformation protocols often call for light- limited outgrowth to allow cellular recovery. Cells are then routinely plated on selective media, and photodegradation of selection markers is common. When combined with the long doubling times of many microalgae (often >12hrs), this can lead to prevalent false positive integrant detection. Algal susceptibility to bacterial, fungal, viral, and predator contamination is also well- documented, adding yet another layer of complexity which may lead to false positive detection

(often due to the contaminant conferring resistance to selection markers and/or giving rise to algal satellite colonies).

3.7 Concluding Remarks

Microalgae have emerged as high-potential biocatalysts for an array of industrial pursuits, ranging from food to fuel production, due to their vast metabolic diversity and potential for direct photosynthetic bioproduction. To date, a series of academic endeavors have demonstrated proof-

44 of-concept algal biocatalysis for an array of bioproducts, and industrial pursuits are now online for production of a series of high value algal products (though commercialization of algal biofuels is notably yet to be achieved). This progress has been driven, in large part, by the development of advanced genetic tools and transformation methodologies, which allow targeted modification to gene expression levels and resultant algal phenotypes. These toolboxes are requisite for effective, genome-scale metabolic engineering strategies, targeting technoeconomically viable production of algal biofuels and co-products. As novel genetic tools continue to emerge, genetic engineering is certain to play an increasingly important role in algal biotechnology.

Here we have highlighted a few examples of recent advances in algal genetic toolbox development, including classical integrative approaches, replicative episomal gene expression, and genome editing (Figure 3.1). We note that this is far from an exhaustive evaluation of algal genetic tools developed to date, or applications thereof; a series of additional genetic modulators, notably including RNAi and miRNA gene silencing strategies, as well as sexual mating, all offer powerful genetic engineering options for basic and applied algal biology pursuits. Alternative methods for indirect engineering of organelles outside the nucleus, such as the use of transit peptides and targeting sequences, also offer promise, allowing for sequence targeting in the mitochondria, endoplasmic reticulum, and nucleus 32–34,48.

Additionally, we have focused exclusively on advances in eukaryotic microalgae; though beyond the scope of this review, we note that there is also a wealth of promising research ongoing in the field of cyanobacterial biotechnology and genetic tool development (for extensive review of cyanobacterial genetic tools, please refer to 49–51. These microbes offer well- established genetic toolkits and often do not face the technical limitations associated with

45 eukaryotic microalgae. To date, engineered cyanobacterial biocatalysts have been leveraged to commercialize algal ethanol production, and have shown promise for the production of a number of the above discussed bioproducts, including industrial platform chemicals, such as ethylene and terpenes 52,53. Lastly, we also acknowledge that a series of computational tools, including bioinformatic suites for genomic query, design of promoter and control elements, gene assembly, and expression analyses, have contributed to, and will play an ongoing, integral role in, the development of the next generation of advanced genetic tools. Integration of these technologies will undoubtedly drive exciting future advances in algal biocatalysis.

3.8 References

1. Guarnieri, M. T. & Pienkos, P. T. Algal omics: unlocking bioproduct diversity in algae cell factories. Photosynth. Res. 123, 255–263 (2015). 2. Rosenberg, J. N., Oyler, G. A., Wilkinson, L. & Betenbaugh, M. J. A green light for engineered algae: redirecting metabolism to fuel a biotechnology revolution. Curr. Opin. Biotechnol. 19, 430–436 (2008). 3. Specht, E., Miyake-Stoner, S. & Mayfield, S. Micro-algae come of age as a platform for recombinant protein production. Biotechnol. Lett. 32, 1373–1383 (2010). 4. Gregory, J. A., Topol, A. B., Doerner, D. Z. & Mayfield, S. Alga-produced cholera toxin- Pfs25 fusion proteins as oral vaccines. Appl. Environ. Microbiol. 79, 3917–25 (2013). 5. Cordero, B. F. et al. Enhancement of Lutein Production in Chlorella sorokiniana (Chorophyta) by Improvement of Culture Conditions and Random Mutagenesis. Mar. Drugs 9, 1607–1624 (2011). 6. Couso, I., Vila, M., Rodriguez, H., Vargas, M. A. & León, R. Overexpression of an exogenous phytoene synthase gene in the unicellular alga Chlamydomonas reinhardtii leads to an increase in the content of carotenoids. Biotechnol. Prog. 27, 54–60 (2011). 7. Guarnieri, M. T. et al. Examination of Triacylglycerol Biosynthetic Pathways via De Novo Transcriptomic and Proteomic Analyses in an Unsequenced Microalga. PLoS One 6, e25851 (2011). 8. Guarnieri, M. T., Nag, A., Yang, S. & Pienkos, P. T. Proteomic analysis of Chlorella vulgaris: Potential targets for enhanced lipid accumulation. J. Proteomics 93, 245–253 (2013). 9. Fan, J., Andre, C. & Xu, C. A chloroplast pathway for the de novo biosynthesis of triacylglycerol in Chlamydomonas reinhardtii. FEBS Lett. 585, 1985–1991 (2011). 46

10. Miller, R. et al. Changes in Transcript Abundance in Chlamydomonas reinhardtii following Nitrogen Deprivation Predict Diversion of Metabolism 1[W][OA]. Plant Physiol. 154, 1737–1752 (2010). 11. Msanne, J. et al. Metabolic and gene expression changes triggered by nitrogen deprivation in the photoautotrophically grown microalgae Chlamydomonas reinhardtii and Coccomyxa sp. C-169. Phytochemistry 75, 50–59 (2012). 12. Work, V. H. et al. Increased lipid accumulation in the Chlamydomonas reinhardtii sta7-10 starchless isoamylase mutant and increased carbohydrate synthesis in complemented strains. Eukaryot. Cell 9, 1251–61 (2010). 13. Dauvillée, D. et al. Biochemical characterization of wild-type and mutant isoamylases of Chlamydomonas reinhardtii supports a function of the multimeric enzyme organization in amylopectin maturation. Plant Physiol. 125, 1723–31 (2001). 14. Posewitz, M. C. et al. Hydrogen photoproduction is attenuated by disruption of an isoamylase gene in Chlamydomonas reinhardtii. Plant Cell 16, 2151–63 (2004). 15. Mussgnug, J. H. Genetic tools and techniques for Chlamydomonas reinhardtii. Appl. Microbiol. Biotechnol. 99, 5407–5418 (2015). 16. Jinkerson, R. E. & Jonikas, M. C. Molecular techniques to interrogate and edit the Chlamydomonas nuclear genome. Plant J. 82, 393–412 (2015). 17. Radakovits, R., Eduafo, P. M. & Posewitz, M. C. Genetic engineering of fatty acid chain length in Phaeodactylum tricornutum. Metab. Eng. 13, 89–95 (2010). 18. Hamilton, M. L., Haslam, R. P., Napier, J. A. & Sayanova, O. Metabolic engineering of Phaeodactylum tricornutum for the enhanced accumulation of omega-3 long chain polyunsaturated fatty acids. Metab. Eng. 22, 3–9 (2014). 19. Ifuku, K. et al. A stable and efficient nuclear transformation system for the diatom Chaetoceros gracilis. Photosynth. Res. 123, 203–211 (2015). 20. Sodeinde, O. A. & Kindle, K. L. Homologous recombination in the nuclear genome of Chlamydomonas reinhardtii. Proc. Natl. Acad. Sci. U. S. A. 90, 9199–203 (1993). 21. Lapidot, M., Raveh, D., Sivan, A., Arad, S. M. & Shapira, M. Stable Chloroplast Transformation of the Unicellular Red Alga Porphyridium Species. Plant Physiol. 129, 7– 12 (2002). 22. Kilian, O., Benemann, C. S. E., Niyogi, K. K. & Vick, B. High-efficiency homologous recombination in the oil-producing alga Nannochloropsis sp. Proc. Natl. Acad. Sci. 108, 21265–21269 (2011). 23. Lozano, J.-C. et al. Efficient gene targeting and removal of foreign DNA by homologous recombination in the picoeukaryote Ostreococcus. Plant J. 78, 1073–1083 (2014). 24. Georgianna, D. R. et al. Production of recombinant enzymes in the marine alga Dunaliella tertiolecta. Algal Res. 2, 2–9 (2013). 25. Remacle, C., Cardol, P., Coosemans, N., Gaisne, M. & Bonnefoy, N. High-efficiency

biolistic transformation of Chlamydomonas mitochondria can be used to insert mutations in complex I genes. Proc. Natl. Acad. Sci. 103, 4771–4776 (2006). 26. Hu, Z., Fan, Z., Zhao, Z., Chen, J. & Li, J. Stable Expression of Antibiotic-Resistant Gene ble from Streptoalloteichus hindustanus in the Mitochondria of Chlamydomonas reinhardtii. PLoS One 7, e35542 (2012). 27. Hu, Z. et al. Successful expression of heterologous egfp gene in the mitochondria of a photosynthetic eukaryote Chlamydomonas reinhardtii. Mitochondrion 11, 716–721 (2011). 28. Karas, B. J. et al. Designer diatom episomes delivered by bacterial conjugation. Nat. Commun. 6, 6925 (2015). 29. Kirk, M. M. & Kirk, D. L. Translational regulation of protein synthesis, in response to light, at a critical stage of Volvox development. Cell 41, 419–28 (1985). 30. Gillham, N. Translational Regulation of Gene Expression in Chloroplasts and Mitochondria. Annu. Rev. Genet. 28, 71–93 (1994). 31. Mayfield, S. P., Yohn, C. B., Cohen, A. & Danon, A. Regulation of Chloroplast Gene Expression. Annu. Rev. Plant Physiol. Plant Mol. Biol. 46, 147–166 (1995). 32. Rasala, B. A. et al. Expanding the spectral palette of fluorescent proteins for the green microalga Chlamydomonas reinhardtii. Plant J. 74, 545–556 (2013). 33. Rasala, B. A., Chao, S.-S., Pier, M., Barrera, D. J. & Mayfield, S. P. Enhanced Genetic Tools for Engineering Multigene Traits into Green Algae. PLoS One 9, e94028 (2014). 34. Rasala, B. A. et al. Robust Expression and Secretion of Xylanase1 in Chlamydomonas reinhardtii by Fusion to a Selection Gene and Processing with the FMDV 2A Peptide. PLoS One 7, e43349 (2012). 35. Iwai, M., Hori, K., Sasaki-Sekimoto, Y., Shimojima, M. & Ohta, H. Manipulation of oil synthesis in Nannochloropsis strain NIES-2145 with a phosphorus starvation–inducible promoter from Chlamydomonas reinhardtii. Front. Microbiol. 6, 912 (2015). 36. Radakovits, R., Jinkerson, R. E., Darzins, A. & Posewitz, M. C. Genetic engineering of algae for enhanced biofuel production. Eukaryot. Cell 9, 486–501 (2010). 37. Sizova, I., Greiner, A., Awasthi, M., Kateriya, S. & Hegemann, P. Nuclear gene targeting in Chlamydomonas using engineered zinc-finger nucleases. Plant J. 73, 873–882 (2013). 38. Carroll, D. Genome Engineering with Targetable Nucleases. Annu. Rev. Biochem. 83, 409–439 (2014). 39. Daboussi, F. et al. Genome engineering empowers the diatom Phaeodactylum tricornutum for biotechnology. Nat. Commun. 5, 3831 (2014). 40. Weyman, P. D. et al. Inactivation of Phaeodactylum tricornutum urease gene using transcription activator-like effector nuclease-based targeted mutagenesis. Plant Biotechnol. J. 13, 460–470 (2015).

41. Gao, H. et al. TALE activation of endogenous genes in Chlamydomonas reinhardtii. Algal Res. 5, 52–60 (2014). 42. Gao, H., Wang, Y., Fei, X., Wright, D. A. & Spalding, M. H. Expression activation and functional analysis of HLA3, a putative inorganic carbon transporter in Chlamydomonas reinhardtii. Plant J. 82, 1–11 (2015). 43. Mali, P., Esvelt, K. M. & Church, G. M. Cas9 as a versatile tool for engineering biology. Nat. Methods 10, 957–963 (2013). 44. Jiang, W., Brueggeman, A. J., Horken, K. M., Plucinak, T. M. & Weeks, D. P. Successful Transient Expression of Cas9 and Single Guide RNA Genes in Chlamydomonas reinhardtii. Eukaryot. Cell 13, 1465–1469 (2014). 45. Scholz, M. J. et al. Ultrastructure and composition of the Nannochloropsis gaditana cell wall. Eukaryot. Cell 13, 1450–64 (2014). 46. Gerken, H. G., Donohoe, B. & Knoshaug, E. P. Enzymatic cell wall degradation of Chlorella vulgaris and other microalgae for biofuels production. Planta 237, 239–253 (2013). 47. Cerutti, H., Johnson, A. M., Gillham, N. W. & Boynton, J. E. Epigenetic silencing of a foreign gene in nuclear transformants of Chlamydomonas. Plant Cell 9, 925–45 (1997). 48. Fischer, N. & Rochaix, J. D. The flanking regions of PsaD drive efficient gene expression in the nucleus of the green alga Chlamydomonas reinhardtii. Mol. Genet. Genomics 265, 888–94 (2001). 49. Koksharova, O. & Wolk, C. Genetic tools for cyanobacteria. Appl. Microbiol. Biotechnol. 58, 123–137 (2002). 50. Taton, A. et al. Broad-host-range vector system for synthetic biology and biotechnology in cyanobacteria. Nucleic Acids Res. 42, e136–e136 (2014). 51. Wang, B., Wang, J., Zhang, W. & Meldrum, D. R. Application of synthetic biology in cyanobacteria and algae. Front. Microbiol. 3, 344 (2012). 52. Ungerer, J. et al. Sustained photosynthetic conversion of CO2 to ethylene in recombinant cyanobacterium Synechocystis 6803. Energy Environ. Sci. 5, 8998 (2012). 53. Davies, F. K., Work, V. H., Beliaev, A. S. & Posewitz, M. C. Engineering Limonene and Bisabolene Production in Wild Type and a Glycogen-Deficient Mutant of Synechococcus sp. PCC 7002. Front. Bioeng. Biotechnol. 2, 21 (2014).

CHAPTER 4

DEVELOPMENT OF A HIGH-PRODUCTIVITY, HALOPHILIC, THERMOTOLERANT MICROALGA, PICOCHLORUM RENOVO.

Lukas R. Dahlina, Alida T. Gerritsenb, Calvin A. Henardc, Stefanie Van Wychenc, Jeffrey G. Lingerc, Yuliya Kunded, Blake T. Hovded, Shawn R. Starkenburgd, Matthew C. Posewitza and Michael T. Guarnieric

4.1 Abstract

Microalgae are promising biocatalysts for applications in sustainable fuel, food, and chemical production. Here, we describe culture collection screening, down-selection, and development of a high-productivity, halophilic, thermotolerant microalga, Picochlorum renovo.

This microalga displays a rapid growth rate and high diel biomass productivity (34 g m-2 day-1), with a composition well-suited for downstream processing. P. renovo exhibits broad salinity tolerance (growth at 107.5 g L-1 salinity) and thermotolerance (growth up to 40°C), beneficial traits for outdoor cultivation. We report complete genome sequencing and analysis, and genetic tool development suitable for expression of transgenes inserted into the nuclear or chloroplast genomes. We further evaluate mechanisms of halotolerance via comparative transcriptomics, identifying novel genes differentially regulated in response to high salinity cultivation. These findings will enable basic science inquiries into control mechanisms governing Picochlorum biology and lay the foundation for development of a microalga with industrially relevant traits as a model photobiology platform. a Department of Chemistry, Colorado School of Mines, Golden, CO 80401 b Computational Science Center, National Renewable Energy Laboratory, Golden, CO 80401 c National Bioenergy Center, National Renewable Energy Laboratory, Golden, CO 80401 d Los Alamos National Laboratory, Los Alamos, NM 87545

4.2 Introduction

Microalgae are a source of renewable biomass and promising photosynthetic biocatalysts for the sustainable production of fuel and chemical intermediates 1. Importantly, they are also valuable model systems for fundamental investigation of mechanistic photobiology 2. These microbes possess a series of unique characteristics that make them well-suited for biotechnological applications, including year-round cultivation capacity in saline water on non- arable land, higher potential biofuel yields than terrestrial crops, and the ability to utilize CO2 as a sole carbon source 3,4. Rising greenhouse gas emissions from anthropogenic sources has led to a resurgent interest in exploiting these organisms for concurrent CO2 capture and renewable biocommodity production 5,6. However, at present, current model algal systems are not suitable for outdoor deployment, displaying low productivity under relevant environmental conditions

(e.g. high light intensity, high temperature, seawater environments) 7. Further, top candidate deployment systems display low genetic throughput, often requiring weeks-to-months to generate and verify transgenic lines, which hinders fundamental mechanistic inquiry and metabolic engineering strategies in deployment-relevant microalgae 7,8.

Since its first classification in 2004, the genus Picochlorum has been recognized for its distinct characteristics of broad thermotolerance, salinity tolerance, compact genome architecture, fast doubling time, and resilience to high light intensity 6,9–13. An alga of the genus

Picochlorum was recently shown to have the highest biomass productivity in a comparative analysis between a series of industrially relevant microalgae, underscoring this genera’s deployment potential 6. However, to date, there are limited insights into Picochlorum halotolerance, biosynthetic capacity, biomass characterization, and genetic tractability, hindering its development as a fundamental platform and for biotechnological applications.

Here, we report the down-selection, characterization, and development of a novel alga of the genus Picochlorum, Picochlorum renovo sp. nov. We characterized the diel biomass productivity (34 g m-2 day-1) of this alga under simulated outdoor cultivation conditions, quantifying the protein, carbohydrate, and lipid content (20%, 60% and 10% ash-free dry cell weight, respectively), thermotolerance (growth capacity up to 40°C), and salinity tolerance

(growth at 107.5 g L-1 salinity). Furthermore, we generated nuclear, chloroplast, and mitochondrial genome sequences and report comparative transcriptomic analyses under low- and high-salt conditions, enabling high-resolution genome annotation and providing novel insight into the mechanisms of halotolerance. Lastly, we developed a set of facile genetic tools that enable expression of multiple transgenes inserted into either the nuclear or chloroplast genomes.

Combined, these data will enable fundamental insights into Picochlorum photobiology and inform targeted genetic engineering strategies to accelerate microalgal biotechnological applications in a deployment-relevant, emerging model microalga.

4.3 Results

4.3.1 Down-Selection, Physiology, and Compositional Analysis

We screened a >300-strain algal culture collection, in order to identify halotolerant strains14,15. Over 100 unique halotolerant isolates were down-selected and screened under simulated (diurnal light and temperature cycling) summer growth conditions using a custom built photobioreactor, described in Dahlin et al. 15. We identified one isolate that exhibited a noticeably faster growth rate relative to other isolates, including control strains Nannochloropsis oceanica (KA32) and Nannochloropsis salina (CCMP 1776), two top-candidate strains currently under evaluation for outdoor deployment (Figure 4.1) 15–17.

Figure 4.1 Representative culture collection growth screening data. The rapid growth and high optical density phenotype of P.

renovo is highlighted in black. Nannochloropsis salina CCMP 1776 is bolded in red for reference.

This rapidly growing strain was down-selected for further analysis and development.

Under batch growth this microalga displayed a diel biomass productivity of 34.3 g m-2 day-1, from hour 6 to 30, representative of one day and night of high productivity growth (Figure 4.3a).

Dark period biomass loss was 0.25 g m-2 hr-1 during the first 11-hr dark period and 0.46 g m-2 hr-

1 in the second. Cell division occurs during both light and dark periods when grown under a diel cycle. Cessation of cell division and biomass accumulation occurs simultaneously. Nitrogen supplementation during stationary phase led to growth reinitiation (Figure 4.2a). We observed peak growth at 35°C under continuous illumination, with growth capacity up to 40°C (Figure

4.2b). Biomass composition varies as a function of growth phase, with fluctuations in carbohydrate and protein content observed throughout diel cycles (Figure 4.3b). Notably there is

Figure 4.2 Physiological characterizations under varied nitrogen, temperature, and salinity regimes. (a) Representative growth curves of P. renovo with and without addition of sodium nitrate to a final concentration of 4.5mM at hour 54, following entry into stationary growth phase. (b) P. renovo growth rate as a function of temperature. Average and standard deviation are from n=3 biological replicates. (c) Growth curve comparison of P. renovo at 8.75 and 35 g/L salinity. Average and standard deviation are from n=2 biological replicates. (d) Endpoint biomass titer following 6 days of growth at varying salinities for P. renovo. Average and standard deviation are from n=2 replicates are reported.

a substantial decrease in glucose (derived from biomass hydrolysis) following inoculation into fresh media, with glucose declining from 52% to 1.4% of AFDW (ash-free dry weight) (Figure

4.3c). Lipid content, as measured by fatty acid methyl esters, varied from 8.5 – 16.2%, with

C16:0, C16:3, C18:1n9, C18:2n6, and C18:3n3 representing major lipid fractions (Figure 4.3d).

30 hours post-inoculation the cells enter stationary phase and have an ash-free composition of

10% FAME (fatty acid methyl ester), 20% protein, 59.5% carbohydrates (measured as hydrolyzed monomeric sugars), and 10.5% unidentified biomass components (Figure 4.3b).

4.3.2 Genomic Analysis and Speciation

We conducted phylogenetic analysis of the isolate’s 18S rRNA, identifying high similarity (>99%) to numerous Picochlorum species, providing initial evidence for taxonomic classification. PacBio genome sequencing generated an assembled nuclear genome containing 29 contigs with 14.4 Mbps and 46.2% GC, similar to previously reported Picochlorum genomes

12,13. 8,902 protein coding sequences were putatively annotated, with an average of 2.2 exons /

1.2 introns per gene. The nuclear genome contains a homolog of the universally conserved meiosis associated gene, spo11-2 (E-value: 2E-22), and homologs of multiple meiotic and/or

55 homologous recombination associated genes (with reported E-values), including four rad51 homologs (2E-107 to 4E-26), dmc1 (2E-121), pol2A (0), rfc1 (4E-34), polD1 (0), mre11 (2E-77), rad50 (0), rad54 (4E-134), mus81 (6E-15), msh4 (2E-10), msh5 (2E-50), rpa1 (3E-69), rpa2

(8E-25), and rpa3 (5E-16) 18. Further evidence of genes putatively involved in meiosis is provided by the identification of oda2 (0) and bug22 (1E-72) homologs, which are flagella associated genes, implicated in gamete pairing prior to mating 19–23. We also identified a putative chlorophyllide-a oxygenase (0), necessary for chlorophyll b production, and cell division was observed to occur by autosporulation (Figure 4.4), providing additional evidence for classification of this strain as a Picochlorum 11. When compared to other available Picochlorum genomes, the novel Picochlorum isolate displayed 87-94% whole genome sequence identity

(Table 4.1). These data support that the isolate is a novel Picochlorum species, henceforth termed Picochlorum renovo sp. nov. 24.

Chloroplast and mitochondria genome maps are presented in Figure 4.5. The 74 kb chloroplast genome displayed a non-canonical chloroplast genome architecture lacking an inverted repeat region, as noted by Krasovec et al. for the genus Picochlorum 12. The 36 kb mitochondrial genome displayed a compact coding architecture, representing the highest mitochondrial coding density reported to date (1.05 CDS per kb) for the class Trebouxiophyceae, in line with the previously reported mitochondrial genome of Picochlorum costavermella 12.

Notably, genes encoding a homing endonuclease and protein of unknown function split the mitochondrial 23s rRNA, contributing, in part, to the higher coding density (Figure 4.5).

Table 4.1 Whole genome alignment analysis of Picochlorum spp.

Figure 4.4 Division of P. renovo highlighting autosporulation and mother cell wall (arrow), a defining trait of the Picochlorum genus. Red coloring indicates chlorophyll autofluorescence.

Figure 4.5 Chloroplast and mitochondria genome maps of P. renovo.

4.3.3 Transcriptome Response to Salinity

We observed broad halotolerance in P. renovo, as reported previously for this genus 10,11, with cultivation capacity in minimal media salinity concentrations ranging from 8.75-107.5 g L-1 sea salts (Figure 4.2c,d). To better understand the genes involved in the salinity response, cultures were grown in 8.75 and 35 g L-1 salinity seawater and assessed via comparative transcriptomics. RNA from triplicate mid-log phase cultures was sequenced, and subsequent differential expression analysis was performed. 3,464 genes were differentially-expressed at 35 g

L-1 salinity (1934 down, 1530 up) at statistically significant values (q<0.05), representing 39 percent of total coding sequences. Gene ontology semantic analyses were used to deconvolute the large number of differentially-expressed transcripts, implicating a subset of processes

59 involved in the high-salt response, including previously reported genes governing proline metabolism (Figure 4.6) 10. A series of previously unreported haloresponsive genes were also observed, including ppsA (E-value: 2E-54), ppsC (E-value: 2E-60), pks1 (E-value: 4E-30) and pks15 (E-value: 1E-28) (polyketide synthases), iput1 (inositol phosphorylceramide glucuronosyltransferase, E-value: 9E-75), cerk (ceramide kinase), rad54 (DNA repair and recombination protein, E-value: 2E-20), and dmc1 (disrupted meiotic cDNA 1, E-value 2E-121), discussed further below.

4.3.4 Nuclear and Chloroplast Engineering

We randomly integrated a linear PCR amplicon comprised of native promoter and terminator elements into the nuclear genome of P. renovo, directing transcription of 2A peptide- linked bleomycin resistance gene and the fluorescent reporter mcherry (Figure 4.7a) via electroporation. mCherry was chosen as a reporter gene for nuclear expression based on prior reports of high signal to noise ratios in microalgae 25. Per transformation, an average of 41 colonies were obtained, representing transformation efficiencies of 14 colonies per µg of DNA, and 9×10-8 colonies per electroporated cell. 75% (9/12) of PCR screened transformants contained the entire transgene construct, while the remaining contained truncated versions (Figure 4.7b).

Transgene integration was also achieved via biolistics, however we observed approximately an order of magnitude lower transformation efficiency relative to electroporation. Positive transformants showed a 2-to-4- fold increase in mCherry fluorescence over wild type (Figure

4.7c), and confocal microscopy confirmed mCherry fluorescence localized to the nucleus and cytoplasm in these cells (Figure 4.7d). Additional promoter configurations, with and without their respective introns, were also evaluated, including elongation factor 1-alpha 2 and photosystem I reaction center subunit II (psaD), both utilizing the eef1A2 (elongation factor 1-

60 alpha 2) terminator, which displayed comparable transformation efficiencies and mCherry fluorescence. P. renovo is also sensitive to G418, which we have successfully used as a selection agent.

Figure 4.8a depicts the construct utilized for targeted engineering of the P. renovo chloroplast via biolistics. The native 16s ribosomal RNA promoter and 3’ UTR were utilized to direct transgene expression. The commonly utilized antibiotics spectinomycin and streptomycin failed to inhibit P. renovo growth, and the above utilized phleomycin (for nuclear transformation) was ineffective for isolation of viable chloroplast transformants. Therefore, we

chose erythromycin for selection, following antibiotic sensitivity screening. Notably, there is

100% homology between the last 9 base pairs (anti-Shine-Dalgarno sequence) of the P. renovo

16S rRNA and the E. coli 16S rRNA 26. Thus, a canonical E. coli ribosomal binding site (RBS,

AGGAGGTTATAAAAA) was used to direct translation. The erythromycin resistance gene

(ereB) was linked to the reporter super folder green fluorescent protein (sfGFP) in an operon 27 for rapid identification of transgenic lines. When fully constructed with targeting homology arms, this plasmid readily yielded transformed microalgae using a conventional biolistic approach 28–30. Transformants could be rapidly identified via the reporter gene by imaging of the

62 bombarded plate in a gel imaging station with filter sets suitable for sfGFP detection. This procedure yielded an average (n=3) of a single colony per transformation with efficiencies of 1.4 colonies per µg delivered DNA and 8×10-9 colonies per microalgal cell. Colonies positive for sfGFP were passaged on selective media and proper integration of the construct into the target region was verified via PCR using primers binding outside the homology region and within the transgene operon, depicted in Figure 4.8(a,b). 38-48-fold greater sfGFP fluorescence was observed over wild type when measured via fluorometry (Figure 4.8c). Epifluorescent microscopy showed sfGFP fluorescence successfully localized to the chloroplast (Figure 4.8d).

4.4 Discussion

P. renovo displayed a distinct rapid growth rate phenotype, in initial screening trials comparing over 100 unique isolates (Figure 1). Peak growth rate at 35°C (Figure 4.2b) and cultivation capacity up to ~3X seawater salinity (Figure 4.2d) indicates this strain is well-suited for outdoor cultivation in high temperature regions with saltwater access. These traits complement those of previously identified winter candidate deployment strains 15,31, laying the foundation for crop rotation strategies 17,32. Given the potential discordance between optical density and biomass density we further characterized P. renovo’s biomass productivity; the diel biomass productivity reported here exceeds the target productivity of 25 g/m-2/day-1 reported by

Davis, et al. 33 for cost-competitive algal biofuels. Higher biomass productivities are likely achievable, given the suboptimal growth temperatures used in this study, which simulated outdoor cultivation in Mesa, Arizona. Future studies will evaluate outdoor productivity metrics to assess translatability of indoor metrics to outdoor systems in geographic regions better suited for high temperature cultivation.

Biomass analysis indicates the primary biomass hydrolysis product in P. renovo is glucose, which is a favorable feedstock for downstream biotechnological applications 34. A drastic depletion of glucose was observed following inoculation into fresh media, similar to outdoor cultivation trends observed in other microalgal genera 15 (Figure 4.3c). Dark period biomass loss is characterized almost exclusively by a decrease in glucose. These phenomena putatively function as a mechanism to remobilize glucose as an energy source for cell division and cellular homeostasis 35. Thus, as previously reported 36, dark period biomass loss is an important parameter to consider when cultivating microalgae for biomass production. Under the conditions tested here, dark period biomass loss ranged from 0.25 to 0.46 g m-2 hr-1. Combined, these data highlight the potential advantage of harvesting P. renovo biomass prior to dark period losses to maximize biomass and storage carbon yields.

Interestingly, cell division occurs during both the light and dark periods when grown under a diel cycle (Figure 4.3a). This is contrary to many microalgae that synchronize cell division to occur at night; such is the case for the genera Chlamydomonas 37, Nannochloropsis 38,

Chlorella, and Scenedesmus 39. This continuous diurnal and nocturnal cell division, coupled with the compact genome(s), may partially explain the rapid doubling time and high biomass productivity of P. renovo, and represents an area warranting further research. Cell division and biomass accumulation cease concurrently, suggesting a non-photosynthetic state when nitrogen- deprived (Figure 4.3a). This is notable as some microalgae will continue to accumulate biomass post-nitrogen deprivation, presenting another avenue for comparative analyses 40. Importantly, addition of nitrogen following growth arrest resulted in reinitiated growth, implicating nitrogen deprivation as the key driver for entry into stationary phase under the conditions evaluated in this

64 study (Figure 4.2a). Optimization of nitrogen levels and harvest point may lead to enhanced productivity and storage carbon content.

Comparative transcriptomic analyses identified a series of previously unreported, haloresponsive genes. dmc1, which is involved in homologous chromosome pairing during meiosis was one of the most highly upregulated transcripts at higher salinity 41. rad54, encoding a putative DMC1-interacting protein known to function during homologous recombination, is concurrently upregulated 42,43. The upregulation of these genes could be attributed to meiosis, or homologous recombination repair of double strand DNA breaks, due to increased double strand breaks at higher salinities 44,45. The observation of differentially-expressed genes putatively associated with meiosis and homologous recombination suggests P. renovo may participate in sexual mating, and is capable of DNA repair via nuclear homologous recombination, both powerful tools for genetic manipulation. Indeed, sexual mating has been leveraged for trait stacking in both microalgae and plants and presents a powerful approach for rapid development of production hosts. Though we acknowledge that gene homology is insufficient evidence to assert functionality, as reviewed by Fučíková et al., multiple morphological/cytological observations of syngamy and/or meiosis have been reported in the class Trebouxiophyceae and high conservation of meiotic genes is found within this class 46–48.

Downregulation of genes encoding proteins putatively relating to lipid remodeling was observed under high salt conditions, including pks1, pks15, ppsA, ppsC, iput1, and cerk. ppsA, ppsC, pks1, and pks15 are involved in the synthesis of diverse lipids and polyketides which have been implicated in cell wall permeability 49. cerk is an enzyme that transfers a phosphate group to ceramide and is potentially acting in coordination with iput1 which transfers a glucuronic acid moiety to glycosyl inositol phosphorylceramides. Ceramides provide the lipid backbone for plant

65 sphingolipids, and are primarily believed to be structural components of cellular membranes; however, ceramides have also been suggested to play a role in plant signaling 50. The above data suggests that P. renovo is potentially using lipid remodeling to tune membrane permeability at differing salinities.

To facilitate P. renovo genetic and metabolic engineering, we developed tools enabling transgene expression in both the nucleus and chloroplast. Interestingly, only 9 of the 12 nuclear transgenic isolates screened showed insertion of the full transgene construct. Of the remaining 3 isolates, 2 were shown to have a truncated promoter or terminator, and one was shown to have an incomplete mCherry coding sequence, observed by the inability to generate a full-length coding sequence PCR product (Figure 4.7b). It is not clear whether these truncated transgene constructs are the result of native P. renovo machinery cleaving the transgene construct or an incomplete

PCR product integrating into the genome. Fluorescent plate reader analysis of the clones revealed increases in mCherry fluorescence over wild type for the 11 clones containing a full length mCherry coding sequence (Figure 4.7b,c). As expected, the single clone without a full length mCherry coding sequence did not show an increase in mCherry fluorescence relative to wild type. The variation in mCherry fluorescence could be due to unique integration sites, or multiple integration events. mCherry fluorescence was primarily localized to the nucleus and cytoplasm of the transformant, with no observable chloroplast localization, as has been previously observed 51. Preliminary analyses indicate stable nuclear transformation; mCherry fluorescence intensity remains constant following passaging on and off the selection marker.

Successful chloroplast transformation was phenotypically observed via high reporter expression, and epifluorescent microscopy confirmed successful localization of the sfGFP to the

P. renovo chloroplast, evident by overlap with chlorophyll autofluorescence (Figure 4.8c,d). The

66 ability to confirm transgenic colonies via direct imaging of the high sfGFP signal will increase the throughput of control element screening, such as varied promoter strengths, in order to optimize metabolic engineering strategies. Additionally, the successful utilization of an E. coli

RBS for operonic expression presents the potential for optimization of mRNA translation via the employ of established RBS prediction software 26. Thus, these tools will be useful for biotechnological applications, such as overexpression of desired industrial enzymes 28,52 or fine- tuned regulation of native and/or synthetic metabolic pathways for bioproduct formation 53.

The transformation procedure presented herein is a facile protocol with relatively rapid turnaround time that can be completed in a few hours. Given the fast growth of this alga, transformant colonies can be generated in as few as 5 days, considerably faster than top- candidate deployment strains such as Nannochloropsis, wherein colonies need ~21 days of growth before verification analyses can be performed 7. We have also provided the sequences of two additional nuclear promoters (elongation factor 1-alpha 2 and photosystem I reaction center subunit II), which we have successfully utilized to generate transformants. These additional promoters could prove useful for expression of multiple transgenes from one nuclear targeting cassette.

4.5 Conclusions

The full biotechnological potential of microalgae has yet to be brought to bear at commercial scale, in part due to the lack of robust, high-productivity strains suitable for outdoor deployment. Further, algal genetics in non-model systems has proven to be a limiting factor in strain development and fundamental mechanistic probing of top-candidate deployment strains.

Here, we characterize a novel high-productivity, halophilic, thermotolerant microalga, and report

67 the development of genomic and genetic tools therein. In addition to possessing a series of traits suitable for outdoor deployment, this strain displays favorable characteristics for development as a model system, in part due to its compact genomic architecture and rapid genetic throughput.

Combined, the above-described traits present a unique complement to extant model systems.

Further genetic development of P. renovo will enable both fundamental and applied insights, including elucidation of key regulatory mechanisms governing rapid growth and halotolerance in microalgae, as well as strain engineering strategies targeting enhanced productivity and carbon partitioning in a deployment-relevant microalga.

4.6 Methods

4.6.1 Strain Screening and Characterization of Algal Growth

Microalgae were screened as previously reported 15, under conditions representative of summer cultivation. Briefly, 100 mL microalgal cultures were sparged with 2% CO2 at 100 mL min-1. Temperature cycled from 21 to 32 °C while lighting cycled from 0 to 965 µmol m-2 s-1 (the maximum output of the utilized lights). This regime was designed to simulate the temperature and lighting diel cycles measured in outdoor raceway ponds located at the Arizona Center for

Algae Technology and Innovation testbed site located in Mesa Arizona, during the time frame of

June 12th to July 21st, 2014. We utilized a modified f/2 medium for cultivation, termed NREL

Minimal Medium (NM2), in seawater (Gulf of Maine, Bigelow Laboratory). The following were added to the indicated final concentrations followed by addition of 12 M HCl to attain pH 8.0:

-3 -3 -4 NH4Cl (5.0 x 10 M), NaH2PO4∙H2O (0.313 x 10 M), Na2SiO3∙9H2O (1.06 x 10 M),

-5 -5 -8 FeCl3∙6H2O (1.17 x 10 M), Na2EDTA∙2H2O (1.17 x 10 M), CuSO4∙5H2O (3.93 x 10 M),

-8 -8 -8 Na2MoO4∙2H2O (2.60 x 10 M), ZnSO4∙7H2O (7.65 x 10 M), CoCl2∙6H2O (4.20 x 10 M),

-7 -7 -9 MnCl2∙4H2O, (9.10 x 10 M), thiamine HCl (2.96 x 10 M), biotin (2.05 x 10 M), cyanocobalamin (3.69 x 10-10 M), Tris base (24.76 x 10-3 M). For genetic engineering, the concentration of seawater was diluted 4-fold with Milli-Q water (Millipore Corporation), ammonium bicarbonate was utilized in the place of ammonium chloride, and 1.5x vitamins

(thiamine HCl, biotin, cyanocobalamin) were utilized. Agar (Bacto) plates were prepared by autoclaving 3% agar in Milli-Q water, followed by addition of an equal volume of sterile filtered

NM2 (seawater diluted 2-fold) with 2x nutrients, trace metals, vitamins, and Tris buffer. Sterile filtered selection antibiotic was added as necessary to appropriate concentrations, defined below.

To obtain a more detailed analysis of P. renovo growth, the above conditions were utilized with a 120 mL culture volume. Mid-log phase seed culture was generated under the above diel conditions, and used to inoculate 36, 120 mL cultures at a starting optical density of

1.0, in biological triplicate. Inoculation occurred approximately halfway through the light cycle, and biomass samples were harvested at the initiation, mid-point, and end-point of the lighting cycle, as indicated in Figure 4.3. Sterile water was added prior to samplings to account for evaporative loses. Cell counts were performed using an Improved Neubauer hemocytometer. To convert volumetric productivities to areal values, the cross-sectional area of the culture tubes

(0.00459 m2) was employed.

Growth at varying salinities for Figure 4.2d were done in the same fashion as culture collection screening except salt levels were varied by addition of sea salts (Sigma S9883). 17.5 g

L-1 salinity was achieved via addition of seawater to milli-Q water and higher salinities utilized addition of sea salts. 100 mL of culture was harvested after 6 days of growth, utilizing the temperature and light cycling from the culture collection screening methods described above.

Temperature optima data, represented in Figure 4.2b was generated by growing strains in NM2,

-2 -1 with culture conditions of constant 400 µmol m s lighting, 2% constant CO2 sparging, and 100 mL volume. To determine growth rates, optical density (750 nm) measurements were taken daily

Figure 4.2b.

4.6.2 Compositional Analysis

Compositional analysis was carried out as reported previously 15, with the following modification; a Carbopac PA1 HPLC column was utilized for sugar monomer (carbohydrate) analysis. Protein was quantified via CHN (carbon, hydrogen, and nitrogen) analysis, utilizing an

Elementar VarioEL cube CHN analyzer according to the manufacture’s specifications. Briefly, a

5 mg sample is combusted at 950°C, and subsequent gasses are carried via helium to reduction and adsorption tubes utilizing an intake pressure of 1200 psi and ultimately detected with a thermal conductivity detector. A nitrogen-to-protein conversion factor of 4.78 was used 54.

4.6.3 Genome Sequencing, Assembly and Annotation

High molecular weight algal genomic DNA was extracted from cells imbedded in agarose, purified and concentrated using AMPure PB beads. The DNA was then fragmented using Covaris g-Tubes. Fragmented and purified DNA was processed for 20 kb SMRT bell library prep. The long insert libraries were size selected using a Blue Pippin instrument (Sage

Sciences, Beverly, MA). The sequencing primer was annealed to the selected SMRT bell templates. The libraries were bound to DNA polymerase and loaded on the PacBio RSII for sequencing. Sequencing was completed using either C2/P4 or C3/P5 chemistry and 3-h movies.

8 SMRT cells of sequencing data were assembled with FALCON, version 0.2.2 55. The final assembly includes 29 contigs with an assembled genome size of 14.4 Mbp. Estimated fold coverage of the PacBio reads was 270X.

Genome annotation was performed using the BRAKER (v2) training and annotation pipeline 56 utilizing the 6 sets of transcriptomic reads (described below in “Transcriptome

Response to Salinity”) to inform AUGUSTUS gene models 57,58. Functional annotation of the

8,902 genes was performed by InterProScan 5 59 and BLASTp searches against the UniProt 60 protein blast database; reported E-values reflect this methodology. The P. renovo genome assembly and annotation is available for download at the Greenhouse Knowledgebase

(greenhouse.lanl.gov).

4.6.4 Transcriptome Response to Salinity

In order to identify genes putatively conferring halotolerance, cells were cultivated under low- and high-salinity conditions, corresponding to 8.75 g L-1 and 35 g L-1 sea salts. Cells grow at approximately the same growth rate under these conditions (Figure 4.2c). Cells adapted to the appropriate salinity level were grown in NM2 medium, utilizing ammonium bicarbonate as a nitrogen source. Biological triplicate culture conditions were as follows: 33°C, 400 µmol m-2 s-1 lighting, and 2% constant CO2 sparging in 100 mL volume. Seawater was employed as a source of salt, as this provides a more accurate proxy for halo-responsiveness compared to NaCl 61.

Seawater was diluted with distilled water to obtain appropriate salinity levels. The data from these methods are reflected in Figure 4.2c and salinity transcriptomics data. The above methods were done with the explicit goal of reducing culture shock, and subsequent global stress response, thus allowing a steady state comparison of RNA transcripts relating to salinity tolerance. RNA was obtained utilizing a QIAGEN RNeasy Plant Mini Kit following the manufacturer’s recommendations, cells were homogenized under liquid nitrogen using a mortar and pestle. Paired-end 150bp Illumina read RNA seq data were received from Genewiz in the form of compressed fastq files. Samples were comprised of 2 conditions and 3 biological

71 replicates of each condition, resulting in 6 total samples. Raw fastq reads were quality trimmed using HTStream 62 and mapped via Salmon 63 to the available genomic assembly. Coding regions were extracted from the full reference assembly prior to mapping. Read counts were formatted into a tab-separated file and migrated to R 64 to perform differential expression using the edgeR

65 package. Low-level transcripts were filtered and removed, and all libraries were normalized to each other. Transcript counts were fit to a generalized linear model and the Cox-Reid profile- adjusted likelihood method was used to estimate the dispersion of each transcript. Differential expression was performed by a quasi-likelihood test between each condition. Transcripts were determined as differentially expressed when the corrected p-value (also known as q-value, or

False Discovery Rate) was less than or equal to 0.05 after a Benjamini-Hochberg correction for multiple hypothesis testing.

Whole genome alignments to other publicly available Picochlorum genomes were done as follows: 6 assemblies of different strains of Picochlorum sp. were compared using the nucmer utility in the large-scale alignment program MUMmer 12,13,66,67. Maximal matches were found and total bases matching between the samples were summed and the percent identity was reported as the average identity among the maximal unique matches.

Gene ontology analysis was performed as follows: differentially expressed genes were assigned putative functions by extracting the FASTA sequence from the original list of genes and aligning the sequence against the available Chlamydomonas reinhardtii annotated assembly

(version 5.5) via BLAST 68. Protein identification numbers and putative annotations were then uploaded to the UniProt 60 database and cross-referenced against the available gene ontology

(GO) terms. GO terms were visualized on a semantic space scatterplot with the online software

Revigo 69.

4.6.5 Nuclear Engineering

A nuclear integration cassette, as depicted in Figure 4.7a, was synthesized and subcloned into the pUC19 plasmid backbone by Genewiz, Inc (South Plainfield, NJ). The selection marker,

2A peptide and mCherry were codon optimized to the P. renovo genome. The final linear PCR product (from primers LRD 49 and 11) for transformation was generated utilizing Q5 2x hot start master mix (NEB) and purified with a PureLink Quick PCR Purification kit (Invitrogen) following the manufacture’s protocol, modified to include a second wash step.

10 OD units (~475 x 106 cells) of early log phase cells per transformation were harvested and washed 3 times at room temperature in 375 mM D-Sorbitol (Sigma S6021). Washing utilized 2 mL Eppendorf tubes, 950 µL of 375 mM D-Sorbitol per wash, centrifuged at 8000 g for 1 min. After washing, cells were resuspended in 100 µL of 375 mM D-Sorbitol; 3 µg of DNA at 850 ng per µL (concentrated on a vacuum centrifuge) was added to the cells and gently mixed.

Cells and DNA were incubated for 3 minutes, transferred to an ice cold 2 mm gap electroporation cuvette (Bulldog Bio) and electroporated with a Gene Pulser Xcell (Bio-Rad) electroporator utilizing a set time constant and voltage protocol of 2200 volts with a 25 ms time constant. Immediately following the pulse, cells were transferred to 400 µL of media supernatant

(from the above utilized cells) and incubated at room temperature for 15 min. Cells were then split equally between 3 selection plates (1.5 % agarose) comprised of NM2 supplemented with

20 µg per mL of phleomycin (InvivoGen). Plates were placed in a Percival incubator with

-2 -1 fluorescent lighting at 33°C, 150 µmol m s , and 1.5% CO2. Colonies were counted and passaged on selection after 5 days for further analysis.

4.6.6 Chloroplast Engineering

Homology arm sequences were PCR amplified from chloroplast genomic DNA using

NEB Q5 Master Mix from New England Biolabs (Ipswich, MA). A promoter-RBS-sfGFP-RBS- ereB-3’ UTR cassette was synthesized by Genewiz, Inc (South Plainfield, NJ) as depicted Figure

4.8a. The chloroplast integration cassette was assembled into a pUC19 backbone using 2X

Gibson Assembly Mix from New England Biolabs, following the manufacturer’s protocol.

Complete vector sequences were confirmed by Sanger sequence analysis (Genewiz, South

Plainfield, NJ).

Biolistic transformation was employed to deliver DNA into the chloroplast, as reported previously 28,70. 10 µg of plasmid DNA (QIAprep spin miniprep kit QIAGEN) was precipitated onto 550 nm gold sphere nanoparticles (Seashell Inc.) under constant vortexing. 10 µL of plasmid DNA (1 µg per µL) was added to 60 µL of gold particles (50 mg per mL), followed by dropwise addition of 50 µL of 2.5 M CaCl2 and 20 µL of 0.1 M spermidine (Sigma S0266-1G).

This was vortexed for 5 min, incubated for 1 min at room temperature, briefly centrifuged, and washed with 140 µL of isopropanol. Following removal of wash supernatant, the gold particles were resuspended in 60 µL of isopropanol and gently sonicated in a bath sonicator to resuspend the pellet. To assay loading efficiency of the DNA onto the gold, a 9 µL aliquot was taken, washed in 9 µL of water and assayed for DNA concentration utilizing a NanoDrop 2000 spectrophotometer.

To transform P. renovo, an overnight culture was grown to early log phase in NM2, concentrated to 2.5 OD units in 170 µL, and spread evenly onto a 100 x 15 mm NM2 agar plate supplemented with 800 µg per mL erythromycin (Sigma E5389-5G). A Biolistic PDS-1000/He

Particle Delivery System (metal case version) (Bio-Rad) was used for bombardment, which was accomplished by drying 9 µL of the above DNA loaded gold particles onto the macrocarrier

(fast, low humidity drying was accomplished by placing the loaded macrocarrier into the bombardment chamber and pulling vacuum 71), and bombarding cells 5 cm below the macrocarrier with a 1100 psi rupture disk. After bombardment, plates were placed into the same growth conditions described above. Following 7 days of growth, the plates were imaged with a

FluorChemQ gel imaging station (Protein Simple) with 475/35 and 573/35 nm respective excitation and emission filters, which allowed direct imaging of sfGFP positive colonies.

To assess construct integration into the genomes of P. renovo, genomic DNA was extracted from cells passaged on the selection marker utilizing a MasterPureTM Yeast DNA

Purification kit (Lucigen). PCR was performed utilizing Q5 Hot Start High-Fidelity polymerase

(New England Biolabs) according to the manufacturer’s recommendations.

4.6.7 Fluorescent Plate Reader Analysis

Colonies were restreaked onto fresh agar plates supplemented with the appropriate selection marker (phleomycin 20 µg per mL and erythromycin 800 µg per mL for the nucleus and chloroplast, respectively) and grown in triplicate in 3 mL of growth media (no selection marker) in standard glass cell culture tubes mixed daily via vortexing. Cultures were grown in

-2 -1 the above described Percival incubator conditions (33°C, 150 µmol m s , 1.5% CO2). Early log phase cells were analyzed for mCherry and sfGFP fluorescence utilizing 200 µL of cell culture in a black 96 well plate and a FLUOstar Omega plate reader v. 5.11 R3 (BMG Labtech). To quantify mCherry a 584 nm excitation filter and 620/10 nm emission filter were utilized with gain set to 2500; to quantify sfGFP a 485/12 nm excitation and 520 nm emission filter set was

75 used with gain set to 1200. Data was normalized to chlorophyll content, which was determined by using a 485/12 nm excitation and 680/10 emission filters, with gain set to 1500. Data is represented as a fold increase over the wild type alga.

4.6.8 Statistics and Reproducibility

All experiments in this study utilized triplicate biological replicates. Error is represented by the sample standard deviation of the replicates, unless otherwise noted.

4.6.9 Microscopy

Mid-log phase chloroplast sfGFP transformants and wild type were imaged with a Nikon

Eclipse 80i microscope, equipped with a Nikon Intensilight C-HGFI mercury lamp light source, a Nikon Plan Apo VC 100x objective lens, and a Nikon DS-QiMc camera. NIS-Elements BR

4.30.01 software was utilized for imaging chlorophyll and sfGFP (31017 – Chlorophyll

Bandpass Emission and 41017 – Endow GFP/EGFP Bandpass, both from CHORMA®). Imaging of wild type and transgenic lines employed equivalent exposure time and gain settings. ImageJ was used for post imaging analysis.

Nuclear mCherry transformants and wild type were imaged with a Nikon C1si confocal microscope, equipped with EZ-C1 3.60 software. Chlorophyll was imaged with a 650 LP filter. mCherry was imaged with a 590/50 filter. Both chlorophyll and mCherry were excited with a

561.4 nm laser. Laser intensity, pin hole size, pixel dwell time, and gain were set using an mCherry clone. Equivalent settings were utilized for imaging wild type cells. ImageJ was used for post imaging analysis.

4.7 Acknowledgements

The authors would like to thank Dr. Brian Vogler for insightful discussions regarding design and delivery of nuclear DNA constructs into eukaryotic microalgae, and Andy Politis,

Bonnie Panczak, and Brittany Thornton for assistance in compositional analyses. This research was supported by the Department of Energy, Office of Energy Efficiency and Renewable Energy

(EERE) under Agreements No. 22000 and DE-NL0029949. The views and opinions of the authors expressed herein do not necessarily state or reflect those of the United States

Government or any agency thereof. Neither the United States Government nor any agency thereof, nor any of their employees, makes any warranty, expressed or implied, or assumes any legal liability or responsibility for the accuracy, completeness, or usefulness of any information, apparatus, product, or process disclosed, or represents that its use would not infringe privately owned rights.

4.8 Data Availability

The algal strain (Picochlorum renovo), DNA elements, and raw data supporting the conclusions of this manuscript will be made available by the authors, without undue reservation, to any qualified researcher. The genome for Picochlorum renovo is publicly available at https://greenhouse.lanl.gov/greenhouse/organisms. Genomic sequence data can also be accessed at NCBI, Project Number PRJNA558990. The raw data supporting the conclusions of salinity transcriptomics is available from the Sequence Read Archive, under Project Number

PRJNA553204 (https://www.ncbi.nlm.nih.gov/Traces/study/?acc=PRJNA553204).

4.9 References

1. Sheehan, J., Dunahay, T., Benemann, J. & Roessler, P. Look back at the U. S. Department of Energy’s Aquatic Species Program: Biodiesel from Algae; Close-Out Report. (1978). 2. Harris, E. H. Chlamydomonas as a model organism. Annu. Rev. Plant Physiol. Plant Mol. Biol. 52, 363–406 (2001). 3. Wijffels, R. H. & Barbosa, M. J. An outlook on microalgal biofuels. Science 329, 796–9 (2010). 4. Dismukes, G. C., Carrieri, D., Bennette, N., Ananyev, G. M. & Posewitz, M. C. Aquatic phototrophs: efficient alternatives to land-based crops for biofuels. Curr. Opin. Biotechnol. 19, 235–240 (2008). 5. Ajjawi, I. et al. Lipid production in Nannochloropsis gaditana is doubled by decreasing expression of a single transcriptional regulator. Nat. Biotechnol. 35, 647–652 (2017). 6. Weissman, J. C. et al. High-light selection produces a fast-growing Picochlorum celeri. (2018). doi:10.1016/j.algal.2018.09.024 7. Kilian, O., Benemann, C. S. E., Niyogi, K. K. & Vick, B. High-efficiency homologous recombination in the oil-producing alga Nannochloropsis sp. Proc. Natl. Acad. Sci. 108, 21265–21269 (2011). 8. Radakovits, R. et al. Draft genome sequence and genetic transformation of the oleaginous alga Nannochloropsis gaditana. Nat. Commun. 3, 686 (2012). 9. Foflonker, F. et al. Genome of the halotolerant green alga Picochlorum sp. reveals strategies for thriving under fluctuating environmental conditions. Environ. Microbiol. 17, 412–426 (2015). 10. Foflonker, F. et al. The unexpected extremophile: Tolerance to fluctuating salinity in the green alga Picochlorum. Algal Res. 16, 465–472 (2016). 11. Henley, W. J. et al. Phylogenetic analysis of the ‘Nannochloris-like’ algae and diagnoses of Picochlorum oklahomensis gen. et sp. nov. (Trebouxiophyceae, Chlorophyta). Phycologia 43, 641–652 (2004). 12. Krasovec, M. et al. Genome analyses of the microalga Picochlorum provide insights into the evolution of thermotolerance in the green lineage. Genome Biol. Evol. (2018). doi:10.1093/gbe/evy167 13. Foflonker, F., Mollegard, D., Ong, M., Yoon, H. S. & Bhattacharya, D. Genomic Analysis of Picochlorum Species Reveals How Microalgae May Adapt to Variable Environments. Mol. Biol. Evol. 35, 2702–2711 (2018). 14. Elliott, L. G. et al. Establishment of a bioenergy-focused microalgal culture collection. Algal Res. 1, 102–113 (2012).

15. Dahlin, L. R. et al. Down-Selection and Outdoor Evaluation of Novel, Halotolerant Algal Strains for Winter Cultivation. Front. Plant Sci. 9, 1513 (2018). 16. McGowen, J. et al. The Algae Testbed Public-Private Partnership (ATP3) framework; establishment of a national network of testbed sites to support sustainable algae production. Algal Res. 25, 168–177 (2017). 17. Lammers, P. J. et al. Review of the cultivation program within the National Alliance for Advanced Biofuels and Bioproducts. Algal Res. 22, 166–186 (2017). 18. Wang, Y. & Copenhaver, G. P. Meiotic Recombination: Mixing It Up in Plants. (2018). doi:10.1146/annurev-arplant-042817 19. Kamiya, R. Functional diversity of axonemal dyneins as studied in Chlamydomonas mutants. Int. Rev. Cytol. 219, 115–155 (2002). 20. Meng, D., Cao, M., Oda, T. & Pan, J. The conserved ciliary protein Bug22 controls planar beating of Chlamydomonas flagella. J. Cell Sci. 127, 281–7 (2014). 21. Silflow, C. D. & Lefebvre, P. A. Assembly and motility of eukaryotic cilia and flagella. Lessons from Chlamydomonas reinhardtii. Plant Physiol. 127, 1500–7 (2001). 22. Pan, J. & Snell, W. J. Signal transduction during fertilization in the unicellular green alga, Chlamydomonas. Curr. Opin. Microbiol. 3, 596–602 (2000). 23. Frenkel, J., Vyverman, W. & Pohnert, G. Pheromone signaling during sexual reproduction in algae. Plant J. 79, 632–644 (2014). 24. Chung, M., Munro, J. B., Tettelin, H. & Dunning Hotopp, J. C. Using Core Genome Alignments To Assign Bacterial Species. mSystems 3, e00236-18 (2018). 25. Rasala, B. A. et al. Expanding the spectral palette of fluorescent proteins for the green microalga Chlamydomonas reinhardtii. Plant J. 74, 545–556 (2013). 26. Tian, T. & Salis, H. M. A predictive biophysical model of translational coupling to coordinate and control protein expression in bacterial operons. Nucleic Acids Res. 43, 7137–7151 (2015). 27. Quesada-Vargas, T., Ruiz, O. N. & Daniell, H. Characterization of Heterologous Multigene Operons in Transgenic Chloroplasts. Transcription, Processing, and Translation 1. (2005). doi:10.1104/pp.105.063040 28. Georgianna, D. R. et al. Production of recombinant enzymes in the marine alga Dunaliella tertiolecta. Algal Res. 2, 2–9 (2013). 29. Randolph-Anderson, B. et al. Sub-micron gold particles are superior to larger particles for efficient Biolistic® transformation of organelles and some cell types. (2015). 30. Bio-Rad Laboratories, M. B. G. Biolistic ® PDS-1000/He Particle Delivery System.

31. Řezanka, T., Nedbalová, L., Lukavský, J., Střížek, A. & Sigler, K. Pilot cultivation of the

green alga Monoraphidium sp. producing a high content of polyunsaturated fatty acids in a low-temperature environment. Algal Res. 22, 160–165 (2017). 32. White, R. L. & Ryan, R. A. Reviews Long-Term Cultivation of Algae in Open-Raceway Ponds: Lessons from the Field. doi:10.1089/ind.2015.0006 33. Davis, R. et al. Process Design and Economics for the Production of Algal Biomass: Algal Biomass Production in Open Pond Systems and Processing Through Dewatering for Downstream Conversion. (2016). 34. Dong, T. et al. Combined algal processing: A novel integrated biorefinery process to produce algal biofuels and bioproducts. Algal Res. (2015). doi:10.1016/j.algal.2015.12.021 35. Graf, A., Schlereth, A., Stitt, M. & Smith, A. M. Circadian control of carbohydrate availability for growth in Arabidopsis plants at night. 107, (2010). 36. Edmundson, S. J. & Huesemann, M. H. The dark side of algae cultivation: Characterizing night biomass loss in three photosynthetic algae, Chlorella sorokiniana, Nannochloropsis salina and Picochlorum sp. Algal Res. 12, 470–476 (2015). 37. Cross, F. R. & Umen, J. G. The Chlamydomonas cell cycle. Plant J. 82, 370–392 (2015). 38. Poliner, E. et al. Transcriptional coordination of physiological responses in Nannochloropsis oceanica CCMP1779 under light/dark cycles. Plant J. 83, 1097–1113 (2015).

39. Bišová, K. & Zachleder, V. Cell-cycle regulation in green algae dividing by multiple fission. J. Exp. Bot. 65, 2585–2602 (2014). 40. Breuer, G., Lamers, P. P., Martens, D. E., Draaisma, R. B. & Wijffels, R. H. The impact of nitrogen starvation on the dynamics of triacylglycerol accumulation in nine microalgae strains. Bioresour. Technol. 124, 217–226 (2012). 41. Nara, T. et al. Isolation of a LIM15/DMC1 homolog from the basidiomycete Coprinus cinereus and its expression in relation to meiotic chromosome pairing. Mol. Gen. Genet. MGG 262, 781–789 (1999). 42. Mazin, A. V., Mazina, O. M., Bugreev, D. V. & Rossi, M. J. Rad54, the motor of homologous recombination. DNA Repair (Amst). 9, 286–302 (2010). 43. Heyer, W.-D., Li, X., Rolfsmeier, M. & Zhang, X.-P. Rad54: the Swiss Army knife of homologous recombination? Nucleic Acids Res. 34, 4115–25 (2006). 44. Dmitrieva, N. I., Bulavin, D. V. & Burg, M. B. High NaCl causes Mre11 to leave the nucleus, disrupting DNA damage signaling and repair. Am. J. Physiol. Physiol. 285, F266–F274 (2003). 45. Kültz, D. & Chakravarty, D. Hyperosmolality in the form of elevated NaCl but not urea causes DNA damage in murine kidney cells. Proc. Natl. Acad. Sci. U. S. A. 98, 1999–2004

(2001).

46. Fučíková, K., Pažoutová, M. & Rindi, F. Meiotic genes and sexual reproduction in the green algal class Trebouxiophyceae (Chlorophyta). J. Phycol. 51, 419–430 (2015). 47. Rasala, B. A., Chao, S.-S., Pier, M., Barrera, D. J. & Mayfield, S. P. Enhanced Genetic Tools for Engineering Multigene Traits into Green Algae. PLoS One 9, e94028 (2014).

48. Russell, G. E. Plant Breeding for Pest and Disease Resistance : Studies in the Agricultural and Food Sciences. (Butterworth & Co, 1978). 49. Camacho, L. R. et al. Analysis of the phthiocerol dimycocerosate locus of Mycobacterium tuberculosis. Evidence that this lipid is involved in the cell wall permeability barrier. J. Biol. Chem. 276, 19845–54 (2001). 50. Hou, Q., Ufer, G. & Bartels, D. Lipid signalling in plant responses to abiotic stress. Plant. Cell Environ. 39, 1029–1048 (2016). 51. Plucinak, T. M. et al. Improved and versatile viral 2A platforms for dependable and inducible high-level expression of dicistronic nuclear genes in Chlamydomonas reinhardtii. Plant J. 82, 717–729 (2015). 52. Rasala, B. A. et al. Robust Expression and Secretion of Xylanase1 in Chlamydomonas reinhardtii by Fusion to a Selection Gene and Processing with the FMDV 2A Peptide. PLoS One 7, e43349 (2012). 53. Galarza, J. I., Gimpel, J. A., Rojas, V., Arredondo-Vega, B. O. & Henríquez, V. Over- accumulation of astaxanthin in Haematococcus pluvialis through chloroplast genetic engineering. Algal Res. 31, 291–297 (2018). 54. Laurens, L. M. L., Olstad, J. L. & Templeton, D. W. Total Protein Content Determination of Microalgal Biomass by Elemental Nitrogen Analysis and a Dedicated Nitrogen-to- Protein Conversion Factor. in 1–10 (Humana Press, 2018). doi:10.1007/7651_2018_126 55. FALCON Assembler — FALCON 0.5 documentation. Available at: https://pb- falcon.readthedocs.io/en/latest/. (Accessed: 25th April 2019) 56. Hoff, K. J., Lange, S., Lomsadze, A., Borodovsky, M. & Stanke, M. BRAKER1: Unsupervised RNA-Seq-Based Genome Annotation with GeneMark-ET and AUGUSTUS. doi:10.1093/bioinformatics/btv661 57. Stanke, M., Diekhans, M., Baertsch, R. & Haussler, D. Using native and syntenically mapped cDNA alignments to improve de novo gene finding. Bioinformatics 24, 637–644 (2008). 58. Stanke, M., Schöffmann, O., Morgenstern, B. & Waack, S. Gene prediction in eukaryotes with a generalized hidden Markov model that uses hints from external sources. BMC Bioinformatics 7, 62 (2006). 59. Jones, P. et al. InterProScan 5: genome-scale protein function classification.

Bioinformatics 30, 1236–1240 (2014). 60. UniProt Consortium, T. UniProt: the universal protein knowledgebase. Nucleic Acids Res. 46, 2699 (2018). 61. Osterhout, W. J. V. EXTREME TOXICITY OF SODIUM CHLORIDE AND ITS PREVENTION BY OTHER SALTS. J. Biol. Chem. 1, 363–369 (1906). 62. HTStream. Available at: https://github.com/ibest/HTStream. 63. Patro, R., Duggal, G., Love, M. I., Irizarry, R. A. & Kingsford, C. Salmon provides fast and bias-aware quantification of transcript expression. Nat. Methods 14, 417–419 (2017). 64. RFFSC R Development Core Team. R: A language and environment for statistical computing. 409 (2011). 65. Robinson, M. D., McCarthy, D. J. & Smyth, G. K. edgeR: a Bioconductor package for differential expression analysis of digital gene expression data. Bioinformatics 26, 139– 140 (2010).

66. Delcher, A. L. et al. Alignment of whole genomes. Nucleic Acids Res. 27, 2369–76 (1999). 67. Gonzalez-Esquer, C. R., Twary, S. N., Hovde, B. T. & Starkenburg, S. R. Nuclear, Chloroplast, and Mitochondrial Genome Sequences of the Prospective Microalgal Biofuel Strain Picochlorum soloecismus. Genome Announc. 6, (2018). 68. Altschul, S. F., Gish, W., Miller, W., Myers, E. W. & Lipman, D. J. Basic local alignment search tool. J. Mol. Biol. 215, 403–410 (1990). 69. Supek, F., Bošnjak, M., Škunca, N. & Šmuc, T. REVIGO Summarizes and Visualizes Long Lists of Gene Ontology Terms. PLoS One 6, e21800 (2011). 70. Pena, L. et al. Transgenic Plants Methods and Protocols. (2005). doi:10.13387/j.cnki.nmld.2013.02.001 71. Smith, F. D., Harpending, P. R. & Sanford, J. C. Biolistic transformation of prokaryotes: factors that affect biolistic transformation of very small cells. Journal of General Microbiology 138, (2019).

CHAPTER 5

CHARACTERIZATION OF A NOVEL SCENEDESMUS

5.1 Abstract

This chapter presents a body of unpublished work, a portion of which is likely to be published after the submission of this thesis. This chapter covers an intriguing strain of

Scenedesmus which displayed remarkably high biomass accumulation when examining endpoint biomass composition in initial screening. Subsequent studies revealed a high biomass accumulation capacity after cell division, with much of the added biomass in the form of lipids and carbohydrates. This strain provides a striking contrast to the phenotype of Picochlorum renovo presented in chapter 4.

5.2 Introduction

Microalgae are a source of renewable biomass that can be converted to biofuels and bioproducts 1. Year-round cultivation, high salinity tolerances, cultivation on non-arable land, and the ability to sequester CO2 from the air and point sources are some of the favorable characteristics of algae, when compared to terrestrial crops. Indeed, microalgal biomass productivity has been demonstrated at levels up to 10-fold higher than terrestrial feedstocks, such as corn grain 2. Additionally, biodiesel yields from microalgae are estimated to be an order of magnitude higher than terrestrial crops currently employed for biodiesel production, such as oil palm 1. The biomass composition of algae can be fine-tuned to generate biomass rich in protein, carbohydrates, or lipids depending on the strain chosen and growth conditions, such as limitation of nitrogen or phosphorus. Furthermore, algae lack recalcitrant biopolymers, such as lignin, that

83 require extensive downstream processing for biotechnological applications 2. As such, microalgae are a valuable source of sustainable biomass, with the potential to reduce the impacts of climate change, while concurrently reducing demand on petroleum resources.

Despite higher potential yields than terrestrial crops, sustained annual high-productivity microalgal cultivation has yet to be achieved. As such, algae have yet to displace conventional terrestrial feedstocks as a biomass source. Indeed, in order to realize the potential of algae, techno-economic analysis has highlighted the need for significant productivity increases; it is estimated that a minimum productivity of 25 g/m2/day will be required to achieve economically- viable microalgal biofuel production in outdoor raceway ponds, nearly double that of current top- candidate strains 3,4. Additionally, current limitations are compounded by limited genetic tools in high productivity halotolerant algal systems, which hinders fundamental mechanistic inquiry and metabolic engineering strategies in non-model algae.

Through culture collection screening and down-selection, I have identified a novel strain of Scenedesmus that displays a remarkably high biomass accumulation (lipids and carbohydrates) capacity post cell division. This phenotype could be particularly useful, for biotechnological applications, and basic science inquires into the ability of this organism to undergo extended periods of photosynthesis without addition of nitrogen.

5.3 Results

5.3.1 Strain Down-Selection, Physiology, and Compositional Analysis

Initial culture collection screening utilized a custom built photobioreactor capable of comparing growth of up to 36 unique cultivars under simulated outdoor growth conditions

(temperature and PAR cycling), as described in 5 (chapters 2 and 4). Ultimately 107 halotolerant

84 strains were screened under summer growth conditions. In terms of final biomass accumulated, one strain showed remarkably high biomass accumulation potential, reaching approximately double the biomass density as control strain Nannochloropsis salina (CCMP 1776), with

Nannochloropsis representing a genus of high interest for outdoor cultivation (Figure 5.1) 6,7.

Given this unique phenotype, this strain was down-selected for further analysis and development.

Optimal growth temperature was determined to be 25°C (Figure 5.3). Growth was observed up to 57.5 g/L salinity (1.6X seawater), with the highest biomass titer observed at the lowest salinity tested, 17.5 g/L salinity (0.5 X seawater, Figure 5.4). Growth at differing nitrogen to phosphorus ratios showed robust growth at a N:P ratio of 16:0.62 (Figure 5.6), considerably lower than the standard 16:1 Redfield ratio 8. Detailed batch growth analysis showed synchronized cell division occurring during the dark period when grown under a diel cycle. Remarkably, there was significant biomass accumulation post cell division (nitrogen limitation, Figure 5.5). Diel (day and night) biomass productivity was 16-30 g/m2/day, with decreases in productivity observed for later stage cultures. However, 90-98% of additional biomass accumulated in later stage cultures

(post cell division) was in the form of lipids and carbohydrates (Figure 5.2). Biomass compositional analysis at hour 136 (final time point evaluated), indicated biomass consisted of

37% and 43% lipids and carbohydrates respectively (Figure 5.2). Putative lipid storage under nitrogen deprivation is predominately in the form of C16:0 and C18:1n9 (Figure 5.7). Glucose and mannose represent the major monomeric sugars after acid hydrolysis, constituting 29% and

14% of the biomass respectively after 136 hours of growth (Figure 5.8).

Figure 5.1 Representative culture collection growth screening data. The high biomass accumulation capacity phenotype of Scenedesmus sp. (46B-D3) following 6 days of growth is highlighted in black. Nannochloropsis salina CCMP 1776 is bolded in red for reference.

5.3.2 Preliminary Genetic Engineering

A similar approach utilized in Chapter 4 to genetically engineer Picochlorum was applied to this novel Scenedesmus isolate, to engineer the nuclear genome. Three constructs were made that utilized native regulatory elements from the rubisco small subunit, photosystem reaction center subunit (psaD), and elongation factor. The promoter, first intron spliced into bleomycin resistance, and terminator from these three genes were used to express bleomycin resistance linked to mCherry via the 2A peptide.

To determine optimal electroporation settings a matrix of voltages and time constants were performed, while mixing the above constructs. Interestingly the same electroporation settings from Picochlorum were able to generate transformed Scenedesmus, despite different cell sizes. Subsequent trials using individual constructs revealed that only the elongation factor construct was giving rise to transformed algae. These were verified by PCR analysis. The fact the same procedure utilized for Picochlorum could be applied to the Scenedesmus suggests this could be a universal protocol for the transformation of eukaryotic algae.

Figure 5.3 Temperature optima of Scenedesmus sp. (46B-D3), data points represent the average and standard deviation of n=3 biological replicates.

Figure 5.4 Salinity tolerance screening of Scenedesmus sp. (46B-D3), data points represent the average and standard deviation of n=2 biological replicates, following 6 days of growth.

Figure 5.6 Nitrogen: Phosphorus ratio screening of Scenedesmus sp. (46B-D3), data points represent the average and standard deviation of n=2 biological replicates, following 6 days of growth.

Figure 5.7 Lipid analysis via fatty acid methyl esters as a function of time, all data points are an average of n=2 biological replicates; error bars depict the standard deviation of the replicates.

5.3.3 Effect of Prolonged Darkness on Biomass

As seen in Figure 5.3 there appeared to slight increases in the total amount of lipids over

dark periods. It was hypothesized that this could be due to a conversion of carbohydrates (starch)

into lipids in the dark. To test this, cells were grown to a late stationary phase, high in lipids and

carbohydrates, then exposed to a prolonged dark period of 261 hours, by turning the lights off.

As shown in Figure 5.9 there was no increase in lipids during this prolonged dark period.

Interestingly carbohydrates decreased in a non-linear fashion. This is presumably due to the

utilization of starch for metabolism, as glucose is the sugar monomer that decreased from

carbohydrate utilization. During the first ~24 hours there appears to be a more rapid utilization of

carbohydrates, when compared to ~ hour 24 to hour 261. This could have implications on the

understanding of algal metabolism, as it does not appear to be a strictly linear process. The data

from ~ hour 24 to 261 could be classified as a basal metabolism, whereas the data from hour 0 to

24 may be more reflective of an active metabolism, as after the lighted period there will be an additional metabolic burden to repair photosynthetic machinery.

5.4 Discussion

The high biomass productivity phenotype of Scenedesmus sp. post cell division offers benefits for commercial biomass production. Operating expenses can be minimized by increasing the time between harvests, ultimately reducing overall harvesting costs. Further, more biomass in the form of carbohydrates and lipids can be obtained from the same nutrient load, reducing the amount of fertilizer required and associated fertilizer costs. Fertilizer costs can be further reduced via the utilization of a non-Redfield ratio of nitrogen and phosphorus, in the case of Scenedesmus sp. an N:P ratio of 16:0.62 supports equivalent growth with the commonly used 16:1 N:P ratio

(Figure 5.6). This phenotype also offers promise for development of a genetically engineered photosynthetic biocatalyst, wherein product can be sustainably produced for multiple days without the need for nutrient input or cell division. The ability of this strain to grow in seawater offers additional benefits for sustainable biomass production, reducing demand on freshwater resources. However, higher biomass titers were observed at lower salinities (0.5 X seawater), offering potential for utilization of brackish water resources, and to obtain higher productivities in areas with abundant low salinity water sources (Figure 5.4). Peak growth at 25°C indicates this strain is well suited for cultivation in temperate regions (Figure 5.3). The high percentage of lipids and carbohydrates (80% AFDW) during later stage growth is a desirable composition for downstream processing into biofuels and bioproducts 9. The productivity reported here exceeds the productivity reported by Davis et al. for cost-competitive microalgal biofuels 10. These data highlight the potential for this strain to be utilized in outdoor cultivation for biofuels and bioproducts.

5.5 Methods

Strain screening used the same methods as reported in Dahlin, 2018 5 except for a change in the temperature and light cycling to be reflective of summer growth conditions instead of winter. Briefly, 100 mL cultures were sparged with 100 mL/min with air containing 2% CO2, temperature cycles from 21 to 32°C, while lighting cycles from 0 to 965 µmol m-2 s-1. Which was designed to simulate the observed summer outdoor conditions of 1000 L algal ponds at the

Arizona Center for Algae Technology and Innovation testbed site located in Mesa Arizona.

Screening utilized a modified f/2 medium, termed NRE Minimal Medium (NM2). To seawater

(Gulf of Maine, Bigelow Laboratory), the following were added to the reported final

-3 concentration followed by adjust of pH to 8.0 with 12 M HCl: NH4Cl (5.0 x 10 M),

-3 -4 -5 NaH2PO4∙H2O (0.313 x 10 M), Na2SiO3∙9H2O (1.06 x 10 M), FeCl3∙6H2O (1.17 x 10 M),

-5 -8 -8 Na2EDTA∙2H2O (1.17 x 10 M), CuSO4∙5H2O (3.93 x 10 M), Na2MoO4∙2H2O (2.60 x 10 M),

-8 -8 -7 ZnSO4∙7H2O (7.65 x 10 M), CoCl2∙6H2O (4.20 x 10 M), MnCl2∙4H2O, (9.10 x 10 M), thiamine HCl (2.96 x 10-7 M), biotin (2.05 x 10-9 M), cyanocobalamin (3.69 x 10-10 M), Tris base

(24.76 x 10-3 M).

Detailed batch growth for Figure 5.2 utilized the above screening methods, except temperature was held constant at 25°C, lighting utilized a 12 hr light/ 12 hr dark cycling, with lighting at an intensity of 350 µmol m-2 s-1, finally ammonium bicarbonate was employed rather than ammonium chloride and salinity was reduced to 8.75 g/L.

Temperature optima data for Figure 5.3 utilized the above screening methods, except temperature and lighting (constant light at 400 µmol m-2 s-1) were held constant.

Salinity tolerance data, from Figure 5.4, were generated similar to the above screening methods, except the basal NM2 media was generated with half seawater and half Milli-Q water

(Millipore Corporation), to generate the 17.5 g/L salinity media. Higher salinities were obtained by adding seasalts (Sigma S9883) to the appropriate final salinity concentration.

Varying nitrogen to phosphorus ratios, to generate Supplemental Figure 5.6, utilized the above screening methods, except sodium nitrate was utilized as a nitrogen source instead of ammonium chloride. Nitrogen levels were held constant at 5 mM, while phosphate concentrations were decreased to achieve the appropriate ratio reported.

Compositional analysis was done as reported in5.

5.6 References

1. Chisti, Y. Biodiesel from microalgae. Biotechnol. Adv. 25, 294–306 (2007). 2. Dismukes, G. C., Carrieri, D., Bennette, N., Ananyev, G. M. & Posewitz, M. C. Aquatic phototrophs: efficient alternatives to land-based crops for biofuels. Curr. Opin. Biotechnol. 19, 235–240 (2008). 3. Quinn, J. C. & Davis, R. The potentials and challenges of algae based biofuels: A review of the techno-economic, life cycle, and resource assessment modeling. Bioresour. Technol. 184, 444–452 (2015). 4. Sheehan, J., Dunahay, T., Benemann, J. & Roessler, P. Look back at the U. S. Department of Energy’s Aquatic Species Program: Biodiesel from Algae; Close-Out Report. (1978). 5. Dahlin, L. R. et al. Down-Selection and Outdoor Evaluation of Novel, Halotolerant Algal Strains for Winter Cultivation. Front. Plant Sci. 9, 1513 (2018). 6. McGowen, J. et al. The Algae Testbed Public-Private Partnership (ATP3) framework; establishment of a national network of testbed sites to support sustainable algae production. Algal Res. 25, 168–177 (2017). 7. Lammers, P. J. et al. Review of the cultivation program within the National Alliance for Advanced Biofuels and Bioproducts. Algal Res. 22, 166–186 (2017). 8. Alfred C. Redfield. The Biological Control of Chemical Factors in the Environment. Am. Sci. 46, 205–221 (1958). 9. Dong, T. et al. Combined algal processing: A novel integrated biorefinery process to produce algal biofuels and bioproducts. Algal Res. (2015).

doi:10.1016/j.algal.2015.12.021 10. Davis, R. et al. Process Design and Economics for the Production of Algal Biomass: Algal Biomass Production in Open Pond Systems and Processing Through Dewatering for Downstream Conversion. (2016).

CHAPTER 6

CONCLUSIONS AND FUTURE DIRECTIONS

6.1 Conclusions and Future Directions

The initial hypothesis pursued during this thesis, that strains beneficial for large scale biomass production could be identified through culture collection screening, was largely supported. A number of promising isolates were ultimately identified for both large scale cultivation and basic science inquires.

Outdoor cultivation of algae is a difficult endeavor, from the data presented in chapter 2 it is clear that in future algal cultivation trials the starting biomass should not be generated in a greenhouse or other temperature-controlled environment. This led to a drastic shift in lipid composition as the algae transitioned from warmer to colder conditions, likely hampering growth by extending the lag phase. The observation of glucose (presumably starch) utilization upon inoculation is intriguing. This potentially indicates higher productivities can be achieved by harvesting ponds once significant starch and lipids have accumulated, which will provide the energy needed for cell division and metabolic remodeling once nutrients are replenished.

Harvesting and subsequent pond resets (re-inoculation with fresh nutrients) should only be done at night, preferably immediately after the sun sets.

Although not presented in this thesis I performed some preliminary pH optimization (via pH controllers and CO2 injection) experiments, from those it was clear (for outdoor cultivation) that ammonia (gas) or ammonium hydroxide (liquid) should be the sole nitrogen source, as

ammonia is relatively cheap and avoids the problems associated with pH decline when using ammonium salts. Future work on genetic engineering to utilize phosphite (potentially as phosphorous acid) will allow selective monoculturing of a desired strain. By using ammonia and phosphorous acid as nutrient sources this will allow more efficient water recycling as counter ions (e.g. sodium and chloride) will not build up. These nutrients are also among the cheapest nitrogen and phosphorus sources. Given the caustic and acidic nature of these compounds continued use of buffers (e.g. Tris) and less hazardous nitrogen and phosphorus sources is appropriate for lab scale work.

The intent of chapter 3 was largely to introduce and put in context the genetic engineering results of chapter 4. Although not highlighted in chapter 3 it is important to recognize the power of sexual mating, and selective breeding of plants with desirable characteristics. For example, this led to the development of teosinte grass into the corn of today.

Few researchers are exploring the power of sexual mating in algae, in part because screening of progeny is exceedingly difficult due to their microscopic nature. Further, sexual mating is poorly understood and inducing mating phenotypes is only routine for the model organism

Chlamydomonas. There is significant evidence that many algae studied to date can undergo meiosis, the development of methods to induce mating and screen progeny is an area of important future work.

The Picochlorum described in chapter 4 is an interesting organism, while it does not produce large quantities of lipids it could be an ideal candidate for starch production. It is my belief that in the future microalgae could be utilized for the production of sugars (starch), outcompeting terrestrial crops such as sugar cane and corn. This will require significant advances in algal cultivation and downstream processing. Accomplishing cost competitive sugar

production could solve several environmental problems and restore much of the land currently used for sugar production to a natural state (e.g. the rainforest and Great Plains).

Future work on this Picochlorum isolate should target development of additional genetic tools. Homologous recombination and CRISPR/Cas systems are an obvious next step as this would allow targeted knock-out of genes, likely essential for the long sought after productivity increases from antenna reduction. Additional nuclear promoters would also be desirable, I have tested a few promoters with and without introns with little success in finding variable strength promoters. In regards to chloroplast engineering attaining homoplasmy will be essential, and development of a promoter library in the chloroplast will allow fine-tuned gene expression.

Because there are multiple genome copies maintained in the chloroplast, development of

CRISPR/Cas systems in the chloroplast would have limited utility, although may be useful in obtaining homoplasmy. Homologous recombination in the chloroplast is prevalent and should allow targeted knockout or knockdown of desired genes. Once additional tools have been developed application of the tools can then be pursued for growth enhancements, or basic science inquires.

There appears to be a need for further classification of the Picochlorum genus. The literature has varying reports of lipid productivity in Picochlorum, and the whole genome alignments presented in chapter 4 suggest a broad diversity in what is currently classified as

Picochlorum. As genome sequencing becomes cheaper, researches need to move past classic 18S or 16S sequencing for identification and find more robust metrics.

Future screening studies should recognize the benefit in growing clonal isolates individually and comparing growth, as this approach led to the down-selection of the novel

Scenedesmus isolate described in chapter 5. Other screening approaches, such as selective cultivation at high dilution rates may not capture strains of this nature. It is my belief that this strain, and strains with similar growth characteristics will ultimately prove to be more advantageous for large scale cultivation, as significantly more biomass can be generated with the same nutrient input while concurrently reducing the need for daily harvesting of outdoor ponds.

Further, once nutrient deprived the described Scenedesmus appears to degrade chlorophyll, thereby shrinking the photosynthetic antenna, a long sought-after goal in algal engineering.

Despite this shrunken antenna the algal remains photosynthetically active, producing lipids and carbohydrates. It is imperative that future studies look closely at this phase of growth to fully characterize exactly what is occurring, then exploit the knowledge gained for improvements in algal cultivation. Development of additional genetic tools, such as those described above for

Picochlorum will allow further advancement of this strain.

This thesis described the identification of a number of promising strains of algae for biotechnological applications, and basic science inquiries. Microalgae are inherently highly productive phototrophs, understanding and utilizing microalgae for applied purposes offers a route to solve many of the challenges facing mankind. It is my hope that the algal community adopts these strains for scientific and commercial endeavors.

APPENDIX A FURTHER WINTER STRAIN CHARACTERIZATION

A.1 Temperature, Salinity, and N:P Optima

Optimal growth temperatures were determined for strains 14-F2 and 4-C12, two promising strains for winter cultivation, discussed in Chapter 2 (Figure A.1). Methods for determining temperature optima are presented in Chapter 4.

Figure A.1 Optimal growth temperature for 14-F2 and 4-C12. Data point represent the average and standard deviation from n=3 biological replicates.

Growth at varying salinities was also tested for these strains. In general, higher biomass accumulation was noted at lower salinities, a potentially notable exception is higher growth at

22.5 g/L salinity vs. 17.5 g/L salinity for strain 4-C12 (Figure A.2). Growth conditions were the same used in winter strain screening describe in Chapter 2.

100

101

Growth at varying nitrogen to phosphorus ratios was also performed for strain 14-F2.

This largely confirmed the canonical Redfield ratio, with decreases in growth observed when less phosphorus was available (Figure A.3).

102

APPENDIX B FURTHER P. RENOVO DEVELOPMENT

B.1 Nitrogen: Phosphorus Ratios

P. renovo growth at varying nitrogen to phosphorus ratios was performed. The data is somewhat variable, and potentially heavily skewed by the fact growth was allowed to occur for

~6 days, which in hindsight is not appropriate for P. renovo as growth appears to have ended after ~4 days. However, the results are worth noting here as follow on experiments may yield interesting results since remarkably high biomass titers were observed for a ratio of ~16:0.35, significantly less than the canonical Redfield ratio of 16:1. Additionally it appears as though there was an increase in growth from a ratio of 16:1 to 16:0.62, which is unexpected as less phosphorus would not be expected to yield higher growth. If accurate this could be due to phosphate toxicity, which is a rare but established phenomenon in plant horticulture. This observation of high growth with relatively small amounts of phosphorus may be attributable to the small genome, presumably DNA is a large sink of phosphorus in the cells and having a small genome would require less phosphorous to maintain. If this data can be refined and replicated it could lead to the utilization of less phosphorus needed for outdoor growth, reducing fertilizer inputs and ultimately reducing costs associated with biomass production. It is worth noting that the nitrogen source in this experiment was nitrate, which may account for overall reduced growth rates (Figure B.1). Future experiments should likely target ammonium addition, however accurate quantification of actual nitrogen added is difficult when using ammonium salts, as nitrogen can be lost as ammonia gas, hence the logic of using nitrate in these initial trials.

Growth conditions for this was the same as those used for summer screening (see Chapter 4).

103

Figure B.1 Growth analysis of P. renovo at varying nitrogen to phosphorus ratios. Top: endpoint biomass titer after 6 days of growth. Bottom: optical density growth curves.

104

B.2 Additional Genetic Engineering Notes During the writing of the manuscript that makes up Chapter 4 I have continued to develop tools in order to genetically engineer P. renovo, there are also some interesting notes that did not make it into the resultant publication that some readers may find beneficial to be aware of. These are highlighted below.

When initially developing tools for chloroplast engineering we (Calvin Henard and I) had initially designed constructs with homology to the tRNA I – tRNA A region of the chloroplast genome. Three constructs were designed with the same homology arms, however had different promoter and terminator pairs, these were the promoter and terminator for the rubisco large subunit (cbbL), photosystem II protein D1 (psbA), and the 16s rRNA operon. In regard to the cbbL and psbA pairs, these yielded colonies at very high efficiencies, and high sfGFP expression, however despite screening 50+ colonies, none had integrated into the tRNA I – tRNA A locus. Surprisingly the DNA was integrating into the native cbbL and psbA genes, presumably through homologous recombination between the relatively small paired promoter and terminator elements, despite having a larger homology region to a different area of the genome. The 16s rRNA promoter/”terminator” did integrate into the desired locus, and hence why this was presented in the resultant publication. After further work on this however I highly doubt the area we had originally thought of as the 16s rRNA terminator is actually a terminator sequence as it was the DNA immediately downstream of the 16s rRNA. The 16s rRNA, tRNA I, tRNA A, and 23s rRNA are almost certainly functioning as an operon, as such any terminator element would fall after the 23s rRNA. To further confuse matters, after additional reading of the literature it is fairly well established that canonical bacterial terminators do not actually cause transcriptional termination in the chloroplast. There is still much to learn about the engineering

105 of plant chloroplasts, they can not be treated as simply as a bacterial system. There are important differences. Some of these are published, others will need to be determined with ongoing research.

In regard to engineering of the nucleus I spent nearly a year working on various aspects of gene expression. In the publication that makes up Chapter 4 I did note that we tried different promoters, with and without their respective introns. As noted in the publication these did not appear to have variable strengths, and I would like to re-iterate that here. Introns did not seem to have any impact on gene expression, nor did the different promoters tested. This is in direct contrast to the published Chlamydomonas work. To increase gene (mCherry) expression I had also tried some other approaches. Expression of mCherry on a separate promoter-terminator construct (not utilizing a 2A peptide) did not increase gene expression. A different codon optimization of mCherry, when using the published 2A construct did not increase gene expression. I also had previously tried gfp and mTagBFP fluorescent proteins, they appeared to have a lower signal to noise ratio than mCherry, making quantification more difficult. In order to determine 2A cleavage efficiency I tried multiple western blots with two different antibodies, I consistently had problems with non-specific binding to other proteins of similar size making quantification difficult. As such I do not recommend trying additional mCherry western blots, using mCherry antibodies. Antibodies to a different protein, or protein tag will likely allow determination of cleavage efficiency.

I have attempted numerous approaches to achieve Cas9 editing. I was able to successfully design and generate guide RNA’s for Cas9, as determined from an in vivo assay utilizing Cas9 purchased from the qb3 core facility at UC Berkeley. I then attempted electroporation of Cas9 protein, complexed to guide RNAs for the nitrate reductase gene. Many colonies were obtained,

106 but did not have edits, either indels or insertion of the selection marker into the targeted locus.

Given the success in editing corn and wheat using a biolistics approach to deliver Cas9 RNPs I also attempted those approaches, these approaches did not yield transgenic algae. I then went onto express the Cas9 protein with a SV40 nuclear localization signal, using the construct described in chapter 4, swapping out mCherry for the Cas9 protein. This was a rather large construct but did yield colonies at a relatively low efficiency, most (10/12) of these were truncated, as expected, and did not contain a full cas9 gene. I was able to identify 2 colonies that did contain the full cas9 gene and was able to sequence verify the construct. Thus, Cas9 toxicity seems unlikely, and these may serve as editor lines. I did try initial trials electroporating in a selection marker (G418) and guide RNAs, these caused arcing, presumably from salts in the guide RNA. Despite arcing I obtained ~100 colonies and screened them for insertion of the selection marker at the targeted locus, I did not observe any insertion. I have yet to sequence these for indels. It is my belief the best approach for additional cas9 editing will be through electroporation of an RNP, as this is in theory simplistic and does not require cas9 expression from the genome. It does not appear this will be an easy endeavor though as my initial trials did not work. I would base future electroporation trials from the methods described in Brian

Vogler’s thesis, a previous graduate student at the Colorado School of Mines.

107

APPENDIX C

Chapter 2

108

109

110

111

Chapter 3

112

Chapter 4

113

114

115

116

117