Wild-type C57BL/6J mice and OT-II C57BL/6J mice (Barnden et al., 1998) were bred in the animal facility of the Institute for Virology and Immunobiology, University of Würzburg. CD55^-/- C57BL/6 mice (Sun et al., 1999) were obtained from the University of Cardiff and also bred in our animal facility. Crry^-/- C57BL/6 (Ruseva et al., 2009) and CD59a^-/- C57BL/6 mice (Holt et al., 2001) were bred at Cardiff University. All mice were kept in a specified pathogen free conventionally housed environment and used for experiments between six and 21 weeks of age.

The following antibodies and reagents were used to stain mouse cells: anti-CD4 Alexa Fluor 647 (clone RM4-5), anti-IFNγ Alexa Fluor 488 (clone XMG1.2), Streptavidin-PerCP (all BioLegend, San Diego, CA, USA), anti-CD25 biotin (clone 7D4, BD Pharmingen, Franklin Lakes, NJ, USA) anti-CD55 unconjugated (RIKO-3, Biolegend), anti-CD11b FITC (clone M1/70), anti-B220 Alexa Fluor 647 (clone RA3-6B2) (all BD Pharmingen), anti-CD3 PerCp (clone 145-2C11, BioLegend).

For staining of Pra1 a polyclonal antibody was raised in rabbits by immunization with purified recombinant Pra1. Aspf2-antiserum was generated by immunization of mice with purified recombinant Aspf2. Secondary polyclonal antibodies for staining of primary antibodies were goat anti-mouse-Ig FITC and donkey anti-rabbit-Ig PE (Jackson Immunoresearch, West Grove, PA, USA). Flow cytometry was performed on a FACSCalibur or LSR II flow cytometer using either CellQuest or DIVA software (BD Bioscience, Franklin Lakes, NJ, USA). We used FlowJo (TreeStar) to further analyze FACS data.

Recombinant Pra1wt and Aspf2 were expressed in Pichia pastoris and isolated via the His-tag (Luo et al., 2009; Bacher et al., 2014). For protein overexpression and purification, the pra1 gene encoding a protein lacking the C-terminal 61-amino acid was amplified from the pPICZαB-Pra1wt clone using the sequence specific forward primers ACTGAATTCTGTGGAGCCATCCGCAGTTTGAAAAAAGCGCGGCACCAGTTACGGTTACC and reverse primer ACTGGTACCGCGCACCCTTCGCCGGGAATTG, containing the restriction sites EcoRI and KpnI. The PCR product and plasmid pPICZαB were enzymatically digested, ligated, and sub-cloned into pPICZαB (Invitrogen, Karlsruhe, Germany). The resulting plasmid pPICZαB-Pra1ΔC61 was transfected and overexpressed in Pichia pastoris X33 (EasySelect^TMPichia Expression Kit, Invitrogen, Karlsruhe, Germany). The Pra1ΔC61 was purified as described (Luo et al., 2009).

Single cell suspensions were generated by mashing cervical, axillary, inguinal and mesenteric lymph nodes or spleens through a cell strainer (Falcon, Pittsburg, PA, USA). Single cell suspensions of splenocytes were then subjected to red cell lysis by hypoosmotic shock. Lymph node and red cell-lysed spleen cells were then resuspended in buffered salt solution (BSS) containing 0.1% (w/v) bovine serum albumin (BSA). Total lymph node or spleen cells were incubated with Pra1 (10 μg/ ml) or Aspf2 (10 μg/ ml) in PBS at 37°C for 30 or 45 min. For investigation of the influence of zinc on Pra1 binding, ZnCl₂ (1, 10, or 100 μM) was added while incubating cells together with Pra1. After washing bound Pra1 or Aspf2 were detected with a polyclonal anti-Pra1- (rabbit) or anti-Aspf2- (mouse) antiserum followed by PE anti-rabbit Ig polyclonal antibody (donkey; Dianova) or FITC anti-rabbit Ig polyclonal antibody (donkey; Dianova). For further stainings the samples were blocked with normal rabbit serum (1:500) or normal mouse Ig (20 μg/ ml, Sigma) followed by incubation with anti-CD4 mAb (Alexa Fluor 647) alone or together with anti-CD3 mAb (PerCp). For Kv1.3 detection, ShK-F6CA (0.3 μg/ml; Bachem AG, Bubendorf, Switzerland) was incubated together with mAb against cell surface proteins for 30 min at room temperature (Beeton et al., 2003).

To test for co-stimulation lymph node cells from WT mice were first enriched for CD4⁺ T cells (MagniSort Mouse CD4 T cell Enrichment Kit, eBioscience, Santa Clara, CA, USA or CD4⁺ T cell isolation kit, Miltenyi) resulting in at least 93% pure CD4⁺ T cells. Afterwards the cells were stained with anti-CD4-Alexa Fluor 647 and CD4⁺ T cells sorted using the FACS Aria III (BD) cell sorter (100% purity). For analysis of cell proliferation cells were incubated for 5 min at RT with 5 μM Vybrant CFDA SE Cell Tracer Kit (CFSE, Life Technologies, Carlsbad, CA, USA). Anti-CD3-mAb (2.5 μg/ ml, clone 145-2C11, BioLegend) was bound to 96-flat bottom-plates (Greiner, Kremsmuenster, Austria) after incubation on the plate o/n at 4°C dissolved in 0.1 M NaHCO₃-buffer (pH 9). After coating of the plate, non-specific binding was blocked by incubation with normal mouse immunoglobulin (20 μg/ml in BSS/0.1% BSA (w/v), Sigma Aldrich, St. Louis, MO, USA) at 37°C for 30 min. 1 × 10⁵ CFSE-labeled CD4⁺ T cells were added per well and anit-CD28 mAb E18 (Exbio) (Dennehy et al., 2006) (1 and 10 μg/ ml) or Pra1 (0.1 pg/ ml – 100 ng/ ml) were added in solution. For the cultures, we used complete RPMI 1640 medium supplemented with 1 mM sodium pyruvate, non-essential amino acids MEM (0.05–2 mM), 100 U/ ml penicillin and 100 μg/ml streptomycin, 30 μM mercaptoethanol, 2 mM L-glutamine (all Gibco) and 10% (v/v) heat-inactivated fetal calf serum. After 3 days, CD4 and CD25 were stained and expression of both markers, together with CFSE dilution, was analyzed by flow cytometry. To determine cytokine secretion, magnetically purified CD4⁺ T cells were cultured with plate-bound anti-CD3 mAb in the presence of soluble Pra1 (1–100 ng/ml) or anti-CD28 mAb (10 μg/ml) as already described. Alternatively, we coated Dynabeads^® Pan Mouse IgG (Invitrogen) with 10 μg/ml anti-CD3 mAb (clone 145-2C11, BioLegend) and 1 μg/ml anti-CD28 mAb (clone E18) according to the manufacturer’s instructions and added the beads at a bead to cell ratio of 5:1 to the cultures. After three days, culture supernatants were harvested and frozen at –70° C for subsequent cytokine analysis.

Lymph node cells and red blood cell-lysed splenocytes from OT-II mice were pooled and 2 × 10⁵ cells seeded per well of a 96-well round bottom plate (Greiner). OVA 323-339 peptide (OVAp, Charité Berlin) was added at 1 or 0.1 μM. For each condition six technical repeats were set up. On day three, culture supernatant was harvested and frozen at –70°C. Cells were stained for CD4 and CD25 expression and absolute cell numbers were determined by FACS using counting beads. To generate Th1 cells pooled spleen/lymph node cells from an OT-II mouse were depleted of CD4⁺ CD25⁺ regulatory T cells using anti-CD25 biotin (5 μg/ml, BD) and Streptavidin-beads and passage over an LD column (both Miltenyi). CD25-depleted spleen/lymph node cells were then seeded at 1 × 10⁶ cells/well (48-well plate, Greiner, final volume: 1 ml, 3 × 10⁶ cells in total), OVAp was added at 1 μM, recombinant mouse IL-12 (R&D Systems, Minneapolis, MN, USA) at 10 ng/ml and goat-anti-mouse IL-4 (R&D Systems) at 10 μg/ml. To obtain ‘Th0’ cells the CD25-depleted spleen/lymph node cells were stimulated with OVAp only. After five days of culture, CD4⁺ T cells were magnetically purified (CD4⁺ Isolation kit, Affymetrix, according to manufacturer’s instructions) and cultured for another two days in the presence 0.1 μM Proleukin^® (Novartis) (48-well plate, 5 × 10⁵ cells/well). Afterward, the cells were harvested and co-cultured with T cell-depleted splenocytes (anti-CD90.2 beads, LD column, Miltenyi) isolated from a WT C57BL/6 mouse (96-well round-bottom plate; 1 x 10⁵ T cell-depleted splenocytes per well, 1 × 10⁴ Th1 cells per well, triplicates). OVAp and Pra1 were added to these cultures in different concentrations. Every day 15 μl of culture supernatant were harvested per well to determine cytokine concentrations. On day three the Th1 cells were restimulated with phorbol myristate acetate (5 ng/ml) and ionomycin (500 ng/ml) for two hours at 37°C/5% CO₂ (v/v) before addition of Brefeldin A (10 μg/ml; all Sigma) and incubation for another two hours followed by cell surface staining for CD4, fixation, and permeabiliziation of the cells (Fix/Perm and Perm buffers, ebioscience) and intracellular staining for IFNγ expression (anti-IFNγ Alexa Fluor 488, clone XMG1.2, Biolegend).

Concentrations of the indicated cytokines were determined in culture supernatants using LEGENGplex^TM (Biolegend) according to the manufacturer’s instructions.

Summary graphs were generated and statistical testing was done using Excel © 14.4.1 (Microsoft) and Prism 4.0c © (GraphPad). P < 0.05 was considered statistically significant.

Stadt Würzburg (City of Würzburg) and UK Home Office (PPL 30/3038) approved breeding of the mice used in this study and the animals were culled by Annex IV approved techniques in accordance with Directive 2010/63/EU.

As Pra1 expression of C. albicans is induced upon contact with human cells and as it has already been shown to strongly modulate innate immunity (Zipfel et al., 2011), we studied direct binding of Pra1 to mouse CD4⁺ T cells in vitro. We used recombinantly expressed Pra1 purified from Pichia pastoris for staining and found that Pra1 bound to all splenocytes in a dose-dependent manner (Figure 1A, left histogram). Among total splenocytes CD11b⁺ CD3^- monocytic cells bound Pra1 particularly well (Figure 1A, middle left histogram), which was expected as complement receptor 3 (CR3, Mac1, CD11b/CD18) had been identified as a cellular receptor for Pra1 on mouse leukocytes (Soloviev et al., 2007, 2011). Splenic B (Figure 1A, middle right histogram) and T cells (Figure 1A, right), i.e., CD4⁺ and CD8⁺ T cells (Figure 1B), also clearly bound Pra1, albeit to a lesser extent than the monocytic cells.

As Pra1 binds zinc (Citiulo et al., 2012) we tested whether zinc influences Pra1 binding to mouse CD4⁺ T cells. Zn²⁺ which is found in serum at a concentration of 10 μM (Feske et al., 2012) and beyond increased Pra1 binding to mouse CD4⁺ T cells (Figures 2A–C) with plateau levels of binding reached after 30 min of incubation (Figure 2D). Moreover, a pra1 deletion mutant encoding a Pra1 protein lacking the putative zinc-binding domain (Pra1 Δ238–299) (Citiulo et al., 2012) showed almost no binding to mouse CD4⁺ T cells (Figure 2E).

The zinc binding capacity of Pra1 is shared by its homolog in A. fumigatus, i.e., the Aspf2 protein (Citiulo et al., 2012). We, therefore, used recombinantly expressed (P. pastoris) and purified Aspf2 and tested whether Aspf2 also directly binds to mouse CD4⁺ T cells. Aspf2, in contrast to Pra1, however, did not bind to the mouse T cells even when ZnCl₂ was added to the buffer (Figure 3). Thus, Pra1, but not Aspf2, directly binds to mouse CD4⁺ T cells and Pra1 binding is enhanced in the presence of extracellular zinc.

So far, only complement receptor 3 (CR3, Mac1, CD11b/CD18) has been identified as a cellular receptor for Pra1 on mouse leukocytes (Soloviev et al., 2007, 2011). As the staining pattern of Pra1 showed that Pra1 binds similarly well to all mouse CD4⁺ T cells (Figure 1B) we hypothesized that a complement regulatory protein expressed by all mouse T cells might be the receptor for Pra1. Therefore, we analyzed Pra1 binding to CD4⁺ T cells of CD55^-/- mice in more detail as CD55, Crry, and CD59a are the three complement-regulatory proteins expressed by mouse T cells (Miwa and Song, 2001). While CD4⁺ T cells of CD55^-/- mice were clearly devoid of CD55 expression at the cell surface (Figure 4A) binding of Pra1 was not reduced in the absence of CD55 (Figure 4B). Moreover, addition of zinc also increased binding of Pra1 to mouse CD4⁺ T cells of CD55^-/- mice (Figure 4B). Apart from CD55^-/- mice we also studied binding of Pra1 to CD4⁺ T cells of Crry^-/- and CD59a^-/- mice, which was also not reduced (Figure 4C). Therefore, Pra1 does not seem to interact with any of the three complement regulatory proteins expressed on the surface of mouse CD4⁺ T cells.

To gain further insight into the functional consequences of Pra1 binding to CD4⁺ T cells, we first studied its impact on T cell activation and proliferation in vitro. To avoid confounding effects through the interaction of Pra1 with CD11b/CD18 expressed by monocytic cells in our cultures, we FACS-sorted mouse CD4⁺ T cells which lack CD11b/CD18 to more than 99% purity. Stimulation of these highly pure CD4⁺ T cells by plate-bound anti-CD3 mAb and titrated amounts of Pra1 led to a dose-dependent increase in proliferation and CD25 expression similar to what we observed by adding an anti-CD28 mAb (Dennehy et al., 2006) (Figures 5A,B). Moreover, Pra1 truly induced a co-stimulatory signal in the T cells as in the absence of CD3 stimulation Pra1 did not activate the cells (Figure 5C). The same effect was observed for the anti-CD28 mAb (Figure 5C). Binding of Pra1 to mouse CD4⁺ T cells, thus, enhanced T cell activation and proliferation, which comprise the first steps of the adaptive immune response.

While the activation of naïve T cells and clonal expansion mark the beginning of the CD4⁺ T cell response, secretion of cytokines such as IFNγ characterize its effector and memory phase. We, therefore, analyzed cytokines in the supernatants of purified CD4⁺ T cells, containing in vivo generated memory T cells, stimulated via plate-bound anti-CD3 mAb and soluble Pra1 or anti-CD28 mAb (Dennehy et al., 2006) (Figure 6A). In contrast to its co-stimulatory effect on T cell activation and proliferation Pra1 suppressed secretion of both Th1 and Th2 cytokines (Figure 6A). Only IL-17 secretion appeared not to be affected, while secretion of IL-10 was below the detection limit in these experiments. Seemingly at odds with our observation concerning expression of the IL-2 receptor α-chain, CD25 (Figure 5), IL-2 concentrations were also reduced in the presence of Pra1. We assume that this reflects increased IL-2 consumption through increased receptor expression rather than reduced IL-2 production (Malek, 2008) uniting these two findings. To further test the capacity of Pra1 to inhibit cytokine secretion we added Pra1 to purified CD4⁺ T cells which we co-stimulated with anti-CD3/anti-CD28 mAb-coated Dynabeads^® (Figure 6B). Even under these conditions, which more faithfully mimic T cell-antigen presenting cell interactions than stimulation via plate-bound antibodies, Pra1 reduced cytokine, i.e., IFNγ, secretion by the CD4⁺ T cells (Figure 6B). The same was true for the supernatant of cultured C. albicans containing the whole array of secreted fungal proteins (Figure 6B).

Apart from binding to CD4⁺ T cells, Pra1 interacts with CD11b/CD18 integrin (Mac1) expressed by monocytic and granulocytic cells (Soloviev et al., 2011). To test whether Pra1 also suppresses IFNγ secretion in the presence of Mac1-expressing antigen-presenting cells (APCs) we stimulated total splenocytes from T cell receptor-transgenic OT-II mice with 1 μM OVA-peptide 323–339 in the presence of 100 or 1 ng/ml Pra1 (Figure 6C). Also under these conditions Pra1 inhibited IFNγ secretion by the OT-II CD4⁺ T cells.

Both in the presence of recombinant Pra1 as well as C. albicans supernatant, secretion of cytokines by mouse CD4⁺ T cells was, thus, reduced.

T cells isolated from healthy mice producing effector cytokines are by definition mostly resting memory T cells. During acute invasive C. albicans infection or C. albicans-induced inflammation the fungus, however, mainly encounters effector T cells. Therefore, we first deliberately generated OT-II Th1 effector cells during a five-day culture in vitro followed by a two-day resting phase and subsequent re-stimulation of the Th1 cells in the presence of APCs and different concentrations of peptide antigen and Pra1 (Figure 7A). Addition of Pra1 to Th1 cells stimulated with 0.1 μM OVA peptide reduced IFNγ secretion into the supernatant (Figure 7A, middle), while this was not the case at 1 μM OVA peptide (Figure 7A, right). Analyzing intracellular IFNγ expression by the Th1 cells after PMA/ionomycin re-stimulation, further, showed that the reduced secretion of IFNγ into the culture supernatant in the presence of Pra1 was not due to a per se lower capacity of the Th1 cells to produce IFNγ. Without OVAp re-stimulation the expression of IFNγ by the Th1 cells was, however, reduced suggesting that Pra1 increases the threshold for stimulation-induced cytokine secretion by CD4⁺ T cells. Incubation of Th1 cells with Pra1 showed, in comparison to OT-II CD4⁺ T cells cultured under Th0 conditions in parallel, that Th1 cells bind Pra1 better than Th0 cells (Figures 7C,D) suggesting that differentiated effector memory Th1 cells are a primary target of Pra1. In autoreactive pathogenic T cells Kv1.3 has been shown to be the main voltage-gated potassium channel and blocking the channel with the ShK peptide inhibits the autoreactive pathogenic T cells in animal models of autoimmunity in vivo (Beeton et al., 2001) and in cell cultures of human T cells in vitro (Wulff et al., 2003). Using a fluorescently labeled ShK peptide we observed that the Th1 cells expressed more Kv1.3 channels than Th0 cells and that Pra1 binding and Kv1.3 expression were positively correlated in Th1 cells (Figure 7E). Pra1, thus, preferentially bound to effector/memory Th1 cells inhibiting IFNγ secretion provided TCR stimulation did not surpass a certain threshold.