Functional Biology, Photophysiology and Nutrient Dynamics of Non-Constitutive Mixotrophic Protists

UNIVERSITY OF COPENHAGEN FACULTY OR DEPARTMENT

PhD Thesis

Kirstine Drumm

Functional biology, photophysiology and nutrient dynamics of non-constitutive mixotrophic protists

Supervisor: Per Juel Hansen

Submitted on: 14 February 2020

UNIVERSITY OF COPENHAGEN FACULTY OR DEPARTMENT

PhD Thesis Kirstine Drumm

Functional biology, photophysiology and nutrient dynamics of non-constitutive mixotrophic protists

Supervisor: Per Juel Hansen

Submitted on: 14 February 2020 University of Copenhagen Department of Biology Marine Biological Section

PhD thesis Kirstine Drumm

Functional biology, photophysiology and nutrient dynamics of non-constitutive mixotrophic protists

Principal Supervisor: Per Juel Hansen, Professor Marine Biological Section Department of Biology University of Copenhagen

This thesis has been submitted to the PhD school of The Faculty of Science, University of Copenhagen

Submitted on: 14 February 2020 Date of defense: 27 March 2020

Chairman of assessment committee: Lasse Riemann, Professor. Marine Biological Section, University of Copenhagen, Helsingør, Denmark

Assessment committee: Christopher J. Gobler, Professor. School of Marine and Atmospheric Sciences, Stony Brook University, New York, USA

Assessment committee: George McManus, Professor. Marine Sciences, University of Connecticut, Connecticut, USA

Cover photos: Sofie B. Binzer, drawings of Dinophysis acuminata, Mesodinium major, M. rubrum and Nusuttodinium aeruginosum

Preface

This thesis represents the work done during my years as a PhD fellow at the University of Copenhagen. Funding was provided by a full PhD stipend from the Danish Council for Independent Research, granted to Per Juel Hansen. This work was done under the supervision of Per Juel Hansen, at the Marine Biological Section in Helsingør, Denmark.

The original title of this thesis was “Gene expression, photophysiology and nutrient dynamics in marine protists with acquired phototrophy”. The aim was to obtain insights to non-constitutive mixotrophs through the combination of developed skills such as growing mixotrophic protists, imaging behavior using epifluorescence, uptake kinetics of nutrients, genetic analysis and modelling. This developed, in my humble opinion, into a very fruitful collaboration and resulted in many fascinating results and challenging discussions.

This journey started for me, in the company of three wise guys who suddenly became very excited about a little blue dinoflagellate. Per, Niels and Øjvind you are a power couple that no one can say no to. I remember Per saying “it’s just out there, you can just go and catch it….”. Well, almost one year later we (Mette and I) had a culture and four years later this PhD fellow began. I will forever be grateful that you convinced us to “just go and catch it”. It continues to amaze me how such a small organism can be so complicated, so fragile and yet so perfectly built and adaptable to its surroundings. Sometimes it seems that the more we learn the less we know.

Helsingør, February 2020

Table of Contents

List of papers ...... 1 Abstract ...... 3 Dansk resumé ...... 5 Introduction ...... 7 Mixotrophy in planktonic protists ...... 7 Non-constitutive mixotrophs ...... 8 Generalist non-constitutive mixotrophs (GNCMs) ...... 9 Specialist non-constitutive mixotrophs ...... 9 Study organisms and organelle retention ...... 10 Dinoflagellates ...... 10 Ciliates ...... 11 Eco-Physiology of SNCMs ...... 13 Growth responses to different irradiance ...... 13 Photosynthetic activity and chlorophyll a ...... 13 SNCMs and their dependence on their prey ...... 14 Nutrients ...... 14 Major nutrients of phototrophic protists ...... 14 N-forms and metabolism ...... 15 Uptake of nutrients by NCM ...... 16 Abbreviations ...... 18 Aims of this thesis ...... 19 Specific aims of each paper ...... 19 Results and discussion ...... 21 The role of sequestered chloroplasts ...... 21 Cell growth and irradiance ...... 21 Photosynthesis, chlorophyll a and photoacclimation ...... 21 Ingestion of prey ...... 22 Integrations of the chloroplasts ...... 22 Prey deprivation ...... 22 Sequestered nuclei ...... 23 Symbiosis or sequestered ...... 24

Effect of inorganic nitrogen ...... 24 Uptake of inorganic nitrogen and urea ...... 24 Assimilation of nitrogen ...... 25 Prey deprivation and N-uptake ...... 26 Future perspectives...... 26 References ...... 28 Acknowledgements ...... 33 Papers ...... 35

List of papers

I. Drumm K, Liebst-Olsen M, Daugbjerg N, Moestrup Ø and Hansen PJ. 2017. Effects of irradiance and prey deprivation on growth, cell carbon and photosynthetic activity of the freshwater kleptoplastidic dinoflagellate Nusuttodinium (= Gymnodinium) aeruginosum (Dinophyceae). PloS one 12(8): e0181751.

II. Drumm K, Norlin A, Altenburger A, Kim M and Hansen PJ. 2020. Physiological responses of Mesodinium major to irradiance, prey concentration and prey starvation. Manuscript.

III. Kim M, Drumm K, Daugbjerg N and Hansen PJ. 2017. Dynamics of sequestered cryptophyte nuclei in Mesodinium rubrum during starvation and refeeding. Frontiers in Microbiology 8: 423.

IV. Johnson MD, Lasek-Nesselquist E, Moeller HV, Altenburger A, Lundholm N, Kim M, Drumm K, Moestrup Ø and Hansen PJ. 2017. Mesodinium rubrum: The symbiosis that wasn’t. Proceedings of the National Academy of Sciences 114(7): E1040-E1042.

V. Drumm K, Glibert PM, Flynn KJ and Hansen PJ. 2020. Effects of nitrogen on two non- constitutive mixotrophs, Mesodinium rubrum and Dinophysis acuminata I: Effects of ammonium concentrations on growth, inorganic carbon uptake and removal of different inorganic nitrogen forms as a function of time after prey deprivation. Manuscript.

VI. Drumm K, Glibert PM, Flynn KJ and Hansen PJ. 2020. Effects of nitrogen on two non- constitutive mixotrophs, Mesodinium rubrum and Dinophysis acuminata II: Potential for uptake of ammonium, nitrate and urea as a function of time after prey deprivation. Manuscript.

List of papers

II. Drumm K, Norlin A, Altenburger A, Kim M and Hansen PJ. 2020. Physiological responses of Mesodinium major to irradiance, prey concentration and prey starvation. Manuscript.

III. Kim M, Drumm K, Daugbjerg N and Hansen PJ. 2017. Dynamics of sequestered cryptophyte nuclei in Mesodinium rubrum during starvation and refeeding. Frontiers in Microbiology 8: 423.

Additional work done during the PhD not to be included in the evaluation of the thesis

Appendix I Altenburger A, Cai H, Drumm K, Kim M, Garcia L, Zhan X, Li S, Hansen PJ, John U, Li Q and Lundholm N. 2020. Limits to the control of sequestered cryptophyte chloroplasts in the marine ciliate Mesodinium rubrum. Manuscript.

Appendix II Hansen PJ, Anderson R, Stoecker DK, Decele J, Altenburger A, Blossom HE, Drumm K, Mitra A and Flynn KJ. 2019. Mixotrophy among freshwater and marine protists. Encyclopedia of Microbiology, Academic Press: 199-210

Abstract

Mixotrophy is the combination of phototrophic and heterotrophic nutrition in a single organism. Mixotrophy is no longer confined to phototrophic protists (i.e. organisms with “built-in” chloroplasts, constitutive mixotrophs), but also includes protists that lack built-in chloroplasts, and instead host symbionts or retain functional chloroplasts from their prey (non-constitutive mixotrophs). This thesis has studied specialist non-constitutive mixotrophs (SNCM), referring to the selective behaviour were only a special type of prey can be used for C-fixation, as appose to generalist (GNCM) where a C- fixation can be acquired from different types of prey organisms. SCMNs are found in both marine and freshwater systems. In this thesis I studied one freshwater dinoflagellate, Nusuttodinium aeruginosum, and three marine species, one dinoflagellate, Dinophysis acuminata, and two ciliates, Mesodinium rubrum and M. major. Light response curves were done on N. aeruginosum and M. major and illustrated a high dependence of light for cell growth and photosynthetic activity. Their light responses were very similar to that of a phototrophic organism. M. major was able to increase the cellular content of chlorophyll a (Chl a) at low irradiances, which is similar to what has been shown previously for M. rubrum, indicating abilities of photoacclimation. Chl a was not measured in N. aeruginosum, so it is unknown if this species has the ability of photoacclimation. Inorganic carbon fixation accounted for up to 26% and 95% of the total C need for N. aeruginosum and M. major, respectively. When these two organisms were subjection to prey deprivation they continued to grow, going through 4-6 cell divisions over a period of 20-25 days, hereafter the cells just sustained themselves. So, N. aeruginosum and M. major both displayed a close relation to the retained prey chloroplasts. The two ciliates, M. rubrum and M. major, keep some genetic material from the prey. In well-fed cultures the ciliates keep a single prey nucleus, which was termed the Centered Prey Nucleus (CPN). This nucleus is placed in close proximity to the ciliates macro and micro nuclei. The CPN nucleus is an enlarged version of a single prey nuclei, which the ciliates sequester from their prey. Besides this enlarged prey nucleus, well-fed ciliates keep some additional prey nuclei, termed “Extra Prey Nuclei” (EPN). In prey starved ciliate cultures only one of the daughter cells receives the CPN. In the other daughter cell one of the EPNs moved into close proximity to the ciliate nuclei and enlarges. This phenomenon was found for both M. rubrum and M. major. A direct correlation between the percentage of cells with a CPN and photosynthetic activity was found, stressing the importance of retaining the CPN. This was also confirm through gene expressing studies, where none of the genes related to photosynthesis were expressed by M. rubrum. The exact relationships between the photosynthetic prey and the host have in many aspects been clarified, but some discussions are still going on with regard to the stability of the symbiotic relationships. Paper IV was part of such dispute, where the conclusion of Qui et al. 2016 that M. rubrum “farms” endosymbiotic T. amphioxeia, was questioned. Here it is was argued that M. rubrum does not have a permanent endosymbiont, but rather rely on prey ingestion for prey chloroplasts, prey nuclei and other cell organelles.

- + Effects of nitrogen (N) availability and the N-form (NO3 , NH4 and urea) was investigated for the SNCMs, D. acuminata and M. rubrum, and compared to their chloroplasts donor, the cryptophyte T. + amphioxeia. Uptake and utilization of NH4 by D. acuminata was demonstrated, which led to - increased cell growth and photosynthetic activity. Uptake of NO3 was shown, but only at very low - rates, and D. acuminata could not utilize NO3 for cellular growth. M. rubrum was able to take up all forms of inorganic N, and cell growth responded positively until N-depletion, very similar to the prey, + T. amphioxeia. Also, additions of NH4 to M. rubrum and T. amphioxeia cultures did not affect + photosynthetic activity. Common for all tree species was the repressive effect by NH4 additions on - - the uptake rate of NO3 and urea, and an increased uptake rate of NO3 with increasing ambient concentration – similar to a phototroph. Finally, the effect of time of prey deprivation was shown to have no effect on potential N-uptake in M. rubrum, while it had some effect on D. acuminata.

Dansk resumé

Mixotrofi er kombinationen af at ernære sig ved fototrofi og heterotrofi i en enkelt organisme. Mixotrofi er ikke længere begrænset til kun at omfatte fototrofe protister (dvs. encellede organismer med ”indbyggede” kloroplastre, i den international litteratur kaldet constitutive mixotrophs, CMs), men inkluderer også protister, der ikke har ”indbyggede” kloroplastre, og i stedet tilbageholder funktionelle kloroplastre fra deres bytte (kaldet non-constitutive mixotrophs, NCMs). I denne afhandling studeres specialiserede non-constitutive mixotroph (SNCM); navnet henviser til deres specialiserede adfærd hvor kun en specifik bytte-type, modsat generalister (GNCM) udnytte forskellige byttetyper. I denne afhandling har jeg studeret én ferskvandsdinoflagellat, Nusuttodinium aeruginosum, og tre marine arter, én dinoflagellat, Dinophysis acuminata, og to ciliater, Mesodinium rubrum og M. major. Lysresponskurver blev udarbejdet for N. aeruginosum og M. major. Disse organismer er stærkt afhængige lys for at kunne gro. Responskurverne er meget lig dem som man kender for almindelige fototrofe alger. Ved lave lysintensiteter kunne M. major forøge mængden af cellulært klorofyl a, hvilket også tidligere er blevet vist for en nærtbeslægtet art, M. rubrum. Denne evne indikerer at denne art ligeledes kan fotoakklimatisering, og dermed tilpasse miljøer med lavt lys. Mængden af cellulært klorofyl a blev ikke målt for N. aeruginosum, så det er uvist om denne også kan fotoakklimatisere. Fiksering af uorganisk C (=fotosyntese) udgjorde op til 26% og 95% af cellens kulstofbehov for henholdsvis N. aeruginosum og M. major. Disse to organismer gennemgik 4-6 celledelinger over en perioden på 20-25 dage efter føden blev frataget dem. Herefter kunne de kun opretholde sig selv gennem længere tid (månedsvis). Både N. aeruginosum og M. major udviste en således en tæt relation til deres tilbageholdte kloroplastre. De to ciliater, M. rubrum og M. major tilbageholder genetisk materiale fra deres bytte. Hos velfodrede kulturer har ciliaterne en enkelt byttekerne (i den international litteratur kaldet Centered Prey Nucleus, CPN). Denne kerne er placeret i tæt forbindelse med ciliaternes egne makro og mikro kerner. CPN kernen er en forstørret version af en enkelt byttekerne, som ciliaten har tilbageholdt fra deres bytte. Ud over denne forstørrede kerne har velfodrede ciliater yderligere byttekerne kaldet ”Extra Prey Nuclei” (EPN). Når ciliaterne fratages deres bytte er det kun den ene datter-celle som modtager en CPN. I den anden datter-celle bliver en af EPNerne flyttet ind til ciliatens kerner og forstørres; dette fænomen blev fundet både hos M. rubrum og M. major. En direkte forbindelse mellem procentdelen af celler med CPN og fotosyntetisk aktivitet blev fundet, hvilket understreger vigtigheden af at tilbageholde en CPN for disse ciliater. Denne forbindelse blev bekræftet gennem studier med genekspression, hvor ingen af generne som er relateret til fotosyntese blev udtrykt af M. rubrum. Det egentlige forhold mellem det fotosyntiserende bytte og værten er til en vis grad blevet afdækket, men der foregår til stadighed diskussioner omkring stabiliteten af det symbiotiske forhold. Artikel IV deltog i en sådan diskussion, da de draget tvivl om konklusioner fremkommet i Qui et al. 2016 om at M. rubrum ”dyrkede” endosymbiotiske T. amphioxeia. Nyere undersøgelser har nemlig vist at M. rubrum ikke har permanente endosymbionter, men i stedet er afhængige af indtage bytte for at få bytte kloroplastre, byttekerner og andre organeller.

- + Effekten af tilgængeligheden af kvælstof (N) og hvilken N-form (NO3 , NH4 og urea) blev undersøgt for udvalgte SNCM-arter, D. acuminata og M. rubrum og sammenlignet med deres kloroplastdonor, + rekylalgen T. amphioxeia. Optag og udnyttelse af NH4 blev vist for D. acuminata og det blev vist at - dette resulterede i en forøget vækst og fotosynteseaktivitet. Optag af NO3 blev vist, men de fundne - rater var meget lave rater, og D. acuminata kunne ikke udnytte NO3 til celle vækst. M. rubrum kunne optage alle former for uorganisk N og positiv celle vækst forekom indtil N var opbrugt, meget lignede + de fundne respons hos byttet, T. amphioxeia. Derudover medførte tilsætningen af NH4 til kulturer med M. rubrum og T. amphioxeia ikke en synlig effekt på deres fotosynteseaktivitet. Fælles for alle + - tre arter var den repressive effekt af tilsætningen af NH4 på optaget af NO3 og urea, samt en forøget - optagsrate af NO3 med stigende koncentrationer i mediet – som hos fototrofe organismer. Afslutningsvis viste M. rubrum sig at være upåvirket af at blive sultet i forhold til det potentielle optag af N, hvorimod D. acuminata var delvis påvirket.

Introduction Mixotrophy in planktonic protists Mixotrophy is the combination of phototrophic and heterotrophic nutritional mode in a single organism. Over the past 40 years the importance and distribution of mixotrophic organisms have been acknowledge and recognized as a common member of the plankton community globally (Stoecker, Hansen et al. 2017). Also, it has become clear that there is not only one form of mixotrophy, but rather it is an ongoing spectra/continuum from almost completely phototrophy through mixotrophy to almost complete heterotrophy (Fig. 1; Andersen, Aksnes et al. 2015). A protist will invest in a trophic strategy somewhere along this spectrum. However, the investment in autotrophy and heterotrophy, respectively is not a fixed ratio, but may change considerable depending on environmental conditions.

Figure 1. Conceptual illustration of the trophic continuum from photosynthesis over mixotrophy to pure heterotrophy. The trophic continuum is defined by the allocation into three harvesting traits: nutrient uptake, photosynthesis and phagotrophy, each leading to uptake of dissolved inorganic nutrients, CO2 and particulate organic matter. A specific organisms’ trophic strategy is defined as a point within the triangle: an organism in the middle will invest equally into all three traits, an organism somewhere along the left side will be a pure phototroph and an organism at the right tip would be a pure heterotroph. (Taken from Anderson et al. 2015, with original figure legend.)

Mixotrophy is no longer confined to phototrophic protists (i.e. organisms with built in chloroplasts), but also includes protists that lack “built in” chloroplasts, and instead host endo-/ectosymbionts or retain functional chloroplasts from their prey (Mitra, Flynn et al. 2016). The latter one is also referred to as having kleptochloroplasts (stolen chloroplasts), enabling the protist to be photosynthetic active (Schnepf and Elbrächter 1992). The sequential loss of cell organelles in protists, from those hosting complete endosymbionts, via those with reduced endosymbionts, to those that just retain the chloroplasts, could provide us with an understanding of the evolutionary steps that are still unknown to us (Wilcox and Wedemayer 1984). Many questions are still open with regard to the functional biology and physiology of protists with kleptochloroplasts, thus the aim of this thesis is to contribute to our knowledge of these organisms. Non-constitutive mixotrophs The terms “constitutive” and “non-constitutive” mixotrophs (CM and NCMs, respectively) were introduced as a way to separate protists that have permanent “built-in” (constitutive) chloroplasts from protists that acquire the ability for inorganic C-fixation through the ingestion and sequestration of photosynthetic prey or prey chloroplasts (non-constitutive; Fig. 2; Mitra, Flynn et al. 2016). Furthermore, the model divides NCMs up in two, depending on the protist selectivity and relation to the prey. If the protist has a loose relation and can acquire C-fixation from different types of prey they are referred to as generalists (GNCMs), whereas the specialists (SNCMs) usually depend on a single type of prey organism and have a close relation to it.

Figure 2. Flowchart showing the pathways used to derive the functional groups to classify the planktonic protists (adapted from Mitra et al. 2016).

Generalist non-constitutive mixotrophs (GNCMs) GNCMs are found mainly within dinoflagellates and oligotrich ciliates and have received very little attention (Stoecker, Hansen et al. 2017). The retained chloroplasts do not divide in the ciliate; instead the chloroplasts are continuously replaced through ingestion. The contribution of inorganic carbon uptake may vary a lot depending of prey type and environmental changes, and inorganic carbon uptake may contribute from 0-70 % of the total carbon needs, and a small contribution from inorganic C-fixation can be sufﬁcient to increase survival and/or enhance growth in many of these species. GNCMs are believed to obtain most of their N and P from ingestion of prey, but quantitative data are lacking on their N and P budgets (Stoecker, Johnson et al. 2009). GNCMs are generally very sensitive to prey deprivation and can usually survive less than a week without food. Specialist non-constitutive mixotrophs Specialist non-constitutive mixotrophs (SNCM) can be divided into two groups (Fig. 2) defined by whether the prey symbiont is kept alive as ecto- or endosymbiont (eSNCM) or the plastids are selectively retained (at times combined with other organelles, pSNCM). The eSNCM is most often exemplified through the green Noctiluca scintillans, that harbor thousands of pedinophytes swimming free inside the cell (Hansen, Miranda et al. 2004). But also members from the ciliates, foraminiferans and radiolarians are found within eSNCMs (Fig 3).

Figure 3. Examples of endosymbiont and ectosymbiont. A) the green Noctiluca scintillans (Photo by Per Juel Hansen), and B) the Foraminifera (Globigerinoides sacculifer) (Photo by Katsunori Kimoto).

The focus of this thesis will be on the pSNCMs. These plastidic specialists are found globally in oceanic open waters, coastal and estuaries and in freshwater. They are mainly found among dinoflagellates and ciliates. The most known and investigated are the red Mesodinium spp., known to cause red tides, and Dinophysis spp. that are responsible for the diarrhetic shellfish poisoning (DSP; Hansen, Nielsen et al. 2013). SNCMs exploit chloroplasts from very specific types of prey. Having a close relation with specific types of prey have both advantages and disadvantages. The main advantage is the ability to control the chloroplasts, with or without the aid of prey nuclei, resulting in a phototrophic trophic niche, and a low requirement of prey ingestion. The main disadvantage is that the SNCMs depend upon a single prey type, and when prey is unavailable, the cell most likely starves. SNCMs are generally, unlike GNCMs, resilient to prey starvation. The current literature has provided

9 us with many insights, especially since it became possible to grow these in the laboratory (e.g. Dinophysis spp. (Park, Kim et al. 2006).

Study organisms and organelle retention Dinoflagellates Dinoflagellates belong to the Alveolates and the group comprises > 2,200 species distributed globally. The large majority of species are marine (90%), but some species are also found in freshwater (Gómez 2012). About half of all dinoflagellates are phototrophic and with “built-in” chloroplasts, but it has been speculated that most of these phototrophic dinoflagellates are in fact constitutive mixotrophs (CMs) (Hansen & Tillmann, in press). Dinoflagellates have been shown to harbour several different types of chloroplasts. The built-in chloroplasts include at least the peridinin-type, fucoxanthin-type and carotene-type chloroplasts (Delwiche 1999). Some dinoflagellates harbour other sorts of chloroplasts via permanent endosymbiosis, like the pedinophyte endosymbionts in the green Noctiluca (refs) and the diatom endosymbionts in most Kryptoperidinum spp. (refs). Dinoflagellates can also harbor temporary and often reduced endosymbionts, like the sequestration of cryptophytes in the Nusuttodium spp., or retention of just the chloroplasts in Dinophysis spp. (sensu stricto). Not surprising, NCMs are found within many genera of dinoflagellates (Hansen and Tillmann, in press). Dinophysis acuminata The genus Dinophysis (sensu stricto) consists of many species with the characteristic wing (enlargement of longitudinal furrow). It is believed to consist of mixotrophic (NCMs) species, all found in marine waters going from tropical to artic waters. Dinophysis spp. are known to produce toxins, diarrhetic shellfish toxins (DSTs) and to cause DSP – a syndrome associated with human consumption of intoxicated shellfish. DSP is a major public health and economic problem for many countries and is among the most important and widely distributed of the harmful algal bloom (HAB) associated poisoning syndromes (Van Dolah 2000). In the case of D. acuminata, the chloroplasts seem to derive from the cryptophyte Teleaulax spp. (mainly T. amphioxeia). D. acuminata is not able to ingest single Teleaulax cells due to their small size. Instead, D. acuminata ingest the chloroplasts from the ciliate Mesodinium rubrum. The contents of the ciliate is sucked out using a peduncle (a feeding tube). The chloroplasts are quickly separated from the other organelles and prey cytoplasm, and the chloroplasts undergo some quick morphological changes and are distributed into the different chloroplast centers inside the D. acuminata. Recently, the division of the kleptochloroplasts was shown for D. acuminata. However, when subjected to prey starvation, D. acuminata cells will lose the pigmentation with time (Hansen, Nielsen et al. 2013, Rusterholz, Hansen et al. 2017).

Figure 4. Photos of dinoflagellates. A) Dinophysis acuminata (picture by GarciaCuetos et al. (2010)) and B) Nusuttodinium aeruginosum (picture from paper I)

Nusuttodinium aeruginosum Nusuttodinium is a genus of marine- and fresh water dinoflagellates that was described in 2014. The genus consists of 6 non-constitutive mixotrophic species, including former Gymnodinium and Amphidinium species (Takano, Yamaguchi et al. 2014). N. aeruginosum (previously G. aeruginosum; Stein 1883) is a large dinoflagellate that vary in size (30 – 80 µm in diameter). It contains blue-green chloroplasts giving it blueish appearance that stands out in natural samples. The cell retains 6-10 chloroplasts that are obtained by feeding on the cryptophytes Chroomonas sp or Komma sp, (Paper I); other cell organelles are ingested as well. N. aeruginosum cells were shown to survive without prey for > 30 days, but the cells, like D. acuminata, became pale after some time without access to prey. In this thesis, we will consider N. aeruginosum as a SNCM, based on the findings in Paper I, although more studies are needed to verify this. If N. aeruginosum is in fact a SNCM, it will represent the first and only example of a naked dinoflagellate with this trophic mode. Ciliates Ciliates also belong to the Alveolates and consist of > 4,500 species (Foissner and Hawksworth 2009). It is a very morphological diverse group that covers a large size span (from 10 µm – 4 mm in length). The common features of ciliates are the cortex, the numerous cilia, complex nuclei – nuclear dualism, and a conjugation stage during the sexual phase of the life cycle. Ciliates never have “built-in” chloroplasts, but acquired phototrophy have been shown to occur in at least 8 major ciliate taxa (Heterotrichea, Hypotrichia, Oligotrichida, Stichotrichia, Litostomatea, Prostomatea, Peniculia, Peritrichia). In 7 of the 8 ciliate taxa, phototrophy is usually acquired by algal endosymbiosis, but in the Oligotrichida it is usually by plastid retention (Stoecker, Johnson et al. 2009). Mixotrophic ciliates can average 30-92% of the total ciliate population in the photic zone (Stoecker, Michaels et al. 1987). Mesodinium rubrum and M. major Mesodinium spp are by far the most studied mixotrophic ciliates. This genus consist of species investing in heterotrophic, mixotrophic and almost completely phototrophic trophic modes. Mesodinium spp are very common and found globally, in both fresh and marine waters. The marine

11 species can be divided into primarily benthic species (M. chamaeleon, M. coatsi, M. pulex and M. pupula) and generally pelagic species (M. rubrum and M. major; Hansen, Nielsen et al. 2013). The red pelagic species have in the recent years, been referred to as the M. rubrum/M. major complex, due to difficulties on separating the two species (Fig. 5). The original description of M. rubrum was by Lohmann (1908), whereas M. major was described recently (Garcia-Cuetos, Moestrup et al. 2012). M. rubrum has often been listed as the species behind the blooms, however, recent studies have shown that in fact some or even most of these blooms may be due to M. major (Johnson, Beaudoin et al. 2016). A common feature for the M. rubrum/M. major complex is their red color, caused by their cryptophyte prey within the Teleaulax/Plagioselmis/Geminigera clade. The red Mesodinium spp often cause of massive non-toxic red tides world-wide in costal marine waters (Lindholm 1985, Hansen, Nielsen et al. 2013). Recent studies have shown that M. rubrum/M. major complex group comprises eight genetic clades (Johnson, Beaudoin et al. 2016), and it is still unsolved how many species are indeed valid species. M. rubrum and M. major are, despite of their size difference, very similar in their morphology and organization of organelles retained (Garcia-Cuetos, Moestrup, & Hansen, 2012).

Figure 5. M. major (A – B) and M. rubrum (C-D), showing their overall differences (size and nuclei arrangement (without CPN)). M. rubrum/M. major possess three nuclei (Fig. 5 C and F), but retains, together with the chloroplast also the prey nucleus, prey nucleomorph, prey mitochondria, prey ribosomes and prey cytoplasm. One of the ingested prey nuclei is positioned close to the ciliates nuclei, and converted into a “symbiont or master nucleus”, also referred to as a centered prey nucleus (CPN). Extra prey nuclei are also obtain when prey is plentiful. When M. rubrum divides, one daughter cell maintains the CPN, while the other daughter cell changes one of the extra prey nuclei into the CPN (Paper III).

A correlation between photosynthetic activity and the CPN was found, recently (Kim et al 2017). This correlation was confirmed when well-fed and starved (<10% of the cells contained CPN) M. rubrum were compared using gene expression (Appendix I). This study showed that photosynthetic genes are only expressed when M. rubrum contains CPN. A similar arrangement of ciliate and prey nuclei was found for M. major in Paper II. Thus, the major morphological differences between M. rubrum and M. major is thus size and the number of chloroplasts retained (Fig. 5 B and E). These difference naturally lead to differences in physiological performances. However, this study (Paper II) represents the only study on M. major done in the laboratory so far. All other laboratory studies have been done with different strains of M. rubrum.

Eco-Physiology of SNCMs Growth responses to different irradiance Common for SNCMs are a strong dependence on irradiance for growth. Whereas heterotrophic species can survived in the dark, as long as given prey, SNCMs do not seem able to grow in the dark, even if fed prey in excess. Thus, the growth responses as a function of irradiance are quite similar to phototrophs (Fig. 6). Some SNCMs can sustain themselves down to very low light levels as has been shown for an Arctic strain of M. rubrum, which was shown to sustain itself at irradiance < 5 µE (Moeller, Johnson et al. 2011).

Figure 6. Simplified conceptual drawing showing the overall difference between the generalists and the specialists in relation to irradiance.

Photosynthetic activity and chlorophyll a Photosynthetic activity is of great importance to SNCM and can constitute > 95% of the cells carbon needs (Smith and Hansen 2007). The level of photosynthetic activity as a function of irradiance can be fitted to Michaelis-Mentens kinetics with activity being saturated, for most studies > 200 µmol photon m-2 s-1. Both Mesodinium spp. and Dinophysis spp. have been shown to photoregulate, hence produce chl a (Hansen, Ojamäe et al. 2016). However, only M. rubrum have been shown to photoacclimate, when

13 subjected to low irradiance, by increasing chlorophyll a concentration in the cell (Moeller, Johnson et al. 2011). The study of M. major also indicated photoacclimation (Paper II). Chl a concentrations in N. aeruginosum was not measured, so we can only speculate as to whether or not it has the ability to photoregulate, but the responses on different light intensities, found for N. aeruginosum combined with their tolerance to prey deprivation could indicate a very close relationship to the chloroplasts, hence indications of photoregulation (Paper I). Why this difference between Mesodinium spp. and Dinophysis spp. exists and if this could represent a general difference among dinoflagellates and ciliates, is presently unknown. SNCMs and their dependence on their prey The ingestion of prey by SNCMs in cultures seems to be stable, and often maximum growth rates are obtained at very low prey ingestion rates. In the case of M. rubrum, the ingestion of only ~ 1 prey cell-1 d-1 is enough (Smith and Hansen 2007). Ingestion rates were not measured properly for N. aeruginosum, but rough estimates suggested < 2 prey cells-1 d-1. M. major was shown to have and ingestion rate at 5-10 prey cells-1 d-1. Ingestion rates will of course depending upon environmental conditions and prey concentrations, but ingestion rates did not seem to be affected much by the irradiance in these 2 species (Paper I and II). Ingestion rates were though greatly affected by periods of prey deprivation. High initial ingestion rates were observed when prey was re-introduced after long time prey starvation (Paper II and III). This suggests a quick recovery of phototrophy in the natural environment, when prey availability increases. When SNCMs are subjected to starvation, growth and inorganic carbon uptake both decrease, as was shown with M. rubrum (Park, Park et al. 2008, Hansen, Nielsen et al. 2013). M. rubrum typically + divide up to four times when subjected to prey starvation. However, when after the addition of NH4 to the M. rubrum culture, an enhanced number of cell divisions was observed. In fact, up to eight cell divisions were observed following prey depletion in these cases (Paper V). Light had the opposite effect on the number of cells divisions in M. rubrum during prey starvation, with less cell divisions and greater loss of cells in culture. Also, the photosynthetic activity was quickly reduced at high irradiance, indicating a degradation of the chloroplasts at high light.

Nutrients Major nutrients of phototrophic protists All living organism needs nutrients to exist. Heterotrophic organisms acquire nutrient by eating and digesting organic material. Autotrophic organisms on the other hand utilize light as an energy source taking up inorganic nutrients producing organic molecules. For aquatic phototrophs inorganic nitrogen (N) and phosphorous (P) are the major macronutrients required for growth. Micronutrients are naturally of great importance, especially iron (Fe) and vitamins have been shown to be limiting factors for growth. These are, however, rarely growth-limiting factors and the focus in this thesis will therefore be on macronutrients, in particular N.

N-forms and metabolism Inorganic nitrogen - The inorganic N source, of greatest importance for growth, is the oxidized form nitrate (NO3 ) and + the reduced form ammonium (NH4 ) (Ryther and Dunstan 1971, Goldman, McCarthy et al. 1979). + NH4 is generally considered the preferred form of N, due to lower energy requirements for the cell and a more easy transport over the cell membrane (Fig. 7; (Raven, Wollenweber et al. 1992, Berges 1997).

Figure 7. Simpliﬁed conceptual relationship of a generic diatom cell and the major metabolic pathways - + of uptake and assimilation of NO3 and NH4 (Adapted from Glibert et al. (2016)).

+ - - Once in the cytoplasm NH4 can be used directly, whereas for NO3 is reduced first into NO2 and + - - secondly into NH4 (fig. 7). The reduction of NO3 to NO2 takes place in the cytosol, for most - phototrophs, but for some dinoflagellates NO3 needs to be transported into the plastids before - + reduction (Berges and Mulholland 2008). The reduction of NO2 in to NH4 occurs in the chloroplasts. + - This difference in assimilation of NH4 can result in a delayed uptake of NO3 , most often caused by - - repression of NO3 transport over the cell membrane or inhibition of NO3 reduction in the cell (Fig. 8A; (Berges 1997, Vergara, Berges et al. 1998, Glibert, Wilkerson et al. 2016). Despite the preference + + of NH4 for most phototrophs, NH4 is rarely used in cultures in the laboratory due to its toxicity at

+ high ambient concentrations. Ambient concentrations of NH4 > 50 µM can be toxic to as least some + + phototrophs (Fig. 8B). When a cell has excess amount of NH4 , the uptake of NH4 typically stops. - This is unlike NO3 that can be present in ambient concentration of ~ 800 µM, like in the f/2 medium, often used in marine cultures (Fig. 8C).

- + Figure 8. Conceptual drawings of uptake kinetics of NO3 and NH4 A) Uptake of N as a + - function of time, showing the repression by NH4 on the uptake of NO3 , B) Uptake as a + function of ambient concentration of NH4 and C) Uptake as a function of ambient - concentration of NO3 (Adapted from McCarthy (1981) and Glibert et al. (2016)).

Inorganic nitrogen is contributing to eutrophication worldwide, in particular in coastal marine systems causing major blooms of harmful algae species (Burkholder, Glibert et al. 2008) and most likely also for NCMs, such as M. rubrum and D. acuminata (Wilkerson and Grunseich 1990, Hattenrath- Lehmann, Marcoval et al. 2015).

Organic nitrogen Organic nitrogen (N) differs from inorganic N simply by containing carbon (C). Historically, dissolved organic nitrogen (DON) was believed to be unavailable for phytoplankton, but this have been proven wrong many time since (sensu Dugdale and Goering 1967). Many forms of organic N, such as nucleic acids and amino sugars, are of great importance, but in present thesis, I have focused on urea. The ambient concentration of urea has been increasing since the 60ies and can constitute up to 50 % of the oceanic ambient N pool (Glibert, Burkholder et al. 2006). Urea as a N source, is broken + down to NH4 and CO2, and has been shown to solemnly sustain growth of some phototrophic protists e.g. dinoflagellates, diatoms and haptophytes (Solomon, Collier et al. 2010), in fact, McCarthy et al. + - (1977) suggested the preference of N-source, for most phototrophic protists, is NH4 > urea > NO3 . Uptake of nutrients by NCM In general very little is known about the uptake or utilization of nutrients by non-constitutive mixotrophs (NCMs). Besides a single study on Strombidium spp. (Schoener and McManus 2017) the only organisms subjected to investigations are Mesodinium rubrum and Dinophysis acuminata (Hattenrath‐Lehmann and Gobler 2015, Tong, Smith et al. 2015).

Mesodinium rubrum Mesodinium rubrum has been shown to take up inorganic N and urea in many field studies (Wilkerson and Grunseich 1990). There exist a general agreement in literature that M. rubrum takes up all forms

16 of N, but the conclusion has primarily based on field studies. Only three studies exist to date that have investigated uptake rates of nutrients in the laboratory (Paper V; Paper VI; Tong, Smith et al. 2015). - + These studies have shown that M. rubrum is able to take up NO3 , NH4 and urea at high uptake rates, - + comparable to phototrophic protists. The uptake of NO3 and NH4 was shown to enhance growth of M. rubrum. Also, the uptake of N was not affected by time after prey deprivation.

Dinophysis spp. Dinophysis acuminata has, similar to M. rubrum, been investigated in many field studies, but the results from these studies have been contradictory. The most recent field study concluded that blooms + of D. acuminata was promoted by NH4 (Hattenrath-Lehmann, Marcoval et al. 2015). Most studies + - have found an uptake of NH4 , but uptake of NO3 has also been indicated. Recently several laboratory + studies have confirmed that D. acuminata can take up NH4 and utilize the N for both increased - growth and photosynthetic activity. The potential to take up NO3 was also shown (but only when no + NH4 was added), but at lower uptake rates, and D. acuminata was unable to convert N into growth (Hattenrath‐Lehmann and Gobler 2015). It seems that D. acuminata is missing the necessary enzyme - + to convert NO3 into NH4 , but further studies are needed to clarify this (Paper V; Paper VI).

Abbreviations Chl a Chlorophyll a CCN complex Cryptophyte-Ciliate nucleus complex CM Constitutive mixotroph CPN Centered prey nucleus DSP Diarrhetic shellfish poisoning EPN Extra prey nucleus eSNCM Endosymbiotic specialist non-constitutive mixotroph GNCM Generalist non-constitutive mixotroph HAB Harmful algae bloom NCM Non-constitutive mixotroph NR Nitrate reductase pSNCM Plastidic specialist non-constitutive mixotroph SNCM Specialist non-constitutive mixotroph

Aims of this thesis My overall aim was to study the functional biology of protists engaging in acquired phototrophy (i.e., the use of chloroplasts from ingested photosynthetic prey), and to evaluate in which ways this mode of nutrition is advantageous. My work centered around 3 main topics:

1) Effects of light on growth, photosynthesis and ingestion of some common ciliates and dinoflagellates without permanent chloroplasts (Papers I and II)

2) The dynamics of sequestered chloroplasts and other ingested cell organelles in eSNC- mixotrophs during prey starvation and refeeding. How important are stolen prey nuclei for the photosynthetic performance and cell growth (Papers III and IV)?

3) The role of inorganic nitrogen and urea uptake in Mesodinium and Dinophysis play (Papers V and VI)

Specific aims of each paper Paper I Effects of irradiance and prey deprivation on growth, cell carbon and photosynthetic activity of the freshwater kleptoplastidic dinoflagellate Nusuttodinium (= Gymnodinium) aeruginosum (Dinophyceae) The aims were to: 1) examine growth and feeding responses as a function of irradiance and 2) investigate the role of photosynthetic activity of the retained chloroplasts during prey deprivation.

Paper II Physiological responses of Mesodinium major to irradiance, prey concentration and prey starvation The aims were to: 1) examine the responses to different irradiances of M. major with focus on growth, photosynthetic activity, cellular Chl a and ingestion rates, 2) investigate effects of prey deprivation for 50 days on the loss of the centered prey nucleus, 3) examine refeeding and recovery after prey starvation, and 4) discuss the physiological similarities and differences between M. rubrum and M. major.

Paper III Dynamics of sequestered cryptophyte nuclei in Mesodinium rubrum during starvation and refeeding The aims were to: 1) investigate the fate of prey nuclei sequestered by M. rubrum during prey starvation and refeeding experiments, 2) determine changes in size and position of prey nuclei inside the ciliate, 3) test evidence of nuclear division or fusion, and 4) study the relationship between the presence of a retained prey nucleus and the photosynthetic efficiency and growth of M. rubrum.

Paper IV Mesodinium rubrum: The symbiosis that wasn’t The aims were to: discussion the suggestion by Qui et al. (2016) that M. rubrum farms symbiotic T. amphioxeia rather than “farming” cryptophyte organelles through continuous prey ingestion.

Paper V Effects of nitrogen on two non-constitutive mixotrophs, Mesodinium rubrum and Dinophysis acuminata I: Effects of ammonium concentrations on growth, inorganic carbon uptake and removal of different inorganic nitrogen forms as a function of time after prey deprivation The aims were to: 1) investigate the ability of Mesodinium rubrum and Dinophysis + - acuminata to take up NH4 and NO3 as a function of time after prey deprivation at + - different NH4 and NO3 concentrations, and 2) examine cell concentration and carbon + - assimilation rate at four different initial NH4 concentrations and two NO3 concentrations.

Paper VI Effects of nitrogen on two non-constitutive mixotrophs, Mesodinium rubrum and Dinophysis acuminata II: Potential for uptake of ammonium, nitrate and urea as a function of time after prey deprivation - The aims were to: 1) examine if NO3 is taken up and incorporated in to growth by M. rubrum and D. acuminata, 2) investigate if time of prey deprivation will effect potential uptake rates of nitrogen by M. rubrum and D. acuminata, (3) test if initial concentrations - + of NO3 and NH4 impact the potential uptake by M. rubrum and D. acuminata, and 4) investigate if the C:N uptake ratio will change as a function of time after prey deprivation for M. rubrum and D. acuminata.

Results and discussion The role of sequestered chloroplasts Cell growth and irradiance The dinoflagellate, Nusuttodinium aeruginosum and the ciliate Mesodinium major were both found to be highly affected by irradiance, and the relationship between growth rate and irradiance could be closely fitted to Michaelis-Menten kinetics in both cases (Fig. 9; paper I and II). Clear differences between the two species were found. The growth rate of N. aeruginosum and M. major saturated at 40 and 75 µmol photons m-2 s-1, respectively. Similar, N. aeruginosum was able to sustain itself at an irradiance of 10 μmol photons m-2 s-1, while M. major required an irradiance of 25 μmol photons m-2 s-1 to sustain itself. Despite being similar in size, an obvious difference between N. aeruginosum and M. major are the numbers of retained chloroplasts of approximately 6-10 and 40-80 cell-1, respectively. This could potentially explain some of the differences in the higher requirements for light in the latter species.

Figure 9. Light response curves for N. aeruginosum (A; paper I) and M. major (B; paper II). These light response curves found for these specialist non-constitutive mixotrophs (SNCMs) are quite similar to light response curves in purely phototrophic organisms. The literature does not offer much data for direct comparison. The general understanding of the species studied of genus Nusuttodinium is that they are generalist non-constitutive mixotrophs (GNCMs). Several species have been shown to grow in complete darkness, when supplied with prey (i.e. Nusuttodinium poechilocroum and N. gracilentum; (Jakobsen, Hansen et al. 2000). The genus of Mesodinium contains the entire mixotrophic spectrum, from almost completely heterotrophs (M. pulex and M. pupula) to the almost phototrophic “red” Mesodinium spp. Only the “red” Mesodinium spp have been shown to depend on light for cell growth (Tarangkoon and Hansen 2011, Hansen, Nielsen et al. 2013, Kim, Kang et al. 2019). Photosynthesis, chlorophyll a and photoacclimation Photosynthetic activity was measured in N. aeruginosum and M. major (paper I and II). For both species the photosynthetic activity increased as a function of irradiance, but the photosynthetic activity at an irradiance of 100 µmol photons m-2 s-1 was ~25 and ~225 pg C cell-1 h-1 for N.

21 aeruginosum and M. major, respectively. N. aeruginosum were found to obtain only 26% of its C from photosynthetic activity, whereas M. major obtained up to 95%. It is noteworthy that M. major does retain almost 10 times as many chloroplasts as N. aeruginosum, thus a difference in photosynthetic activity would be expected. It is then odd that N. aeruginosum cannot sustain themselves to a higher degree on photosynthetic activity. The cellular Chl a concentration was only measured for M. major, but the results from this study indicated that M. major was able to photoacclimate, indicating a close relation with the sequestered chloroplasts. The chloroplasts seem well integrated in the cell (paper I). Ingestion of prey The results of photosynthetic activity by N. aeruginosum would lead us to anticipate notable ingestions rates; however this was not the case. We were unable to measure ingestion rates for N. aeruginosum in our experimental setup, but rough calculations indicated that ingestion rates were < 2 cells predator-1 d-1, hence below our detection levels in this study. The ingestion rates of M. major were much higher, on average ˜12 cells predator-1 d-1. According to the current literature, however scares, the ingestion rates vary a lot depending on the trophic mode of the organism and we would have expected lower ingestion rates for M. major also due to the high photosynthetic activity.

In paper II and III we subjected M. major and M. rubrum to prey deprivation and re-feeding at different predator: prey ratios. After long time of prey starvation, initial ingestion rates were > 10 times higher those measured in well-fed cells for both species. This underlines the necessity of proper evaluation of pre-incubation history when analyzing prey ingestion data. According to a recent conceptual review paper, one of the main differences between GNCMs and SNCMs is prey selectivity. Investigations of prey selectivity have unfortunately not been carried out on either N. aeruginosum or M. major. Most of the physiological measurements gathered so far indicate, however, that they are pSNCMs.

Integrations of the chloroplasts Prey deprivation The only NCMs that can tolerate > 30 days of prey deprivation, prior to this thesis, are M. rubrum and Dinophysis spp (Park, Park et al. 2008, Myung, Kim et al. 2013). Both N. aeruginosum and M. major could sustain > 30 days of prey deprivation (paper I and II). They both showed great tolerance to prey starvation and went through 4-6 cell divisions before growth eventually stopped. The chloroplasts were not enumerated, but according to their photosynthetic activity the results indicated that both species were able to divide the chloroplasts, just like Dinophysis spp and M. rubrum (Rusterholz, Hansen et al. 2017). The tolerance to prey deprivation indicates that the chloroplasts of both N. aeruginosum and M. rubrum well integrated. This is in contrast with the GNCM where e.g. Strombidium rassoulzadegani have been found to decrease in cell numbers immediately after prey deprivation (Schoener and McManus 2012). Intermediate forms between GNCM and SNCM, such as M. chamaeleon, was shown to undergo 1 cell division and could sustain itself for > 2 weeks (Moeller and Johnson 2018). These differences have immense implications the role of these

22 mixotrophic organisms in the ecosystem. The GNCMs will quickly disappear when prey availability is very low or depleted, whereas the SNCM can sustain themselves for > a month, depending on the water temperature. If this is a general phenomenon among the two NCM groups is unknown and more information is required in the future. Sequestered nuclei One of the main differences between the pSNCM dinoflagellates and ciliates are the sequestered nuclei. The ciliates, in particular M. rubrum, has been shown to obtain not only one, but several prey nuclei and speculations on whether or not M. rubrum could divide these nuclei have been ongoing. In paper III, we presented a different explanation, suggesting that the centered prey nucleus (CPN) is not divided when the cell undergoes cell division. Rather only one of the daughter cell keeps the (CPN), while the other daughter cell relocate and transforms an extra prey nucleus (EPN) into a CPN. Paper II looked for a similar pattern for M. major, and though CPNs were found within M. major (Fig. 5C), suggesting the same arrangement, the EPNs were not located. Due to high ingestion rate, we firmly believe that M. major retains EPNs, but that these were difficult to locate due to the intense fluorescence signal from the numerous chloroplasts.

Figure 10. Prevalence of a centered prey nucleus (CPN) and number of chloroplasts cell-1 for M. rubrum plotted as a function of photosynthetic rate (Paper III). The chloroplasts were enumerated in M. rubrum during prey deprivation to investigate a possible correlation with photosynthetic activity (paper III). Such a relationship was however not found. A clear correlation, however, between the photosynthetic activity and the presence of a CPN was found for both M. rubrum and M. major (Fig 10; paper II and III). This suggests that the CPN has active DNA that supports photosynthetic activity. Appendix I investigated this and found no photosynthetic

23 genes were transferred to the host genome, but that M. rubrum have considerable control over gene expression of the acquired T. amphioxeia nucleus. This confirms the theory from paper III.

- Traditionally all experiments on M. rubrum are carried out in NO3 -based growth medium (like f/2). + In paper V, an experiment with the additions of NH4 was carried out. The results showed that M. rubrum was able to undergo > 8 cell divisions, when subjected to prey starvation. The numbers of CPNs were followed simultaneously, and the results indicated that either the CPNs are in fact divided within M. rubrum or additional EPNs were available that we could not locate. This opens up the discussion of why M. rubrum stops functioning as a phototroph, if it can divide the chloroplasts and the CPN, why then does it have to eat? Symbiosis or sequestered Our knowledge of pSNCM is still limited. The terminology is still undergoing changes, and in many cases it is still discussed how to interpret our observations. Paper IV brings up one such problem after Qui et al. (2016) observed a M. rubrum cell with an intact prey cell, T. amphioxeia and suggested that M. rubrum “farms” T. amphioxeia as an intact endosymbiont, rather than retain its organelles. In paper IV this was questioned and refuted with the lack of cell membrane of T. amphioxeia. Here it was shown that M. rubrum can express many of the cryptophyte gene pathways even when only prey organelles remain. Together with the findings in paper III, where new CPNs can be acquired during prey deprivation, it is difficult to support the idea that T. amphioxeia is “farmed” as an intact endosymbiont. So what to call the relationship between the red Mesodinium spp and their prey? M. rubrum only sequesters the prey organelles but also the cytosol of T. amphioxeia with its ribosome (Hansen, Moldrup et al. 2012). We lack a term for this “reduced endosymbiont”, which is not really an endosymbiont. This discussion is similar to a previous debate on the origin of the chloroplasts in Dinophysis spp (Garcia-Cuetos, Moestrup et al. 2010). What was believed to be permanent cryptophyte chloroplasts with an ultrastructure that is very different from its prey, M. rubrum, turned out to be later proven wrong by (Kim, Nam et al. 2012) (see discussion in Hansen et al. 2013). It stresses the necessity of being open-minded when it comes to the functional biology of these organisms.

Effect of inorganic nitrogen Uptake of inorganic nitrogen and urea - + M. rubrum was shown to take up and deplete ambient NO3 , NH4 and Urea from the growth medium. + The uptake rates were affected by the combination of nutrients, and NH4 had a repressive effect on - the uptake of NO3 and urea. The results indicated that M. rubrum utilizes nitrogen in the same manner + as its phototrophic prey, T. amphioxeia (Papers V and VI). D. acuminata took up NH4 , but were to + + a lesser degree able to deplete NH4 from the medium, in particular at high ambient NH4 - concentrations. The NO3 was never deleted by D. acuminata in the experiments, but it was able to - + take up a fraction of the ambient NO3 , in particular when NH4 was not available; probably due to + - the repressive effect of NH4 on the uptake of NO3 and urea, as seen for M. rubrum. The uptake rates of inorganic N by M. rubrum and D. acuminata (Papers V and VI) were generally higher than previous published rates (Hattenrath‐Lehmann and Gobler 2015, Tong, Smith et al. 2015,

García-Portela, Reguera et al. 2020). We did multiple measurements of both types of cells as a function of incubation time and found some regulation of rates, while previous measurements only have been single measurements of well-fed and starved cells. Thus, time course effects may at least explain some of these differences between studies. If we assume that the chloroplasts are the uptake sites of N, then the chloroplasts have very similar + uptake rates of NH4 in T. amphioxeia, M. rubrum or D. acuminata (Fig. 11 and Paper V). The uptake - rates of NO3 on the other hand was higher in T. amphioxeia compared to that of M. rubrum or D. acuminata.

- + Figure 11. Measured maximum uptake rates of NO3 (A) and NH4 (B) by Teleaulax amphioxeia, Mesodinium rubrum and Dinophysis acuminata converted into uptake rate chloroplast-1 (pM N chloropl-1 h-1). It’s assumed that both M. rubrum and D. acuminata contain 20 chloroplasts pr. cell. Paper V

- + M. rubrum and D. acuminata showed an increased uptake of NO3 , and to some extent NH4 , with increasing ambient concentrations (paper VI). This is basic kinetics for phototrophic organisms, but these results indicate that it is also true for some pSNCMs. Assimilation of nitrogen M. rubrum was shown to assimilate both forms of inorganic nitrogen into cell growth. D. acuminata + - could utilize NH4 for growth, but could apparently not assimilate NO3 as we saw not cell divisions - when subjected to prey deprivation and NO3 as the only N-source (Fig 12; paper V). The differences in cell abundances for M. rubrum is most likely the result of higher initial N concentrations, thus the + + treatment with additions of 50 µM N-NH4 had a total of ~100 µM N (50 µM N-NH4 + 50 µM N- - NO3 ). Furthermore, D. acuminata showed a significant increase in photosynthetic activity when + supplied with NH4 , whereas this was not the case for M. rubrum nor T. amphioxeia.

Figure 12. Time course of numbers of cells during incubation M. rubrum (A) and D. acuminata (B). + + + The different treatments (control, [10] µM NH4 , [25] µM NH4 and [50] µM NH4 ) are presented in each graph indicated by symbols shown in A. Points are means while error bars represent standard error of triplicate measurements. Paper V

Prey deprivation and N-uptake - + M. rubrum was unaffected by prey deprivation in their potential uptake of NO3 , NH4 and urea for - 14 days. D. acuminata could upregulate the potential uptake of NO3 through prey deprivation, but without an effect on growth. We speculate that this could be explained by the lack of the enzuyme + nitrate reductase, but further studies are needed (paper VI). The uptake of NH4 by D. acuminata + could, to some degree, be upregulated and utilized, but the cells were either full of NH4 , causing a feedback mechanism halting the uptake or they lost the ability.

Future perspectives The terminology within NCMs is still undergoing changes and while many aspects have been clarified, the role of prey nuclei in the functioning is still not fully understood. Among the NCMs we have organisms that only sequester the prey chloroplast, often referred to as kleptoplastidic organisms (sensu stricto). At the other end of the spectrum we have protists that “farm” intact endosymbionts. In between, we have protists that have, what has been called “reduced endosymbionts”. In these cases, the ingested cells are retained and reduced. Often prey nuclei, prey mitochondria and cytoplasm with ribosomes are retained along with the prey chloroplasts, at least for some time. One of the main questions addressed in this thesis was; why do pSNCM stop functioning as phototrophic organisms after prey depletion? Despite constructive additions to answering this question, we have not been able fully answer it. To me, this question can split into four additional questions:

I. Why are the kleptochloroplasts (sensu stricto) being diluting through division SNCMs in the long run under prey depletion? We often see an initial division of the chloroplasts in mixotrophs like Dinophysis spp, but somehow this ability is lost over time of prey deprivation. Is it due to the age of the retained chloroplasts? Could it be due to lack of ability to repair the kleptochloroplasts? Or due to lack of ability to produce certain enzymes, like RuBISCo?

II. What is the exact role of the retained nuclei by Mesodinium (both CPN and EPNs)? Even though a correlation between the CPN and photosynthetic activity has been found, and the fact that all the expression of the photosynthetic genes is carried out by the sequestered prey nuclei, it seems unlikely that photosynthetic activity alone is the purpose. What role does the CPN play in chloroplast division?

III. The effects of nutrients? To answer some of the above questions, looking at nutrients could very well indicate in which - direction we should focus future research. It seems that D. acuminata lacks enzymes for NO3 + reductase, and that the ability to take up NH4 is related to time of prey deprivation, but we need more evidence. Future experiments should also study potential N-uptake in M. rubrum that has been subjected to long-term prey starvation; i.e. situations where the percentage of cells with prey nuclei is very low. How will this affect potential N-uptake?

IV. What about the GNCM? While little is known about SNCMs, even less is known about the functional biology and physiology of GNCMs. These organisms do not seem to retain prey nuclei, they seem to have a much looser relationship to their retained chloroplasts. Nevertheless, they are still able to fix inorganic carbon. Even though many of these species relatively quickly loose the chloroplasts after prey depletion, some species can sustain themselves for weeks without the addition of prey. How is that possible, when they do not retain prey nuclei? Furthermore, we know very little about inorganic nutrient uptake in GNCMs, and of the effects of nutrients on their growth? Finally, what is it that makes GNCMs different from pure heterotrophs? What determines this difference?

References

Andersen, Ken H, Dag L Aksnes, Terje Berge, Øyvind Fiksen, and Andre Visser. 2015. 'Modelling emergent trophic strategies in plankton', Journal of Plankton Research, 37: 862-68.

Berges, John A. 1997. 'Miniview: algal nitrate reductases', European Journal of Phycology, 32: 3-8.

Berges, John A, and Margaret R Mulholland. 2008. 'Enzymes and nitrogen cycling', Nitrogen in Marine Environment: 1385-444.

Burkholder, J. M., P. M. Glibert, and H. M. Skelton. 2008. 'Mixotrophy, a major mode of nutrition for harmful algal species in eutrophic waters', Harmful Algae, 8: 77-93.

Delwiche, Charles F. 1999. 'Tracing the thread of plastid diversity through the tapestry of life', the american naturalist, 154: S164-S77.

Dugdale, RC, and JJ Goering. 1967. 'Uptake of new and regenerated forms of nitrogen in primary productivity 1', Limnology and Oceanography, 12: 196-206.

Foissner, Wilhelm, and David Leslie Hawksworth. 2009. Protist diversity and geographical distribution (Springer Science & Business Media).

Garcia-Cuetos, L., O. Moestrup, and P. J. Hansen. 2012. 'Studies on the genus Mesodinium II. Ultrastructural and molecular investigations of five marine species help clarifying the taxonomy', J Eukaryot Microbiol, 59: 374-400.

Garcia-Cuetos, Lydia, Øjvind Moestrup, Per Juel Hansen, and Niels Daugbjerg. 2010. 'The toxic dinoflagellate Dinophysis acuminata harbors permanent chloroplasts of cryptomonad origin, not kleptochloroplasts', Harmful Algae, 9: 25-38.

García-Portela, María, Beatriz Reguera, Jesús Gago, Mickael Le Gac, and Francisco Rodríguez. 2020. 'Uptake of Inorganic and Organic Nitrogen Sources by Dinophysis acuminata and D. acuta', Microorganisms, 8: 187.

Glibert, Patricia M, JoAnn M Burkholder, Matthew W Parrow, Alan J Lewitus, and Daniel E Gustafson. 2006. 'Direct uptake of nitrogen by Pfiesteria piscicida and Pfiesteria shumwayae, and nitrogen nutritional preferences', Harmful Algae, 5: 380-94.

Glibert, Patricia M, Frances P Wilkerson, Richard C Dugdale, John A Raven, Christopher L Dupont, Peter R Leavitt, Alexander E Parker, JoAnn M Burkholder, and Todd M Kana. 2016. 'Pluses and minuses of ammonium and nitrate uptake and assimilation by phytoplankton and implications for productivity and community composition, with emphasis on nitrogen‐ enriched conditions', Limnology and Oceanography, 61: 165-97.

Goldman, Joel C, James J McCarthy, and Dwight G Peavey. 1979. 'Growth rate influence on the chemical composition of phytoplankton in oceanic waters', Nature, 279: 210.

Gómez, F. 2012. 'A checklist and classification of living dinoflagellates (Dinoflagellata, Alveolata)', Cicimar Oceánides, 27: 65-140.

Hansen, Per J, Lilibeth Miranda, and Rhodora Azanza. 2004. 'Green Noctiluca scintillans: a dinoflagellate with its own greenhouse', Marine Ecology Progress Series, 275: 79-87.

Hansen, Per J, Karin Ojamäe, Terje Berge, Erik CL Trampe, Lasse T Nielsen, Inga Lips, and Michael Kühl. 2016. 'Photoregulation in a kleptochloroplastidic dinoflagellate, Dinophysis acuta', Front Microbiol, 7: 785.

Hansen, Per Juel, Morten Moldrup, W Tarangkoon, L Garcia-Cuetos, and Øjvind Moestrup. 2012. 'Direct evidence for symbiont sequestration in the marine red tide ciliate Mesodinium rubrum', Aquatic Microbial Ecology, 66: 63.

Hansen, Per Juel, Lasse Tor Nielsen, Matthew Johnson, Terje Berge, and Kevin J Flynn. 2013. 'Acquired phototrophy in Mesodinium and Dinophysis -A review of cellular organization, prey selectivity, nutrient uptake and bioenergetics', Harmful Algae, 28: 126-39.

Hattenrath-Lehmann, Theresa K, Maria A Marcoval, Heidi Mittlesdorf, Jennifer A Goleski, Zhihong Wang, Bennie Haynes, Steve L Morton, and Christopher J Gobler. 2015. 'Nitrogenous nutrients promote the growth and toxicity of Dinophysis acuminata during estuarine bloom events', PloS one, 10: e0124148.

Hattenrath‐Lehmann, Theresa, and Christopher J Gobler. 2015. 'The contribution of inorganic and organic nutrients to the growth of a North American isolate of the mixotrophic dinoflagellate, Dinophysis acuminata', Limnology and Oceanography, 60: 1588-603.

Jakobsen, Hans Henrik, Per Juel Hansen, and Jacob Larsen. 2000. 'Growth and grazing responses of two chloroplast-retaining dinoflagellates: effect of irradiance and prey species', Marine Ecology Progress Series, 201: 121-28.

Johnson, Matthew D., David J. Beaudoin, Aitor Laza-Martinez, Sonya T. Dyhrman, Elizabeth Fensin, Senjie Lin, Aaron Merculief, Satoshi Nagai, Mayza Pompeu, Outi Setälä, and Diane K. Stoecker. 2016. 'The Genetic Diversity of Mesodinium and Associated Cryptophytes', Frontiers in Microbiology, 7.

Kim, Miran, Misun Kang, and Myung Gil Park. 2019. 'Growth and Chloroplast Replacement of the Benthic Mixotrophic Ciliate Mesodinium coatsi', Journal of Eukaryotic Microbiology, 66: 625- 36.

Kim, Miran, Seung Won Nam, Woongghi Shin, D Wayne Coats, and Myung Gil Park. 2012. 'Dinophysis caudata (dinophyceae) sequesters and retains plastids from the mixotrophic ciliate prey mesodinium rubrum', J Phycol, 48: 569-79.

Lindholm, Tore. 1985. 'Mesodinium rubrum-a unique photosynthetic ciliate', Advances in aquatic microbiology, 3: 1-48.

Mitra, Aditee, Kevin J Flynn, Urban Tillmann, John A Raven, David Caron, Diane K Stoecker, Fabrice Not, Per J Hansen, Gustaaf Hallegraeff, and Robert Sanders. 2016. 'Defining planktonic protist functional groups on mechanisms for energy and nutrient acquisition: incorporation of diverse mixotrophic strategies', Protist, 167: 106-20.

Moeller, Holly V, and Matthew D Johnson. 2018. 'Preferential plastid retention by the acquired phototroph Mesodinium chamaeleon', Journal of Eukaryotic Microbiology, 65: 148-58.

Moeller, Holly V, Matthew D Johnson, and Paul G Falkowski. 2011. 'Photoacclimation in the phototrophic marine ciliate Mesodinium rubrum (Ciliophora)', J Phycol, 47: 324-32.

Myung, Geumog, Hyung S Kim, Jong Woo Park, Jong Soo Park, and Wonho Yih. 2013. 'Sequestered plastids in Mesodinium rubrum are functionally active up to 80 days of phototrophic growth without cryptomonad prey', Harmful Algae, 27: 82-87.

Park, Myung Gil, Sunju Kim, Hyung Seop Kim, Geumog Myung, Yi Gu Kang, and Wonho Yih. 2006. 'First successful culture of the marine dinoflagellate Dinophysis acuminata', Aquatic Microbial Ecology, 45: 101-06.

Park, Myung Gil, Jong Soo Park, Miran Kim, and Wonho Yih. 2008. 'Plastid dynamics during survival of Dinophysis caudata without its ciliate prey', Journal of phycology, 44: 1154-63.

Raven, John A, Bernd Wollenweber, and Linda L Handley. 1992. 'A comparison of ammonium and nitrate as nitrogen sources for photolithotrophs', New Phytologist, 121: 19-32.

Rusterholz, Pernille Møller, Per Juel Hansen, and Niels Daugbjerg. 2017. 'Evolutionary transition towards permanent chloroplasts?-Division of kleptochloroplasts in starved cells of two species of Dinophysis (Dinophyceae)', PloS one, 12: e0177512.

Ryther, John H, and William M Dunstan. 1971. 'Nitrogen, phosphorus, and eutrophication in the coastal marine environment', Science, 171: 1008-13.

Schnepf, Eberhard, and Malte Elbrächter. 1992. 'Nutritional strategies in dinoflagellates: a review with emphasis on cell biological aspects', European Journal of Protistology, 28: 3-24.

Schoener, DM, and GB McManus. 2012. 'Plastid retention, use, and replacement in a kleptoplastidic ciliate', Aquatic Microbial Ecology, 67: 177.

Schoener, Donald M, and George B McManus. 2017. 'Growth, grazing, and inorganic C and N uptake in a mixotrophic and a heterotrophic ciliate', Journal of Plankton Research, 39: 379-91.

Smith, Morten, and Per Juel Hansen. 2007. 'Interaction between Mesodinium rubrum and its prey: importance of prey concentration, irradiance and pH', Marine Ecology Progress Series, 338: 61-70.

Solomon, Caroline M, Jackie L Collier, Gry Mine Berg, and Patricia M Glibert. 2010. 'Role of urea in microbial metabolism in aquatic systems: a biochemical and molecular review', Aquatic Microbial Ecology, 59: 67-88.

Stein, FR. 1883. 'Der Organismus der Infusionsthiere. III Abt', Der Organismus der Arthodelen Flagellaten. II. Hälfte. Die Naturgeschichte der Arthrodelen Flagellaten. Einleitung und Eklärung der Abbildungen. Wilheim Engelmann, Leipzig, 25.

Stoecker, Diane K, Per Juel Hansen, David A Caron, and Aditee Mitra. 2017. 'Mixotrophy in the marine plankton', Annual Review of Marine Science, 9: 311-35.

Stoecker, Diane K, Matthew D Johnson, Colomban de Vargas, and Fabrice Not. 2009. 'Acquired phototrophy in aquatic protists', Aquatic Microbial Ecology, 57: 279-310.

Stoecker, Diane K, Ann E Michaels, and Linda H Davis. 1987. 'Large proportion of marine planktonic ciliates found to contain functional chloroplasts', Nature, 326: 790-92.

Takano, Yoshihito, Haruyo Yamaguchi, Isao Inouye, Øjvind Moestrup, and Takeo Horiguchi. 2014. 'Phylogeny of five species of Nusuttodinium gen. nov.(Dinophyceae), a genus of unarmoured kleptoplastidic dinoflagellates', Protist, 165: 759-78.

Tarangkoon, Woraporn, and Per Juel Hansen. 2011. 'Prey selection, ingestion and growth responses of the common marine ciliate Mesodinium pulex in the light and in the dark', Aquatic Microbial Ecology, 62: 25-38.

Tong, Mengmeng, Juliette L Smith, David M Kulis, and Donald M Anderson. 2015. 'Role of dissolved nitrate and phosphate in isolates of Mesodinium rubrum and toxin-producing Dinophysis acuminata', Aquatic Microbial Ecology, 75: 169.

Van Dolah, Frances M. 2000. 'Marine algal toxins: origins, health effects, and their increased occurrence', Environmental health perspectives, 108: 133-41.

Vergara, Juan J, John A Berges, and Paul G Falkowski. 1998. 'Diel periodicity of nitrate reductase activity and protein levels in the marine diatom Thalassiosira weissflogii (Bacillariophyceae)', Journal of phycology, 34: 952-61.

Wilcox, Lee W, and Gary J Wedemayer. 1984. 'Gymnodinium acidotum Nygaard (Pyrrophyta) a dinoflagellate with an endosymbiotic cryptomonad', Journal of phycology, 20: 236-42.

Wilkerson, Frances P, and Gary Grunseich. 1990. 'Formation of blooms by the symbiotic ciliate Mesodinium rubrum: the significance of nitrogen uptake', Journal of Plankton Research, 12: 973-89.

Acknowledgements

Per Juel Hansen! Basically, I could just stop there, for you are the reason this has come together. Thank you for your inspirational curiosity, and your never-ending enthusiasm. You have the powers to create a unique work environment, where, despite different backgrounds and experiences, you make us all feel equal, and that we all have something to contribute. You know when to push and when to support and despite being under a continued work overload, you always make time for the big and small problems. Thank you for believing in me and making me believe in me.

Where good people are, good people will follow and you, Per have surrounded yourself with a committed work group. Sofie, Hannah and Ruth – aka the baby-bunch. Thank you for your friendship, enlightening talks, coffee breaks, for listening to complaints about babies, you’re wise advices and Sofie, for your help in the end; that meant a lot –and of course for the lucky lemon. Also, a big thank you to everyone in Pers group, Nicola, Andreas, Per, Erin and finally Maira, my fellow NCM lover/hater, I wish we could have done even more studies together.

A huge thank you to my dear colleges and post docs on the same project; Andreas and Miran. Thank you, Miran for showing and teaching me about the fantastic world of epiflourescence microscopy. I have the deepest admiration and respect for your work. You were the funniest teacher I have ever had and it always makes me smile when I think of you. Andreas, thank you for being my go-to guy when I did not understand the power of transcriptomes, for your gift for explaining your research so the rest of us could follow and for always being the greatest optimist.

To Patricia M. Glibert and Kevin J. Flynn, thank you for your patience and for introducing me to basic kinetics of N uptake and the terrible truth that nothing is just simple. Pat, thank you for following Per enthusiasm and designing a comprehensive experiment that took more than three years to execute. On my travels to further investigate NCM I had the pleasure to meet Johan Decell. Thank you for introducing me to NanoSIMS and even though our collaboration did not produce a paper I like to think that we started something. On another journey, I had the pleasure to travel and work together with Kristine, Torkel, Lorenz and Josephine, I could not have asked for a better companionship and I will be forever grateful for the chance to go on a cruise on Sanna in the fjords of Ilulissat. Thank you for a lovely trip and for adding a completely different approach to the world of protists.

Thank you to all at UP4. Øjvind, I will always be grateful to you for introducing me to dinoflagellates, you are a living Wikipedia. Thank you, Niels for your moral support and your expertise in microscopy. Thank you, Brett for always being a sweetheart and to 2*Pernille, for making me feel normal in my love of Dinos and paranoid approach to them.

Thank you to everyone at MBS for making it was it is -a unique, safe little piece of heaven. I might be feeling a bit sentimental, but that is what MBS is to me. The technical staff, Sofie, Helle, Marianne, Lilian, Jens, Rikke, Jacob and Thomas, thank you for being the backbone of the institute, and for all your continuous help, your company is always refreshing. A special thank you to Sofie for being the perfect office mate, a friend, and for being available to help out when needed.

To my friends and family, without your support this would have never happened! Thank you for your endless support and encouragement. Thank you for listening to nonsense about my tiny, perfect babies and letting me persuade you that they are fantastic and beautiful. Thank you for helping out every time experiments were running out of hand and during the final stages of the making of this thesis. To Anne and the boys, you keep me grounded and you are the reason that I look forward to coming home -every day. I love you.

Papers

RESEARCH ARTICLE Effects of irradiance and prey deprivation on growth, cell carbon and photosynthetic activity of the freshwater kleptoplastidic dinoflagellate Nusuttodinium (= Gymnodinium) aeruginosum (Dinophyceae)

Kirstine Drumm*, Mette Liebst-Olsen, Niels Daugbjerg, Øjvind Moestrup, Per Juel Hansen a1111111111 Marine Biological Section, Department of Biology, University of Copenhagen, Copenhagen, Denmark a1111111111 * [email protected] a1111111111 a1111111111 a1111111111 Abstract

The freshwater dinoflagellate Nusuttodinium aeruginosum lacks permanent chloroplasts. Rather it sequesters chloroplasts as well as other cell organelles, like mitochondria and OPEN ACCESS nuclei, from ingested cryptophyte prey. In the present study, growth rates, cell production Citation: Drumm K, Liebst-Olsen M, Daugbjerg N, and photosynthesis were measured at seven irradiances, ranging from 10 to 140 μmol pho- Moestrup Ø, Hansen PJ (2017) Effects of tons m-2s-1, when fed the cryptophyte Chroomonas sp. Growth rates were positively influ- irradiance and prey deprivation on growth, cell enced by irradiance and increased from 0.025 d-1 at 10 μmol photons m-2s-1 to maximum carbon and photosynthetic activity of the -1 -2 -1 freshwater kleptoplastidic dinoflagellate growth rates of ~0.3 d at irradiances 40 μmol photons m s . Similarly, photosynthesis Nusuttodinium (= Gymnodinium) aeruginosum ranged from 1.84 to 36.9 pg C cell-1 h-1 at 10 and 140 μmol photons m-2s-1, respectively. (Dinophyceae). PLoS ONE 12(8): e0181751. The highest rates of photosynthesis in N. aeruginosum only corresponded to ~25% of its https://doi.org/10.1371/journal.pone.0181751 own cell carbon content and estimated biomass production. The measured rates of photo- Editor: Ross Frederick Waller, University of synthesis could not explain the observed growth rates at high irradiances. Cultures of N. aer- Cambridge, UNITED KINGDOM uginosum subjected to prey starvation were able to survive for at least 27 days in the light. Received: March 20, 2017 The sequestered chloroplasts maintained their photosynthetic activity during the entire Accepted: July 6, 2017 period of starvation, during which the population underwent 4 cell divisions. This indicates Published: August 1, 2017 that N. aeruginosum has some control of the chloroplasts, which may be able to replicate. In conclusion, N. aeruginosum seems to be in an early stage of chloroplast acquisition with Copyright: © 2017 Drumm et al. This is an open access article distributed under the terms of the some control of its ingested chloroplasts. Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.

Data Availability Statement: All relevant data are within the paper and its Supporting Information Introduction files. About half the known species of free-living dinoflagellates lack built-in chloroplasts and are Funding: ND thanks the Carlsberg Foundation for referred to as heterotrophic dinoflagellates. However, during the past three decades a growing an equipment grant. This study was partly number of dinoflagellate species have been shown to exploit chloroplasts from their prey supported by the Danish Council for Independent Research, grant number 4181-00484 to PJH. The and thereby supplement their carbon needs via photosynthesis [1]. Exploitation of ingested funders had no role in study design, data collection

PLOS ONE | https://doi.org/10.1371/journal.pone.0181751 August 1, 2017 1 / 13 Effects of irradiance and prey deprivation on a freshwater kleptoplastidic dinoflagellate

and analysis, decision to publish, or preparation of chloroplasts can be found among several dinoflagellate genera and is organized in many differ- the manuscript. ent ways. Some species utilize intact algal cells as ectosymbionts (e.g. Ornithocercus [2, 3] and Competing interests: The authors have declared Amphisolenia [4]) or endosymbionts (e.g. the green Noctiluca [5–7]). In other species, only the that no competing interests exist. chloroplasts are utilized (i.e. Dinophysis [8, 9]), and genes allowing the dinoflagellate to utilize the chloroplasts have been transferred to the host genome [10]. In between these two extremes, species are found that sequester prey nuclei and mitochondria together with the chloroplasts, allowing the dinoflagellates to utilize the acquired chloroplasts for days or weeks. This ability has been reported from a number of freshwater species: Nusuttodinium acidotum, N. amphidinoides and N. aeruginosum [11–13], as well as marine species: Amphidinium wigrense, Amylax buxus, A. triacantha, Cryptoperidiniopsis sp., Gymno- dinium gracilentum, Nusuttodinium latum, N. myriopyrenoides, N. poecilochroum, Pfiesteria piscicida and a yet undescribed Antarctic dinoflagellate [14–22]. Our knowledge of the ecophysiology of these organisms is still quite sparse. The dependency of feeding and growth on irradiance has so far only been studied in two marine species G. gracilentum and N. poecilochroum [17, 23]. These species can grow in complete darkness if supplied with fresh prey. However, their growth rates depend on irradiance, and rates increase by a factor of 2–3 when at irradiances > 50 μmol photons m-2s-1. Ingestion rates are also light dependent and the rates also increase by a factor of 2–3. This indicates that increased-light-dependent digestion rates are the explanation for the observed elevations in growth rates at higher irradiances. Rates of photosynthesis are only available for G. gracilentum at a few irradiances [17] and suggest that photosynthetic activity decreases rapidly after prey starvation. Hardly any photosynthesis could be measured in prey-starved cultures after 48h. We found the species N. aeruginosum to be common in small lakes in Denmark, established it into culture and studied its ecophysiology. Species identification was based on light microscopy and a sequence comparison of the nuclear-encoded LSU rDNA gene (1400 base pairs) between our strain and a Japanese strain of N. aeruginosum studied by Onuma et al. [24] (Genbank accession number LC027055). The sequence divergence was 0.5% (data not shown). In the present work we aimed to study growth, cell production and photosynthesis of the dinoflagellate at different irradiances when fed a cryptophyte identified as Chroomonas sp. This work represents the first ecophysiological study of a freshwater kleptoplastidic species. Specifically, we examined (1) growth and feeding responses as a function of irradiance and (2) the role of photosynthetic activity of the retained chloroplasts during prey starvation.

Materials and methods Isolation and maintenance of cultures Nusuttodinium aeruginosum was isolated from a surface water sample from Horsekær, Tibirke, Denmark in July 2012. Chroomonas sp. (SCCAP K-1623) was obtained from Scandinavian Culture Collection of Algae & Protozoa. Cells were grown in L16 medium and were maintained in TPP 24 well test plates. The cultures were kept at an irradiance of 65 μmol photons m-2s-1, a light:dark cycle of 16:8 h and temperature of 15±1˚C. Irradiance was measured using a LI-COR LI-1000 radiation sensor equipped with a spherical probe.

Light microscopy Live cells (Fig 1A and 1B) were observed using an Axio imager.M2 (Zeiss, Germany) equipped with Nomarski interference contrast and a x63 oil objective (NA = 1.4). Micrographs were taken with an Axiocam HRc digital camera (Zeiss, Germany).

PLOS ONE | https://doi.org/10.1371/journal.pone.0181751 August 1, 2017 2 / 13 Effects of irradiance and prey deprivation on a freshwater kleptoplastidic dinoflagellate

Fig 1. Light microscopy of Nusuttodinium aeruginosum. (A) Large chloroplast-containing cell. (B) Small cell showing chloroplast degradation. (C) Epifluorescence microscopy revealing numerous chloroplasts. https://doi.org/10.1371/journal.pone.0181751.g001

Epifluorescence microscopy Well-fed cells, experiencing prey concentrations > 5000 Chroomonas sp. cells ml-1, were fixed in 2% glutaraldehyde in L16 medium and filtered onto a black filter with a pore size of 0.22 μm (Osmonic Inc). Kleptochloroplasts were examined with an Olympus IX81 motorized inverted microscope equipped with epifluorescence illumination and a disc-spinning unit for confocal imaging. Stacks of 36 images were taken of three individual cells using a digital camera (Soft Imaging System F View II). Chloroplasts from a 3D-reconstructed cell is shown in Fig 1C. Two experiments on N. aeruginosum were conducted. The first experiment aimed to quantify the effects of irradiance on growth, photosynthesis and cell carbon while the second experiment was carried out to study the effects of prey starvation on photosynthetic activity of N. aeruginosum. Both experiments were carried out in TPP 96 well test plates.

Experiment 1: Effects of irradiance on growth, photosynthesis and cell volume The effect of irradiance on the growth of Nusuttodinium aeruginosum was studied at seven different irradiances: 10, 20, 25, 40, 65, 100 and 140 μmol photons m-2s-1. Seven plates, each containing 12 replicates of mixed cultures and 6 replicates of mono cultures of prey were initiated at the same time at the different irradiances. Each replicate was initiated with ten well-fed N. aeruginosum which were picked individually with a drawn Pasteur pipette and transferred to a well with 100 μl L16 medium. Adding 200 μl of 500 cells ml-1 Chroomonas sp.-culture gave an initial 10:100 relationship between predator (N. aeruginosum) and prey (Chroomonas sp.). Sub- sequently one plate was fixed for every doubling time (varied between 2 and 4 days). Cells were enumerated directly from wells using an inverted microscope. For Chroomonas sp. a minimum of 200 cells was counted (if the total was less than 200, then all cells were counted). All N. aeruginosum cells were counted and cell volume measured, giving datasets of 12

PLOS ONE | https://doi.org/10.1371/journal.pone.0181751 August 1, 2017 3 / 13 Effects of irradiance and prey deprivation on a freshwater kleptoplastidic dinoflagellate

replicates for mixed and 6 replicates for mono cultures. Photosynthesis was measured once at each light intensity after approximately 4 generations (see below for details).

Experiment 2: Survival and photosynthetic activity of endosymbionts during prey deprivation The second experiment applied the same set up as experiment 1, but the initial number of Nusuttodinium aeruginosum was changed to 50 cells per well and prey was not added. Prior to the commencement of the experiment the predator was separated from the prey, as described under photosynthesis. To ensure that N. aeruginosum was in exponential growth phase when transferred, cells were taken on day eight from a 96-well plate which had been initialized as in experiment 1. The experiment was carried out at 65 μmol photons m-2s-1. Measurements of photosynthesis, cell numbers and cell volumes were done on day 0, 2, 4, 6, 9, 14 and 27 (see below for details).

Measurements of growth and cell volume The cell concentrations of Nusuttodinium aeruginosum were determined by total cell count of cells growing in 96-well plates. The growth rate (μ, d-1) was calculated using:

À 1 my ¼ ðlneðNt À lneðN0ÞÞt ð1Þ

where Nt is the number of cells at time t and t is the experimental time (days). Growth rate (μy) as a function of irradiance was fitted to Michaelis-Menten kinetics.

Cell volume (Vy) were computed using the geometric formula for a prolate spheroid (length > width = depth) on Lugol-fixed cells. Calculations of cell carbon (Cy) of N. aeruginosum were based on Menden-Deuer and Lessard (2000):

0:819 Cy ¼ 0:76 Â Vy ð2Þ

where Vy is cell volume of N. aeruginosum. Biomass production rate was computed as:

Biomass production ¼ my Â Cy ð3Þ

-1 -1 where μy is growth rate (μ, d ) and Cy is (pg C cell ). As with growth rate, both cell volume, cell carbon and biomass production as functions of irradiance were fitted to Michaelis-Menten kinetics.

Photosynthesis Photosynthetic activity was measured by a modification of the ’single-cell method’ [24, 25]. The prey Chroomonas sp. is photosynthetic, and it was therefore necessary to separate predators from prey prior to photosynthesis measurements. This was done by picking individual Nusuttodinium aeruginosum cells with a hand drawn Pasteur pipette and rinsing each cell in L16 medium. Four replicates of 23-ml glass scintillation vials were filled with 2 ml L16 14 - medium, and 50 rinsed cells were transferred to each vial. 20 μl NaH CO3 stock solution (specific activity 100 μCi ml-1) was added in each vial, resulting in a specific activity of ~1.0 μCi ml-1. The vials were left for 3 h at the matching light intensity. Vials were accompanied by dark vials, which were treated similarly, but wrapped in tin foil during incubation. After incubation the specific activity was determined by transferring 100 μl from each vial to new vials containing 200 μl phenylethylamine. The remaining sample was acidified with 2 ml 10% glacial acetic acid in methanol and left overnight for evaporation on a 65˚C heat plate. A volume of 1.5 ml distilled water was added to the dried samples, followed by 10 ml of Packard

PLOS ONE | https://doi.org/10.1371/journal.pone.0181751 August 1, 2017 4 / 13 Effects of irradiance and prey deprivation on a freshwater kleptoplastidic dinoflagellate

Insta-Gel scintillation cocktail, and radioactivity was determined using a Packard 1500 Tri- Carb liquid scintillation analyzer with automatic quench correction. Rates of photosynthetic activity (PA) were calculated as follows:

À 1 À 1 DPM Â ½DIC PA ðmgC Â ml Â h Þ ¼ 14 ð4Þ C a Â h

Where DPM is disintegrations min-1ml-1, DIC is the concentration of inorganic carbon (μg -1 14 -1 -1 C × ml ), Ca is the specific activity in disintegrations min ml and h is the incubation time in hours. To obtain daily photosynthesis, the rates per hour were multiplied by 16. DIC concentrations were measured on 1 ml subsamples using an infrared gas analyzer (ADC 225 Mk3 Gas analyzer, Analytic Development Co. Ltd., Hoddesdon, England) as described in detail else- where [26]. Glass vials with screw caps were used for DIC samples allowing no headspace, and the samples were analyzed within a few hours.

Results Qualitative observations of Nusuttodinium aeruginosum fed Chroomonas sp. using light and epifluorescence microscopy Well-fed cells of N. aeruginosum were 27–57 μm long and 22–46 μm wide (Fig 1A and 1B) and contained 6–10 chloroplasts (Fig 1C). The cell size and coloration decreased with starvation time, and chloroplasts in a stage of apparent degradation were observed (Fig 1B).

Experiment 1. Effects of irradiance on growth, photosynthesis and cell volume The growth response of Nusuttodinium aeruginosum was greatly stimulated by irradiance (S1 Fig). At an irradiance of 10 μmol photons m-2s-1, the cultures more or less sustained themselves during the duration of the experiment (14 days). At higher irradiances, a short lag phase was observed, before cultures went into exponential growth, which lasted for 6–16 days (non- shaded areas in S1 Fig) depending upon irradiance. The growth rate data fitted the Michaelis- Menten kinetics closely (R2 = 0.91) and indicate that the growth rate of N. aeruginosum saturated at an irradiance of ~40 μmol photons m-2s-1 at a growth rate (μ) of 0.3 d-1 (Fig 2A). Cell concentrations of the prey (Chroomonas sp.) were monitored in both mixed and mono cultures to evaluate the grazing responses of N. aeruginosum at the different irradiances. In all treatments, no significant differences were observed in prey concentrations, indicating low grazing rates (S2 Fig). To evaluate the extent to which growth rates of N. aeruginosum were affected by prey concentration at the different irradiances, the relationship between cell concentrations of Chroomonas sp. and growth rates of N. aeruginosum were tested (data not shown). The prey concentration increased from approximately 0.5–165.0x103 cells ml-1, equivalent to a predator:prey ratio of 1:9 to 1:87. Only in one case, 20 μmol photons m-2s-1, a significant relationship was observed between growth rate of N. aeruginosum and concentration of the prey (p = 0.016, R2 = 0.95). At all other irradiances tested, the slope of the correlation was not statistically different from zero. Hence within the prey concentrations used no influence of prey concentration on N. aeruginosum (0.13 < p < 0.83) was recorded. Photosynthetic activity of N. aeruginosum increased linearly with irradiance up to 140 μmol photons m-2s-1, which was the highest irradiance used in this experiment (Fig 2B p < 0.005, R2 = 0.96). The relationship between photosynthesis and growth could also be fitted to Michaelis-Menten kinetics well (R2 = 0.89). Thus, growth rates of N. aeruginosum saturated at

PLOS ONE | https://doi.org/10.1371/journal.pone.0181751 August 1, 2017 5 / 13 Effects of irradiance and prey deprivation on a freshwater kleptoplastidic dinoflagellate

Fig 2. Effect of irradiance on growth and photosynthetic activity of Nusuttodinium aeruginosum. (A) Growth rates as a function of irradiance. The curve was numerically fitted to Michaelis-Menten kinetics. μ = 0.45*(I-10)/(26.99+(I-10)), R2 = 0.91. Data points represent means ± SE (n = 12). (B) Photosynthetic activity as a function of irradiance. The data was fitted to a linear line, R2 = 0.96. Data points represent means ± SE (n = 4, except irradiance = 10 where n = 3). (C) Growth rates of N. aeruginosum as a function of photosynthetic activity. The curve was numerically fitted to Michaelis-Menten kinetics. μ = 0.41*(I-1.5)/(5.54 +(I-1.5)), R2 = 0.89. https://doi.org/10.1371/journal.pone.0181751.g002

PLOS ONE | https://doi.org/10.1371/journal.pone.0181751 August 1, 2017 6 / 13 Effects of irradiance and prey deprivation on a freshwater kleptoplastidic dinoflagellate

a photosynthetic activity of ~12 pg C cell-1h-1 or 192 pg C d-1 taking the light:dark period into account (Fig 2C). Cell carbon content estimates of N. aeruginosum gave values ranging from 1.1x103 ± 58 to 2.4x103 ± 168 pg C cell-1 at 10 and 65 μmol photons m-2s-1, respectively (Fig 3A). Cell carbon and daily biomass production (growth rate x cell carbon) (Fig 3B) could be fitted to Michaelis- Menten kinetics. The low R2, in the cell carbon relationship is due to large variations in cell volume among cells.

Experiment 2. Survival and maintenance of endosymbiont during prey deprivation Cells of Nusuttodinium aeruginosum subjected to prey starvation at an irradiance of 65 μmol photons m-2s-1 divided approximately four times during the first 14 days of the experiment (S3A Fig). Net mortality of cells was observed after day 14. The cell volume of N. aeruginosum and the calculated cell carbon decreased immediately after the first cell division. From day 4 the cells remained at almost half their initial cell size in linear dimensions, which is equivalent to 1500 pg C cell-1 (S3B Fig). Nonetheless, there was no major decrease in the photosynthetic activity which sustained at 300 pg C cell-1 d-1 the first 10 days and stayed constant at ~160 pg C cell-1 d-1 for the remainder of the experiment (Fig 4). Biomass production was reduced (due to the close relation to cell carbon/volume) to more than half of the initial production from day 2 to 4. From then on the production showed a similar response as the photosynthetic activity. However when growth became negative so did biomass production (Fig 4).

Discussion Effects of irradiance on growth and prey ingestion Growth rates of the blue-green freshwater dinoflagellate Nusuttodinium aeruginosum depended strongly on irradiance. The light compensation point, where the dinoflagellate was able to sustain itself, was achieved at an irradiance of ~10 μmol photons m-2s-1. Maximum growth rates of 0.3 (d-1) were obtained at irradiances 40 μmol photons m-2s-1 at predator: prey ratios of 1:9 to 1:87. Data on growth rates could be fitted to Michaelis-Menten kinetics taking this light compensation point into account. Thus, N. aeruginosum is highly dependent upon light for growth and it has a growth response very similar to an entirely phototrophic species. This is unlike other closely related species, like the marine N. poecilochroum and Gym- nodinium gracilentum, which can grow in the dark at a rate corresponding to ~40% of the maximum growth rate obtained in light [23]. While the planktonic species N. aeruginosum and G. gracilentum attain their maximum growth rates at about 50–80 μmol photons m-2s-1, the benthic species N. poecilocroum obtains its maximum growth rate at an irradiance as low as 10 μmol photons m-2s-1 [23]. The differences in responses towards light among these closely related dinoflagellates seem to reflect the light dependence of the algal prey. The prey used in the experiments on N. poecilochroum was a Chroomonas sp., which was well adapted to a low- light environment and sustained 50% of its maximum growth rate at only 2.9 μmol photons m-1s-1 [23]. In the present study we could not get reliable estimates of ingestion rates. The concentrations in mixed cultures were almost constantly lower than the concentration of mono cultures, but the differences between Chroomonas sp. in mixed and mono cultures were too small to allow for computation of ingestion rates. Since we initiated the experiments at a predator:prey ratio of 10:1, this indicates low ingestion rates (initially less than ~1–2 prey cells ingested per day).

PLOS ONE | https://doi.org/10.1371/journal.pone.0181751 August 1, 2017 7 / 13 Effects of irradiance and prey deprivation on a freshwater kleptoplastidic dinoflagellate

Fig 3. Effect of irradiance on cell volume and biomass production of Nusuttodinium aeruginosum. (A) Cell volume and cell carbon of N. aeruginosum as a function of irradiance. Data points represent means ± SE (n = 30±40). The curves were numerically fitted to Michaelis-Menten kinetics: cell volume = 22048*I/(21.04+I), R2 = 0.25; and cell carbon = 2668*I/(15.45+I), R2 = 0.28. Note different ordinate scales. (B) Biomass production (BP) and photosynthetic activity (PA) as a function of irradiances. The BP curve was numerically fitted to Michaelis-Menten kinetics: BP = 1144*(I-13)/(38.62+(I-13)), R2 = 0.93. The data for PA were fitted to a linear line, R2 = 0.97. Data for biomass production are the same as in Figs 2A and 3A. https://doi.org/10.1371/journal.pone.0181751.g003

PLOS ONE | https://doi.org/10.1371/journal.pone.0181751 August 1, 2017 8 / 13 Effects of irradiance and prey deprivation on a freshwater kleptoplastidic dinoflagellate

Fig 4. Biomass production (BP) and photosynthetic activity (PA) of Nusuttodinium aeruginosum as a function of time under prey deprivation. Data points represent means ± SE (BP; n = 30, PA; n = 4). https://doi.org/10.1371/journal.pone.0181751.g004

Importance of inorganic carbon uptake for the carbon budget We observed a positive linear relationship between photosynthesis and irradiance within the studied irradiance levels, which is very different from the Michaelis-Menten relationship observed in the light response of growth and estimated biomass production. The potential contribution of photosynthesis varied with irradiance, with the lowest contribution at low irradiance and the highest at high irradiance. The daily inorganic carbon fixation corresponded at most to ~26% of the cell carbon content (Table 1). Reported values of inorganic carbon uptake are sparse for other dinoflagellates with kleptochloroplasts. Dinophysis acuminata, which only sequesters chloroplasts, has been found to have a daily specific inorganic carbon fixation corresponding to ~59% of its cell carbon content at an irradiance of 100 μmol photons m-2s-1 [27]. For dinoflagellates that also retain prey nuclei and prey mitochondria, such as the closely related species, Gymnodinium gracilentum, the specific inorganic carbon uptake rates were

Table 1. Comparisons of maximum photosynthetic rate, cellular carbon and cell carbon fixed per day (%) in three dinoflagellate and one ciliate that all carry out acquired phototrophy. Data from the literature. All data are from experiments carried out at 15ÊC, except for M. rubrum (3ÊC).

a c Irradiance Light:Dark cycle Pmax Cellular carbon Cell carbon fixed per day (μmol photons m-2s-1) (pgC cell-1 d-1) (pgC cell-1) (%) Gymnodinium gracilentum [17] 90 16:8 123±5 104 118 Dinophysis acuminata [27] 100 14:10 532±84 b 895±66 59 Nusuttodinium aeruginosum (Present 140 16:8 592±16 22.9x103±153 26 study) Mesodinium rubrum [28] 25 16:8 726±64 597 122 a Pmax is the highest photosynthetic rate measured, it might saturate at another irradiance. b Given numbers are from well-fed cultures. c Calculated using [31]

https://doi.org/10.1371/journal.pone.0181751.t001

PLOS ONE | https://doi.org/10.1371/journal.pone.0181751 August 1, 2017 9 / 13 Effects of irradiance and prey deprivation on a freshwater kleptoplastidic dinoflagellate

higher, ~118% [17]. Similarly high inorganic carbon fixation values (~184%) were found in the ciliate Mesodinium rubrum, which contains prey nuclei and mitochondria [28–30]. The fact that the measured photosynthesis only to a minor degree could explain the observed increased growth rates at higher irradiances raises some questions concerning our measurements and calculations. First, cell carbon was not measured but estimated from published relationships between cell volume and cell carbon. Second, it is possible that we may have underestimated the uptake of inorganic carbon using the 14C technique. This may happen

if the CO2 produced during the prey digestion is reused by the dinoflagellate. Third, it is possible that light may have a positive effect of light on growth rates. This has been documented in a number of both heterotrophic and non-constitutive mixotrophic species of ciliates and dinoflagellates [17, 32–35]. A hypothesis for this phenomenon could be that high irradiance induces and aids digestion of phytoplankton. Some studies have shown that especially phototrophic prey is more easily digested in the light [33, 36]. This is supported by a recent study which showed that the oxidation of carbohydrates by enzymes substantially increased in light [37]. Another pathway to utilize solar energy could be through the protein, rhodopsin, which was recently found in Oxyrrhis marina [35]. We suggest three explanations for the positive effect of light: (1) Algal pigments are highly labile in the presence of light and oxygen, making ingested material more readily utilizable for growth [38]. (2) Complex systems like enzymes or other proteins are expressed to greater extent in light. (3) The ATP produced by photosynthetic activity aids digestion and ingestion of prey, rather than to produce glucose. With increasing irradiance the photosynthesis increases, resulting in greater amounts of ATP being produced, allowing faster digestion, hence greater ingestion rates.

Nusuttodinium aeruginosum during prey starvation Previous studies have demonstrated that the kleptochloroplasts of N. aeruginosum can be retained for > 1 month, when cells are subjected to prey starvation [39]. In the present experiments, N. aeruginosum cells doubled more than 3 times during the first 9 days when subjected to prey starvation, despite this, N. aeruginosum cells were able to obtain its maximum cellular photosynthetic rate in the same period. This may indicate that N. aeruginosum is able to divide the kleptochloroplasts. However, since we did not perform either measurements of chlorophyll a or enumerated the chloroplasts, this is speculation and thus future experiments are needed. Cultures of N. aeruginosum survived for more than 27 days without prey and the chloroplasts were photosynthetic active. This is a long period compared to most other dinoflagellates with kleptochloroplasts. In Gymnodinium gracilentum the photosynthetic activity decreased quickly with starvation time and was essentially zero after 48 h of prey starvation [17]. Simi- larly, although no actual rates of photosynthesis have been reported on Pfiesteria piscicida, autoradiographical studies have indicated photosynthetic activity in this species for a least 7 days [18]. The only dinoflagellates retaining kleptochloroplasts for longer time are species of Dinophysis, but they only sequester the chloroplasts, not other cell constituents [1, 27]. The ability of N. aeruginosum to retain kleptochloroplasts and other cell organelles in many ways resemble the case of the red-pigmented ciliate Mesodinium rubrum. This ciliate may retain and divide the kleptochloroplasts [29], using the sequestered prey nucleus to farm the chloroplasts and produce photosynthetic pigments, even performing photo acclimation [40–42].

Conclusion This investigation clearly shows that Nusuttodinium aeruginosum has some control of its kleptochloroplasts. The kleptochloroplasts were photosynthetic active for at least 27 days in prey-starved cultures, and N. aeruginosum maintained a high photosynthetic rate for the first

PLOS ONE | https://doi.org/10.1371/journal.pone.0181751 August 1, 2017 10 / 13 Effects of irradiance and prey deprivation on a freshwater kleptoplastidic dinoflagellate

9 days, despite going through 3 cell divisions. These observations clearly indicate that the dinoflagellate is able to divide the sequestered chloroplasts, but details need to be investigated further. The growth rate of N. aeruginosum was strongly dependent upon irradiance, yet measured rates of photosynthesis only contributed by 15–25% of the total carbon requirements of N. aeruginosum under well-fed conditions. Thus, the sequestered chloroplasts may contribute something in addition to carbon. It is not the first time this phenomenon has been described as similar observations have been done in other heterotrophic and kleptoplastidic species [23, 33, 34]. Further studies are needed to elucidate the contribution from sequestered chloroplasts.

Supporting information S1 Fig. Development in cell concentration of Nusuttodinium aeruginosum (in a mixed culture) through time under seven different irradiances (10, 20, 25, 40, 65, 100 and 140 μmol photons m-2 s-1). Data points represent 12 replicates. Shaded areas are considered lag phase and steady state, respectively and were not used in calculations of the growth rates. Arrows indicate the time of photosynthesis measurements. (TIFF) S2 Fig. Development in cell concentration of Chroomonas sp. as prey (n = 12) and as control (n = 6) through time, under six different irradiances (10, 20, 25, 40, 65 and 140 μmol photons m-2 s-1). Data points represent means ± SE. (TIFF) S3 Fig. Effect of prey deprivation on growth and cell volume of Nusuttodinium aeruginosum. (A) Development in cell concentration and growth rate of N. aeruginosum during prey deprivation. Data points represent means ± SE (n = 8, except day 6 and 9 = where n = 4 and 5 respectively). (B) Cell volume and cell carbon of N. aeruginosum as a function of time under prey deprivation. Data points represent means ± SE (n = 30). (TIFF)

Acknowledgments We thank Gert Hansen for a culture of Chroomonas sp. (SCCAP K-1623).

Author Contributions Conceptualization: Kirstine Drumm, Mette Liebst-Olsen, Niels Daugbjerg, Øjvind Moestrup, Per Juel Hansen. Data curation: Kirstine Drumm, Mette Liebst-Olsen. Formal analysis: Kirstine Drumm, Mette Liebst-Olsen. Funding acquisition: Niels Daugbjerg, Per Juel Hansen. Investigation: Kirstine Drumm, Mette Liebst-Olsen, Niels Daugbjerg, Øjvind Moestrup, Per Juel Hansen. Methodology: Kirstine Drumm, Mette Liebst-Olsen, Niels Daugbjerg, Øjvind Moestrup, Per Juel Hansen. Project administration: Per Juel Hansen. Resources: Niels Daugbjerg, Per Juel Hansen.

PLOS ONE | https://doi.org/10.1371/journal.pone.0181751 August 1, 2017 11 / 13 Effects of irradiance and prey deprivation on a freshwater kleptoplastidic dinoflagellate

Software: Niels Daugbjerg, Per Juel Hansen. Supervision: Niels Daugbjerg, Øjvind Moestrup, Per Juel Hansen. Validation: Kirstine Drumm, Mette Liebst-Olsen, Niels Daugbjerg, Øjvind Moestrup, Per Juel Hansen. Visualization: Kirstine Drumm, Mette Liebst-Olsen, Niels Daugbjerg, Øjvind Moestrup, Per Juel Hansen. Writing – original draft: Kirstine Drumm, Mette Liebst-Olsen. Writing – review & editing: Kirstine Drumm, Mette Liebst-Olsen, Niels Daugbjerg, Øjvind Moestrup, Per Juel Hansen.

References 1. Hansen PJ, OjamaÈe K, Berge T, Trampe EC, Nielsen LT, Lips I, et al. Photoregulation in a kleptochloroplastidic dinoflagellate, Dinophysis acuta. Front Microbiol. 2016; 7: 785. https://doi.org/10.3389/fmicb. 2016.00785 PMID: 27303378 2. Foster RA, Carpenter EJ, Bergman B. Unicellular cyanobionts in open ocean dinoflagellates, radiolarians, and intinnids: ultrastructural characterization and immuno-localization of phycoerythrin and nitro- genase. J Phycol. 2006; 42: 453±63. 3. Tarangkoon W, Hansen G, Hansen PJ. Spatial distribution of symbiont-bearing dinoflagellates in the Indian Ocean in relation to oceanographic regimes. Aquat Microb Ecol. 2010; 58: 197±213. 4. Daugbjerg N, Jensen MH, Hansen PJ. Using nuclear-encoded LSU and SSU rDNA sequences to identify the eukaryotic endosymbiont in Amphisolenia bidentata (Dinophyceae). Protist. 2013; 164: 411±22. https://doi.org/10.1016/j.protis.2012.10.001 PMID: 23266212 5. Sweeney BM. Laboratory studies of a green Noctiluca from New Guinea. J Phycol. 1971; 7: 53±8. 6. Sweeney BM. Pedinomonas noctilucae (Prasinophyceae), the flagellate symbiotic in Noctiluca (Dino- phyceae) in Southeast Asia. J Phycol. 1976; 12: 460±4. 7. Hansen PJ, Miranda L, Azanza R. Green Noctiluca scintillans: a dinoflagellate with its own greenhouse. Mar Ecol Prog Ser. 2004; 275: 79±87. 8. Kim M, Kim KY, Nam SW, Shin W, Yih W, Park MG. The effect of starvation on plastid number and photosynthetic performance in the kleptoplastidic dinoflagellate Amylax triacantha. J Eukaryot Microbiol. 2014; 61: 354±63. https://doi.org/10.1111/jeu.12115 PMID: 24734883 9. Kim M, Nam SW, Shin W, Coats DW, Park MG. Dinophysis caudata (Dinophyceae) sequesters and retains plastids from the mixotrophic ciliate prey Mesodinium rubrum. J Phycol. 2012; 48: 569±79. https://doi.org/10.1111/j.1529-8817.2012.01150.x PMID: 27011072 10. Wisecaver JH, Hackett JD. Transcriptome analysis reveals nuclear-encoded proteins for the maintenance of temporary plastids in the dinoflagellate Dinophysis acuminata. BMC genomics. 2010; 11: 1. 11. Schnepf E, Winter S, Mollenhauer D. Gymnodinium aeruginosum (Dinophyta): a blue-green dinoflagellate with a vestigial, anucleate, cryptophycean endosymbiont. Plant Syst Evol. 1989; 164: 75±91. 12. Fields SD, Rhodes RG. Ingestion and retention og Chroomonas spp. (Cryptophyceae) by Gymnodi- nium acidotum (Dinophyceae). J Phycol. 1991; 27: 525±9. 13. Takano Y, Yamaguchi H, Inouye I, Moestrup Ø, Horiguchi T. Phylogeny of five species of Nusuttodinium gen. nov. (Dinophyceae), a genus of unarmoured kleptoplastidic dinoflagellates. Protist. 2014; 165: 759±78. https://doi.org/10.1016/j.protis.2014.09.001 PMID: 25460229 14. Wilcox LW, Wedemayer GJ. Dinoflagellate with blue-green chloroplasts derived from an endosymbiotic eukaryote. Science. 1985; 227: 192±4. https://doi.org/10.1126/science.227.4683.192 PMID: 17843078 15. Larsen J. An ultrastructural study of Amphidinium poecilochroum (Dinophyceae), a phagotrophic dinoflagellate feeding on small species of cryptophytes. Phycologia. 1988; 27: 366±77. 16. Horiguchi T, Pienaar R. Amphidinium latum Lebour (Dinophyceae), a sand-dwelling dinoflagellate feeding on cryptomonads. Jpn J Phycol. 1992; 40: 353±63. 17. Skovgaard A. Role of chloroplast retention in a marine dinoflagellate. Aquat Microb Ecol. 1998; 15: 293±301. 18. Lewitus AJ, Glasgow HB, Burkholder JM. Kleptoplastidy in the toxic dinoflagellate Pfiesteria piscicida (Dinophyceae). J Phycol. 1999; 35: 303±12.

PLOS ONE | https://doi.org/10.1371/journal.pone.0181751 August 1, 2017 12 / 13 Effects of irradiance and prey deprivation on a freshwater kleptoplastidic dinoflagellate

19. Eriksen NT, Hayes KC, Lewitus AJ. Growth responses of the mixotrophic dinoflagellates, Cryptoperidi- niopsis sp. and Pfiesteria piscicida, to light under prey-saturated conditions. Harmful Algae. 2002; 1: 191±203. 20. Gast RJ, Moran DM, Dennett MR, Caron DA. Kleptoplasty in an Antarctic dinoflagellate: caught in evolutionary transition? Environ Microbiol. 2007; 9: 39±45. https://doi.org/10.1111/j.1462-2920.2006. 01109.x PMID: 17227410 21. Koike K, Takishita K. Anucleated cryptophyte vestiges in the gonyaulacalean dinoflagellates Amylax buxus and Amylax triacantha (Dinophyceae). Phycol Res. 2008; 56: 301±11. 22. Yamaguchi H, Nakayama T, Kai A, Inouye I. Taxonomy and phylogeny of a new kleptoplastidal dinoflagellate, Gymnodinium myriopyrenoides sp. nov. (Gymnodiniales, Dinophyceae), and its cryptophyte symbiont. Protist. 2011; 162: 650±67. https://doi.org/10.1016/j.protis.2011.01.002 PMID: 21497133 23. Jakobsen HH, Hansen PJ, Larsen J. Growth and grazing responses of two chloroplast-retaining dinoflagellates: effect of irradiance and prey species. Mar Ecol Prog Ser. 2000; 201: 121±8. 24. Stoecker D, Silver M, Michaels A, Davis L. Obligate mixotrophy in Laboea strobila, a ciliate which retains chloroplasts. Mar Biol. 1988; 99: 415±23. 25. Skovgaard A, Hansen PJ, Stoecker DK. Physiology of the mixotrophic dinoflagellate Fragilidium subglobosum. I. Effects of phagotrophy and irradiance on photosynthesis and carbon content. Mar Ecol Prog Ser. 2000; 201: 129±36. 26. Nielsen LT, Lundholm N, Hansen PJ. Does irradiance influence the tolerance of marine phytoplankton to high pH? Marine Biology Research. 2007; 3: 446±53. 27. Riisgaard K, Hansen PJ. Role of food uptake for photosynthesis, growth and survival of the mixotrophic dinoflagellate Dinophysis acuminata. Mar Ecol Prog Ser. 2009; 381: 51±62. 28. Johnson MD, Stoecker DK. Role of feeding in growth and photophysiology of Myrionecta rubra. Aquat Microb Ecol. 2005; 39: 303±12. 29. Hansen PJ, Fenchel T. The bloom-forming ciliate Mesodinium rubrum harbours a single permanent endosymbiont. Mar Biol Res. 2006; 2: 169±77. 30. Smith M, Hansen PJ. Interaction between Mesodinium rubrum and its prey: importance of prey concentration, irradiance and pH. Mar Ecol Prog Ser. 2007; 338: 61±70. 31. Menden-Deuer S, Lessard EJ. Carbon to volume relationships for dinoflagellates, diatoms, and other protist plankton. Limnol Oceanogr. 2000; 45: 569±79. 32. Stoecker D, Putt M, Davis L, Michaels A. Photosynthesis in Mesodinium rubrum: species-specific measurements and comparison to community rates. Mar Ecol Prog Ser. 1991; 73: 245±52. 33. Strom SL. Light-aided digestion, grazing and growth in herbivorous protists. Aquat Microb Ecol. 2001; 23: 253±61. 34. Tarangkoon W, Hansen PJ. Prey selection, ingestion and growth responses of the common marine ciliate Mesodinium pulex in the light and in the dark. Aquat Microb Ecol. 2011; 62: 25±38. 35. Guo Z, Zhang H, Lin S. Light-promoted rhodopsin expression and starvation survival in the marine dinoflagellate Oxyrrhis marina. PloS One. 2014; 9: e114941. https://doi.org/10.1371/journal.pone.0114941 PMID: 25506945 36. Klein B, Gieskes WW, Krray GG. Digestion of chlorophylls and carotenoids by the marine protozoan Oxyrrhis marina studied by HPLC analysis of algal pigments. J Plankton Res. 1986; 8: 827±36. 37. Cannella D, MoÈllers K, Frigaard N-U, Jensen P, Bjerrum M, Johansen K, et al. Light-driven oxidation of polysaccharides by photosynthetic pigments and a metalloenzyme. Nat Commun. 2016; 7: 11134. https://doi.org/10.1038/ncomms11134 PMID: 27041218 38. Porra R, PfuÈndel E, Engel N. Metabolism and function of photosynthetic pigments. Jeffrey SW M R, Wright SW, editor. Phytoplankton pigments in oceanography: UNESCO, Paris; 1997. 85±126 p. 39. Onuma R, Horiguchi T. Kleptochloroplast enlargement, karyoklepty and the distribution of the cryptomonad nucleus in Nusuttodinium (= Gymnodinium) aeruginosum (Dinophyceae). Protist. 2015; 166: 177± 95. https://doi.org/10.1016/j.protis.2015.01.004 PMID: 25771111 40. Johnson MD, Oldach D, Delwiche CF, Stoecker DK. Retention of transcriptionally active cryptophyte nuclei by the ciliate Myrionecta rubra. Nature. 2007; 445: 426±8. https://doi.org/10.1038/nature05496 PMID: 17251979 41. Johnson MD, Tengs T, Oldach D, Stoecker DK. Sequestration, performance, and functional control of cryptophyte plastids in the ciliate Myrionecta rubra (Ciliophora). J Phycol. 2006; 42: 1235±46. 42. Moeller HV, Johnson MD, Falkowski PG. Photoacclimation in the phototrophic marine ciliate Mesodi- nium rubrum (Ciliophora). J Phycol. 2011; 47: 324±32. https://doi.org/10.1111/j.1529-8817.2010. 00954.x PMID: 27021864

PLOS ONE | https://doi.org/10.1371/journal.pone.0181751 August 1, 2017 13 / 13

Supporting Information

S1 Fig. Development in cell concentration ofNusuttodiniumaeruginosum(in a mixed culture) through time under seven different irradiances (10, 20, 25, 40, 65, 100 and 140μmol photons m-2 s-1). Data points represent 12 replicates. Shaded areas are considered lag phase and steady state, respectively and were not used in calculations of the growth rates. Arrows indicate the time of photosynthesis measurements.

S2 Fig. Development in cell concentration ofChroomonassp. as prey (n = 12) and as control (n = 6) through time, under six different irradiances (10, 20, 25, 40, 65 and 140μmol photons m-2 s-1). Data points represent means±SE.

S3 Fig. Effect of prey deprivation on growth and cell volume ofNusuttodiniumaeruginosum. (A) Development in cell concentration and growth rate ofN.aeruginosum during prey deprivation. Data points represent means±SE (n = 8, except day 6 and 9 = where n = 4 and 5 respectively). (B) Cell volume and cell carbon ofN.aeruginosum as a function of time under prey deprivation. Data points represent means±SE (n = 30).

Prepared for Journal of Eukaryotic Microbiology

Physiological responses of Mesodinium major to irradiance, prey concentration and prey starvation

Kirstine Drumm1, Andreas Norlin1,2, Miran Kim1,3, Andreas Altenburger4, Per Juel Hansen1

1University of Copenhagen, Department of Biology, Helsingør, Denmark

2Swansea University, Department of Biosciences, Wales, UK

3Chonnam National University, Research Institute for Basic Science, South Korea

4University of Copenhagen, Natural History Museum of Denmark, Copenhagen, Denmark

Abstract Mesodinium rubrum/M. major are non-constitutive mixotrophs harboring chloroplasts and other cell organelles from specific cryptophyte prey. They are ecologically important due to the formation of red tides globally and their function as prey for the DSP producing Dinophysis spp. M. major has recently been described and new studies have indicated that red tides may be due to M. major to same extent as M. rubrum. Despite this significant role, the physiology of M. major has never been studied and compared to M. rubrum. In this study, the growth, food uptake, chlorophyll a and inorganic carbon uptake was measured at six different irradiances (25, 38, 50, 75, 100 and 200 µmol m-2 s-1) when fed the cryptophyte Teleaulax amphioxeia. Experiments were also carried out on the response of M. major to prey deprivation and refeeding. Maximum growth rates for M. major of 0.39 d-1 were achieved > 75 µmol photons m-2 s-1, whereas a growth rate of only 0.07 d-1 was achieved at an irradiance of 25 µmol photons m-2 s-1. Similarly, photosynthesis ranged from 79.2 to 323 pg C cell-1 h-1 at irradiances of 25 and 200 µmol photons m-2 s-1, respectively. M. major ingested T. amphioxeia at a rate of 9-15 preys predator-1 d-1 at all irradiances. The highest rates of photosynthesis covered 95% of M. major carbon needs. M. major was well adapted to prey starvation and tolerated 51 days of prey deprivation without mortality. M. major quickly recovered starvation when re-fed, especially due to high ingestion rates of >150 prey predator-1 day-1. 3D imaging of M. major cells stained with Hoechst 3325 and CellMask Green in combination revealed a range of 40-80 chloroplasts. M. major has two macro nuclei, one micro nucleus at all times and, when well fed, an enlarged prey nucleus, similar what has been found in M. rubrum. Generally, M. major functions quite similar to M. rubrum. However, it contains a higher number of chloroplasts allowing it to have a up to four times higher photosynthesis, and it requires more light to growth.

Keywords: Mesodinium major, light response, growth, C-fixation, Chl. a, prey deprivation, sequestered chloroplasts, sequestered nucleus

Introduction The ciliate genus Mesodinium is globally distributed in marine and freshwaters. In marine waters, six Mesodinium species have been formally described (Garcia-Cuetos, Moestrup et al. 2012, Moestrup, Garcia- Cuetos et al. 2012, Nam, Shin et al. 2015). The genus is a physiologically diverse group consisting of both heterotrophs (M. pupula and M. pulex) and mixotrophs (M. chamaeleon and M. coatsi, M. major and M. rubrum) (Smith and Hansen 2007, Tarangkoon and Hansen 2011, Garcia-Cuetos, Moestrup et al. 2012, Kim, Kang et al. 2019). Recent research using molecular techniques suggest quite a few additional species may be present within this M. rubrum/M. major complex (Herfort, Peterson et al. 2011, Johnson, Beaudoin et al. 2016, Johnson and Beaudoin 2019, Kim and Park 2019).

The M. rubrum/major species complex has received the most attention, because they form non-toxic “red tide” blooms worldwide (Packard, Blasco et al. 1978, Lindholm 1985, Herfort, Peterson et al. 2011). The blooms were attributed to M. rubrum in older records, but recent papers have suggested that many of these red Mesodinium blooms are due to M. major (Yang, Löder et al. 2015, Johnson, Beaudoin et al. 2016, Herfort, Maxey et al. 2017). In fact, one paper found M. major to be the most widely encountered Mesodinium species in red tides (Johnson, Beaudoin et al. 2016). Despite this, almost all laboratory studies on the red Mesodinium spp have been done on M. rubrum, probably due to failed attempts to culture M. major (Garcia-Cuetos, Moestrup et al. 2012, Rial, Laza-Martínez et al. 2015). The red Mesodinium spp have also received considerable interest because they serve as prey for the toxic dinoflagellates, Dinophysis spp (Park, Kim et al. 2006). Dinophysis spp produce shellfish toxins that may damage the aquaculture industry due to the accumulation DST toxins in mussels (Reguera, Velo-Suárez et al. 2012).

Species within the M. rubrum/M. major complex differ from the other mixotrophic Mesodinium spp in their association with the ingested cryptophytes. These red Mesodinium spp only seem to utilize cryptophytes within the Teleaulax/Plagioselmis/Geminigera clade, while the other mixotrophic species, M. coatsi and M. chamaeleon, are more flexible, and can utilize a wide range of cryptophyte species. M. rubrum and M. major also differ from the other mixotrophic Mesodinium species in that they upon ingestion of a cryptophyte prey cell separate the prey nuclei from the rest of the ingested cell, while the chloroplasts, nucleomorph, cryptophyte mitochondria, cryptophyte ribosomes, and cryptophyte cytoplasm are kept together as entities. One of the ingested prey nuclei is made the “master”, often referred to as the “symbiont nucleus” or “the centered prey nucleus”. This nucleus is transported into close proximity of the ciliate nuclei (two macronuclei and a single micronucleus) and enlarged. Often some extra cryptophyte nuclei can be found in the periphery of the Mesodinium cell.

At present only the physiology of M. rubrum has been studied in details in Antarctic and temperate strains from Europe, Asia and Antarctic (Moeller, Johnson et al. 2011, Johnson, Beaudoin et al. 2016, Kim, Kang et al. 2019), while the physiology of M. major and other members of mixotrophic members of this species

2 complex is still unexplored. Results from the M. rubrum strains indicate that they acquire most of their carbon via photosynthesis, and that up to ~ 98% of the carbon need in this species can be covered by photosynthesis at low prey concentrations, which is enough to support high growth rates (Smith and Hansen 2007, Mitra, Flynn et al. 2016). It has also been shown that M. rubrum is able to photoacclimate, thereby allowing them to exploit low light environments. Finally, it has been shown that M. rubrum can survive extended periods of prey starvation (Johnson and Stoecker 2005, Smith and Hansen 2007)

Our aim was to explore ecophysiology of M. major. We have maintained a culture of M. major for > two years using Teleaulax amphioxeia as prey and investigated: (i) The responses to different irradiances with focus on growth of M. major, photosynthetic activity, cellular chl a and ingestion rates (ii) Effects of prey deprivation for 50 days on the loss of centered prey nucleus. (iii) Refeeding and recovery after prey starvation. These experiments allowed us to study the physiological similarities and differences between M. rubrum and M. major. Our results indicate some similarities, but also some differences that suggest a necessity to differentiate them with regard to physiological performance.

Method and materials Isolation and cultures Single cells of Mesodinium major (MM-DK2016) were isolated from Helsingør harbor, Denmark, in 2016, using a drawn Pasteur and transferred to 6-well multidishes containing sterile filtered seawater from the location three times to remove all other protists. Dilute concentrations of the cryptophyte, Teleaulax amphioxeia (SCCAP K-0434) cells were supplied as prey. After some months the cultures were scaled up to blue cap bottles (0.25L), containing 100 mL f/4 medium based on heat-treated seawater (95 C, 90 min) with a salinity of 15. M. major was transferred weekly to new medium and supplied T. amphioxeia as prey at a predator:prey ratio of ~1:5. Both species were grown at 15 °C in a temperature regulated room, under a photon irradiance of 70 µmol photons m2 s1 (PAR, 400–700 nm), and on a light:dark cycle of 16:8. Irradiance was measured at the level of incubation flasks using a light meter equipped with a spherical quantum sensor (ULM and US-SQS/L, Walz GmbH, Germany) and pH was followed using a SenTix®41 pH electrode (WTW, Germany) connected to a pH meter (WTW, pH 3210, Germany) and calibrated with pH 7 and 10 standard buffers.

Experiment 1. Effect of irradiance on cellular chlorophyll a, photosynthesis and growth rate M. major was exposed to six different irradiances (25, 38, 50, 75, 100 and 200 µmol photons m-2s-1) at a prey concentration of 500 cells mL-1, and a predator:prey ratio of ~1:15 in 0.5 L blue cap flasks containing 300 mL f/4 medium. Subsamples for enumeration of M. major and T. amphioxeia were withdrawn every second day for two weeks. Experiments with monocultures of T. amphioxeia were also carried out, allowing the calculation

3 of ingestion rates (see below). After ~three cell divisions (most often after 7 days of incubation) samples for photosynthetic activity and chlorophyll a (Chl a) were withdrawn. Refeeding and dilution of the cultures were carried out when prey:predator ratio were around 1:1 and cultures exceeded 1000 cells mL-1, respectively.

Experiment 2: Effect of prey depletion and starvation Experiment 2 was initiated with the cultures from experiment 1 grown at an irradiance of 100 µmol photons m2 s1. The triplicates from experiment 1 were mixed in one bottle, diluted and transferred into four new bottles, functioning as replicates to prevent elevated pH complications. Cell numbers, photosynthetic activity, Chl a, and number of centered prey nuclei (CPN) were monitored for 51 days (Table 1, see for details). The cultures were diluted when densities exceeded 1000 cells mL-1 with fresh f/4 medium to avoid complications with elevated pH.

Experiment 3: Effects of refeeding Experiment 3 was carried out directly following experiment 2. Subsamples from experiment 2 were withdrawn after 22 days of starvation from each replicate flasks and mixed in a single bottle. Subsequently, this suspension was divided into two different treatments using Teleaulax amphioxeia as prey. Treatment 1 and 2 were initiated using a prey to predator ratio of 30:1, and 100:1, respectively. Each treatment was further subdivided into three new replicate bottles, functioning as replicates. Measurements of photosynthetic activity, Chl a, number of CPNs, and cell number was monitored for 10 days. The setup was accompanied with monocultures of T. amphioxeia, allowing the calculation of ingestion rates (see below). Dilution of the cultures were done when densities exceeded 1000 cells mL-1.

Cell Abundance and Growth Rate Aliquots (2 mL) withdrawn from each flask were fixed with acid Lugol’s solution (final concentration 1%). Abundances of M. major and T. amphioxeia were enumerated using a Sedgewick-Rafter chamber under the inverted microscope (Olympus CK40) at 100X and 200X. At least 200 cells were enumerated each time. Growth rates were calculated during the exponential portion of the growth phase using the following exponential growth equation:

( ) ( ) = 𝑙𝑙𝑙𝑙 𝑁𝑁2 − 𝑙𝑙𝑙𝑙 𝑁𝑁1 𝜇𝜇 2 1 Where N1 and N2 are cell concentrations at time 1 and𝑡𝑡 time− 𝑡𝑡2, respectively.

Ingestion rate The ingestion rate of M. major was calculated from the reduction in prey concentrations over 48-72 h periods as compared with prey control cultures according to (Jakobsen and Hansen 1997). This method assumes that

4 the prey grows exponentially at the same rate as in predator-free prey controls. Ingestion rate (U) was calculated for each sample point using following equation:

( ) If µx ≠ µy , then = ( 𝜇𝜇)𝑥𝑥𝑇𝑇 𝜇𝜇𝑦𝑦−𝜇𝜇𝑥𝑥 �𝑋𝑋0− 𝑋𝑋𝑇𝑇𝑒𝑒 � 𝜇𝜇𝑥𝑥𝑇𝑇 𝜇𝜇𝑦𝑦𝑇𝑇 𝑈𝑈 𝑌𝑌0 𝑒𝑒 − 𝑒𝑒 Where prey (X) are ingested by grazer (Y), µy are exponential growth of M. major, µx are exponential growth of T. amphioxceia without predator, X0 and XT are cell concentrations at time 0 and time T, respectively.

Photosynthetic Activity (14C) Two 2-mL aliquots were withdrawn from each flask, transferred to each of two 23-mL glass scintillation vials.

14 - -1 Twenty microliter of NaH CO3 stock solution (specific activity 100 µCi mL ) was added to each vial. One vial of each pair was incubated for 3 h in the same place as the experimental flask, and the other vial was kept in complete darkness by wrapping in aluminum foil. After incubation, a 100 mL sub-sample was withdrawn from each vial and added to a new vial containing 200 mL phenylethylamine for measurements of specific activity (Skovgaard, Hansen et al. 2000). The remaining 1.9 mL was acidified with 2 mL 10% glacial acetic acid in methanol, and evaporated overnight at 60 °C to remove all inorganic carbon. The residue in the vial was re-dissolved in 2 mL Milli-Q water before adding 10 mL of scintillation cocktail (Insta-Gel Plus, Packard, USA). All vials were analyzed using a liquid scintillation counter (Tri-Carb 2910 TR, Perkin-Elmer). Rates of photosynthetic activity PA (µgC × mL-1 × h-1) were calculated from the equation:

× = 14 𝐷𝐷𝐷𝐷𝐷𝐷× ×𝐼𝐼𝐼𝐼 𝑃𝑃𝑃𝑃 Where DMP is disintegrations min-1mL-1, IC is the concentration𝐶𝐶𝑎𝑎 ℎ 𝑁𝑁𝑡𝑡 of inorganic carbon (µgC × mL-1), 14Cₐ is the -1 -1 specific activity in disintegrations min mL , h is the incubation time in hours and Nt is the total number of cells in the vail. IC was measured using a Shimadzu Total Organic Carbon (TOC) analyzer.

Chlorophyll a measurements 2 mL subsample was filtered onto a 24mm GF/F (Whatman) glass fiber filter, which were subsequently transferred to 5 mL ethanol (96%) in 23-mL glass scintillation vials. The vials were wrapped with tinfoil and left overnight in the refrigerator at 4 °C. Aliquots were transferred to 2 mL glass vials and measured on a Turner ® Trilogy Fluorometer using non-acidification method. Throughout all handling of these sample extraction and measurements, light was eliminated.

Enumeration of centered prey nuclei and imaging of Mesodinium major Prey nuclei and chloroplasts were stained using the fluorescent nuclear stain Hoechst 33258 (Invitrogen, USA) and plasma membrane stain using CellMask Green (Life technologies, Carlsbad, CA, USA). Two mL of culture were fixed in 4% glutaraldehyde in filtrated seawater and left cold (4°C). The samples were stained

5 with a mix of 25 mg mL-1 Hoerst and 0.25 mg mL-1 Cell mask, and left for 15 min, before collected on a 0.2 mm black polycarbonate membrane filter (Frisenette, Denmark) using filtration. A single drop of immersion oil was added to microscopy slide and the filter was placed on top of this. Another drop of immersion oil was added to the top of the filter, before a cover slip was added. The slides were kept at 4 °C in the dark, before analysis. Epifluorescence micrographs of stained M. major cells were taken at 1,000X magnification using a digital camera coupled to the Olympus BX51 microscope equipped with differential interference contrast. 3D images were generated using IMARIS software program (Bitplane, Zürich, Switzerland) to assess the number of chloroplasts of M. major.

Figure 1. Epifluorescence micrographs of Mesodinium major. A-C Micrographs from disc-spinning unit revealing arrangement of the chloroplasts in M. major, (A) in the form of a 3D-reconstructed cell, (E) seen from the top, and (C) seen from the side. D-E M. major stained with Hoechst 3325 and CellMask Green in combination, (D) Starved M. major containing two ciliate macronuclei and one ciliate micronucleus, (E) Well- fed M. major containing centered prey nucleus (CPN), (F) Overview of M. major and the numerous chloroplasts. The scale bar in D is 20 µm and applies to all panels.

Results Direct observations Large variations in size and number of chloroplasts were observed in Mesodinium major (Fig. 1A-C and F). Cells ranged from 30-80 µm in length, and the chloroplasts were mainly located in the periphery of the cell (Fig. 1B and C). Attempts were made to enumerate the chloroplasts using confocal microscopy together with the Software “Imaris”. However, the large size of cells, and the large number of chloroplasts retained, made it difficult. The chloroplasts on the top of the preparation shadowed for the chloroplasts located on the other side of cell (Fig. 1A). Nevertheless, assuming that the chloroplasts are equally distributed in the cell, estimated number of chloroplasts ranged from 40-80 chloroplasts (see supplementary video). Well-fed cells of the ciliate had two macronuclei, one micronucleus and a centered prey nucleus (CPN) (Fig. D and E). Extra prey nuclei were rarely observed.

Figure 2. Experiment 1. Effects of irradiance. (A) Growth rates as a function of irradiance. The curve was numerically fitted to Michaelis-Menten kinetics. Y = 0.45*(X-21.5)/(19.99+(X-21.5)), R2 = 0.84. Data points represent means ±SE (n = 3). (B) Photosynthetic activity as a function of irradiance. The data was fitted to Michaelis-Menten kinetics. Y = 530.7*X/(48.9+X), R2 = 0.76. Data points represent means ±SE (n = 3). (C) Chlorophyll a as a function of irradiance. The data could neither be fitted linear line nor exponential decay kinetics; dotted line represents drawn trend line. (D) Ingestion rate as a function of irradiance. The data was fitted to a linear line, R2 = 0.0024. The line is not significant from zero, P = 0.9267.

Experiment 1. Effect of irradiance on the physiology of Mesodinium major The growth rates of Mesodinium major in well-fed cultures were highly affected by irradiance and the growth rates as a function of irradiance fitted Michaelis-Menten kinetics very well (R2 = 0.84; Fig. 2A). Maximum growth rates of 0.39 d-1 were achieved > 75 µmol photons m-2 s-1. At lower irradiances the growth was reduced, and a growth rate of only 0.07 d-1 was achieved at an irradiance of 25 µmol photons m-2 s-1. Photosynthetic activity also increased as a function of irradiance from 79.2 pg C cell-1 h-1 at 25 µmol photons m-2 s-1 to 323 pg C cell-1 h-1 at an irradiance of 200 µmol photons m-2 s-1. The data also fitted Michaelis-Menten kinetics well (R2 = 0.77, Fig. 2B). The photosynthetic activity of the prey, Teleaulax amphioxeia could not be fitted to Michaelis-Menten kinetics, because it kept increasing with irradiance (suppl. 1).

Figure 3. Experiment 1. Effect of irradiance.

(A) Growth as a function of photosynthetic activity. The curve was numerically fitted to Michaelis- Menten kinetics. Y = 0.508*(X-70)/(70.49+(X- 70)), R2 = 0.90. (B) Daily carbon obtained from photosynthetic activity (converted into daily uptake) and from ingestion of prey (ingestion rate d- 1 * 38 pgC (carbon in T. amphioxeia from Smith and Hansen (2007))), as a function of irradiance.

The cellular Chl a concentration of ~75 pg cell-1 was found the be the same within irradiances of 35-200 µmol photons m-2 s-1. However, at an irradiance of 25 µmol photons m-2 s-1 the cellular Chl a was significantly higher, 135 pg cell-1 (p< 0.005). M. major ingested T. amphioxeia at a rate of 9-15 preys predator-1 d-1, and ingestion rates were not affected by irradiance in the range of 25-200 µmol photons m-2 s-1 (Fig. 2D). Growth rates as a function of photosynthetic activity could be fitted to Michaelis-Menten kinetics (R2 = 0.99, Fig. 3A). The relationship indicated that maintenance requirements were ~70 pg C cell-1 h-1, which is equivalent 25-33 % of photosynthesis at its maximum growth rate (Fig. 3A). Comparison of the carbon uptake via ingestion and

8 photosynthesis of M. major indicates that the primary carbon source in M. major comes from photosynthesis at irradiances ≥ 50 µmol photons m-2 s-1 (Fig. 3B). At lower light levels, the contribution of carbon from ingestion was half of the amount from photosynthesis.

The Chl a-specific photosynthetic capacity (pg C pg Chl a-1 h-1) of M. major is similar to that of Teleaulax amphioxeia (Fig. 4A). Thus, the number of chloroplasts inside M. major can then be estimated from measured photosynthetic activity and Chl a of M. major compared with rates measured on Teleaulax amphioxeia (Fig. 4B). These calculations indicated that M. major may harbor between 200-350 T. amphioxeia. This was, however, not supported from the direct observations, where numbers were in the range of 40-80 chloroplasts.

Figure 4. Experiment 1. Effects of irradiance. (A) Chlorophyll a-specific photosynthetic capacity for M. major and T. amphioxeia as a function of light. (B) Estimated number of chloroplasts derived from measured photosynthetic activity and Chl a of M. major and T. amphioxeia.

Experiment 2: Effect of prey depletion/starvation on Mesodinium major Mesodinium major divided 5 times during 51 days of prey deprivation (Fig. 5A). Cell divisions occurred mostly in the beginning of starvation period (16 days) and only one division was recorded during the remaining period (last 40 days of the experiment). During the first 10 days, cells with a central prey nuclei (CPNs) fell from 70 % to 15% and this number remained stable at 10% after day 17 (Fig. 5B). Similarly, the photosynthetic activity fell from 119 pg C cell-1 h-1, during the first 10 days and leveled out at ~70 pg C cell-1 h-1 during the remaining experimental period (Fig. 5C). Levels of cellular Chl a were in the range of 60-90 pg Chl a cell-1, and no significant relationship with time was observed (Fig. 5D). This resulted in an increase Chl a-specific

Figure 5. Experiment 2. Prey deprivation. (A) Time course of numbers of cells during incubation of 51 days. Arrowheads indicate cell divisions of Mesodinium major. Dashed line indicates a mixture of the four cultures and resuspension in the 3 replicates. (B) Time course of percentage of cells with centered prey nucleus during incubation. The data was fitted to “One phase decay kinetics”. Y = (54.3-7.68)*exp(-0.16*X)+7.68, R2 = 0.9113. (C) Time course of photosynthetic activity during incubation. The data was fitted to “One phase decay kinetics”. Y = (124.8-67.14)*exp(-0.20*X)+67.14, R2 = 0.471. (D) Time course of amount of Chlorophyll a during incubation. The data was fitted to a linear line, R2 = 0.0058. The line is not significant from zero, P = 0.8336. photosynthetic capacity for the first two measurements (Day 0 and 1), and hereafter an unchanged capacity during the rest of the experimental period (Fig. 6A). An exponential relationship between photosynthetic activity and numbers of cells with CPNs (Fig. 6B) was found and the data could be fitted to exponential function (R2 = 0,784).

Experiment 3: Refeeding of Mesodinium major After 22 days of prey deprivation M. major was fed two different predator:prey ratios, 1:30 and 1:100, respectively, and cell abundances of both M. major and Teleaulax amphioxeia were then monitored over 10 days (Fig. 7 A). T. amphioxeia was completely depleted in both treatments at day 8, and M. major divided 5 times in both treatments during the 10 days of observation. Ingestion rates for treatment 1:30 varied ingestion rates increased from initially ~20-25 preys predator-1 d-1 during the first 3 days of incubation, but as the prey got depleted, these rates decreased, and at day 8 no prey cells were left (Fig. 7 B). Initial ingestion rates were

Figure 6. Experiment 2. Prey deprivation. (A) Time course of Chlorophyll a-specific photosynthetic capacity. The data was fitted to “One phase decay kinetics”. Y = (0.144- 0.073)*exp(-0.08*X)+ 0.073, R2 = 0.659. (B) Percentage of cells with centered prey nucleus (CPN) as a function of photosynthetic activity. The data was fitted to exponential growth kinetics. Y = 1.388*exp(0.03*X), R2 = 0.784. Please note that percentage of cells with CPN only goes up to 55%, and that photosynthetic activity was measured > 200 pgC cell-1 h-1 in exp. 1 at the same irradiance.

low (not measurable) in the 1:100 treatment. However, the following days, ingestion rates increased to reach 85 and 167 preys M. major-1 d-1. No differences in photosynthetic activity, Chl a, or percentage of cells with CPNs were found between the two treatments (Fig. 8 A, B and C). The photosynthetic activity and cellular Chl a increased from 50 to 200 pg C cell-1 h-1 and 60 to 120 pg Chl a cell-1, respectively, from day 0 to day 8, in both treatments. The percentage of cells with CPNs increased over the first four days, as long as prey was available; thereafter the number of CPNs decreased (Fig. 8 C).

Discussion The ecophysiology of Mesodinium major Dependence on light for growth The investigated strain of Mesodinium major depended on light for growth. At 25 µmol photons m-2 s-1 M. major was barely able to sustain itself with a growth rate of 0.07 d-1, even when offered prey in excess. Maximum growth rate of ~ 0.35 d-1 was achieved at irradiances > 75 µmol photons m-2 s-1 at 15°C. The data fitted very well to Michalis-mentens kinetics (r2 = 0.86; Fig. 2A). This strain of M. major was not able to grow in complete darkness. Only few data are available on the light responses of other planktonic Mesodinium spp,

Figure 7. Experiment 3. Refeeding Mesodinium major. M. major was refed at two different prey ratios 1:30 and 1:100 (predator:prey). (A) Time course of number of cells of M. major (solid lines, left y-axis) and Teleaulax amphioxeia (dashed lines, right y-axis) during incubation. (B) Time course of ingestion rate during incubation.

which utilize cryptophytes within the Teleaulax/Geminigera clade as prey and donor of cell organelles. A temperate strain of M. rubrum had a maximum growth rate of 0.23 and 0.49 d-1 at irradiances of 20 and 100 µmol photons m-2 s-1, respectively (see Table 1 for a comparison of M. major and M. rubrum) (Smith and Hansen 2007). An Antarctic strain of M. rubrum was able to grow at irradiances as low as ~2.5 µmol photons m-2 s-1 with a growth rate of ~ 0.11 d-1 at a temperature of 4°C (Johnson and Stoecker 2005). None of the investigated M. rubrum strains could sustain growth in complete darkness. Thus, even though both the investigated M. major and M. rubrum strains required light for growth, the requirements for light was higher for this isolate of M. major compared to the closely related M. rubrum strains studied. The light compensation point for growth was found to be 21.5 µmol photons m-2 s-1 for M. major (Fig. 2A). For an Arctic strain of M. rubrum the light compensation point was found to be as low as 0.5 µmol photons m-2 s-1 (table 1; (Moeller, Johnson et al. 2011). Data are also available for other Mesodinium spp, like the benthic heterotrophic M. pulex and the mixotrophic M. coatsi and M. chamaeleon. The heterotrophic M. pulex has been show to grow in both light (Irradiance of100 µmol photons m-2 s-1) and in complete darkness at growth rates of 1.41 d-1 and 1.19 d- 1, respectively, when supplied food in excess at 15 °C (Tarangkoon and Hansen 2011). The benthic mixotrophic M. coatsi has been shown to achieve a growth rate of 0.22 d-1 in complete darkness when supplied

the cryptophyte Rhodomonas sp, but this species also grows faster in the light (Kim, Kang et al. 2019). No data are available for mixotrophic M. chamaeleon in complete darkness, but this species has been shown to grow as fast as 0.35 d-1at an irradiance of 4 µmol photons m-2 s-1 at temperature of 18°C (Moeller and Johnson 2018, Kim, Kang et al. 2019). Thus, despite the limited data sets on light dependency on growth among Mesodinium spp, they point to some significant differences among the Mesosdinium spp. The Mesodinium spp that only utilize prey from the Teleaulax/Gemenigera/Plagioselmis clade seem to be dependent in light for growth, while the other benthic mixotrophic and heterotrophic Mesodinium spp that can utilize a larger variety of prey species are able to grow in the dark, although at a reduced growth rate. Light dependence on growth rates of other marine planktonic mixotrophic ciliates are sparse. A study of the mixotrophic Strombidium rassoulzadegani revealed growth rates (over three days) of ~0.6 d-1 in complete darkness (McManus, Schoener et al. 2012), and a maximum growth rate of 1.0 d-1 at an irradiance of 100 µmol photons m-2 s-1. S. rassoulzadegani died however if exposed to complete darkness for > 10 days, even if supplied with fresh food.

Figure 8. Experiment 3, Refeeding of Mesodinium major at two different ratios 1:30 and 1:100 (predator:prey). (A) Time course of photosynthetic activity during incubation. (B) Time course of Chlorophyll a during incubation. (C) Time course of percentage of cells with centered prey nucleus during incubation.

Dependence on light for prey ingestion, photosynthesis and cellular Chla Ingestion rates of Teleaulax amphioxeia by M. major were in the range of 8-15 cells predator-1 d-1, but no significant relationship to irradiance was observed (Fig 2D). Previous studies of a temperate strain of M. rubrum have found a maximum ingestion rates between 3-4 prey cells ciliate-1 d-1, if fed in excess at 15°C (Hansen and Fenchel 2006), and no significant difference was obtained at the two studied irradiances (20 and 100 µmol photons m-2 s-1). The mixotrophic M. chamaeleon and M. coatsi were found to have maximum ingestion rates similar to M. major, at ~10-25 prey cells ciliate-1 day-1, whereas the heterotropic M. pulex ingested up to 30 prey cells ciliate-1 day-1 at 15 °C (Tarangkoon and Hansen 2011, Moeller and Johnson 2018, Kim, Kang et al. 2019).

Photosynthetic activity increased as a function of irradiance and measured values ranged from 89.2 to 323 pg C cell-1 h-1 at 15°C, at irradiances ranging of 25 - 200 µmol photons m-2 s-1 and prey concentrations of ~4,000 cells mL-1. The data could be fitted to Michalis-Menten kinetics, suggesting a saturation of photosynthesis > 75 µmol photons m-2 s-1. This is, to our knowledge, one of the highest photosynthetic rate measured for a ciliate. It is only exceeded by the large Laboea strobila, in which a photosynthetic rate of 465 pg C cell-1 h-1 was achieved at an irradiance of 100 µmol photons m-2 s-1 and a temperature at 12°C on cells picked from natural plankton samples (Stoecker, Taniguchi et al. 1989). For comparison, published maximum photosynthetic rates ~88 pg C cell-1 h-1 have been found for M. rubrum at irradiances of >200 µmol photons m-2 s-1 (table 1; (Stoecker, Putt et al. 1991). Photosynthetic rates of benthic mixotrophic M. chamaeleon has been shown to be much lower, ~6.3 pg C cell-1 h-1 (Moestrup, Garcia-Cuetos et al. 2012). Maximum photosynthetic rates are only available for a few other mixotrophic ciliates. In the genus Strombidium, S. conicum and S. riticulatum maximum photosynthetic rates at 12 and 31 pg C cell-1 h-1, respectively, have been reported (Jonsson 1987). The high photosynthetic rates of M. major may not be surprising. It is a very large ciliate and it contains a larger number of chloroplasts. The maintenance requirements for M. major could be estimated from a plot of growth versus photosynthetic activity calculated for M. major to be ~70 pg C cell-1 h- 1 (fig 3A; table 1). The current literature does not provide this information for any other ciliates.

If we compare the obtained carbon from photosynthetic activity and from ingestion of prey, the photosynthetic activity account for ten times more carbon than ingestion (Fig 3B). At lower irradiances the two carbon sources are however closer to one another, suggesting that ingestion at low irradiances play a significant role in low irradiance. For comparison, M. rubrum obtains a similar contribution of carbon from photosynthesis ˜95% and ˜80 % at irradiance 100 and 20 µmol m-2 s-1, respectively (table 1; (Hansen and Fenchel 2006, Smith and Hansen 2007). The benthic M. chamaeleon on the other hand gets less of its carbon from photosynthesis (0-70 %)(Moeller and Johnson 2018).

Cellular Chl a levels decreased from 130 to 80 pg Chl a cell-1 as the irradiances increased from 25 µmol photons to 200 µmol photons m-2s-1, indicating possible photoacclimation in M. major (Fig. 2C). Similar

15 observations have been found in an Antarctic strain of M. rubrum. Here, the cellular levels of Chl a decreased from 60 to 30 pg Chl a cell-1 when irradiances increased from 2.5 to 55 µmol photons m-2 s-1, respectively (Johnson and Stoecker 2005). Thus, it is possible that M. major may also be able to photoacclimate. More and detailed studies of the photosynthetic apparatus are however necessary, like the ones carried out on M. rubrum (Moeller, Johnson et al. 2011). The dependency of photosynthetic activity is likewise illustrated by the Chl a- specific photosynthetic capacity. Here we show that M. major is able to preserve the capacity of the chloroplasts to the same extent as T. amphioxeia (Fig. 4A). For comparison, the mixotrophic Strombidium rassoulzadegani was found be able to maintain ~50 % of Chl a-specific photosynthetic capacity compared to its prey (McManus, Schoener et al. 2012). It is possible that M. rubrum has the same ability, but such information is not available in the literature.

It was impossible to count the total the number of chloroplasts inside the large M. major cells directly. However, enumeration of the number chloroplasts in one-half of the cell could be obtained, and assuming that the cell contains the double amount, estimates suggest ~40-80 chloroplasts cell-1. If instead the photosynthetic activity was used to estimate the number of chloroplasts inside M. major, the results indicate that the cells would have > 120 cryptophytes at an irradiance of 100 µmol photons m-2 s-1 (Fig. 4B). However, such higher numbers did not match at all the number of chloroplasts counted in microscope. Similar results were found using Chl a cell-1. A possible reason for this mismatch could be that the cryptophyte chloroplasts enlarge inside the cell, as reported in M. rubrum (Hansen and Fenchel 2006).

Responses of Mesodinium major to prey deprivation and refeeding The studied strain of Mesodinium major coped very well with prey starvation. Within 5 days of prey starvation, the Chl a-specific photosynthetic capacity decreased by 50%, but thereafter remained constant throughout 51 days of prey starvation. In that period, the culture went through 5 cell divisions. Similar results have been found for several strains of M. rubrum. This ciliate likewise is able to starve for long when subjected to prey starvation, and be photosynthetic active for > two months (table 1; (Johnson and Stoecker 2005, Smith and Hansen 2007, Myung, Kim et al. 2013). The ability to survive without prey for this amount of time gives organisms great advances to other similar organisms, and could explain the annual presence that we find in many coastal areas (like in Helsingør harbor, personal observations and at Helgoland (Yang, Löder et al. 2014). Starvation of M. pulex and M. chamaeleon/coatsi shows up to two weeks tolerance with 0-1 cell division (Tarangkoon and Hansen 2011, Moeller and Johnson 2018). Starvation experiments with other mixotrophic ciliates are sparse. In the case of the ciliate, S. capitatum, little tolerance to prey deprivation was observed. After only 40 hours, the cell populations were declining quickly in numbers (Stoecker and Silver 1990).

M. major was able to recover quickly from 22 days of starvation. After only 6 days of refeeding, photosynthetic activities were back to the levels as before prey starvation. Cellular Chl a increased from 58 to 122 pg cell-1 after 8 days of exposure to prey cells (Fig. 8B). This is surprising since we did not find a decrease in the cellular

Chl a of the ciliates during starvation. However, the data could suggest a higher level of Chl a in “freshly caught” chloroplasts. Also, M. major quickly started to ingest prey cells when refed, and ingestion rates were very high (> 150 prey predator-1 day-1) after some days of refeeding (fig 7B). After six days of refeeding both prey treatments had depleted all available prey (Fig. 7A). M. rubrum has previously been subjected to a similar starvation exposure, and here it took M. rubrum 10-13 days to deplete all available prey (Kim, Drumm et al. 2017). Similar, M. rubrum was found to ingest ~1.3 prey predator-1 h-1, after being starved for 14 days, but this high ingestion rate was only found within the first hours after refeeding, where after ingestion rates rapidly decreased (Gustafson, Stoecker et al. 2000). For comparison, M. major could maintain ingestion rates of 7.0 prey cells predator-1 h-1 for > 24 hours.

Control of retained chloroplasts and nuclei during prey starvation A positive correlation between the centered prey nucleus (CPN) and photosynthetic performance in M. major was found during the starvation experiment. The correlation was similar to result found for M. rubrum (Kim, Drumm et al. 2017). This had some implications for the photosynthetic rates in this study. For instance, in the case of the of the starvation experiment the M. major culture was initiated when the cells were not at preforming optimally (119 pg C cell-1 h-1), compared to photosynthetic activity measured at day 8 in the acclimation experiment leading up the starvation experiment (238 pg C cell-1 h-1). This could very well be correlated with that the percentage of cells with CPNs were 55 % at day 0. Thus, a comparison with a similar study on M. rubrum by Kim et al. (2017) should take this into account. For both M. major and M. rubrum a decrease in the percentage of cells with a CPN from 55 to 10 of cells results in a 50 % reduction of the photosynthetic activity. We therefore suggest that the relationship between CPN and photosynthetic activity in M. major is very similar to M. rubrum.

Despite very high ingestion rates in the M. major when fed at high prey concentrations, no difference in the percentage of acquired CPNs within the two prey treatments was observed (Fig. 8C). Unlike M. rubrum, the culture of M. major only managed to obtain a percentage of approximately 50 % of cells with CPN, whereas M. rubrum obtain 90 % in the same time interval (4 days) (Kim, Drumm et al. 2017). This could suggest that M. rubrum has a more quickly retention of the CPN than M. major. The reason for this difference could be explained a difference in the ingestion and retention of prey nuclei, since we rarely found extra prey nuclei (EPNs) in M. major. We can only speculate if the missing observations of EPNs in M. major is caused by M. major not retaining them or if it is an effect of their size, shadowing in the epifluorescence microscope.

Conclusion Mesodinium major functions quite similar to M. rubrum. However, it needs more light to sustain and to grow. Chl a-specific photosynthesis capacity was similar to that of Teleaulax amphioxeia at all irradiances, and some indications of possible photoacclimation were found. The primary carbon source in M. major is from photosynthesis, similar to what has been shown for M. rubrum. M. major is well adapted to prey starvation

17 and tolerates 51 days of prey deprivation without any mortality. Refeeding after prey starvation leads high ingestion rates for short time. Thus, M. major differs from M. rubrum by having a higher demand of light, a photosynthetic activity which is up to four times higher than M. rubrum, and a maximum ingestion rate that exceeds M. rubrum by more than a tenfold.

Acknowledgements This research was founded by the DDF 2 project, project number 4181-00484. The authors would like to sincerely thank Niels Daugbjerg for proving access and assistance to the use of his epifluorescence microscope.

References Garcia-Cuetos, L., O. Moestrup and P. J. Hansen (2012). "Studies on the genus Mesodinium II. Ultrastructural and molecular investigations of five marine species help clarifying the taxonomy." J Eukaryot Microbiol 59(4): 374-400.

Gustafson, D. E., D. K. Stoecker, M. D. Johnson, W. F. Van Heukelem and K. Sneider (2000). "Cryptophyte algae are robbed of their organelles by the marine ciliate Mesodinium rubrum." Nature 405(6790): 1049-1052.

Hansen, P. J. and T. Fenchel (2006). "The bloom-forming ciliate Mesodinium rubrum harbours a single permanent endosymbiont." Mar Biol Res 2(3): 169-177.

Herfort, L., K. Maxey, I. Voorhees, H. M. Simon, K. Grobler, T. D. Peterson and P. Zuber (2017). "Use of highly specific molecular markers reveals positive correlation between abundances of Mesodinium cf. major and its preferred prey, Teleaulax amphioxeia, during red water blooms in the Columbia river estuary." Journal of Eukaryotic Microbiology 64(6): 740-755.

Herfort, L., T. D. Peterson, L. A. McCue, B. C. Crump, F. G. Prahl, A. M. Baptista, V. Campbell, R. Warnick, M. Selby and G. C. Roegner (2011). "Myrionecta rubra population genetic diversity and its cryptophyte chloroplast specificity in recurrent red tides in the Columbia River estuary."

Jakobsen, H. H. and P. J. Hansen (1997). "Prey size selection, grazing and growth response of the small heterotrophic dinoflagellate Gymnodinium sp. and the ciliate Balanion comatum--a comparative study." Marine ecology progress series 158: 75-86.

Johnson, M. D. and D. J. Beaudoin (2019). "The genetic diversity of plastids associated with mixotrophic oligotrich ciliates." Limnology and Oceanography.

Johnson, M. D., D. J. Beaudoin, A. Laza-Martinez, S. T. Dyhrman, E. Fensin, S. Lin, A. Merculief, S. Nagai, M. Pompeu, O. Setälä and D. K. Stoecker (2016). "The Genetic Diversity of Mesodinium and Associated Cryptophytes." Frontiers in Microbiology 7(2017).

Johnson, M. D. and D. K. Stoecker (2005). "Role of feeding in growth and photophysiology of Myrionecta rubra." Aquat Microb Ecol 39(3): 303-312.

Jonsson, P. (1987). "Photosynthetic assimilation of inorganic carbon in marine oligotrich ciliates (Ciliophora, Oligotrichina)." Mar Microb Food Webs 2: 55-68.

Kim, M., K. Drumm, N. Daugbjerg and P. J. Hansen (2017). "Dynamics of sequestered cryptophyte nuclei in Mesodinium rubrum during starvation and refeeding." Frontiers in Microbiology 8: 423.

Kim, M., M. Kang and M. G. Park (2019). "Growth and Chloroplast Replacement of the Benthic Mixotrophic Ciliate Mesodinium coatsi." Journal of Eukaryotic Microbiology 66(4): 625-636.

Kim, M. and M. G. Park (2019). "Unveiling the hidden genetic diversity and chloroplast type of marine benthic ciliate Mesodinium species." Scientific Reports 9(1): 1-10.

Lindholm, T. (1985). "Mesodinium rubrum-a unique photosynthetic ciliate." Advances in aquatic microbiology 3: 1-48.

McManus, G. B., D. M. Schoener and K. Haberlandt (2012). "Chloroplast symbiosis in a marine ciliate: ecophysiology and the risks and rewards of hosting foreign organelles." Molecular and functional ecology of aquatic microbial symbionts: 55.

Mitra, A., K. J. Flynn, U. Tillmann, J. A. Raven, D. Caron, D. K. Stoecker, F. Not, P. J. Hansen, G. Hallegraeff and R. Sanders (2016). "Defining planktonic protist functional groups on mechanisms for energy and nutrient acquisition: incorporation of diverse mixotrophic strategies." Protist 167(2): 106-120.

Moeller, H. V. and M. D. Johnson (2018). "Preferential plastid retention by the acquired phototroph Mesodinium chamaeleon." Journal of Eukaryotic Microbiology 65(2): 148-158.

Moeller, H. V., M. D. Johnson and P. G. Falkowski (2011). "Photoacclimation in the phototrophic marine ciliate Mesodinium rubrum (Ciliophora)." J Phycol 47(2): 324-332.

Moestrup, O., L. Garcia-Cuetos, P. J. Hansen and T. Fenchel (2012). "Studies on the genus Mesodinium I: ultrastructure and description of Mesodinium chamaeleon n. sp., a benthic marine species with green or red chloroplasts." J Eukaryot Microbiol 59(1): 20-39.

Myung, G., H. S. Kim, J. W. Park, J. S. Park and W. Yih (2013). "Sequestered plastids in Mesodinium rubrum are functionally active up to 80 days of phototrophic growth without cryptomonad prey." Harmful Algae 27: 82-87.

Nam, S. W., W. Shin, M. Kang, W. Yih and M. G. Park (2015). "Ultrastructure and molecular phylogeny of Mesodinium coatsi sp. nov., a benthic marine ciliate." Journal of Eukaryotic Microbiology 62(1): 102-120.

Packard, T., D. Blasco and R. Barber (1978). Mesodinium rubrum in the Baja California upwelling system. Upwelling ecosystems, Springer: 73-89.

Park, M. G., S. Kim, H. S. Kim, G. Myung, Y. G. Kang and W. Yih (2006). "First successful culture of the marine dinoflagellate Dinophysis acuminata." Aquatic Microbial Ecology 45(2): 101-106.

Reguera, B., L. Velo-Suárez, R. Raine and M. G. Park (2012). "Harmful Dinophysis species: A review." Harmful Algae 14: 87-106.

Rial, P., A. Laza-Martínez, B. Reguera, N. Raho and F. Rodríguez (2015). "Origin of cryptophyte plastids in Dinophysis from Galician waters: results from field and culture experiments." Aquatic Microbial Ecology 76(2): 163.

Skovgaard, A., P. J. Hansen and D. K. Stoecker (2000). "Physiology of the mixotrophic dinoflagellate Fragilidium subglobosum. I. Effects of phagotrophy and irradiance on photosynthesis and carbon content." Marine Ecology Progress Series 201: 129-136.

Smith, M. and P. J. Hansen (2007). "Interaction between Mesodinium rubrum and its prey: importance of prey concentration, irradiance and pH." Mar Ecol Prog Ser 338: 61-70.

Stoecker, D., M. Putt, L. Davis and A. Michaels (1991). "Photosynthesis in Mesodinium rubrum: species- specific measurements and comparison to community rates." Mar Ecol Prog Ser 73: 245-252.

Stoecker, D. and M. Silver (1990). "Replacement and aging of chloroplasts in Strombidium capitatum (Ciliophora: Oligotrichida)." Marine Biology 107(3): 491-502.

Stoecker, D. K., A. Taniguchi and A. E. Michaels (1989). "Abundance of autotrophic, mixotrophic and heterotrophic planktonic ciliates in shelf and slope waters." Marine ecology progress series. Oldendorf 50(3): 241-254.

Tarangkoon, W. and P. J. Hansen (2011). "Prey selection, ingestion and growth responses of the common marine ciliate Mesodinium pulex in the light and in the dark." Aquat Microb Ecol 62(1): 25-38.

Yang, J., M. G. J. Löder, G. Gerdts and K. H. Wiltshire (2015). "Structural composition and temporal variation of the ciliate community in relation to environmental factors at Helgoland Roads, North Sea." Journal of sea research 101: 19-30.

Yang, J., M. G. J. Löder and K. H. Wiltshire (2014). "A survey of ciliates at the long-term sampling station “Helgoland Roads”, North Sea." Helgoland Marine Research 68(2): 313-327.

Supporting Information

S1. Fig Photosynthetic activity by Teleaulax amphioxeia as a function of irradiance

III

fmicb-08-00423 March 18, 2017 Time: 15:45 # 1

ORIGINAL RESEARCH published: 21 March 2017 doi: 10.3389/fmicb.2017.00423

Dynamics of Sequestered Cryptophyte Nuclei in Mesodinium rubrum during Starvation and Refeeding

Miran Kim1*, Kirstine Drumm1, Niels Daugbjerg2 and Per J. Hansen1

1 Marine Biological Section, Department of Biology, University of Copenhagen, Helsingør, Denmark, 2 Marine Biological Section, Department of Biology, University of Copenhagen, Copenhagen, Denmark

The marine mixotrophic ciliate Mesodinium rubrum is known to acquire chloroplasts, mitochondria, nucleomorphs, and nucleus from its cryptophyte prey, particularly from species in the genera, Geminigera and Teleaulax. The sequestered prey nucleus and chloroplasts are considered to support photosynthesis of M. rubrum. In addition, recent

Edited by: studies have shown enlargement of the retained prey nucleus in starved M. rubrum Senjie Lin, and have inferred that enlargement results from the fusion of ingested prey nuclei. University of Connecticut, USA Thus far, however, little is known about the mechanism underlying the enlargement of Reviewed by: the prey nucleus in M. rubrum. Here, we conducted starvation and refeeding studies Matthew David Johnson, Woods Hole Oceanographic to monitor the fate of prey nuclei acquired by M. rubrum when feeding on Teleaulax Institution, USA amphioxeia and to explore the influence of the retained prey nucleus on photosynthesis Dajun Qiu, South China Sea Institute Of of M. rubrum. Results indicate that enlargement of the prey nucleus does not result Oceanology (CAS), China from fusion of nuclei. Furthermore, the enlarged prey nucleus does not appear to divide Francisco Rodriguez, during cell division of M. rubrum. The presence of a prey nucleus significantly affected Instituto Español de Oceanografía, Spain photosynthetic performance of M. rubrum, while the number of retained chloroplasts *Correspondence: had little influence on rate of carbon fixation. We interpret results within the context of a Miran Kim model that considers the dynamics of ingested prey nuclei during division of M. rubrum. [email protected] Keywords: nucleus enlargement, photosynthesis, sequestered chloroplasts, sequestered nucleus, Teleaulax Specialty section: amphioxeia This article was submitted to Aquatic Microbiology, a section of the journal INTRODUCTION Frontiers in Microbiology Received: 14 November 2016 Mesodinium rubrum (=Myrionecta rubra) is a common ciliate in coastal waters worldwide, where Accepted: 28 February 2017 it sometimes causes red tides (Taylor et al., 1971; Lindholm, 1985). It is an obligate mixotroph, Published: 21 March 2017 requiring both light and prey uptake for sustained growth and survival (Gustafson et al., 2000; Yih Citation: et al., 2004; Johnson and Stoecker, 2005; Hansen and Fenchel, 2006). Growth of M. rubrum is to Kim M, Drumm K, Daugbjerg N and a large extent phototrophic and closely linked to irradiance (Johnson and Stoecker, 2005; Johnson Hansen PJ (2017) Dynamics et al., 2006; Smith and Hansen, 2007; Moeller et al., 2011). Under culture conditions, M. rubrum of Sequestered Cryptophyte Nuclei feeds specifically on cryptophytes belonging to the genera Geminigera and Teleaulax (Hansen and in Mesodinium rubrum during Starvation and Refeeding. Fenchel, 2006; Park et al., 2007; Myung et al., 2011; Hansen et al., 2012; Raho et al., 2014). Front. Microbiol. 8:423. It has long been known that Mesodinium rubrum contains chloroplasts of cryptophyte origin doi: 10.3389/fmicb.2017.00423 (Taylor et al., 1971; Hibberd, 1977). Recent studies have shown that these chloroplasts are

Frontiers in Microbiology| www.frontiersin.org 1 March 2017| Volume 8| Article 423 fmicb-08-00423 March 18, 2017 Time: 15:45 # 2

Kim et al. Dynamics of Acquired Prey Nucleus in Mesodinium rubrum

genetically similar to the prey on which M. rubrum is fed, MATERIALS AND METHODS indicating that chloroplasts are sequestered from the prey (Johnson et al., 2006, 2007; Hansen et al., 2012; Myung et al., Cultures 2013). Furthermore, cross-feeding experiments in which prey Cultures of Mesodinium rubrum (MBL-DK2009) and Teleaulax were switched from one species to another led to the replacement amphioxeia (SCCAP K-0434; SCCAP) were established using of chloroplasts in M. rubrum, depending upon the available single cells isolated from Helsingør harbor, Denmark, in 2009. prey species (Hansen et al., 2012; Raho et al., 2014). In starved Both species were grown in f/2 medium based on autoclaved M. rubrum, sequestered prey chloroplasts have been shown to seawater (Guillard, 1975) with a salinity of 30 and maintained divide along with cell divisions of the ciliate (Johnson et al., in 24-well tissue culture plates (TPP, Switzerland). Both species 2006; Hansen and Fenchel, 2006; Kim et al., 2016). Nevertheless, were grown at 15◦C ± 1.0 in a temperature regulated room, M. rubrum requires continuous acquisition of new chloroplasts under a photon irradiance of 100 µmol photons m−2 s−1 and other cell organelles acquired through feeding for sustained (PAR, 400–700 nm), and on a light:dark cycle of 14:10. Light growth. was provided by cool white fluorescent tubes (OSRAM, 58W, Early works found that M. rubrum contains a single 840). Irradiance was measured (in seawater) at the level of enlarged nucleus of cryptophyte origin, along with cryptophyte incubation flasks using a light meter equipped with a spherical chloroplasts, other cell organelles, and cytoplasm (Taylor quantum sensor (ULM and US-SQS/L, Walz GmbH, Germany). et al., 1969, 1971; Hibberd, 1977; Oakley and Taylor, 1978). T. amphioxeia was supplied as prey at a predator:prey ratio of This condition was believed to represent an incomplete approximately 10:1 when M. rubrum was transferred weekly to endosymbiont, and the enlarged cryptophyte nucleus was new medium. referred to as a ‘symbiont nucleus’ (Hibberd, 1977; Oakley and Taylor, 1978). Later, however, it was verified that the symbiont nucleus, like the cryptophyte chloroplasts, is acquired Experiment 1: Starvation of Well-Fed by M. rubrum after feeding on prey (Gustafson et al., 2000; Mesodinium rubrum Johnson et al., 2007; Johnson, 2011; Hansen et al., 2012; Kim The first experiment was designed to monitor the change in et al., 2016) and was termed ‘kleptokaryon’ by Johnson et al. photosynthetic performance and the number, size and position (2007). For simplicity, the ‘enlarged acquired prey nucleus’ of prey nuclei during starvation of M. rubrum. M. rubrum cells will be called the ‘symbiont nucleus’ in the remainder of the were kept well-fed for 2 weeks by adding sufficient prey every introduction. 3 days to a culture grown in a 750-ml tissue culture flask (TPP, The single symbiont nucleus observed in well-fed M. rubrum Switzerland). After 2 weeks of frequent feeding, the M. rubrum cultures (Johnson et al., 2007) is eventually lost following culture was allowed to deplete the prey, and absence of prey cell divisions of the ciliate subjected to prolonged starvation in the culture was confirmed under an inverted microscope (Johnson and Stoecker, 2005; Johnson et al., 2007; Kim et al., at x40 magnification (Olympus CK2, Japan). A portion of 2016). However, the symbiont nucleus was still observed in all the well-fed, but prey-free, M. rubrum and stock culture of cells after the first cell division in prey starved experiments, T. amphioxeia were then added to a 750 ml tissue culture indicating that the acquired prey nucleus divide at least one flask to achieve an initial predator:prey ratio to 1:5 and a time inside M. rubrum (Hansen and Fenchel, 2006). Gustafson M. rubrum cell concentration of ∼200 ml−1. From this stock et al.(2000), as well as recent papers by Kim et al.(2016) culture, a 150-ml aliquot was transferred to each of three and Nam et al.(2016) showed that well-fed M. rubrum 270-ml tissue culture flasks. The flasks were placed on a shelf retained additional prey nuclei which were smaller than the with light coming from the side at an irradiance of 60 µmol symbiont nucleus and of size similar to the nucleus of the photons m−2s−1. Position of the flasks was changed sequentially prey species. Also, it has been observed that smaller prey once a day to minimize difference in light exposure between nuclei become enlarged over time during prey starvation in the flasks. Subsamples for cell enumeration and assessment M. rubrum (Johnson et al., 2007; Kim et al., 2016; Nam of photosynthetic performance were withdrawn from each et al., 2016). The interpretation made by the authors of these flask on 13 occasions during the experiment: on Day 0, 2, recent studies was that acquired prey nuclei fused to make 4, 6, 8, 11, 14, 16, 18, 20, 23, 26, and 29 (Figure 1). For the symbiont nucleus. However, this interpretation was not confocal microscopy, a single Day 0 sample was taken prior to experimentally tested and the exact mechanism underlying distributing stock culture to replicate flasks, with samples taken enlargement of the acquired cryptophyte nuclei in M. rubrum from replicate flasks on all other days. pH was monitored at remains unknown. each sampling occasion directly in the flasks with a SenTix R 41 In the present study, the fate of prey nuclei sequestered by pH electrode (WTW, Germany) connected to a pH meter M. rubrum was monitored during prey starvation and refeeding (WTW, pH 3210, Germany), and calibrated with pH 7 and experiments using confocal microscopy. Changes in size and pH 10 standard buffers (WTW, Technischer, NIST, buffers). To position of prey nuclei inside the ciliate were determined, and avoid physiological effects of elevated pH in laboratory cultures evidence of nuclear division or fusion was noted. Furthermore, (Hansen, 2002), half the volume of each experimental cultures the relationship between the presence of a retained prey nucleus of M. rubrum was removed and replaced with fresh f/2 medium and the photosynthetic efficiency and growth of M. rubrum was on Day 8, 11, 14, and 20, when pH values were approaching studied. to 8.5.

Frontiers in Microbiology| www.frontiersin.org 2 March 2017| Volume 8| Article 423 fmicb-08-00423 March 18, 2017 Time: 15:45 # 3

Kim et al. Dynamics of Acquired Prey Nucleus in Mesodinium rubrum

FIGURE 1 | Schematic illustration of experimental design used for starvation and refeeding experiments, respectively. The two experiments were conducted over 7 weeks. Experiment 1, starvation of well-fed Mesodinium rubrum, was carried out for 4 weeks, with the starved cells at the end of that experiment immediately use in Experiment 2, refeeding of starved M. rubrum.

Cell Abundance and Growth Rate Photosynthetic activity (PA, pg C cell−1 h−1) was calculated as Aliquots (2.3 ml) withdrawn from each flask were fixed with follows: acid Lugol’s solution (final concentration 1%). Abundance 14 of M. rubrum and T. amphioxeia was enumerated using PA = DPM × [DIC]/ C × h × N a Sedgewick-Rafter chamber under the inverted microscope − where DPM is disintegrations min 1 (in 1.9 ml) in the light (Olympus CK40) at 100X and 200X. At least 400 cells were corrected for dark value, DIC is the concentration of dissolved enumerated. Growth rates were calculated during the exponential − inorganic carbon (pg C ml 1), 14C is the specific activity portion of the growth phase using the following exponential − − (disintegrations min 1 ml 1), h is the incubation time, and N growth equation: is the number of cells in the vial (1.9 ml). DIC concentrations were measured within a few hours using a total organic carbon µ = lnN1 − lnN0/t1 − t0 analyzer (TOC-L, Shimadzu, Japan). Confocal Microscopy where N1 and N0 are cell concentrations at time t1 and time 0, Location, number, and changes in size of prey nuclei inside respectively, and t1-t0 is the time interval between samplings. M. rubrum were studied by confocal microscopy. Nuclei were stained using the fluorescent nuclear stain Hoechst 33258 Measurement of Photosynthetic Activity (14C) (Invitrogen, USA) and plasma membrane stain using CellMask Two 2-ml aliquots were withdrawn from each flask, transferred Green (Life technologies, Carlsbad, CA, USA). Subsamples to each of two 20-ml glass scintillation vials, and used for (5–10 ml) taken from each flask were fixed in 1% glutaraldehyde measurements of photosynthesis. Twenty microliter of NaH14 (EMD Millipore, USA) at 4◦C for 1 h. Fixed samples were −1 −1 CO3 stock solution (specific activity = 100 µCi ml ; Carbon-14 stained with a combination of 25 µg ml Hoechst 33258 Centralen, Denmark) was added to each vial. One vial of and 0.25 X CellMask for 15 min in a dark chamber, then each pair was incubated for 3 h in the same place as the filtered through a 0.2 µm black polycarbonate membrane filter experimental flask, and the other vial was kept in complete (Frisenette, Denmark), and finally washed with fresh autoclaved darkness by wrapping in aluminum foil. After incubation, seawater to remove excess dye. A drop of immersion oil placed a 100 µl sub-sample was withdrawn from each vial and on both sides of a membrane filter was used to attach the added to a new vial containing 200 µl phenylethylamine for filter to the microscope slide and coverglass. Nuclear size was measurements of specific activity (Skovgaard et al., 2000). measured directly from images taken at 600X magnification using The remaining 1.9 ml were acidified with 2 ml 10% glacial a FViewII digital camera (Olympus Soft Imaging System, Tokyo, acetic acid in methanol, and evaporated overnight at 60◦C Japan) linked to the inverted microscope (Olympus IX81, Japan) to remove all inorganic carbon. The residue in the vial was equipped with a disk-spinning unit (DSU, Olympus, Japan). re-dissolved in 2 ml Milli-Q water before adding 10 ml Epifluorescence micrographs of stained M. rubrum cells were of scintillation cocktail (Insta-Gel Plus, Packard, USA). All taken at 1,000X magnification using a digital camera coupled vials were vigorously shaken and then analyzed using a to the Olympus BX51 microscope equipped with differential liquid scintillation counter (Tri-Carb 2910 TR, Perkin-Elmer). interference contrast. Twenty cells were examined for each

Frontiers in Microbiology| www.frontiersin.org 3 March 2017| Volume 8| Article 423 fmicb-08-00423 March 18, 2017 Time: 15:45 # 4

Kim et al. Dynamics of Acquired Prey Nucleus in Mesodinium rubrum

sample. 3D images were generated using IMARIS software (diameter 2.08 µm ± 0.04 µm, Table 1) of the prey, Teleaulax program (Bitplane, Zürich, Switzerland) to assess the number and amphioxeia,(Figures 2A,C). We will subsequently refer to the volume of chloroplasts of M. rubrum. former and latter type of ingested prey nuclei as “CPN” and “extra prey nucleus” (EPN), respectively (Figure 2C). Experiment 2: Refeeding of Starved After staining with a combination of Hoechst 3325 and Mesodinium rubrum CellMask Green, the ciliate micronucleus always emitted stronger The second experiment was designed to monitor the changes in fluorescence than the two macronuclei (Figures 2B,C). EPNs the number, size and position of prey nuclei upon refeeding and were typically spherical (Figure 2C), while CPNs varied from subsequent prey starvation. After taking subsamples on Day 29, spherical to irregular in shape (Figure 2C). EPNs were never the three cultures from the first experiment were poured together clustered close together, and none of 181 EPNs examined during in a 750-ml tissue culture flask and then distributed to three the experiment showed evidence of fusing with another EPN 750-ml tissue flasks. T. amphioxeia and new f/2 medium was or a CPN. Dividing M. rubrum cells were common in stained added to each of the three flasks to give predator:prey ratios of preparations, but none of 388 CPNs examined for the experiment 1:2.5, 1:5, and 1:10 and then a 150-ml subsample of each flask appeared to be undergoing nuclear division. The location, size, was transferred to each of three 270-ml tissue culture flasks. and shape of CPNs present in dividing M. rubrum cells (Figure 9) The triplicate flasks for each treatment were maintained as in were indistinguishable from that of CPNs occurring in non- Experiment 1, with subsamples for estimating cell abundance dividing cells. (2.3 ml) withdrawn on 10 occasions during the experiment (Day Cell Divisions 0, 0.5, 1, 4, 6, 8, 10, 13, 16, and 19) (Figure 1). Samples for measuring the size of prey nuclei inside of M. rubrum (5–10 ml) The culture of Mesodinium rubrum that was mixed with were taken on Day 0.5 to Day 19, with Day 29 data from Teleaulax amphioxeia at an initial ratio of 1:5 had almost depleted Experiment 1 used as Day 0 data for Experiment 2. All nine flasks the cryptophyte prey at Day 8, and the prey were depleted below < −1 Figure 3A were supplied with fresh f/2 medium after removal of half of the detection limit ( 1 cell ml ) by Day 11 ( ). During the ‘old’ medium on Day 10 for the same reason as above. first 8 days of the incubation, M. rubrum divided every second day (µ = 0.37 ± 0.01 d−1). After that period, growth declined. Statistical Analyses From Day 8 to Day 14, M. rubrum divided every third day (µ = 0.25 ± 0.02 d−1), and from Day 14 to end of the experiment Relationships of photosynthesis with prevalence of the centered − cells stopped dividing (µ = 0.10 ± 0.02 d 1). A total of eight prey nucleus (CPN; see below) and number of chloroplasts cell divisions were observed, four of which occurred in the period were analyzed and plotted using non-linear regression analysis without available prey. (Sigma Plot v. 10.0). Data reported in the text as means are ± given standard error of the mean (SE). Error bars provided in Number and Volume of Chloroplasts figures also represent SE. Even though M. rubrum divided several times during the experiment, the number of chloroplasts cell−1 remained relatively constant at 20 ± 0.39 cell−1 (n = 12) for 26 days, despite RESULTS the fact that prey cells were depleted after Day 8 (Figure 3B). After Day 26, the number of chloroplasts decreased rapidly to Experiment 1: Starvation of Well-Fed reach 14 ± 0.10 cell−1 at the end of the experiment. Chloroplast Mesodinium rubrum volume cell−1 remained more or less steady from Day 0 to Day 18 Nuclei of Prey and Well-Fed Mesodinium rubrum (444 ± 33 µm3 cell−1; n = 9), thereafter gradually decreasing to Well-fed Mesodinium rubrum cells (Day 0 to Day 11) contain a mean of 274 ± 8 µm3 cell−1 on the last day of the experiment. two ciliate macronuclei and one ciliate micronucleus, all of which were closely positioned at the center of the cell (Figure 2B). Changes in Prevalence and Linear Dimensions of The two macronuclei (diameter 3.54 µm ± 0.07 µm; Table 1) Sequestered Prey Nuclei Inside M. rubrum Cells were placed close to each other in the middle of the cell, More than 90% of the well-fed M. rubrum cells possessed a CPN while the micronucleus (diameter 2.63 µm ± 0.09 µm) was at the initiation of the experiment (Figure 3C) and a similar located just posterior to the two macronuclei. Well-fed cells of percentage was observed for the following 6 days (92.4 ± 0.2%; M. rubrum contained additional nuclei of cryptophyte origin. n = 4). The proportion of ciliates having a CPN decreased A solitary cryptophyte nucleus (diameter 4.35 µm ± 0.04 µm; to ∼70% from Day 8 to 11 as prey cells were depleted, in Table 1) was located in the center of the cell in close association which time M. rubrum underwent two additional cell divisions. with the ciliate nuclei and always anterior to the two ciliate Subsequently, the proportion of cells with a CPN gradually macronuclei (Figure 2C). This arrangement is subsequently declined to ∼13.3% on Day 29. referred to as the cryptophyte-ciliate nuclear complex (CCN Centered prey nucleus diameter was relatively stable over the complex; Figure 2D). We also observed smaller cryptophyte first 11 days of the experiment (Figure 3C), showing a minimum nuclei (diameter 2.53 µm ± 0.04 µm), typically located at the value of 4.01 ± 0.15 µm (n = 3) for samples taken on Day 4 periphery of the ciliate and usually in the anterior part of the and a maximum of 4.73 ± 0.14 µm on Day 11 (n = 3). During cell (Figure 2C). These nuclei were similar in size to the nucleus starvation (Day 11–29), however, the size of the CPN gradually

Frontiers in Microbiology| www.frontiersin.org 4 March 2017| Volume 8| Article 423 fmicb-08-00423 March 18, 2017 Time: 15:45 # 5

Kim et al. Dynamics of Acquired Prey Nucleus in Mesodinium rubrum

FIGURE 2 | Epiﬂuorescence micrographs of the cryptophyte Teleaulax amphioxeia (A) and the ciliate Mesodinium rubrum (B,C) stained with both Hoechst 3325 and CellMask Green in combination. (A) T. amphioxeia showing the brightly stained cryptophyte nucleus. (B) Starved M. rubrum containing two ciliate macronuclei and one ciliate micronucleus. (C) Well-fed M. rubrum containing two types of prey nuclei (EPN and CPN) as well as the three nuclei. (D) Schematic drawing of the well-fed M. rubrum. Macro, ciliate macronucleus; Micro, ciliate micronucleus; EPN, extra prey nucleus; CPN, centered prey nucleus. Scale bar in (A) is 10 µm and applies to (B,C).

TABLE 1 | Diameter of Teleaulax amphioxeia nuclei, Mesodinium rubrum nuclei, and ingested prey nuclei (EPNs and CPNs) during Experiments 1 and 2.

Species Type of nucleus Sample ID Nuclear Range Total # Total # diameter1 ± SE (µm) nuclei cells (µm) examined observed

Teleaulax amphioxeia Nucleus Experiment 1, Day 0 2.08 ± 0.04 1.8 – 2.5 27 27 Mesodinium rubrum Ciliate micronucleus Experiment 1, Day 0 2.63 ± 0.09 2.3 – 3.8 21 21 Ciliate macronuclei Experiment 1, Day 0 3.54 ± 0.07 2.9 – 4.5 42 21 EPN Well-fed cells Experiment 1, Day 0 – Day 8 2.35 ± 0.04 1.3 – 3.9 181 240 Starved/refeed cells Experiment 2, Day 0.5 – Day 13 2.17 ± 0.01 1.2 – 4.6 1959 1260 CPN Well-fed cells Experiment 1, Day 0 – Day 11 4.35 ± 0.04 2.5 – 7.7 255 300 Starved/refeed cells Experiment 2, Day 0 6.83 + 0.26 5.8 – 8.0 8 60 Experiment 2, Day 0.5 – Day 4 3.80 + 0.07 1.8 – 8.4 293 540 Experiment 2, Day 19 5.75 + 0.12 3.9 – 7.6 42 180

Mean values are for nuclear diameters pooled for speciﬁed sample times. 1Nuclear diameter estimated as the mean of length and width.

Frontiers in Microbiology| www.frontiersin.org 5 March 2017| Volume 8| Article 423 fmicb-08-00423 March 18, 2017 Time: 15:45 # 6

Kim et al. Dynamics of Acquired Prey Nucleus in Mesodinium rubrum

increased to 6.8 ± 0.1 µm (n = 3) for samples taken on Day 26 (Figure 3C), in direct contrast to the change in prevalence of cells with a CPN. The largest CPN observed during the experiment was encountered in samples taken at Day 26 and measured 8.9 µm in diameter. Extra prey nucleus were only observed in M. rubrum cells when prey cells were present (Day 0–8; Figure 3D). The percentage of ciliates with EPN was ∼75% at Day 2 and 4, and on average a little more than 1 EPN was found per cell (1.24 ± 0.05, n = 2, the number relative to all observed cells). Both values dropped rapidly after Day 4. Inorganic Carbon Uptake Mesodinium rubrum maintained an inorganic carbon uptake of 45.2 ± 2.5 pg C cell−1 h−1 (n = 6) until Day 11 (Figure 3C). Subsequently, photosynthetic activity steadily decreased in conjunction with prey depletion, reaching a value of 16 ± 1.4 pg C cell−1 h−1 at Day 29. Over the course of the experiment, inorganic carbon uptake displayed a direct relationship to the prevalence of cells with a CPN (r2 = 0.955, p < 0.0001; Figure 4). No relationship was observed between inorganic carbon uptake and the number of chloroplasts per cell (r2 = 0.0, p = 1; Figure 4). Experiment 2: Refeeding of Starved Mesodinium rubrum Cell Divisions Predator:prey ratio calculated from mean concentration of the predator during the first 48 h of incubation and initial prey density, was 1:2.5, 1:5, and 1:10 for the different treatments. On Day 10, prey concentrations were less than 1% of initial concentration in all treatments. On Day 13, the prey were depleted below detection limit (<1 cell ml−1) in the 1:2.5 and 1:5 treatments, and on Day 16 in the 1:10 treatment (Figures 5A–C). M. rubrum abundance increased rapidly 4 days after refeeding with prey, showing a doubling in cell concentration every second day until Day 10. All cultures were diluted with fresh f/2 medium to a new initial concentration of 500 cells ml−1 on Day 10, resulting in two additional cell divisions every third day, after which growth stopped. Changes in Prevalence of Ingested Prey Nuclei in M. rubrum at Different Prey Concentrations Immediately after the addition of prey, EPNs were observed in M. rubrum cells in all treatments. Likewise, when prey FIGURE 3 | Experiment 1. Starvation of well-fed Mesodinium rubrum. was depleted, EPNs were no longer observed (Figures 5D–F). (A) Abundance of M. rubrum and T. amphioxeia as a function of incubation time. (B) Chloroplast number and volume (cell–1) as function of incubation Changes in the prevalence and number of EPNs were directly time for M. rubrum. (C) Percentage of M. rubrum cells with a centered prey related to prey concentration. The highest percentage of cells nucleus (CPN), CPN diameter, and photosynthetic rate for M. rubrum as a having EPNs was observed on Day 4 or Day 6 in the three function of incubation time. (D) Percentage of M. rubrum cells with one or treatments. At low prey concentration (predator to prey ratio more extra prey nuclei (EPNs) and number of EPNs cell–1 over incubation of 1:2.5), 66.6% of M. rubrum cells had a mean of 1.12 ± 0.38 time. Data for cell abundance and photosynthesis represent mean ± SE for = −1 triplicate flasks (n = 3). Data for other parameters represent mean ± SE for (n 3) EPNs cell on Day 6 (Figure 5D). At moderate prey triplicate flasks (n = 3), except for Day 0, when cells were examined from a concentration (predator to prey ratio of 1:5), about 88.3% of single sample taken prior to distribution of stock culture to experimental − M. rubrum cells had a mean of 2.18 ± 0.19 (n = 3) EPNs cell 1 flasks. For Day 0, n = 18 for chloroplasts number cell–1 and chloroplasts on Day 4 (Figure 5E), while at the high initial prey concentration volume cell–1, n = 21 for CPN prevalence, EPN prevalence, CPN number –1 –1 (predator to prey ratio of 1:10), 100% of M. rubrum cells retained cell , and EPN number cell , and n = 19 for CPN diameter. Dashed vertical ± = −1 lines in (B–D) denote the point at which prey were depleted. Arrowheads a mean of 5 0.16 (n 3) EPNs cell on Day 6 (Figure 5F). The indicate a doubling in cell number relative to prior values. largest number of EPNs retained by a single M. rubrum cell (11)

Frontiers in Microbiology| www.frontiersin.org 6 March 2017| Volume 8| Article 423 fmicb-08-00423 March 18, 2017 Time: 15:45 # 7

Kim et al. Dynamics of Acquired Prey Nucleus in Mesodinium rubrum

FIGURE 4 | Prevalence of a centered prey nucleus (CPN) and number of chloroplasts cell−1 for Mesodinium rubrum in Experiment 1 plotted as a function of photosynthetic rate.

FIGURE 5 | Experiment 2. Refeeding of starved Mesodinium rubrum (A–C). Abundance of M. rubrum and T. amphioxeia in low, medium, and high prey treatments, respectively, plotted as a function of incubation time. (D–F) Percentage of M. rubrum cells with ingested prey nuclei in low, medium, and high prey treatments, respectively, plotted as a function of incubation time. Plots show mean ± SE for triplicate ﬂasks (n = 3), with 20 cells examined sample-1. EPN, extra prey nucleus; CPN, centered prey nucleus; arrows indicate dilution of experimental cultures; arrowheads indicate a doubling in cell number relative to prior values; stars indicate that prey was present at less than 1% of predator abundance; dotted horizontal lines represent half of the former ratio value.

was observed on Day 1 in the high prey treatment. The ‘n’ refers At the start of the experiment, less than 10% of M. rubrum to triplicate samples and EPNs were scored in at least 20 cells in cells had a CPN, but the prevalence of cells with a CPN each sample. increased after refeeding, exceeding 80% on Day 4 in all

Frontiers in Microbiology| www.frontiersin.org 7 March 2017| Volume 8| Article 423 fmicb-08-00423 March 18, 2017 Time: 15:45 # 8

Kim et al. Dynamics of Acquired Prey Nucleus in Mesodinium rubrum

FIGURE 6 | Prevalence of a centered prey nucleus as a function of the number of cell divisions for low, medium, and high prey treatments of Experiment 2.

treatments (Figures 5D–F). Changes in the occurrence of Changes in Size of Sequestered Prey Nuclei in cells with a CPN, however, showed different patterns across M. rubrum at Different Prey Concentrations the treatments of prey concentration. At the lowest prey Extra prey nucleus observed in all three treatments were density, the prevalence of a CPN declined after Day 6, as relatively constant in size (Figure 8A), with a mean diameter cells underwent division (Figure 5D), while at moderate prey of 2.17 ± 0.01 µm for measurements pooled across treatment density, prevalence of a CPN remained above 90% from and time (Table 1). By contrast, CPNs were initially large in Day 4 to 10 as cells divided 4 times and then decreased size (Figure 8B), having a mean diameter of 6.83 ± 0.26 µm (Figure 5E). At the high prey concentration, prevalence of on Day 0 (Table 1). Treatment means showed decreased CPN a CPN was maintained above 80% while cells divided five size from Day 0.5 to Day 4 as prey were ingested and new times (Day 4 to Day 13) and then decreased (Figure 5F). CPNs formed, with CPN diameter for data pooled across Changes in the prevalence of CPNs relative to the number treatments for those days averaging 3.08 ± 0.07 µm. Like in of cell divisions were dependent on initial prey density. The Experiment 1, the size of CPNs increased over time after prey percent cells with CPNs declined after three cell divisions in the were depleted (Figures 8B, 9), with data pooled across treatments low prey treatment after 4 cell divisions in the medium prey for Day 19 averaging 5.75 ± 0.12 µm. CPN size at the end of treatment, and after five cell divisions in the high prey treatment the experiment was similar across treatment, but pronounced (Figure 6). enlargement occurred sooner (Day 6–8) in the lowest prey Extra prey nucleus abundance (i.e., number ml−1) increased treatment compared to in the medium and high prey treatments to a peak on Day 6 in the low prey treatment and on Day 8 (Day 10–13). in the medium and high prey treatments, before declining to Frequency distributions for CPN diameter showed distinct undetectable levels as prey was depleted (Figure 7). Abundance size classes of ‘old’ and ‘new’ CPNs on Day 0 to Day 4. M. rubrum of total ingested prey nuclei (EPNs + CPNs) increased to a peak on Day 0 were derived from pooled culture remaining on Day on Day 10 in all three treatments, but remained relatively stable 29 of Experiment 1 and thus contained ‘old’ CPNs that fell following dilution of the cultures and depletion of prey. After into the 5.5–6.4 µm to 7.5–8.4 µm size classes (Supplementary dilution of the high prey treatment on Day 10, EPN abundance Figure S1). On Day 3.5, smaller, presumably ‘new’ CPNs were was about half that of total ingested nuclei (Figure 7C). Over the observed, 94% of which fell into the 1.5–2.4 µm and 2.5–3.4 µm following 6 days, EPN abundance in that treatment dropped to size classes and had a mean diameter of 2.58 ± 0.06 µm (n = 29). zero without influencing abundance of total ingested nuclei. Given the mean diameter of EPNs on Day 0.5 (2.06 µm, Table 1),

Frontiers in Microbiology| www.frontiersin.org 8 March 2017| Volume 8| Article 423 fmicb-08-00423 March 18, 2017 Time: 15:45 # 9

Kim et al. Dynamics of Acquired Prey Nucleus in Mesodinium rubrum

FIGURE 8 | Mean nuclear diameter ± SE for triplicate ﬂasks (n = 3) during Experiment 2. (A) Extra prey nuclei (EPNs) for low, medium, and high prey treatments. (B) Centered prey nuclei (CPNs) for low, medium, and high prey treatments. Dotted line in (A) denotes the nuclear diameter for the prey, Teleaulax amphioxeia, (mean 2.08 µm).

2.5–3.4 µm size class on Day 1 and in the 3.5–4.4 µm size class on Day 4. A total of 1959 ENPs were examined for the experiment, none of which appeared to be in the process of fusing with another EPN or a CPN. While dividing M. rubrum were common in FIGURE 7 | Abundances of ingested prey nuclei (nuclei ml–1) for (A) Experiment 2, none of the 335 CPNs examined for Day 0.5 to low, (B) medium, and (C) high prey treatments of Experiment 2 as a function Day 19 appeared to be undergoing division. of incubation time. Mean ± SE for nuclear abundance calculated as mean abundance. M. rubrum abundance at each sample time multiplied by number of ingested prey nuclei cell–1 for the triplicate ﬂasks (n = 3). EPNs, extra prey DISCUSSION nuclei; CPNs, center prey nuclei; arrows indicate dilution of cultures; stars indicate relatively constant values for (EPN + CPN) ml–1 following dilution of cultures. Presence of Extra Prey Nuclei (ENPs) and the Centered Prey Nucleus The ciliate Mesodinium rubrum has for long been known to harbor prey cytoplasm and cell organelles, including chloroplasts, the volume of ‘new’ CPNs (8.9 µm3) was about twice that of mitochondria, and nuclei through feeding on cryptophytes EPNs (4.5 µm3) present in M. rubrum at the same time. Over (Taylor et al., 1969, 1971; Hibberd, 1977; Oakley and Taylor, the following 3.5 days, the frequency distribution for ‘new’ CPNs 1978). Initially, it was believed that this represented a reduced shifted toward larger size classes, with the peak occurring in the permanent “symbiont,” but recent molecular studies have

Frontiers in Microbiology| www.frontiersin.org 9 March 2017| Volume 8| Article 423 fmicb-08-00423 March 18, 2017 Time: 15:45 # 10

Kim et al. Dynamics of Acquired Prey Nucleus in Mesodinium rubrum

FIGURE 9 | Schematic diagram showing the ideal dynamics of acquired prey nuclei following cell division in the ciliate Mesodinium rubrum and epifluorescence micrographs of M. rubrum cells from the high prey treatment of Experiment 2 after staining with Hoechst 33258 and CellMask Green. (1) M. rubrum feeds on three cryptophyte Teleaulax amphioxeia cells as prey, and obtains the nuclei of each prey (EPN). One of the EPN moves toward the center of the ciliate (or anterior to the three ciliate nuclei) and becomes enlarged (CPN). The other two EPN transfer to daughter cells through cell division, respectively. After first cell division, one daughter cell (2.1) has a CPN derived from one of the EPN. This daughter cell is divided into new daughter cells having a CPN (3.1) or not (3.2) by a second cell division. Meanwhile, the other daughter cell (2.2) has a CPN with one EPN. In case of this daughter cell, it is divided into new daughter cells through a second cell division; two daughter cells (3.3 and 3.4) have a CPN, respectively. Each CPN moves in the same way as before following a third cell division. As a result, only three cells among the eight daughter cells have a CPN after the third cell division. The series of epifluorescence images show temporal differences in the arrangement, number, and size of ingested prey nuclei after feeding starved M. rubrum with T. amphioxeia. (A) Day 0: 18 days starved M. rubrum derived from Day 29 of Experiment 1. (B) Day 1, (C) Day 4, (D) Day 8, (E) Day 13, and (F) Day 19 after feeding. Macro, macronucleus; Micro, micronucleus; EPN, extra prey nucleus; CPN, centered prey nucleus. Scale bar in (A) is 10 µm and applies to all images.

indicated that this is not true. There is now clear evidence that individual M. rubrum cells and provide little information about the cryptophyte organelles and cytoplasm are identical to the how M. rubrum retains diﬀerent sizes of prey nuclei. Kim et al. prey that the ciliate is fed (e.g., Johnson et al., 2007) and that the (2016), however, suggested that small prey nuclei fuse to form chloroplasts from one cryptophyte species can be exchanged with a large prey nucleus. Similarly, Nam et al.(2016) reported that chloroplasts from another species (Hansen et al., 2012). Previous small nuclei sequestered from prey were enclosed in a single publications have used transmission electron microscopy and membrane and suggested that small nuclei moved to the center of ﬂuorescence microscopy to document the presence of small the M. rubrum cell and fused to form a single large prey nucleus. (1.9–4 µm in diameter) or large (4.5–10 µm in diameter) prey Our study documents that along with two ciliate macronuclei nuclei (Hibberd, 1977; Oakley and Taylor, 1978; Johnson et al., and one ciliate micronucleus, well-fed M. rubrum cells can 2007; Kim et al., 2016; Nam et al., 2016). Those reports are simultaneously have multiple small, peripherally positioned prey limited to observations of only one type of prey nucleus inside nuclei that we call extra prey nuclei (EPNs), and a single large,

Frontiers in Microbiology| www.frontiersin.org 10 March 2017| Volume 8| Article 423 fmicb-08-00423 March 18, 2017 Time: 15:45 # 11

Kim et al. Dynamics of Acquired Prey Nucleus in Mesodinium rubrum

centrally positioned prey nucleus (CPN). EPNs were slightly undetectable levels. Enlargement of CPNs when prey cells were larger (mean diameter 2.53 µm) than the mean size of the not available to be ingested and when EPNs were not present prey, Teleaulax amphioxeia, nucleus (mean diameter 2.08 µm). in M. rubrum cells, indicates that CPNs can enlarge without The CPN was always located at the center of the cell, as also fusion with EPNs. Third, once prey were reduced to very low shown for the large prey nucleus reported in previous studies or undetectable levels, the number of total ingested prey nuclei (Hibberd, 1977; Hansen and Fenchel, 2006; Kim et al., 2016; (EPNs + CPNs) ml−1 remained stable as the number of EPNs Nam et al., 2016), and was generally much larger than the ml−1 decreased. Were the disappearance of EPNs to result from EPNs. The CPN, however, showed considerable variation in size fusion with CPNs, then the total number of ingested prey nuclei during our experiments, depending on prey concentration and ml−1 would be expected to decrease. Fourth, if enlargement incubation time. The CPN size was relatively stable in well-fed of CPNs were to result from fusion with EPNs, then it would M. rubrum supplied with plentiful prey (mean diameter 4.35 µm) be reasonable to expect the size of the CPNs and the rate of but increased dramatically during starvation, averaging 6.8 µm in CPN enlargement to depend on the number of EPNs that are diameter 18 days after depletion of prey. retained. In our starvation/refeeding study, however, the size CPNs was similar in all treatments even though the maximum Formation and Enlargement of the CPN mean number of EPNs cell−1 showed a fivefold difference. In Eighteen days after depletion of prey, ∼85% of the M. rubrum addition, enlargement of CPNs progressed faster in our lowest in our study lacked a CPN and none possessed EPNs, but prey treatment where maximum mean number of EPNs cell−1 both types of nuclei were reacquired by a majority of the cells was one, than in our moderate and high prey treatment where upon refeeding. EPNs appeared in most cells within 12 h and maximum mean number of EPNs cell−1 was two and five, increased in number over time, with maximum mean number respectively (Figure 8B). cell−1 ranging from one to five depending on the amount of prey Our results support the alternative hypothesis that CPNs are provided. Up to 11 EPNs were observed in individual M. rubrum. formed by the relocation of a single EPN from the periphery EPNs occupied a peripheral position in the cell, whereas recently to the center of the M. rubrum cell where it becomes part of sequestered ‘new’ CPNs were located at the center of the cell had the ciliate-cryptophyte nuclear complex and increases in size a mean diameter (2.58 µm), similar to that of recently acquired without fusing with EPNs. Under that scenario, EPNs remaining EPNs (2.06 µm) present in cells at the same time. Four days at the periphery of cells after formation of a ‘new’ CPN might be post feeding, 80 – 90%, depending on prey concentration, of digested, or might be distributed to daughter cells during division M. rubrum cells had a CPN, with mean diameter of 3.80 µm, of M. rubrum where they would be available to form a ‘new’ CPN. larger than an EPN, but smaller than the CPN of well-fed cells The latter possibility assumes that CPNs do not divide along with (mean 4.35 µm). Calculations based on mean diameter indicate the ciliate, as our data suggest (see below). If EPNs are distributed that the volume of ‘new’ CPNs is about twice that of recently to daughter cells to form ‘new’ CPN, then the mean number of acquired EPNs, raising the possibility that ‘new’ CPNs arise from EPNs cell−1 would influence the number of cell divisions that fusion of two EPNs. However, fusing of multiple EPNs to form could occur without a decline in the frequency of M. rubrum a ‘new’ CPN seems unlikely, since none of the more than 2100 with a CPN. That seems to be the case in our starvation/refeeding EPNs observed in our experiments were closely clustered or experiment, as M. rubrum in the low, medium, and high prey appeared to be fusing with another EPN. It seems more probable treatments had a maximum mean number EPNs cell−1 of 1, 2, that ‘new’ CPNs form by the relocation of a single EPN to the and 5 divided 3, 4, and 5 times, respectively, before showing a center of the cells accompanied by and/or followed by an increase drop in CPN frequency. in size that does not result from fusion with another EPN. One may wonder how the CPN of M. rubrum increases in size Lack of Division and Disappearance of the CPN over time. As mentioned in the introduction, enlargement of the Our results imply that the CPN of M. rubrum does not divide. centrally positioned prey nucleus was reported by Johnson et al. Not only did we fail to observe indications of division in any of (2007), Kim et al.(2016), and Nam et al.(2016) and suggested the more than 600 CPNs examined, CPNs of dividing M. rubrum in the latter two papers to result from the fusion of smaller were indistinguishable from the CPNs of non-dividing cells. Also, ingested prey nuclei. Data from our starvation/refeeding study CPN prevalence decreased with division of host cells as prey were (Experiment 2), however, do not support that hypothesis, for depleted, suggesting dilution of CPN cell−1 due to lack of CPN several reasons. First, of the more than 2100 EPNs and over 600 division. As mentioned above, the total number of ingested prey CPNs examined during our experiments, fusions of an EPN with nuclei (EPNs + CPNs) ml−1 as prey were depleted remained a CPN was never observed. If enlargement of the ‘new’ CPN relatively constant, due to apparent transformation of EPNs were to be a slow process, fusion events might occur infrequently into CPNs. Were CPNs to have divided during that time, total and thus not be observed in our samples. Enlargement of ‘new’ ingested prey nuclei ml−1 should have increased. Hence, during CPNs in our starvation/refeeding study, however, appeared to cell division of M. rubrum, the CPN appears to be inherited be a rather rapid process, as indicated by the observed upward by only one of the two daughter cells. The lack of the ability shift in size classes of ‘new’ CPNs from Day 0.5 to Day 1 and of M. rubrum to divide the CPN has previously been reported Day 4. Second, during both of our experiments, CPNs showed by Johnson et al.(2007). During prey starvation they found a a dramatic increase after prey had been depleted and continued disappearance of the CPN (termed kleptokaryon in their study) to enlarge even after the number of EPNs cell−1 had dropped to in M. rubrum over time and could estimate a half time of its

Frontiers in Microbiology| www.frontiersin.org 11 March 2017| Volume 8| Article 423 fmicb-08-00423 March 18, 2017 Time: 15:45 # 12

Kim et al. Dynamics of Acquired Prey Nucleus in Mesodinium rubrum

disappearance. Hansen and Fenchel(2006) and Kim et al.(2016), ciliate M. rubrum (Johnson and Stoecker, 2005; Johnson et al., asserted that the CPN in M. rubrum is able to divide at least once 2007; Kim et al., 2014, 2016; Onuma and Horiguchi, 2015). in prey starved cultures, since cells that had undergone one cell A few molecular and transcriptome studies focusing on the division all had a CPN. Based on our observations, however, their photosynthetic ability of M. rubrum in association with the results could be explained by the retention of EPNs prior to cell sequestered prey nucleus (Johnson et al., 2007; Lasek-Nesselquist division and distribution of EPN to daughter cells to generate et al., 2015; Kim et al., 2016) have shown expression of nuclear- CPNs, rather than by division of the CPNs. encoded plastid-targeted algal genes. Subsequently, the retained Through the present study, enlargement of the CPN was prey nucleus was suggested to mainly contribute to sustained inferred to result from an increase in size of only one EPN. chloroplasts function (Johnson et al., 2007; Hansen et al., However, one question still remains to be answered: what causes 2012; Lasek-Nesselquist et al., 2015; Kim et al., 2016). It has enlargement of the CPN. Prior studies (Hibberd, 1977; Kim et al., previously been shown that photosynthetic activity in different 2016; Nam et al., 2016) have provided ultrastructural images M. rubrum strains declines in prey starved cultures (Johnson of the enlarged prey nucleus (i.e., CPN) within M. rubrum. and Stoecker, 2005; Hansen and Fenchel, 2006; Johnson et al., Here the chromosomes seemed to be untangled or less 2007). It has also been shown that the declines in photosynthetic dense, the nucleolus had a large size and large amounts of parameters coincided with the loss of CPN (called prey nuclei) nucleoplasm surrounded the expanded nuclear envelope. We from M. rubrum cells, implicating a possible functional role also observed similar morphological changes of the CPN under for retained prey nuclei (Johnson and Stoecker, 2005; Johnson the confocal microscope, with less dense chromosomes and et al., 2007). Our work confirms that the presence of a CPN an enlarged nucleolus present in large irregular shaped CPNs substantially affects the photosynthetic performance of the (Supplementary Figure S2). The chromosome just seemed to M. rubrum chloroplasts. A reduction in prevalence of the CPNs be loose or swollen, but the possibility of replication cannot in starved populations of M. rubrum led to a significant decline in be ruled out. Surprisingly, two enlarged nucleoli were found in inorganic carbon uptake while chloroplast number cell−1 showed one CPN, a phenomenon that also appeared in the study of little change. This result might help to explain the formation Johnson et al.(2007). Nevertheless our quantitative data do not of the ‘CCN’ complex; the position could facilitate the gene indicate that M. rubrum can replicate the CPN and the results exchange related with nuclear-encoded, chloroplasts targeted seem inconsistent with a recent study by Qiu et al.(2016). They genes for stable photosynthesis, between the host nuclei and observed gene transcripts for prey nucleus replication on field prey nucleus (Martin et al., 1998; Martin and Herrmann, 1998). populations of M. rubrum. Since their study was carried out on Even though our results imply that a CPN is involved with field populations, it is difficult to interpret their data. It is possible photosynthetic ability of M. rubrum, it cannot be ruled out that that M. rubrum population they studied function differently than the ciliate nuclei also participate in the photosynthetic ability of those isolates from Korea, Denmark, and Antarctica that have M. rubrum. We do not know whether retention of a CPN and been studied in detail in laboratory culture. After all six clades chloroplasts derived from prey in M. rubrum is an evolutionary (Clades A–F) of the M. rubrum/M. major species complex have step toward establishing permanent chloroplasts, but M. rubrum been described recently (Herfort et al., 2011; Johnson et al., 2016). is notable for showing the unique photosynthetic performance However, it is also possible that the M. rubrum population studied from the acquired chloroplasts and nucleus of prey. by Qiu et al.(2016) functions similar to the ones studied in Hansen et al.(2012) proved that M. rubrum chloroplasts detail in the laboratory, but the authors may have simply caught derived from T. amphioxeia can be replaced by chloroplasts M. rubrum cells that recently ingested a diving cryptophyte cell. derived from T. acuta. Whether or not the sequestered Future studies will show which of the two interpretations are prey nucleus was simultaneously replaced, however, remains right. unknown. Addressing the possibility of replacement of the prey nucleus in M. rubrum may enhance our understanding of Prey Nucleus Effects on Photosynthetic Ability of sequestration, enlargement, and function of the prey nucleus M. rubrum inside M. rubrum. Mesodinium rubrum is unique among the marine alveolates for its ability to sequester prey nuclei and chloroplasts along with other organelles and show enlargement of the prey nucleus once Model for CPN Maintenance and Increase in Size sequestered. While sequestration of nuclei and chloroplasts along Based on the results of our study, we propose the following with other prey organelles is well known for a few dinoflagellates model to explain the dynamics of acquisition, enlargement, and species (Dodge, 1971; Farmer and Roberts, 1990; Horiguchi and distribution of prey nuclei to daughter cells in Mesodinium Pienaar, 1992; Okamoto and Inouye, 2005; Gast et al., 2007; rubrum. When M. rubrum having only three ciliate nuclei (one Yamaguchi et al., 2011; Onuma and Horiguchi, 2013, 2015; Kim micronucleus and two macronuclei positioned at the center of et al., 2014), enlargement of the sequestered prey nucleus has not the cell) feeds on the cryptophyte Teleaulax amphioxeia, prey been reported in any of studies. nuclei (i.e., EPNs) are acquired at the periphery of the cell. The The sequestered prey nucleus has been inferred to allow the number of acquired nuclei is equal to the number of prey ingested host cell to exploit photosynthetic performance of its sequestered (Figure 9). Subsequently, one of the EPNs relocates to the center prey chloroplasts; e.g., the dinoflagellates Amylax triacantha, of the cell (or anterior to the three ciliate nuclei) to become a Nusuttodinium aeruginosum, and N. myriopyrenoides and the CPN, thus forming a CCN complex. The newly formed CPN

Frontiers in Microbiology| www.frontiersin.org 12 March 2017| Volume 8| Article 423 fmicb-08-00423 March 18, 2017 Time: 15:45 # 13

Kim et al. Dynamics of Acquired Prey Nucleus in Mesodinium rubrum

is small at first, but continuously increases in size over time, AUTHOR CONTRIBUTIONS without fusing with EPNs that persist in the cell. With division of M. rubrum, the enlarged CPN does not divide and is inherited by All authors were involved in the design and planning of the only one of the daughter cells, with the other daughter cell having experiments. MK carried out the experiments, collected data, and the possibility of receiving one or more of the persisting EPNs. performed data analysis with help from the co-authors. MK wrote If the daughter cell lacking a CPN receives one or more EPNs, the first draft of the manuscript based on discussions with all then a single EPN relocates to the center of the cell to form a the co-authors. All co-authors were involved in revision of the CCN complex and a CPN which enlarges over time. The inherited manuscript and all the co-authors approved final manuscript. CPN continues to enlarge until reaching maximum size as seen in starved M. rubrum. Once the ‘old’ inherited CPN senesces, it can be replaced by a new CPN if one or more EPNs are present in FUNDING the peripheral cytoplasm of M. rubrum or are acquired through feeding. This work was supported by the Danish Council for Independent Research, grant number 4181-00484 (PH). ND thanks the Villum Kann Rasmussen Foundation for an equipment grant. CONCLUSION

Our study show that the sequestered prey nucleus (CPN) ACKNOWLEDGMENT associated with the CCN complex (Cryptophyte-Ciliate Nuclear complex) of M. rubrum is derived from a single prey nucleus, We are indebted to D. Wayne Coats for comments and discussion enlarges over time without fusing with other ingested prey nuclei on an earlier manuscript draft. (EPNs), does not divide, and is inherited by only one daughter cell when M. rubrum divides. Also, EPNs present in M. rubrum possessing a CPN can be inherited by and form a CPN in the SUPPLEMENTARY MATERIAL daughter cell that does not receive the parental CPN. How the EPN-CPN system of M. rubrum works when the ciliate is fed The Supplementary Material for this article can be found a mixture of prey species is unknown. Will M. rubrum then online at: http://journal.frontiersin.org/article/10.3389/fmicb. contain two CPNs or can one CPN control the chloroplasts from 2017.00423/full#supplementary-material two species? To elucidate this, development of species specific FIGURE S1 | Frequency distributions for diameter of centered prey nuclei molecular techniques is required. Our study was conducted using (CPNs) on (A) Day 0, (B) Day 0.5, (C) Day 1, and (D) Day 4 of Experiment 2. “Old M. rubrum isolated from Danish waters, while reports of fusion of CPNs” on Day 0 are enlarged nuclei remaining in M. rubrum cells after 18 days of ingested prey nuclei to form the enlarged prey nucleus (i.e., CPN) starvation (Day 29 of Experiment 1). “New CPNs” represent nuclei acquired after of the CCN complex (Kim et al., 2016; Nam et al., 2016) was based addition of prey at the beginning of Experiment 2. In (B–D), CPNs in the three on a Korean strain (MR-MAL01) of Mesodinium cf. rubrum. largest size classes were assumed to be “old” CPNs. Since our Danish M. rubrum and the Korean M. cf. rubrum differ FIGURE S2 | Confocal images showing differences in size and shape of at the strain level and may even represent different species, it ingested prey nucleus in M. rubrum (A) 2 days, (B–D) 29 days after feeding on is possible that the two cultures process ingested prey nuclei in prey. (A) Ingested prey nucleus located at the periphery of the ciliate (EPN) similar to the nucleus of the prey species. (B) Enlarged prey nucleus located at the center different way. Additional studies using the Korean isolate of M. of the cell (CPN) seems to have untangled or less dense chromosomes and cf. rubrum and other isolates of M. rubrum from other parts of enlarged nucleolus. (C) The CPN shows morphological change. (D) The CPN has the world may help to resolve this issue. two enlarged nucleoli. Scale bar in (A) is 5 µm and applies to all images.

REFERENCES Hansen, P. J. (2002). Effect of high pH on the growth and survival of marine phytoplankton: implications for species succession. Aquat. Microb. Ecol. 28, Dodge, J. D. (1971). A dinoflagellate with both a mesocaryotic and a eucaryotic 279–288. doi: 10.3354/ame028279 nucleus I. Fine structure of the nuclei. Protoplasma 73, 145–157. doi: 10.1007/ Hansen, P. J., and Fenchel, T. (2006). The bloom-forming ciliate Mesodinium BF01275591 rubrum harbours a single permanent endosymbiont. Mar. Biol. Res. 2, 169–177. Farmer, M. A., and Roberts, K. R. (1990). Organelle loss in the endosymbiont doi: 10.1080/17451000600719577 of Gymnodinium acidotum (Dinophyceae). Protoplasma 153, 178–185. Hansen, P. J., Moldrup, M., Tarangkoon, W., Garcia-Cuetos, L., and Moestrup, doi: 10.1007/BF01354002 Ø. (2012). Direct evidence for symbiont sequestration in the marine red tide Gast, R. J., Moran, D. M., Dennett, M. R., and Caron, D. A. (2007). Kleptoplasty ciliate Mesodinium rubrum. Aquat. Microb. Ecol. 66, 63–75. doi: 10.3354/ame in an Antarctic dinoflagellate: caught in evolutionary transition? Environ. 01559 Microbiol. 9, 39–45. doi: 10.1111/j.1462-2920.2006.01109.x Herfort, L., Peterson, T. D., McCue, L. A., Crump, B. C., Prahl, F. G., Guillard, R. R. (1975). “Culture of phytoplankton for feeding marine invertebrates,” Baptista, A. M., et al. (2011). Myrionecta rubra population genetic diversity in Culture of Marine Invertebrate Animals, eds W. L. Smith and M. H. Chanley and its cryptophyte chloroplast specificity in recurrent red tides in the (New York, NY: Springer), 29–60. doi: 10.1007/978-1-4615-8714-9_3 Columbia River estuary. Aquat. Microb. Ecol. 62, 85–97. doi: 10.3354/ame Gustafson, D. E., Stoecker, D. K., Johnson, M. D., Van Heukelem, W. F., and 01460 Sneider, K. (2000). Cryptophyte algae are robbed of their organelles by the Hibberd, D. J. (1977). Observations on the ultrastructure of the cryptomonad marine ciliate Mesodinium rubrum. Nature 405, 1049–1052. doi: 10.1038/ endosymbiont of the red-water ciliate Mesodinium rubrum. J. Mar. Biol. Assoc. 35016570 U.K. 57, 45–61. doi: 10.1017/S0025315400021226

Frontiers in Microbiology| www.frontiersin.org 13 March 2017| Volume 8| Article 423 fmicb-08-00423 March 18, 2017 Time: 15:45 # 14

Kim et al. Dynamics of Acquired Prey Nucleus in Mesodinium rubrum

Horiguchi, T., and Pienaar, R. N. (1992). Amphidinium latum Lebour Okamoto, N., and Inouye, I. (2005). A secondary symbiosis in progress? Science (Dinophyceae), a sand-dwelling dinoflagellate feeding on cryptomonads. Jpn. 310, 287–287. J. Phycol. 40, 353–363. Onuma, R., and Horiguchi, T. (2013). Morphological transition in Johnson, M. D. (2011). Acquired phototrophy in ciliates: a review of cellular kleptochloroplasts after ingestion in the dinoflagellates Amphidinium interactions and structural adaptations. J. Eukaryot. Microbiol. 58, 185–195. poecilochroum and Gymnodinium aeruginosum (Dinophyceae). Protist doi: 10.1111/j.1550-7408.2011.00545.x 164, 622–642. doi: 10.1016/j.protis.2013.06.003 Johnson, M. D., Beaudoin, D. J., Laza-Martinez, A., Dyhrman, S. T., Fensin, Onuma, R., and Horiguchi, T. (2015). Kleptochloroplast enlargement, karyoklepty E., Lin, S., et al. (2016). The genetic diversity of Mesodinium and associated and the distribution of the cryptomonad nucleus in Nusuttodinium Cryptophytes. Front. Microbiol. 7:2017. doi: 10.3389/fmicb.2016.02017 (= Gymnodinium) aeruginosum (Dinophyceae). Protist 166, 177–195. Johnson, M. D., Oldach, D., Delwiche, C. F., and Stoecker, D. K. (2007). Retention doi: 10.1016/j.protis.2015.01.004 of transcriptionally active cryptophyte nuclei by the ciliate Myrionecta rubra. Park, J. S., Myung, G., Kim, H. S., Cho, B. C., and Yih, W. (2007). Growth responses Nature 445, 426–428. doi: 10.1038/nature05496 of the marine photosynthetic ciliate Myrionecta rubra to different cryptomonad Johnson, M. D., and Stoecker, D. K. (2005). Role of feeding in growth and strains. Aquat. Microb. Ecol. 48, 83–90. doi: 10.3354/ame048083 photophysiology of Myrionecta rubra. Aquat. Microb. Ecol. 39, 303–312. Qiu, D., Huang, L., and Lin, S. (2016). Cryptophyte farming by symbiotic doi: 10.3354/ame039303 ciliate host detected in situ. Proc. Natl. Acad. Sci. U.S.A. 113, 12208–12213. Johnson, M. D., Tengs, T., Oldach, D., and Stoecker, D. K. (2006). Sequestration, doi: 10.1073/pnas.1612483113 performance, and functional control of cryptophyte plastids in the ciliate Raho, N., Jaén, D., Mamán, L., Rial, P., and Marín, I. (2014). psbA based molecular Myrionecta rubra (Ciliophora). J. Phycol. 42, 1235–1246. doi: 10.1111/j.1529- analysis of cross-feeding experiments suggests that Dinophysis acuta does not 8817.2006.00275.x harbour permanent plastids. Harmful Algae 35, 20–28. doi: 10.1016/j.hal.2014. Kim, G. H., Han, J. H., Kim, B., Han, J. W., Nam, S. W., Shin, W., et al. 03.003 (2016). Cryptophyte gene regulation in the kleptoplastidic, karyokleptic ciliate Skovgaard, A., Hansen, P. J., and Stoecker, D. K. (2000). Physiology of the Mesodinium rubrum. Harmful Algae 52, 23–33. doi: 10.1016/j.hal.2015.12.004 mixotrophic dinoflagellate Fragilidium subglobosum. I. Effects of phagotrophy Kim, M., Kim, K. Y., Nam, S. W., Shin, W., Yih, W., and Park, M. G. (2014). and irradiance on photosynthesis and carbon content. Mar. Ecol. Prog. Ser. 201, The effect of starvation on plastid number and photosynthetic performance in 129–136. doi: 10.3354/meps201129 the kleptoplastidic dinoflagellate Amylax triacantha. J. Eukaryot. Microbiol. 61, Smith, M., and Hansen, P. J. (2007). Interaction between Mesodinium rubrum and 354–363. doi: 10.1111/jeu.12115 its prey: importance of prey concentration, irradiance and pH. Mar. Ecol. Prog. Lasek-Nesselquist, E., Wisecaver, J. H., Hackett, J. D., and Johnson, M. D. (2015). Ser. 338, 61–70. doi: 10.3354/meps338061 Insights into transcriptional changes that accompany organelle sequestration Taylor, F. J. R., Blackbourn, D. J., and Blackbourn, J. (1969). Ultrastructure of the from the stolen nucleus of Mesodinium rubrum. BMC Genomics 16:805. chloroplasts and associated structures within the marine ciliate Mesodinium doi: 10.1186/s12864-015-2052-9 rubrum (Lohmann). Nature 224, 819–821. doi: 10.1038/224819a0 Lindholm, T. (1985). Mesodinium rubrum-a unique photosynthetic ciliate. Adv. Taylor, F. J. R., Blackbourn, D. J., and Blackbourn, J. (1971). The red-water ciliate Aquat. Microbiol. 3, 1–48. Mesodinium rubrum and its “incomplete symbionts”: a review including new Martin, W., and Herrmann, R. G. (1998). Gene transfer from organelles to ultrastructural observations. J. Fish. Res. Board Can. 28, 391–407. doi: 10.1139/ the nucleus: How much, what happens, and why? Plant Physiol. 118, 9–17. f71-052 doi: 10.1104/pp.118.1.9 Yamaguchi, H., Nakayama, T., Kai, A., and Inouye, I. (2011). Taxonomy Martin, W., Stoebe, B., Goremykin, V., Hansmann, S., Hasegawa, M., and Kowallik, and phylogeny of a new kleptoplastidal dinoflagellate, Gymnodinium K. V. (1998). Gene transfer to the nucleus and the evolution of chloroplasts. myriopyrenoides sp. nov. (Gymnodiniales, Dinophyceae), and its cryptophyte Nature 393, 162–165. doi: 10.1038/30234 symbiont. Protist 162, 650–667. doi: 10.1016/j.protis.2011.01.002 Moeller, H. V., Johnson, M. D., and Falkowski, P. G. (2011). Photoacclimation in Yih, W., Kim, H. S., Jeong, H. J., Myung, G., and Kim, Y. G. (2004). Ingestion the phototrophic marine ciliate Mesodinium rubrum (Ciliophora). J. Phycol. 47, of cryptophyte cells by the marine photosynthetic ciliate Mesodinium rubrum. 324–332. doi: 10.1111/j.1529-8817.2010.00954.x Aquat. Microb. Ecol. 36, 165–170. doi: 10.3354/ame036165 Myung, G., Kim, H. S., Park, J. S., Park, M. G., and Yih, W. (2011). Population growth and plastid type of Myrionecta rubra depend on the kinds of available Conflict of Interest Statement: The authors declare that the research was cryptomonad prey. Harmful Algae 10, 536–541. doi: 10.1016/j.hal.2011. conducted in the absence of any commercial or financial relationships that could 04.005 be construed as a potential conflict of interest. Myung, G., Kim, H. S., Park, J. W., Park, J. S., and Yih, W. (2013). Sequestered plastids in Mesodinium rubrum are functionally active up to 80 days of The reviewer MJ declared a past co-authorship with several of the authors phototrophic growth without cryptomonad prey. Harmful Algae 27, 82–87. (MK, KD, PJH) to the handling Editor, who ensured that the process met the doi: 10.1016/j.hal.2013.05.001 standards of a fair and objective review. Nam, S. W., Park, J. W., Yih, W., Park, M. G., and Shin, W. (2016). The fate of cryptophyte cell organelles in the ciliate Mesodinium cf. rubrum subjected to Copyright © 2017 Kim, Drumm, Daugbjerg and Hansen. This is an open-access starvation. Harmful Algae 59, 19–30. doi: 10.1016/j.hal.2016.09.002 article distributed under the terms of the Creative Commons Attribution License Oakley, B. R., and Taylor, F. J. R. (1978). Evidence for a new type of (CC BY). The use, distribution or reproduction in other forums is permitted, provided endosymbiotic organization in a population of the ciliate Mesodinium rubrum the original author(s) or licensor are credited and that the original publication in this from British Columbia. BioSystems 10, 361–369. doi: 10.1016/0303-2647(78) journal is cited, in accordance with accepted academic practice. No use, distribution 90019-9 or reproduction is permitted which does not comply with these terms.

Frontiers in Microbiology| www.frontiersin.org 14 March 2017| Volume 8| Article 423

Supporting Information

FIGURE S1 | Frequency distributions for diameter of centered prey nuclei (CPNs) on (A) Day 0, (B) Day 0.5, (C) Day 1, and (D) Day 4 of Experiment 2. “Old CPNs” on Day 0 are enlarged nuclei remaining in M. rubrum cells after 18 days of starvation (Day 29 of Experiment 1). “New CPNs” represent nuclei acquired after addition of prey at the beginning of Experiment 2. In (B–D), CPNs in the three largest size classes were assumed to be “old” CPNs.

FIGURE S2 | Confocal images showing differences in size and shape of ingested prey nucleus in M. rubrum (A) 2 days, (B–D) 29 days after feeding on prey. (A) Ingested prey nucleus located at the periphery of the ciliate (EPN) similar to the nucleus of the prey species. (B) Enlarged prey nucleus located at the center of the cell (CPN) seems to have untangled or less dense chromosomes and enlarged nucleolus. (C) The CPN shows morphological change. (D) The CPN has two enlarged nucleoli. Scale bar in (A) is 5 μm and applies to all images.

LETTER LETTER Mesodinium rubrum:Thesymbiosisthatwasn’t

Matthew D. Johnsona,1, Erica Lasek-Nesselquistb, Holly V. Moellerc, Andreas Altenburgerd, Nina Lundholmd, Miran Kime, Kirstine Drumme, Øjvind Moestrupe, and Per Juel Hansene

Qiu et al. (1) report that a red tide of the photosynthetic genes in an Antarctic M. rubrum culture (clade A) (8). In ciliate Mesodinium rubrum in Long Island Sound “farms” addition, because genomes or transcriptomes of the symbiotic Teleaulax amphioxeia cells within its cyto- target organisms were not used for annotation (1) there plasm. M. rubrum has long been studied for causing is a high degree of uncertainty in assigning transcript red tides (2–5), and laboratory culture work on multiple identity. strains from around the world has shown that M. rubrum Second, Qiu et al. (1) report intact cryptophyte cells extracts organelles from ingested cryptophyte algae, in- inside M. rubrum. However, their TEM images are incon- cluding chloroplasts, mitochondria, cytoplasm, and a clusive due to (i) low resolution (ii), extraordinarily poor transcriptionally active nucleus, or kleptokaryon (6, 7). fixation quality, and (iii) unusually small cryptophyte or- M. rubrum functions like a true phototroph, with the ganelles, compounding the interpretation of the low-res- ability to regulate and divide chloroplasts (7). olution images. No clear cell membrane, which would The conclusions of Qiu et al. (1), based on a single include cytoplasm completely surrounding the chloro- field sample, contrast sharply with these previously plast, is visible around the cryptophyte organelles in their published studies of M. rubrum. Their conclusions images. Rather, they seem to be organelle complexes are based on (i) their inference that “complete” prey that are packed into a membrane, consistent with pre- metatranscriptomes indicate metabolically intact prey vious observations (9). Furthermore, in other M. rubrum cells and (ii) their visual observation, using transmis- recently ingested intact cryptophytes seem to be in a sion electron microscopy (TEM), of intact prey cells. vacuole before organelle extraction (Fig. 1) (10). However, we believe that these findings do not pro- M. rubrum-like ciliates are complex organisms that vide sufficient evidence to support the extraordinary do not fit into established “boxes” for trophic modes claim by Qiu et al. (1) that M. rubrum farms prey cells. or cellular organization. However, we have previously First, the authors argue that the expression of shown that their unique mode of acquired phototro- genes involved in membrane transporters, nucleus- phy is capable of “farming” cryptophyte organelles to-cytoplasm RNA transporters, and all major meta- when the kleptokaryon is present (7), and there is no bolic pathways is evidence of intact cryptophyte evidence for maintenance of intact symbionts within symbionts. Here we show data from a temperate strain any cultures of the ciliate. We firmly believe that the of M. rubrum (clade G) that indicate that many crypto- conclusions of Qiu et al. (1) do not represent a new phyte gene pathways are expressed at levels equal to association in M. rubrum but rather illustrate the actual or greater than in T. amphioxeia even when only prey difficulties of accurately interpreting “snapshots” of organelles remain (Table 1 and Fig. 1). One exception natural populations. is expression levels of ABC-like transporters, which were observed to be at even lower numbers in Qui Acknowledgments et al. (1). Furthermore, we have previously shown similar M.D.J. and E.L.-N. were supported by National Science Founda- transcriptional patterns of highly expressed cryptophyte tion Integrative and Organismal Systems Award 1354773.

aBiology Department, Woods Hole Oceanographic Institution, Woods Hole, MA 02543; bWadsworth Center, New York State Department of Health, Albany, NY 12208; cDepartment of Ecology, Evolution & Marine Biology, University of California, Santa Barbara, CA 93106; dNatural History Museum of Denmark, University of Copenhagen, 1350 Copenhagen, Denmark; and eMarine Biological Section, Department of Biology, University of Copenhagen, 2200 Copenhagen, Denmark Author contributions: M.D.J. designed research; M.D.J. and E.L.-N. performed research; M.D.J. and E.L.-N. analyzed data; and M.D.J., E.L.-N., H.V.M., A.A., N.L., M.K., K.D., Ø.M., and P.J.H. wrote the paper. The authors declare no conflict of interest. 1To whom correspondence should be addressed. Email: [email protected].

E1040–E1042 | PNAS | February 14, 2017 | vol. 114 | no. 7 www.pnas.org/cgi/doi/10.1073/pnas.1619247114 Downloaded at LAEGEVIDENSKABELIGE BIBLIO on January 25, 2020 Fig. 1. Transmission electron micrograph of Mesodinium rubrum (CBJR05; clade G) fed T. amphioxeia (GCEP01). The image shows a lateral cross- section of an M. rubrum cell revealing at least nine plastid complexes and a recently ingested T. amphioxeia cell (also a cross-section) within a vacuole (white arrow) in the center. Note the ingested cell’s periplast membrane and cytoplasm surrounding the chloroplast, and the vacuolar space surrounding it. In this image the cytoplasm of the cryptophyte organelle complexes is lighter than that of the ciliate, revealing that large portions of M. rubrum cells are devoted to hosting stolen organelles. Ciliate cytoplasm and many of the organelles in the image of Qiu et al. (1) are missing or unrecognizable, respectively.

Johnson et al. PNAS | February 14, 2017 | vol. 114 | no. 7 | E1041 Downloaded at LAEGEVIDENSKABELIGE BIBLIO on January 25, 2020 Table 1. Comparison of key metabolic pathways (reads per kilobase of transcript per million mapped reads) in the T. amphioxeia-derived kleptokaryon of M. rubrum (KN) and free-living T. amphioxeia (TA) KEGG gene orthology groups* Metabolic pathway KN TA KN/TA

Metabolism ko01100 Metabolic pathways 223,716 169,969 1.32 ko00195 Photosynthesis 64,880 53,849 1.20 ko00196 Photosynthesis-antenna proteins 34,946 4,731 7.39 ko00860 Porphyrin and chlorophyll metabolism 10,007 13,685 0.73 Transport and protein processing ko04141 Protein processing in endoplasmic reticulum 59,957 62,823 0.95 ko04130 SNARE interactions in vesicular transport 1,670 1,451 1.15 ko02010 ABC transporters 699 3,686 0.19 DNA and RNA pathways ko03010 Ribosome 288,272 714,136 0.40 ko03040 Spliceosome 29,099 27,447 1.06 ko03013 RNA transport 29,037 18,314 1.59 ko03030 DNA replication 10,621 1,583 6.71 ko03015 mRNA surveillance pathway 10,253 6,909 1.48 ko03020 RNA polymerase 10,124 7,932 1.28 Cell cycle and cytoskeleton ko04110 Cell cycle 17,241 5,403 3.19 ko04810 Regulation of actin cytoskeleton 5,460 4,578 1.19 Signaling ko04010 MAPK signaling pathway 19,891 18,091 1.10 ko04020 Calcium signaling pathway 17,312 8,949 1.93

*Kyoto Encyclopedia of Genes and Genomes.

1 Qiu D, Huang L, Lin S (2016) Cryptophyte farming by symbiotic ciliate host detected in situ. Proc Natl Acad Sci USA 113(43):12208–12213. 2 Crawford DW (1989) Mesodinium rubrum: The phytoplankter that wasn’t. Mar Ecol Prog Ser 58(1-2):161–174. 3 Hansen PJ, Nielsen LT, Johnson M, Berge T, Flynn KJ (2013) Acquired phototrophy in Mesodinium and Dinophysis–A review of cellular organization, prey selectivity, nutrient uptake and bioenergetics. Harmful Algae 28:126–139. 4 Herfort L, Peterson TD, Campbell V, Futrell S, Zuber P (2011) Myrionecta rubra (Mesodinium rubrum) bloom initiation in the Columbia River estuary. Estuar Coast Shelf Sci 95(4):440–446. 5 Taylor FJR, Blackbourn DJ, Blackbourn J (1969) Ultrastructure of the chloroplasts and associated structures within the marine ciliate Mesodinium rubrum (Lohmann). Nature 224:819–821. 6 Gustafson DE, Jr, Stoecker DK, Johnson MD, Van Heukelem WF, Sneider K (2000) Cryptophyte algae are robbed of their organelles by the marine ciliate Mesodinium rubrum. Nature 405(6790):1049–1052. 7 Johnson MD, Oldach D, Delwiche CF, Stoecker DK (2007) Retention of transcriptionally active cryptophyte nuclei by the ciliate Myrionecta rubra. Nature 445(7126):426–428. 8 Lasek-Nesselquist E, Wisecaver JH, Hackett JD, Johnson MD (2015) Insights into transcriptional changes that accompany organelle sequestration from the stolen nucleus of Mesodinium rubrum. BMC Genomics 16:805. 9 Garcia-Cuetos L, Moestrup Ø, Hansen PJ (2012) Studies on the genus Mesodinium II. Ultrastructural and molecular investigations of five marine species help clarifying the taxonomy. J Eukaryot Microbiol 59(4):374–400. 10 Nam SW, Park JW, Yih W, Park MG, Shin W (2016) The fate of cryptophyte cell organelles in the ciliate Mesodinium cf. rubrum subjected to starvation. Harmful Algae 59:19–30.

E1042 | www.pnas.org/cgi/doi/10.1073/pnas.1619247114 Johnson et al. Downloaded at LAEGEVIDENSKABELIGE BIBLIO on January 25, 2020

Effects of nitrogen on two non-constitutive mixotrophs, Mesodinium rubrum and Dinophysis acuminata I: Effects of ammonium concentrations on growth, inorganic carbon uptake and removal of different inorganic nitrogen forms as a function of time after prey deprivation

Kirstine Drumm1, Patricia M. Glibert2, Kevin J. Flynn3 and Per J. Hansen4

1University of Copenhagen, Department of Biology, Helsingør, Denmark

2 University of Maryland, Center for Environmental Science, Cambridge, Maryland, US

3Swansea University, Department of Biosciences, Wales, UK

Abstract Current knowledge on the ability of non-constitutive mixotrophs (NCM) to take up inorganic nutrients is very limited and published results are contradictory with respect to relative importance of ammonium and nitrate. In this study, we investigated the ability of two NCMs, Mesodinium rubrum and Dinophysis acuminata, to + - + - take up NH4 and NO3 at different NH4 and NO3 concentrations. While M. rubrum sequesters the chloroplasts and other cell organelles from its cryptophyte prey, Teleaulax amphioxeia, D. acuminata sequesters only the chloroplasts, and it does that by feeding on M. rubrum. It cannot sequester chloroplasts from the cryptophyte directly. The ability of the two NCMs to take up inorganic nitrogen was compared to that of T. amphioxeia, the chloroplast donor for both species. In all cases, growth and inorganic carbon uptake were monitored + - simultaneously. M. rubrum showed a preference for uptake of NH4 over NO3 , with a repressive effect of + - NH4 on NO3 uptake, similar to its prey, T. amphioxeia. M. rubrum was able to obtain significantly higher cell + concentrations and inorganic carbon uptake rates with increasing addition of NH4 however, the rates were + affected by N-depletion, not by the addition of NH4 . Similar results were found for T. amphioxeia. M. rubrum + was able to divide 8 times without prey, at the highest NH4 addition, while only 4 cell divisions were observed + + when no NH4 was added to the growth medium, indicating that cell division was promoted by NH4 uptake. Despite this difference in cell divisions, the percentage of M. rubrum cells with a centered prey nucleus (CPN) -1 + and number of CPNs cell were not significantly different among the NH4 treatments, indicating that the M. + rubrum cells that had received more NH4 were able to divide the CPNs. The outcome of N-depletion (culture conditions that also resulted in higher pH) resulted in a significantly lower inorganic carbon uptake rate for - both T. amphioxeia and M. rubrum. Uptake of NO3 by Dinophysis acuminata was primarily measurable in the + - control treatments, to which no NH4 was added. In all other treatments, the ambient NO3 concentration + + remained unchanged during the experiments with added NH4 , suggesting repression. High rates of NH4 removal (> 20 pM N cell-1 d-1) by D. acuminata were found, which resulted in both significant higher growth - + and inorganic carbon uptake rates. In conclusion, both NO3 and NH4 can support the growth of M. rubrum, + whereas only NH4 supports the growth of D. acuminata.

Keywords: Dinophysis, Mesodinium, nitrate, ammonium, uptake rates, growth rates, inorganic C-uptake

Introduction Non constitutive mixoplankton (NCM; Flynn et al. 2019) obtain energy and nutrition by a combination of autotrophy and phago-heterotrophy. However, unlike constitutive mixoplankton, which can synthesize and have full control over their plastids, NCM protists do not have chloroplasts of their own. Instead, they sequester chloroplasts by ingesting phototrophic prey organism and retaining chloroplasts together with other organelles, depending on species (Mitra, Flynn et al. 2016). NCMs are further divided into two types, depending on whether the chloroplasts are derived from specific prey organisms (the specialized SNCMs), or whether the chloroplasts are derived from non-specific prey organisms (the generalists GNCMs). The dinoflagellate, Dinophysis acuminata and the ciliate, Mesodinium rubrum are both SNCMs acquiring their chloroplasts ultimately from the same donor, Teleaulax amphioxeia (Hansen, Nielsen et al. 2013).

Mesodinium rubrum retains both chloroplasts and the nucleus directly from T. amphioxeia (Gustafson, Stoecker et al. 2000). The chloroplasts become swollen after ingestion, leading to increased chlorophyll a and primary production (Smith and Hansen 2007). Furthermore, one of the ingested prey nuclei is transported to the cryptophyte-ciliate nucleus complex (CCN-complex) and become a centered prey nucleus (CPN) (Kim, Drumm et al. 2017). The CPN has been shown to play a key role in inorganic carbon uptake in M. rubrum. M. rubrum can divide and uphold the number of chloroplasts, including when food deprived (Kim, Drumm et al. 2017). In contrast to Mesodinium, D. acuminata is unable to take up chloroplasts directly from T. amphioxeia, but by using its peduncle, it sequesters chloroplasts from M. rubrum (other organelles are ingested but quickly digested) (Park, Kim et al. 2006). After ingestion, the chloroplast goes through a transition orchestrated by D. acuminata. The chloroplast are organized into kleptochloroplast centers, with terminal pyrenoids and the thylakoids are rearranged from triplicates to pairs (Garcia-Cuetos, Moestrup et al. 2010, Kim, Nam et al. 2012, Rusterholz, Hansen et al. 2017). When D. acuminata undergo cell division, the number of chloroplasts decreases, but recent studies have shown that D. acuminata is also able to divide its ingested chloroplasts (Rusterholz, Hansen et al. 2017). Even over long periods of food deprivation, D. acuminata can retain viable chloroplasts (Riisgaard and Hansen 2009, Nielsen, Krock et al. 2012). At least 2 separate studies have shown that plastidic encoding genes derive from at least five different algae sources (Wisecaver and Hackett 2010, Hongo, Yabuki et al. 2019).

The ciliate M. rubrum is known for causing non-toxic red tides, but it also functions as prey for the dinoflagellate Dinophysis spp. (Park, Kim et al. 2006). Dinophysis spp. is often the cause of diarrhetic shellfish poisoning (DSP) and blooms of this species have thus received considerable attention due to this problem. Despite their importance, the nutritional ecology of both M. rubrum and D. acuminata are poorly understood. Speculations on whether nutritional conditions are related to bloom of either species have motivated field studies (Delmas, Herbland et al. 1992, Johansson, Graneli et al. 1996, Seeyave, Probyn et al. 2009, Hattenrath- Lehmann, Marcoval et al. 2015). These studies have shown inconsistent results as to the extent to which

3 inorganic nitrogen supports growth of D. acuminata, and to some extent M. rubrum, during blooms. Comparatively few laboratory data exist on the use of inorganic nitrogen by M. rubrum and D. acuminata (Hattenrath‐Lehmann and Gobler 2015, Tong, Smith et al. 2015, García-Portela, Reguera et al. 2020), and they too have reported inconsistent results. Hattenrath-Lehmann & Gobler (2015) and García-Portela et al. (2020) + - found that D. acuminata takes up NH4 and to a lesser extent NO3 . Both papers also reported ~200% higher uptake rates of NH4+ in well-fed cells than by starved cells. Furthermore, Hattenrath-Lehmann & Gobler + (2015) observed a higher growth rate for D. acuminata cultures grown with addition of 25-µM NH4 relative + to cultures without added NH4 . In contrast, Tong et al. (2015) reported that D. acuminata was unable to - + - + remove neither ambient NO3 nor NH4 , while they reported that M. rubrum took up NO3 , but not NH4 . These studies applied different measures of nitrogen uptake and thus some variability between studies is to be expected.

Accordingly, knowledge of the nitrogen nutritional ecology of M. rubrum and D. acuminata remains unclear. The effect of different ambient concentrations of nutrients over time after prey deprivation has not, previously been investigated. In present study, we investigated the ability of Mesodinium rubrum and Dinophysis + - + - acuminata to take up NH4 and NO3 as a function after prey deprivation, at different NH4 and NO3 concentrations. The cell concentration and carbon assimilation rate were followed when these cultures were + - grown in monoculture conditions at four different initial NH4 concentrations and two NO3 concentrations. The chloroplasts donor, Teleaulax amphioxeia, was investigated under the same conditions, simultaneously. This study represents the first part of two, where paper no. 2 deals with the potential for uptake of ammonium, nitrate and urea as a function of time after prey deprivation in M. rubrum and D. acuminata.

Method and materials Protist strains and culture conditions The strains of Dinophysis acuminata (Da-DK2007), Mesodinium rubrum (MBL-DK2009) and Teleaulax amphioxeia (SCCAP K- 0434) used in the present study all originate from Danish waters. The cultures were grown in f/2 growth medium based on heat treated (95 °, 90 min) natural seawater (30‰), and received an irradiance of 100 µM photons m-2 s-1 (day:night cycle 16:8h). M. rubrum was fed T. amphioxeia in a predator:prey cell ratio of 1:2, twice a week. Dinophysis acuminata was fed M. rubrum once a week using a predator:prey ratio of 1:1. The cultures of M. rubrum were starved for prey a minimum of 2 weeks to avoid transferring any T. amphioxeia to D. acuminata culture.

Experiments 1 and 2. Effects of nitrate concentrations ([50] µM and [15] µM N, respectively) and four differentiated ammonium concentrations on growth, photosynthesis and removal. These experiment were designed to assess rates of growth and inorganic carbon uptake when cells were grown + + + at four different initial NH4 concentrations (T0[NH4 ]): [0], [10], [25] and [50] µM NH4 , respectively in the

- - - presence of NO3 (T0[NO3 ]) at [50] (experiment 1) and at [15] µM NO3 (experiment 2). For both experiments, 3- concentrations of phosphate were held at (T0[PO4 ]): [10] µM, thus yielding initially varying N:P conditions for the different treatments, but high enough to ensure that only N and not P stress is involved. The initial inoculum of T. amphioxeia was obtained by diluting an actively growing culture of T. amphioxeia in f/2 media - with heat-treated seawater to a concentrations equal to f/40 (~50 µM NO3 ). In the cases of M. rubrum and D. acuminata, actively growing cultures were diluted using inverse filtration (mesh size 10 and 15 µm, respectively) using heat-treated seawater. This was done to remove prey organisms and to optimize cell concentrations; neither D. acuminata nor M. rubrum can grow to high cell densities, due to elevated pH. Twelve litres of diluted culture of each organism were gently mixed and distributed into four, 5 L flasks (with 3- + 3 L culture in each flask). Stock solutions of PO4 and NH4 were then subsequently added to reach the selected treatment concentrations (Fig. 1). Each treatment was again distributed into three new flasks, functioning as triplicates.

- Figure 1. Schematic illustration of experimental design startup, of exp. 1 ([50] µM NO3 ), including the four + - different treatments of NH4 . The exact same setup was used for exp. 2 except for the initial ambient NO3 concentration was at [15] µM N.

Sampling time points were selected based on the cell division rates for each species (Table 1). At each sampling

+ - point, samples were withdrawn for cell concentration, inorganic carbon uptake and ambient NH4 , NO3 and 3- PO4 concentrations (see below for details). Inorganic carbon (IC) were measured at day 0 and approximately on every second sampling point. pH was monitored using a SenTix 41 pH electrode (WTW, Germany) connected to a pH meter (WTW, pH 3210, Germany) and calibrated with pH 7 and 10 standard buffers.

Table 1 Time of sampling for the three experiments.

Experiment 1 and 2 Experiment 3

Teleaulax Mesodinium Dinophysis Mesodinium amphioxeia rubrum acuminata rubrum

Cell concentration Day 0, 1, 2, 3, 4 Day 0, 2, 4, 7, 10 Day 0, 3, 5, 7 Day 4, 7, 9, 11, 14, and 7 and 14 and 10 17

Ambient nutrients Day 0, 1, 2, 3, 4 Day 0, 2, 4, 7, 10 Day 0, 3, 5, 7 Day 4, 7, 9, 11, 14, and 7 and 14 and 10 17

Inorganic carbon Day 0, 1, 2, 3, 4 Day 0, 2, 4, 7, 10 Day 0, 3, 5, 7 - uptake and 7 and 14 and 10

Nitrogen and Day 0 and 3 Day 0 and 7 Day 0 and 7 - carbon content

Enumeration of - Day 0, only for - Day 7, 9, 11, 14 nucleus (CPN) exp. 1 and 17

pH Day 0, 3, 4 and Day 0, 4, 7 and Day 0, 3, 5, 7 Day 4, 7, 11, 14 7 10 and 10 and 17

Experiment 3. Effects of initial ammonium concentration on cell division and retention of prey nuclei in M. rubrum. + Initial studies indicated an increase in cell divisions in Mesodinium rubrum with initial NH4 concentration. To investigate this further, 100 mL of M. rubrum were withdrawn on day 4 from experiment 1 from each experimental flask, and mixed with 300 mL seawater to which vitamins and micronutrients were added + corresponding to f/40, and NH4 additions made corresponding to the previous treatment. Hence, [50] µM + + + NH4 was added to treatments which had been grown in [50] µM NH4 and [25] µM NH4 was added to + treatments with [25] µM NH4 and so forth. Cell concentrations and nutrient contents were measured at T4, T7, T9, T11, T14 and T17. In addition, 2 mL culture were fixed in 4% glutaraldehyde for enumeration of nuclei (see below for details) at T0 (sample from experiment 1), T7, T9, T11 and T17 (Table 1).

Cell concentration Aliquots (2mL) withdrawn from each flask were fixed in Lugol’s solution (final concentration 1%). Enumeration of Teleaulax amphioxeia, Mesodinium rubrum and Dinophysis acuminata were done with 1 mL Sedgewick-Rafter chambers, using an inverted microscope (Olympus CK40).

Ambient nutrients and uptake rates For ambient nutrients, subsamples (25mL) withdrawn from each flask were filtered through precombusted GF/F (Whatman) filters, and stored at -20°C until subsequent analysis on a Seal Analytical Autoanalyzer, model AA3HR according to (Koroleff 1970) and (Solórzano and Sharp 1980). Nutrient depletion rates per cell θ (pM cell-1 d-1) were calculated using the formula:

( ) = ( )( ) 𝑁𝑁1 − 𝑁𝑁2 θ 𝐶𝐶 𝑡𝑡2 − 𝑡𝑡1 where N1 and N2 are the dissolved ambient contents of nutrient, at time 1 and time 2 (μM), respectively, and is the natural logarithm (ln) average of the cell concentrations, which was calculated using the formula𝐶𝐶 (Anderson, Kulis et al. 1990):

= 𝐶𝐶ln2 (− 𝐶𝐶)1 𝐶𝐶 𝐶𝐶2 Where C1 and C2 are the initial and final cell concentrations.𝐶𝐶1 Inorganic carbon uptake (14C) Subsamples (2 mL) were withdrawn from each experimental flasks and added to 23 mL glass scintillation 14 - -1 vials. 20 µl NaH CO3 stock solution (specific activity 100 µCi mL ) were added to each vial, resulting in a specific activity of ~1.0Ci mL-1. The vials were incubated for 3 h under the same light conditions as the experiment flasks. Parallel dark vials were treated similarly, except that they were wrapped in alufoil during incubation. After incubation the specific activity was determined by transferring 100 µl from each vial to new vials containing 200 µl phenylethylamine. The remaining sample was acidified with 2 mL 10% glacial acetic acid in methanol, and left overnight for evaporation on a 65°C heat plate. A 1.5 mL of distilled water were added the dried samples followed by 10 mL of Packard Insta-Gel scintillation cocktail and radioactive activity was determined using liquid scintillation counter (Tri-Carb 2910 TR, Perkin-Elmer). Rates of inorganic carbon uptake (µgC × mL-1 × h-1) were calculated from the equation:

× = 14 𝐷𝐷𝐷𝐷𝐷𝐷× ×𝐼𝐼𝐼𝐼 𝑖𝑖𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛 𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐 𝑢𝑢𝑢𝑢𝑢𝑢𝑢𝑢𝑢𝑢𝑢𝑢 Where DMP is disintegrations min-1mL-1, IC is the concentration of𝐶𝐶𝑎𝑎 inorganicℎ 𝐶𝐶𝑡𝑡 carbon (µgC × mL-1), 14Cₐ is the -1 -1 specific activity in disintegrations min mL , h is the incubation time in hours and Ct is the total number of cells in the vail. Inorganic carbon (IC) was measured using a Shimadzu Total Organic Carbon (TOC) analyzer.

- - Figure 2, Experiment 1 ([50] µM NO3 ). Time course of ambient NO3 (light grey) and NH4 (dark grey) for + + + treatment: control (A), [10] µM NH4 (B), [25] µM NH4 (C) and [50] µM NH4 (D) during incubation of Teleaulax amphioxeia. Solid lines are uptake rates (pM cell-1 d-1), calculated from ambient concentrations. Points are means while error bars represent standard error of triplicate measurements. Enumeration of the number of centered prey nuclei (CPNs) For enumeration of the percentage of cells with a centered prey nucleus (CPN) in Mesodinium rubrum, aliquots (5 mL) from each flask were fixed in 1% glutaraldehyde in seawater, and left in the refrigerator (at 4°C) in the dark overnight. Fixed samples were stained with a combination of 25 µg mL-1 Hoechst 33258 (Invitrogen, USA) and 0.25 µg mL-1 CellMask (Life technologies, Carlsbad, CA, USA) for 15 min in a dark chamber, then filtered onto a 0.2 µm black polycarbonate membrane filter (Frisenette, Denmark). CPNs were quantified using Olympus BX-50 epifluorescence microscope using filter sets U-FUW (Olympus CO., Japan). A minimum of fifty cells were examined for each sample.

Statistical analysis

- + Cell concentration, inorganic carbon uptake and removal of NO3 and NH4 were evaluated, over time with Two-Way ANOVA using GraphPad.

- - Figure 3, Experiment 2 ([15] µM NO3 ). Time course of ambient NO3 (light grey) and NH4 (dark grey) for + + + treatment: control (A), [10] µM NH4 (B), [25] µM NH4 (C) and [50] µM NH4 (D) during incubation of Teleaulax amphioxeia. Solid lines are uptake rates (pM cell-1 d-1), calculated from ambient concentrations. Points are means while error bars represent standard error of triplicate measurements.

Results Experiment 1 and 2. Effects of initial ammonium concentrations on removal rates of nitrate, ammonium and urea in the three species, as well as on rates of growth and photosynthesis Nitrogen removal and calculated uptake rates of ammonium and nitrate - + Teleaulax amphioxeia took up both NO3 and NH4 and both these nitrogen forms were depleted over a period + - of 3-7 days, depending on the initial availability (Fig. 2 and 3). NH4 was depleted first, while NO3 was left + almost unchanged until NH4 was depleted (fig. 2 and 3 A, B and C). Measured maximum uptake rates of NO3 - -1 -1 for the [50] µM NO3 experiment (exp. 1) were 0.60, 0.47, 0.65 and 0.88 pM cell d , for treatments 1) control, + + + 2) [10] µM NH4 , 3) [25] µM NH4 and 4) [50] µM NH4 , respectively. Measured maximum uptake rates of - -1 -1 NO3 for the [15] µM NO3 experiment (exp. 2) were lower: 0.15, 0.11, 0.05, 0.04 pM cell d , respectively. Within the experiments the uptake rates did not differ significantly, but between the two experiments the uptake

- - Figure 4, Experiment 1 ([50] µM NO3 ). Time course of ambient NO3 (light grey) and NH4 (dark grey) for + + + treatment: control (A), [10] µM NH4 (B), [25] µM NH4 (C) and [50] µM NH4 (D) during incubation of Mesodinium rubrum. Solid lines are uptake rates (pM cell-1 d-1), calculated from ambient concentrations. Points are means while error bars represent standard error of triplicate measurements.

- rates of NO3 were significantly different (p<0.05, one-way ANOVA on measured maximum rating). Measured + + maximum uptake rates of NH4 increased with NH4 concentration in both experiments; the rates were 0.007, -1 -1 + 0.30, 0.50 and 0.71 pM cell d , respectively for the control, [10], [25] and [50] µM NH4 treatments. Uptake + - -1 -1 rates of NH4 in the [15] µM NO3 experiment were 0.03, 0.50, 0.67 and 0.81 pM cell d , respectively for the four treatments, and the results from exp. 2 did not differ significantly exp. 1 (p<0.05, one-way ANOVA on measured maximum ratings).

+ As for T. amphioxeia, for M. rubrum grown without prey, NH4 was depleted from the medium within the first - 4-5 days, except for the control which did not contain NH4 (Figs. 4,5). Thereafter, NO3 was removed from - + the medium. Uptake rates of NO3 were generally lower than those of NH4 . The measured maximum uptake - - -1 -1 rates of NO3 for the [50] µM NO3 experiment were 5.58, 4.48, 3.66 and 2.41 pM cell d , for treatment + control, [10], [25] and [50] µM NH4 , respectively (p<0.05, one-way ANOVA on measured maximum rating), - - resulting in a decrease in NO3 uptake with increasing NH4 concentrations (Fig. 4). For the [15] µM NO3

- - Figure 5, Experiment 2 ([15] µM NO3 ). Time course of ambient NO3 (light grey) and NH4 (dark grey) for + + + treatment: control (A), [10] µM NH4 (B), [25] µM NH4 (C) and [50] µM NH4 (D) during incubation of Mesodinium rubrum. Solid lines are uptake rates (pM cell-1 d-1), calculated from ambient concentrations. Points are means while error bars represent standard error of triplicate measurements.

- experiment, the measured maximum uptake rate of NO3 were not different between treatments and highest -1 -1 - uptake was around 2 pmol cell d (fig. 5). The uptake rates of NO3 were significantly higher in the [50] µM - - + NO3 experiment (exp. 1) compared to the [15] µM NO3 experiment, except for treatment [50] µM NH4 from exp. 1 (p<0.05, one-way ANOVA on measured maximum rating). The measured maximum uptake rates of + + + NH4 increased with increasing NH4 concentration i.e. treatments 0, [10], 25 and [50] µM NH4 had rates of -1 -1 - 0.27, 5.6, 7.36 and 10.8 pmol cell d , respectively for the [50] µM NO3 experiment. Similar results were - -1 found for the [15] µM NO3 experiment with measured maximum rates of 0.15 , 5.02, 9.13 and 8.07 pM cell -1 + - d , respectively. The uptake of NH4 was, unlike NO3 , not different between experiments 1 and 2.

+ + Dinophysis acuminata removed NH4 in both experiments, but could not fully deplete NH4 in the [50] µM + + NH4 treatment (Fig. 6D and 7D). The uptake rates of NH4 of D. acuminata increased with elevated amounts + - of NH4 , similar to what was observed in T. amphioxeia and M. rubrum (Fig. 6 and 7). However, NO3 was

- - Figure 6, Experiment 1 ([50] µM NO3 ). Time course of ambient NO3 (light grey) and NH4 (dark grey) for + + + treatment: control (A), [10] µM NH4 (B), [25] µM NH4 (C) and [50] µM NH4 (D) during incubation of Dinophysis acuminata. Solid lines are uptake rates (pM cell-1 d-1), calculated from ambient concentrations. Points are means while error bars represent standard error of triplicate measurements.

seemingly left unused by D. acuminata, and NO3 uptake rates were barely measurable during the [50] µM - NO3 experiment in any of the treatments, a finding very different from those of T. amphioxeia and M. rubrum. - Only for the control treatment could an initial removal of NO3 could be detected (Fig. 6A). Measured - - -1 -1 maximum NO3 uptake rates in the [50] µM NO3 experiment were 3.91, 0.29, 0.24 and 0.28 pM cell d for + - the treatments control, [10], [25] and [50] µM NH4 , respectively. In the [15] µM NO3 experiment significantly - -1 -1 higher NO3 uptake rates were found, 12.50, 5.15, 1.68, 1.81 pM cell d , respectively (p<0.05, one-way ANOVA on measured maximum rating). Generally, these uptake rates were not high enough to have an actual - - effect on the ambient NO3 , except for the [15] µM NO3 experiment control (fig. 7A), (fig. 6 and 7). Measured + + - maximum uptake rates of NH4 increased with increasing NH4 concentration with rates for the [50] µM NO3 -1 -1 + experiment of 0.82, 14.6, 17.05 and 24.75 pM cell d for treatments control, [10], [25] and [50] µM NH4 , - respectively. Similar results were found in the [15] µM NO3 experiment with rates of 0.0, 9.79, 20.64 and

- - Figure 7, Experiment 2 ([15] µM NO3 ). Time course of ambient NO3 (light grey) and NH4 (dark grey) for + + + treatment: control (A), [10] µM NH4 (B), [25] µM NH4 (C) and [50] µM NH4 (D) during incubation of Dinophysis acuminata. Solid lines are uptake rates (pM cell-1 d-1), calculated from ambient concentrations. Points are means while error bars represent standard error of triplicate measurements.

-1 -1 + 16.58 pM cell d , respectively. The uptake of NH4 did not differ significantly between experiment 1 and 2 (p<0.05, one-way ANOVA on measured maximum ratings).

Effects of dissolved nutrients on cellular growth rate and inorganic carbon uptake

+ The cell concentration of Teleaulax amphioxeia was unaffected by the different NH4 treatments used in the - + - [50] µM NO3 experiment (Fig. 8 A and B). The effect of NH4 addition had a clear effect in the [15] µM NO3 + experiment (Fig. 8 B). A decrease in cell concentrations of T. amphioxeia observed in the [50] µM NH4 treatment at the end of the experiment coincided with elevated pH (~9.3) that developed in this treatment + (sup.1). Effects of added NH4 on inorganic carbon uptake could not be detected in T. amphioxeia (Fig. 9 A - and B). Initially a small increase in inorganic carbon uptake was observed in the [15] µM NO3 experiment, - but this was for all treatments and, as in the [50] µM NO3 experiment, a decrease in activity was generally seen.

Figure 8, Cell numbers from experiment 1 (A, C and D) and 2 (B, D and F). Time course of numbers of cells during incubation for T. amphioxeia (A and B), M. rubrum (C and D) and D. + + + acuminata (E and F). The different treatments (control, [10] µM NH4 , [25] µM NH4 and [50] µM NH4 ) are presented in each graph indicated by symbols shown in A. For D. acuminata an initial presence of M. rubrum (and T. amphioxeia for exp. 2) was counted and are presented as dotted lines. Points are means while error bars represent standard error of triplicate measurements.

Figure 9, Inorganic carbon uptake for experiment 1 (A, C and D) and 2 (B, D and F). Time course of inorganic carbon uptake (pgC cell-1 h-1) during incubation of Teleaulax amphioxeia (A and B), Mesodinium rubrum (C and D) and Dinophysis acuminata (E and F). The different treatments (control, [10] + + + µM NH4 , [25] µM NH4 and [50] µM NH4 ) are presented in each graph indicated by symbols shown in A. Points are means while error bars represent standard error of triplicate measurements.

+ The presence of NH4 had an effect on Mesodinium rubrum cell concentration in both experiments. In general, + + cell concentration (Fig. 8 C and D) increased with the addition of NH4 . No significant effects of NH4 on inorganic carbon uptake could be detected in M. rubrum (Fig. 9 C and D). Both experiments showed a slight increase in the beginning of the experiments and then, like T. amphioxeia, decreased over time.

+ A significant effect of NH4 was seen on the cell concentration of Dinophysis acuminata (Fig. 8 E and F). In - the [50] µM NO3 experiment, the control treatment did not divide at all (Fig 8 E). For the other three treatments - the cells divided approximately 1.5 times over the duration of the study. In the [15] µM NO3 experiment, the + pattern seemed more similar to M. rubrum, where an increase of NH4 resulted in increased cell numbers (Fig. - 8 F). The inorganic carbon uptake initially increased in the [50] µM NO3 experiment, but did not decrease to - the extend shown for M. rubrum and T. amphioxeia (Fig. 9 E). This was even more clear in the [15] µM NO3 experiment, where the inorganic carbon uptake basically was unchanged throughout the experiment (Fig. 9 F). - The inorganic carbon uptake in the treatment control of the [50] µM NO3 experiment was only significantly lower than the rates measured for all nutrient addition treatments, hence D. acuminata was the only species + where the inorganic carbon uptake rate was affected by the NH4 addition.

Experiment 3. Effects of initial ammonium concentration on cell division and retention of prey nuclei in M. rubrum. + Additions of NH4 to M. rubrum did not affect the number of cell divisions up to day 6 among the different treatments. At day 6, cells with CPN had decreased ~20 % in all treatments. After day 6, all treatments were significantly different from the control treatment (p < 0.0001), which did not divide more than 4 times. Cells + + in the [10] µM NH4 treatment divided twice more and those in both the 25 and [50] µM NH4 treatments divided 4 times more than the control. Despite of this difference in cell divisions, the percentage of cells with CPN were approximately the same across all treatments (p < 0.765). From day 6, the percentage of cells with CPN was reduced to ~10% at day 9, and ~5% at day 11. From day 11 and onwards the percentage of cells with a CPN stayed unchanged at 5%. Over the course of the experiment, nitrogen became depleted in all treatments + except for that receiving [50] µM NH4 (sup. 2). Nitrogen depletion was noted by day 6 for the control and + + [10] µM treatments, and by day 12 for the [25] µM NH4 treatment. In the [50] µM NH4 treatment, nitrogen concentration was reduced to <2 µM N by day 14.

Figure 10. Experiment 3. Time course of percentage of cells with centered prey nucleus (solid lines, left y- axes) and number of cell divisions (dotted lines, right y-axes), for each ammonium treatment (control, [10] + + + µM NH4 , [25] µM NH4 and [50] µM NH4 ). Points are means while error bars represent standard error of triplicate measurements.

Discussion Nutrient uptake by Teleaulax amphioxeia Teleaulax amphioxeia is a cryptophyte that typically is grown as a classic phototrophic protist in monocultures. Most of our knowledge on T. amphioxeia derives from experiments using this cryptophyte as prey, especially for Mesodinium spp (Yih, Kim et al. 2004, Hansen and Fenchel 2006, Smith and Hansen 2007, Moeller, Johnson et al. 2011). Data are lacking on rates of different nitrogen forms on the growth and inorganic carbon uptake rates of T. amphioxeia.

+ + In the present experiments, T. amphioxeia readily took up NH4 and displayed a repressive effect of NH4 on - + - NO3 uptake at all NH4 and NO3 additions (Fig. 2 and 3). However, growth rates were unaffected in the [50] - - µM NO3 experiment (Fig. 8), while significant differences in growth rates were observed in the [15] µM NO3 + experiment at the [50] µM NH4 treatment. Inorganic carbon uptake rates by T. amphioxeia, except for in - + single case in the [15] µM NO3 experiment, were similarly unaffected by NH4 .

Rates of cell divisions of T. amphioxeia were the same over the first three days for all treatments and experiments. Thereafter the ability of T. amphioxeia to divide depended on the availability of N in any form + - (NH4 or NO3 ). Despite species-specific preferences, this is unlike most other phototrophic algae (e.g. green algae, Raphiophyceae, Cryptophyceae), which not only show a preference for ammonium over nitrate, but also experience increased growth when ammonium is available (Raven, Wollenweber et al. 1992, Wood and Flynn 1995, Giordano 1997, Thoisen, Vu et al. 2018). Rates of inorganic carbon uptake were likewise unaffected by

17 the different treatments but decreased over time. The decrease in cell numbers and to some extent the inorganic + - carbon uptake, for treatment [50] µM NH4 , the [15] µM NO3 experiment, is most likely a result of pH, not N-depletion.

Nutrient uptake by Mesodinium rubrum - + In the present experiments, M. rubrum was able to take up both NO3 and NH4 . Uptake of inorganic nitrogen forms by M. rubrum has previously been studied in both field studies and laboratory cultures (Wilkerson and Grunseich 1990, Herfort, Peterson et al. 2012, Tong, Smith et al. 2015). Wilkerson and Grunseich (1990) were - + the first to demonstrate uptake of both NO3 and NH4 in field samples. Our findings are in general agreement - + with Wilkerson and Grunseich (1990) with high uptake rates of both NO3 and NH4 (Table 2). Repression of - + + NO3 uptake by NH4 was clear in all treatments where NH4 was added. The uptake rates were calculated from - + + removal of NO3 and NH4 , and the data indicated that NH4 was taken up at a rate which was twice of the - + NO3 (Table 2). The higher uptake rates of NH4 could be explained as by the fact that, at least in experiment + - 2, concentrations of NH4 were generally higher than those of NO3 .

+ - Table 2. Reported values of measured maximum uptake rates of NH4 and NO3 in Mesodinium rubrum and Dinophysis acuminata. Data from the literature. Mesodinium rubrum Dinophysis acuminata - + - + Method NO3 NH4 NO3 NH4 (pM N cell-1 d-1) (pM N cell-1 d-1) (pM N cell-1 d-1) (pM N cell-1 d-1) Wilkerson and 15N 1.18 * 1.33 ** - - Grunseich (1990)a Seeyave et al. 15N - - 7.52 * 69.6 ** (2009) Tong et al. Removal 1.38 0.0 0.0 0.0 (2015) Hattenrath‐ 15N - - 0.08 * 12.48 ** Lehmann and Gobler (2015) García-Portela, 15N - - 1.6 * 22.8 ** et al. (2020) Present study Removal 5.58 10.79 12.50 24.75 a Assumed average cell concentration of 2000 cells mL-1 *Original uptake rate x 16 (h of light). **Original uptake rate x 24 (h of light).

Tong et al. (2015), in a study of M. rubrum under similar conditions (irradiance of 65 µM photons m-2 s-1 and - 15˚C), initiated their experiment with a concentration of ~200 µM NO3 , but despite this high initial concentration, the highest uptakes rate they reported were ~1.3 pM cell-1 d-1. In present experiments the highest

- -1 -1 - uptake rates of NO3 were ~5.6 pM cell d . The uptake rate of NO3 in present study increased significantly - with increasing addition of NO3 (Fig. 4 and 5), so it would be expected that the uptake rates reported by Tong et al (2015) were higher than ours. So, why Tong et al. (2015) did not find higher uptake rates could be related the M. rubrum strain used. It is now known that at least 6 different subclades of M. rubrum/M. major species complex exist (Johnson, Beaudoin et al. 2016). The specific strain of M. rubrum used by Tong et al. (2015) was isolated from Japanese waters. Since we worked with a Danish strain, strain differences between the - Danish and the Japanese strain might explain the different uptake rates of NO3 measured. This would also + explain why Tong et al. (2015) did not manage to measure any uptake of NH4 by their strain of M. rubrum. + Another reason for the lack of utilization of NH4 in that study, could however also be due to their quite low + + NH4 additions (~1 µM NH4 ).

+ Cell concentration of M. rubrum significantly increased with increasing additions of NH4 in both experiments - in the present study ([50] and [15] µM NO3 experiment). Cell divisions stopped shortly after N-depletion in + all experiments, however the total initial [N] concentration varied depending on the amount of NH4 added. Thus, the significant increase in cell number seemed to be a direct effect of nitrogen-depletion, rather than an + effect of addition of NH4 , (Fig. 8 C and D). Cell divisions in M. rubrum after prey deprivation have been investigated previously and most studies suggest approximately four cell divisions can occur without prey (Hansen and Fenchel 2006, Smith and Hansen 2007, Kim, Drumm et al. 2017). Approximately four cell - - divisions were found for M. rubrum in the [50] µM NO3 experiment, whereas in the [15] µM NO3 experiment + the cells treated with no or low amount of NH4 divided only 3 times.

+ To study the role of NH4 in cell divisions in M. rubrum in further details we carried out an additional - experiment (exp. 3). A subsample of M. rubrum, from each treatment from the [50] µM NO3 experiment, was + suspended after four days into new triplicates and with additions of NH4 concentration corresponding to the previous treatment. In contrast, we found that in the third experiment, M. rubrum divided eight times with + treatment [25] and [50] µM NH4 . This is an unusual high amount of cell division by M. rubrum, compared to previous results showing up to four cell divisions under prey deprivation (Johnson and Stoecker 2005, Hansen and Fenchel 2006, Kim, Drumm et al. 2017). Speculations on a connection between the percentage of cells with a centered prey nucleus (CPN) and growth, led us to measure the amounts of percentage of cells with a CPN in exp. 3. Despite eight cell divisions for the treatments 25 and 50, all treatments obtained the same amount of CPNs cell-1 as the control that only divided 4 times. While phosphate was never less than 2 µM P for any of the treatments, the N:P availability did vary across treatments. This indicate that M. rubrum that had + received more NH4 is able to divide the CPN. However, the numbers of extra prey nuclei (EPNs) were not enumerated during this experiment. If those have developed differently in the different treatments we cannot rule out that some of the M. rubrum cells had enough ENPs to maintain the percentage of CPNs during the last + three cell divisions for treatment [25] and [50] µM NH4 .

- Inorganic carbon uptake for M. rubrum responded differently in exp. 1 and 2 ([50] and [15] µM NO3 ). A common trend in both experiments was a very low inorganic carbon uptake (< 20 pgC cell-1 h-1) by the end of the experiments. This could be regarded as an effect of N-depletion in combination with a low percentage cells - + with a CPN. The repressive effect of NO3 uptake by NH4 made a potential relation between the percentage - of cells with a CPN and NO3 uptake impossible to explore in detail.

Nutrient uptake by Dinophysis acuminata + We found significant uptake of NH4 that affected both cell concentration and inorganic carbon uptake of D. - + acuminata. Uptake rate of NO3 was demonstrated, but primarily in the control treatments, where no NH4 was added, and in the beginning of the experiment. Uptake of inorganic nitrogen and the effect on cell concentration by D. acuminata has been studied in both field and laboratory studies previously, but inorganic carbon uptake has not previously been included. Field studies on Dinophysis sp. have shown different results whether nutrients affected growth (Delmas, Herbland et al. 1992, Seeyave, Probyn et al. 2009, Hattenrath-Lehmann, Marcoval et al. 2015). Similarly, a few studies on the uptake of inorganic nitrogen and the effects on the growth rate have been carried out on D. acuminata in cultures with mixed results (Hattenrath‐Lehmann and Gobler 2015, Tong, Smith et al. 2015, García-Portela, Reguera et al. 2020).

- NO3 assimilation by D. acuminata has been found previously, but at much lower rates than measured here (see table 2 for specific rates) (Seeyave, Probyn et al. 2009, Hattenrath‐Lehmann and Gobler 2015, García- Portela, Reguera et al. 2020). García-Portela et al. (2020) showed, using transcriptomes, that D. acuminata - express NO3 - transporter, however at values similar to heterotrophic dinoflagellatese. Tong et al. (2015) found - + - + no evidence of NO3 or NH4 assimilation, and no effects of NO3 or NH4 on cell concentration. Hattenrath- + - Lehmann & Gobler (2015) and Seeyave et al. (2009), however, found uptake of NH4 and to less extent NO3 and a significant increase of the nutrients on growth rates. To allow comparisons of the different measurements of uptake rates, all rates were converted into daily rates (Table 2). To be able to do this comparison the effects - + of light on uptake rates need to be considered. The enzymes involved in NO3 and NH4 uptake differ in the - + sense that uptake of NO3 is usually light-dependent, whereas uptake of NH4 is less affected by light (Glibert, + - Wilkerson et al. 2016). Higher uptake rates for NH4 over NO3 were expected due to a simpler metabolism of + + + NH4 . D. acuminata was however unable to deplete NH4 in treatment 25 and [50] µM NH4 , despite the uptake + rates. This response could suggest that 1) the cells have assimilated excess amount of NH4 and explicit a + repressive effect on NH4 uptake (Glibert, Kana et al. 2013) or 2) the cells somehow lose the ability to remove + NH4 over time. Hattenrath-Lehmann & Gobler (2015) and García-Portela et al. (2020) bothe tested uptake of + - NH4 and NO3 once in cultures labeled “well fed” and “starved”, and found ~200% higher uptake rates of + NH4 for “well fed”. Together with our finding, this could suggest that D. acuminata somehow loses the ability + to assimilate NH4 .

+ Cell division of D. acuminata was significantly promoted by the addition of NH4 (Fig. 8 E and F). The inorganic carbon uptake was likewise affected with significant higher rates found in the treatments with + addition of NH4 . The control treatments had less than one cell division during our experiment, hence D. - acuminata were not able to assimilate NO3 into an increase of cell concentration nor inorganic carbon uptake (Fig. 8 E and F ; 9 E and F).

Comparison of the three different organisms with regard to uptake and utilization of different nitrogen forms Mesodinium rubrum and its prey Teleaulax amphioxeia M. rubrum and T. amphioxeia showed very similar responses in regards to uptake rates, depletion of N and - utilization. Basically our data cannot differentiate the two, in response to removal and utilization of NO3 and + - NH4 . The only difference observed among the two species was for M. rubrum in the [15] µM NO3 experiment, + where the inorganic carbon uptake was significantly higher for the two high NH4 treatments (Fig 9D). To evaluate M. rubrum further, the assimilation ability of the chloroplasts, after acquisition, were compared to the original donor, T. amphioxeia. M. rubrum contains an average of ~20 chloroplasts cell-1, during starvation (Kim, Drumm et al. 2017). If we convert uptake rates found by M. rubrum, to uptake rate pr. chloroplast - (divided by 20), the uptake rates of NO3 are slightly below the rates found by T. amphioxeia, whereas the + uptake rates of NH4 were very similar (Fig 11). This implies that the chloroplasts, after acquisition by M. rubrum, were slightly decreased in assimilation capacity. M. rubrum has previously been known to enhance the photosynthetic ability T. amphioxeia-chloroplast-1. The chloroplasts became swollen in size and the chlorophyll a chloroplast-1 likewise increased (Hansen and Fenchel 2006).

Dinophysis acuminata and its prey Mesodinium rubrum Dinophysis acuminata reacted generally differently with regard to uptake rates, depletion of N and utilization, - compared to both M. rubrum and T. amphioxeia. Despite an initial uptake of NO3 by D. acuminata it was - + unable to deplete NO3 . On the other hand D. acuminata was able to assimilate NH4 and a direct utilization of + NH4 into both cell concentration and increased inorganic carbon uptake was found. In fact, the inorganic + carbon uptake for D. acuminata did not decrease, but was maintained despite depletion of ambient NH4 in the + control and [10] µM NH4 treatment, unlike M. rubrum and T. amphioxeia. Hence, the ability to maintain inorganic carbon uptake by D. acuminata seemed to be less dependent on N availability than for T. amphioxeia and M. rubrum. To do similar comparisons of assimilation ability of the chloroplasts by D. acuminata, we used data found by Rusterholz et al. (2017) that estimated the number of chloroplasts ~20 chloroplasts cell-1. The - - uptake rate of NO3 by D. acuminata was significantly higher for the control treatment in the [15] µM NO3 + experiment and the uptake rates of NH4 were generally higher than the rates found for M. rubrum. It thus seems that the potential for uptake rates, similar to T. amphioxeia, could be found within D. acuminata. Why this potential is not utilized by D. acuminata is unknown at present.

- + Figure 11. Measured maximum uptake rates of NO3 (A and B) and NH4 (C and D) by Teleaulax amphioxeia, Mesodinium rubrum and Dinophysis acuminata converted into uptake rate chloroplast-1 (pM N chloropl-1 h-1). It’s assumed that both M. rubrum and D. acuminata contain 20 chloroplasts pr. cell.

The main differences between M. rubrum and D. acuminata when grown without prey were that M. rubrum - + can support cell growth on both NO3 and NH4 , while D. acuminata can only support cell growth when + supplied with NH4 . We can only speculate, but an obvious reason for this difference would be the retention of cryptophyte nuclei (CPN) by M. rubrum. We know that D. acuminata transcripts different genes in relation to chloroplast functioning/nursing and divisions, but that these genes are not always derived from cryptophytes, but from other algal groups, including haptophytes, chlorarachniophytes and cyanobacteria (Wisecaver and Hackett 2010, Hongo, Yabuki et al. 2019).

References Anderson, D., D. Kulis, J. Sullivan, S. Hall and C. Lee (1990). "Dynamics and physiology of saxitoxin production by the dinoflagellatesAlexandrium spp." Marine biology 104(3): 511-524.

Delmas, D., A. Herbland and S. Y. Maestrini (1992). "Environmental-Conditions Which Lead to Increase in Cell- Density of the Toxic Dinoflagellates Dinophysis Spp in Nutrient-Rich and Nutrient-Poor Waters of the French Atlantic Coast." Marine Ecology Progress Series 89(1): 53-61.

Garcia-Cuetos, L., Ø. Moestrup, P. J. Hansen and N. Daugbjerg (2010). "The toxic dinoflagellate Dinophysis acuminata harbors permanent chloroplasts of cryptomonad origin, not kleptochloroplasts." Harmful Algae 9(1): 25-38.

García-Portela, M., B. Reguera, J. Gago, M. L. Gac and F. Rodríguez (2020). "Uptake of Inorganic and Organic Nitrogen Sources by Dinophysis acuminata and D. acuta." Microorganisms 8(2): 187.

Giordano, M. (1997). "Adaptation of Dunaliella salina (Volvocales, Chlorophyceae) to growth on NH4+ as the sole nitrogen source." Phycologia 36(5): 345-350.

Glibert, P. M., T. M. Kana and K. Brown (2013). "From limitation to excess: the consequences of substrate excess and stoichiometry for phytoplankton physiology, trophodynamics and biogeochemistry, and the implications for modeling." Journal of Marine Systems 125: 14-28.

Glibert, P. M., F. P. Wilkerson, R. C. Dugdale, J. A. Raven, C. L. Dupont, P. R. Leavitt, A. E. Parker, J. M. Burkholder and T. M. Kana (2016). "Pluses and minuses of ammonium and nitrate uptake and assimilation by phytoplankton and implications for productivity and community composition, with emphasis on nitrogen‐ enriched conditions." Limnology and Oceanography 61(1): 165-197.

Hansen, P. J. and T. Fenchel (2006). "The bloom-forming ciliate Mesodinium rubrum harbours a single permanent endosymbiont." Mar Biol Res 2(3): 169-177.

Hansen, P. J., L. T. Nielsen, M. Johnson, T. Berge and K. J. Flynn (2013). "Acquired phototrophy in Mesodinium and Dinophysis -A review of cellular organization, prey selectivity, nutrient uptake and bioenergetics." Harmful Algae 28: 126-139.

Hattenrath-Lehmann, T. K., M. A. Marcoval, H. Mittlesdorf, J. A. Goleski, Z. Wang, B. Haynes, S. L. Morton and C. J. Gobler (2015). "Nitrogenous nutrients promote the growth and toxicity of Dinophysis acuminata during estuarine bloom events." PloS One 10(4): e0124148.

Hattenrath‐Lehmann, T. and C. J. Gobler (2015). "The contribution of inorganic and organic nutrients to the growth of a North American isolate of the mixotrophic dinoflagellate, Dinophysis acuminata." Limnology and Oceanography 60(5): 1588-1603.

Herfort, L., T. D. Peterson, F. G. Prahl, L. A. McCue, J. A. Needoba, B. C. Crump, G. C. Roegner, V. Campbell and P. Zuber (2012). "Red waters of Myrionecta rubra are biogeochemical hotspots for the Columbia River estuary with impacts on primary/secondary productions and nutrient cycles." Estuaries and Coasts 35(3): 878-891.

Hongo, Y., A. Yabuki, K. Fujikura and S. Nagai (2019). "Genes functioned in kleptoplastids of Dinophysis are derived from haptophytes rather than from cryptophytes." Scientific reports 9.

Johansson, N., E. Graneli, T. Yasumoto, P. Carlsson and C. C. Legrand (1996). "Toxin production by Dinophysis acuminata and D. acuta cells grown under nutrient sufficient and deficient conditions."

Johnson, M. D. and D. K. Stoecker (2005). "Role of feeding in growth and photophysiology of Myrionecta rubra." Aquat Microb Ecol 39(3): 303-312.

Kim, M., K. Drumm, N. Daugbjerg and P. J. Hansen (2017). "Dynamics of sequestered cryptophyte nuclei in Mesodinium rubrum during starvation and refeeding." Frontiers in Microbiology 8: 423.

Kim, M., S. W. Nam, W. Shin, D. W. Coats and M. G. Park (2012). "Dinophysis caudata (dinophyceae) sequesters and retains plastids from the mixotrophic ciliate prey mesodinium rubrum." J Phycol 48(3): 569- 579.

Koroleff, F. (1970). "The above paper revised, Int. Con. Explor. Sea, Information on techniques and methods for sea water analysis." Inter lab Repor t(3): 19-22.

Moeller, H. V., M. D. Johnson and P. G. Falkowski (2011). "Photoacclimation in the phototrophic marine ciliate Mesodinium rubrum (Ciliophora)." J Phycol 47(2): 324-332.

Nielsen, L. T., B. Krock and P. J. Hansen (2012). "Effects of light and food availability on toxin production, growth and photosynthesis in Dinophysis acuminata." Marine Ecology Progress Series 471(3750): 22.

Park, M. G., S. Kim, H. S. Kim, G. Myung, Y. G. Kang and W. Yih (2006). "First successful culture of the marine dinoflagellate Dinophysis acuminata." Aquatic Microbial Ecology 45(2): 101-106.

Raven, J. A., B. Wollenweber and L. L. Handley (1992). "A comparison of ammonium and nitrate as nitrogen sources for photolithotrophs." New Phytologist 121(1): 19-32.

Riisgaard, K. and P. J. Hansen (2009). "Role of food uptake for photosynthesis, growth and survival of the mixotrophic dinoflagellate Dinophysis acuminata." Mar Ecol Prog Ser 381: 51-62.

Rusterholz, P. M., P. J. Hansen and N. Daugbjerg (2017). "Evolutionary transition towards permanent chloroplasts?-Division of kleptochloroplasts in starved cells of two species of Dinophysis (Dinophyceae)." PloS one 12(5): e0177512.

Seeyave, S., T. Probyn, G. Pitcher, M. Lucas and D. Purdie (2009). "Nitrogen nutrition in assemblages dominated by Pseudo-nitzschiaspp., Alexandrium catenellaand Dinophysis acuminataoff the west coast of South Africa." Mar Ecol Prog Ser 379: 91-107.

Smith, M. and P. J. Hansen (2007). "Interaction between Mesodinium rubrum and its prey: importance of prey concentration, irradiance and pH." Mar Ecol Prog Ser 338: 61-70.

Solórzano, L. and J. H. Sharp (1980). "Determination of total dissolved nitrogen in natural waters 1." Limnology and Oceanography 25(4): 751-754.

Thoisen, C., M. T. T. Vu, T. Carron-Cabaret, P. M. Jepsen, S. L. Nielsen and B. W. Hansen (2018). "Small-scale experiments aimed at optimization of large-scale production of the microalga Rhodomonas salina." Journal of Applied Phycology 30(4): 2193-2202.

Tong, M., J. L. Smith, D. M. Kulis and D. M. Anderson (2015). "Role of dissolved nitrate and phosphate in isolates of Mesodinium rubrum and toxin-producing Dinophysis acuminata." Aquatic Microbial Ecology 75(2): 169.

Wilkerson, F. P. and G. Grunseich (1990). "Formation of blooms by the symbiotic ciliate Mesodinium rubrum: the significance of nitrogen uptake." Journal of Plankton Research 12(5): 973-989.

Wisecaver, J. H. and J. D. Hackett (2010). "Transcriptome analysis reveals nuclear-encoded proteins for the maintenance of temporary plastids in the dinoflagellate Dinophysis acuminata." BMC genomics 11(1): 1.

Wood, G. J. and K. J. Flynn (1995). "Growth of Heterosigma carterae (Raphidophyceae) on nitrate and ammonium at three photon flux densities: evidence for N stress in nitrate‐growing cells." Journal of phycology 31(6): 859-867.

Yih, W., H. S. Kim, H. J. Jeong, G. Myung and Y. G. Kim (2004). "Ingestion of cryptophyte cells by the marine photosynthetic ciliate Mesodinium rubrum." Aquatic Microbial Ecology 36(2): 165-170.

Supporting Information

S1 Fig. PH measurements Time course of pH measurements during incubation of Teleaulax amphioxeia (A and B), Mesodinium rubrum (C and D) and Dinophysis acuminata (E and F), for the four different treatment (control, [10] µM + + + NH4 , [25] µM NH4 and [50] µM NH4 ) and for experiment 1 and 2. Points are means while error bars represent standard error of triplicate measurements.

S2 Fig. Experiment 3.

Time course of ambient NO3 (light grey) and NH4 (dark grey) for all treatments treatment: control (A), [10] + + + µM NH4 (B), [25] µM NH4 (C) and [50] µM NH4 (D) during incubation of Mesodinium rubrum. Solid lines are uptake rates (pM cell-1 d-1), calculated from ambient concentrations. Points are means while error bars represent standard error of triplicate measurements.

Effects of nitrogen on two non-constitutive mixotrophs, Mesodinium rubrum and Dinophysis acuminata II: Potential for uptake of ammonium, nitrate and urea as a function of time after prey deprivation

Kirstine Drumm1, Patricia M. Glibert2, Kevin J. Flynn3 and Per J. Hansen4

1University of Copenhagen, Department of Biology, Helsingør, Denmark

2 University of Maryland, Center for Environmental Science, Cambridge, Maryland, US

3Swansea University, Department of Biosciences, Wales, UK

Abstract The ability of non-constitutive mixoplankton (NCMs) to take up inorganic nitrogen has received very little attention in the literature. Recent studies have shown that specialist NCMs, like Mesodinium rubrum and Dinophysis acuminata, which both can retain kleptochloroplasts for a considerable time, can use inorganic - + nitrogen sources. While M. rubrum can take up NO3 and NH4 , D. acuminata has so far only been shown to + take up NH4 at rates significant for growth. In this study, we investigated the potential uptake (i.e., the initial - + transport rate) of NO3 , NH4 and urea into M. rubrum and D. acuminata; these potentials were measured as a - function of time after prey deprivation at two different initial NO3 concentrations ([50], and [15] µM N) in + combination with four different initial NH4 concentrations ([0], [10], [25] and [50] µM N). The potential of nutrient uptake rates of the two NCMs were compared to that of Teleaulax amphioxeia, the chloroplast donor for both species. Our study revealed no negative effect of prey deprivation on the potential uptake of nutrient- N by M. rubrum and D. acuminata; on the contrary, potential uptake rates were enhanced in most experiments. - + M. rubrum was able to upregulate the potential uptake of NO3 , NH4 and urea over the first 5 days, similar to - its prey, T. amphioxeia. D. acuminata was also able to upregulate the potential uptake of NO3 and urea, but + + only for the treatments to which < 10 NH4 µM N was added. Potential uptake of NH4 by D. acuminata was - + not upregulated over time in the [50] µM NO3 experiment. An increase in uptake of NH4 was measured in - + [15] µM NO3 experiment, but only for the two treatments with no or little NH4 was added. The measured - - maximum potential uptake rates of NO3 for all species occurred in experiments with the highest ambient NO3 + concentration. The measured maximum potential uptake rates of NH4 in M. rubrum and D. acuminata - experiments were positively affected by higher ambient NO3 concentrations, while maximum potential uptake + - of NH4 by T. amphioxeia was not affected by the ambient NO3 concentration. The C:N uptake ratios of D. acuminata were unaffected or increased with incubation time, unlike M. rubrum and T. amphioxeia, where the + C:N uptake ratios were found to decrease throughout the experiments. The initial concentration of NH4 in the - - [15] µM NO3 experiments had a significant repressive effect on the potential uptake of NO3 and urea for all species. Finally, by using potential uptake, we showed that despite depletion of inorganic N in the culture medium this did not affect the ability of M. rubrum and D. acuminata to take up N, indicating that potential N uptake is not related to the prey starvation.

Keywords: Dinophysis, Mesodinium, nitrate, ammonium, urea, uptake rates, 15N incubations, time of prey deprivation

Introduction Nitrogen (N) is often the limiting macro-nutrient for growth of phytoplankton in marine coastal and oceanic - waters. The inorganic N-forms that are of greatest importance for growth of phytoplankton are nitrate (NO3 ) + + and ammonium (NH4 ) (Ryther and Dunstan 1971, Goldman, McCarthy et al. 1979). NH4 is generally considered the preferred form of inorganic N for phytoplankton growth due to the lower energy requirements for its transport over the cell membrane and assimilation into the cell (Raven, Wollenweber et al. 1992, Berges - 1997). This difference in assimilation of inorganic N can result in a delayed uptake of NO3 , most often caused - - by repression of NO3 transport over the cell membrane or inhibition of NO3 reduction in the cell (Berges 1997, Vergara, Berges et al. 1998, Glibert, Wilkerson et al. 2016).

Recently, the dichotomy of plankton organisms into phytoplankton (plants) and zooplankton (animals) has been challenged as many protist plankton organisms are mixotrophic, mixing phototrophy and phagotrophy (Mitra, Flynn et al. 2016, Flynn, Mitra et al. 2019). Mixoplankton can be separated into constitutive mixotrophs (CMs) that are organisms with the capability to synthesis and regulate chloroplasts, and non-constitutive mixotrophs (NCMs) that are organisms that acquire functional chloroplasts from their prey for a short or longer time. Little is known about nitrogen uptake among the NCMs.

The red tide ciliate Mesodinium rubrum and the toxic dinoflagellate Dinophysis acuminata are common and important NCMs worldwide (Hansen, Nielsen et al. 2013). Both these mixotrophs are known to survive and even go through 4-8 cell divisions following prey depletion (Hansen, Nielsen et al. 2013, Drumm et al. 2020, paper 1). Also, they are known to have considerable control of their chloroplasts such as the ability of division of the kleptochloroplasts and production of photosynthetic pigments. The two species differ in the sense that D. acuminata only sequesters the chloroplasts, while M. rubrum also retain prey nuclei and nucleomorphs, prey mitochondria, prey ribosomes and prey cytoplasm (Gustafson, Stoecker et al. 2000, Park, Kim et al. 2006). Nevertheless, the reason for the termination of cell growth and utilization of the retained chloroplasts in the 2 species is unknown. Speculations on a correlation between uptake of nutrients and time of prey deprivation have been proposed for D. acuminata (Hattenrath‐Lehmann and Gobler 2015).

Studies have so far shown contrasting results on the possible utilization of dissolved nutrient-N by M. rubrum and D. acuminata (Delmas, Herbland et al. 1992, Johansson, Graneli et al. 1996, Seeyave, Probyn et al. 2009, Hattenrath-Lehmann, Marcoval et al. 2015, Hattenrath‐Lehmann and Gobler 2015, Tong, Smith et al. 2015, - Gao, Hua et al. 2018, García-Portela, Reguera et al. 2020). M. rubrum has been reported to assimilate NO3 , + + NH4 and dissolved organic N, however the uptake of NH4 was questioned by (Tong, Smith et al. 2015). + Several studies have indicated that D. acuminata can assimilate NH4 and dissolved organic N, while results - are inconclusive with regard to NO3 assimilation (Hattenrath‐Lehmann and Gobler 2015, García-Portela, Reguera et al. 2020). - + Recently, it was shown that M. rubrum takes up both NO3 and NH4 (Drumm et al. 2020, paper 1). M. rubrum was able to completely deplete the nitrogen forms from the culture medium over a period of 4-6 days and incorporate them into cell growth; very similar to the chloroplast donor, the cryptophyte Teleaulax amphioxeia. + D. acuminata, which does not retain the prey nucleus, was only able to remove part of the supplied NH4 . + - Nevertheless, NH4 was incorporated into cell growth and stimulated inorganic carbon uptake. Uptake of NO3 by D. acuminata was shown, but no signs of utilization of this N-form for growth were observed. Also, the study showed that the initial ambient N concentrations had a significant effect on the measured uptake rates. The relation between the uptake of N and carbon fixation has not previously been investigated, but the study indicated that the inorganic carbon uptake was coupled to the uptake of N in M. rubrum, whereas this was not the case of D. acuminata (Drumm et al. 2020, paper 1).

Many questions still remain open with regard to the utilization of the different N-forms in the two species following prey depletion. In the previous study (Drumm et al. 2020, paper 1) inorganic nitrogen were depleted by the organisms during the experiments, making it difficult differentiate actual uptake (calculated by removal) from the potential to take up the nutrients and how prey depletion impacts this in the two organisms.

Stable isotopes (15N) have previously been used to measure potential uptake rates (i.e., the initial transport - + rate) of NO3 and NH4 (and urea) in both M. rubrum and D. acuminata. However, these studies only measured the rates once during their experiment (Wilkerson and Grunseich 1990, Hattenrath‐Lehmann and Gobler 2015, García-Portela, Reguera et al. 2020). The studies found a potential for uptake of N for both species, but how the potential uptake evolves over time during prey starvation or at different amounts of inorganic N concentrations have not been investigated previously. Similar, preferences of N source and metabolic + - interactions such as repression by NH4 on NO3 and urea remain unclear.

In the present study, we monitored the potential nutrient uptake rates as a function of time after prey deprivation in an NCM that retains prey nuclei, M. rubrum and a NCM that does not, D. acuminata. These rates were evaluated by comparison with the chloroplast/nuclei donor (Teleaulax amphioxeia). The two species were - subjected to two different initial concentrations of NO3 (exp. 1 – [50] µM N and exp. 2 – [15] µM N) and four + different concentrations of NH4 ([0], [10], [25] and [50] µM N, referred to as treatments). We hypothesized - that: (1) NO3 is taken up and incorporated in to growth by M. rubrum, but not by D. acuminata; (2) Potential uptake rates of nitrogen by M. rubrum and D. acuminata will decrease as a function of time after prey - + deprivation; (3) Initial concentrations of NO3 and NH4 impact the potential uptake by M. rubrum and D. acuminata; and (4) If inorganic carbon uptake is coupled to the uptake of N, then the C:N uptake ratio will not change as a function of time after prey deprivation for M. rubrum, but it will for D. acuminata. The experiments carried out in present study were done simultaneously with the experiments in Drumm et al. (2020, paper 1). Method and materials A short resume of the overall setup is presented below; full details are given in Drumm et al. (2020, paper 1).

Protist strains and culture conditions The strains of Dinophysis acuminata (Da-DK2007), Mesodinium rubrum (MBL-DK2009) and Teleaulax amphioxeia (SCCAP K- 0434), were grown in f/2 medium based on heat treated (95 °, 90 min) natural seawater (30‰), and received an irradiance of 100 µM photons m-2 s-1 (day:night cycle 16:8h). Mesodinium rubrum was fed T. amphioxeia in a predator:prey ratio of 1:2, twice a week. Dinophysis acuminata was fed M. rubrum once a week using a predator:prey ratio of 1:1. The cultures of M. rubrum were starved for prey a minimum of 2 weeks to avoid transferring any T. amphioxeia to D. acuminata culture.

Experiment 1 and 2. Effects of nitrate concentrations ([50] µM and [15] µM, respectively) and four differentiated ammonium concentrations on potential uptake of nitrate, ammonium and urea + + The experiment was designed to monitor the responses to four different initial NH4 concentrations (T0[NH4 ]): + - - [0], [10], [25] and [50] µM NH4 , respectively in the presence of NO3 (T0[NO3 ]) at [50] (experiment 1) and - 3- at [15] µM NO3 (experiment 2). For both experiments, concentrations of phosphate were held at (T0[PO4 ]): [10] µM, thus yielding initially varying N:P conditions for the different treatments, but high enough to ensure that only N and not P stress is involved. To obtain a medium concentration of f/40, T. amphioxeia culture in f/2 medium were diluted with heat treated seawater. For M. rubrum and D. acuminata the cultures were diluted using inverse filtration (mesh size 10 and 15 µm, respectively) with heat treated seawater. From here all three cultures were treated in the same way. 12 L culture (f/40) was distributed into four flasks. To each of these 3- + flasks stock solution of PO4 and NH4 were added to reach the selected treatment concentrations. Each treatment was again distributed into three new flasks, functioning as triplicates.

Sampling time points were selected based on the cell division rates for each species; thus app. Every second to third day of prey deprivation for M. rubrum and D. acuminata and every day for T. amphioxeia. At each - + 15 sampling point cell concentration, potential uptake of NO3 , NH4 and urea (using N), inorganic carbon uptake

+ - 3- and ambient NH4 , NO3 and PO4 were measured. pH was followed using a SenTix 41 pH electrode (WTW, Germany) connected to a pH meter (WTW, pH 3210, Germany) and calibrated with pH 7 and 10 standard buffers.

- + 15 Potential uptake rates of NO3 , NH4 and urea ( N) To quantify cell specific uptake rates, 3 × 20-100 mL subsamples (depending on cell density) withdrawn from each flask were distributed into sterile tissue culture flasks (TPP, Switzerland). Tracer additions (20 µM) of highly enriched (98%) 15N-labeled compounds of ammonium, nitrate or urea were added to the flasks and incubated for 1 h as the experiment flasks. After incubation the samples were filtered onto a pre-combusted (3h at 450°C) 24mm GF/F (Whatman) glass fiber filter. The filters were dried overnight and placed in separate tinfoil packages. 15N analyses were carried out by mass spectrometry at the University of Maryland Center for Environmental Science.

Cell specific uptake ω (pM N cell-1 h-1) was calculated as:

V PN = ∗ 𝜔𝜔 𝑡𝑡 𝐶𝐶 -1 Where Ct is total amount of cells on the filter, PN is particulate nitrogen (pM L ) found using Shimadzu Total Organic Carbon (TOC) analyzer and V is velocity (V h-1) found as (Glibert and Capone 1993):

atom% excess = Atom% enrichment × incubation time( T) 𝑉𝑉 Where atom%excess is measured 15N minus 0.366 and atom%enrichment found∆ as

% = 15+ 100 𝑁𝑁𝑎𝑎𝑎𝑎𝑎𝑎 𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎 𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒ℎ𝑚𝑚𝑚𝑚𝑚𝑚𝑚𝑚 15 �𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴 𝑁𝑁 𝑁𝑁𝑎𝑎𝑎𝑎𝑎𝑎� ∗ Where 15Nadd is the added 15N tracer and Ambient N is the ambient concentrations of N analyzed on a Seal Analytical Autoanalyzer (see Drumm et al. 2020, paper 1 for more details).

Statistical analysis

- + Measured maximum potential uptake rates as well as the potential uptake rates of NO3 , NH4 and urea as a function of time were evaluated using One- and Two-Way ANOVA (GraphPad), respectively.

Results

- + Potential uptake of NO3 , NH4 and Urea Teleaulax amphioxeia: Potential nitrogen uptake rates (i.e., the initial transport rate, pM N cell-1 h-1) for T. amphioxeia as a function of incubation time after prey deprivation in monocultures were similar in all the - treatments (Fig. 1). Potential uptake rates of NO3 increased from an average of treatments of 0.02 and 0.005 -1 -1 - pM N cell h for the [50] and [15] µM NO3 experiment, respectively, during the first three days to 0.09 and 0.015 pM N cell-1 h-1 after which they declined to values that were slightly less than the initial uptake rates + (Fig. 1 A and B). Potential uptake rates of NH4 increased initially from an average of treatments of 0.004 to 0.02 pM N cell-1 h-1 for both experiments, after which the uptake rates stabilized ~ 0.015 pM N cell-1 h-1 in all treatments (Fig. 1 C and D). Finally, potential uptake rates of urea increased throughout the experiment for all

Figure 1. Potential nitrogen uptake in Teleaulax amphioxeia -1 -h - + Potential uptake rates (pM N cell h ) of NO3 (A and B), NH4 (C and D) and Urea (E and F) during + + incubation of T. amphioxeia for the four treatments (control, [10] µM NH4 , [25] µM NH4 and [50] µM + - NH4 ) initiated with [50] and [15] µM NO3 in exp. 1 and 2, respectively. Points are means, while error bars represent standard error of triplicate measurements. Please note the differentiate y-axis for fig A and B, and E and F.

Figure 2. Potential nitrogen uptake in Mesodinium rubrum -1 -h - + Potential uptake rates (pM N cell h ) of NO3 (A and B), NH4 (C and D) and Urea (E and F) during + + + incubation of M. rubrum, for the four treatments (control, [10] µM NH4 , [25] µM NH4 and [50] µM NH4 ) - initiated with [50] and [15] µM NO3 in exp. 1 and 2, respectively. Points are means while error bars represent standard error of triplicate measurements. Please note the differentiate y-axis for fig A and B, and x-axis between exp. 1 and 2.

Figure 3. Potential nitrogen uptake in Dinophysis acuminata -1 -h - + Potential uptake rates (pM N cell h ) of NO3 (A and B), NH4 (C and D) and Urea (E and F) during incubation + + + of D. acuminata for the four treatments (control, [10] µM NH4 , [25] µM NH4 and [50] µM NH4 ) initiated - with [50] and [15] µM NO3 , in exp. 1 and 2, respectively. Points are means while error bars represent standard error of triplicate measurements. Pease note the differentiate y-axis. - -1 -1 treatments in the [15] µM NO3 experiment from an average of treatments of 0.02 to 0.075 pM N cell h (Fig. - 1F), whereas a decrease in uptake rates were found in the [50] µM NO3 experiment after day three (Fig. 1E).

- Mesodinium rubrum: Potential uptake rates of NO3 increased from an average of treatments of 0.5 to 1.2 and -1 -1 - from 0.05 to 0.25 pM N cell h for the [50] and [15] µM NO3 experiment, respectively in the beginning of the experiments. The rates thereafter decreased to initial uptake rates after 4-6 days without prey (Fig. 2 A and + B). The potential uptake of NH4 by M. rubrum increased from an average of treatments of 0.25 to 1.5 and -1 -1 - from 0.2 to 0.7 pM N cell h for the [50] and [15] µM NO3 experiment, respectively. However, a decrease - was seen after day 6 for the [50] µM NO3 experiments, while the increase continued throughout the [15] µM - NO3 experiments (Fig. 2 C and D). Potential uptake of urea increased initially from an average of treatments -1 -1 - of 0.15 to 0.3 and from 0.03 0.15 pM N cell h for [50] and [15] µM NO3 experiment, respectively, except - for treatment [50] µM NH4 in the [15] µM NO3 experiments, where the rates were unchanged (Fig. 2 E and F).

- Dinophysis acuminata: Potential uptake rates of NO3 increased significantly in the beginning of the -1 -1 + experiment, from 0.19 to 0.45 and 0.33 pM N cell h for treatment control and [10] µM NH4 , respectively, - - in the [50] µM NO3 experiments (Fig. 3A and 4A). The potential uptake rates of NO3 in treatment [25] and + - [50] µM NH4 , were unchanged in the [50] and [15] µM NO3 experiments (Fig. 3 A and B). The measured + - potential uptake of NH4 were quite different in the [50] and [15] µM NO3 experiments. Increased potential + uptake rates were observed initially as a function of elevated additions of NH4 , of 0.35, 0.49, 0.77 and 1.04 -1 -1 + + pM N cell h for the control, [10], [25] and [50] µM NH4 , respectively. However, the uptake of NH4 + - suddenly stopped after day 4 for the treatments with additions of NH4 in the [50] µM NO3 experiment (Fig. + - 3C). The potential uptake of NH4 likewise stopped after day 7 in the control treatment in the 50-µM NO3 + - experiment. The potential uptake rates of NH4 in the [15] µM NO3 experiment showed two patterns (Fig. 3D). + + The potential uptake rates of NH4 in the control and [10] µM NH4 treatments increased rapidly from an -1 -1 + average 0.075 to 2.2 pM N cell h , where after the rates stabilized. For [25] and [50] µM NH4 treatments, the increase was more gradual from 0.11 to 2.17 and from 0.01 to 0.10 pM N cell-1 h-1, respectively. Potential - uptake of urea showed similar responses in the [50] and [15] µM NO3 experiments, with a decreased uptake + with increasing NH4 (Fig. 3E and F). The potential uptake rates of urea were unchanged for [25] and [50] µM + -1 -1 - NH4 treatments of an average of 0.3 and 0.04 pM N cell h for the [50] and [15] µM NO3 experiments, respectively, throughout the experiments. Significant increases were observed for the control and [10] µM + -1 -1 - NH4 treatments up to an average of 0.8 and 0.13 pM N cell h for the [50] and [15] µM NO3 experiments, respectively, followed by a gradual decreased towards the end of the experiments.

-1 -1 - + Figure 4. Measured maximum potential uptake rates (pM N cell h ) of NO3 (A and B), NH4 (C and D) + and urea (E and F) by the three species, comparing the four treatments (control, [10] µM NH4 , [25] µM + + - NH4 and [50] µM NH4 ) in experiments initiated with [50] and [15] µM NO3 in exp. 1 and 2, respectively. Letters indicate significant differences. Data based on measured maximum potential uptake fund for each treatments from fig. 1, 2 and 3. Measured maximum uptake rates of N M. rubrum and T. amphioxeia showed similar response curve for the different treatments as a function of incubation time, however significant differences in the measured maximum potential uptake rates were - + observed (Fig. 4). A comparison of the measured maximum potential uptake rates of NO3 , NH4 and urea + - between control and treatment [50] µM NH4 in the [15] µM NO3 experiments were all significantly different (p<0.05, one-way ANOVA on measured maximum rating). The measured maximum potential uptake rates of - + NO3 and urea declined with increasing additions of NH4 for M. rubrum and T. amphioxeia (Fig 4 B and F). + + Increases potential uptake rates of NH4 were observed with increasing NH4 (Fig. 4D). The measurements of - + measured maximum potential uptake rates of NO3 , NH4 and urea by M. rubrum were significantly different - from each other in the [50] µM NO3 experiments (Fig. 4A, C and E). Similar observations were made for T. + amphioxeia. For this species, significant differences were observed only for the potential uptake rates of NH4 .

D. acuminata showed significantly differentiated potential uptake rates for all treatments in both experiments - + (Fig. 4). The measured maximum potential uptake of NO3 and urea declined with increasing additions of NH4 - in both experiments, as seen for M. rubrum and T. amphioxeia but only in the [15] µM NO3 experiments. The + + - measured maximum potential uptake of NH4 increased with increasing additions of NH4 in the [50] µM NO3 - experiment, whereas the opposite was true for the [15] µM NO3 experiment.

- + Measured maximum potential uptake rates of NO3 , NH4 and urea for the three species measured in the [50] - - µM NO3 experiments were all significantly higher than those measured in the [15] µM NO3 experiments (Fig. + 4; p<0.05, one-way ANOVA on measured maximum rating), except for the potential uptake rates of NH4 in T. amphioxeia.

Uptake of N as a function of ambient N concentration - - The potential uptake rates of NO3 as a function of ambient NO3 concentration showed a similar pattern for all - three species (Fig. 5 A, C and E). The potential uptake increased with increasing ambient NO3 concentration. This was confirm/illustrated with grey linear lines, based on trend lines for the dataset, where measurements + when NH4 was present (grey points) were ignored (Fig. 5 A and C).

+ + Potential uptake rates of NH4 as a function of ambient NH4 concentration were scattered in the cases of T. - amphioxeia and M. rubrum, and the data from the two experiments ([50] and [15] µM NO3 ) does not show - different pattern (Fig. 6 A and C). The [50] µM NO3 experiments with D. acuminata clearly gave the highest + + potential uptake rates of NH4 , indicating an increase of uptake with increased ambient NH4 concentration (Fig. 6E).

- - Figure 5. Uptake of NO3 as a function of ambient NO3 concentration. - - Potential uptake rates (A, C and E) and actual uptake rate (B, D and F) of NO3 as a function of ambient NO3 + concentration. Grey points are uptake rates where ambient NH4 was present in a concentration < 2 µM + NH4 . Dotted lines are drawn to illustrate the uptake kinetics, if we disregard the grey points. Points are means while error bars represent standard error of triplicate measurements. Graphs are a merge of suppl. 1 and 2.

+ + Figure 6. Uptake of NH4 as a function of ambient NH4 concentration. - Potential uptake rates (A, C and E) and actual uptake rates (B, D and F) of NO3 as a function of ambient - 2 NO3 concentration. Michealis-Mentens kinetics were established for Fig D, solid line (r = 0.91). Dotted + lines are drawn to illustrate the uptake kinetics for the three treatments with NH4 , [10], [25] and [50] µM + NH4 , respectively (B, D and F). Points are means while error bars represent standard error of triplicate measurements. Graphs are a merge of suppl. 3 and 4. Carbon : nitrogen (C:N) uptake ratios The calculated C:N uptake ratios were based on the uptake rates of C through inorganic carbon uptake (data - found in Drumm et al. (2020, paper 1) and through uptake of N as the sum of potential uptake rates of NO3 + and NH4 . The C:N uptake ratio as a function of incubation time in both T. amphioxeia and M. rubrum decreased throughout the experiments and for all treatments (Fig. 7 A, B, C and D). A small initial increase could be detected in the C:N uptake ratio, but hereafter the C:N uptake ratio decreased in all treatments to <1, - + except for T. amphioxeia in the [50] µM NO3 experiments where the [50] µM NH4 treatment led to increased C:N uptake ratios compared to ~initial levels (Fig. 7F). The development of the C.N ratios in M. rubrum and T. amphioxeia should though be evaluated taking the elevated pH that developed into account (see Drumm et al. 2020, paper 1).

The C:N uptake ratios in D. acuminata developed different that the two other organisms. The C:N uptake ratios + were unaffected as a function of incubation time in the control and in the treatments with [10] µM NH4 in - + [50] µM NO3 experiments, whereas the C:N uptake ratios in the [25] and [50] µM NH4 treatments increased + from a uptake ratio of 2 to 10. The C:N uptake ratio in the [25] and [50] µM NH4 treatments increased initially + from a C:N uptake ratio of 18 to 70, but later decreased to the same ratio as the control and the [10] µM NH4 - treatments, which both decreased slightly over time in [15] µM NO3 experiments (to an uptake ratio of app 10).

C:N uptake ratios as a function of growth rates showed similar results for all three species (Fig. 8 A-F). Clear positive correlations between C:N uptake ratio and growth rates were generally observed for all species and - experiments. The only exceptions were the T. amphioxeia in the [50] µM NO3 experiments, where an initial correlation was found only no correlation was found (Fig. 8A).

Discussion Hypothesis 1: Nitrate is taken up and incorporated in to growth by M. rubrum, but not by D. acuminata Mesodinium rubrum has previously been shown to have nitrate reductase enzymatic activity (Packard, Blasco - et al. 1978), and earlier studies have shown uptake of NO3 in cells collected from field samples as well as for 2 strains of M. rubrum in the laboratory (Tong, Smith et al. 2015, Drumm et al. 2020, paper 1). The ability of take up nitrate by a Danish strain of M. rubrum strain (MBL-DK2009) was confirmed in Drumm et al. (2020, paper 1) via removal of nitrate form the medium as well as in this paper by the use of stable isotopes.

- Dinophysis acuminata (strain Da-DK2007) in the present study was shown to take up NO3 and confirming - the transport of NO3 over the cell membrane of D. acuminata. In fact, the potential uptake increased over - + time during the [50] µM-NO3 experiment (exp. 1) for the control treatment that had no NH4 addition (Fig. 3A) Figure 7. Development of C:N uptake ratios for the three species during the incubations based on potential uptake measurements. Calculated by converting inorganic carbon uptake into pM C and aggregation of - + potential uptake of NO3 and NH4 . Note that the y-axes are different between experiments 1 and 2.

Figure 8. C:N uptake ratio as a function of growth rates, for each species and treatment, comparing - - experiments 1 ([50] µM NO3 ) and 2 ([15] µM NO3 ). Light grey points are a part of the cells initial lag phase. Dark grey points are from the final day in the experiments. C:N uptake ratios are based on inorganic carbon - + uptake and potential uptake of N (NO3 + NH4 ). Dotted lines are drawn to illustrate the uptake kinetics, if we disregard the grey points. This is contrasting to the recent results that showed indications that D. acuminata was strongly depleted of - nitrate transporters (García-Portela, Reguera et al. 2020). Despite the ability of D. acuminata to take up NO3 , - D. acuminata is apparently unable to utilize NO3 (Drumm et al. 2020, paper 1). Which parts of the metabolic - - pathways of NO3 utilization are malfunctioning in D. acuminata are presently unknown. NO3 reductase has been found to occur in the chloroplasts in many dinoflagellates (Berges and Mulholland 2008), but it is - unknown where this enzyme is located in D. acuminata. Thus, three options are possible: 1) NO3 reductase is - not found in the cytosol of Dinophysis, and it is not acquired together with the chloroplasts, and, 2) NO3 - reductase is found in the cytosol, but the NO2 may not be transferred into the chloroplasts/plastids, or 3) the - + reduction of NO2 into NH4 in the chloroplasts does not take place. The latter two seem however unlikely due - - to the toxicity of NO2 (Flynn and Flynn 1998), and thus NO3 reductase is most likely not found within D. acuminata nor acquired together with the chloroplasts.

- Future studies should investigate causes of the lack of ability of D. acuminata to utilize NO3 as a nitrogen source. This could be done either by measuring enzyme activities or through gene expression studies of D. - acuminata. It could be of particular interest to investigate which NO3 metabolic related enzymes are expressed in D. acuminata as opposed to M. rubrum.

Hypothesis 2: Potential uptake of N by M. rubrum and D. acuminata will decrease as a function of time after prey deprivation Mesodinium rubrum was able to maintain and in most cases increase the potential uptake of all nitrogen sources - + (NO3 , NH4 and urea) through prey deprivation (Fig. 2A-F). These data could suggest an increased uptake of nutrients by M. rubrum through starvation, but the increase was temporary. For most experiments the potential uptake decreased after a week, although never to values below the initial uptake rates (Fig. 2 A-F). The results for M. rubrum were very similar to those obtained for T. amphioxeia, where increased potential uptake rates - + of NO3 , NH4 and urea over time were also observed (Fig. 1 A-F). The potential uptake rates found in present - study for M. rubrum were significantly higher than rates previously published. The potential NO3 uptake rates in present study were in the range of 0.05 to 1.68 pM N cell-1 h-1, whereas Wilkerson & Grunseich (1990) and - -1 -1 Tong et al. (2017) found actual (calculated by removal) NO3 uptake rates of 0.074 and 0.086 pM N cell h , + respectively. Similar differences were found for potential uptake rates of NH4 . The measured maximum uptake rates found in Drumm et al. (2020, paper 1) of 0.35 pM N cell-1 h-1 were more comparable to our + findings. Similar high variations among potential NH4 uptake rates have been found, varying from no uptake found by Tong et al. (2017) and 0.055 pM N cell-1 h-1 by Wilkerson & Grunseich (1990) to 0.67 and 1.6 pM N cell-1 h-1 found by Drumm et al. (2020, paper 1) and present study, respectively. The reasons for these differences in potential N uptake rates differences are unknown, but could be due to strain differences or methodology. The M. rubrum strain studied by Tong et al., was from Inokushi Bay, Japan and belonging to another subclade of M. rubrum. The study by Wilkerson & Grunseich (1990) was on a field population of M.rubrum that might have contained other protist organisms.

- + Dinophysis acuminata was, in most cases, able maintain the potential uptake rates of NO3 , NH4 and urea during 10 days of prey starvation in the present experiments (Fig. 3 A-F). D. acuminata more than doubled the - - potential uptake of NO3 over a period of 7 days, but only for the control treatment, where no NH4 was added, - - -1 -1 in the [50] µM NO3 experiments (Fig 3A). The potential uptake of 0.49 pM NO3 cell h measured in present - -1 -1 study was significantly higher than the rates of 0.005 pM NO3 cell h found previously by Hattenrath‐ Lehmann & Gobler (2015) and García-Portela et al. (2020), but similar to Seeyave et al. (2009) and Drumm - -1 et al. (2020, paper 1) at 0.47 and 0.78 pM NO3 h , respectively.

+ + -1 - The present maximum uptake rates of NH4 by D. acuminata at (1.01 pM NH4 cell h 1) were comparable to + -1 -1 previous findings. Seeyave et al. (2009) found higher rates of 5.8 pM NH4 cell h when measured on field samples, whereas Hattenrath-Lehmann & Gobler (2015) and García-Portela et al. (2020) only measured the + -1 -1 uptake rates once, but on starved (0.096 and 0.6 pM NH4 cell h , respectively) and fed (0.52 and 0.9 pM + -1 -1 NH4 cell h , respectively) D. acuminata, and found very similar potential uptake rates as present study. The + - potential uptake rates of NH4 by D. acuminata gave two very different results in the [15] and [50] µM NO3 - experiments. The results from the [50] µM NO3 experiments gave initially high potential uptake rates. + However, after 4 days of prey starvation, the potential uptake of NH4 stopped (Fig. 3C). Such a complete stop + - in in potential uptake of NH4 was not found in the [15] µM NO3 experiments (Fig. 3D). Instead, we found a + continued increase in the potential uptake of NH4 during this experiment. The two experiments were incubated - at predator:prey ratios of ~280:15 and ~250:50, for the [50] and [15] µM NO3 experiments, respectively, - despite attempts to wash the cultures free from M. rubrum. Thus, D. acuminata cells in the [15] µM NO3 experiments had higher amount of preys available relatively, and it is therefore likely that the chloroplasts - found inside the D. acuminata cells were more recently acquired than in the [50] µM NO3 experiments.

Hattenrath-Lehmann & Gobler (2015) and García-Portela et al. (2020) also found that freshly fed D. acuminata + had higher rates of NH4 uptake in well-fed compared to prey starved cultures. We can only speculate as to + - why NH4 uptake stopped in D. acuminata during prey starvation in the [50] µM NO3 , and decreased in the experiments done by Hattenrath-Lehmann & Gobler (2015). It is possible that either: 1) that the cells had + assimilated excess amounts of NH4 during the first 4 days and that that led to the reduction, or 2) that D. + acuminata had lost the ability to take up NH4 during prey starvation due to ageing of the chloroplasts as also suggested by Hattenrath-Lehmann & Gobler (2015). At present, it is not possible to judge which of the explanations are true. - + Hypothesis 3: Initial concentrations of NO3 and NH4 impact potential uptake by Mesodinium rubrum and Dinophysis acuminata

- Nitrate (NO3 )

- - Measured maximum potential uptake rates of NO3 were affected by the initial NO3 concentration in both M. rubrum and D. acuminata (Fig. 4 C-F). The same hold true for T. amphioxeia (Fig. 4 A and B). This is in accordance with results found for phototrophic protists in general (Glibert, Wilkerson et al. 2016). A positive - effect of ambient NO3 concentration was also observed on the potential uptake rates in the two experiments as a function of ambient nitrate concentration for M. rubrum, as well as the plastid donor alga T. amphioxeia (Fig. + 5 A, C and E), if we disregard uptake rates in treatments where NH4 was available in concentrations < 2µM. - The same holds true if we use the data of actual uptake of NO3 (measured by removal), from Drumm et al. (2020, paper I), (Fig. 5 B and D), except for D. acuminata, where no correlation between uptake and ambient - - NO3 concentration was found (Fig. 5 F). The ambient NO3 concentrations also positively affected the potential + uptake of NH4 and urea in both M. rubrum and D. acuminata (Fig. 4 C-F), as well as in T. amphioxeia. Such + an effect was, however, not observed on actual uptake of NH4 in M. rubrum and D. acuminata (Drumm et al. + 2020, paper 1), where the measured maximum uptake rates of NH4 , calculated by removal, was found not affected by the initial nitrate concentration.

+ Ammonium (NH4 )

+ + The data on the effects of initial NH4 concentration on the measured maximum potential uptake rates of NH4 gave mixed results with regard to M. rubrum and D. acuminata (Fig. 4 C and D). In some cases, an increase + in potential uptake rates were observed at increased NH4 concentrations, while in other cases, a decrease with + + increasing additions of NH4 . (Fig. 4 C and D). For example, the potential uptake rates of NH4 by M. rubrum - + in the [50] µM NO3 experiment increased with increasing additions of NH4 , while this was not observed in - the [15] µM NO3 experiment. The opposite was true for D. acuminata. No correlations with increasing + additions NH4 additions were found for potential uptake for neither D. acuminata nor T. amphioxeia (Fig. 6 A and E). However, the uptake rates calculated from removal found in Drumm et al. (2020, paper 1) showed patterns of high initial uptake, followed by a steep decrease in uptake, similar to our findings (Fig. 6 B and F, + dotted lines). For e.g. D. acuminata, treatments with 50 µM NH4 addition, showed a significantly lower uptake + + + rate, when NH4 concentration was around 30 µM NH4 as appose to treatment [25] µM NH4 , that was initiated + + at a concentration of NH4 at 25 µM (Fig. 6F). An initial amount of NH4 in the experiments with D. acuminata and T. amphioxeia appeared to be taken up for each treatment, followed by a halt/radical decrease in uptake + + rates. This indicates that the cells are either harmed by the available NH4 (NH4 can be toxic to cells (Collos + and Harrison 2014)), or that they have an initial high uptake rate, assimilating excess amount of NH4 , resulting in a self-protecting strategy and closing of uptake (Glibert, Wilkerson et al. 2016). The latter explanation seems + + most likely, since Drumm et al. (2020, paper 1) found the highest utilization of NH4 at highest NH4 treatment. + M. rubrum did not show this behavior, the uptake rates were related to ambient NH4 concentration and the uptake rates could be fitted to Michealis-Mentens kinetics (r2 = 0.91) (Fig. 6D).

+ - NH4 has previously been shown to repress uptake of NO3 in photosynthetic organisms. The concentrations of + - NH4 required to repress NO3 uptake, however, are very different among different protist groups as well as species (Morris and Syrett 1963, Collos and Slawyk 1980, Dortch 1990, Flynn, Fasham et al. 1997). The effects + of NH4 also seem to depend on the ambient nitrate concentration (Carpenter and Dunham 1985). Similar + results have been found on the repression of urea uptake by NH4 (Horrigan and McCarthy 1982, Solomon, Collier et al. 2010). In the present experiments on the measured maximum potential uptake of nitrogen, no + - repressive effect of NH4 on potential NO3 and urea uptake was found for M. rubrum nor for T. amphioxeia, - - - when the ambient NO3 concentration was ≥ [50] µM NO3 (exp. 1) (Fig. 4A and E). However, when NO3 was + - < 15 µM (exp. 2) there was a significant repression by NH4 on urea and NO3 uptake (Fig. 4B and F). In the + - case of D. acuminata the repressive effect of NH4 on potential NO3 and urea uptake was found in both - + experiments (Fig. 4 A, B, E and F). The potential uptake of NO3 was to a lesser extent repressed by the NH4 - concentration than observed from studies of NO3 uptake via removal from the medium (Drumm et al. 2020, - + paper 1). In those experiments, the repressions of NO3 uptake throughout the experiments until NH4 was depleted.

The uptakes rates of urea by M. rubrum and D. acuminata confirm our understanding of these organisms as primarily phototrophic organisms (Hansen, Nielsen et al. 2013, Mitra, Flynn et al. 2016). We would have expected higher uptake rates of urea if its trophic mode was more dependent upon feeding. It could therefore be interesting to investigate other Mesodinium species with a different trophic mode to investigate this hypothesis further. The repressive effect of ammonium on urea uptake rates has been observed in other studies, but never for Dinophysis spp and M. rubrum. The N preferences, based on measured maximum potential + - uptake, showed a preference of NH4 > urea > NO3 for D. acuminata and T. amphioxeia, while M. rubrum - preferred NO3 over urea. The reason why M. rubrum differs from the other two organisms is unknown at present.

Hypothesis 4: C:N uptake ratio will not change as a function of time after prey deprivation for M. rubrum, but it will for D. acuminata Drumm et al. (2020, paper 1) showed the uptake of N stops together with carbon fixation for M. rubrum, whereas D. acuminata were able to uphold inorganic carbon uptake over time. This led us to our initial hypothesis, however the results from potential uptake gave a different insight.

The cellular C:N ratio is, on average found around 16:1 in most phototrophic protists (e.g (Geider and La Roche 2002) and the uptake rates should large reflect this ratio, however this was rarely the case. The C:N uptake ratio of M. rubrum decreased as a function of incubation time from app 4 to 0.2 and 30 to 0.2 in the - [50] and [15] µM N-NO3 experiments, respectively (Fig. 7 C-D). This is most likely due to the decreased inorganic carbon uptake rates observed (Drumm et al. 2020, paper 1). The C:N uptake ratio was based on inorganic carbon uptake (C), which decreased continuously during the experiment and the potential uptake (N), which was never found to be lower than the initial rate (Fig. 1 and 3), resulting in a decreasing C:N uptake ratio. The reason for the observed decrease in the inorganic carbon uptake during the experiments were most likely a combination of low ambient N concentration and high pH (> 8.5) caused by cell density (Drumm et - al. 2020, paper 1). The initial C:N uptake ratio differed by a factor of 10 during [50] µM N-NO3 experiment, - - most likely due to the high NO3 concentration. This led to significantly higher potential uptake of NO3 by M. rubrum that was not accompanied with a similar increase in inorganic carbon uptake (Drumm et al. 2020, paper 1). A similar pattern was found for T. amphioxeia, whereas rates of photosynthesis in M. rubrum - decreased. T. amphioxeia had assimilated excess amount of N in the 50-µM NO3 experiments, resulting in inorganic carbon uptake and an increased C:N uptake ratio at the end of the experiment (Drumm et al. 2020, paper 1).

The C:N uptake ratio in D. acuminata decreased as a function of time unlike in M. rubrum and T. amphioxeia. + Instead the C:N uptake ratio increased for the [25] and [50] µM NH4 treatments, representing an unchanged or increasing inorganic carbon uptake together with an stable potential uptake of N. Unlike M. rubrum and T. + + amphioxeia pH was not at a limiting level and NH4 was not depleted for to [25] and [50] µM NH4 treatments. This supports our hypothesis that D. acuminata assimilated excess amount of N to support inorganic carbon uptake.

A positive correlation between high C:N uptake ratio and growth rate was found in all experiments for M. rubrum and D. acuminata (Fig. 8). This correlation is what we would anticipated for a phototropic protist that depends upon uptake of nutrients and photosynthetic activity. This also confirms our understanding of M. + rubrum and D. acuminata as primarily phototrophic organisms. The fact that all NH4 treatments fall on a similar line indicates that the amount added are above saturation levels. However, interpretations should be made with care, since these results also represent a halt of metabolism; growth rate and inorganic carbon uptake decreased in combination with continued potential uptake, hence very low C:N uptake ratios.

Other factors that affected the potential uptake of nitrogen in the experiments Cell membrane transporters

- Potential uptake of NO3 increased significantly for M. rubrum and T. amphioxeia during the experiments, + regardless of the different treatments with NH4 (Fig. 1 and 2 A and B). Since M. rubrum and T. amphioxeia had been grown in F/2 medium (with high amounts of nitrate) prior to the experiments, our data indicate that the potential uptake of nitrate cell-1 h-1 increased with decreasing concentration of nitrate (nitrate levels given in Drumm et al. (2020, paper 1). Thus, it seems likely that more transporters are activated to insure measured maximum nitrogen uptake, when nutrients are limited (Flynn, Page et al. 1999). This does seem to collide with - - previous comments on increased uptake of NO3 with increased ambient NO3 concentration. However, we speculate that for M. rubrum and T. amphioxeia it is a combination of the two we are witnessing here. Hence, - that potential uptake is increased with elevated ambient NO3 concentration and that the cells optimize the - uptake, by upregulation of transporters and enzymes, with decreasing ambient NO3 concentration (Flynn, Fasham et al. 1997).

Cell density and pH

- The potential uptake rates of NO3 by M. rubrum and T. amphioxeia were found to decrease by the end of the experiments. We can only speculate but we suggest that either: 1) the high density of the cells together with high pH measurements (> 9, data shown in Drumm et al. (2020, paper 1)), resulted in a decreased metabolism of the cells, thereby also affecting uptake of nutrients. M. rubrum and T. amphioxeia had stopped dividing, and the inorganic carbon uptake was likewise significantly reduced (Drumm et al. 2020, paper 1), or 2) that the enzymes related to the uptake of nitrate became inactive, due to very low ambient nitrate concentrations. A combination of the two cannot be ruled out.

Conclusion

- + + D. acuminata is able to take up both NO3 and NH4 , but only NH4 can be assimilated leading to an increased - cell growth and inorganic carbon uptake. D. acuminata is able to upregulate the uptake of NO3 through time - - of prey deprivation, however assimilations of NO3 is most likely hindered by the lack of NO3 reductase enzymes. Investigation of C:N uptake ratios showed that inorganic carbon uptake is unrelated to the uptake of + N for D. acuminata. Indication of a negative effect of prey deprivation on the uptake of NH4 was found in one experiment with D. acuminata. Inorganic nitrogen uptake in M. rubrum is very similar to its prey, T. - + amphioxeia and uptake of both NO3 and NH4 lead to increased cell growth. Nitrogen uptake rates is upregulated as function of prey starvation time, and a clear correlation between the uptake of inorganic carbon - and uptake of N is found, similar to T. amphioxeia. The ambient concentration of NO3 is positively correlated - + - with the uptake rate of NO3 and NH4 has a repressive effect on the uptake of NO3 and urea in both M. - rubrum and D. acuminata. Future studies should investigate to which extent D. acuminata lack NO3 reductase + enzymes, as well as reasons behind the apparent stop in NH4 uptake in this species. Could it be that the age of the chloroplasts plays a role in this or is something else going on? Also, it could be interesting to look into a possible role of the retained cryptophyte nucleus in uptake of inorganic N in M. rubrum.

Acknowledgment This work was supported by the Danish Council for Independent Research, grant number 4181-00484 (PH). The authors would like to thank Marianne Saietz for assistance using the TOC analyzer. References Berges, J. A. (1997). "Miniview: algal nitrate reductases." European Journal of Phycology 32(1): 3-8.

Berges, J. A. and M. R. Mulholland (2008). "Enzymes and nitrogen cycling." Nitrogen in Marine Environment: 1385-1444.

Carpenter, E. J. and S. Dunham (1985). "Nitrogenous nutrient uptake, primary production, and species composition of phytoplankton in the Carmans River estuary, Long Island, New York 1." Limnology and Oceanography 30(3): 513-526.

Collos, Y. and P. J. Harrison (2014). "Acclimation and toxicity of high ammonium concentrations to unicellular algae." Marine pollution bulletin 80(1-2): 8-23.

Collos, Y. and G. Slawyk (1980). Nitrogen uptake and assimilation by marine phytoplankton. Primary productivity in the sea, Springer: 195-211.

Delmas, D., A. Herbland and S. Y. Maestrini (1992). "Environmental-Conditions Which Lead to Increase in Cell-Density of the Toxic Dinoflagellates Dinophysis Spp in Nutrient-Rich and Nutrient-Poor Waters of the French Atlantic Coast." Marine Ecology Progress Series 89(1): 53-61.

Dortch, Q. (1990). "The interaction between ammonium and nitrate uptake in phytoplankton." Marine ecology progress series. Oldendorf 61(1): 183-201.

Flynn, K. J., M. J. Fasham and C. R. Hipkin (1997). "Modelling the interactions between ammonium and nitrate uptake in marine phytoplankton." Philosophical Transactions of the Royal Society of London. Series B: Biological Sciences 352(1361): 1625-1645.

Flynn, K. J. and K. Flynn (1998). "Release of nitrite by marine dinoflagellates: development of a mathematical simulation." Marine Biology 130(3): 455-470.

Flynn, K. J., A. Mitra, K. Anestis, A. A. Anschütz, A. Calbet, G. D. Ferreira, N. Gypens, P. J. Hansen, U. John and J. L. Martin (2019). "Mixotrophic protists and a new paradigm for marine ecology: where does plankton research go now?" Journal of Plankton Research.

Flynn, K. J., S. Page, G. Wood and C. R. Hipkin (1999). "Variations in the maximum transport rates for ammonium and nitrate in the prymnesiophyte Emiliania huxleyi and the raphidophyte Heterosigma carterae." Journal of Plankton Research 21(2).

Gao, H., C. Hua and M. Tong (2018). "Impact of Dinophysis acuminata feeding Mesodinium rubrum on nutrient dynamics and bacterial composition in a microcosm." Toxins 10(11): 443.

García-Portela, M., B. Reguera, J. Gago, M. L. Gac and F. Rodríguez (2020). "Uptake of Inorganic and Organic Nitrogen Sources by Dinophysis acuminata and D. acuta." Microorganisms 8(2): 187.

Geider, R. J. and J. La Roche (2002). "Redfield revisited: variability of C [ratio] N [ratio] P in marine microalgae and its biochemical basis." European Journal of Phycology 37(1): 1-17.

Glibert, P. M. and D. G. Capone (1993). "Mineralization and assimilation in aquatic, sediment, and wetland systems." Nitrogen isotope techniques: 243-272.

Glibert, P. M., F. P. Wilkerson, R. C. Dugdale, J. A. Raven, C. L. Dupont, P. R. Leavitt, A. E. Parker, J. M. Burkholder and T. M. Kana (2016). "Pluses and minuses of ammonium and nitrate uptake and assimilation by phytoplankton and implications for productivity and community composition, with emphasis on nitrogen‐enriched conditions." Limnology and Oceanography 61(1): 165-197.

Goldman, J. C., J. J. McCarthy and D. G. Peavey (1979). "Growth rate influence on the chemical composition of phytoplankton in oceanic waters." Nature 279(5710): 210.

Horrigan, S. and J. McCarthy (1982). "Phytoplankton uptake of ammonium and urea during growth on oxidized forms of nitrogen." Journal of Plankton Research 4(2): 379-389.

Johansson, N., E. Graneli, T. Yasumoto, P. Carlsson and C. C. Legrand (1996). "Toxin production by Dinophysis acuminata and D. acuta cells grown under nutrient sufficient and deficient conditions."

Morris, I. and P. Syrett (1963). "The development of nitrate reductase in Chlorella and its repression by ammonium." Archiv für Mikrobiologie 47(1): 32-41.

Packard, T., D. Blasco and R. Barber (1978). Mesodinium rubrum in the Baja California upwelling system. Upwelling ecosystems, Springer: 73-89.

Park, M. G., S. Kim, H. S. Kim, G. Myung, Y. G. Kang and W. Yih (2006). "First successful culture of the marine dinoflagellate Dinophysis acuminata." Aquatic Microbial Ecology 45(2): 101-106.

Raven, J. A., B. Wollenweber and L. L. Handley (1992). "A comparison of ammonium and nitrate as nitrogen sources for photolithotrophs." New Phytologist 121(1): 19-32.

Ryther, J. H. and W. M. Dunstan (1971). "Nitrogen, phosphorus, and eutrophication in the coastal marine environment." Science 171(3975): 1008-1013.

Solomon, C. M., J. L. Collier, G. M. Berg and P. M. Glibert (2010). "Role of urea in microbial metabolism in aquatic systems: a biochemical and molecular review." Aquatic Microbial Ecology 59(1): 67-88. Tong, M., J. L. Smith, D. M. Kulis and D. M. Anderson (2015). "Role of dissolved nitrate and phosphate in isolates of Mesodinium rubrum and toxin-producing Dinophysis acuminata." Aquatic Microbial Ecology 75(2): 169.

Vergara, J. J., J. A. Berges and P. G. Falkowski (1998). "Diel periodicity of nitrate reductase activity and protein levels in the marine diatom Thalassiosira weissflogii (Bacillariophyceae)." Journal of Phycology 34(6): 952-961.

Wilkerson, F. P. and G. Grunseich (1990). "Formation of blooms by the symbiotic ciliate Mesodinium rubrum: the significance of nitrogen uptake." Journal of Plankton Research 12(5): 973-989.

Supporting Information

- - S1 Fig. Potential uptake rates of NO3 as a function of ambient NO3 concentration for the three species and the four treatments, comparing experiments 1 and 2. Pease note the differentiate y-axis for fig A and B.

S1 - - S2 Fig. Uptake, calculated from ambient concentrations, of NO3 as a function of ambient NO3 concentration for the three species and the four treatments, comparing experiments 1 and 2. Data from article 1.

+ + + S3 Fig. Potential uptake rates of NH4 as a function of NH4 concentration. Since we added 20 µM NH4 to measure potential uptake, a bar like structure is found here, making an actual pattern difficult to detect.

+ + S4 Fig. Uptake, calculated from ambient concentrations, of NH4 as a function of concentration of NH4 for the three species and the four treatments, comparing experiment 1 and 2. Data from article 1. Dotted lines were fitted by eye and only show trend lines for the treatments with ammonium.

S5 Fig. Time course over C:N uptake ratios, based on ambient measurements. Calculated by converting - + inorganic carbon uptake into pM C and aggregation of uptake of NO3 and NH4 .

Appendix

App. I

Limits to the control of sequestered cryptophyte chloroplasts in the marine ciliate Mesodinium rubrum

Running title: Mesodinium rubrum transcriptome

Nature Ecology and Evolution

Andreas Altenburger1,2*, Huimin Cai3*, Kirstine Drumm4, Miran Kim4,5, Lydia Garcia1, Xiaoyu Zhan3, Shuaicheng Li6, Per Juel Hansen4, Uwe John7, Qiye Li3¤, Nina Lundholm1¤

1University of Copenhagen, Natural History Museum of Denmark, Copenhagen, Denmark

2present address: UiT - The Arctic University Museum of Norway, Tromsø, Norway

3Beijing Genomic Institute, Shenzhen, China

4University of Copenhagen, Department of Biology, Helsingør, Denmark

5Chonnam National University, Research Institute for Basic Science, South Korea

6City University of Hong Kong, Department of Computer Science, China

7Alfred-Wegener-Institute, Bremerhaven, Germany

*shared first authorship

¤shared last authorship

Abstract

The ciliate Mesodinium rubrum is famous for its ability to acquire and use chloroplasts and other cell organelles from some cryptophyte species. Genomes and transcriptomes of free-swimming Teleaulax amphioxeia were sequenced and transcriptional changes upon sequestration by M. rubrum determined under different time and light conditions. 68% of the genes expressed within well-fed M. rubrum originated from T. amphioxeia. These genes contributed 48% to the global transcriptome of M. rubrum. 46% of the T. amphioxeia transcripts were significantly differentially expressed after sequestration. Genes related to metabolic processes were upregulated, whereas genes related to membranes and cytoskeleton were downregulated. Hence, M. rubrum has considerable transcriptional control over the acquired cryptophyte. This control is not fine-tuned, as transcriptional changes that were observed for different time and light conditions in free-swimming T. amphioxeia got lost after sequestration. This means that M. rubrum can only force one expression pattern out of its acquired cryptophyte organelles and is not able to induce time and light dependent transcriptional changes. This tertiary endosymbiosis system functions for several weeks when deprived of prey. After this time, the ciliate dies if no fresh cryptophytes are available to prey upon. Time will tell whether this system is an evolutionary dead-end or a transitionary stage towards a permanent tertiary endosymbiosis.

Introduction

Endosymbiotic events have enabled eukaryotes to photosynthesize. More than a billion years ago, during a primary endosymbiosis event, a photosynthesizing cyanobacterium was retained by a plastid-lacking protist. Since then plastids have spread throughout the eukaryotic tree of life by secondary and tertiary endosymbiosis. When the cryptophyte Teleaulax amphioxeia is preyed upon by the ciliate Mesodinium rubrum, the chloroplast of T. amphioxeia remains active within the ciliate. In contrast to most other acquired phototrophy systems, M. rubrum does not only keep the chloroplast, but also the nucleus, mitochondria and ribosomes of its prey. Here we show that 1) M. rubrum has considerable control over gene expression of the acquired T. amphioxeia nucleus. 2) This control is not fine-tuned i.e. it does not respond to changing time and light conditions, indicating that M. rubrum can only force one particular pattern of gene expression out of the acquired nucleus.

The cryptophyte T. amphioxeia is an ecologically important, phototrophic marine unicellular protist with a worldwide distribution1. It is 8 – 11 µm long and a member of the enigmatic group of cryptophytes, a group that is challenging to place in the evolutionary tree of life2. Most cryptophytes have permanent chloroplasts, originating from a secondary endosymbiosis event between a red alga and a phylogenetically distinct, non-photosynthetic host3,4. Due to this origin, cryptophyte chloroplasts have a complex bounding membrane topology with four membranes. The nucleomorph, an organelle only found in cryptophytes and chlorarachniophytes, is a highly reduced remnant of the endosymbiotic red algal nucleus. It is situated between the outer two and the inner two membranes of the chloroplast5-7. Cryptophytes hence possess DNA of different origin: red algal nuclear DNA in the nucleomorph, plastid DNA, cryptophyte mitochondrial DNA, and cryptophyte nuclear DNA8.

Being primary producers, phototrophic cryptophytes are at the base of the marine food web, and grazed upon by heterotrophic and mixotrophic protists9. One of these grazers is Mesodinium rubrum, an abundant and ecologically important ciliate. M. rubrum is widely distributed in coastal ecosystems and famous for causing non-toxic red tides10-12. Acquisition of phototrophy by retaining a chloroplast that originated from a secondary endosymbiosis event is regarded as tertiary endosymbiosis13,14. M. rubrum preys on cryptophytes belonging to the genera Geminigera, Teleaulax and Plagioselmis. M. rubrum cells keep around 20 chloroplasts from its cryptophyte prey, 3 and usually a single enlarged prey nucleus located close to the nuclei of the ciliate (ciliates have two macronuclei and one micronucleus)15-17. In order to sustain its maximum growth rate of ~0.5 per day, M. rubrum have to ingest ~one cryptophyte per day18,19. M. rubrum covers > 98% of its carbon need via photosynthesis at natural prey concentrations, and can replicate the acquired chloroplasts approximately four times, after which they are degraded and M. rubrum dies if no new cryptophyte nuclei and chloroplasts are provided with new cryptophyte prey18-21. Thus, this tertiary endosymbiosis between a cryptophyte and M. rubrum is not permanent and stable22.

The regulation of cryptophyte genes within M. rubrum has previously been studied using RNA-seq or EST and microarray approaches23,24. These studies found a remarkable cellular and metabolic chimerism between host and prey, and showed that M. rubrum not only sequesters the organelle machinery of its prey, but also the anabolic potential of the sequestered organelles24. Most cryptophyte genes involved in the photosynthesis were up-regulated after sequestration of the cryptophyte nucleus and chloroplasts into the ciliate23. Previous studies lacked biological replicates and had the challenge to distinguish between transcripts originating from M. rubrum and transcripts originating from the prey cryptophytes. We used genomic DNA (gDNA) data from free-swimming T. amphioxeia and starved M. rubrum to overcome this problem. By screening for k-mers shared between gDNA reads and transcripts, we were able to assign transcripts to the right species. Using this approach, we are able to enhance our understanding of transcriptional changes upon sequestration, as well as to elucidate how the acquisition of chloroplasts is regulated after sequestration by M. rubrum. We investigated changes in the level of genes expressed before and after ingestion by M. rubrum and compared those with starved M. rubrum cells that have lost the prey nucleus (Fig. 1). We explored regulation of the sequestered chloroplasts in response to changing light and time conditions (night, morning and day) corresponding to darkness, 20 minutes after turning on the light, and full light.

Figure 1. Light micrographs of Teleaulax amphioxeia and Mesodinium rubrum with corresponding cartoons. a, free swimming T. amphioxeia with chloroplast, nucleomorph, mitochondrion and nucleus. The outer membrane of the nucleus is connected to the outer membrane of the chloroplast. b, well-fed M. rubrum with two macronuclei and one micronucleus, and one enlarged cryptophyte nucleus. M. rubrum contains its own mitochondria, cryptophyte mitochondria, and cryptophyte chloroplasts that are arranged along the periphery of the cell. c, starved M. rubrum with two macronuclei, one micronucleus and ciliate mitochondria. Note: starved M. rubrum cultures where at least 90% of cells had lost the cryptophyte nucleus. Note also: Well-fed cells of M. rubrum have one enlarged cryptophyte nucleus, which is always located in the center of the cell, termed CPN (centered prey nucleus)23. Well-fed cells might keep some extra prey nuclei in the periphery of the cell. Upon ciliate cell division, one of the two daughter cells receives the CPN, while in the other, one of the extra prey nuclei migrate close to the ciliate nuclei and enlarges16. Scale bar in a equals 5 µm, in b and c 10µm.

Results

Transcriptomes of Mesodinium rubrum and Teleaulax amphioxeia

We identified 72,061 and 22,250 genes from T. amphioxeia and M. rubrum, respectively (Table 1, supplementary tables 1 and 2). 63.2% of T. amphioxeia and 51.1% of M. rubrum genes were annotated. Interestingly the GC content of T. amphioxeia is 59%, and thus considerably higher than the one of M. rubrum with a GC content of 34%. M. rubrum keeps all the genetic material from the acquired cryptophyte (Fig. 2a). Considering all sampling points and replicates together, 95% of T. amphioxeia genes were actively transcribed in free-swimming T. amphioxeia at some time-point during night, morning, or day (Fig. 2b). After sequestration by M. rubrum, 77% of the T. amphioxeia genes were actively transcribed at some time- point during the sampling cycle (Fig. 2b, supplementary table 3.).

Table 1. Summary of the transcriptome assembly. The number genes, their total and length, N50, GC content and the percentage of genes with ORF are listed (nt, nucleotides; N50, minimum contiq length needed to cover 50% of the genome; ORF, open reading frame).

T. amphioxeia M. rubrum

Assembly

Number of genes 72,061 22,250

Total length (nt) 78,195,118 19,532,541

Mean length (nt) 1,057 856

N50 (nt) 2,187 1,326

GC content 59% 34%

Genes with ORF [≥ 50 aa] 52,545 (72.9%) 21,109 (94.9%)

Annotation

KEGG 37,675 8,878

GO 21,991 6,827

UniProtKB/Swiss-Prot and TrEMBL 44,786 11,159

Annotated genes 45,567 (63.2%) 11,374 (51.1%)

Figure 2. Transcriptome features of Teleaulax amphioxeia and Mesodinium rubrum. a. Screening of M. rubrum DNA read sets against the 21-mer sketches of T. amphioxeia transcripts with mash [identity >= 0.95 and p-value < 1e-5]. b. proportion of actively transcribed T. amphioxeia genes before and after sequestration. c. global transcriptome of M. rubrum with proportion of T. amphioxeia and M. rubrum genes. d. global transcriptome of M. rubrum with proportion of transcript abundance originating from T. amphioxeia or M. rubrum.

The majority of genes expressed within M. rubrum originated from T. amphioxeia 67.8 ± 1.2% (well fed, all time points) and 51.0 ± 1.3% (starved all time points) (Fig. 2c). However, the contribution of T. amphioxeia transcripts to the global transcriptome of M. rubrum ranged - depending on condition - from 47.9 ± 1.4% (well 7 fed morning) to 11.2 ± 0.8% (starved night) (Fig. 2d). This means that M. rubrum uses fewer genes with higher expression levels, whereas T. amphioxeia contributes a high diversity of genes at lower expression levels to the global transcriptome.

Differential gene expression after chloroplast sequestration

A principal component analysis identified well-defined clusters for free-swimming T. amphioxeia, well-fed M. rubrum and starved M. rubrum (Fig. 3a). After filtration for lowly expressed genes (mean normalized count <= 10), 52,565 T. amphioxeia genes were passed to weighted gene correlation network analyses (WGCNA), which identified 29 modules. Module 1 represented 50% of the genes included in the analysis; they were prevailingly downregulated after sequestration by M. rubrum. Module 2 contained 29% of genes included in the analysis; they were predominately upregulated after sequestration (Fig. 2b). As one might expect for a cell that is taken apart by its host, genes related to the flagella and cytoskeleton are enriched in module 1 (supplementary table 4). Similarly, genes involved in interaction with the environment, related to membranes, signaling, ion channels and ribosomes are found in this module of prevailingly downregulated genes. Genes enriched in module 2 are related to metabolic processes, photosynthesis, cell cycle and transporters (supplementary table 5).

Confirming these results, differentially expressed genes (DEG) analyses showed that 45.7% of T. amphioxeia transcripts were significantly differentially expressed upon sequestration in well-fed M. rubrum. Most DEGs (27.9% in free-living vs inside well- fed M. rubrum cells) were downregulated, while fewer DEGs (17.7% in free-living cells vs inside well-fed M. rubrum cells) were upregulated (Fig. 3c). The difference in gene expression was substantial upon sequestration, and thus in free-swimming vs inside well-fed or starved M. rubrum groups (Fig. 3c). Compared to that, only few T. amphioxeia genes were significantly differentially expressed in well-fed vs starved M. rubrum cells (Fig. 3c). The majority of the DEGs were shared (DEGs with the same regulated direction, up or down) between free-living-vs-inside well-fed M. rubrum and free-living-vs-inside starved M. rubrum (Fig. 3d), although the amount of DEGs was slightly smaller in starved M. rubrum (Fig. 3d). The 26,066 T. amphioxeia genes shared between well-fed and starved M. rubrum were used for GO term enrichment analyses. Genes related to the membranes and cytoskeleton were enriched in downregulated DEGs (supplementary table 6), genes related to metabolic processes, 8 catalytic activity, and DNA repair were enriched in the genes that got upregulated upon sequestration (supplementary table 7).

Fig. 3. Changes in gene expression of Teleaulax amphioxeia in response to sequestration. a. Principal component analysis of T. amphioxeia genes show a clear segregation between free-living, well-fed, and starved samples. b. modules 1 (downregulated upon sequestration) and 2 (upregulated upon sequestration) as identified weighted gene correlation network analysis (WGCNA). c. amount of differentially expressed genes upon sequestration. d. proportion of differentially expressed genes T. amphioxeia genes in free-living vs well-fed and free-living vs starved condition.

Differential gene expression according to time and light condition

Free-swimming T. amphioxeia differentially expressed genes according to time and light conditions during night, morning and day (Fig 4a), which is expected in a photosynthetic organism that downregulates genes related to photosynthesis at night. This differential expression according to time and light condition ceases upon sequestration by M. rubrum (Fig. 4b). Hence, M. rubrum has control over the cryptophyte nucleus, but the time and light dependent expression pattern gets lost. The expression pattern of T. amphioxeia genes in M. rubrum is time and light independent after sequestration (Figs. 4c, e, supplementary table 8), so apparently fine-tuning of genes is lost upon sequestration. Photosynthesis related genes got upregulated upon sequestration (supplementary table and Fig. 9).

Fig. 4. Changes in light and time controlled gene expression of free-swimming T. amphioxeia and after sequestration. a. Pearson correlation analysis of T. amphioxeia genes among different samples show differences according to time and light condition. b. Pearson correlation analysis of T. amphioxeia genes after sequestration by Mesodinium rubrum. c. Amount of T. amphioxeia genes that were differentially expressed according to time and light condition in free-swimming cells and after sequestration by M. rubrum. d. heat map of T. amphioxeia genes that get differentially expressed according to time and light condition, before and after sequestration.

Discussion

Global transcriptome

In contrast to previous studies that argued for culturing of intact T. amphioxeia cells within M. rubrum25, our results confirm that the cryptophyte nucleus within M. rubrum has no autonomy, a finding that has also been put forward previously23,24,26. In other words, M. rubrum has a high degree of control over the acquired prey nucleus, and the relationship between cryptophyte nucleus and M. rubrum is best described as an enslavement, a concept that has previously been used to describe the nucleomorph of red algal origin in cryptophytes and of green algal origin in chlorarachnio- phytes6,7,27. What is new in our findings is that even though, M. rubrum has considerable control of the cryptophyte nucleus and its gene expression; it has no fine-tuned control of the cryptophytes gene expression which otherwise responds to changes in time and light condition in free-swimming cryptophytes.

The contribution of cryptophyte genes to the global transcriptome of well-fed M. rubrum is with 67.8% substantial and higher than previously estimated (13.5% in23, 58-62% in24). One can speculate that many of those transcripts are by-products and not actually needed by M. rubrum. In a well-integrated endosymbiotic system, one would expect to find the expression of plastid targeted genes only. We interpret this broad expression of cryptophyte genes within M. rubrum as an indication that M. rubrum has no fine-tuned control over the acquired nucleus, but can only force a coarse range of transcripts out of the acquired nucleus due to limited transcriptional regulation.

Photosynthesis

Apart from M. rubrum, plastid sequestration is known from other ciliates28, dinoflagellates29,30, sacoglossan sea slugs31-33, and marine flatworms34. Until recently, M. rubrum was the only ciliate known to be capable of functional phototrophy and chloroplast division28. Recently, a large red Mesodinium species, M. major, was described to also contain cryptophyte chloroplasts, mitochondria, and prey nuclei.35 At present, however, eight clades (A-H) have been identified within the M. rubrum/M. major species complex36. Besides the M. rubrum/M. major species complex, also M. coatsi and M. chamaeleon contain both chloroplasts and prey nuclei35,37, but in these cases the ingested cryptophytes are kept as individual

12 packages that do not have the centered prey nucleus as in in the M. rubrum/M.major species complex.

Functional chloroplasts of higher plants and algae use approximately 2000 proteins, but their genomes contain only 60-250 protein-coding genes38-40. The remaining genes are encoded by the host nucleus41. The chloroplast-encoded proteins in T. amphioxeia are necessary for photosynthesis and contribute among others to photosystems, ATP synthesis, and rubisco1. Other plastid-encoded proteins perform essential plastid functions not directly linked to photosynthesis such as cofactor biosynthesis, protein import, and plastid gene expression40. It is estimated that a chloroplast needs more than thousand proteins, in order to remain active and functional42-44.

We found no photosynthesis related genes originating from M. rubrum. Hence, in order to keep the cryptophyte chloroplasts functional and to provide them with essential proteins, M. rubrum is dependent on the cryptophyte nucleus. Nuclear- encoded photosynthesis related genes are strongly upregulated upon sequestration. The real challenge is to transport gene products that are vital for chloroplast function and survival from the host to the chloroplast42. Photosynthesis is a harsh chemical endeavor that involves many reactive radicals, especially the protein components of the photosystems in active algal plastids, which are very short-lived and need constant replacement in bright light45-47. The biliproteins are embedded in the thylakoid lumen, the thylakoids of cryptophyte chloroplasts are generally thicker compared with chloroplasts of other photosynthetic eukaryotes48. T. amphioxeia have thylakoids arranged in loose groups of three49.

We are not aware of any other system that is comparable to M. rubrum, where a similar loss in the ability to regulate genes according to time and light condition was observed after sequestration. This system might be a general example of how endosymbionts are integrated into a new host, and how the transition from engulfment of a whole cell, over a transitionary endosymbiont, towards a permanent endosymbiont takes place. In a first step a cell is taken up by a host and not digested. The second step is that the host gets some control over the gene expression of the acquired cell – that is where M. rubrum is right now. In a third step, the genes from the host and acquired cell need to align in order to fine-tuning the

13 gene expression according to environmental conditions. Whether or not M. rubrum is on a way towards a permanent tertiary endosymbiosis is speculative.

It is also not clear whether M. rubrum can divide chloroplasts without the presence of cryptophyte nuclei or not. Experimental data showed that M. rubrum is not able to photosynthesize without prey nuclei present16. However, even starved M. rubrum that have lost the prey nucleus, will usually have some chloroplasts remaining in the cell. An answer to the survival of the chloroplasts within M. rubrum might be in them being robust. Chloroplast genomes from diverse groups of algae and plants share a similar gene set50. Red algal chloroplast genomes harbor 1.5 to 2 times as many genes as green algae and land plants51. The red algal derived plastid genome of T. amphioxeia contains 143 genes1. That is less than a red algae plastid (183-233 plastid encoded genes) but more than most Viridiplantae with very few to up to 174 plastid encoded genes51. In addition, the T. amphioxeia chloroplast has its nucleomorph, which in the cryptophyte Guillardia theta encodes 487 genes6. It might be true that the T. amphioxeia chloroplast is a favored endosymbiont when compared to chloroplasts with smaller gene sets52.

Cryptophyte nuclei

Many questions about the acquired cryptophyte nuclei remain unanswered. In typical well-fed M. rubrum cells, the cryptophyte nucleus is enlarged and located at more less the same position anterior to the two macronuclei within M. rubrum16. It has been proposed that several nuclei fuse in order to form the enlarged nucleus23. It is however unclear how this should be organized physiologically. A more likely scenario is that M. rubrum is not able to pack the foreign chromosomes tightly. This would explain why the prey nucleus gets larger and larger over time until it eventually disappears. It has also been proposed that M. rubrum keeps extra prey nuclei for later use16.

Metabolism

As the sequestered organelles are no longer part of an intact cryptophyte cell, a reduction in gene expression was expected for many pathways that are not related to the chloroplast24. This was confirmed in our study. In order for acquired phototrophy to work, the host needs to ensure that it does not digest the acquired plastid. M. rubrum partially bypasses the typical heterotrophic digestion to allow interaction with

14 foreign organelles28. We found an upregulation of metabolic pathways in the T. amphioxeia genes after sequestration.

The genetic code

Ciliates show deviations in the genetic code and it has been suggested that these deviations have occurred multiple times independently53. M. rubrum uses a genetic code that is different from cryptophytes and other eukaryotes as it translates UAA and UAG into Tyrosine and not into STOP codons54. In addition to this, M. rubrum does not have the genomic information necessary to transcribe the genes and translate the proteins necessary to serve the plastid. The solution to this problem is to retain the nucleus from its cryptophyte prey in addition to the plastids16,23. M. rubrum thus uses the cryptophyte nucleus as a dictionary to serve the chloroplast gene products using its own genetic code (the standard code). Given this is possible for several cryptophytes such as Teleaulax amphioxeia, T. acuta and Geminigera cryophila, the question remains as to why only these taxa are preyed upon and no other cryptophytes can be exploited15. M. rubrum does feed on other cryptophyte species, but for unknown reasons cannot utilize them for growth and photosynthesis15,55.

Outlook

Our study provides new insights into the interplay between M. rubrum and one of its cryptophyte prey. Many details, however, such as the actual communication between the different organelles, remain essentially unknown. Many scientific steps forward are required to fully understand such complex systems in which DNA of different origin are functioning within a single cell. The field is moving forward quickly, for example, a re-evaluation of nucleomorph encoded proteins recently found 215 new annotations out of 826 uncharacterized open reading frames of cryptophytes56, and it was possible to identify tetrapyrroles as mediators of a chloroplast-to-nucleus signaling pathway57.

Materials and Methods

Cultures

Cultures were established from single-cell isolates of Teleaulax acuta (SCCAP K-1486, collected in Nivå Bay, Denmark), Teleaulax amphioxeia (SCCAP K-1837, collected in Elsinore Harbor, Denmark), and Mesodinium rubrum (MBL-DK2009 collected in September 2009 in Elsinore Harbor, Denmark). Cultures (T. acuta, T. amphioxeia, M. rubrum fed T. acuta, M. rubrum fed T. amphioxeia) were kept in triplicates and grown in F/2 medium at 15 °C in a light/dark cycle of 16/8h with a light intensity of 100 µmol photons m−2s−1. During the exponential phase of growth, the ciliates were transferred to new media when cell concentrations reached 5000 ml-1 or more.

RNA extraction

For RNA extraction, cultures were harvested in full light (7 hours into the light cycle), in darkness (6 hours into dark cycle) and in the transition between dark and light (20 minutes into the light cycle). Cells of M. rubrum were harvested in a well-fed and a starved stage.

For the well-fed stage we checked before extraction that no free cryptophyte cells remained in the medium and that at least 90% of all M. rubrum cells contained a cryptophyte nucleus. This was done by staining the nuclei with Hoechst reagent (#33342, Thermo Fisher Scientific, Waltham, USA), and checking 20 stained cells under a fluorescent microscope. Harvesting of starved cells was done approximately four weeks after the last cryptophytes had been seen in the culture. We confirmed the loss of cryptophyte nuclei by staining with Hoechst reagent and checking for symbiont nuclei under a fluorescence microscope. Cells were harvested after at least 90% of all M. rubrum cells had lost their cryptophyte nucleus. Cells were harvested by centrifugation in 10ml glass tubes at 16.100 g for 10 minutes (see supplementary material S1 for cell numbers in each harvest). Pellets were transferred to 1.5 mL LoBind Eppendorf tubes and liquid nitrogen was directly added onto the pellets. The Eppendorf tubes were stored on ice without allowing the pellets to thaw until the lysis buffer was added. RNA was extracted using the column based Exiqon Cell and Plant RNA Isolation Kit (Exiqon, Vedbæk, Denmark, cat# 300110) following the ‘plant’ protocol. In addition, a separate round of harvest has been transferred to hot Trizol and stored at -80 °C as backup. Two samples (10 and 11) from this backup have been used for RNA extraction using the Trizol method. Extracted RNA was stored at -80 ˚C until library preparation for sequencing.

Sequencing

RNA was sequenced using the BGISeq 500 and Illumina Hiseq 4000 platforms following the respective protocols. The raw sequence reads were quality controlled with FastQC 58. Low 16 quality reads were removed and trimmed with SOAPnuke 1.5.3. De novo assembly was done with Trinity 2.4.059,60. Transcriptome assembly was quality checked using transrate v.1.0.361, and BUSCO62.

Acknowledgements

The Danish Council for Independent Research funded this project (grant number 4181- 00484).

Author contributions

PJH, NL, UJ and AA conceived the work. AA, KD, MK, LG did the culturing. AA and LG extracted RNA; HC and QL did the sequencing, HC, AA and QL analyzed the data, AA and HC drafted the manuscript. All authors contributed to writing of the final version of the manuscript.

Conflict of Interest

The authors declare no conflict of interest.

References

1 Kim, J. I. et al. The Plastid Genome of the Cryptomonad Teleaulax amphioxeia. PLoS One 10, e0129284 (2015). 2 Burki, F. et al. Untangling the early diversification of eukaryotes: a phylogenomic study of the evolutionary origins of Centrohelida, Haptophyta and Cryptista. Proc. R. Soc. B 283, (2016). 3 Cavalier-Smith, T. Membrane heredity and early chloroplast evolution. Trends Plant Sci. 5, 174-182 (2000). 4 Douglas, S. E., Murphy, C. A., Spencer, D. F. & Gray, M. W. Cryptomonad algae are evolutionary chimaeras of two phylogenetically distinct unicellular eukaryotes. Nature 350, 148-151 (1991). 5 Cavalier-Smith, T. Principles of protein and lipid targeting in secondary symbiogenesis: euglenoid, dinoflagellate, and sporozoan plastid origins and the eukaryote family tree. J. Eukaryot. Microbiol. 46, 347-366 (1999). 6 Curtis, B. A. et al. Algal genomes reveal evolutionary mosaicism and the fate of nucleomorphs. Nature 492, 59-65 (2012). 7 Douglas, S. et al. The highly reduced genome of an enslaved algal nucleus. Nature 410, 1091- 1096 (2001). 8 Hoef-Emden, K. & Archibald, J. M. in Handbook of the Protists (eds John M. Archibald, Alastair G. B. Simpson, & Claudio H. Slamovits) Ch. 24, 851-891 (Springer International Publishing, 2017). 17

9 Ward, B. A. & Follows, M. J. Marine mixotrophy increases trophic transfer efficiency, mean organism size, and vertical carbon flux. Proc. Natl. Acad. Sci. U.S.A. 113, 2958-2963 (2016). 10 Herfort, L., Peterson, T. D., Campbell, V., Futrell, S. & Zuber, P. Myrionecta rubra (Mesodinium rubrum) bloom initiation in the Columbia River estuary. Estuar. Coast. Shelf Sci. 95, 440-446 (2011). 11 Johnson, M. D. et al. The Genetic Diversity of Mesodinium and Associated Cryptophytes. Front. Microbiol. 7, 2017 (2016). 12 Lindholm, T., Lindroos, P. & Mork, A. C. Ultrastructure of the photosynthetic ciliate Mesodinium rubrum. BioSyst. 21, 141-149 (1988). 13 Nowack, E. C. & Melkonian, M. Endosymbiotic associations within protists. Phil. Trans. R. Soc. B 365, 699-712 (2010). 14 Johnson, M. D., Oldach, D., Delwiche, C. F. & Stoecker, D. K. Retention of transcriptionally active cryptophyte nuclei by the ciliate Myrionecta rubra. Nature 445, 426-428 (2007). 15 Hansen, P. J., Moldrup, M., Tarangkoon, W., Garcia-Cuetos, L. & Moestrup, O. Direct evidence for symbiont sequestration in the marine red tide ciliate Mesodinium rubrum. Aquat. Microb. Ecol. 66, 63-75 (2012). 16 Kim, M., Drumm, K., Daugbjerg, N. & Hansen, P. J. Dynamics of Sequestered Cryptophyte Nuclei in Mesodinium rubrum during Starvation and Refeeding. Front. Microbiol. 8, 1-14 (2017). 17 Nam, S. W., Park, J. W., Yih, W., Park, M. G. & Shin, W. The fate of cryptophyte cell organelles in the ciliate Mesodinium cf. rubrum subjected to starvation. Harmful Algae 59, 19-30 (2016). 18 Juel Hansen, P. & Fenchel, T. The bloom-forming ciliate Mesodinium rubrum harbours a single permanent endosymbiont. Mar. Biol. Res. 2, 169-177 (2006). 19 Smith, M. & Hansen, P. J. Interaction between Mesodinium rubrum and its prey: importance of prey concentration, irradiance and pH. Mar. Ecol. Prog. Ser. 338, 61-70 (2007). 20 Matthew, D. J. & Diane, K. S. Role of feeding in growth and photophysiology of Myrionecta rubra. Aquat. Microb. Ecol. 39, 303-312 (2005). 21 Fenchel, T. & Hansen, P. J. Motile behaviour of the bloom-forming ciliate Mesodinium rubrum. Mar. Biol. Res. 2, 33-40 (2006). 22 Gustafson, D. E., Stoecker, D. K., Johnson, M. D., Van Heukelem, W. F. & Sneider, K. Cryptophyte algae are robbed of their organelles by the marine ciliate Mesodinium rubrum. Nature 405, 1049-1052 (2000). 23 Kim, G. H. et al. Cryptophyte gene regulation in the kleptoplastidic, karyokleptic ciliate Mesodinium rubrum. Harmful Algae 52, 23-33 (2016). 24 Lasek-Nesselquist, E., Wisecaver, J. H., Hackett, J. D. & Johnson, M. D. Insights into transcriptional changes that accompany organelle sequestration from the stolen nucleus of Mesodinium rubrum. BMC Genomics 16, 805 (2015). 25 Qiu, D., Huang, L. & Lin, S. Cryptophyte farming by symbiotic ciliate host detected in situ. Proc. Natl. Acad. Sci. U.S.A. 113, 12208-12213 (2016). 26 Johnson, M. D. et al. Mesodinium rubrum: The symbiosis that wasn't. Proc. Natl. Acad. Sci. U.S.A. 114, E1040-E1042 (2017). 27 Kawach, O. et al. Unique tRNA introns of an enslaved algal cell. Mol. Biol. Evol. 22, 1694-1701 (2005). 28 Johnson, M. D. Acquired phototrophy in ciliates: a review of cellular interactions and structural adaptations. J. Eukaryot. Microbiol. 58, 185-195 (2011). 29 Schnepf, E. & Elbrachter, M. Cryptophycean-Like Double Membrane-Bound Chloroplast in the Dinoflagellate, Dinophysis Ehrenb.: Evolutionary, Phylogenetic and Toxicological Implications. Bot. Acta 101, 196-203 (1988). 30 Takishita, K., Koike, K., Maruyama, T. & Ogata, T. Molecular evidence for plastid robbery (Kleptoplastidy) in Dinophysis, a dinoflagellate causing diarrhetic shellfish poisoning. Protist 153, 293-302 (2002).

31 Bhattacharya, D., Pelletreau, K. N., Price, D. C., Sarver, K. E. & Rumpho, M. E. Genome analysis of Elysia chlorotica Egg DNA provides no evidence for horizontal gene transfer into the germ line of this Kleptoplastic Mollusc. Mol. Biol. Evol. 30, 1843-1852 (2013). 32 de Vries, J. et al. Is ftsH the key to plastid longevity in sacoglossan slugs? Genome Biol. Evol. 5, 2540-2548 (2013). 33 Wagele, H. et al. Transcriptomic evidence that longevity of acquired plastids in the photosynthetic slugs Elysia timida and Plakobranchus ocellatus does not entail lateral transfer of algal nuclear genes. Mol. Biol. Evol. 28, 699-706 (2011). 34 Van Steenkiste, N. W. L. et al. A new case of kleptoplasty in animals: Marine flatworms steal functional plastids from diatoms. Sci. Adv. 5, eaaw4337 (2019). 35 Garcia-Cuetos, L., Moestrup, O. & Hansen, P. J. Studies on the genus Mesodinium II. Ultrastructural and molecular investigations of five marine species help clarifying the taxonomy. J. Eukaryot. Microbiol. 59, 374-400 (2012). 36 Kim, M. & Park, M. G. Unveiling the hidden genetic diversity and chloroplast type of marine benthic ciliate Mesodinium species. Sci. Rep. 9, 14081 (2019). 37 Nam, S. W., Shin, W., Kang, M., Yih, W. & Park, M. G. Ultrastructure and Molecular Phylogeny of Mesodinium coatsi sp. nov., a Benthic Marine Ciliate. J. Eukaryot. Microbiol. 62, 102-120 (2015). 38 Timmis, J. N., Ayliffe, M. A., Huang, C. Y. & Martin, W. Endosymbiotic gene transfer: organelle genomes forge eukaryotic chromosomes. Nat. Rev. Genet. 5, 123-135 (2004). 39 Turmel, M., Otis, C. & Lemieux, C. The complete chloroplast DNA sequence of the green alga Nephroselmis olivacea: insights into the architecture of ancestral chloroplast genomes. Proc. Natl. Acad. Sci. U.S.A. 96, 10248-10253 (1999). 40 Dorrell, R. G. & Howe, C. J. Integration of plastids with their hosts: Lessons learned from dinoflagellates. Proc. Natl. Acad. Sci. U.S.A. 112, 10247-10254 (2015). 41 de Vries, J. & Archibald, J. M. in Adv. Bot. Res. Vol. 85 (eds Shu-Miaw Chaw & Robert K. Jansen) 1-28 (Academic Press, 2018). 42 Archibald, J. M. The puzzle of plastid evolution. Curr. Biol. 19, R81-88 (2009). 43 Gould, S. B., Waller, R. F. & McFadden, G. I. Plastid evolution. Annu. Rev. Plant Biol. 59, 491- 517 (2008). 44 Kleine, T., Maier, U. G. & Leister, D. DNA transfer from organelles to the nucleus: the idiosyncratic genetics of endosymbiosis. Annu. Rev. Plant Biol. 60, 115-138 (2009). 45 Faleev, N. G. et al. Tyrosine phenol-lyase from Citrobacter intermedius. Factors controlling substrate specificity. Eur. J. Biochem. 177, 395-401 (1988). 46 Vass, I. et al. Reversible and irreversible intermediates during photoinhibition of photosystem II: stable reduced QA species promote chlorophyll triplet formation. Proc. Natl. Acad. Sci. U.S.A. 89, 1408-1412 (1992). 47 Warner, M. E., Fitt, W. K. & Schmidt, G. W. Damage to photosystem II in symbiotic dinoflagellates: a determinant of coral bleaching. Proc. Natl. Acad. Sci. U.S.A. 96, 8007-8012 (1999). 48 Tanifuji, G. & Onodera, N. T. in Adv. Bot. Res. Vol. 84 (ed Yoshihisa Hirakawa) Ch. 8, 263-320 (Academic Press, 2017). 49 Deane, J. A., Strachan, I. M., Saunders, G. W., Hill, D. R. A. & McFadden, G. I. Cryptomonad evolution: Nuclear 18S rDNA phylogeny versus cell morphology and pigmentation. J. Phycol. 38, 1236-1244 (2002). 50 Allen, J. F., de Paula, W. B. M., Puthiyaveetil, S. & Nield, J. A structural phylogenetic map for chloroplast photosynthesis. Trends Plant Sci. 16, 645-655 (2011). 51 Qiu, H., Lee Jun, M., Yoon Hwan, S. & Bhattacharya, D. Hypothesis: Gene-rich plastid genomes in red algae may be an outcome of nuclear genome reduction. J. Phycol. 53, 715- 719 (2017). 52 Grzebyk, D., Schofield, O., Vetriani, C. & Falkowski, P. G. The Mesozoic Radiation of Eukaryotic Algae: The Portable Plastid Hypothesis. J. Phycol. 39, 259-267 (2003).

53 Tourancheau, A. B., Tsao, N., Klobutcher, L. A., Pearlman, R. E. & Adoutte, A. Genetic code deviations in the ciliates: evidence for multiple and independent events. EMBO J. 14, 3262- 3267 (1995). 54 Heaphy, S. M., Mariotti, M., Gladyshev, V. N., Atkins, J. F. & Baranov, P. V. Novel Ciliate Genetic Code Variants Including the Reassignment of All Three Stop Codons to Sense Codons in Condylostoma magnum. Mol. Biol. Evol. 33, 2885-2889 (2016). 55 Johnson, M. D., Beaudoin, D. J., Frada, M. J., Brownlee, E. F. & Stoecker, D. K. High grazing rates on cryptophyte algae in Chesapeake Bay. Front. Mar. Sci. 5, 1-13 (2018). 56 Zauner, S., Heimerl, T., Moog, D. & Maier, U. G. The Known, the New, and a Possible Surprise: A Re-Evaluation of the Nucleomorph-Encoded Proteome of Cryptophytes. Genome Biol. Evol. 11, 1618-1629 (2019). 57 Shimizu, T. et al. The retrograde signaling protein GUN1 regulates tetrapyrrole biosynthesis. Proc. Natl. Acad. Sci. U.S.A. 116, 24900-24906 (2019). 58 Andrews, S. FastQC: a quality control tool for high throughput sequence data, (2010). 59 Haas, B. J. et al. De novo transcript sequence reconstruction from RNA-seq using the Trinity platform for reference generation and analysis. Nat. Protoc. 8, 1494-1512 (2013). 60 Brocks, J. J. et al. The rise of algae in Cryogenian oceans and the emergence of animals. Nature 548, 578-581 (2017). 61 Smith-Unna, R., Boursnell, C., Patro, R., Hibberd, J. M. & Kelly, S. TransRate: reference-free quality assessment of de novo transcriptome assemblies. Genome Res. 26, 1134-1144 (2016). 62 Simao, F. A., Waterhouse, R. M., Ioannidis, P., Kriventseva, E. V. & Zdobnov, E. M. BUSCO: assessing genome assembly and annotation completeness with single-copy orthologs. Bioinformatics 31, 3210-3212 (2015).

App. II

Mixotrophy Among Freshwater and Marine Protists Per J Hansen and Ruth Anderson, University of Copenhagen, Helsingør, Denmark Diane K Stoecker, University of Maryland Center for Environmental Science, Cambridge, MD, United States Johan Decelle, University of Grenoble Alpes, CNRS, CEA, INRA, Grenoble, France Andreas Altenburger, University of Copenhagen, Copenhagen, Denmark Hannah E Blossom and Kirstine Drumm, University of Copenhagen, Helsingør, Denmark Aditee Mitra and Kevin J Flynn, Swansea University, Swansea, United Kingdom

Glossary bi-flagellated and possess a dinokaryon, a special nucleus Autotroph An organism capable of synthesizing its own characterized by the presence of a permanent nuclear food from inorganic substances. envelope and chromosomes in a permanently compacted Bacterivory The ingestion of bacteria as food or as an state. Half of the species lack built-in chloroplasts. energy supply. Immature successional stages A term used here to Chlorophytes A group of microalgae that are coccoid or describe the colonization of pelagic waters by a few fast bi-flagellated, solitary or in colonies. They are green and growing species. Immature successional stages are typically possess chlorophylls a and b and store starch in the found during the spring bloom in temperate and polar chloroplast. Very common in freshwater, but can also be regions, and upwelling areas where nutrient rich water is found in marine waters. brought up from deep waters into the light. Chloroplast Membrane-bound structure (¼plastid) Kleptoplastidy Nutritional strategy in which phototrophy containing photosynthetic pigment found within a cell. is aquired by sequestration and subsequent utilization of Chrysophytes A group of protists with two flagella (one plastids and other cell organelles from prey organisms. short and one long) that belong to the Heterokonta. They Mature successional stages A term used here to describe have chloroplasts with chlorophylls a and c. They are the successional stages dominated by many slower-growing yellowish due to carotenoids. They are common in both species. Mature successional stages are typically found marine and freshwaters. during summer periods in temperate and polar regions, and Ciliates A group of protists belonging to the Aveolata. year round in tropical open marine waters. Concentrations They are characterized by the presence of hair-like organelles of inorganic nutrients are low during these stages. called cilia, which are identical in structure to eukaryotic Mixotroph As used here, a planktonic organism that mixes flagella, but are in general shorter and present in much photosynthesis and food uptake via phagotrophy; larger numbers. sometimes referred to as a photo-phago-mixotroph. Constitutive mixotrophs (CMs) Protists with built-in Non-constitutive mixotrophs (NCMs) Mixotrophic chloroplasts that are capable of phagotrophy. protists which lack built-in chloroplasts. Some NCMs utilize Cryptophytes A group of protists in which most of the ecto- or endosymbionts, while others sequester chloroplasts species contain chloroplasts with chlorophylls a and c, and sometimes other cell organelles from their prey. Species together with phycobili-proteins and other pigments; they in the latter group are often referred to as klepto(chloro) vary in color (brown, red to blueish-green). Cells contain an plastidic. extra reduced nucleus called a nucleomorph, which is Pelagic Describing or referring to organisms that are living derived evolutionarily from a previous endosymbiosis in the water column in fresh and marine waters, either occasion. swimming or passively suspended; opposite to benthic Dinoflagellates A group of protists belonging to the (living on the bottom). Alveolata, very common in marine and freshwaters. They are Phagotrophy Ingestion of particulate matter.

Introduction

A wide diversity of marine and freshwater unicellular eukaryotes (protists) are capable of acquiring energy, carbon and nutrients through both photosynthesis and phagotrophy of prey. Among these so-called photo-phago-mixotrophs, two major groups can be distinguished: (1) Protists with permanent, built-in chloroplasts, termed ‘Constitutive mixotrophs (CMs)’ (Fig. 1). These are often referred to as phytoplankton that can eat microbial prey (e.g., bacteria or other protists). (2) Protists that lack permanent chloroplasts (and often regarded as microzooplankton), but use photosynthetic machinery derived from their algal prey. These are termed, ‘Non-constitutive mixotrophs (NCMs)’. Some NCMs keep the engulfed algal cells intact as ecto- or endosymbionts (Fig. 2), while other NCMs, often termed klepto(chloro)plastidic species, retain only the chloroplasts or the chloroplasts along with prey nuclei, mitochondria, and cytoplasm (Fig. 3). Some NCMs are prey specialists and can only exploit very speciﬁc types of prey (Specialist NCM ¼ SNCMs), while others can exploit a wide range of prey (Generalist ¼ GNCMs).

Encyclopedia of Microbiology, 4e doi:10.1016/B978-0-12-809633-8.20685-7 1 2 Mixotrophy Among Freshwater and Marine Protists

Fig. 1 A panel of light and epifluorescent microscopy images showing different marine CMs. (a) Alexandrium sp.(ca.45mm in width) found in Greenland with a large food vacuole (fv), and a reduced chloroplast (chl). Photo by Hannah Blossom. (b) A. pseudogonyaulax (ca. 40 mm in width) found in Denmark with many food vacuoles (arrow). Photo by Hannah Blossom. (c) Epifluorescent microscopy image of A. pseudogonyaulax (red), which had eaten Teleaulax acuta (visible as yellow food vacuoles). The food vacuoles (yellow) are easy to distinguish against the red autofluorescence of dinoflagellates. Photo by Hannah Blossom. (d) A. andersonii (ca. 20 in width) fed Teleaulax sp. which are visible as food vacuoles (arrows). Photo by Kyong Ha Lee, courtesy of Hae Jin Jeong. (e) Tripos menueri (¼ Ceratium tripos)(ca.60mm wide and 200–250 mm long). Photo by Per Juel Hansen. (f) Tripos menueri found in Denmark with two food vacuoles (yellow). Photo by Hannah Blossom. (g) Karlodinium armiger (Ka, ca. 18 mm in length) swarming around and feeding on Prorocentrum minimum (Pm). Modified from Berge, T., Hansen, P.J., Moestrup, Ø., 2008. Feeding mechanism, prey specificityandgrowthinlightanddark of the plastidic dinoflagellate Karlodinium armiger. Aquatic Microbial Ecology 50, 279–288, with permission from Inter-Research. (h) Karenia mikimotoi (20 mm in width). Photo by Gert Hansen. (i) Lingulodinium sp. found in the field in Greenland with a food vacuole (fv). Photo by Hannah Blossom. (j) The haptophyte Prymnesium parvum (ca. 10 mm long). Photo by Gert Hansen. (k) Prorocentrum minimum (15 mm in length). Photo by Per Juel Hansen. (l) Confocal laser scanning micrograph image of Prorocentrum minimum (15–20 mminlength)fedwithTeleaulax amphioxei. Four food vacuoles are visible, one of which is indicated with an arrow. Modified from Johnson, M.D., 2014. Inducible mixotrophy in the dinoflagellate Prorocentrum minimum. Journal of Eukaryotic Microbiology 62, 431–443, with permission from John Wiley and Sons. (m) Five Fragilidium subglobosum (Fs, 40 mm) engulfing Tripos menueri (Ct). Adapted from Skovgaard, A., 1996. Engulfment of Tripos spp. (Dinophyceae) by the thecate photosynthetic dinoflagellate Fragilidium subglobosum. Phycologia 35, 490–499. All these examples are dinoflagellates except Prymnesium parvum (j).

How Common are Photo-Phago-Mixotrophic Protists in Different Aquatic Environments Compared to Photoautotrophs and Heterotrophs? Photo-phago-mixotrophic protists are common, and often very abundant, in both freshwater and marine environments ranging from eutrophic to highly oligotrophic, and from tropical to polar. However, the relative contributions of CMs and different NCM groups to standing protist biomass and production across different scales of time and space are mostly unknown. This knowledge gap is mainly explained by the difﬁculty in identifying which organisms are actively mixotrophic in the ﬁeld. Below we summarize current knowledge.

Constitutive Mixotrophs (CMs)

In principle, all eukaryotic microalgal lineages (chlorophytes, euglenophytes, cryptophytes, chrysophytes, haptophytes and dinoﬂagellates) have the capacity to take up prey, though it is not always present in all species of each lineage. Only diatoms are thought to lack phagotrophic capabilities as a group; but like all microalgae, they are able to utilize dissolved organic nutrients. Mixotrophy Among Freshwater and Marine Protists 3

Fig. 2 A panel of light microscopy images showing different marine planktonic protists (NCMs) in symbiosis with microalgae, a common relationship in oceanic waters worldwide. (a,b,c) The rhizarian Radiolaria hosts symbiotic microalgal cells (yellow cells), which can be haptophytes or dinoflagellates. Radiolarians can build mineral skeletons of about 200–300 mm in size. (a) polycystines, (b) acantharians, (c) collodarians. Photos by Johan Decelle. (d) The green dinoflagellate Noctiluca scintillans (500–1000 mm in diameter), which can proliferate in Southeast Asia. Photo by Per Juel Hansen. (e) The rhizarian Foraminifera (Globigerinoides sacculifer) living with dinoflagellates (200 mm in diameter). Photo by Katsunori Kimoto. (f) The ciliate Tiarina sp. that builds a shell in calcium carbonate (80 mm in length) and hosts dinoflagellate symbionts (Symbiodinium, also found in corals) in oceanic waters. Photo by Johan Decelle. (g) The dinoflagellate Amphisolenia sp. that hosts pelagophyte endosymbionts along its 500 mm-long cell. Photo by John Dolan.

The major phytoplankton groups have a global distribution with smaller species present in greater proportion in oligotrophic waters, while larger species, such as dinoflagellates, will gain relative abundance in coastal waters. We can view the dominance of diatoms (autotrophs) vs CMs in an environment in the context of the supply and demand of nutrients and energy that can be fulfilled by phagotrophy. The balance between photoautotrophy and photo-phago-mixotrophy in plankton communities can be considered in significant part as determined by a combination of supply and demand of inorganic (dissolved inorganic nitrogen (DIN) and dissolved inorganic phosphorus (DIP)) and organic nutrients (including prey availability). A schematic representation of the expected seasonal progression of non-mixotrophic plankton to one dominated by mixotrophs is shown in Fig. 4. There are two significant challenges affecting our interpretation of the role of CMs in plankton ecology. The first is the quality of plankton survey data, which usually group diverse organisms into different size classes and only identify larger cells (420 mm) with easily recognizable morphology. Small cells are not usually sampled or identified effectively. In consequence, we know very little even in quasi qualitative terms about the global distribution of organisms that we expect to be capable of photo-phago- mixotrophy. Secondly, we do not know the extent to which these phototrophic organisms are engaging in significant levels of mixotrophy. Experimental approaches on cultured representatives are required to actually measure rates of primary production and phagotrophy, but many planktonic organisms are not in culture. Coupled to this challenge is that a degree of feeding which may not appear significant in a community ecology context, could be essential for growth if it supplies a limiting nutrient.

Non-Constitutive Mixotrophs (NCMs)

These can be found mainly among ciliates, heterotrophic dinoﬂagellates, foraminiferans and radiolarians. The NCMs are often relatively large cells (20–1000 mm), and some can build carbonate or silicate mineral skeletons, such as the foraminiferans and 4 Mixotrophy Among Freshwater and Marine Protists

Fig. 3 A panel of light microscopy images showing different NCMs in marine ecosystems, which sequester the chloroplasts of their prey (kleptoplastidy). (a) The toxic dinoflagellate Dinophysis acuta (60 μm in length), which harbors chloroplasts of cryptophyte origin. Photo by Per Juel Hansen. (b) The ciliate Mesodinium major (50 μm in diameter), one of the red Mesodinium species. Photo by Niels Daugbjerg. (c) The oligotrich ciliate Strombidium rassoulzadegani (60 μm in length) with sequestered chloroplasts from different algal prey (green algae and cryptophytes). Modified from Schoener, D.M., McManus, G.B., 2012, Plastid retention, use, and replacement in a kleptoplastidic ciliate. Aquatic Microbial Ecology 67, 177–187, with permission from Inter-Research. radiolarians. Despite a high dependency on photosynthesis, some NCMs need to maintain heterotrophic activity. For others, the type of prey (specificity) and the contribution of predation to the overall metabolism is still poorly understood. Different types of NCMs inhabit different biomes (Fig. 5), though collectively they are present across the global seas. Most inhabit mature ecosystems primarily due to the distribution of their prey, which can be very specific. Examples are large symbiotic NCMs like radiolarians and foraminiferans, which are the most abundant NCMs in the surface waters of the oligotrophic open ocean. However, some NCMs that maintain symbionts or sequester prey nuclei along with the chloroplasts can inhabit waters during less mature stages. Examples of the latter are the ciliate Mesodinium rubrum, which is heavily influenced by upwelling, and the green dinoflagellate Noctiluca, which responds to eutrophication.

Feeding Mechanism, Prey Selection and Toxin Assisted Feeding

Prey Capture and Feeding Mechanisms Suspended and motile prey can be difficult for a single-celled organism to capture and consume, but many mixotrophic protists have come up with truly spectacular ways to achieve this. Many swim and catch prey upon collision, with the prey engulfed directly into a food vacuole. Others, typically slower swimming species, have evolved various sophisticated ways of capturing prey, using a haptonema (many haptophytes), a capture filament (many dinoflagellates) or a mucus trap (some dinoflagellates). Prey are usually taken up whole by direct engulfment, but some species use a feeding tube (a type of straw) to suck out the contents, thereby enabling these mixotrophs to exploit larger prey items (Fig. 6).

Toxin-Assisted Feeding Some mixotrophic protists use toxins to immobilize swimming prey. Examples of these include the haptophyte Prymnesium parvum, which releases toxins to the water to immobilize swimming prey that it would otherwise not be able to catch. Excreted toxins from the dinoﬂagellate Karlodinium armiger have been shown to be involved in capturing prey organisms that would normally be their predators, thereby reversing the food web. Excreted toxins may also be used in combination with a mucus trap, as described in the dinoﬂagellate Alexandrium pseudogonyaulax (Fig. 6), to keep effective toxin concentrations high and in close proximity. Toxin-Assisted prey capture could be more common than we realize; it may not be coincidental that most toxin producing harmful algal bloom forming species are mixotrophic.

Prey Selection Most mixotrophic protists are thought to feed on a wide range of prey and the diet usually includes bacteria and/or other protists, but it may in some cases also include copepods and mollusc larvae. In contrast, some protists are quite selective and rely on specific species of prey. For example, the ciliate Mesodinium rubrum utilizes a specific clade of cryptophytes, which are incorporated into the cell body of M. rubrum allowing the ciliate to rely on photosynthesis as its major carbon source. In turn, M. rubrum is specifically consumed by Dinophysis spp. (dinoflagellate) in order to acquire the cryptophyte chloroplasts from the ciliate. Mixotrophy Among Freshwater and Marine Protists 5

Fig. 4 Schematic diagram of the changes to the planktonic food web over a year, with transitions between immature and mature ecosystem states. The upper panels show changing patterns of light, inorganic nutrients and particle density (i.e., total plankton biomass) over the temperate year. Transitions from mature to immature (spring or autumn “blooms”) to mature again, are as indicated; green dashed line indicting conditions optimal for phototrophy, orange dashed lines for phagotrophy. Other periods are sub-optimal for strict phototrophs and/or strict phagotrophs, and preferable for mixotrophs. The lower panel shows in detail the transition from immature to mature, with changes in selection priorities from so- called “r-select” phototrophs and phagotrophs in immature ecosystems, to a mature ecosystem more optimal for “K-select” mixotrophs. Reprinted from Mitra et al., 2014. The role of mixotrophic protists in the biological carbon pump. Biogeosciences 11, 995–1005, with permission from Copernicus Publications.

Abiotic and Biotic Factors Inﬂuencing Photosynthesis, Growth and Feeding Rates in CM and NCMs

Light and Prey Concentration CMs Most studied CMs require light for sustained growth and will not grow in complete darkness even if fed. Irradiance is required to at least ﬁx enough carbon to match respiration needs. At higher irradiances protist growth increases until saturation, as it does with pure phototrophs such as diatoms and cyanobacteria. If prey are abundant, some CMs can grow faster mixotrophically than phototrophically at low irradiance (e.g., Karlodinium veneﬁcum) and ingestion rates increase as a function of irradiance until it also satiates (Fig. 7). In consequence, mixotrophic growth rates of CMs are often much higher than when grown purely 6 Mixotrophy Among Freshwater and Marine Protists

Fig. 5 Global distribution of protists with acquired phototrophy (non-constitutive mixotrophs, NCMs). Functional groups identify protists which acquire plastids from a variety of prey (generalist NCMs, GNCMs; blue (a)), from specific prey (plastidic specialist NCMs, pSNCMs; red (b)), or enslave entire specific autotrophic prey (size as length): (a) GNCMs Laboea (100 mm) and Strombidinium (50 mm); (b) pSNCMs, Mesodinium (60 mm) and Dinophysis (40 mm); (c) eSNCMs, Sphaerozoum (200 mm) and Noctiluca (500 mm). On maps, symbols correspond to the exact location where mixotrophic species/taxa were found (from more than 110,000 records); the grid indicates biogeographic provinces. Color-cast provinces indicate the presence of NCMs and white provinces correspond to the absence. Provinces marked with asterisk indicate that studies conducted in these areas did not record the presence of mixotrophic species; unmarked white provinces indicate a lack of field studies providing information on acquired phototrophy among microzooplankton. Reproduced with permission from Leles et al., 2017. Oceanic protists with different forms of acquired phototrophy display contrasting biogeographies and abundance. Proceedings of the Royal Society B 284, 20170664.

phototrophically. There are, however, exceptions. Some species can grow in complete darkness when fed a suitable prey (e.g., some chrysophyte species of the genera Poterioochomonas/Ochromonas and the dinoﬂagellate Fragilidium spp.) and light has a very limited effect on growth and prey uptake. In other cases the growth rate increases with irradiance despite the fact that ingestion is unaffected (e.g., the dinoﬂagellate Gymnodinium resplendens).

NCMs Most NCMs are dependent on both light and prey for sustained growth, but some species can grow completely heterotrophically in the dark if prey is available (e.g., the dinoﬂagellate Noctiluca scintillans; the ciliate Strombidium rassoulzadegani). Similar to CMs, the growth rates of NCMs generally increase with irradiance until saturated (e.g., Dinophysis acuminata; Mesodinium rubrum), while ingestion rates may (Dinophysis) or may not (Mesodinium) increase as a function of irradiance. Mixotrophy Among Freshwater and Marine Protists 7

Fig. 6 A panel of light microscopy images showing examples of feeding mechanisms exhibited by mixotrophic protists. (a–c) Prymnesium parvum (10–12 mm long). Photos by Hannah Blossom. (a–b) P. parvum with arrows showing the characteristic short haptomena which prevents P. parvum from using a haptonema to catch mobile prey, unlike many other haptophytes. (c) Two Teleaulax acuta prey (arrow), which P. parvum swarm around and feed upon. P. parvum releases toxic substances which immobilize prey, prior to capture. (d–g) Karlodinium armiger (Ka; ca. 20 mm in length) using tube-feeding to suck the cytoplasm (cp) of the prey (in this case, Fibrocapsa japonica; Fj). The cytoplasm separates from the plasma membrane of the prey. (g) Karlodinium armiger after complete ingestion of prey. Modified from Berge, T., Hansen, P.J., Moestrup, Ø., 2008. Feeding mechanism, prey specificity and growth in light and dark of the plastidic dinoflagellate Karlodinium armiger. Aquatic Microbial Ecology 50, 279–288, with permission from Inter-Research. (h) Karlodinium armiger tube-feeding on a much larger organism, in this case an unidentified polychaete trochophore larva. Modified from Berge et al., 2012. Marine microalgae attack and feed on metazoans. The ISME Journal 6, 1926–1936, with permission from Springer-Nature. (i) K. armiger swarming and feeding on an immobilized polychaete larva (ca. 500 mmin length). K. armiger uses toxins during prey capture. Photo by Terje Berge. (j–l) Alexandrium pseudogonyaulax ingesting whole cell prey through the sulcus. Heterocapsa rotundata prey (arrow). Modified from Blossom, H., Daugbjerg, N., Hansen, P.J., 2012. Toxic mucus traps: A novel mechanism that mediates prey uptake in the mixotrophic dinoflagellate Alexandrium pseudogonyaulax. Harmful Algae 17, 40–53 with permission from Elsevier. (m) A mucus trap made by A. pseudogonyaulax to capture prey, in this case Heterocapsa rotundata. A. pseudogonyaulax (Ap) is attached, behind the trap. Photo by Hannah Blossom. (n) A. pseudogonyaulax (Ap) with two food vacuoles (fv) attached to its mucus trap with Heterocapsa triquetra (Ht) entrapped. Photo by Hannah Blossom. A. pseudogonyaulax is approximately 35–40 mm wide.

Major and Minor Nutrients and Complex Macromolecules Major and minor nutrients Nutrients are essential for the growth and biochemical reactions of protists; particularly, photosynthetic activity relies on pigments and proteins that require nitrogen and metals (iron). It is generally assumed that inorganic or dissolved organic sources are the preferred source of nutrients for most microalgal groups. Thus, their availability at concentrations limiting growth is thought to promote prey ingestion in mixotrophs (Fig. 8). Indeed, many CMs appear to consume prey exclusively to obtain nutrients, maintaining photosynthesis as their primary source of carbon. However, most studies have focused on general nutrient limitation, with organisms grown in diluted growth media until limitation hampers growth, but exactly which nutrient is limiting is 8 Mixotrophy Among Freshwater and Marine Protists

Fig. 7 The dinoflagellate Karlodinium veneficum (¼Gyrodinium galatheanum). (a) Growth rate as a function of irradiance. K. veneficum was grown phototrophically in mono-specific cultures (circles) and mixotrophically in cultures with daily addition of cryptophyte prey Storeatula major (triangulars). (b) Ingestion rate of K. veneficum in food-saturated cultures as a function of irradiance. Error bars indicate 71 SE, n¼3. PFD: photon flux density. Adapted from Li et al., 1999. Feeding, pigmentation, photosynthesis and growth of the mixotrophic dinoflagellate Gyrodinium galatheanum. Aquatic Microbial Ecology 19, 163–176, with permission ©Inter-Research 1999. unknown. The few studies available focusing on limitation for specific nutrients have demonstrated that inter- and intra-species variability may occur in the mixotrophic response when limited for different nutrients. As both major (nitrogen and phosphorous) and minor (e.g., iron) nutrients are known to be important factors constraining planktonic primary production, the lack of clarity on interactions between inorganic nutrient stress and predation severely hampers our understanding of CM ecology. Nutrient uptake in NCMs is largely unexplored. However, uptake of inorganic and dissolved major nutrients has been shown recently in Dinophysis spp. and Mesodinium rubrum.

Complex macromolecules Many CMs are known to be auxotrophic for complex vitamins, amino acids and fatty acids, due to the loss or absence of part or all of the genes required for synthesis. This phenomenon is so prevalent that certain vitamins, such as B12, are routinely added to isolation and growth media to ensure microalgal growth. Paradoxically, very few studies have assessed the inﬂuence of complex macromolecules on feeding. For example, the haptophyte Prymnesium parvum has been shown to take up fatty acids from phagocytized ciliate and bacterial prey and is thought to use them as building blocks for membrane biosynthesis.

New Techniques in the Study of Mixotrophs to Assess Feeding and Nutrient Uptake

Omics During the last decade, the application of omics techniques has revolutionized microbial oceanography. Detailed genome information is necessary to understand the evolutionary history of protists and important events in eukaryotic evolution such as Mixotrophy Among Freshwater and Marine Protists 9

Fig. 8 The dinoflagellate Tripos furca (¼ Ceratium furca). Ingestion rates of cultures compared to inorganic phosphorus (PO4) and nitrogen (NO2 þ NO3 and NH4) concentrations over time. (a) f/2-P medium, (b) f/2-N medium. Data are presented as mean 7 SE. Reproduced from Smalley et al., 2003. Feeding in the mixotrophic dinoflagellate Ceratium furca is influenced by intracellular nutrient concentrations. Marine Ecology Progress Series 262, 137–151, with permission ©Inter-Research 2003. lateral or horizontal gene transfer. Many groups of aquatic protists still have no, or only very few, genomes sequenced. The data available indicate that phagotrophy and photosynthesis occur in most major lineages, thus the capacity for mixotrophy among many aquatic protists would be an ancestral adaptation/capacity. Transcriptomics is the analysis of gene expression. In this method, RNA is extracted from single cells or cultures and sequenced. As RNA is the messenger of genetic information from the nucleus to the cytoplasm of the cell, the sequenced RNA gives an overview over which genes are expressed at a certain point in time within single cells. This can improve our understanding of cell physiology, particularly how they respond to different environmental constraints. Using this method it was found that the availability of light has a minimal effect on population growth of the CM, Ochromonas sp. (chrysophyte). An upregulation of phototrophy- and phagotrophy-related genes was observed, while the availability of bacteria as prey led to major changes in carbon and nitrogen metabolic pathways. Further, transcriptomics of Prymnesium parvum showed that bacteria and ciliate prey were sources of organic carbon, macro- and micro-nutrients, organic nitrogen, and fatty acids. However, a large part of expressed genes in transcriptomes typically remains unassigned to a specific function, thus preventing a full understanding of physiological mechanisms.

High Resolution Chemical Imaging (NanoSIMS: Nanoscale Secondary Ion Mass Spectrometry) Based on cell lysis and extraction of molecules, omics data only provide averaged information from cell population and their organelles. This is an especially important consideration in the study of interactions between heterotrophs and their algal symbionts or between heterotrophs and their sequestered chloroplasts, since it is generally difficult to identify the origin of a gene function/metabolite, and to distinguish the role of each partner and/or organelles. Thus, intra-cell spatial information is needed to further decipher the complex nutritional strategies of mixotrophs. Recent advances in high-resolution single-cell chemical imaging coupled with stable isotopes now allow the observation and quantification of the metabolic activity within a single cell. In particular, nanoscale mass spectrometry imaging (nanoSIMS) is one of the most powerful chemical imaging platforms, available to biologists, with high mass and high spatial resolution capabilities (o100 nm) and excellent sensitivity (detection limit in the ppm or ppb range). NanoSIMS can document metabolic transfers between closely interacting cells and organelles, as well as the subcellular distribution of major nutrients (N, P and N, P and S) and metals. More 10 Mixotrophy Among Freshwater and Marine Protists specifically, following isotope incubation with 13Cor15N, nanoSIMS can elucidate the incorporation and flux of the isotope in cells at the subcellular level. This can be used to characterize and quantify the nutritional exchange between symbiotic partners, but also to quantify the contributions of photosynthesis and heterotrophy to the metabolism and growth of mixotrophs. As an example, nanoSIMS demonstrated that the chrysophyte Ochromonas sp. predominantly acquires carbon and nitrogen from bacterial prey when available, whereas inorganic C and N is assimilated primarily when bacterial abundance is low. This technique should be applicable for determining the relative role of phagotrophy and photosynthesis in both CMs and NCMs over a range of conditions.

Measurements of “In Situ” Feeding Rates in Mixotrophs: Variation in Time and Space (Seasonality, Vertical and Diel Variations)

Experiments on cultured organisms are required to understand the ecology and physiology of aquatic protists. However, scaling results of these microcosm experiments to large-scale natural systems is challenging for a number of reasons: (i) In contrast to simplified and strictly controlled cultures, natural ecosystems are immensely complex, with protist behavior influenced by interacting and fluctuating factors. (ii) Some external drivers for protist behavior, such as small scale turbulence, are very difficult to mimic under laboratory conditions. (iii) Isolation and culturing selects for species that survive under standard culture conditions, but these do not necessarily represent the most important species in the field. (iv) Culture studies tend to be with single strains or species within a genus, potentially ignoring significant inter-strain and inter-species variability. To understand the dynamics of mixotrophic protists in aquatic systems, culture studies need to be complemented by in situ studies. Field studies provide generalized patterns of how natural communities respond to environmental shifts; while culture studies help us understand these patterns by concentrating on selected populations subjected to specific environmental cues.

Bacterivorous CMs and NCMs Bacterivory rates in mixotrophs are most commonly measured by tracking the ingestion of fluorescently labeled bacteria (FLB). The assumption behind this technique is that mixotrophs feed indiscriminately, ingesting FLB at the same rate as co-occurring natural bacteria. This technique is relatively simple, with the possibility to target specific protist groups based on pigment signatures or gene probes. However, it has some caveats: (i) Not all mixotrophs feed indiscriminately, and selection for or against the FLB can occur, leading to biased results. (ii) Bacterivory rates can fall below detection limits when feeding rates are low. (iii) It is laborious and time-consuming. Alternative methods exist, but they present their own limitations, making the development of more accurate methods to determine mixotrophic bacterivory a major goal for future studies. As an example, food vacuole markers, such as acidotropic probes, can be used to determine the fraction of feeding cells in a population. This method is fast and highly sensitive, but cannot provide feeding rates since the number of ingested bacteria is unknown. Despite these complications bacterivory by mixotrophs, primarily CMs, has been generally shown to account for 30%–60% of bacterial consumption in diverse freshwater and marine systems. Mixotrophic bacterivory can thus equal or surpass that of heterotrophic protists, the organisms previously held as the principal cause of bacterial mortality, along with viral lysis. The principal bacterivorous mixotrophs are small (2–10 mm diameter) phytoflagellates of diverse phylogeny. Haptophytes and chrysophytes are well established as major bacterivores, but there are increasing indications that cryptophytes and chlorophytes could also play a significant role. Most information available is extrapolated from punctual studies, with very few assessments existing of bacterivory along spatial and temporal scales. The relative importance and regulation of bacterivory by different algal groups and the spatio-temporal variability of bacterivory are major fields for future study, and vital information for correctly incorporating these organisms into the models that predict productivity of aquatic systems.

CMs and NCMs That Feed on Other Protists Current methods fall short in measuring general in situ feeding rates of CMs and NCMs that feed on other protists, due to the much wider range of prey and predator sizes and the enormous differences in prey motility. Fluorescently labeled protistan prey can be used, but only to measure grazing rates of specific mixotrophs on very specific prey. This has been done to estimate grazing of Tripos furca (dinoflagellate CM) on small ciliates and of M. rubrum (ciliate NCM) on specific cryptophytes. However, this technique is not very useful for estimating the grazing rates of mixotrophs that ingest a variety of protistan prey, for example many of the CM dinoflagellates, plastid retaining oligotrich ciliates (NCMs), and mixotrophic foraminiferans and radiolarians (NCMs). Thus, a major future challenge is the development of methods to measure in situ feeding rates of these protists.

Incorporation of Mixotrophs in Food Web and Biogeochemical Models – Challenges for the Future

Models for support of ﬁsheries, harmful algal bloom science, and biogeochemistry are based almost exclusively on the traditional paradigm of nutrient-phytoplankton-zooplankton (i.e., NPZ) with links to higher trophic levels. Having established that the bulk of the “phytoplankton” and half the “microzooplankton” (which as a collective form ca. 60% of “zooplankton”) are photo-phago- Mixotrophy Among Freshwater and Marine Protists 11

Fig. 9 Conceptual models for the four main photo-phago-mixotrophs. Constitutive mixotroph (CMs) contain their own innate ability to synthesize photosynthetic machinery, use inorganic nutrients, and eat prey. Non-constitutive mixotrophs (NCM) do not possess an innate ability to do photosynthesis but must acquire the ability; they follow three patterns. The generalists (GNCM) can exploit photosystems from many prey types but have little ability to maintain and control them. Plastidic specialists (pSNCM) acquire plastids and some allied nucleic material from very few prey species (though they can eat many prey types); they can achieve a much enhanced level of phototrophy maintenance and control compared to GNCMs. The endosymbiotic specialists (eSNCM) contain fully functional phototrophic algae; they operate in a way akin to a floating greenhouse. Of these, the “perfect beast” model of Flynn and Mitra (2009) can describe all but the eSNCM format. DIM – dissolved inorganic matter (nutrients); DOM – dissolved organic matter. Reproduced from Flynn, K.J., Mitra, A., 2012. Building the “perfect beast”: Modelling mixotrophic plankton. Journal of Plankton Research 31, 965–992, with copyright permission from Oxford Academic. mixotrophs, radically alters the conceptual framework for such simulation models. In theory, the “phytoplankton” are no longer simply restrained by nutrients and light, and the “microzooplankton” are less likely to starve. Mixotrophy has long provided a basis for explorations of trait-trade-offs in theoretical ecology. However, mixotrophs in such models have almost exclusively been configured as CMs, often including trade-offs that have not been substantiated by physiology studies. If such trade-offs are not implemented, then modelled mixotrophs come to dominate; ironically, as we now better understand their physiology and ecology, that is indeed sometimes true. Models including mixotrophs typically fail to describe the variety of mixotrophic formats and the relative roles of different nutritional modes operating under different conditions. This situation, in part, reflects the lack of data and understanding of mixotroph physiology, and also of their prey that concurrently inhabit the environment. This lack of understanding has led some to propose conceptual frameworks for plankton ecosystem modelling. One framework is based on scaling laws which indicates an allometric scaling for osmotrophy (bacteria), phototrophy (phytoplankton), mixotrophy and then particle feeding (zooplankton). The fallacy of the argument is revealed by the fact that some of the very smallest protist “phytoplankton”, as exemplified by Micromonas, are mixotrophic. At the other extreme there are mixotrophs (foraminiferans, radiolarians, green Noctiluca) that are amongst the largest single celled organisms, and are larger than many metazoan zooplankton. We now know that mixotrophy is important in planktonic ecosystems and there is a drive for better marine ecosystem models for climate change prediction. However, we lack the physiological understanding and numeric data to properly construct and constrain simulation models. To get to where we need to be will take a significant and coordinated effort. For field work we need survey protocols that adequately and routinely sample the plankton down to 2 mm. We also need experimental protocols that can provide data on the contributions of phototrophy and phagotrophy in different organism groups. How molecular biology can contribute to our understanding in the quantitative way required to support modelling is at present unclear. An allied issue is that molecular biology provides high resolution details down to sub-species or individual cell level, while models have to clump organisms together as functional groups. It is not yet clear how best to build those functional groups. Laboratory experiments have significant potential to aid model construction, at least of the few species that are cultured. Notably such work requires simultaneous studies of both prey and mixotrophs growing alone, and together, with time series 12 Mixotrophy Among Freshwater and Marine Protists analysis of changes in nutrients, biomass, C:N:P and pigments. Fully integrated physiological-molecular-modelling studies likely offer the best route to enhancing conceptual, practical and numeric understanding; modelling efforts need to be fully embedded in the research effort and not left to the end. What we do have is sufficient conceptual understanding to build alternative structures, models that can be switched to better represent different physiological strategies, and that can be used to test against experimental data and to help us better interpret those data. Thus the “perfect beast” model, which is a simulation model (Fig. 9) has been used variously to explore CMs and NCMs in experimental, field and conceptual scenarios. Integrating molecular biology in such laboratory studies would also help ground omics approaches, providing the missing links between RNA transcripts and function mentioned above. This would, in turn, improve the linkage between field studies and models which is essential for building robust simulation platforms. What we can say for sure is that the complexity of the control of physiology within a given species (responses to light, nutrients, prey etc.), and differences in physiology between different mixotroph groups, argue strongly against attempts to deploy simple descriptions. Indeed, outputs from such models should likely be viewed with as much skepticism as we would now view simple NPZ models given what we now know about the role of mixotrophy in aquatic ecology.

Appendix A Supplementary Material

Supplementary data associated with this article can be found in the online version at doi:10.1016/B978-0-12-809633-8.20685-7.

Further Reading de Vargas, C., Audic, S., Henry, N., et al., 2015. Ocean plankton. Eukaryotic plankton diversity in the sunlit ocean. Science 348, 1261605. Decelle, J., Colin, S., Foster, R.A., 2015. Photosymbiosis in marine planktonic protists. In: Ohtsuka, S., Suzaki, T., Horiguchi, T., Suzuki, N., Not, F. (Eds.), Marine Protists: Diversity and Dynamics. Tokyo: Springer, pp. 465–500. Decelle, J., Probert, I., Bittner, L., et al., 2012. An original mode of symbiosis in open ocean plankton. Proceedings of the National Academy of Sciences of the United States of America 109, 18000–18005. Hansen, P.J., 2011. The role of photosynthesis and food uptake for the Growth of Marine Mixotrophic Dinoﬂagellates. Journal of Eukaryotic Microbiology 58, 203–214. Hartmann, M., Grub, C., Tarran, G.A., et al., 2012. Mixotrophic basis of Atlantic oligotrophic ecosystem. Proceedings of the National Academy of Sciences of the United States of America 109, 5756–5760. Keeling, P.J., Burki, F., Wilcox, H.M., et al., 2014. The Marine Microbial Eukaryote Transcriptome Sequencing Project (MMETSP): Illuminating the functional diversity of eukaryotic life in the oceans through transcriptome sequencing. PLOS Biology 12, e1001889. Leles, S.G., Mitra, A., Flynn, K.J., et al., 2017. Oceanic protists with different forms of acquired phototrophy display contrasting biogeographies and abundance. Proceedings of the Royal Society B 284, 20170664. Mitra, A., Flynn, K.J., Burkholder, J.M., et al., 2014. The role of mixotrophic protists in the biological carbon pump. Biogeosciences 11, 995–1005. Mitra, A., Flynn, K.J., Tillmann, U., et al., 2016. Deﬁning planktonic protist functional groups on mechanisms for energy and nutrient acquisition; incorporation of diverse mixotrophic strategies. Protist 167, 106–120. Musat, N., Musat, F., Weber, P.K., et al., 2016. Tracking microbial interactions with NanoSIMS. Current Opinion in Biotechnology 41, 114–121. Selosse, M.A., Charpin, M., Not, F., 2017. Mixotrophy everywhere on land and in water: The grand écart hypothesis. Ecology Letters 20, 246–263. Stoecker, D.K., Hansen, P.J., Caron, D.A., et al., 2017. Mixotrophy in the Marine Plankton. Annual Reviews in Marine Science 9, 311–335. Stoecker, D.K., Johnson, M.D., de Vargas, C., et al., 2009. Acquired phototrophy in aquatic protists. Aquatic Microbial Ecology 57, 279–310. Terrado, R., Pasulka, A.L., Lie, A.A., et al., 2017. Autotrophic and heterotrophic acquisition of carbon and nitrogen by a mixotrophic chrysophyte established through stable isotope analysis. The ISME Journal 11, 2022–2034.

Relevant Website http://www.mixotroph.org/ MixITiN.