Plasma proteomic profiles differ between European and North American myotid bats colonized by Pseudogymnoascus destructans

Emerging fungal diseases have become challenges for wildlife health and conservation. North American hibernating bat species are threatened by the psychrophilic fungus Pseudogymnoascus destructans (Pd) causing the disease called white‐nose syndrome (WNS) with unprecedented mortality rates. The fungus is widespread in North America and Europe, however, disease is not manifested in European bats. Differences in epidemiology and pathology indicate an evolution of resistance or tolerance mechanisms towards Pd in European bats. We compared the proteomic profile of blood plasma in healthy and Pd‐colonized European Myotis myotis and North American Myotis lucifugus in order to identify pathophysiological changes associated with Pd colonization, which might also explain the differences in bat survival. Expression analyses of plasma proteins revealed differences in healthy and Pd‐colonized M. lucifugus, but not in M. myotis. We identified differentially expressed proteins for acute phase response, constitutive and adaptive immunity, oxidative stress defence, metabolism and structural proteins of exosomes and desmosomes, suggesting a systemic response against Pd in North American M. lucifugus but not European M. myotis. The differences in plasma proteomic profiles between European and North American bat species colonized by Pd suggest European bats have evolved tolerance mechanisms towards Pd infection.

However, several fungal pathogens are known to be major threats to plants (Anderson, Cunningham, et al., 2004). Although fungi are generally considered of minor concern for animals, there is an increasing number of examples of fungal EIDs negatively impacting wildlife species (Fisher et al., 2012;Scheele et al., 2019). For example, pathogenic fungi have led to massive declines in snakes (Ophidiomyces ophiodiicola; Lorch et al., 2016) and amphibians (Batrachochytrium dendrobatidis and B. salamandrivorans ;Berger et al., 2016). White-nose syndrome (WNS), caused by the psychrophilic fungus Pseudogymnoascus destructans (Pd), is an emerging fungus of cave-hibernating bats. Since its initial outbreak in 2006, the fungus has killed millions of North American bats, fatalities reaching up to 99% per hibernaculum . The associated mortality has resulted in local extinction of several previously common myotid bat species Coleman & Reichard, 2014). Thirteen bat species have been diagnosed with WNS in North America and six more species exhibit a fungal growth without disease confirmation (www.white noses yndro me.org). Susceptibility varies among bat species, with the little brown bat (Myotis lucifugus) exhibiting mortality rates as high as 91% Turner, Reeder, & Coleman, 2011). In contrast, the big brown bat (Eptesicus fuscus) appears to be resistant to Pd (Frank et al., 2014;Moore et al., 2018).

. Clinical pathology of Pd infection in North
American species is associated with abnormal behaviour, a disturbed natural hibernation cycle characterized by increased arousal frequency resulting in premature depletion of fat stores during hibernation (Reeder et al., 2012). Diseased animals show disturbed electrolyte and hydration balance (Cryan et al., 2013;Willis, Menzies, Boyles, & Wojciechowski, 2011), oxidative stress (Moore et al., 2013), chronic respiratory acidosis (Verant et al., 2014), altered complement protein activity (Moore et al., 2011) and fever response (Mayberry, McGuire, & Willis, 2018). Pd infection does not produce primary inflammatory cellular infiltration into infected tissues , which would be associated with leukopenia (i.e., low white blood cell counts) during mammalian hibernation (Bouma, Carey, & Kroese, 2010). The expression of inflammatory, wound healing and metabolic genes increases without the recruitment of neutrophils and T cells to the site of invasion . WNS may also elicit an exaggerated inflammatory response during arousal (Meteyer, Barber, & Mandl, 2012), which might further complicate the pathology of the disease. How WNS causes death in bats is not clear (Wibbelt, 2018), but is probably a combination of various physiological disturbances provoked by Pd resulting in a multistage progression of the disease (Verant et al., 2014).
In contrast to North America, Pd is widely distributed among hibernating bat species in Europe (Puechmaille et al., 2011;Zukal et al., 2014), but without associated clinical signs or mortality (Wibbelt et al., 2013). There are 17 bat species which are colonized by the fungus in Europe, with the greater mouse-eared bat (Myotis myotis) being most often infected (Wibbelt, 2018). Recently, four additional species tested positive for Pd in Northeast China extending the host and the geographic range of the pathogen (Hoyt et al., 2016). Lesions associated with Pd infection in European bats are less pronounced than in North America (Wibbelt et al., 2013).
However, deep cutaneous invasion associated with neutrophil infiltration has been observed in some species (Bandouchova et al., 2015;Wibbelt et al., 2013). The differences in epidemiology and pathology of Pd infection in European and North American bats suggest that European bats have co-evolved with the fungus developing immunologic and perhaps behavioural defence mechanisms to the fungus (Bandouchova et al., 2018;Lilley et al., 2019;Puechmaille et al., 2011;Wibbelt et al., 2010;Zukal et al., 2016).
Natural and experimental infection studies of the physio-and immunopathological aspects of WNS have been performed on susceptible and resistant North American bat species. However, similar studies are rare for European species. The different approaches and experimental setups used among studies complicate data interpretation and across-study comparisons Moore et al., 2018;Davy et al., 2017, but see Lilley et al., 2019.
Several studies have demonstrated weak correlations between transcriptome and proteome data (e.g., Gygi, Rochon, Franza, & Aebersol, 1999;Lu et al., 2004;Maier, Güell, & Serrano, 2009;Moritz, Mühlhaus, Tenzer, Schulenborg, & Friauf, 2019). The information available on pathophysiological aspects of Pd colonization is generally based on transcriptome data (Davy et al., 2017;Field et al., 2015Field et al., , 2018Lilley et al., 2019) and it would be reasonable to expect lack of transcriptome and proteome concordance. These methods are both necessary and complementary to understand the underlying molecular changes associated with the various disease states in different hosts. To address the paucity of comparative European and North American Pd infection studies and to get further insights on the molecular mechanisms associated with the infection, we compared the plasma proteomic profile of hibernating North American M. lucifugus and European M. myotis.
We hypothesized that differences between healthy and Pd-colonized individual proteomic profiles will vary among species from different continents and reflect patho-physiological changes associated with the colonization by Pd.

| Ethics statement
Capture, handling and sample collection protocols for this study

| Sample collection
During March-April of 2012, greater mouse-eared (M. myotis) and little brown (M. lucifugus) bats were hand collected from hibernacula in Germany and Canada, respectively. In Germany, bats were collected from two mines and two cellars in Northern Bavaria (n = 12, six female and six male). In Québec, Canada, 12 (six female and six male) individuals were collected from the Trou de la Fée caverne and Laflèche caves. Equal numbers of bats with and without clinical signs of Pd colonization were collected for both species. To avoid cross-contamination, gloves were changed after processing a given animal and all clothes, shoes and gear were disinfected before moving to the next hibernaculum following accepted decontamination protocols (www.white noses yndro me.org).
Immediately after removing the bats from the hibernacula, we used adhesive tape to collect samples for mycological analysis (Wibbelt, 2018;Wibbelt et al., 2010). We recorded the localization of fungal colonization and afterwards bats were euthanized using isoflurane overdose followed by exsanguination. Necropsy was performed locally and blood and tissues were taken. Blood plasma was separated by centrifugation and all samples were stored in liquid nitrogen. Samples were transported to the Leibniz Institute of Zoo and Wildlife Research Berlin, Germany (IZW), where they were stored at -80°C until further analysis.
Although histopathological lesions characteristic of WNS have been described in European species, European bats cannot be considered diseased (Wibbelt, 2018). We assigned collected bats from Europe and North America to healthy and Pd colonized groups according to their Pd colonization status. Colonization status was defined by the results of analysing the samples collected with adhesive tapes. After collection, the tape samples were transferred to glass slides which were examined under light microscope for characteristic Pd conidia. If observed, isolation and mycological confirmation of the fungus was subsequently performed (Wibbelt et al., 2010). In the case of M. lucifugus, Pd infection was also diagnosed histologically ).

| 2D fluorescence difference gel electrophoresis (2-D DIGE)
Plasma proteomic profiles of 12 M. myotis individuals (six healthy and six Pd-colonized) and 12 M. lucifugus individuals (six healthy and six diseased; Pd-colonized and WNS-positive) were determined. Albumin depletion was not performed but albumin was excluded during the mass spectrometry identification process, as described in a prior study on hibernating M. myotis (Hecht et al., 2015).
We performed 2-D DIGE analyses across six SDS gels for each species. Samples were labelled using the G-Dye Refraction-2D labelling kit (NH DyeAGNOSTICS GmbH, Germany) according to the manufacturer's protocol. Briefly, after we determined the total plasma protein concentration using a NanoDrop, samples were diluted to the required concentration of 5 µg protein/µl in labelling buffer (Bio-Rad, USA). Isoelectric focusing was performed using the following conditions: step 1, 300 V, 150 V/hr rapid; step 2, 600 V, 300 V/h rapid; step 3, 1,500 V, 750 V/hr rapid; step 4, 3,000 V, 48,000 V/hr rapid; step 5, 6000 V, 10,000 V/hr rapid; step 6, 300 V, 5 hr; total 60,700 V/hr.
After the IEF run, the IPG stripes were equilibrated in equilibration buffer (EB: 6 M Urea, 2% SDS, 0.375 M Tris, 20% v/v glycerol) with first 20 mg/ml DTT for 15 min, followed by EB with 25 mg/ ml iodoacetamide (IAA) for 15 min. For running the second dimension, stripes were placed on 12.5% SDS gels in 27.5 × 22 cm low fluorescence glass cassettes (NH DyeAGNOSTICS GmbH, Germany) and overlaid with 1% agarose including bromophenol blue. Gel electrophoresis was performed in a SE900 electrophoresis unit (Hoefer Inc., USA) for a minimum of 2,400 V/hr and a maximum of 2,550 V/ hr at 80 mA/gel, 100 W and 100 V. Imaging of the gels was performed by fluorescence scanning on a Typhoon 9,400 Imager (GE Healthcare, USA) at excitation/emission wavelengths of 498/524 nm (G-Dye100), 554/575 nm (G-Dye200) and 648/663 nm (G-Dye300).
To evaluate the expression pattern of protein spots separated by 2-D DIGE for healthy and Pd-colonized individuals, sample gels were analysed using the Delta2D software (DECODON, Germany). Briefly, an IS G-Dye100 image was designated as the master gel based on the largest number of detectable spots, and then connected to all images by a "sample-in-gel" warping strategy in the Delta2D software. For warping of gels, we defined matched vectors between distinct protein spots. They were chosen automatically and manually. For expression analysis of protein spots, a fused image of all sample images (G-Dye200 and G-Dye300) was generated and a consensus spot pattern for normalization against the internal standard images was applied. Matched protein spots present in all sample images with a minimum of 1.5fold change between the healthy and Pd-colonized state were statistically analysed using a nonparametric Wilcoxon Rank Sum test (alpha: p < .05) with the Delta2D statistic software TMeV (Decodon). Protein spots on the edges and in the upper region of the gels where no adequate protein separation was achieved were excluded from spot normalization and analysis in the Delta2D software. According to this, pooled plasma was labelled with G-Dye300 and mixed with unlabelled sample. A total protein concentration of 900 µg was applied per IPG stripe (pI 3-10 Nl) and passively rehydrated. We performed isoelectric focusing on an IEF 100 Focusing Unit (Hoefer Inc, USA) and second dimension separation on a SE900 electrophoresis unit (Hoefer Inc., USA) with a 12.5% SDS gel. After 2-D electrophoresis, the gel was stained with Coomassie Brilliant Blue and fixed (40% ethanol, 10% acetic acid). Preparative gel images were matched with sample images in the Delta2D software.

| Preparative 2D gel for protein identification
According to analysed expression profiles proteins spots with a fold change difference of ≥1.5 and a statistical significance of p < .05 between healthy and Pd-colonized individuals were picked for protein identification. It was not possible to pick all differentially expressed spots as not all spots were distinguishable on the Coomassie blue stained gel.

| Protein identification by liquid chromatography-mass spectrometry (LC-MS)
Excised gel spots were washed with water, 25 mM ammonium bicarbonate in acetonitrile/water (1:1) and 50 mM ammonium bicarbonate, before they were shrunk by dehydration in acetonitrile and

| RE SULTS
In the European bat species M. myotis, we observed 157 matched protein spots in all samples based on consensus spot patterns of analysed expression profiles. We did not detect different levels of protein expression between healthy and Pd-infected in M. moytis. Therefore, proteins were not identified in detail.
In the North American bat species M. lucifugus, we also observed 157 matched protein spots, of which eleven protein spots (7%) showed a significant (p < .05) differential expression with a minimum of a 1.5-fold difference between healthy and diseased conspecifics.
We detected a significant upregulation of eight protein spots and a significant downregulation of three protein spots in WNS-positive M. lucifugus compared to healthy conspecifics ( Figure 1 and Table 1).
We were able to pick and identify eight out of the 11 proteins with differential expression profiles (Ml1, Ml2, Ml3  These findings support an active host response against Pd on a systemic level and concur with transcriptomic findings at the site of infection of WNS-affected M. lucifugus  and previous functional studies ( Moore et al., 2011Moore et al., , 2013. Pd infection in M. lucifugus shows similarities to other fungal skin infections in euthermic animals . Host response against fungi is initiated by the recognition of fungal proteins by epithelial and innate immune cells resulting in the activation of the immune system through the production of inflammatory cytokines (Romani, 2011). In euthermic animals, the production of proinflammatory cytokines activates the acute phase response (APR), a systemic immune reaction characterized by the production of acute phase proteins (APPs), fever, leucocytosis (i.e., increase in white blood cell counts) and the display of sickness behaviour (anorexia, lethargy; Cray, Zaias, & Alman, 2009;Schneeberger, Czirják, & Voigt, 2013).
TA B L E 1 Differential protein expression in the Myotis lucifugus plasma proteome. Protein spots exhibiting differential expression (p < .05; minimum fold-change 1.5) using Delta2D software are shown. Fold change reflects differences in protein spot volume comparing healthy to diseased individuals. Identified proteins via LC-MS/MS are displayed with the total MS protein score based on MASCOT searches on NCBI database. Protein IDs listed are the top 5 ranked protein matches based on the MASCOT score excluding protein matches of serum albumin. Spots Ml6, Ml7, and Ml11 were not distinguishable on the preparative gels, thus could not be picked and identified the site of Pd infection despite the apparent expression of inflammatory cytokines and chemokines . We identified several APPs, such as A2M-like, TF and protein chains of PF (α, β, γ), in the majority of differentially regulated protein spots in M. lucifugus plasma. The physiological function of APPs is to reestablish homeostasis, promote healing, kill different pathogens and mitigate damage by reactive oxygen species. They also have important roles in the promotion of the host's adaptive immune response (Cray et al., 2009). Iron-binding glycoprotein TF helps maintain iron homeostasis, which is altered during infections and which influences modulation of innate immune defences and prevention of pathogen survival (Ganz & Nemeth, 2015). It is functionally similar to haptoglobin, another important APP in bats (Costantini, Czirják, Bustamante, Bumrungsri, & Voigt, 2019;Fritze et al., 2019), which has been identified in the wing tissue transcriptome of WNS positive M. lucifugus . As a major protease inhibitor, A2M is involved in regulating inflammatory processes by scavenging defensins and inhibiting proteases of host and nonhost origin, which protects inflamed tissues against excessive damage (Rehmann, Ahsan, & Khan, 2013). Furthermore, glycoprotein PF has been described as having regulatory properties during inflammation in different tissues (Davalos & Akassoglou, 2012).
Upregulated C3 complement component, together with other complement factors, is part of the constitutive innate immune system, considered to be the first line of defence against pathogens (Becker, Czirják, Rynda-Apple, & Plowright, 2019). C3 has a central role in the complement system, as it is involved in phagocytosis, inflammation and cell lysis. It is required in both classical and alternative complement activation pathways (Sarma & Ward, 2011).
Complement-associated defence is one of the few parts of the immune system that is not affected by hibernation (Maniero, 2002).
Increased C3 production may represent an attempt to control Pd infection using one of the only available active immune components during hibernation.
The observed differential regulation of innate immune proteins   and showed altered complement activity (Moore et al., 2011). An activated systemic immune response may also explain the detection of upregulated gene product IGLL5 associated with different immune functions in cancer (Ascierto et al., 2013;White et al., 2018) and downregulated IGHV4OR15-8-like protein, a gene product of the immunoglobulin heavy chain variable region associated with immune response (Dammalli et al., 2017;Matsuda et al., 1998).
Activation of the immune system by fungal and bacterial pathogens initiates phagocytosis associated with the production of reactive oxygen species (ROS) (Brown, 2011). Increasing level of ROS during host defence leads to oxidative imbalances in affected tissues and thus an upregulation of oxidative stress markers (Schneeberger et al., 2013). For Pd infected North American bats, it was observed that the expression of genes related to oxidative stress changes  and that the total antioxidant defence of animals was compromised (Moore et al., 2013). We found upregulated plasma proteins DJ-1-X1, 6PGD and downregulated AMBP-X1, which are associated with host protection during oxidative stress. LPSinduced inflammatory conditions in mice showed an upregulation of protein DJ-1 in response to ROS (Mitsumoto & Nakagawa, 2001).
One of the energetic costs of the APR during infection is changed lipid metabolism (Khovidhunkit et al., 2004). The identified upregulated plasma proteins APO-AI and PON-X1 are associated with the high-density lipoprotein complex, which is essential for cholesterol trafficking between peripheral cells and the liver (Tall, 1990). An increased level of lipoproteins in WNS-infected M. lucifugus may suggest changed triglyceride metabolism. Gene expression of apolipoproteins has also been found to be increased in infected wing membranes .
Other differentially expressed proteins identified in Pd-infected M. lucifugus plasma included SIAT8-F and MAT2, which are associated with metabolic cycles of oligosaccharides (Wang, Liu, Wu, & Sun, 2016) and the amino acid methionine (Finkelstein & Martin, 2000). Downregulated desmosome proteins included DP and JPG-X1, which are both involved in cell signaling (Johnson, Najor, & Green, 2014). Similarly, genes involved in metabolic and cell signaling pathways were altered at the transcriptome level displaying a changed metabolism most probably associated with disturbance of host homeostasis due to WNS Verant et al., 2014).
European M. myotis in contrast with North American M. lucifugus did not exhibit any changes in proteomic profile in response to Pd infection. This suggests that M. myotis tolerates Pd colonization. Tolerance in a host-pathogen system is context dependent and influenced by both host and pathogen factors (King & Li, 2018;Mandl, Schneider, Schneider, & Baker, 2018). Pd in Europe and North America are genetically very similar (Drees et al., 2017), suggesting that Pd was recently introduced to North America (Leopardi, Blake, & Puechmaille, 2015). Therefore, pathogen factors probably play little role in the differences observed in host mortality and host response between North American and European bats. The lack of response of M. myotis to Pd probably represents a long-term co-existence of pathogen and host, which has resulted in a balance that yields low rates of host mortality.
The change in proteomic profile of North American M. lucifugus and extremely high mortality rates suggests this balance has not been reached in North American bats.

ACK N OWLED G EM ENTS
We are grateful to Katja Pohle for her crucial fieldwork and laboratory assistance, to Heike Stephanowitz for carrying out the MS analysis, to Arturo Zychlinsky for guidance on the initial study design and discussion on the first preliminary findings and to