Atypical Membrane-Anchored Cytokine MIF in a Marine Dinoflagellate

Macrophage Migration Inhibitory Factors (MIF) are pivotal cytokines/chemokines for vertebrate immune systems. MIFs are typically soluble single-domain proteins that are conserved across plant, fungal, protist, and metazoan kingdoms, but their functions have not been determined in most phylogenetic groups. Here, we describe an atypical multidomain MIF protein. The marine dinoflagellate Lingulodinium polyedra produces a transmembrane protein with an extra-cytoplasmic MIF domain, which localizes to cell-wall-associated membranes and vesicular bodies. This protein is also present in the membranes of extracellular vesicles accumulating at the secretory pores of the cells. Upon exposure to biotic stress, L. polyedra exhibits reduced expression of the MIF gene and reduced abundance of the surface-associated protein. The presence of LpMIF in the membranes of secreted extracellular vesicles evokes the fascinating possibility that LpMIF may participate in intercellular communication and/or interactions between free-living organisms in multispecies planktonic communities.


Introduction
Macrophage Migration Inhibitory Factors (MIF) are pivotal cytokines/chemokines for vertebrate immune systems. They are involved in numerous infectious diseases and pathological disorders, such as septic shock, cancer, and rheumatic and cardiovascular diseases [1][2][3][4]. These multifunctional proteins control major cell functions such as migration, proliferation, and p53-mediated apoptosis [1,5,6]. MIFs are conserved across plant, fungal, protist, and metazoan kingdoms [7], but their functions have not been determined in most phylogenetic groups. The MIF proteins that have been described to date are small, conserved single-domain proteins (approx. 12.5 kDa). In a recent survey across 800 species of plants, protists, fungi, and metazoan, we reported the existence of typical MIF proteins in all kingdoms and phylogenetic groups [7]. Interestingly, we also identified a few atypical MIF proteins, including a putative membrane-anchored MIF protein from the Genbank TSA database (Genbank accession number JO713726.1) in the dinoflagellate Lingulodinium polyedra, but their existence remained uncertain as they may result from assembly or automatic annotation errors [7]. The possible occurrence of a transmembrane cytokine in a free living unicellular eukaryote would be unexpected, as MIFs are typically soluble proteins that are stored in vesicles of immune and non-immune cells, and massively Microorganisms 2020, 8,1263 2 of 14 secreted upon challenge [3,[8][9][10][11]. In addition, previous studies on MIFs from unicellular species have been restricted to parasitic protists [12] such as Plasmodium species [13], Leishmania, [14], or Entamoeba histolitica [15]. MIFs from these species are secreted into the blood of the vertebrate host and participate in immune evasion and pathogenesis [12,16]. The aim of this study was therefore to investigate the existence of a transmembrane MIF protein in the dinoflagellate L. polyedra, and to start exploring its potential function.

Lingulodinium polyedra
The dinoflagellate Lingulodinium polyedra (MCCV 130) was originally collected from the "Baie de Vilaine" and cultured at the Mediterranean Culture Collection of Villefranche (MCCV). Cultures were maintained in 75 mL of L1 medium [17] prepared with sterile and aged sea water adjusted to a salinity of 34. Cultures were maintained at 22 • C, under a 14/10 light/dark cycle with a light intensity of 250 µmol m −2 .s −1 . Cells of L. polyedra used for the experiments were collected from the cultures after 14 days. They were counted in 12 mL aliquots (triplicates) fixed with acidic lugol (4% v/v) using a liquid particle counter (HIAC/Royco 9703, Pacific Scientific Instruments, Jakarta Barat, Indonesia) with a size range of 2-80 µm 2 .

cDNA Sequencing and Expression Studies
For RNA extraction, L. polyedra were flash-frozen and ground using the TissueLyser LT (Qiagen Inc., Venlo, the Netherlands) for 2 min at 50Hz. Samples were centrifuged at 13,000× g, for 20 min, at 4 • C to remove debris. Total RNA was extracted with the PureLink™ RNA Mini Kit (Ambion-Life Technologies/Thermo Fisher Scientific, Waltham, MA, USA) following the manufacturer's instructions and were quantified using a Nanodrop (Life Technologies/Thermo Fisher Scientific, Waltham, MA, USA). RNAs were reverse-transcribed using iScript™ cDNA Synthesis Kit (Bio-Rad Laboratories, Hercules, CA, USA) for 30 min at 42 • C. The reverse transcriptase was inactivated for 5 min at 85 • C. The cDNAs were amplified by Phusion DNA Polymerase (New England Biolabs, Ipswich, UK) with the 5× Phusion GC Buffer (New England Biolabs, Ipswich, UK) and specific primers designed according to the GenBank sequence JO713726.1. Amplified cDNA was sequenced by Eurofins and deposited as LpMIF under the GenBank accession number MN911288.

Sequence Analysis
The physico-chemical parameters of molecular weight and theoretical isoelectric point (pI) of the deduced protein were computed using the ProtParam tool (http://web.expasy.org/protparam/) [22]. The functional domains and important sites of the protein were predicted by InterPro [23] and PFAM [24] softwares. The transmembrane regions and their topology were predicted using a set of 5 transmembrane prediction algorithms: TMHMM Server [25], HMMTOP [26], TMpred [27], ∆G Prediction server [28] and Phobius [29].
Agrobacterium-mediated transient expression was carried in N. benthamiana. Epidermal leaf cells were syringe-infiltrated with A. tumefaciens at a final OD 600 = 0.1, according to a classical procedure described classically such in Evangelisti et al. [32]. Two days after infiltration, leaf discs were observed by a laser confocal microscope (Carl Zeiss MicroImaging GmbH, Jena, Germany, #LSM880) using an excitation at 488 nm.

Mass Spectrometry
Mass spectrometry was performed by the Protein Science Facility (SFR Biosciences, Lyon, France). Briefly, following protein precipitation step (using trichlororoacetic acid 20% in volume, overnight at 4 • C), samples were washed twice in acetone and solubilized in 8M urea. Samples were reduced (Tris(2-CarboxyEthyl)Phosphine (Sigma-Aldrich, Saint Louis, MO, USA, #C4706), 5 mM, 57 • C, 1 h), alkylated (iodoacetamide (Sigma-Aldrich, Saint Louis, MO, USA, #I1149) 10 mM, RT, 45 min), and digested overnight at 37 • C with trypsin (1/100 ratio). Peptides digest was next desalted using C18 spin column (Thermo Fisher Scientific, Waltham, MA, USA). Peptides were dried in a speed-vac and suspended in 50 µL 0.1% HCOOH before nanoLC-MS/MS analysis. Samples were analyzed using an Ultimate 3000 nano-RSLC (Thermo Fisher Scientific, Waltham, MA, USA) coupled on line with a Q Exactive HF mass spectrometer via a nano-electrospray ionization source (Thermo Fisher Scientific, Waltham, MA, USA). 1 µL of peptide mixtures was loaded on a C18 PepMap100 trap-column (300 µm ID × 5 mm, 5 µm, 100Å, Thermo Fisher Scientific, Waltham, MA, USA) for 3.0 min at 20 µL/min with 2% ACN, 0.05% TFA in H 2 O and then separated on a C18 Acclaim PepMap100 nano-column, 50 cm × 75 µm i.d, 2 µm, 100 Å (Thermo Fisher Scientific, Waltham, MA, USA) with a 60 min linear gradient from 3.2% to 40% buffer B (A: 0.1% FA in H 2 O, B: 0.1% FA in ACN) and then from 40 to 90% of B in 2 min, hold for 10 min and returned to the initial conditions in 1 min for 15 min. The total duration was set to 90 min at a flow rate of 300 nL/min. The oven temperature was kept constant at 40 • C. Sample were analysed with TOP20 HCD method: MS data were acquired in a data dependent strategy selecting the fragmentation events based on the 20 most abundant precursor ions in the survey scan (350-1600 Th). The resolution of the survey scan was 60,000 at m/z 200 Th. The Ion Target Value for the survey scans in the Orbitrap and the MS 2 mode were set to 3E6 and 1E5 respectively and the maximum injection time was set to 60 ms for both scan modes. Parameters for acquiring HCD MS/MS spectra were as follows; collision energy = 27 and an isolation width of 2 m/z. The precursors with unknown charge state or a charge state of 1 were excluded. Peptides selected for MS/MS acquisition were then placed on an exclusion list for 20 s using the dynamic exclusion mode to limit duplicate spectra.
Proteins were identified by database searching using Sequest HT with Proteome Discoverer 2.2 software (Thermo Fisher Scientific, Waltham, MA, USA) against the uniprot Lingulodinium polyedra database (50 entries, June 2019) and LpMIF sequence. Precursor mass tolerance was set at 10 ppm and fragment mass tolerance was set at 0.02 Da, and up to 2 missed cleavages were allowed. Oxidation (M), acetylation (Protein N-terminus) were set as variable modification, and Carbamidomethylation (C) as fixed modification. Proteins were filtered with a fixed value PSM validator.

Stress Bio-Assays
To investigate L. polyedra stress responses, three experimental samples were compared: (i) L. polyedra cells; (ii) L. polyedra exposed to the copepods for 24 h; and (iii) L. polyedra exposed to the copepods for 24 h and then maintained under normal conditions (without copepods) for a week. Each treatment was replicated 3 times. All samples were maintained in 75 mL flasks containing 50 mL of autoclaved, aged and 0.2-µm filtered seawater (salinity 34), at 22 • C. Prior to the experiments, copepods (adults and late copepodites) were collected from the culture tanks, rinsed twice in seawater and transferred to each flask (50 individuals per flask). They were left for 2 days without food to allow gut clearance and salinity acclimation before the experiments. Lingulodinium polyedra cells were added in each flask at a final concentration of 100 cells mL −1 . All samples were maintained in 75 mL flasks containing 50 mL of autoclaved, aged and 0.2-µm filtered seawater (salinity 3.4%), at 22 • C. After 24 h of exposure, 2 mL of L. polyedra were sampled for each replicate to study the photosystem II (PSII) function using a Multi-Color Pulse-Amplitude-Modulated analyser (MC-PAM, Heinz Walz Gmbh, Effeltrich, Germany) by measuring the maximum quantum yield of PSII (Fv/Fm ratio) [33,34]. Prior to fluorescence measurement, samples were dark-acclimated at 24 • C for 15 min to re-open PSII reaction centers and relax non-photochemical quenching.
Images of fluorescent cells were converted by the LSM Image Browser software, ZEN 2011, Blue edition (Carl Zeiss, Jena, Germany) into JPEG files (8 bit black and white quality). Fluorescence was measured with ImageJ (cut off >10 pixels) [35]. Results are shown as the ratio of fluorescent vs total area. Nine images were analysed per condition, corresponding to the examination of approximately 20 independent dinoflagellates. All images, including those converted for ImageJ analyses, have been deposited in Mendeley Data (http://dx.doi:10.17632/5j4k6r79ys.1).

Immunogold Labelling and Transmission Electron Microscopy (TEM)
For ultrastructural immunocytochemistry, cells were fixed with 4% paraformaldehyde (PAF) or 4% PAF and 0.2% glutaraldehyde (PG) in phosphate buffer 0.1 M (pH 7.4) (Sigma Aldrich, Saint Louis, MO, USA). They were then dehydrated in ethanol series and embedded in acrylic resin (LR-WHITE (Electron Microscopy Sciences-EMS, Hatfield, PA, USA) before sectioning. For immunogold labelling, grids were deposited face down on the top of small drops of the following solutions: PBS containing 50 mM NH4Cl for 10 min, PBS containing 1% BSA and twin 20 for 10 min, PBS containing the relevant mAbs in 1% BSA and tween 20 for 1 h, PBS containing 0.1% BSA for 10 min, PBS 10 min, PBS containing 1% BSA and tween 20 with 15 nm protein-A gold conjugated (PAG-15 nm Cell Microscopy Core, Utrecht, The Netherland) for 30 min, PBS containing 0.1% BSA for 5 min, PBS for 5 min twice, PBS containing 1% glutaraldehyde for 5 min and distilled water for 5 min. For controls, only the primary antibody was omitted. The sections were then contrasted with uranyl acetate (4% in water) and visualized using a JEM 1400 Electron Microscope (JEOL, Tokyo, Japan) operating at 100 kV equipped with a MORADA SIS camera (Olympus, Tokyo, Japan).

Scanning Electron Microscopy (SEM)
Cell pellets were fixed with a 1.6% glutaraldehyde solution in 1:1 mixture of 0.2 M sodium cacodylate buffer (pH 7.4) and artificial sea water at room temperature and then stored at 4 • C. After three washes in distilled water, cells were filtered on a 0.2 µm isopore filter (Merck Millipore, Carrigtwohill, Ireland). Samples on filters were subsequently dehydrated in a series of ethanol baths (70%, 96%, 100% three times, 15 min each). After a final bath in hexamethyldisilazane (Carl Roth GmbH, Karlruhe, Germany) (HMDS, 5 min), samples were left to dry overnight. Samples on filters were mounted on SEM stubs with silver paint and coated with platinum (3 nm) prior to observation. SEM observations were performed with a JSM-6700F SEM (JEOL, Tokyo, Japan) at an accelerating voltage of 3 kV.

Quantification and Statistical Analysis
All experiments were repeated three times independently. Kruskal-Wallis tests were used to compare the photosystem II activity between the different experimental conditions (control, stress, post-stress). Fluorescence signal intensities between the three different conditions (control, stress, unstressed) were compared using the Kruskal-Wallis statistical test (* = p ≤ 0.05, ** = p ≤ 0.01, *** = p ≤ 0.001). Data are presented as bars of relative fluorescence. Differences in gene expression levels were tested for statistical significance by one-way ANOVA and Tukey-Kramer tests (Software Prism v.5.0, GraphPad Software, San Diego, CA, USA). Expression levels of the target genes are presented as a relative expression normalized to internal reference genes and to expression in the control sample.  Figure 1a), previously shown to determine the human MIF oxidoreductase (TPOR; thiol-protein oxidoreductase) activity [36]. This site is associated with the regulation of cellular redox homeostasis. In humans, the biological functions of these two enzymatic activities remain largely unknown, in particular because their endogenous substrates have not been identified yet [3]. However, there is increasing evidence for their roles in the pro-inflammatory effects of MIF [37]. Mutations in or inhibition of the enzymatic active sites lead to the loss of major MIF biological functions [37], possibly due to conformational changes in the catalytic domains [38]. Conservation of these active sites in LpMIF indicates that the protein could have typical MIF biological activities.

Lingulodinium polyedra Expresses a Transmembrane MIF Protein
In addition to the MIF domain, transmembrane (TM) helix and general hydrophobicity predictions support the existence of three TM helices in the N-terminal part of the protein (Figure 1b). Although the exact positions of the transmembrane regions are determined slightly differently (Figure 1b), all prediction tools agree on the calculated topology, which predicts a non-cytoplasmic localization for the MIF domain (Figure 1c). As a first approach to ascertain the transmembrane nature of LpMIF, we prepared the membranes of L. polyedra cells and observed a major protein band (Figure 1d). The presence of an LpMIF peptide in this fraction was confirmed by mass spectrometry (Figure 1a). The apparent molecular mass for LpMIF was revealed to be approximately 60 kDa in the immunoblots, which might correspond to a dimer (expected at 50 kDa) as estimated molecular masses of proteins are not accurate under non-denaturing conditions. Typical MIF proteins can co-exist in several oligomeric forms (mono-, di-, or trimers), although MIF trimers represent the most prominent and stable form [10,39,40]. Future structural studies should reveal whether homodimers are the predominant form of the transmembrane LpMIF. Further evidence for the membrane association of LpMIF was provided by Agrobacterium tumefaciens-mediated transient expression assays to produce LpMIF in cells of Nicotiana bentamiana leaves. The use of this heterologous plant expression system confirmed that LpMIF has a non-cytoplasmic localization and accumulates in the membranes of vesicles and in the plasma membrane of the plant cells (Figure 1e). In contrast, LpMIF without the transmembrane domains shows a nucleo-cytoplasmic expression in N. benthamiana cells, which is typical for small, soluble proteins (Figure 1f).

LpMIF Is Present in the Membranes of Vesicular Bodies, Cell Wall and Extracellular Vesicles
The subcellular localization of LpMIF in L. polyedra cells was investigated by transmission electron microscopy (TEM) with immunogold labelling. LpMIF-bound gold particles were observed

LpMIF Is Present in the Membranes of Vesicular Bodies, Cell Wall and Extracellular Vesicles
The subcellular localization of LpMIF in L. polyedra cells was investigated by transmission electron microscopy (TEM) with immunogold labelling. LpMIF-bound gold particles were observed in the membrane of vesicular bodies (Figure 2a), further supporting the membrane localization of this atypical MIF protein. Extensive observation of L. polyedra cells did not reveal the presence of labelled LpMIF in any particular organelle (Figure 2a and Figure S1b) other than the vesicular bodies. The highest concentrations of LpMIF-bound gold particles were observed in the cell wall or theca (Figure 2b), particularly in the vicinity of the apical pore ( Figure S1a). The cell wall of thecate dinoflagellates (or theca) consists of multiple membrane layers strengthened by polysaccharides [41]. This membranous composition therefore supports the presence of membrane-anchored proteins such as LpMIF. Confocal laser-scanning microscopy (CLSM) confirmed the presence of LpMIF in the theca, and further showed specific accumulation at the apical pore and at secondary thecal pores (Figure 2c,d; compare with scanning electron micrographs in Figure 2e,f). Dinoflagellates are extensively studied for their toxicity because of their roles in marine red tides and shellfish poisoning [42,43], yet their ultrastructure and physiology remain poorly understood. However, dinoflagellate pores, and apical pores in particular, are known to be involved in mucilage excretion on the cell surface [44]. In addition, thecal pores extrude fibrous and mucous trichocysts (cf Figure S1b) over the entire cell surface, which contribute to dynamic streamlining and defense against grazing [45].
To determine if LpMIF is secreted, we repeated the immunofluorescence labelling using milder washing conditions (identical procedure but without Triton). Confocal analyses subsequently showed MIF-labelled extracellular vesicles at the surface of L. polyedra cells (Figure 2g) with particular accumulations in mucus outside the apical pore (Figure 2h).
MIF proteins of metazoa are soluble and present in the cytoplasm of numerous cell types, stored in vesicular bodies and released in extracellular fluids (blood, haemolymph) in response to various stimuli such as bacterial lipopolysaccharides (LPS), tumour necrosis factor, or hypoxia [8,11]. Despite the decades of functional studies on human MIF, the exact mechanism of MIF secretion remains unclear. Human MIF, which does not localize to the endoplasmic reticulum or Golgi apparatus, is secreted through unconventional secretory pathways [11]. Some studies showed that MIF secretion was sensitive to inhibitors of ATP-binding cassette (ABC) transporters, while others reported the release of MIF via exosomes [3]. While the trafficking processes of LpMIF clearly requires dedicated studies, the presence of this atypical transmembrane MIF in intra-and extracellular vesicles is reminiscent of the localization of vertebrate MIFs.

LpMIF Is Under-Expressed during a Stress Response
Predicting putative functions for a transmembrane MIF in a free-living unicellular organism is particularly challenging. Previous studies on MIFs from unicellular species have been restricted to parasitic protists [12] such as Plasmodium species [13], Leishmania [14], Giardia lamblia [46], or Entamoeba histolitica [15]. MIFs from these species are secreted into the blood of the vertebrate host and participate in immune evasion, host invasion, and pathogenesis [12,16]. Such functions are not relevant for a free-living marine dinoflagellate. Furthermore, the major function of MIFs as innate immune regulators, which has been reported in free-living metazoan species, including invertebrates [10,16,47,48], is hardly applicable and measurable on non-model protists such as L. polyedra. We therefore hypothesized that LpMIF may be involved in a stress response, since functional links between immune and stress responses are well established in both animal and plant model species [49][50][51][52][53] and since MIFs from vertebrates are also involved in stress responses [2]. Under natural conditions, L. polyedra lives in phytoplankton communities that are exposed to grazing copepod predators. Copepods are known to induce defense responses, such as an increase in bioluminescence [54,55], a change of swimming behaviour [56], an increase of toxin production [57], and a decrease in photosynthetic activity [34,58]. It has been suggested that these induced defenses may provide slow-growing L. polyedra with the required advantage to co-exist with faster-growing competitors in phytoplankton communities [55].
Exposure of L. polyedra to the copepod Sarsamphiascus cf. propinquus for 24h correlated with a significant decrease in photosynthetic activity (Figure 3a), confirming that the copepod did induce defense responses in our experiment. This stress response is transient, since the fluorescence recovered and increased above control levels seven days after copepod exposure (Figure 3a). A decrease in LpMIF transcript content (Figure 3b) and labelled LpMIF protein content (Figure 3c,d) occurred during this stress response. This response indicates that, in this context, the transmembrane MIF from L. polyedra is indeed a stress-responsive protein.  (Figure 2b), particularly in the vicinity of the apical pore ( Figure S1a). The cell wall of thecate dinoflagellates (or theca) consists of multiple membrane layers strengthened by polysaccharides [41]. This membranous composition therefore supports the presence of membrane-anchored proteins such as LpMIF.
Confocal laser-scanning microscopy (CLSM) confirmed the presence of LpMIF in the theca, and further showed specific accumulation at the apical pore and at secondary thecal pores (Figure 2c,d; compare with scanning electron micrographs in Figure 2e,f). Dinoflagellates are extensively studied for their toxicity because of their roles in marine red tides and shellfish poisoning [42,43], yet their ultrastructure and physiology remain poorly understood. However, dinoflagellate pores, and apical pores in particular, are known to be involved in mucilage excretion on the cell surface [44]. In addition, thecal pores extrude fibrous and mucous trichocysts (cf Figure S1b) over the entire cell surface, which contribute to dynamic streamlining and defense against grazing [45].

Conclusions
MIFs are enigmatic proteins with extremely complex trafficking processes, biological activities, functions, and signalling pathways [3,11]. Here, we provide evidence for a new dimension in the complexity of MIF structure and possible functions. The presence of LpMIF in the membranes of secreted extracellular vesicles evokes the fascinating possibility that LpMIF participates in intercellular communication and/or regulation in the marine environment. There is increasing evidence that phytoplankton communities, including a variety of protist and bacterial species, produce substantial amounts of extracellular vesicles that are involved in interspecies communication, nutrient acquisition and exchange, biofilm formation, and cellular defense processes [59][60][61]. The multiple functions of typical MIF proteins, such as the effects on cell migration, cell division, apoptosis, and induction, of other immune or hormonal regulators, may all have a role in coordinating stress responses within a dinoflagellate colony. In addition, since MIF proteins are prone to interspecies and inter-kingdom interactions, as shown by their involvement in host-parasite interactions [16,62,63], they may also help regulate the interactions of multispecies planktonic communities. Such possibilities can now be explored by investigating LpMIF functions under biotic stress conditions. Supplementary Materials: The following are available online at http://www.mdpi.com/2076-2607/8/9/1263/s1, Figure S1: Immunogold-labelled electron micrographs of L. polyedra, Figure S2: Negative control samples of unstressed and stressed L. polyedra, Table S1: Accessions, names and primer sequences for transcripts analyzed in this study.