Life cycle studies and transmission mechanisms of myxozoan parasites

Den Naturwissenschaftlichen Fakultäten der Friedrich-Alexander- Universität Erlangen-Nürnberg zur Erlangung des Doktorgrades

vorgelegt von

Dennis Marc Kallert

aus Erlangen

Als Dissertation genehmigt von den Naturwissenschaftlichen Fakultäten der Friedrich-Alexander-Universität Erlangen-Nürnberg

Tag der mündlichen Prüfung: 3. März 2006

Vorsitzender der Prüfungskommission: Prof. Dr. D.-P. Häder

Erstberichterstatter: Prof. Dr. W. Haas

Zweitberichterstatter: Prof. Dr. Dr. M. El-Matbouli -Contents-

Contents

ABBREVIATIONS 5

I. INTRODUCTION 6

1. The Phylum Myxozoa Grassé 1970: Common features 6

2. Current knowledge on myxozoan life cycles 8

3. Host invasion by actinospores 10

II. MATERIALS & METHODS 12

1. Animals and parasite cultivation 12 1.1. Fish for experimental infections 12 1.2. Oligochaetes 12 1.3. Parasites 13 1.4. Animals for substrate isolation 14

2. Experimental infections 14 2.1. Henneguya nuesslini Schuberg & Schröder 1905 14 2.2. Myxobolus parviformis sp.n. 16

3. Viability assay 17

4. Discharge experiments 18 4.1. Test substrates 18 4.2. Experimental set-up 19 4.2.1. Test for mechanical and chemical stimuli 19 4.2.2. Bulk experiments 20 4.2.3. Frequency dependency 21 4.2.4. Ca2+ dependency 22

5. Cinematography of polar filament discharge 22

6. Sporoplasm emission 22

7. Mucus fractionation 23 7.1. Heat and incubation 23 7.2. Ashing 23 7.3. Acetone precipitation 24

- 1 - -Contents-

7.4. Alcian blue precipitation 24 7.5. Fluorescamin derivatisation 24 7.6. Extraction with activated charcoal 25 7.7. Lipid extraction 25 7.7.1. Ether-extraction 25 7.7.2. Lipid isolation 26 7.8. Chemical fractionation 26 7.8.1. TFMS Deglycosylation 26 7.8.2. Sialic acid extraction 27 7.9. Enzymatic fractionation 28 7.10. Chromatographic fractionation 29 7.10.1. Ion-exchange 29 7.10.2. Lichroprep RP 18 30

8. Pure chemicals 30

9. Chemical analyses 32 9.1. Proteins 32 9.2. Amino compounds 32 9.3. Neutral sugars 32 9.4. Urea 32 9.5. Sialic acids 33 9.6. TLC-separation of lipid classes 33 9.7. Hydrophilic TLC 34 9.8. RP-HPLC 35 9.9. GC/MS 35 9.10. UV-Spectroscopy and fluorescence detection 36 9.11. HPIC-IPAD and HPLC-MDD 36 9.12. NMR-spectroscopy 37

10. Chemicals 38

11. Statistical methods 38

III. RESULTS 39

1. Life cycle experiments 39 1.1. Henneguya nuesslini 39 1.2. Myxobolus parviformis sp. n. 43

2. Viability assay 51

3. Polar filament discharge 51 - 2 - -Contents-

3.1. Visualisation of polar filament discharge 51 3.2. Mechanical and chemical stimuli 53 3.3. Experimental conditions for bulk experiments 53 3.4. Ca2+ dependency 54 3.5. Host specificity 55 3.5.1. Myxobolus cerebralis 55 3.5.2. Henneguya nuesslini 58 3.5.3. Myxobolus parviformis 58 3.6. Analysis of chemical signals for polar filament discharge 59 3.6.1. Ultrafiltration 59 3.6.2. Lipids 60 3.6.3. Amino compounds 61 3.6.4. Nucleotides 62 3.6.5. Carbohydrates 63 3.6.6. Proteins 68 3.6.7. Stability and small molecular compounds 70 3.6.8. Extraction with activated charcoal 72 3.6.9. Chromatographic fractionation 72

4. Sporoplasm emission 77 4.1. Emission time course 77 4.2. Dependence from discharge 78 4.3. Host specificity 78 4.4. Emission stimuli 78

5. Analyses 79 5.1. Substrate osmolality 79 5.2. Urea 80 5.3. Lipids 81 5.4. Hydrophilic TLC Analyses 83 5.5. Gradient chromatography 87 5.6. UV-Spectroscopy 89 5.7. HPIC-detection 89 5.8. Amino acids 91 5.9. Proteins 92 5.10. Carbohydrates 93 5.10.1. Neutral sugars 93 5.10.2. Sialic acids 94 5.10.3. GC-MS for monosaccharides 94 5.11. NMR-spectroscopy 95

- 3 - -Contents-

IV. DISCUSSION 96

1. Myxozoan lifecycles 96

2. Host invasion by actinospores 103

3. Host signals for polar filament discharge 107

4. Impacts on myxozoan transmission 113

V. SUMMARY 116

VI. ZUSAMMENFASSUNG 118

VII. ACKNOWLEDGEMENTS 120

VIII. REFERENCES 121

IX. APPENDIX 130

- 4 - -Abbreviations-

Abbreviations

1 J One bond AA Amino acid ATP Adenosine triphosphate BSA Bovine submaxillary mucin cAMP Cyclic adenosine monophosphate cGMP Cyclic guanosine monophosphate cIMP Cyclic inosine monophosphate

D2O Deuterium oxide df Degrees of freedom FDA Fluorescein-diacetate GAGs Glycosaminoglycans GC/MS Gas chromatography/mass spectroscopy H & E Haematoxylin/eosin dye HMQC Hetero multiple quantum coherence HPIC-IPAD High performance ion exchange chromatography with integrated pulsed amperometric detection HPLC High performance liquid chromatography l Litres MDD Metal dye detection MW Molecular weight MWCO Molecular weight cut-off NBD-Cl 4-chloro-7-nitrobenzofurazan NMR Nuclear magnetic resonance p.e. Post exposure p.i. Post infection PBS Phosphate buffered saline PCR Polymerase chain reaction PI Propidium iodide RFLP Restriction fragment length polymorphism SD Standard deviation SDBS Data base system for organic compounds SEM Standard error of the mean SFW Standard fresh water TAM Triactinomyxon TFMS Trifluoromethanesulfonic acid TLC Thin-layer chromatography v/v Volume per volume w/v Weight per volume

- 5 - -I. Introduction-

I. Introduction

1. The Phylum Myxozoa Grassé 1970: Common features

About 1350 species in 52 genera belong to the Myxozoa, an obligate parasitic group forming a separate phylum of multicellular metazoan parasites mainly of teleosts. Invertebrates like oligochaetes, bryozoans and polychaetes serve as secondary hosts. Despite being well-known as fish parasites, Myxozoa was also discovered in trematodes (Overstreet 1976, Siau et al. 1981), reptiles (Lom 1990) and amphibians (McAllister et al. 1995). Developmental stages were found in waterfowl (Lowenstine et al. 2002), in nervous systems of mammalians (Friedrich et al. 2000) and myxospores were even detected in human feces (Lebbad and Wilcox 1998, Moncada et al. 2001). The members of the most abundant genus, Myxobolus (Myxobolidae), have recently been reviewed by Eiras et al. (2005). Myxosporea were first described 1841 by Müller (named ‘Psorospermien’) and Štolc (1899) already noted the metazoan character of these organisms. Nevertheless, until the second half of the last century, the Myxozoa were commonly assigned to the protozoans. Due to the morphological variability within species and their highly reduced body organization, the taxonomy and the phylogenetic position of these obscure parasites are still subject of numerous controversies (Kent et al. 1994, 2001, Siddall et al. 1995, Anderson et al. 1998, Canning & Okamura 2004). An outgrowth of the ongoing phylogenetical work was the development of comprehensive PCR-based assays suitable for diagnostics. Although rather few species exert problematic symptoms, some members are severe pathogens of teleosts. Several species have a significant ecological and economic impact on freshwater and marine fish populations in Europe and the USA. Whirling disease, caused by the cosmopolitan parasite Myxobolus cerebralis, is still considered as one of the most devastating diseases among salmonid populations (Nehring & Walker 1996, Gilbert and Granath, 2003). It has been responsible for declines of the wild rainbow trout population in more than 22 northern american states (Hedrick et al. 1998). Other important pathogens of cultured fish include Tetracapsula bryosalmonae (proliferative kidney disease of salmonids) and Sphaerospora renicola (swimbladder inflammation of carp). Myxospores have been shown to resist freezing and passage through alimentary tracts of poikilothermic animals (e.g. birds, see El-

- 6 - -I. Introduction-

Matbouli and Hoffmann 1991) and are therefore highly adapted to environmental changes. Morphologically, the myxozoans share some unique features defining phylum affiliation. The class Malacosporea, probably an ancestral myxozoan group with few species described that partially resemble different (nemathelminth) characteristics, will be excluded in this study. Myxosporean specimens are uniformly small sized (myxospores 10-25 µm, actinospores up to 300µm), thereby showing an immense biodiversity in shape and morphological variation. Actinospores differ from myxospores by their triradial symmetry and the softer valves. Developmental stages lack cilia and cellular fission seems to occur without centrioles. Especially peculiar is the development including secondary and tertiary cell stages formed by endogeny during proliferation, a feature that is very rare in animals and otherwise found only in some protozoans. Most denominating are the polar capsules, specialised cell organelles reminiscent of cnidarian nemtocysts that are used for attachment to the host. Though recent findings revealed strong evidence for triploblast bilateral ancestors of the myxozoan phylum (Schlegel et al., 1996; Anderson et al. 1998; Okamura et al., 2002), there are arguments for a cnidarian origin of the polar capsules (discussed in Canning and Okamura, 2004). For example, the polar filaments were compared to the stinging threads of the parasitic cnidarian Polypodium hydriforme by Ibragimov (2001).

Figure 1. Schematical structure of a myxozoan triactinomyxon-type actinospore.

- 7 - -I. Introduction-

2. Current knowledge on myxozoan life cycles

About 25 whole myxozoan life cycles have been elucidated today (see Kent et al. 2001). With few exceptions, they include two alternating hosts, an aquatic invertebrate (oligochaetes or bryozoans) and a vertebrate host, mainly teleost fish (Wolf & Markiw, 1984). Life cycle studies of myxozoans are difficult to conduct, most research has been carried out with M. cerebralis, which is available through continuous cultivation (El-Matbouli et al. 1992, Hedrick & El-Matbouli 2002). The complete chronological development of a myxozoan has only been obtained for M. cerebralis (El-Matbouli & Hoffmann 1989, El-Matbouli et al. 1995). Waterborne actinospores (Fig. 1) released from oligochaetes infect the fish via gills or skin. Amoeboid sporoplasms containing the infective secondary cells leave the actinospore valve construct and actively penetrate the host integument. In the respective target tissue, they develop from trophozoits into sporogenic plasmodia. Coelozoic or histozoic parasitism occurs mostly extracellular, while some species penetrate host cells for multiplication. Some myxosporeans form extrasporogonic (e.g. blood-) stages and multiply several times by internal cleavage. Most species tend to infect specific kinds of organs, tissues or cell types (Molnar 2002, Eszterbauer 2004). M. cerebralis developmental stages migrate through several tissues along peripheral nerves to reach the cartilage (El-Matbouli et al. 1995). During growth, somatic and generative nuclei are found. At the end of the intrapiscine development, myxospores composed of at least six cells are formed shedding at least two rigid (Myxosporea) or soft (Malacosporea) protective shell valves. They are formed in pansporoblasts inside plasmodia or develop in coelozoical or intercellular in pseudoplasmodia. Although the existing knowledge on transmission is limited, mxospores are thought to be released from the fish hosts by death of the host (e.g. predation), but there is an indication that release may also occur from live fish by Myxobolus artus (Ogawa et al. 1992) and most likely some gill-dwelling species. They infect the intestine of invertebrate hosts by ingestion. Vegetative and sporogonic stages of actinosporeans are located intercellularly in the intestinal epithelium or in the coelom. Actinospores released from the oligochaete host inflate to develop their final habitus and initiate a new development in fish (Fig. 2). The release of myxo- and actinospores is often seasonal; therefore the whole life cycle may last for 1-2 years although the

- 8 - -I. Introduction-

Figure 2. Simplified schematical life cycle of M. cerebralis showing important developmental processes of both transmission stages in each host. Multiple arrows indicate parasite multiplication. development takes only about 2 months in each host (El-Matbouli et al., 1992, 1995; Molnár, 1994; El-Matbouli & Hoffmann, 1998). Factors shown to influence myxozoan development include temperature and fish age. The majority of myxozoans is described to be rather specific regarding susceptible teleost host taxa (species, genus or family specificity) (Lom & Dyková 1992) and the oligochaete hosts (e.g. Stevens et al. 2001). With 11 susceptible salmonid species, M. cerebralis is an example for a polyxenous taxon, Enteromyxum leei (Padrŏs et al. 2001) and some Sphaerospora spp. are restricted to a single teleost species. On the other hand, susceptibility among salmonids to M. cerebralis and Ceratomyxa shasta is determined by strain (Bartholomew 1998, Hedrick et al. 1999, Thompson et al. 1999). Although these obvious borders in susceptibility exist, there is no information about their origin yet.

- 9 - -I. Introduction-

3. Host invasion by actinospores

With the growing importance of fish diseases caused by Myxozoa, studies of the life-cycles, parasite-host-relationships and developmental mechanisms of the phylum resulted in a better understanding of life history traits. Nevertheless, the mechanisms of host recognition and invasion of this successful group are to a large extent unclear (Yokoyama 2003). The passively floating actinospores play a key role in the spread of the infective stages, but myxozoan actinospores do not remain infective for as long as myxospores, which are able to retain their infectivity for more than 20 years (El- Matbouli et al. 1992). Infectivity of actinospores is assumed to last only for several days (Yokoyama et al. 1993, Xiao & Desser 2000b). Waterborne M. cerebralis actinospores have to invade their host in less than 60 h after release (Markiw 1992), and thus would benefit from functional adaptations for adequate host recognition. Host attachment is mediated by the extrusion of an eversible tubule from apically located polar capsules (Fig. 1). The capsulogenic cells are tightly joined with the valve cells through septae complexes or gap/tight junctions. The mouth of the polar capsule is capped with a plug-like structure that closes the opening of the invaginated polar filament. The discharge of the polar filaments anchors the actinospore to the fish surface. Following attachment, the infective amoeboid sporoplasm actively penetrates the host integument, preferably through the opening of mucous cells in the skin (El-Matbouli et al. 1999). Within one min after exposure, M. cerebralis sporoplasms were detected in the host epidermis (El-Matbouli et al. 1999a). The physiological mechanisms and host cues that initiate the invasion reactions have long been a matter of speculation and remain unknown (Kent et al. 2001, Yokoyama 2003). Gills, skin and the buccal cavity were confirmed as portals for entry for M. cerebralis, while Thelohanellus hovorkay prefers entry through the gills (Yokoyama & Urawa 1997) and Henneguya ictaluri infects via the intestine (Yokoyama 2003). Oral uptake of infected oligochaetes may also serve as a route for infection in some species. For actinospores, it would be of great benefit to avoid interference (esp. polar filament discharge) not only with incompatible hosts, but also with other aquatic organisms and dead organic matter frequently encountered. Some researchers could show reactivity of several actinospore types to fish mucus (Yokoyama et al. 1993;

- 10 - -I. Introduction-

Uspenskaya, 1995; Yokoyama et al. 1995a&b, Yokoyama & Urawa 1997; McGeorge et al. 1997; Xiao & Desser, 2000; Ozer & Wootten 2002). Such investigations draw an ambiguous picture on parasite-derived host preference upon exposure to mucus of different fish (sporoplasm emission was evaluated in most cases). However, no consistent assay was employed and thus an indication towards host specificity and the involved stimuli was not possible. There have been indications of host specificity in the very first steps of fish host invasion by M. cerebralis, based on the absence of parasite stages in histological examinations and unsuccessful infection in various experimentally exposed non-salmonids (El-Matbouli et al. 1999). Unfortunately, M. cerebralis actinospores could not be stimulated to discharge filaments or release their sporoplasms when incubated with salmonid mucus or epidermal tissue (El-Matbouli et al., 1999; Wagner, 2001a). Up to date, there are no hints whether host specificity regarding the teleost host is a result of parasite choice upon contact, parasite adaptation to the host immune system or adaptation by the host. So far it was proposed, that actinosporeans specifically recognise their respective hosts. The search for the discharge triggering host signal for actinospores has resulted in some effort to isolate such compounds from fish skin mucus. Yokoyama et al. (1995a) reported that a low molecular fraction (< 6 kDa) from Carrasius auratus mucus caused actinospores of M. cultus to release their sporoplasms. A random screening of chemicals that discharge nematocysts in order to find host-related substances which stimulate the discharge reaction of actinospores was unsuccessful (Wagner, 2001b). Furthermore, discharge reactions of both spore types to various chemicals were reviewed and compared to cnidae of coelenterates by Cannon & Wagner (2003), including morphological aspects. However, no convenient answers could be given on the naturally occurring interactions between the parasites and the host, especially the polar filament discharge reaction under physiological conditions. The present thesis should serve as a step forward for a better understanding by which means myxozoan parasites ensure transmission success during the difficult task to recognise and invade new fish hosts by actinospores. Parts of the study investigated whether the invasion behaviour of actinospores is initiated by a host- specific attachment reaction by actinospores. For this purpose, different myxozoan species were included in the experiments, enabling a more general statement on host specificity. Therefore, life cycles of two species were described initially during the study. The main focus of the work was to identify host cues that trigger polar filament

- 11 - -I. Introduction- discharge and sporoplasm emission. The dependence of polar filament discharge on mechanical and mucus-derived chemical cues was investigated and experimental procedures were developed to record polar filament discharge rates. Discharge triggering host signals for M. cerebralis actinospores were isolated by extensive biochemical fractionation and analysis of surface mucus from susceptible fish species.

II. Materials & Methods

1. Animals and parasite cultivation

1.1. Fish for experimental infections

Four to seven month old specimens (3-4 cm) of brook trout Salvelinus fontinalis and brown trout Salmo trutta were obtained from a local hatchery (Bavaria, Germany) and used for infection in the Henneguya nuesslini transmission experiments. The fish were raised parasite free by spring water in plastic containers. Common carp (Cyprinus carpio) fingerlings (2-4 cm) for infection experiments originated from a federal hatchery in Bavaria (Germany). Adult common bream (Abramis brama, 30-45 cm), naturally infected with gill developing Myxobolus spp., were caught by line in the river Aisch near Höchstadt, Bavaria (Germany). Two year old specimens (7-10 cm) of common bream for infection were collected from a carp-rearing pond and were kept at 18 °C in 50 l plastic tanks without a flow-through system. Four bream specimens were dissected before use to check for possible enzootic gill infections by myxozoans and they were all negative. All fish were kept at 16-18 °C under a constant flow of well water and fed on conventional fish flakes.

1.2. Oligochaetes

For the H. nuesslini infections, naturally infected oligochaetes were sampled from a mud deposition pond of the salmonid hatchery. The mud cultures containing oligochaetes were kept in aerated tubs (300-500 l) under constant flow of well water with a mean temperature of 12 °C. For infection, aliquots of these were transferred in 10 l plastic beakers with a bottom mixture of coarse cooked sand and sterilised mud (1:1) without a flow-through system.

- 12 - -II. Materials and Methods-

For the M. parviformis sp.n. infections, commercially obtained mixed tubificid oligochaetes (from a pet shop) were kept in aerated 4 l water tanks with a bottom substrate of washed coarse grained sand at 16-18 °C and constantly monitored for myxozoan infections for over one year. To obtain naïve oligochaetes, offspring from these cultures were collected manually by sorting out cocoons and freshly hatched oligochaetes, which then were raised in a separate container (500 ml) with sterile sand as bottom substrate. For M. cerebralis cultivation, laboratory infected Tubifex tubifex cultures were obtained from the Institute of Zoology, Fish Biology and Fish Diseases of the University Munich (Germany). Alternatively, aliquots of oligochaetes (> 200 000) from the salmonid hatchery were transferred to aerated plastic beakers (4 l) and used for infection after a monitoring period. The infected oligochaete cultures were kept at 12- 15 °C in the dark. All oligochaete cultures were provided a mixture of frozen Artemia, spray-dried Spirulina (MaBitech) and frozen lettuce as food weekly. Minced and cooked horse dung and cooked mud from the hatchery outlet was given occasionally. Only copper and chlorine free well water was used for oligochaete cultures.

1.3. Parasites

H. nuesslini originated from naturally infected oligochaetes from a salmonid hatchery in Aufseß (Bavaria). Myxospores were either obtained from muscle-tissue during transmission experiments or (for bulk infections) cut tailpieces of large, naturally infected brown trout specimens from a pond of the hatchery. Myxospores were isolated by homogenisation for 2 - 5 min at 11500 rpm using an Ultra Turrax (IKA Labortechnik, Staufen, Germany). Remnants were then removed by passing the homogenate through 100 µm mesh filters and the suspension was allowed to sediment at 4 °C overnight. M. parviformis originated from naturally infected bream from the river Aisch. To harvest myxospores, infected bream were anaesthetised by a sharp blow on the cranium, killed by spinal severance and gill arches and lamellae were homogenized after microscopic detection of pseudoplasmodia under addition of small amounts of tap water. Single pseudoplasmodia from the repeated infection were thoroughly dissected from gill lamellae under a stereomicroscope using sterile needles for

- 13 - -II. Materials and Methods- individual screening. They were transferred to Eppendorf tubes with 30 µl of tap water and could be stored for several days at 6 °C. Laboratory cultures of Tubifex tubifex that were infected with M. cerebralis originated from the Institute of Zoology, Fish Biology and Fish Diseases of the Ludwig Maximilian University Munich, Germany. Besides, hatchery oligochaete cultures were infected by the addition of 2-3 Mio M. cerebralis myxospores after incubating the oligochaetes at 25 °C for at least one week. Non-exposed control cultures did not produce actinospores. Triactinomyxon spores (TAMs) were filtered through 20 µm nylon mesh or, during the life cycle experiments, obtained from single oligochaetes isolated in cell well plates according to Yokoyama et al. (1991).

1.4. Animals for substrate isolation

Adult rainbow trout (Oncorhynchus mykiss) and carp were obtained from a local distributor, and common bream were caught in the river Aisch. Lymnaea stagnalis originated from a laboratory stock. Frogs (Rana esculenta) were caught from the field during a former project around the institute facilities.

2. Experimental infections

2.1. Henneguya nuesslini Schuberg & Schröder 1905

Brown and brook trout were infected in a first trial by cohabitation as groups of five in the oligochaete containers for five days. In a second trial, the fish were individually incubated overnight in 200 ml aerated water containing 1500 to 2000 actinospores collected from isolated oligochaetes, shedding only actinospores that showed the characteristic morphological structures described below. Altogether, 25 fish of each species were used in the first trial, 15 in the second trial and 20 specimen served as controls. Common carp (Cyprinus carpio) fingerlings (2-4 cm) were exposed as in the second trial. Fish were examined from day 80 post-infection (p.e.). Muscle tissue, heads, spinal cord and fins were homogenised separately for 20 min at 11 500 rpm using an Ultra Turrax. Ten carp fingerlings were completely homogenised. Tissue remnants were removed by passing homogenates through 100 µm mesh filters and the suspension

- 14 - -II. Materials and Methods- was then allowed to settle at 4 °C overnight. For histological studies (conducted at the Institute of Zoology, Fish Biology and Fish Diseases of the LMU Munich), samples of infected fish were fixed in 5% formalin, embedded in paraffin, sectioned at 3-4 µm, stained using haematoxylin and eosin (H & E) and mounted in Eukit (Kindler, Freiburg, Germany). Sections were screened by light microscopy for myxozoan infections. Photomicrographs and measurements were taken from a series of digital high- resolution data sets of fresh material obtained with an Axiophot (Zeiss, Jena, Germany). Oligochaete hosts were identified by fixation of four infected tubificid host specimens in Bouin´s fluid, dehydrated in xylol, epon-mounted on microscope slides, and examined under a light microscope. To supplement this information, a fifth individual showing partial maturity was preserved in 80% ethanol and divided into two parts. The anterior part (including the genital segments) was cleared in glycerine under a cover-slip and examined microscopically. DNA was then extracted from the posterior part of the oligochaete and a fragment of the mitochondrial 16S rDNA gene (520 bp) was amplified and sequenced by Prof. C. Erséus at the Department of Zoology of the University of Göteborg (Sweden) as described in Kallert et al. (2005a). The sequence obtained was aligned with those of Tubifex tubifex, Potamothrix hammoniensis and Ilyodrilus templetoni obtained from tubificid cultures belonging to T. Timm (Võrtsjärv Limnological Station of the Institute of Zoology and Botany of Rannu, Tartumaa, Estonia), using Clustal X, version 1.8 (Thompson et al. 1997), with default settings; the latter three oligochaetes, originally identified by Timm, were previously sequenced by C. Erséus (unpublished data). For molecular identification of the parasite, DNA was extracted from fish tissue homogenates containing myxospores and from actinospore samples collected by the cell well method. DNA amplification by a nested PCR system (SphF-SphR followed by SphF-MB5r and MB5-SphR primers, see Table 1, IX. Appendix) and sequencing after cloning of 18S fragments was performed by Dr. Eszterbauer at the Veterinary Medical Research Institute of the Hungarian Academy of Sciences (Budapest, Hungary) as described in Kallert et al. (2005a).

- 15 - -II. Materials and Methods-

2.2. Myxobolus parviformis sp.n.

Wild bream initially harboured at least three Myxobolus morphotypes in their gills, but those differing from the herein described were not further investigated. Large fused pseudoplasmodia (length up to seven mm) from gill lamellae were excised and ruptured. The isolated myxospores were kept in 5 ml copper-free well water that was exchanged every two days. Myxospores obtained from naturally infected bream gills were used for initial oligochaete infection. An oligochaete culture was divided into two equal parts (~2500 individuals each) one of which was exposed to a pool of myxospores (> 4 million), while the second part served as control. The water of all oligochaete cultures was not exchanged for 3 weeks after exposure. Actinospore producing oligochaetes were sorted by the cell well method (Yokoyama et al. 1991) based on actinospore morphological features (proportions, number of sporoplasm cells). Triactinomyxon spores with the herein specified morphological features were used for infection of 5 young bream (pure infection). To obtain a higher infection intensity for further parasite cultivation (bulk infection), ten bream specimens were exposed by repeated addition of an actinospore suspension from whole culture filtrates containing > 3000 TAMs to the aquaria every third day for 3 weeks (cohabitation) and further ten fish by individual incubation for 3 h in 3 l filtrate containing 1500 to 2000 actinospores. Ten fish served as controls. Gill arches and gill lamellae of the five specimens (pure infection) were excised and myxospores extracted as described above. For the second trial, myxospores from selected single pseudoplasmodia that originated from the first trial were screened for the morphological features presented in this study. Matching isolates were partially used for PCR-RFLP detection and compared to the actinosporean counterpart. Their remnants were pooled and used for experimental infection of a naïve oligochaete culture (pure oligochaete infection) at an estimated dose of 1000 spores per oligochaete. For bulk infection of oligochaetes (derived from the former control group), one culture (~2000 oligochaetes) was inoculated with total gill homogenate. In this trial, only actinospores from the ‘pure oligochaete infection’ were used for further infection of three bream specimens (mode of infection was cohabitation). The basic

- 16 - -II. Materials and Methods- steps for the isolation of a myxozoan species from sympatric species as performed in this part of the study are shown in Fig 2 (Appendix) schematically. From all stages that were used for infection, PCR samples from filtrates or tissue homogenates were prepared by centrifugation (10000 g, 15 min) and mixing of the pellet with ethanol (> 70% v/v). A myxospore sample was prepared by pooling four individually excised pseudoplasmodia from bream of the first trial, which revealed mature myxospores with identical M. parviformis morphometrics. For parasite identification, a nested PCR amplification (18e-18g’ universal primer pair and second round PCR with the MX5-MX3 primer pair, see Table 1, IX. Appendix) followed by PCR-RFLP (restriction enzymes HinfI, MspI and TaqI) was conducted by Dr. Eszterbauer as described in Kallert et al. (2005b). Purified PCR fragments of a pooled TAM sample from the first infection trial were then cloned and sequenced in both directions as already described (Kallert et al. 2005b). For identification of oligochaete hosts, three infected tubificid host specimens were fixed in Bouin’s fluid, two other oligochaetes were preserved directly in 80% ethanol. The posterior end of one ethanol-preserved individual was set aside for DNA extraction, while the rest of the material, including the anterior end of the partitioned oligochaete, were mounted on microscope slides and examined under a light microscope. DNA was then extracted from the selected oligochaete and a fragment of the mitochondrial 16S rDNA gene, about 520 bp long, was amplified and sequenced by Prof. C. Erséus as described in Kallert et al. (2005b). The sequence obtained (GenBank Accession No. AY836151) was aligned with those of 14 other representatives of the subfamily Tubificinae (covering seven different genera) as described for the H. nuesslini infection experiments. Other oligochaete sequences for comparisons were previously obtained by C. Erséus (unpublished information).

3. Viability assay

To assess the physiological condition and infectivity, the viability status of the cellular components of M. cerebralis actinospores (up to 24 h of age) used in the discharge experiments was exemplarily assayed by fluorescein-diacetate/propidium iodide (FDA/PI) double staining as described by Yokoyama et al. (1997) and Markiw (1992) from spore isolates in three different weeks.

- 17 - -II. Materials and Methods-

4. Discharge experiments

4.1. Test substrates

Fish were killed by a sharp blow to the head and mucus was scraped off with a blunt knife while rubbing small amounts of deionised water onto the fish surface. The collected mucus was homogenised using an Ultra Turrax and suspended by sonication for 10 min by an ultrasonic processor (VP 50H, Dr. Hielscher GmbH) (amplitude 80%, cycle 0.8). Centrifugation at 2200 g and 4 °C for 10 min to remove insoluble components yielded a fully soluble supernatant. The supernatants, hereinafter referred to as ‘mucus homogenate’, were stored at -75 °C. When necessary, the substrate was concentrated by removing water from the solution by partial lyophilisation with a SpeedVac (Vacuum Concentrator, Bachofer). Mucus of Lymnaea stagnalis was collected by scraping the walls of a small PVC box containing 50 large snails in 15 ml water for up to 3 h, the remaining faeces being removed by one additional centrifugation. Frog epidermal extract was obtained by pulverising Stratum corneum patches of freshly killed Rana esculenta specimens in liquid nitrogen, followed by homogenisation after addition of 0.4 times the wet weight of deionised water. Trout muscle tissue homogenate was prepared from muscle sections of fresh dissected fish (0.6 g/ml deionised water), the concentrated muscle tissue solutions (5 mg/ml final concentration) were desalted as described below. The substrates were further treated as described for the mucus homogenates. The concentration of the homogenates was adjusted referring to the initial dry weight after complete lyophilisation. Different molecular size fractions were obtained by ultrafiltration at 4 °C and 2200 g using Centricon or Centriprep concentrators (Amicon) with a cut-off (MWCO) at 3 kDa. The retentates were washed with at least half the original filling volume of deionised water to remove low molecular weight remnants. To adjust mucus isolates to the same carbohydrate content, neutral sugar concentration was measured by the H2SO4/Resorcinol-method of Monsigny et al. (1988) as described under 12.3. Substrate osmolality was determined by freeze-point lowering using an osmometer (Osmometer automatic, Knauer) with two-point calibration in triplicates.

- 18 - -II. Materials and Methods-

4.2. Experimental set-up

4.2.1. Test for mechanical and chemical stimuli

By observing the responses of single M. cerebralis actinospores, the influence of mechanical stimulation on polar filament discharge was examined before and after subsequent treatment under a stereomicroscope. The spores were immobilised by sucking in one caudal process into the opening of a glass capillary on a moulded slide (Fig. 3). Test and control substrates were applied directly covering the whole spore by two separate Hamilton glass syringes (10 µl volume). All elements were movable by micromanipulators. The mechanostimulus was administered by gently touching the apex three times at its polar capsule bearing apical ending with the tip of a stainless steel needle (0.5 mm diameter). Each actinospore received a mechanical stimulus after application of 1.5 µl control substrate (buffered deionised water). Then 1.5 µl crude trout mucus were added, followed by another mechanostimulus. After each stimulation-step (chemical and mechanical), the actinospores were checked for extruded filaments for 10 s and considered discharged if at that time one of the three filaments was extruded. In a control trial, control water was offered as second chemostimulus prior to the second mechanostimulation to exclude an effect of repeated stimulation. The experiment was conducted with actinospore isolates from 10 different days, each replicate was only included if the specimens remained intact throughout the whole procedure.

- 19 - -II. Materials and Methods-

Figure 3. Experimental setup to evaluate the dependency between mechanical stimulation and trout mucus application to trigger polar capsule discharge of individual Myxobolus cerebralis actinospores. All elements were freely moveable by micromanipulators. (Kallert et al. 2005c).

4.2.2. Bulk experiments

Bulk determination of polar filament extrusion rates was performed on microscope glass slides (Fig. 4). An amount of 9 µl of buffered test substrate was added to 21 µl buffered spore suspension on slides and covered with a cover slip (20 x 20 mm). To provide a mechanical stimulus, the preparation was placed on a vibration apparatus (Mini Shaker Type 4810, Brüel & Kjœr, Copenhagen), which was driven by an amplified laboratory frequency generator (FG 200, HTronic GmbH, Hirschau). The slide was vertically shaken for 3 s at 50 Hz (amplitude 0.5 mm), which was permanently controlled by an oscilloscope (Tektronix, Cologne) with triggered differential amplification. Spores were considered discharged if at least one polar filament was extruded. All experiments were conducted at room temperature under a microscope using magnifications of 200 x to 400 x and phase contrast optics. Only undamaged, viable spores that were not older than 48 h were included in the evaluation. All media, containing parasites or test substrates were adjusted to pH 7.5 (5 mM sodium phosphate buffer, PBS). The spore suspension was stored at 15 °C and the test substrates were kept on ice during the experiment. Except for the paired sample test

- 20 - -II. Materials and Methods- procedure and the frequency response test, a blinded protocol was used by allocating the test substrates codes unknown to the investigators. Buffered deionised water served as control in all experiments, untreated mucus homogenate (1 mg/ml final concentration unless otherwise noted) allowed judgement on the maximum achievable discharge rate.

Figure 4. Setup for the measurement of polar capsule discharge rates by actinospores in the bulk experiments.

4.2.3. Frequency dependency

Polar filament discharge of M. cerebralis actinospores using different mechano- stimulatory frequencies was examined as described for the bulk experiments (7.2.2.), whereas only 4 µl of trout mucus (1 mg/ml) were added to 16 µl spore suspension to stabilise the system at higher frequencies. The frequency and amplitude settings were altered according to calibrated values to retain the same vertical deflection amplitude (0.5 mm).

- 21 - -II. Materials and Methods-

4.2.4. Ca2+ dependency

To investigate the role of extracellular Ca2+ in the surrounding medium on polar filament discharge, H. nuesslini actinospores were brought from their original medium after filtration to artificial water (‘SFW 100’ according to Meier-Brook 1978, containing 100 ppm Ca2+) with and without Ca2+-ions. The water of the filtrated actinospore suspension was replaced by transferring the filters in 5 ml of the respective artificial water for 30 min with a change of medium for four times. In one amount of Ca2+–free water (7 mOsm/kg), the osmolality was restored according to the value of the Ca2+- containing mixture (12 mOsm/kg) by addition of NaCl, which was controlled by a repeated measurement. The discharge rates were measured as described under 4.2.2. using carp mucus homogenate (1 mg/ml final concentration) as the stimulating substrate. Individual controls were included showing the discharge rates for each artificial water mixture separately.

5. Cinematography of polar filament discharge

To visualize the process of polar filament discharge in detail, video sequences were recorded using an Axiophot (Zeiss) and a video camera (Canovision EX1 8 mm, Canon). M. cerebralis actinospores were placed on a glass slide in an excessive amount of medium, covered with a coverslip and 4-12 µl of a 30% aqueous NH3 solution were added slowly from the sides. The scenes were recorded digitally and stills were taken from screenshots using Ulead Video Studio 7.0 SE software.

6. Sporoplasm emission

To find out whether an intrinsic mechanism upon polar filament discharge causes the actinospores to open their apical sutures for sporoplasm release, H. nuesslini actinospores were observed after addition of rainbow trout mucus homogenate (1 mg/ml final concentration). The suspension (40 µl) was covered with a cover slip including plastilin spacers and discharged and undischarged spores were observed for 15 min. Sporoplasm release (emission of at least 1/3 of the primary cell mass) with and without preceding polar filament discharge was recorded. The period from

- 22 - -II. Materials and Methods- substrate mixing until sporoplasm release occurred was recorded to measure the duration the active sporoplasm emergence requires under undisturbed conditions. During selected bulk experiments measuring polar filament discharge rates, the rate of sporoplasms actively leaving the apical valve scaffold was additionally determined. Hereby, both discharged and undischarged spores not emitting sporoplasms were summarised. These data were intended to show whether the signals for sporoplasm emission are different from the discharge triggering mucus components.

7. Mucus fractionation

All treatments were carried out with rainbow trout mucus homogenate and included controls such as similarly treated and control (water) probes. Similarly treated samples were parallel handled substrates that did not contain reactive agents. Wherever possible, an appropriate monitoring technique to ensure successful fractionation was chosen. The substrates obtained were tested using the bulk experiments for measurement of discharge rates.

7.1. Heat and incubation

Mucus homogenate was incubated for 3 min at 60 °C or 37 °C for 3 h respectively in a water bath and was then immediately frozen at –75 °C. A small molecular weight fraction (< 3 kDa obtained by ultrafiltration) was heated to 100 °C for 5 min. Additionally, one aliquot was completely lyophilised (5 h) in a SpeedVac to examine the effect of volatile compound removal.

7.2. Ashing

Inorganic, non-volatile components from whole trout mucus homogenate were obtained by heating the substrate over a Bunsen burner until all organic substances were reduced to CO2 and water. The remaining white salt was resuspended in the original volume of deionised water and conductivity was measured using a conductivity measuring device (Albert Lang) for comparison with the initial value. The

- 23 - -II. Materials and Methods- resistance changed from 0.025 to 0.02 MΩ, with the difference resulting from withdrawn charged organic substances.

7.3. Acetone precipitation

For non-selective globular protein removal by acetone (see Nakagawa et al. 1988), mucus homogenate (7.25 mg/ml) was mixed 1:1 with acetone followed by centrifugation at 3600 g. The detached supernatant and the pellet obtained were lyophilised to remove acetone. Both fractions were resuspended in deionised water.

7.4. Alcian blue precipitation

The cationic dye alcian blue binds polyanionic Glycosaminoglycans (GAGs) and acidic proteoglycans (Whitehead, 1978). The same amount of a 1 mg/ml alcian blue solution (in 100 mM ammonium acetate and 50 mM MgCl2, titrated to pH 5.0 with acetic acid) was added to mucus homogenate forming a precipitate after 30 min of mixing on a rotating wheel. The pellet yielded by centrifugation was discarded and the supernatant was lyophilised under 3-fold resuspension to remove ammonium acetate. The blue dye interferes with most colorimetric monitoring procedures, so the sulphuric acid/resorcinol-method was chosen. The decrease in neutral sugar content in the extracted supernatant was 7% lower than in the similarly treated probe along with formation of visibly more precipitated substance.

7.5. Fluorescamin derivatisation

Selective, irreversible blocking of amino and thiol groups by fluorescamin (Castell et al. 1979) was conducted by resuspending lyophilised trout mucus homogenate in 25 mM ammonium bicarbonate (pH 9.0) and addition of 0.7 mg Fluorescamin per mg mucus homogenate in 25 µl of acetone. After a reaction period of 15 min at RT the reaction was stopped by addition of 250 µl 0.1 N HCl and the samples were frozen at –20 °C and lyophilised overnight.

- 24 - -II. Materials and Methods-

7.6. Extraction with activated charcoal

Trout mucus homogenate < 3 kDa was extracted by activated charcoal to quantitatively remove most aromatic compounds (esp. nucleotides, alcohols, most sugars, lipids and several amino acids) by hydrophobic interaction. Amongst others, inositols were not removed from the solution. The 0.4–fold amount of an aqueous 20% (w/v) activated charcoal solution (Norit A, Sigma, cooked in 3 N HCl, washed neutral on a Büchner funnel and dried at 120 °C prior to use) was added to the sample. After incubation for 15 min under repeated mixing, the suspension was centrifuged (6000 rpm, 10 min, 4 °C) and the supernatants from this and a subsequent washing step with 500 µl of deionised water were pooled and concentrated in a SpeedVac. To prepare the aliquot used for inositol analysis (II. 9.11.), 100 mM NaCl and 50 mM sodium acetate were added to an activated charcoal solution to avoid a major loss of phospho-inositol. The extraction was conducted as described above. The probe was further treated with an excess of IWT TMD-8 (Sigma, = Amberlite MB-3), a mixed bed ion exchanger to remove salts and acetate otherwise interfering with HPIC-IPAD and HPLC-MDD. The UV-absorbance spectra (200 - 380 nm) of these samples (100 µl) were detected by a microtiter plate reader (µQuant, Bio-Tek).

7.7. Lipid extraction

7.7.1. Ether-extraction

To obtain a lipid-free hydrophilic fraction, mucus homogenate was mixed with an equal amount of diethylether and mixed vigorously for 5 min in a separation funnel. Phase separation was achieved within 15 min, the hydrophilic phase was separated and extraction was repeated two times yielding an essentially lipid-free fraction without protein precipitation after evaporation of solvent remnants in a SpeedVac.

- 25 - -II. Materials and Methods-

7.7.2. Lipid isolation

Lipids were extracted according to Folch et al. (1957) by adding 40 x the volume of chloroform/methanol (2:1 v/v) to mucus homogenate followed by stirring on ice for 20 min with an Ultra Turrax (2500 rpm). After addition of 0.2% the volume of deionised water and extended shaking, the suspension was filtrated and allowed to stand overnight at 4 °C in a separation funnel. The organic phase was drained and evaporated with a rotation evaporator (RV 05, Janke & Kunkel GmbH); remaining humidity was removed under a nitrogen stream. Lipids were stored in chloroform/methanol (2:1 v/v) at –75 °C. The solvent was removed for experiments under a nitrogen stream and an aqueous suspension was prepared by warming dry lipid to 50 °C and addition of 5% ethanol for solubilisation. Deionised water heated to 65 °C was added and samples were homogenised using a sonicator for 3 min (amplitude 80%, cycle 0.8).

7.8. Chemical fractionation

7.8.1. TFMS Deglycosylation

Deglycosylation of mucins was conducted by treatment with trifluoromethanesulfonic acid under anhydrous conditions using a modified protocol according to Edge et al. (1981). After a preceding sialidase digestion (0.25 U/ml sialidase from Clostridium perfringens per probe dissolved in 100 µl 50 mM ammonium acetate giving a pH of 5.5, 15 min incubation at room temperature on a mixing wheel followed by 10 min reclining on ice and ultrafiltration (MWCO 3 kDa) under addition of 0.05% sodium azide and 1 mM Pefabloc protease inhibitor), retentates > 3 kDa of trout mucus homogenate and controls (deionised water) were completely lyophilised for 24 h and stored in glass tubes with pierce-able gas-tight plugs (Vacutainer, Brand) under nitrogen. To each dry sample (37 mg) 2.9 ml of reaction mixture (TFMS/anisole 1:0.17 v/v, anisole served as scavenger) was added using gas-tight glass syringes under nitrogen and subsequent cooling to –20 °C. The reactants were left on ice for 3 h and vortexed repeatedly. A mixture of ice-cold pyridine solution (pyridine/deionised water 3:2 v/v) for

- 26 - -II. Materials and Methods-

TFMS complexation was added and pH was adjusted with 9.5 ml of neutralisation solution (1.5% ammonium bicarbonate). Samples were then transferred to Zellu Trans dialysis tubes (cut-off MW 2 kDa, Roth) and excessively dialysed against 4 x 1 l of deionised water containing 0.05% w/v sodium azide and Pefabloc (5 µg/ml). To one set of probes, deionised water was added instead of all chemicals. They were otherwise treated as the deglycosylated probes to reveal biological effects of reagent remnants. Before use in experiments, the substrates were again desalted by ultrafiltration. Monitoring with the sulphuric acid/resorcinol method could show that the deglycosylated samples contained only 24.9% of neutral carbohydrates that were present in similarly treated substrates (both > 3 kDa).

7.8.2. Sialic acid extraction

For preparative release of sialic acids from mucins, trout mucus homogenate (4 ml > 3 kDa-fractions corresponding to an initial dry weight of 16.5 mg) was mixed with a 2.25-fold amount of 2 M propionic acid and incubated at 80 °C for 3 h (Mawhinney & Chance, 1994). Following acid hydrolysis, samples were freeze-dried (Alpha I-5, Christ) and the sialic acid residues were separated from proteins by ultrafiltration as described. For purification, Dowex 1 x 8 anion exchange resin (mesh 200-400) was brought to the HCOOH-form by suspending the resin in 2 M NaOH for 30 min, neutral washing with water and functional group replacement by addition of 2 M formic acid. After washing with 500 ml of water, the filtrates (1.5 ml) were adjusted to pH 4 by addition of NH3 and added to a column loaded with 7.5 ml bed volume of resin. The bound constituents were washed with 30 ml of deionised water (F I) and sialic acids were eluted in a first step with 30 ml 1 M propionic acid (F II) and in a third step with 30 ml 2 M formic acid (F III). Solvents were removed by freeze-drying, the remainder was resuspended in deionised water. Sialic acid contents were measured using the colorimetric method described under III. 9.3.

- 27 - -II. Materials and Methods-

7.9. Enzymatic fractionation

Terminal sialic acid residues from high MW mucin glycoconjugates were removed by enzymatic hydrolysis. Samples were diluted 1:1 with ammonium acetate (100 mM) and the pH was adjusted to 5.5 by acetic acid addition (final reaction volume was 10 ml). To suppress intrinsic proteolytic activity, protease inhibitor (1 mM Pefabloc, Roth) was added to the samples. Neuraminidase (Type V from Clostridium perfringens) was added (0.27 units per mg solid substrate in 25 µl 50 mM ammonium acetate) prior to incubation at 37 °C for 3 h. Low molecular components were removed by ultrafiltration to 1/3 of the volume including a washing step with an amount of 2/3 the remaining volume of deionised water before resuspension of the retentate. Serine-positions of proteins were cleaved by adding 0.04 units Proteinase K (from Tritirachium album, preincubated for 30 min at 37 °C, in 10 µl 50 mM PBS, pH 7.5) per mg dry solid to 4 ml reaction volume containing 25 mM PBS (pH 7.5). Incubation time was 3 h at 37 °C with a repeated enzyme addition (boost) after 2 h. After incubation, the reaction was stopped by heating to 70 °C for 5 min. Additionally, a solution of BSA (5 mg/ml) was also digested for monitoring. All samples were filtrated (MWCO 3 kDa) overnight at 4 °C and the retentates were resuspended in deionised water. Total proteolysis was conducted by Pronase E (Sigma) digestion as described for proteinase K digestion using 0.1 units of enzyme per mg solid substrate without enzyme preincubation. To break down small peptide structures in the low molecular mucus fraction, peptidase (from porcine intestinal mucosa) and proteinase K digestion was carried out. Substrates (< 3 kDa) were treated as described above (enzyme concentration was 0.07 units per mg initial dry weight without PBS), incubating the reactions for only 1 h on a mixing wheel at room temperature and without addition of PBS. Pefabloc (1 mM) was used as a stopping reagent.

- 28 - -II. Materials and Methods-

7.10. Chromatographic fractionation

7.10.1. Ion-exchange

Fractionation by Macro-Prep High Q support (a basic reversed phase anion exchanger) utilises quaternary amine groups, providing effective protein and peptide retention. Macro-Prep High Q Econo-Pac cartridges (Bio-Rad, 5 ml bed volume) were connected to a Bio-Rad Econo-system and conditioned following the manufacturer’s protocol. Before each run the support was purified by 20 ml 1 M NaOH followed by washing with 100 ml deionised water (1.5 ml/min) and equilibration with deionised water for 1 h at 0.2 ml/min flow rate. The column was loaded with sample volumes of 2 ml substrate (trout mucus homogenate < 3 kDa) that were eluted (flow rate 0.5 ml/min) in a first step with 60 ml 50 mM ammonium acetate (low salt buffer) and in a second step with 60 ml 1 M ammonium acetate (high salt buffer). Both solvents were adjusted to pH 7.5 with NH3 and acetic acid. The resulting fractions were completely freeze- dried under five-fold resuspension to evaporate the buffer. Low molecular ionic substances were removed with a pre-packed AG18 A11 (Bio- Rad) mixed bed resin column for inorganic ion retardation chromatography. This ion retardation material separates salts from organic materials by absorbing both anions and cations in equivalent amount while allowing the organic compounds to pass through the column. Ionic species of organic character (acidic and basic amino acids) are not absorbed by the resin. The retarded ionic substances, except the hydrogen ion, leave the column later when eluted with water. Sample volumes of 3 ml were loaded on the column after washing with 30 ml of deionised water. The sample was separated in 5 fractions by subsequent elution with 10 ml of water each and one ensuing washing step with 50 ml water, the eluates were freeze-dried. Organic anions were directly extracted from trout mucus homogenate < 3 kDa using Dowex 1x8 resin (formiate-form, 200-400 mesh) as described under III. 7.8.2., while sialic acids, that normally do not displace strong anionic functional groups bind to this phase as well.

- 29 - -II. Materials and Methods-

7.10.2. Lichroprep RP 18

Separation by hydrophobic interaction was conducted by Lichroprep RP 18 reversed phase chromatography. For solid phase extraction (SPE), a self-packed glass column (10 g support, grain 40-63 µm, Merck; 7 ml bed volume, slurry packed after swelling in methanol) was washed with 4 bed volumes of deionised water and loaded with 3 ml substrate (< 3 kDa). Stepwise elution by gravity (0.4 ml/min) was performed by the addition of 20 ml deionised water, 10% 2-propanol, 50 % 2-propanol and 100% methanol respectively. For improved separation of compounds, another column (14 g support, 10 ml bed volume, including an upper frit and complete sealing) was packed and connected to a Bio-Rad Econo-system. After addition of 3 ml sample, the system was pre-run for loading with 4 ml of deionised water at a flow rate of 0.2 ml/min. Bound components were washed with 40 ml of deionised water followed by a linear gradient of 0 – 25 % 2- propanol and washing with 20 ml 50% 2-propanol at a flow rate of 0.5 ml/min. A fraction collector took samples of 880 µl each. All fractions were lyophilised in a SpeedVac and resuspended in deionised water.

8. Pure chemicals

Several commercially obtained substances were tested for their stimulating effect on polar filament discharge. The mixtures requiring explanation are described hereafter. Amino acids were mixed according to the data obtained by HPLC detection (9.8.). Different mixtures from pure AAs were analysed and the one most closely resembling the quantitative data (Fig. 26) was then used for experiments. An electrolyte mixture was prepared according to the data for rainbow trout mucus by

Handy (1989). The salts contained were 2 mM NaCl, 0.09 mM CaCl2, 0.12 mM CaCO3 and 0.42 mM K2CO3.

- 30 - -II. Materials and Methods-

The following fluorescein-labelled neoglycoproteins (BSA-conjugates, Sigma) were tested for discharge triggering activity:

Albumin, bovine β-lactosylphenyl isothiocyanate (BSA-Lac) Albumin, bovine α-D-mannopyranosylphenyl isothiocyanate (BSA-Man) Albumin, bovine α-D-glucopyranosylphenyl isothiocyanate (BSA-Glu) Albumin, bovine β-D-galactopyranosylphenyl isothiocyanate (BSA-Gal) Albumin, bovine α-L-fucopyranosylphenyl isothiocyanate (BSA-Fuc)

To test combinatory effects, a mixture of the major free compounds (except nucleosides) that could be quantitatively detected in this study was prepared by using electrolyte solution obtained from mucus ashing preparations as a basis. Neuraminic acid (Neu5Ac) was used as substitute for all sialic acid derivatives. The mixture used in the experiments (corresponding to a final concentration of 2 mg/ml initial dry weight) included the following constituents per ml (paragraph number of the analytical procedures given in brackets):

Inorganic electrolytes - - N-acetyl-neuraminic acid 13.5 ng (9.5.) Myo-inositol 8.5 ng (9.11.) Scyllo-inositol 0.3 ng (9.11.) Chiro-inositol 0.3 ng (9.11.) D-Glucose 6.7 ng (9.9.) D-Galactose 0.2 ng (9.9.) D-Mannose 0.4 ng (9.9.) N-Acetyl-Glucosamine 0.2 ng (9.9.) N-Acetyl-Galactosamine 0.1 ng (9.9.) Amino acids (see Results section) (9.8.) Urea 20.8 ng (9.4.)

- 31 - -II. Materials and Methods-

9. Chemical analyses

9.1. Proteins

For the quantitative analysis of proteins according to Bradford (1976), 160 µl of each probe were analysed in 96 well plates with 40 µl Roti-Quant (Roth). Absorbance was measured with a plate reader within 30 min after mixing at 595 nm. An aqueous BSA dilution series (10-50 µg/ml) was used for calibration.

9.2. Amino compounds

As a sensitive colorimetric method to detect small amounts of peptides and amino acids, a rapid assay for amines according to Starcher (2001) was used. To sample volumes of 100 µl, the same amount of ninhydrin reagent (25 mg SnCl2 and 200 mg ninhydrin in 8 ml ethylene glycol) was added in a 96 well microtiter plate. The plate was heated to 100 °C for 15 min floating in a water bath and absorbance was measured at 575 nm with a plate reader.

9.3. Neutral sugars

Concentrations of neutral sugars were measured by the H2SO4/Resorcinol-method of Monsigny et al. (1988) in 96-well microtiter plates. To 20 µl sample 20 µl resorcinol (6 mg/ml) were added and carbohydrates were hydrolysed to furfurals by 100 µl

H2SO4 (75%). Surface tension was lowered by 40 µl Pristane and the plates were heated for 30 min to 95 °C and cooled to room temperature for 30 min. A mannose series (6.25 to 100 µg/ml) served as standard and the absorbance was measured at 430 nm.

9.4. Urea

The determination of urea concentration in mucus homogenates was conducted by the phenol-hypochlorite assay of Bernt & Bergmeyer (1970). Ammonia released by urease reacts with phenol and hypochlorite to indophenol. Urease suspension (200

- 32 - -II. Materials and Methods-

U/ml in 50% glycerine) was mixed 1:2 with the samples followed by incubation at 37 °C for 15 min in a water bath. Then the 10-fold amount of phenol reagent (0.106 M phenol and 0.17 M nitroprusside-sodium) and hypochlorite solution (11 mM NaOCl and 0.125 M NaOH) were added. Absorbance at 630 nm was measured after 30 min of incubation at room temperature. To assess the content of naturally occurring ammonia, samples without enzyme were included in the test.

9.5. Sialic acids

The Svennerholm resorcinol method for the detection of sialic acids by Jourdian et al. (1971) was adapted for use with microtiter plates by Bhavanandan & Sheykhnazari (1993). A sample (40 µl) was added 10 µl of 0.032 M periodic acid and the solution was mixed for 5 min. The plate was placed on ice for 35 min, then 100 µl resorcinol reagent (600 mg resorcinol and 25 µmol Cu2SO4 in 60 ml of 28% HCl and 40 ml of deionised water. The plate was heated to 80 °C for 1 h and absorbance was measured at 630 nm.

9.6. TLC-separation of lipid classes

Mucus lipids were divided into their major constituents by a four-step thin layer chromatographic method using Kieselgel 60 layers (0.25 mm, 20 x 20 cm glass plates, Merck). The plates were activated at 130 °C for 30 min. References (0.1 mg in 20 µl chloroform/methanol 2:1 each) and samples were applied in 5 µl steps as dots on the starting line. After evaporating of the solvent the plate was placed into a TLC chamber (double trough chamber, CAMAG) for 10 min in which before a run the front chamber was filled with solvent to saturate the air in the chamber. After the solvent front had migrated, the plate was taken from the chamber and dried. The solvents and separation track lengths were:

n-hexane: 19 cm Toluol: 19 cm n-hexan/diethylether/acetic acid (70:30:1 v/v) 10 cm (two separations)

- 33 - -II. Materials and Methods-

Polar lipids and hydrophilic compounds were separated using the same protocol as for the whole lipid TLC using chloroform/methanol/water (40:10:1 or 65:25:4 v/v) as sole solvent. For detection, spots were charred by spraying with 50-75% sulphuric acid and heating to 160 °C for several min.

9.7. Hydrophilic TLC

To analyse small molecular mucus compounds (< 3 kDa) and several mucus fractions obtained (e.g. by Lichroprep RP 18 SPE, III. 7.10.2.), a TLC protocol for carbohydrates according to Siegenthaler & Ritter (1977) was used. Acetonitrile/deionised water (7:3) served as solvent for use with Kieselgel 60 plates (0.25 mm, 20 x 20 cm, glass plates, Merck). The spots were visualised on a UV imaging table at 312 nm excitation, optionally followed by spray detection using either a ninhydrin reagent (0.2% in ethanol, Sigma), carbohydrate staining by anilinephtalate (No. 1266, Merck) or charring by sulphuric acid as already described. To see whether a better separation could be achieved, some plates were developed again after the first fluorescence detection. In a preparative approach, five bands detected by UV were scraped off, the remaining lane was pooled. As a solvent for re-extraction, water and methanol were used subsequently. After solvent addition, reaction vials were shaken on an Eppendorf mixer (Type 5432) for 50 min followed by centrifugation (7000 rpm, Type 5417 R). After repeated extraction, the supernatants for each fraction were pooled, concentrated in a Speed Vac and analysed by TLC again. Various reference substances were used for comparison of retention values with compounds from mucus fractions showing bioactivity for polar filament discharge. For this purpose, mycosporine-like amino acids (MAAs) were prepared from freeze-dried Spirulina sp., referring to a protocol by Sinha et al. (2002). Briefly, 20 mg pulverised algae (MaBitech) were extracted in 1 ml 20% methanol at 45 °C for 2 h. After centrifugation at 10 000 rpm for 10 min a supernatant could be taken off and was lyophilised in a SpeedVac. The remainder was solubilised in 1 ml of 100% methanol and centrifuged again. The resulting supernatant was again lyophilised and the remainder was solubilised in 200 µl of water.

- 34 - -II. Materials and Methods-

9.8. RP-HPLC

For high performance liquid chromatography of free amino acids in trout-, carp-, and snail-mucus, a LiChrospher 100 column (5 µm, 250 x 4mm, Merck) was connected to a Kontron HPLC system with a fluorescence detector (BT 6630, Biotronik). O-phtaldialdehyde was used for pre-column derivatisation (see Fürst et al. 1990) of lyophilised samples. Samples (10 µl in 1 M sodium borate buffer) were added 5 µl of OPA/MPA (10 mg o-phtaldialdehyde and 10 µl 3-mercapto-propionic acid in 1 ml ice-cold methanol). The reaction was stopped after 150 s by addition of 15 µl

KH2PO4, and 20 µl of the derivatised sample were injected onto the column after 60 s. Elution was conducted during 15 min at a flow rate of 1.5 ml/min with a linear gradient from solvent A, consisting of a 10:3:3:84 mixture of 10 x PBS (125 mM Na2HPO4, 125 mM NaH2PO4)/tetrahydrofuran/acetonitrile/water and solvent B (7:30:63 PBS/acetonitrile/water). Amino acid standards (AA-S-18, A-6282, A-6407, Sigma) in different dilutions were used for 3-point calibration. Because imino acids like proline and hydroxy-proline react with OPA only after oxidization by chlorinated agents to a primary amino acid, pre-column derivatisation with 4-chloro-7-nitrobenzofurazan (NBD-Cl) according to Carisano (1985) was used. The derivatisation reagent (5 µl of 10 mg NBD-Cl in 2 ml methanol) was added to a 10 µl OPA-containing sample (50 µl sodium borate and 10 µl OPA) after 2 min heating to 60 °C followed by incubation at 60 °C for 10 min. The reaction was stopped by addition of 24 µl of 0.5 N HCl before injection. Solvent A was 10 x PBS/tetrahydrofuran/acetonitrile/water 10:3:3:84 (v/v) brought to a pH of 6.5 with acetic acid. Solvent B was a mixture of methanol/water (65:35). Elution time was 15 min at a flow rate of 1.5 ml/min with a linear gradient from solvent A to B.

9.9. GC/MS

Monosaccharide spectra of small molecular fractions of trout-, carp-, and snail- mucus could be obtained by gas chromatography/mass spectrometry in the laboratory of Prof. R. Geyer at the Biochemical Institute of the University of Gießen by P. Kaese. For determination of free sugars, dry samples (trout mucus homogenate < 3 kDa) were solubilised in 500 µl 0.1% NaBH4 and left for 16 h. After the samples were neutralised with 100 µl 2 N acetic acid and dried in a SpeedVac. To remove boracic

- 35 - -II. Materials and Methods- acid methyl ester, the samples were diluted 4 times with 1% methanol/acetic acid solution (1:1 v/v) and dried under a stream of nitrogen. Acetylation was achieved by addition of 100 µl pyridine and 400 µl acetic acid anhydride followed by incubation for 16 h under argon. Dried by nitrogen again, samples were extracted 3 times with di- chloromethane/water (1:1 v/v) with the aqueous phase being rejected. After desired concentration by gaseous nitrogen, sampled were analysed by GC/MS (Trace GC and Polaris Q MS, Thermo Electron). For determination of bound sugars, dry samples were incubated for 4 h at 100 °C in 4 N TFA prior to the above protocol.

9.10. UV-Spectroscopy and fluorescence detection

Ultraviolet absorbance spectra (200-400 nm) of various pure substances, small molecular mucus components and chromatographic fractions were measured using a plate reader. Samples (100 µl aqueous solutions) were analysed using UV-lucent microtiter plates (Costar, Corning). Additionally, as the fractions F I and F II obtained from the Lichroprep RP 18 SPE (III. 7.10.2.) exhibited a strong luminescence upon excitation at 312 nm, grey scale (brightness) values measured from digital images of the plates were recorded for each fraction. Browsing each fraction well using Paint Shop Pro (Version 7.0, Jasc Software) software, only the largest values were included, yielding an additional chromatogram.

9.11. HPIC-IPAD and HPLC-MDD

This analysis was done by Dr. S. Adelt at the Biochemistry Institute of the Bergische University Wuppertal. Briefly, high-performance anion-exchange chromatography (HPIC) of inositols employed a linear gradient up to 250 mM NaOH on a high-capacity column (CarboPac MA1, DIONEX) with integrated pulsed amperometric detection (IPAD). At high pH, the compounds are electrocatalytically oxidised at the surface of a gold electrode by application of a positive potential, and the current generated is proportional to their concentration. Additionally, phosphate and sulphate content could be measured by conductivity detection of the eluents using corresponding standards. In all separations, conductivity could be measured in an extra channel.

- 36 - -II. Materials and Methods-

By the additional HPLC-MDD (metal dye detection) procedure, inositol phosphates were analysed as described by Adelt et al. (1999). The compounds were separated by anion-exchange chromatography on a MonoQ HR10/10 column (Pharmacia). To purify inositol phosphates, a linear gradient of HCl was applied (0 min 0.2 mM HCl; 70 min 0.5 M HCl; flow rate 1.5 ml/min). Photometric detection at 546 nm was achieved with a metal-dye reagent [2 M Tris/HCl (pH 9.1), 200 µM 4-(2-pyridylazo) resorcinol, 30 µM

YCl3 ,10% (v/v) methanol at a flow rate of 0.75 ml/min.

9.12. NMR-spectroscopy

Prof. Dr. Bauer of the Institute of Organic Chemistry of the FAU Erlangen-Nürnberg kindly conducted this analysis to identify the contents that were isolated in the fraction that effectively triggered polar filament discharge. A combination of one-dimensional 1H- and 13C-spectra, as well as two-dimensional COSY and HMQC spectra for homo- and heteronuclear shift correlation was employed. Spectra were recorded on a JEOL Alpha500 spectrometer (1H = 500 MHz). The amount of 0.56 mg of the lyophilised subfraction peak (elution volume 61-67 ml) from the LichroPrep RP 18 gradient chromatography (7.10.2.) was dissolved in 0.4 ml of D2O. A 5 mm multinuclear probe head was employed for the recording of the one-dimensional 13C-spectrum. The one- dimensional 1H-spectrum as well as the field gradient-based COSY and HMQC spectra were recorded on an inverse probe head with actively shielded gradient coils.

- 37 - -II. Materials and Methods-

10. Chemicals

All chemicals were purchased by Sigma, Fluka and Roth. Chromatography media were purchased from Merck and Supelco.

11. Statistical methods

Statistical analysis was performed using SPSS for Windows (11.05). Relative abundances were arcsine square root transformed to obtain approximately normally distributed data. Mean values and standard errors were computed from the transformed data and were retransformed. Differences between means were tested for statistical significance by a multivariate comparison procedure (Tukey HSD multiple t- test after One-Way-ANOVA and the Levene-test for homogeneity). When only two samples were compared, a t-test based on Student`s distribution was used. Differences in change of response of the nonparametric data that resulted from the paired samples test of individual spores were determined by the McNemar χ²- procedure for nominal dichotomous variables in experimental before-after designs. Correlation between binominal variables was tested by crosstabulation followed by calculation of Yates` continuity correction as well as the Phi-value and the contingency coefficient as symmetric measures. Mantel-Haenszel statistics (homogeneity of the odds ratio) was used to test for independence in response between the binary variables.

- 38 - -III. Results-

III. Results

1. Life cycle experiments

1.1. Henneguya nuesslini

Infection experiments

During the attempt to infect brown and brook trout with actinospores from naturally infected oligochaetes, up to day 80 p.e. no myxospores were detected in homogenates of three fish from each salmonid species. Mature myxospores were initially found in tissue homogenates of both species 102 days p.e. in the first infection trial, in which the fish were infected by cohabitation in the oligochaete containers. At this time the fish were up to 7 cm in length. All fish infected in the first trial harboured mature myxospores, but showed no pathological symptoms. The numbers of spores found in individual fish varied greatly during examinations over 3 months and were not estimated. Surprisingly, all except two of the individually infected brown trout and brook trout from the second trial died during the first 28 days p.e. after showing severe neurological disorders, i.e. swimming imbalance, dyspnoea, secondary infections and cessation of feeding. The surviving specimens were heavily infected with H. nuesslini when examined after 172 days. No other myxosporean infections were detected in homogenates of two examined individuals of both fish species infected only with H. nuesslini actinospores from single oligochaetes. No visible pseudoplasmodia were observed in the trials. No myxospores were found in any homogenised tissue samples from carp and control salmonid fish at 115–160 days p.e.

Oligochaete identification

Actinospores were shed by oligochaetes of the species Tubifex tubifex (Müller, 1774, Annelida, Clitellata, Tubificidae). All the examined oligochaetes were sexually immature, but their chaetal morphology and distribution suggest that they were T. tubifex. Their general body outline indicated they were not Potamothrix hammoniensis (Michaelsen, 1901) or Ilyodrilus templetoni (Southern, 1909), species which have chaetae similar to T. tubifex (Timm 1999). In the ethanol fixed specimen, although its

- 39 - -III. Results- male ducts were not well visible, there was no sign of the characteristic modified genital chaetae of P. hammoniensis or the narrow, cuticular penis sheaths of I. templetoni (Timm 1999). The host sequence showed a 99.6% match with that of T. tubifex from a culture specified as ‘B178’.

Parasite identification

The universal 18e-18g’ primers and the specific primer pair SphF-SphR amplified fragments of about 1900 and 1400 bp of the 18S rRNA gene from every sample examined respectively. Nested PCR with SphF-SphR primers followed by SphF-MB5r primers generated a 660 bp DNA fragment, while the SphF-SphR/MB5-SphR nested PCR system gave a 760 bp product. Assembled sequences of all samples obtained from TAMs and infected fish tissue examined were 100% identical. A 1417 bp long consensus DNA sequence was deposited in GenBank (Accession Number AY669810). Comparing this sequence to myxosporean sequences available in GenBank, H. zschokkei (AF378344) collected from muscle of the mountain whitefish Prosopium williamsoni (Girard, 1856) in Canada was found to be 99.0% similar to the species examined, while H. salminicola (AF031411) and H. nuesslini shared 96.6% of identical nucleotides.

Description of actinospores

The typical anchor-shaped triactinomyxon-type spores (Fig. 5) had long, tenuous styles and large processes (measurements, Table 1), the latter being arcuate with pointed tips (Fig. 6 (A)). Three pyriform polar capsules notably surmounted the spore body apex (Fig. 6 (B)) and enclosed a five-fold coiled filament (6 individuals examined). Spores lacking one polar capsule were frequently found. Sutural lines along the valve cells were well visible; the respective nuclei were situated at the process base. The cylindrical shaped sporoplasm contained most frequently 16 secondary cells (15 individuals examined), sometimes less (12 or 14), embedded in a very dense granular matrix. Sporoplasm cells were not visible in uncompressed fresh mounts (Fig. 6 (B)). Valve cell nuclei were located at the base of the caudal processes or in the style median. TAM measurements varied in total size, style length, and, as a result of being bent, caudal processes length.

- 40 - -III. Results-

Table 1. Measurements (µm) of Henneguya nuesslini triactinomyxon spores (26 individuals) from various infected Tubifex tubifex

Measurements Mean ± SD Minimum Maximum

Total length 164.8 ± 8.5 146 180 Polar capsules length 6.6 ± 0.5 5 7 width 3.8 ± 0.4 3 4 Sporoplasm length 27.1 ± 3.0 21 32 width 11.1 ± 0.8 9 12 Spore body length 34.6 ± 3.7 26 41 width 11.1 ± 0.8 9 12 Style length 138.4 ± 9.5 114 156 width 17.2 ± 2.2 14 22 Caudal processes length 227.5 ± 27.9 180 317 width 12.9 ± 1.8 10 16

Figure 5. Line drawing of the triactinomyxon spore of Henneguya nuesslini; bar = 100 µm.

- 41 - -III. Results-

Description of myxospores

Spores were drop-like in valvular view and biconvex in sutural view (Fig. 6 (C)), and both valves exhibited a bifurcate caudal process (‘tails’) that measured about 1.5 times the spore body length. The valve-derived tails were sometimes slightly dehisced in sutural view, but appeared to be closely attached to each other in valvular view (Fig. 6 (C), inset). The spore body was rather thick (~8.0 µm), and the sporoplasm filled more than half of the spore body cavity. The suture line was readily visible in fresh material. At least five filament coils were counted (5 individuals examined) in the equally sized pyriform polar capsules. All proportions are summarised in Table 2.

Table 2. Measurements (µm) of Henneguya nuesslini myxospores (25 individuals) from Salvelinus fontinalis muscle tissue.

Measurements Mean ± SD Minimum Maximum

Total length 31.3 ± 2.8 27 37 Polar capsules length 4.8 ± 0.5 3 5 width 3.2 ± 0.3 2 3 Spore body length 11.8 ± 1.3 10 17 width 9.2 ± 1.4 8 12 thickness 7.9 ± 0.8 7 9

Histology

Sections of connective tissue of brown trout, obtained after 164 days p.i., revealed spherical, microscopic trophozoites (Fig. 6 (D)) harbouring mature myxospores in connective tissue. The round to ellipsoid trophozoites were encapsulated, their size and shape was variable. They were mostly located in the caudal part of the host, but trophozoites were also detected in cranial regions, where plasmodia had migrated between cartilage plates. No myxozoan stages were found in parenchymal organs, gills or muscle septa.

- 42 - -III. Results-

Figure 6. Photomicrographs of developmental stages of Henneguya nuesslini. (A) Triactinomyxon spore in bright field (bar, 100 µm). (B) Apical region of the triactinomyxon spore. Note the dense sporoplasm matrix and protruding polar capsules (bar, 10 µm). (C) Myxospore in valvular and sutural view (inset) (bar, 10 µm). (D) Subcutaneous cysts of H. nuesslini in posterior connective tissue of Salmo trutta 164 days after exposition (H & E stain, bar, 100 µm).

1.2. Myxobolus parviformis sp. n.

Infection experiments

Oligochaetes produced actinospores (~300000 per week) 136 days p.e. in the first trial. The infected annelids produced two actinosporean types in the initial trial, but M. parviformis TAMs could clearly be distinguished by their diminutive size and constituted over 80% in TAM number throughout the production. The control oligochaetes did not produce any actinosporeans. Bream specimens infected with

- 43 - -III. Results- whole filtrate actinospore suspensions were initially found to carry pseudoplasmodia containing mature myxospores in secondary gill lamellae 147 days p.e. The size of the pseudoplasmodia was always less than 300 µm. Specimens that were infected by cohabitation yielded higher infection intensity (mean 3.2 identifiable pseudoplasmodia per primary gill filament, 10 individuals examined) than the individually infected ones (mean 0.54 identifiable pseudoplasmodia per primary gill filament, 9 individuals examined). In the second trial to infect oligochaetes using myxospores obtained from the first trial, actinospore production from naïve parasite-free specimens started after 114 days, then lasted over three months at 18 °C and could be prolonged in selected host specimens by keeping the oligochaetes at 4-6 °C for over 300 days. Bream that were exposed to actinospores only from M. parviformis (‘pure infections’) harboured mature myxospores after 67 days p. e. and developed only M. parviformis myxospores in numerous pseudoplasmodia. No other spore type was found in gill homogenates.

Oligochaete identification

Although none of the tubificids was sexually mature, they matched the morphological characteristics of either young Tubifex tubifex Müller, 1774 (3 specimens) or young L. hoffmeisteri Claparéde, 1862 (2 specimens, including the partitioned one, see above). The single oligochaete host sequence obtained was most similar to that of a positively identified L. hoffmeisteri specimen (with a 10.4% sequence divergence), and second most similar with that of a Limnodrilus udekemianus (with a 14.4% sequence divergence). There are great differences in the 16S sequences of ubiquitous, cosmopolitan tubificid taxa; e.g., an average divergence of 12.2% was noted between various Limnodrilus spp. (Beauchamp et al. 2001). Therefore, it is likely that the sequenced host oligochaete indeed was L. hoffmeisteri.

Parasite identification

Sequence data of the myxospore samples obtained from the first trial were compared with actinospores shed by various isolated oligochaetes and a pool of actinospores from ‘pure infection’ (second trial) culture filtrates. The universal 18e-18g’ primers and the specific primer pair MX5-MX3 successfully amplified approximately

- 44 - -III. Results-

1900 and 1600 bp fragments of the 18S rDNA from every sample examined respectively. In the case of the myxospore-sample, PCR with MX5-MX3 produced a weak unspecific fragment. For sequencing, the DNA fragment of the expected size was isolated from the agarose gel. The unspecific band, about 1400 bp in size, was also isolated, purified and the 5’-end of the fragment was sequenced directly. With BLAST search, this DNA fragment was determined as virtually identical to the 18S rDNA of a planktonic alga, Spumella elongata (Chrysophyceae) (AJ236859); only a 1 nucleotide difference was observed in a 690 bp long overlapping DNA sequence.

Figure 7. RFLP patterns of the Myxobolus parviformis sp. n. SSUrDNA PCR products digested with (A) HinfI, (B) MspI and (C) TaqI enzymes. Lane 1, TAMs from the first infection cycle (pooled filtrate); lane 2, TAMs from isolated positive T. tubifex (from first infection cycle); lane 3, collected cysts from bream gills; lane 4-8, TAMs from the second infection cycle (from single isolated infected T. tubifex) and lane M, 100 bp DNA ladder.

The restriction fragment patterns of all samples examined were identical using the three restriction enzymes (Fig. 7). Assembled sequences of the myxosporean and triactinomyxon developmental stages of M. parviformis were also 100% identical. A 1586 bp DNA sequence from the myxosporean stage was deposited in the GenBank (Accession No. AY836151). Two sequences of cloned fragments of TAM samples from the first infection trial were 100% identical with pooled actinosporean samples obtained from selected oligochaetes from the second trial and the myxospore sample, while one to three nucleotide differences (0.06 - 0.19%) at different positions were observed among sequences of the other four clones. Using BLAST search, one of the seven clones was determined as the partial 18S rDNA of a planktonic alga belonging to the genus Chrysosaccus (Chrysophyceae). Contaminating myxozoan DNA fragments were not - 45 - -III. Results- amplified by the nested PCR in any of the samples. In comparisons with sequences of other myxozoans available in GenBank, that of triactinomyxon ‘type 1’ (AY495704), reproduced by Hallett et al. (2005) from similar TAMs released by commercially purchased T. tubifex specimens, was found to be 99.94% similar to the M. parviformis sequence. There is only a single nucleotide difference within 1586 bp.

Description as new species

Myxobolus parviformis sp. n. Triactinomyxon type 1; Hallett et al. (2005)

Etymology: The epithet ‘parviformis’ refers to the comparatively small size of all lifecycle stages that were produced in the experimental infections. This includes gill plasmodia, triactinomyxon spores and myxospores. Type material: Syntype myxospores mounted in glycerine-gelatine (and a photo series) are deposited in the protozoan collection of the Zoological Department of the Hungarian Natural History Museum (Coll. No. HNHM-69902/1-2). Geographic source: River Aisch near Erlangen, Bavaria, Germany. Vertebrate host (type host): Abramis brama L. Invertebrate hosts: Limnodrilus hoffmeisteri Claparéde, 1862, and Tubifex tubifex Müller, 1774. Site of infection: Myxospore stages in respiratory lamellae of gills, in distal region of gill filaments up to the tip, forming spherical pseudoplasmodia. Actinospore stages developed in the gut epithelium of the oligochaete hosts. Actinospores: For detailed proportions see Table 3. Triactinomyxon-type spores (Fig. 8; 111.0 – 142.5 µm total length; mean 127 +/- 6 µm, 31 individuals examined) with stout styles (100.3 µm long, 10.3 µm wide) and sharply pointed processes with acuminate, often tapering tips (101.7 µm long). Three pyriform polar capsules (7.3 µm long) with five filament coils (6 individuals examined) and a dense, vacuole-like inclusion. Valve cell nuclei always located in median or caudal process positions, not at process bases or along style. Sporoplasm cell count generally 32 (> 80%, 10 individuals examined from different oligochaete hosts), sometimes less, embedded in an ellipsoid to barrel-shaped sporoplasm, visible in uncompressed fresh mounts (Fig.

- 46 - -III. Results-

9 (A), inset). Sporoplasm (26.7 µm long, 11.8 µm wide) enclosed in secondary detachable sheath (Fig. 9 (B)). The sporoplasm sheath is a soft, pouch-like structure that carries the polar capsules apically. It is expelled after spore valves are opened along deep hyphenation ridges (up to half of the style length) emanating from the sutural lines of the valves. After 5 sec of vortexing a spore suspension, the opening of the actinosporean valves (Fig. 9 (A)) could be observed in several specimens, which released the sheath unit (Fig. 9 (B)). This reaction plus the emergence of the amoeboid germ could also be generated by incubation of the spores in homogenised bream skin mucus. The polar capsules were attached to the immobile sheath unit; the sporoplasm germ had to creep out of the detachable sheath to become freely motile (Fig. 10). The emerging mass was identified as the amoeboid primary cell by fluorescein-diacetate staining of its demarcating membrane (not shown) according to Markiw (1992).

Table 3. Mean measurements (µm) of Myxobolus parviformis triactinomyxon spores (31 individuals) from various oligochaete hosts.

Measurements Mean ± SD Minimum Maximum

Total length 127.0 ± 6.0 111.0 142.5 Polar capsules length 7.3 ± 0.9 5.8 9.5 width 3.9 ± 0.4 3.7 4.7 Sporoplasm length 26.7 ± 3.1 20.0 34.2* width 11.8 ± 1.4 9.5 14.7 Spore body length 33.9 ± 3.1 28.9 42.1 width 11.8 ± 1.4 9.5 14.7 Style length 100.3 ± 6.3 84.7 112.6 width 10.3 ± 1.0 6.3 12.1 Caudal processes length 101.7 ± 6.8 86.2 118.3 width 9.3 ± 1.1 6.3 12.1

- 47 - -III. Results-

Figure 8. Line drawing of the triactinomyxon spore of Myxobolus parviformis sp. n.; bar, 50 µm.

Myxospores: For detailed proportions see Table 4. Development in subspherical plasmodia (< 300 µm), presumably of vascular type (Fig. 9 (C)), in secondary gill lamellae. Lens-like spores with two valves, biconvex in sutural view, reverse ovoid proportions in valvular view, tapering posteriorly (Fig. 9 (D)). Around posterior sutural ridge usually four (maximum six) sutural edge markings. Two pyriform polar capsules of equal size, slight inward inclination of about 25°, five filament coils (8 individuals examined), occupy about half of the interior space (Fig. 11). Triangular intercapsular process present between capsules. No mucus coat, surface ridges or iodinophilous vacuole was observed.

- 48 - -III. Results-

Figure 9. (A) Uncompressed fresh mount of a Myxobolus parviformis sp. n. triactinomyxon spore; note the location of the valve cell nuclei in caudal process regions; bar, 40 µm. Inset: Apical region of the style; bar, 10 µm. (B) Apically opened triactinomyxon with released sheath unit; bar, 30 µm. Inset: Uncompressed fresh mount of intact sheath unit containing the sporoplasm, polar filaments are extruded (arrowhead); bar, 10 µm. (C) Mature Myxobolus parviformis plasmodia (pseudocysts) in Abramis brama gill filaments; bar, 500 µm. (D) Myxobolus parviformis myxospore from dissected pseudocyst from gill tissue of experimentally infected Abramis brama; bar, 5 µm.

- 49 - -III. Results-

Figure 10. Wet mount of a Myxobolus parviformis sp. n. triactinomyxon spore. Emergence of the sporoplasm after spore valve opening and sheath unit emission upon incubation in crude homogenized bream skin mucus; bar, 50 µm. (A) Inner sheath opening after polar filament discharge (arrowhead). (B) Sporoplasm release from inner sheath (arrowhead). (C) Empty inner sheath carrying the three polar capsules; note sporoplasm cells (arrowhead) emerging from the primary cell.

Figure 11. Line drawing of the myxospore of Myxobolus parviformis sp. n.; bar, 5 µm.

- 50 - -III. Results-

Table 4. Mean measurements (µm) of Myxobolus parviformis sp. n. myxospores from Abramis brama gill lamellae (33 individuals examined).

Measurements Mean ± SD Minimum Maximum

Total length 11.2 ± 0.5 9.9 12.1 Polar capsules length 5.1 ± 0.2 4.6 5.8 width 3.3 ± 0.3 2.7 3.8 Spore body length 5.1 ± 0.6 4.2 6.7 width 9.6 ± 0.3 8.7 10.4 thickness+ 7.2 ± 0.7 6.1 8.5 +20 individuals examined

2. Viability assay

FDA/PI double-staining showed that averages of 48.5% (± 3.3) were fully viable on the cellular level of the spores that were expelled less than 24 h before filtration. Single discharged or damaged polar capsules accounted for more than 90% of the not fully viable individuals (202 spores counted, 3 replicates). In virtually all actinospores the sporoplasms including the secondary cells were viable, about 5% showed dead amoeboid germ cells along with considerable damage of the spore valves. As a conclusion, 48 h old actinospores were assumed to be suitable for use in the discharge experiments.

3. Polar filament discharge

3.1. Visualisation of polar filament discharge

Video analysis of polar filament discharge showed that the extrusion takes place in less than 10 msec (Fig. 12 (A)-(B)). Most notably, the polar filament first elongates, then retracts shortly after discharge to about half the length (depends on the degree of eversion extension) in less than one second (Fig. 12 (D) and (E)). The actinospores thus have developed an excellent mechanism to pull their apical region tightly towards the host surface, enabling the sporoplasm to directly enter the epidermal layer. This ‘dragging mechanism’ could be observed to pull the whole spore body several µm along the slide by their sticky filament.

- 51 - -III. Results-

Figure 12. Polar filament discharge of Myxobolus cerebralis actinospores induced by addition of 30% ammonia. (A) Actinospore apex with polar capsules (arrowhead). (B) The same spore having discharged one filament (arrowhead) 0.01 s later. (C) Discharge of a second filament 0.01 s later. (D) Actinospore during polar filament discharge showing elongation of the filament. (E) Retracted discharged filaments of the same spore 10 s later. Bar, 20 µm, counter included from the original footage shows ms (small scale) and s (large scale).

- 52 - -III. Results-

3.2. Mechanical and chemical stimuli

The subsequent exposure of 47 individual spores to host mucus and mechanical stimulation revealed a combination of sensitising chemostimuli and a following touch signal to elicit polar filament discharge. In a first step, water was applied, which caused no discharge of polar filaments. Thereafter, the same actinospores received the mechanical stimulation, where one of the 47 actinospores reacted. The ensuing mucus application alone did not cause discharge in any specimen. Among hitherto non-extruded spores that were fully covered in trout mucus at this step, 54% of the tested individuals extruded one or more filaments only after further mechanostimulation (25 spores, significant change of response vs. subsequent control/mechanostimulus/trout mucus-application, P < 0.001, McNemar χ²-procedure). In the control trial, where water replaced trout mucus prior to the second mechanostimulus, discharge of the polar capsules was not stimulated; only two individuals reacted after the first mechanostimulus and none after the second mechanostimulus (17 spores, P < 0.001 vs. the same step in the mucus trial, McNemar χ²-procedure).

3.3. Experimental conditions for bulk experiments

In the bulk experiments, the hypo-osmolality of control samples (mean osmolality 18 mosm/kg ± 0.7 when buffered with 5 mM PBS) may inhibit polar filament discharge, leading to false results. Therefore, water from spore filtrations (mean osmolality 27 mosm/kg ± 0.0 when buffered with 5 mM PBS) was tested. The reaction rate was equal to that with a deionised water-based control (P = 0.97). In a modified bulk experiment (less total volume), a frequency dependency of the responses was not observed, but trout mucus raised the reaction rates by about 20% over the whole frequency range (Fig. 2; P < 0.05 with mucus vs. control at all frequencies). Interestingly, an increased discharge rate could already be noted when no vibrations, but mucus was applied (discharge rate at 0 Hz, Fig. 13).

- 53 - -III. Results-

M E Discharged actinospores [%] ± S

Vibration frequency [Hz]

Figure 13. Effect of vibration frequency on polar filament discharge rates (%) by Myxobolus cerebralis actinospores after exposure to trout mucus (solid symbols) and water (open symbols) in the bulk experiment (Actinospores (16 µl) and substrates (4 µl) were mixed on a slide); 285-493 actinospores counted per frequency, 7 replicates.

3.4. Ca2+ dependency

H. nuesslini actinospores discharged their polar filaments less frequently when the surrounding medium (‘SFW’) was Ca2+-deficient (P < 0.05, Table 5). Actinospores in artificial medium with Ca2+ showed a significant discharge rate (P < 0.001). When the osmolality of the Ca2+-deficient medium (7 mosm/kg) was adjusted to the value of Ca2+-containing water (12 mosm/kg), the discharge rate was indiscernible from that measured in Ca2+-containing water (P = 0.57, Table 5). Nevertheless, the responses in Ca2+-deficient water (with and without adjusted osmolality) were not significantly different from the respective controls (without mucus addition) (P = 0.3 and 0.08 respectively). The responses to controls (without mucus) differed among the media (Table 5). Therefore, the discharge rates obtained in controls were subtracted from the responses caused by mucus addition in the respective media. As a consequence, the discharge rate in Ca2+-containing water was significantly higher than that in Ca2+- deficient water with adjusted osmolality (P < 0.05). The responses in both Ca2+-free media were not significantly different from each other (P = 0.7).

- 54 - -III. Results-

Table 5. Effect of Ca2+-ions on polar filament discharge of Henneguya nuesslini actinospores during response to carp mucus homogenate (1 mg/ml). The actinospores were transferred to artificial medium (‘Standard fresh water’ (SFW) according to Meier-Brook 1978). NaCl was used for osmolality equilibration of Ca2+-deficient water. Actinospores and substrates were mixed on a slide (21 + 9 µl) with application of vibrations (3 s, 50 Hz, Amplitude 0.5 mm). Letters indicate a significant difference (ab, P ≤ 0.001, ABCD, P ≤ 0.05 Tukey HSD). Total number of actinospores counted per substrate 286-398 (8 replicates).

Substrate Discharge rate [%] SEM

SFW + Ca2+ 21.3 a + 3.5/- 3.3 SFW + Ca2+ + mucus 42.1 abAB + 4.4/- 4.3 SFW - Ca2+ 15.0 bCD +2.4/- 2.3 SFW - Ca2+ + mucus 26.8A +3.3/- 3.1 SFW - Ca2+ + NaCl 23.9B +2.7/- 2.6 SFW - Ca2+ + NaCl + mucus 33.5CD +3.5/- 3.4

3.5. Host specificity

3.5.1. Myxobolus cerebralis

When trout mucus homogenate was offered in different concentrations (Fig. 14), it became obvious that there was no difference in discharge rates using 1 mg/ml or 5 mg/ml (P = 0.99). Although all concentrations were significantly different from control (P < 0.05), fewer spores reacted at 0.1 mg/ml (P = 0.07 vs. 1 mg/ml). Carp mucus triggered polar filament discharge as effectively as trout mucus (Table 6 (A); P = 0.74 vs. trout mucus). Bream mucus was similarly effective (P = 0.63 vs. trout mucus at the same concentration), though it did not reach the same discharge percentage as trout mucus, even at higher concentrations (Table 6 (B)).

- 55 - -III. Results-

M E Discharged actinospores [%] ± S

Concentration [mg/ml]

Figure 14. Effect of trout mucus homogenate concentration on discharge rates of M. cerebralis actinospores in the bulk experiment (Actinospores (16 µl) and substrates (4 µl) were mixed on a slide); 80-151 actinospores counted per substrate in 5 replicates.

Freshwater snail mucus seemed to lack the required triggers (Table 6 (C)). Lymnaea-mucus neither stimulated in the same dry weight concentration as trout mucus, nor when adjusted to the same carbohydrate content as trout mucus (Table 6 (D); P = 0.002) or even higher concentration. Amphibian epidermal substrate (Rana epidermis) stimulated discharge only at increased dry weight concentration (Table 6 (E). Bovine submaxillary mucin was a weaker chemostimulant (2 mg/ml, P = 0.25 vs. control, 1.00 vs. frog (1 mg/ml), 0.19 vs. frog (5 mg/ml) and 0.03 vs. trout mucus) and increased concentrations led to even fewer responses (Table 6 (F); 5 mg/ml, P = 0.678 vs. control, P = 0.038 vs. frog (5 mg/ml)). To designate the recognised signal as mucus-specific, a supernatant of homogenised trout muscle tissue was offered. This substrate caused relatively low discharge rates even at very high concentrations (5 mg/ml, P = 0.32 vs. trout mucus Table 6 (G)).

- 56 - -III. Results-

Table 6. Host specificity of polar filament discharge of Myxobolus cerebralis actinospores using host- and non-host mucus homogenates (method see legend of Table 5). Letters indicate a significant difference (abcde, P ≤ 0.001, ABCD, P ≤ 0.05; Tukey HSD). Total number of individuals counted per substrate (A) 216-355 (7 replicates), (B) 713-1179 (6 replicates), (C) 290-481 (6 replicates), (D) 209- 337 (12 replicates), (E) 515-674 (6 replicates), (F) 657-733 (6 replicates), (G) 546-928 (8 replicates).

Substrate Discharge rate [%] SEM

(A) Non-host fish mucus (Cyprinus carpio) Carp mucus 1 mg/ml 46.8b + 5.2/- 5.3 Trout mucus 1 mg/ml 51.7a + 5.2/- 5.1 Control 20.3ab + 3.1/- 2.9 (B) Non-host fish mucus (Abramis brama) Bream mucus 1 mg/ml 48.0 + 8.8/- 8.8 Bream mucus 2 mg/ml 51.0 + 7.1/- 7.2 Bream mucus 3 mg/ml 55.7B + 6.6/- 6.7 Trout mucus 1 mg/ml 61.8A + 7.3/- 7.6 Control 27.2AB + 5.7/- 5.3 (C) Snail mucus (Lymnaea stagnalis) Snail mucus 1 mg/ml 13.8b + 7.7/- 6.9 Trout mucus 1 mg/ml 51.0ab + 2.9/- 2.9 Control 11.4a + 1.7/- 1.3 (D) Snail mucus (concentrated) Snail mucus 1.56 mg/ml+ 23.0b + 2.7/- 2.6 Snail mucus 5 mg/ml 30.8A + 5.2/- 4.9 Trout mucus 1 mg/ml 53.2abA + 3.6/- 3.6 Control 22.0a + 3.8/- 3.6 (E) Frog epidermis (Rana esculenta) Frog skin homogenate 1 mg/ml 39.0A + 4.6/- 4.5 Frog skin homogenate 5 mg/ml 59.2b + 4.8/- 4.9 Trout mucus 1 mg/ml 65.7aA + 5.9/- 6.2 Control 25.1ab + 3.1/- 2.9 (F) Bovine submaxillary mucin Bovine submaxillary mucin 2 mg/ml 41.1A + 7.3/- 7.1 Bovine submaxillary mucin 5 mg/ml 35.6B + 5.3/- 5.1 Trout mucus 1 mg/ml 65.7aAB + 5.9/- 6.2 Control 25.1a + 3.1/- 2.9 (G) Host muscle tissue Trout muscle 1 mg/ml 32.0 + 6.4/- 6.2 Trout muscle 5 mg/ml* 39.0 + 8.8/- 8.3 Trout mucus 1 mg/ml 61.2A + 8.3/- 8.9 Control 31.5A + 7.8/- 7.4 +Adjusted to same total carbohydrate content as trout mucus; *Desalted by ultrafiltration (MWCO 3 kDa) - 57 - -III. Results-

Bovine submaxillary mucin was a weaker chemostimulant (2 mg/ml, P = 0.25 vs. control, 1.00 vs. frog (1 mg/ml), 0.19 vs. frog (5 mg/ml) and 0.03 vs. trout mucus) and increased concentrations led to even fewer responses (Table 6 (F); 5 mg/ml, P = 0.678 vs. control, P = 0.038 vs. frog (5 mg/ml)). To designate the recognised signal as mucus-specific, a supernatant of homogenised trout muscle tissue was offered. This substrate caused relatively low discharge rates even at very high concentrations (Table 6 (G)).

3.5.2. Henneguya nuesslini

Although rainbow trout have not been shown to be susceptible hosts for H. nuesslini, homogenate of their mucus was used to ensure comparability with other experiments. The actinospores of this myxozoan species also responded to mucus homogenate of non-susceptible fish (Table 7, carp mucus homogenate vs. control, P < 0.05). Bovine submaxillary mucin and snail mucus were without effect (P < 0.001 vs. trout mucus and carp mucus).

Table 7. Host specificity of polar filament discharge of Henneguya nuesslini actinospores using host- and non-host mucus homogenates (1 mg/ml) (method see legend of Table 5). Letters (abcdef) indicate a significant difference (P ≤ 0.001, Tukey HSD). Total number of individuals counted per substrate 267- 375 (7 replicates).

Substrate Discharge rate [%] SEM

Rainbow trout mucus 50.2abc + 4.7/- 4.7 Carp mucus 52.0def + 5.4/- 5.4 Bovine submaxillary mucin 17.9ad + 3.6/- 3.3 Snail mucus 22.6be + 3.0/- 2.9 Control 19.7cf + 2.5/- 2.3

3.5.3. Myxobolus parviformis

M. parviformis is a myxozoan species that is very likely specific for bream and possibly other Leuciscinae (Cyprinidae). Therefore it was not surprising, that the actinospores were not able to distinguish between bream and carp mucus by polar filament discharge activation, but they responded weaker to trout substrate (Table 8, P < 0.05 vs. control). The response to trout mucus increased to over 40% when its

- 58 - -III. Results- concentration was doubled (data not shown). In sum, M. parviformis actinospores showed comparatively low reaction rates that massively fluctuated during the experiments, so that even the response to host (bream) mucus did not significantly differ from control.

Table 8. Polar filament discharge rates of Myxobolus parviformis actinospores in response to host- and non-host substrates (1 mg/ml) (method see legend of Table 5). Letter (a) indicates a significant difference (P ≤ 0.001, Tukey HSD). Total number of individuals counted per substrate 211-267 (11 replicates).

Substrate Discharge rate [%] SEM

Rainbow trout mucus 28.4 + 3.8/- 3.6 Carp mucus 44.2a + 4.2/- 4.2 Bream mucus 32.3 + 6.3/- 6.0 Control 17.7a + 3.6/- 3.3

3.6. Analysis of chemical signals for polar filament discharge

3.6.1. Ultrafiltration

The molecular size of the chemostimulus could not be satisfactorily determined. When mucus homogenate was split into molecular size classes using a MWCO of 3 kDa, the stimuli were present in both MW fractions. Only a mixture (‘refractionated’) of both fractions could restore the full effect (Table 9 (A)). Ultrafiltration using MWCO volumes of 10 and 30 kDa did not yield further information; the stimulating compounds were similarly distributed among the filtrates and the retentates (Table 9, (B), data for MWCO 10 kDa not shown). Although the discharge rate using mucus < 30 kDa was equal to that of untreated mucus at the same concentration (P = 1.0 vs. untreated substrate), the refractionated mixture gained more than a 20% increase in the number of discharged actinospores.

- 59 - -III. Results-

Table 9. Molecular weight of polar filament discharge triggers for Myxobolus cerebralis actinospores obtained by ultrafiltration of trout mucus (1 mg/ml) (method see legend of Table 5). Letters (abcde) indicate a significant difference ((A) P ≤ 0.001, (B) P ≤ 0.05 Tukey HSD). Total number of individuals counted per substrate (A) 166-365 (10 replicates), (B) 305-408 (6 replicates).

Substrate Discharge rate [%] SEM

(A) MWCO 3 kDa Mucus untreated 51.8ab + 5.1/- 5.1 > 3 kDa 31.1c + 3.3/- 3.2 < 3 kDa 22.6ad + 5.3/- 4.9 Refractionated 63.4cde + 5.6/- 5.8 Control 23.6be + 3.3/- 3.1 (B) MWCO 30 kDa Mucus untreated 39.5a + 7.3/- 7.1 > 30 kDa 29.8 + 6.6/- 6.2 < 30 kDa 38.5 + 3.0/- 3.9 Refractionated 50.3b + 7.1/- 7.1 Control 17.4ab + 3.6/- 3.3

3.6.2. Lipids

The average content of lipids in trout mucus as obtained by the Folch method was 12% of the dry weight (4 extractions, minimum 9%, maximum 16%). The discharge triggering activity of hydrophilic (from ether-extraction III. 7.7.1.) and lipophilic (from Folch-extraction III. 7.7.2.) components resembling the respective content in untreated mucus homogenate (including an estimated 30% loss during extraction) was compared with a mixture of both extracts. Both fractions showed increased triggering activity (not significant, Table 10 (A)), but only a refractionated mixture was as effective as untreated mucus homogenate (Table 10 (A), P = 0.98 vs. untreated substrate). The discharge triggering substances were quantitatively split into both fractions, as the initial effect could be restored by concentration adjustment (Table 10 (B)). Similar results were obtained with other lipid extractions as well but discharge rates observed varied greatly throughout the experiments.

- 60 - -III. Results-

Table 10. Role of lipids from trout mucus (corresponding to 1 mg/ml initial dry weight) in polar filament discharge of Myxobolus cerebralis actinospores in response to (method see legend of Table 5). 2x, double concentration corresponding to a 2 mg/ml dilution of untreated trout mucus. Letters indicate a significant difference (abcd, P ≤ 0.001, ABCD, P ≤ 0.05 Tukey HSD). Total number of individuals counted per substrate (A) 172-287 (6 replicates), (B) 160-363 (6 replicates).

Substrate Discharge rate [%] SEM

(A) Lipids as contained in mucus homogenate Mucus untreated 69.3aA + 5.9/- 6.2 Lipid fraction 48.5B + 7.0/- 7.0 Hydrophilic fraction 45.7AC + 5.1/- 5.0 Refractionated 73.3bBC + 4.7/- 5.0 Control 26.5ab + 4.2/- 4.0 (B) Comparison with concentrated lipids/hydrophilics Mucus untreated 36.6aA + 3.7/- 3.6 Lipid fraction 22.5ABCD + 3.7/- 3.5 Hydrophilic fraction 26.5c + 1.8/- 1.8 Lipid fraction 2x 39.6dD + 3.1/- 3.1 Hydrophilic fraction 2x 38.8C + 2.3/- 2.3 Refractionated 36.6bB + 4.3/- 4.2 Control 17.3abcd + 1.7/- 1.6

3.6.3. Amino compounds

Free amino acids (AA) as contained in mucus homogenate were mixed from pure substances according to HPLC results (III. 9.8.). Even twice the concentration of the determined amino acid pattern could not cause discharge in M. cerebralis actinospores (mucus homogenate 49.9%, AA mixture 16.7%, P = 1 for AA mix vs. control; total number of individuals counted per substrate 535-737, 6 replicates). Higher concentrations were equally ineffective (data not shown). In another approach, amino groups were irreversibly derivatised by fluorescamin. To exclude the possibility that actinospore chemoreception is impaired by the reagent itself, an aliquot of actinospores was pre-incubated in control substrate containing fluorescamin for 30 min and tested with mucus homogenate. The derivatisation did not lower the discharge rate (Table 11, P = 1 vs. similarly treated substrate) and did not affect spore reactivity.

- 61 - -III. Results-

Table 11. Effect of blocking of amino goups (trout mucus (1 mg/ml) treated with fluorescamin) on polar filament discharge of Myxobolus cerebralis actinospores (method see legend of Table 5). Letters indicate a significant difference (a, P ≤ 0.001, AB, P ≤ 0.05 Tukey HSD). Total number of individuals counted per substrate 251-644 (6 replicates).

Substrate Discharge rate [%] SEM Mucus untreated (actinospores + fluorescamin)+ 56.6 + 6.0/- 6.1 Mucus untreated 52.1a + 7.6/- 7.7 Mucus + fluorescamin 43.2B + 3.7/- 3.7 Mucus similarly treated 44.2A + 3.1/- 3.1 Control 19.9aAB + 1.2/- 1.2 + Pre-incubated actinospores; 3 replicates before the experiment, 154 individuals counted.

3.6.4. Nucleotides

The nucleotide base guanine (abundant in fish epidermis and responsible for the “silvering”) had no discharge triggering effect at a final concentration of 0.1 mg/ml, (discharge rate 29.0% ± 1.6, P = 0.98 vs. control, 372 individuals counted in 6 replicates) when mixed with the same concentration of inosine. The discharge rate was also not significantly different from that caused by trout mucus homogenate < 3 kDa (discharge rate 31.4% + 7.2/- 6.7, P = 1). However, the activity of untreated mucus homogenate could not be reached (discharge rate 44.3% ± 5.2, P = 0.26). Inosine, a predominant compound in the discharge triggering subfraction obtained by LichroPrep RP 18 gradient chromatography, had no effect at a final concentration complying with the calculated content of inosine in trout mucus homogenate < 3 kDa (concentration calculated from NMR data; 66% of whole peak dry weight, III. 5.11.; concentrated corresponding to 2 mg/ml initial dry weight) (Table 12). Although the effect of its mono-phosphate derivative (IMP) was not significantly different from control, it was significantly different from the rate observed with the final nucleotide degradation product, uric acid (same molarity as inosine) (Table 12). This shows that inosine in its standard nucleoside configuration is not the effective molecule for triggering discharge and that phosphorylation is not sufficient to yield considerably higher discharge rates.

- 62 - -III. Results-

Table 12. Effect of nucleoside derivatives (as contained in trout mucus homogenate < 3 kDa of 2 mg/ml initial concentration) on polar filament discharge of Myxobolus cerebralis actinospores (method see legend of Table 5). Inosine was concentrated as present in trout mucus homogenate < 3 kDa at an initial concentration of 2 mg/ml). Letters indicate a significant difference (abcdefg, P ≤ 0.001, AB, P ≤ 0.05 Tukey HSD). Total number of individuals counted per substrate 183-257 (6 replicates).

Substrate Discharge rate [%] SEM

Mucus untreated 63.4abcd + 4.6/- 4.7 Trout mucus homogenate < 3 kDa 57.0efg + 2.6/- 2.7 Inosine 36.6bfA + 2.3/- 2.3 5’-inosine monophosphate 42.8cB + 0.4/- 0.4 Uric acid 22.5dgAB + 3.7/- 3.5 Control 31.1ae + 1.8/- 1.8

3.6.5. Carbohydrates

The amino sugar N-acetylgalactosamine is a terminal carbohydrate structure that comprises up to 25% of total sugars of rainbow trout mucus glycoprotein (Nakagawa 2001). As pure substance (17.5 µg/ml final concentration), this sugar did not cause a significant response with actinospores (discharge rate 13.0%, P = 0.001 vs. mucus homogenate). As the GC/MS analysis of mucus homogenate < 3 kDa showed (III. 5.10.3.), glucose is the major free hexose present. Its discharge triggering activity was tested including the measurements for rainbow trout mucus by other authors. Ferguson (1992) had calculated the highest content of 1.53 mg/ml in his mucus preparation. Roberts and Powell (2004) detected a glucose concentration of 0.03 mg/ml in his substrate and put this also in proportion to protein content (yielding an amount of 0.53 mg/ml for the crude mucus homogenate (9.22 mg/ml) used in this study). None of these concentrations could reach a significant discharge rate (Table 13). Several neoglycoproteins (BSA-conjugates) were also offered (0.15 mg/ml final concentration). None of the glycoconjugates showed a discharge rate different from BSA alone (P > 0.98 vs. BSA, 388-458 individuals counted per substrate in 5 replicates).

- 63 - -III. Results-

Table 13. Effect of glucose on polar filament discharge of Myxobolus cerebralis actinospores (method see legend of Table 5). Letters indicate a significant difference (ab, P ≤ 0.001, A, P ≤ 0.05 Tukey HSD). Total number of individuals counted per substrate 126-212 (6 replicates).

Substrate Discharge rate [%] SEM

Mucus untreated 45.2abA + 3.9/- 3.9 Glucose as in Ferguson (1992) 32.2 + 5.1/- 4.9 Glucose as in Roberts & Powell (2004) 18.4b + 2.2/- 2.1 Glucose as in Roberts & Powell (2004)+ 21.2A + 4.2/- 3.9 Control 19.7a + 3.9/- 3.6 +Concentrated as calculated protein-glucose-ratio

To test for a discharge triggering effect of an osmolality similar to that of buffered mucus homogenate < 3 kDa, myo-inositol (the major free carbohydrate in mucus homogenate < 3 kDa as analysed by GC/MS) was chosen (2.94 mM/ml, 40.67 mosm/kg whithout buffer). Despite its very high concentration, the solution did not trigger discharge of polar filaments (P = 0.99 vs. control, 615 individuals counted in 6 replicates). Polyanionic GAGs and highly negatively charged glycoproteins were removed from trout mucus by alcian blue. The bulk experiment with this substrate was conducted with H. nuesslini actinospores (369-499 individuals counted per substrate in 7 replicates), as no M. cerebralis actinospores were available at this time. Negatively charged glycans were no stimulating cues for polar filament discharge, as the supernatant from extracted substrate was as effective as similarly treated or untreated mucus homogenate (48.5% discharge rate, P ≤ 0.001 vs. control). Alcian blue dye in control did not show an effect per se (21.5% discharge rate, P ≤ 0.001 vs. untreated substrate). Deglycosylated high molecular mucus did not loose its ability to trigger polar filament discharge (Table 14, P = 0.98 vs. untreated mucus homogenate > 3 kDa) and remnants of reagents had no effect on the response of actinospores as shown by pre- incubation. Preparative deglycosylation of mucins with TFMS involved, despite TFMS, an array of chemicals that might interfere with actinospore chemoreception. Therefore, control and mucus substrates were treated as the deglycosylated probe (without addition of chemicals). A mucus sample receiving all chemicals except TFMS was not included as manifold reactions could be expected after reactant addition in absence of TFMS (neutralising solution, pyridine). A batch of actinospores was pre-incubated in

- 64 - -III. Results- control substrate for 30 min prior to use to exclude impairment of spore reactivity, followed by testing with untreated mucus homogenate. Deglycosylation was monitored using the sulphuric acid/resorcinol method. Neutral carbohydrates decreased by 75.5% in treated substrate > 3 kDa compared to similarly treated substrate (prior to dialysis).

Table 14. Effect of deglycosylation of trout mucus > 3 kDa (corresponding to 2 mg/ml initial dry weight) on polar filament discharge of Myxobolus cerebralis actinospores (method see legend of Table 5). + R, all reactants added. Letters indicate a significant difference (abcd, P ≤ 0.001, ABCD, P ≤ 0.05 Tukey HSD). Total number of individuals counted per substrate 242-483 (6 replicates).

Substrate Discharge rate [%] SEM

Mucus untreated (actinospores + control + R)+ 78.5 + 3.1/- 3.3 Mucus untreated 61.6abA + 6.2/- 6.4 Mucus untreated > 3 kDa 57.3cB + 5.3/- 5.4 Mucus similarly treated > 3 kDa 42.9A + 3.0/- 3.0 Deglycosylated mucus homogenate > 3 kDa 52.8CD + 4.2/- 4.2 Control + R 29.5acC + 2.5/- 2.4 Control 30.6bBD + 3.9/- 3.7 +Three replicates before the experiment, 154 individuals counted.

Sialic acids were released by acidic hydrolysis and the hydrolysate was chromatographically separated into 3 different fractions (Elution with water (FI), 1 M propionic acid (F II) and 2 M formic acid (F III). When the sialic acid content of these fractions was measured colorimetrically (quadruplicates) according to Bhavanandan & Sheykhnazari (1993), the fractions F II and F III showed a high content of released sialic acid (238 and 327 µg/ml respectively), whereas in the washing fraction F I and in a control chromatography of deionised water no sialic acid could be detected (threshold 2 µg/ml). A control treatment of bovine submaxillary mucin had a 15 x higher sialic acid signal in fractions F II and III than in F I. The pooled sialic acid fractions (F II + F III) did not stimulate discharge (P = 0.67 vs. F I) (Table 15 (A)). In a second approach, the preparations were concentrated according to the measured amount of freely available sialic acid in mucus < 3 kDa to avoid physiological impairment of the actinospores. The diluted sialic acid containing substrates could also not stimulate discharge (Table 15 (B)). A third test was conducted to exclude effects of the additional buffering that was necessary in the experiments. Adjusting the pH value to 6.9 (5 mM PBS buffer), the sialic acid containing fraction F II showed intermediate

- 65 - -III. Results- discharge triggering activity (Table 15 (C), P = 0.067 vs. control and 0.064 vs. untreated mucus). However, the increased response was not only due to the pH value as the reaction rate to control substrate indicated. Neuraminidase digestion of mucus homogenate was an effective method to remove most sialic acid residues. Monitoring of the high molecular substrates showed an OD decrease of 49.2% in BSM and 26.4% in trout mucus homogenate (both > 3 kDa). The digested substrate was equally effective in triggering discharge as similarly treated material (Table 15 (D), P = 0.99 vs. similarly treated substrate). Pre-incubation of spores (30 min) was necessary for the known receptor-destroying potency of the enzyme, but had no effect on the reaction. The pure sialic acids NeuGc (200 µg/ml absolute concentration) and 5,9 O-NeuAc (200 µg/ml absolute concentration, a sialic acid abundant in fish mucins and a kind gift from Prof. R. Schauer) were without effect (Table 15 (E), P = 1.0 and 0.99 vs. control respectively).

- 66 - -III. Results-

Table 15. Effect of sialic acids from trout mucus > 3 kDa on polar filament discharge of Myxobolus cerebralis actinospores (method see legend of Table 5). (A) (B) sialic acids released from mucus > 3 kDa (corresponding to 2 mg/ml; conc. < 3, concentrated to measured content in trout mucus < 3 kDa corresponding to 2mg/ml), (C) to mucus homogenate after neuraminidase digestion, (D) as pure substances (initial concentration in brackets) (pH of the substrates was adjusted with 17.5 mM of PBS in (A) and (B)). Letters indicate a significant difference (abc, P ≤ 0.001, ABCDEFG, P ≤ 0.05 Tukey HSD). Total number of individuals counted per substrate (A) 718-1365 (8 replicates), (B) 824-1499 (6 replicates), (C) 208-376 (6 replicates), (D) 301-601 (6 replicates), (E) 208-331 (6 replicates).

Substrate Discharge rate [%] SEM

(A) Sialic acids released from mucus > 3 kDa Mucus untreated 68.4abA + 4.5/- 4.6 F II + FIII 48.2AB + 3.3/- 3.3 F I 42.4a + 3.1/- 3.1 Control+ 33.8bB + 3.0/- 2.9 (B) Sialic acids released from mucus > 3 kDa (diluted) Mucus untreated 48.8abABC + 2.8/- 2.8 Mucus < 3 kDa 46.6DEFG + 3.7/- 3.7 F II + F III (conc. < 3) 34.9C + 3.3/- 3.2 F II + F III (conc. < 3) 10 x diluted 28.6bG + 3.0/- 2.9 F I (conc. < 3) 31.5AD + 1.3/- 1.3 F I (conc. < 3) 10 x diluted 31.0BE + 4.2/- 4.0 Control+ 29.3aF + 1.4/- 1.3 (C) Sialic acids released from mucus > 3 kDa (low pH) Mucus untreated 73.1ab + 4.3/- 4.6 F II pH 6.9 53.5 + 5.2/- 5.3 Water pH 6.9 37.0b + 5.7/- 5.5 Control 33.4a + 4.8/- 4.6 (D) Neuraminidase digested mucus Mucus untreated (actinospores + neuraminidase)* 70.1 + 4.1/- 4.2 Mucus untreated 43.4a + 5.2/- 5.1 Mucus > 3 kDa similarly treated 52.7b + 3.2/- 3.2 Desialylated mucus > 3 kDa 55.0c + 3.4/- 3.4 Control 25.1abc + 2.3/- 2.2 (E) Sialic acids as pure chemicals Mucus untreated 73.1abc + 4.3/- 4.6 NeuGc (200 µg/ml) 34.1b + 3.4/- 3.3 5,9 O-AcNeu (200 µg/ml) 37.7c + 5.2/- 5.0 Control 33.4a + 4.8/- 4.6 + Pooled fractions of a chromatography of deionised water * Pre-incubated actinospores; 3 replicates before the experiment, 154 individuals counted.

- 67 - -III. Results-

3.6.6. Proteins

Commercially obtained protein samples were tested for their discharge triggering activity. Bovine serum albumin did not cause a significant discharge rate even at 2 mg/ml final concentration (discharge rate 22.1% + 3.2/- 3.1, P = 0.8 vs. control, 579 individuals counted in 6 replicates). A mixture of two randomly chosen peptides (Val- Tyr-Val and Gly-Tyr; 0.1 mg/ml each) was also without effect (discharge rate 15.3% + 3.2/- 3.0, P = 0.67 vs. control, 386 individuals counted in 6 replicates). Digestion of trout mucus homogenate with Proteinase K and Pronase E could not withdraw discharge triggering activity from the high molecular (> 3 kDa) mucus homogenate (Table 16 (A) and (B), P = 0.36 and P = 0.39 respectively vs. similarly treated sample). A frequently occurring problem digesting the high molecular weight fraction (> 3 kDa) of mucus was the effect of the incubation conditions on the stimulating activity, decreasing the discharge rates caused by similarly treated substrates. Monitoring of the enzymatical cleavage showed a decrease in protein signal in all digested > 3 kDa fractions and proved the presence of serine-protease sensitive structures, as the protein content could be lowered by 65% with Proteinase K digestion. Proteinase K treatment of the ultrafiltration filtrate (MWCO 3 kDa) resulted in a lower discharge rate (not significant; (Table 16 (C), P = 0.46 vs. similarly treated substrate). The influence of peptidase digestion of the filtrate on the discharge triggering activity was difficult to assess due to the high response to the control substrate (Table 16 (D), P = 0.83 for untreated substrate vs. control). Most likely, the enzyme itself exerted a discharge triggering activity, as only the digested sample showed a significantly increased discharge rate (Table 16 (D)). The Biuret-reaction (quantification of peptide-bonds) and the Bradford-Coomassie method were unsuitable for monitoring due to a too high detection threshold. The absorbance spectra (200-320 nm) were not altered by enzymatic treatment of mucus homogenate < 3 kDa. Additionally, globular proteins from trout mucus were precipitated by acetone and removed by centrifugation. The discharge triggering compounds were only partially co- precipitated and remained also in the supernatant (Table 16 (E), both P > 0.15 vs. control). This suggests that the discharge triggering substances could bind or are attached to high molecular (protein) structures.

- 68 - -III. Results-

Table 16. Effect of mucus-derived proteins on polar filament discharge of Myxobolus cerebralis actinospores (method see legend of Table 5). (A), (B), (C), (D) digestion of trout mucus homogenate with proteolytic enzymes (< and > 3 kDa fractions are correspond to a concentration 2 mg/ml), (E) acetone precipitation (1 mg/ml). Letters indicate a significant difference (abc, P ≤ 0.001, AB, P ≤ 0.05 Tukey HSD). Total number of individuals counted per substrate (A) 492-779 (6 replicates), (B) 185-349 (7 replicates), (C) 284-511 (9 replicates), (D) 231-482 (9 replicates), (E) 39-67 (9 replicates).

Substrate Discharge rate [%] SEM

(A) Proteinase K digestion of mucus > 3 kDa Pre-incubated actinospores + untreated mucus+ 54.7 + 12.6/- 13.0 Mucus untreated 55.1aA + 5.6/- 5.6 Similarly treated > 3 kDa 44.5b + 4.8/- 4.7 Digested > 3 kDa 36.2cA + 4.9/- 4.7 Control 12.1abc + 1.3/- 1.2 (B) Pronase E digestion of mucus > 3 kDa Mucus untreated 61.5aA + 4.0/- 4.1 Similarly treated > 3 kDa 49.2b + 5.6/- 5.6 Digested > 3 kDa 37.5AB + 4.0/- 3.9 Control 16.9aB + 3.6/- 3.3 (C) Proteinase K digestion of mucus < 3 kDa Pre-incubated actinospores + untreated mucus+ 52.8 + 3.7/- 3.7 Mucus untreated 44.6A + 3.5/- 3.4 Similarly treated < 3 kDa 48.9a + 3.2/- 3.2 Digested < 3 kDa 37.8 + 4.6/- 4.5 Control 24.5aA + 2.6/- 2.5 (D) Peptidase digestion of mucus < 3 kDa Pre-incubated actinospores + untreated mucus+ 52.4 + 5.3/- 5.3 Mucus untreated 44.6 + 3.5/- 3.4 Similarly treated < 3 kDa 48.9 + 3.2/- 3.2 Digested < 3 kDa 56.1A + 3.4/- 3.4 Control 36.7A + 3.9/- 3.8 (E) Protein precipitation Mucus untreated 72.2aA + 4.9/- 5.2 Precipitate 46.3A + 10.5/- 10.2 Supernatant 46.0 + 7.8/- 7.7 Control 25.6a + 6.0/- 5.5 +Pre-incubation 30 min, three replicates before experiment, over 150 individuals counted each.

- 69 - -III. Results-

3.6.7. Stability and small molecular compounds

The stimulating activity of trout mucus decreased when it was incubated at 37 °C (2 h) and 60 °C (5 min) (Table 17 (A)). Lyophilisation did not affect the discharge triggering effect (Table 17 (A), P = 0.95 vs. untreated mucus homogenate). Although the discharge rate caused by mucus homogenate < 3 kDa that was heated to 100 °C was not significantly different from that by controls (Table 17 (B)), there was no difference between treated and untreated substrate (P = 1). Ashing of trout mucus homogenate, retaining only the inorganic electrolytes from mucus, resulted in total abolishment of discharge triggering activity (Table 17 (C). A discharge triggering effect of electrolytes from mucus could be rejected, as a mixture of salts present in rainbow trout mucus according to Handy (1989) was also not effective (Table 17 (D)). A mixture to investigate the possibility of combinatory effects of electrolytes and other compounds (complete mixture (CM), contents resembled quantitative measurements of various compounds in mucus homogenate < 3 kDa) did not cause a significant response, but the batch including both urea and ashed salts elicited an increased discharge rate (Table 17 (E)). Removal of the ashed salts (solution prepared with deionised water instead of ashed mucus) resulted in a drop of the discharge rate to control level (P < 0.05 vs. mucus homogenate < 3 kDa).

- 70 - -III. Results-

Table 17. Effect of (A) lyophilisation and heat, (B) boiling (mucus homogenate < 3 kDa), (C) electrolytes from ashed mucus, (D) electrolytes as pure chemicals (according to Handy 1989) and (D) a mixture of small molecular compounds (as contained in trout mucus < 3 kDa; 2 mg/ml) of trout mucus (1 mg/ml) on polar filament discharge rates of Myxobolus cerebralis actinospores (method see legend of Table 5). Letters indicate a significant difference (abcde, P ≤ 0.001, ABC, P ≤ 0.05; Tukey HSD). Total number of individuals counted per substrate (A) 452-684 (6 replicates), (B) 142-240 (6 replicates), (C) 535-753 (6 replicates), (D) 344-700 (6 replicates), (E) 543-654 (6 replicates).

Substrate Discharge rate [%] SEM

(A) Moderate temperature incubation Mucus untreated 46.8aAB + 2.1/- 2.0 Mucus lyophilised 49.4bcC + 3.4/- 3.4 Mucus 60 °C (5min) 36.0dAC + 3.1/- 3.0 Mucus 37 °C (2 h) 32.9beB + 1.0/- 1.0 Control 16.1acde + 1.8/- 1.7 (B) High temperature incubation Mucus untreated 69.7a + 5.9/- 6.3 Mucus < 3 kDa 100 °C (5 min) 47.4 + 4.3/- 4.3 Mucus < 3 kDa similarly treated 45.4 + 3.3/- 3.3 Control 29.4a + 1.9/- 1.9 (C) Ashed mucus homogenate Mucus untreated 49.9ab + 5.3/- 5.2 Mucus ashed 18.2a + 1.3/- 1.3 Control 17.1b + 2.4/- 2.3 (D) Electrolyte mixture Mucus untreated 50.7ab + 6.4/- 6.4 Electrolyte mixture 20.8a + 2.5/- 2.4 Control 20.9b + 2.9/- 2.7 (E) Mixture of small molecular compounds (CM)+ Mucus < 3 kDa 38.6AB + 2.3/- 2.3 CM 30.3 + 2.3/- 2.2 CM – ashed salts +urea 25.6B + 3.7/- 3.5 CM + ashed salts -urea 29.5 + 2.4/- 2.3 Control 23.8A + 2.1/- 2.1 +containing inorganic electrolytes (ashed mucus), N-acetyl-neuraminic acid, myo-inositol, scyllo-inositol, chiro-inositol, D-glucose, D-galactose, D-mannose, N-acetyl-glucosamine, N-acetyl-galactosamine, urea and a mixture of free amino acids as contained in mucus < 3 kDa.

- 71 - -III. Results-

3.6.8. Extraction with activated charcoal

Activated charcoal readily binds various hydrated (esp. aromatic) compounds due to hydrophobic interaction. Mucus homogenate < 3 kDa treated with activated charcoal (142-252 individuals counted per substrate in 6 replicates) lost its discharge triggering activity (discharge rate 28.6% ± 1.5, P = 1 vs. control; discharge rate caused by a similarly treated sample was 51.7% ± 8.1).

3.6.9. Chromatographic fractionation

AG18 A11resin retarded inorganic cations and anions from mucus homogenate < 3 kDa, that were subsequently eluted in 5 fractions by water. F I (not retarded compounds) did not show a significant discharge triggering activity (Table 18 (A)). While the weakly retarded compounds (pooled fractions F II-V) elicited a higher response, the column washing fraction F VI (strongly retarded ions) contained no stimulating compounds. The elevated discharge rate in the control is explained by the use of an increased PBS concentration (9 mM) to adjust the basic pH (due to complete hydrogen ion removal) of the fractions. Dowex 1 x 8 extraction of anionic compounds including organic substances (e.g. free sialic acids) from mucus homogenate < 3 kDa yielded three fractions (FI-III) in accordance with the final step during sialic acid extraction (II. 7.8.2.). The highest response was actuated by the first fraction (Table 18 (B)) indicating that either no binding to the column occurred or the triggers were degraded by acidic elution. For each fraction a control (deionised water as probe) was prepared and tested separately to exclude effects of reactant remnants and formic acid groups; a refractionated mixture was not included. MacroPrep High Q anion exchange support strongly binds anions even of organic nature (e.g. proteins). Chromatography of small molecular (< 3 kDa) mucus homogenate yielded a fraction of non-retained compounds (low salt) that contained effective stimuli for polar filament discharge (Table 18 (C), P = 0.94 vs. refractionated substrate and ≤ 0.001 vs. control and high salt eluate). The fact that the stimuli did not bind to the column was verified by rising the concentration 3-fold (Table 18 (D)), so the low salt fraction became as effective in triggering discharge as untreated substrate (P = 0.99).

- 72 - -III. Results-

Table 18. Effect of ionic compounds from trout mucus < 3 kDa on polar filament discharge of Myxobolus cerebralis actinospores separated by ion retardation chromatography with (A) AG11A8 resin (water as eluent) and ion exchange chromatography with (B) Dowex 1 x 8 (solvents for elution in brackets) (2 mg/ml referring to the initial dry weight) and (C) & (D) MacroPrep High Q (eluted with 50 mM ammonium acetate (low salt buffer) and 1 M ammonium acetate (high salt buffer)) (method see legend of Table 5). Letters indicate a significant difference (abcdefgh, P ≤ 0.001, ABCD, P ≤ 0.05, Tukey HSD). Total number of individuals counted per substrate (A) 341-614 (6 replicates), (B) 297-457 (6 replicates), (C) 692-1251 (6 replicates), (D) 266-407 (6 replicates).

Substrate Discharge rate [%] SEM

(A) AG11A8 chromatography fractions Mucus untreated 64.8abcdA + 4.1/- 4.2 Refractionated (F I – F V) 47.0b + 1.7/- 1.7 F I 42.7c + 1.8/- 1.8 F II-V 48.9AB + 2.6/- 2.6 F VI 40.7d + 3.1/- 3.0 Control+ 37.2aB + 1.8/- 1.8 (B) Dowex 1 x 8 chromatography fractions Mucus untreated 69.0abcdAB + 4.2/- 4.4 F I (water) 51.7A + 4.0/- 4.1 F II (1 M propionic acid) 47.2B + 3.3/- 3.3 F III (2 M formic acid) 38.5a + 2.0/- 2.0 Control F I (water) 38.9b + 4.7/- 4.6 Control F II (1 M propionic acid) 37.1c + 3.7/- 3.6 Control F III (2 M formic acid) 38.2d + 3.4/- 3.3 (C) MacroPrep High Q chromatography fractions (1 mg/ml) Mucus untreated 49.6abcd + 6.8/- 6.8 Refractionated (low salt + high salt fraction) 38.7aef + 1.5/- 1.5 Low salt fraction 33.2bgh + 3.6/- 3.5 High salt fraction 14.4ceg + 1.6/- 1.5 Control+ 15.9dfh + 1.2/- 1.2 (D) MacroPrep High Q chromatography fractions (3 mg/ml) Mucus untreated (1 mg/ml) 55.6AB + 7.7/- 7.8 Refractionated (low salt + high salt fraction) 44.0 + 4.2/- 4.2 Low salt fraction 57.7CD + 8.2/- 8.5 High salt fraction 28.9BC + 3.4/- 3.3 Control+ 24.5AD + 2.3/- 2.3 + Pooled fractions of a chromatography of deionised water

- 73 - -III. Results-

Stepwise column separation (SPE) of mucus homogenate < 3 kDa by hydrophobic interaction with LichroPrep RP 18 support yielded four fractions. The first contained polar molecules, the ensuing fractions were comprised of compounds that had bound to the alkyl-side chains, thus having a more hydrophobic character. Only F II showed a significant response (Table 19 (A)) and was the only fraction not significantly different from the refractionated mixture (P = 0.91). To achieve a better separation of the discharge triggering fraction F II, gradient liquid chromatography using a BioRad LC-system with the same bed support was conducted. The chromatograms showed luminescence and carbohydrate signals (Fig 15 (A)) as well as several UV-absorbing substances (Fig. 15 (B)). Further analysis of the fractions by TLC is included under III. 5.4. The fraction F V consisted of the column washing volume (50% 2-propanol) and was handled as an extra fraction. The highly UV-absorbing fractions were shown to contain the discharge triggering compounds. F III showed the maximum discharge rate (Table 19 (B), P = 0.8 vs. mucus homogenate < 3 kDa), the two adjacent fractions F II and F IV could also increase the discharge rate.

- 74 - -III. Results-

Table 19. Effect of fractions with different polarity obtained by (A) stepwise chromatography (SPE) of trout mucus < 3 kDa and (B) gradient chromatography of the SPE fraction F II from (A) with LichroPrep RP 18 on polar filament discharge rates of Myxobolus cerebralis actinospores (2 mg/ml referring to the initial dry weight, solvents for elution in brackets) (method see legend of Table 5). Letters indicate a significant difference (abcdef, P ≤ 0.001, ABCDE, P ≤ 0.05 Tukey HSD). Total number of individuals counted per substrate (A) 331-690 (6 replicates), (B) 759-847 (10 replicates).

Substrate Discharge rate [%] SEM

(A) Solid-phase extraction (SPE) of mucus < 3 kDa Mucus untreated 48.6abcdA + 6.3/- 6.3 Refractionated 34.8efBC + 2.7/- 2.6 F I (water) 18.7bB + 2.0/- 1.9 F II (10% 2-propanol) 29.4ADE + 3.9/- 3.8 F III (50% 2-propanol) 17.2cC + 2.1/- 2.0 F IV (100% methanol) 14.7deD + 2.7/- 2.5 Control+ 11.6afE + 0.7/- 0.7 (B) Gradient chromatography of the SPE-fraction F II Mucus < 3 kDa 49.4abA + 2.6/- 2.6 F I 33.0bB + 2.2/- 2.2 F II 41.5C + 2.9/- 2.9 F III 43.9cB + 3.0/- 2.9 F IV 41.3D + 3.0/- 3.0 F V 34.1A + 1.7/- 1.6 Control 26.7acCD + 3.1/- 3.0 +Pooled fractions of a chromatography of deionised water

- 75 - -III. Results-

A

I II III IV

B

Figure 15. Chromatograms of the fraction F II from the stepwise Lichroprep RP 18 SPE separated by gradient liquid chromatography. Fractions were elutet with deionised water up to an elution volume of 20 ml, then a linear gradient from 0-25% 2-propanol was used (flow rate 0.5 ml/min). (A) Amino group content was measured by ninhydrin staining, neutral carbohydrate content by the sulphuric acid- resorcinol method and fluorescence from images of emission values after excitation at 312 nm. (B) UV- absorbance at the given wavelengths. Lines indicate pooling into subfractions I-IV. All values are relative contents or absorbance.

- 76 - -III. Results-

4. Sporoplasm emission

4.1. Emission time course

The period the amoeboid sporoplasm requires to leave the valve structure after a preceding stimulation by mucus was determined for H. nuesslini actinospores. Most sporoplasms left the valves within 6 min, whereas most were emitting around 5 min (Fig 16).

Figure 16. Duration of the emission response of H. nuesslini sporoplasms after stimulation with mucus sorted in time categories representing the minute in which the emission movement was observed (64 actinospores counted). A normal distribution curve has been plotted without categorization showing the mean time of emission around 5.4 min (SD 2.8).

- 77 - -III. Results-

4.2. Dependence from discharge

The question whether actinospores can only be activated to leave their valve shell when polar filament extrusion has occurred was investigated by observing discharged and undischarged H. nuesslini actinospores for 15 min after stimulation by mucus homogenate. It was found that from the discharged actinospores, 60.6% kept their germ inside the spore valve (109 individuals counted), and that among non-discharged actinospores 28.4% released their sporoplasms (95 individuals counted). The Phi- value of 0.114 and the contingency coefficient of 0.113 (with a significance of 0.104) show that there is very weak correlation between the responses Additionally, according to Mantel-Haenszel`s chi-square analysis, there is no strong relationship between the two reactions (P = 0.14). Therefore, the sporoplasm emission reaction is generally independent from polar filament discharge (no linear correlation) in H. nuesslini actinospores. Nevertheless, 39.4% more sporoplasms were released, when polar filaments were discharged before, but release is not an imperative consequence after filament discharge.

4.3. Host specificity

Emission of M. cerebralis sporoplasms occurred also when mucus of non- susceptible host fish was given, although the number of emerging sporoplasms was low during the screening of the slides (< 5 min). For example, the emission rate after addition of bream mucus (1 mg/ml) was different from control (12.4% + 3.8/- 3.3, P < 0.001, 949 - 1179 individuals counted per substrate), but could not reach the rate measured in trout mucus (P < 0.05). H. nuesslini sporoplasms also left their valves when carp mucus (1 mg/ml) was offered (16.5% + 4.8/- 4.3, P < 0.001 vs. control 443 individuals counted).

4.4. Emission stimuli

To answer the question whether the stimulus for the activation of the sporoplasm movement is different from the discharge signal, sporoplasm emission rate to fractionated mucus homogenate was measured. Emission of H. nuesslini sporoplasms was observed in response to carp mucus < 3 kDa (14.8% ± 3.7, P < 0.001 vs. control

- 78 - -III. Results- and < 0.05 vs. > 3 kDa, 371-460 individuals counted in 6 replicates), trout mucus homogenate < 3 kDa stimulated some sporoplasms to emerge, but emission rate was low (3.2% ± 0.8, 1083 individuals counted in 6 replicates). Untreated rainbow trout mucus homogenate could elicit significant emission rates of M. cerebralis sporoplasms (27.5% ± 1.5, 3426 individuals counted in 26 replicates). In most treated mucus fractions, almost no sporoplasm emission was observed (emission never occurred in controls). In mucus > 3 kDa digested with Proteinase K, 4.3% ± 0.7 of the actinospores released their sporoplasm. This was significantly less than in the similarly treated substrate (9.7% ± 1.9, P < 0.001, 742 and 874 individuals counted in 6 replicates). After anion exchange chromatography (Macro Prep High Q), the low salt fraction caused sporoplasm emission (3.6% ± 0.8 compared to 0.1% ± 0.08 with the high salt fraction, P = 0.002, 636-770 individuals counted per substrate in 6 replicates), notably the same fraction that triggered discharge as well. Among the fractions from the LichroPrep RP 18 SPE the fraction that stimulated discharge (F II) had an effect on sporoplasm emission (5.0% ± 2.2, not significantly different from a mixture of all fractions, 636-770 individuals counted per substrate in 6 replicates).

5. Analyses

5.1. Substrate osmolality

As the substrate osmolality is a crucial factor for polar filament discharge and its underestimation might lead to a false interpretation of results, some values for selected solutions are shown in Table 20. Osmolality of mucus could only be measured using the small molecular fraction, as the viscosity of components > 3 kDa was too high to be measured by the method.

- 79 - -III. Results-

Table 20. Osmolality [mosm/kg] of solutions used in the bulk experiments measured by freeze-point lowering of 150 µl aliquots in triplicates (0 was deionised water, calibration with 400 mosm/kg NaCl). All values of mucus homogenates and fractions are calculated for a corresponding final concentration of 2 mg/ml unless otherwise noted. Buffered substrates contain 5 mM PBS.

Solution Osmolality SD

Actinospore medium (filtrate) water 8.25 1.09 Actinospore medium (filtrate) water (buffered) 27.00 0.00 SFW +Ca2+ 12.00 0.00 SFW -Ca2+ 7.00 0.71 Trout mucus homogenate < 3 kDa (buffered) 45.00 0.82 Snail mucus homogenate < 3 kDa 39.04 4.97 Control (deionised water, buffered) 18.00 0.82 CM - ashed salts + urea (buffered) 33.00 0.47 CM + ashed salts + urea (buffered) 48.00 0.47 BSA-Lac (buffered)* 12.00 0.47 BSA- Glc (buffered)* 15.75 0.50 Amino acid-mixture (free AAs as in trout mucus) 3.20 0.82 Trout muscle tissue homogenate < 3 kDa 17.05 0.44 Myo-inositol solution (2.94 mM/ml) 40.67 0.47 Mucus homogenate treated with activated charcoal 27.30 0.25 Mucus homogenate not treated with activated charcoal 46.76 0.00 * Working solutions as used in the bulk experiment

5.2. Urea

The content of free urea was determined for different mucus homogenates. L. stagnalis mucus homogenate in the same dilution (7.25 mg/ml dry weight) contained 6.77 µg/ml urea compared to 5.99 µg/ml in trout mucus. Therefore, urea could be excluded as the discharge triggering agent in fish mucus. Carp mucus contained only 1.57 µg/ml, which could be a result of the higher density of carp mucus and its lower protein content at the same dry weight (III. 5.9.). The detected ammonia concentration (obtained by heat inhibition of urease) was 3.2 times higher in snail mucus than in trout mucus (28.5 µg/ml).

- 80 - -III. Results-

5.3. Lipids

Trout mucus homogenate showed a varying content of lipophilic compounds that constituted between 11.4 and 15.9% of the dry weight (5 extractions). The analytical TLC-separation of these lipids revealed numerous compounds belonging to several lipid classes. The hydrophobic phase of the Folch-extraction contained especially high amounts of cholesterol, fatty acids and even triglycerides (Fig. 17). Extraction with diethylether could not fully remove lipid components, as weak cholesterol and fatty acid signals were detected in the hydrophilic phase by TLC (lane 6, Fig. 17). Losses were avoided heating the dried lipid during solubilisation from 50 up to 70 °C (lane 3, Fig. 17).

1 2 3 4 5 6 7 8 9

Figure 17. TLC-separation of lipids in different extracts from trout mucus homogenate on silica gel plates

(silica gel 60, 0.25 mm) detected by charring at 160 °C after spraying with 50% H2SO4. Lane 1 and 9: References; lane 2: whole lipids from Folch-extraction (chloroform/methanol 2:1); lane 3: whole lipids from Folch-extraction without heating during solubilization; lane 4: lipids extracted with diethylether; lane 5: lipids extracted with a 1:1 mixture of diethylether and chloroform/methanol; lane 6: hydrophilic fraction after diethylether-extraction; lane 7: hydrophilic fraction after chloroform/methanol-extraction; lane 8: hydrophilic fraction after diethylether/chloroform/methanol-extraction. All lipid fractions isolated from 10 mg dry mucus. References (from bottom to top): Desoxycholic acid, dipalmitine, cholesterol (dark spot), fatty acids (lauric and palmitic acid), tripalmitine; 0.1 mg each. Solvent system: n-hexane, toluol, n- hexane/diethylether/acetic acid 70:30:1 (2x).

- 81 - -III. Results-

1 2 3 4 5 6 7 8 9 10 11 12

Figure 18. TLC-separation of polar lipids in extracts from trout mucus homogenate on silica gel plates

(silica gel 60, 0.25 mm) detected by charring at 160 °C after spraying with 50% H2SO4. Lane 1: whole lipids from Folch-extraction (chloroform/methanol 2:1); lane 2: hydrophilic fraction after diethylether- extraction; lane 3: lipids extracted with diethylether; lane 4: lipopolysaccharides; lane 5: glucocerebrosides; lane 6: ceramides; lane 7: phosphatidylcholine; lane 8: cholesterol; lane 9: desoxycholic acid; lane 10: monolauryl-glycerol; lane 11: 1,2-dipalmitoyl-sn-glycerol; lane 12: sphingomyelin. All lipid fractions represent isolated compounds from 6 mg dry mucus. Lane 4-12 (references) 0.1 mg each. Solvent system: Chloroform/methanol/water 60:25:4.

A special TLC for polar lipids (Fig. 18) showed that the phospholipids (brownish spots in the lower half) were not fully extracted by the ether extraction and could be seen in the hydrophilic phase as well. They were detected as strong spots in the lipid phase obtained by Folch-extraction. Mucus homogenate obviously did not contain glucocerebrosides, sphingomyelin and bile acids such as desoxycholic acid. Another TLC analysis was conducted to show whether lipophilic compounds are present in mucus homogenate < 3 kDa and in the discharge triggering fraction FII from the LichroPrep RP 18 SPE. Mucus homogenate < 3 kDa contained a significant amount of a lipophilic substance that shared their Rf value range with sphingo- and

- 82 - -III. Results-

phospholipids (Fig. 19). Surprisingly, no phosphatidylcholine and fatty acids could be detected. These substances were probably lost during ultrafiltration. F II did not contain lipophilic compounds and remained on the starting line like the peptide and inositol references. By spotting the same fractions on a TLC for full lipids, the luminescent compounds therein were neither free fatty acids nor other non-polar lipophilic substances (not shown).

Figure 19. TLC-separation of polar lipids (silica gel 60, 0.25 mm plates detected by charring at 160 °C after spraying with 50% H2SO4) showing lipid content in trout mucus homogenate fractions. Lane 1: sphingomyelin, lane 2: γ-globulins (human), lane 3: peptide (Gly-Tyr), lane 4: phosphatidylcholine, lane 5: myo-inositol, lane 6: phosphatidyl-inositol + 11,14 eicosadienic acid, lane 7: phosphatidyl-inositol, lane 8: glucocerebrosides, lane 9: trout mucus homogenate < 3 kDa (respective content from 750 µg initial dry weight), lane 10: F II from LichroPrep RP 18 SPE (respective content from 750 µg initial dry weight), lane 11: cholesterol. References are 0.1 mg each. Solvent system: Chloroform/methanol/water 60:25:4.

5.4. Hydrophilic TLC Analyses

To investigate the contents of the fractions obtained by LichroPrep RP 18 SPE, they were analysed in various hydrophilic TLC separations including numerous - 83 - -III. Results- reference substances. Most astonishingly, a pattern of luminescent compounds was observed in the discharge triggering fraction FII when irradiated with 312 nm on a UV- table (Fig. 20).

1 2 3 4 5 6 7 8 9 10 1112 1 2 3 4 5 6 7 8 9 10 11 12 13 14

A B

Figure 20.TLC-separation of LichroPrep RP 18 SPE fractions (silica gel 60, 0.25 mm plates) detected by

(A) charring at 160 °C after spraying with 50% H2SO4 following two solvent runs (B) transillumination at 312 nm (negative photography) after the first solvent run. Lane 1-7 (references, 0.1 mg each): N-acetyl- neuramini c acid and D-ribose (upper spot), glucose-6-phosphate, inosine, 5’-adenosine- monop hosphate, 5’-inosine-monophosphate, guanine, hypoxanthine; lane 8: trout mucus homogenate < 3 kDa (respective content from 750 µg initial dry weight), lane 9-12: F I, F II, F III, F IV from LichroPrep

RP 18 SPE (respective content from 1.5 g initial dry weight), lane 13: F II (preparative amount), lane 14: mycospo rine-like amino acid extract from Spirulina sp. Solvent system: Acetonitrile/water 70:30.

These components shared their Rf value range with pentoses (ribose), nucleosides like inosine and mono-phosphorylated adenosine. One spot in F II co-migrated exactly with inosine (Fig 20 (A)). This substance was the only one that was sensitive to charring by H2SO4 in F II. Hypoxanthine and guanine, very common nucleotides in fish epidermis, were neither luminescent under UV, nor did the solvent system move them from the starting line. Upon repeated development with the same solvent, the sharp demarcation of the five main spots (Fig. 20 (B)) decreased without achieving a better separation of the luminescent substances (not shown). Luminescent compounds were also detected in the not bioactive fraction F I, whereby these substances shared mobility with the MAA extract to a great extent.

- 84 - -III. Results-

1 2 3 4 5 6 7 8 9

A B

Figure 21. Single TLC-separation of the bioactive F II fraction of a LichroPrep RP 18 solid phase extraction (silica gel 60, 0.25 mm plates) detected by (A) charring at 160 °C after spraying with 50%

H2SO 4/Cu2SO4 (20%) (B) transillumination at 312 nm after the first solvent run (negative grey scale displ ay). Lane 1-6 (references, 0.1 mg each) from left to right: saccharose, dextran, D-glucose & D- lactose (lower spot), D-mannose, N-acetyl-glucosamine, glucuronic acid; lane 7 & 8: F II from LichroPrep

RP 18 SPE of trout mucus homogenate < 3 kDa (respective content from 550 µg and 3 mg initial dry weig ht respectively); lane 9: trout mucus homogenate < 3 kDa. Solvent system: Acetonitrile/water 70:30 (2x).

The compounds in the active fraction F II showed Rf values which differed from those of hexoses abundant in fish mucus (Fig. 21). Due to the higher resistance to acidic charring, the substances in F II could best be detected by UV prior to charring of the whole TLC-plate. Further references that were used in the TLC-analysis of the bioactive fraction F II were salmon sperm DNA, urocanic acid, trimethylamine N-oxide, different amino acid mixtures, acetylcholine and various fatty acids. None of the chemicals showed similar mobility or luminescence upon detection. In trout mucus homogenate < 3 kDa that was treated with activated charcoal, none of the luminescent compounds that were present in the bioactive fraction F II were detected by TLC (Fig 22).

- 85 - -III. Results-

1 2 3 4 5 6 7 8 9

Figure 22. TLC-separation of the trout mucus homogenate fractions (silica gel 60, 0.25 mm plates) detected by transillumination at 312 nm after the first solvent run (negative grey scale display). Lane 1:

11,14 eicosadienic acid; lane 2: F II from LichroPrep RP 18 SPE of trout mucus homogenate < 3 kDa (resp ective content from 1.5 mg initial dry weight); lane 3: F II from LichroPrep RP 18 SPE of trout mucu s homogenate < 3 kDa (concentrated as used in experiments); lane 4: 2-propanol extract from silica gel scrapings containing all UV-luminescent compounds from F II; lane 5: sialic acid preparation (F II & III concentrated as used in experiments); lane 6: sialic acid preparation (F I); lane 7: trout mucus homogenate < 3 kDa heated to 100 °C for 5 min (respective content from 300 µg initial dry weight); lane

8: mucus homogenate < 3 kDa treated with activated charcoal; lane 9: mucus homogenate < 3 kDa not treated with activated charcoal. References are 0.1 mg each. Solvent system: Acetonitrile/water 70:30.

Some nucleotides from fish skin can be extracted by Dowex 1 x 8 chromatography in its formiate form as well and thus might also be present in the sialic acid extracts.

TLC-analysis showed that compounds with similar Rf values were indeed present in the two sialic acid containing fractions F II and F III. However, they produced a smear over the whole lane extent (Fig. 22), indicating a possible degradation by heat or acid treatment during sialic acid isolation. Heating to 100 °C influenced the mobility properties of trout mucus homogenate < 3 kDa, the resulting components remained on the spotting line. As this substrate still triggered discharge, the heat-labile conformation of the effective substance seems to be irrelevant.

- 86 - -III. Results-

5.5. Gradient chromatography

Liquid gradient chromatography with LichroPrep RP 18 showed a high amount of amines and carbohydrate compounds in the early eluting fractions (elution volume 1- 25 ml) of mucus homogenate < 3 kDa (Fig. 23 (A) and (B)) followed by a wide peak of luminescent substances (elution volume 26-37 ml). The first peak area is likely comprised of neutral sugars and amino acids (also concluded from its absorbance at 295 nm) that are not expected to interfere with the column material greatly and thus are easily eluted with water. The luminescent peak could be derived from nucleic acid strands or peptides with aromatic amino acids. These signals were followed by six prominent carbohydrate peaks that could represent acetylated sialic acids, as these compounds become slightly hydrophobic upon O-acetylation. In the last third of the chromatogram (elution volume 61-69 ml), a large signal was present that showed strong luminance at 312 nm excitation, carbohydrate reactivity (resorcinol condensation) and absorbance at 250 nm. This was the component which triggered polar filament discharge by M. cerebralis actinospores and was the only signal present when the bioactive fraction F II from the LichroPrep RP 18 SPE was separated by the same method (Fig. 23 (C)). The substances eluting in these fractions (elution volume 61-72 ml) are always composed of a first large peak (61-65 ml) and one or two smaller subsequent luminescent signals of varying amount. In another separation, a similar pattern was observed, also yielding a smaller follow-up peak, but with intermediate peak area (not shown). In Fig. 23 (C), the subsequent peaks are very weak, indicating that the signal may disintegrate in some separations (e.g. by repeated lyophilisation), thereby forming the secondary signals. This is most remarkably shown in Fig. 15, where another fragment eluting closely after the first large signal was observed. The emerging second signal showed a characteristic absorbance at 259 nm and distinctively incorporated the carbohydrate signal of the whole compound. It is possible, that the large peak volume contains more compounds that were not separated.

- 87 - -III. Results-

A

B

C

Figure 23. Chromatograms of (A and B) trout mucus homogenate < 3 kDa (8.3 mg initial dry weight) and (C) the F II fraction from Lichroprep RP 18 SPE (16.5 mg initial dry weight) separated by gradient liquid chromatography with a BioRad LC-system. Solvents were 20 ml deionised water before a linear gradient from 0-25% 2-propanol (flow rate 0.5 ml/min). Analysis: amino groups: ninhydrin staining, neutral carbohydrates: sulphuric acid/resorcinol method, luminescence occurred upon excitation at 312 nm, UV- absorbance at the given wavelengths.

- 88 - -III. Results-

5.6. UV-Spectroscopy

Treatment of mucus homogenate < 3 kDa with activated charcoal caused a notable change in absorbance spectra (concentration corresponding to 4 mg/ml dry weight). A wide hunchback with a maximum absorbance at 295 nm (ranging from 250 to 320 nm) vanished after the extraction (Fig. 24). This spectral shift was also observed in the substrates that were used in the corresponding bulk experiment.

Figure 24. Absorbance spectra of trout mucus homogenate < 3 kDa (concentrated corresponding to 4 mg/ml initial dry weight) before (grey line) and after (bold line) treatment with activated charcoal.

5.7. HPIC-detection

High-performance anion-exchange chromatography (HPIC) was conducted to measure the content of free inositols in mucus homogenate < 3 kDa. Pulsed amperometric detection (IPAD) identified three inositol derivatives (Fig. 25 (A)) by replenishment calibration and could be quantified using standard dilution series. The probe contained 6.9 nM myo-inositol, 0.3 nM scyllo-inositol and 0.3 nM chiro-inositol per mg initial dry weight. Phosphorylated inositols (IPs) were not found by HPLC- MDD. HPIC of the F II fraction from Lichroprep RP 18 SPE showed four anionic signals by conductivity detection (Fig. 25 (B)) and two UV-absorbing compounds at 245 nm (not shown).

- 89 - -III. Results-

2 A 150

100 l [nC] a

Potenti 50

5 1 3 4

-10 0 10 20 30 40 50 8

B

6

4

2 Conductivity [µS]

-1 0 10 20 30

Elution time [min]

Figure 25. (A) HPIC-IPAD of trout mucus homogenate < 3 kDa treated with activated charcoal. Identified peaks were glycerine (contamination) (1), myo-inositol (2), scyllo-inositol (3), chiro-inositol (4), glucose (5); nC = nano-Coulomb. (B) HPIC-CD of the F II fraction from Lichroprep RP 18 SPE, large signal = phosphate; µS = micro-Siemens. Analysed amount 0.9 mg (A) and 1.2 mg (B) referring to the initial dry weight. - 90 - -III. Results-

By conductivity detection during HPIC-CD, a concentration of free phosphate (second peak in Fig. 25 (B)) of 114.2 nmol/mg could be determined in the F II fraction from Lichroprep RP 18 SPE. This value could also be confirmed for the ashed mucus homogenate.

5.8. Amino acids

The spectrum of free amino acids detected in trout and carp mucus homogenate is shown in Table 21. Neither free proline nor hydroxy-proline were detected, free cystein was also not present. The patterns of trout and carp substrate differ merely in the content of glutamic acid, glycine and taurine. A comparison between the analysed content of amino acids and the respective content in the artificial mixture made of pure substances is given in Fig. 26.

Figure 26. Free amino acids from trout mucus homogenate (µmoles per mg dry weight) and a mixture of pure AAs as analysed by RP-HPLC using OPA-derivatisation). = Mixture of pure substances = Analysed content in mucus sample.

- 91 - -III. Results-

Table 21. Content of free L-amino acids in mucus homogenates (nmoles/mg dry weight) as determined by RP-HPLC.

Amino acid Trout mucus SD Carp mucus SD

Alanine 57 6.2 76 10.5

Arginine 8 2.0 8 1.4

Asparagic acid 17 4.8 13 1.9

Asparagine 25 1.6 28 3.1

Glutamic acid 49 6.8 22 2.5

Glycine 89 13.9 60 4.6

Histidine 33 4.1 43 2.6

Isoleucine 12 1.8 11 1.8

Leucine 13 1.9 12 1.7

Lysine 22 2.3 20 0.1

Methionine 9 1.8 7 0.7

Ornithine 4 0.8 3 0.1

Phenylalanine 13 2.4 12 0.6

Phospho-ethanolamine 14 3.0 14 2.9

Serine 62 13.4 48 8.4

Taurine 265 5.8 166 9.2

Threonine 30 5.1 23 4.1

Tyrosine 8 1.4 9 2.5

Valine 18 3.6 15 1.3

5.9. Proteins

The Bradford-Coomassie assay was used to measure protein concentration in the mucus homogenates (Table 22). Trout substrate had the largest content of almost 70% protein of its dry weight. Snail substrate showed a comparatively small content of only 35% of its dry weight. With a protein ratio of 67.9% from its dry weight, the hydrophilic ether-extracted trout mucus homogenate maintained most of its protein content.

- 92 - -III. Results-

Table 22. Protein content in rainbow trout, common carp and Lymnaea stagnalis mucus homogenates (µg/mg dry weight) as measured by the Coomassie brilliant blue assay according to Bradford (1976).

Substrate Means SD

Trout 696.0 20.1 Carp 402.8 18.4 Snail 351.8 16.2

5.10. Carbohydrates

5.10.1. Neutral sugars

The condensation of resorcinol to hydrolysed sugars was used to obtain basic data on the glycosylation of mucus homogenate preparations. Snail mucus homogenate had the highest degree of glycosylation (Table 23), especially when compared with the relatively low protein content (III. 5.9.). Ultrafiltration of trout mucus homogenate resulted in a shift of 34.9% of resorcinol-reactive components into the filtrate < 3 kDa.

Table 23. Neutral sugar content (µg/mg dry weight) in rainbow trout, common carp and snail (Lymnaea stagnalis) mucus homogenates as measured by the sulphuric acid/resorcinol assay (Monsigny et al. 1988).

Substrate Means SD

Trout 42.8 3.9 Carp 39.0 2.5 Snail 79.3 3.7

In the hydrophilic ether-extracted trout mucus homogenate, 90.8% of the initial content of neutral sugars was still present. However, considering all other quantitative data on carbohydrates, these values cannot be explained by the amount of hexose alone. Substances including free sugars, pentoses (e.g. in nucleosides) and inositols significantly contributed to these data. The distribution of neutral carbohydrates was also monitored in some trout mucus fractions (Table 24).

- 93 - -III. Results-

Table 24. Relative sugar content (%) in different fractions from trout mucus homogenates as measured by the sulphuric acid/resorcinol assay.

Substrate Relative content (%)

(A) Anion-exchange chromatography Low salt 76.6 High salt 23.4 (B) LichroPrep RP 18 SPE F I 75.8 F II 18.0 F III 0.0 F IV 6.2 (C) Sialic acid extraction F I 19.7 F II 50.2 F III 30.1

5.10.2. Sialic acids

The colorimetric method of Bhavanandan & Sheykhnazari (1993) detected 6.5 nmoles (SD ± 0.05) sialic acid in trout mucus homogenate < 3 kDa and 14.9 nmoles in the high molecular fraction per mg dry weight (three independent measurements). Also, the low salt fraction of the anion exchange chromatography and the fraction F I of the LichroPrep RP 18 SPE contained sialic acids. In all other fractions of these separations, no sialic acid residues were found.

5.10.3. GC-MS for monosaccharides

Gas chromatography for monosaccharides showed different contents in small molecular fish and snail mucus homogenates (Table 25). The most abundant hexose in fish and also in snail mucus was glucose (in half the sugar/dry weight ratio). The ultrafiltrate of snail mucus homogenate contained no N-acetyl-glucosamine and galactose. When oligosaccharides within the substrates were hydrolysed, the content of amino sugars and glucose increased. In snail mucus homogenate, four not further identified pentoses were detected, whereas in trout and carp mucus one was found (presumably modified ribose according to MS spectra).

- 94 - -III. Results-

Table 25. Content of neutral hexoses in rainbow trout, common carp and Lymnaea stagnalis mucus homogenates < 3 kDa (nmoles/mg dry weight) as measured by GC/MS before and after acidic hydrolysis (H), n.d., not detected.

Sugar Trout Trout H Carp Carp H Snail Snail H

D (+) Mannose 0.35 0.37 0.13 0.15 n.d. 0.14 α-D (+) Glucose 4.90 5.70 3.97 6.35 0.55 2.19 D (+) Galactose n.d. 0.17 n.d. 0.16 n.d. n.d. N-acetyl-D-glucosamine 0.04 0.17 n.d. 0.13 n.d. n.d. N-acetyl-D-galactosamine 0.07 0.09 n.d. 0.23 n.d. 0.08

5.11. NMR-spectroscopy

The one-dimensional 1H-spectrum of the subfraction peak (elution volume 61-67 ml) from the LichroPrep RP 18 gradient chromatography demonstrated the presence of a major component (ca. 66% from spectrum integration). Comparison with literature data (SDBS data bank) identified this compound as inosine. Further typical signals showed a second nucleoside. With precaution, this second species might be adenosine, based on the 1H-chemical shifts given in literature. The one-dimensional 13C-spectrum (measuring time 28 h) showed the signals of the main component (inosine) ribose unit, as well as the signals of the two protonated positions of the N- heterocycles. Other signals were too weak and disappeared in the spectrum noise. The SDBS-data again identified the main component as inosine. A two-dimensional COSY spectrum (recorded within 14 h) exhibited the coupling pathway of the protons at the ribose unit of inosine. Again, these data are in complete agreement with the proposed structure of the main component. Further evidence for inosine as the prevailing compound was provided by a two- dimensional 1H,13C-correlation spectrum, recorded by the established HMQC sequence (hetero multiple quantum coherence) during a detection time of 24 h. Therein, cross peaks via 1J-couplings identify connected protons and carbons. Again, all data obtained from the HMQC experiment were completely consistent with the assignment of inosine as the prevailing structure.

- 95 - -IV. Discussion-

IV. Discussion

1. Myxozoan lifecycles

Species concepts, although they have been successfully employed in self-fertilizing taxa, might be limited in myxozoans, given the probable absence of intraspecific sexual reproduction. Mating was found to be reduced to postmeiotic autogamy of single clonal lineages in M. cerebralis (El-Matbouli et al.1995, El-Matbouli & Hoffmann 1998). There is also the possibility of the formation of cryptic or sibling species in Myxozoa. As part of the speciation process, genetically similar organisms form distinct local communities (‘subspecies’) as a result of environmental differences. These incipient ‘new’ species could equally concur or differ in morphology, host range and other infection characteristics. The phylogenetic species recognition concept (Taylor et al. 2000; Agapow et al. 2004), if applied to Myxozoa, would solve many problems of morphological approaches by clarifying genetic distance. To avoid subjectivity (e.g. species limitations based on one gene genealogy or a randomly chosen percentage of base pair alterations), multiple target gene sequences should underlie congruence analysis in the future. Despite the improvement of methods to elucidate myxozoan life cycles due to the increase in molecular data available (Kent et al. 2001), attention must be drawn on the reproducibility of transmission experiments. Mixtures of parasites from infected oligochaetes often corrupt the results of such approaches. Depletion and preliminary exclusion of contaminations by a selective use of well discerned transmission stages are imperative for a comprehensive life cycle description. To exclude misinterpretations caused by (molecular) linking of myxosporean stages to possible actinosporean counterparts, the complete development should be followed more than once including the consideration of simple DNA contamination (e.g by persisting myxospores).

- 96 - -IV. Discussion-

Henneguya nuesslini

The genus Henneguya Thélohan, 1892 comprises about 150 species and represents the second largest group in the Myxozoa. Information on the genus has recently been thoroughly reviewed by Eiras (2002). Although new Henneguya spp. are frequently described (e.g. Azevedo & Matos 2003; Casal et al. 2003; Reed et al. 2003; Vital et al. 2003), the life cycles of only two species, Henneguya exilis Kudo, 1929 (Lin et al. 1999) and Henneguya ictaluri (Pote et al. 2000) have been fully elucidated. Both include aurantiactinomyxon actinospores and develop in catfish, Ictalurus punctatus, as the teleost host. The results of the molecular and morphological analysis and the transmission experiments allow the conclusion that the species described here is H. nuesslini. This less known species was originally described from fixed material from just two alcohol-fixed pseudoplasmodia, but the outstanding morphological description and illustration of the original authors (Schuberg & Schröder 1905) are not compromised and largely match the characteristics observed here. The only discrepancies are that the original authors found the polar filaments to bear six to seven coils and the mature pseudoplasmodia (cysts) to be up to several millimetres in size. The mature myxospores of H. nuesslini are similar to those of Henneguya zschokkei (Gurley, 1894) Doflein 1901 and H. salminicola in general spore morphology. However, the measurements of the myxosporean differ from those of either of these species. The polar capsules were about 1.0 µm larger than those of H. salminicola and 1.8 µm smaller than those of Henneguya salvelini Zandt, 1923. In H. zschokkei the tails are much longer. In H. nuesslini, although varying considerably (12 – 23 µm), the tails can be regarded as the shortest in a salmonid infecting sarcohistozoic Henneguya sp. In H. nuesslini, the tails do not diverge as widely and sutural furcation does not emanate from the tail basis, as Zandt (1923) described for H. salvelini. Furthermore, the polar capsules of H. nuesslini are not in contact with each other as they are for H. zschokkei. In sutural view the spore body is almost twice as thick as for H. salminicola. The myxospore morphology was similar to that of H. cutanea, but this species was obtained from common bream and its 18S rDNA sequence (GenBank Accession Number AY676460) differs considerably from that of H. nuesslini (75.0% similarity over a 1348 bp long fragment after aligning with BioEdit

- 97 - -IV. Discussion-

Sequence Identity Matrix program) (Eszterbauer, unpublished data). Indeed, the measurements most closely complying with the observed morphotype were those of the original description of H. nuesslini provided by Schuberg & Schröder (1905). Although it parasitises a different fish host, the recognition of H. salvelini as a valid species (Zandt 1923) is questionable, as there are no significant differences from the species described here, except the tail length and polar capsule size. In contrast to the morphological findings, the 18S sequence data initially suggested H. zschokkei as the species involved in this study, but its original description differs considerably from this species. Moreover, several authors (e.g. Lom & Dyková 1992; Eiras 2002) consider H. salminicola, H. tegidiensis Nicholas & Jones, 1959 and H. kolesnikovi Gurley, 1886 as further synonyms of H. zschokkei. The continuing doubt whether H. zschokkei and H. salminicola are synonymous (Kent et al. 2001) indicates that the 18S rDNA sequence is of limited use in clarifying the taxonomic status quo of closely related myxozoans. Even though GenBank provides sequence data for both species, classification on the basis of sequence homology remains difficult (M. Kent, personal communication) and there is still no evidence for both being valid species. Eventually additional molecular markers, e.g. ITS sequences, will help to clarify relationships between these parasites. One important reason not to rely completely on molecular results was that H. zschokkei and H. salminicola both form large intermuscular pseudoplasmodia, not tiny aggregates in connective tissues. The site of infection assists in the identification of myxozoan parasites, as tissue tropism and the location of sporogonic stages in some cases are more important than morphology, geographical distribution or host specificity (Eszterbauer 2004). For example, members of the genus Kudoa Meglitsch, 1947 cluster more by host specificity and local distribution than by myxospore morphology (Hervio et al. 1997). Although not yet backed by transmission experiments, molecular evidence (Eszterbauer et al. 2002; Hallett et al. 2002, 2004) suggests that morphology per se is not a decisive character for establishing a myxozoan species and therefore should not be used alone. Spore measurements may alter depending on pseudoplasmodia location, infection site, fish age or host species, host size and trophic state, and thus fail to provide comparable information between holospecies. Wagner (2002) noted that the tail length varies within one myxospore and, even more, between individual spores of H. salminicola. In the experiments of this study, the fish were still small at the time they were killed, which might have resulted in formation of shorter ‘tails’ than usual. Nevertheless, the combination of

- 98 - -IV. Discussion- tissue tropism, uniform appearance of short ‘tails’, morphometrics, fish host species and 18S rDNA sequence data (considering the possible limitations of this marker) provide sufficient reason not to designate the isolated species as H. zschokkei. The possibility of cross-contaminations between H. zschokkei and H. nuesslini must be considered in future molecular diagnosis. Several aurantiactinomyxon and echinactinomyxon infective stages bearing 16 secondary sporoplasm cells have been described (Janiszewska 1955, Marques 1984, Xiao & Desser 1998a, Özer, Wootten & Shinn 2000). Although TAMs with 16 secondary sporoplasm cells were also recorded, those of H. nuesslini clearly differ from Triactinomyxon magnum Granata, 1922 and Triactinomyxon sp. 1 sensu Oumouna et al. (2003) in respect of its relative proportions and spore body features. TAM type “F” of Xiao & Desser (1998b) had 16 sporoplasm cells, but was shed from Limnodrilus hoffmeisteri Claparéde, 1862, and had a remarkably short style. Therefore this actinosporean stage is regarded as previously undescribed. The present study shows that Henneguya spp. do not always produce aurantiactinomyxon-type actinospores like H. ictaluri and H. exilis. The latter species show a different (more elongate, spindle shaped) myxospore morphology compared with, e.g. H. zschokkei, characterised by proportions very similar to Myxobolus spp. that are mainly altered by the valve-derived ‘tails’. This might indicate a convergent origin of tail-like projections in myxosporean spores, or represent a link between Myxobolus spp. and phylogenetically ‘younger’ Henneguya spp., like H. exilis. However, it surely is an argument for the polyphyletic status of the genus Henneguya. It might be speculated that the genus within the Myxobolidae might have emerged from a Myxobolus-like ancestor. In this case, aurantiactinomyxon-type actinospores would have developed from triactinomyxon predecessors, perhaps through style reduction and caudal process growth. It is surprising in this context, that aurantiactinomyxon-type spores are also involved in life cycles of genera with rather different myxospore morphology, such as Myxidium or Hoferellus. This supports molecular studies that place the H. zschokkei/H. salminicola branch closer to a neighbouring Myxobolus sister group within the Myxobolidae than to other Henneguya spp. (Kent et al. 2001).

- 99 - -IV. Discussion-

Although the optimal conditions for maintenance of its life cycle have to be defined, H. nuesslini is potentially a suitable model for laboratory culture. It is easily identifiable and its hosts are readily available. Its value as an interesting comparison with M. cerebralis lies in the possibility of maintaining it in the same hosts. Its fish host range provides an excellent basis for research on host specificity and fish susceptibility, whilst its tissue tropism allows investigation of the infection pathway of cyst-forming myxozoans and the cellular interactions within the host.

Myxobolus parviformis sp. n.

The life cycle of this species represents an example of a myxozoan, whose actinospore was already described (Hallett et al. 2005), without further knowledge about the myxospore counterpart. Even the 18S rDNA sequence was already available through GenBank. Hallett et al. (2005) identified T. tubifex as the oligochaete host, which supports the observation that the parasite is not restricted to L. hoffmeisteri. The species differs from hitherto known myxosporeans found in bream gills in its myxospore morphology. It most closely resembles the myxospores of Myxobolus exiguus, Myxobolus encephalicus Landsberg & Lom, 1991 and Myxobolus muelleri. However, the 18S rDNA sequences of all these other species differ considerably from the one of M. parviformis. Furthermore, M. exiguus and M. muelleri also infect other tissues and hosts. The actinospores of M. parviformis are very similar to those of Myxobolus macrocapsularis (Székely et al. 2002), which also develop in bream gills but form large clubbed plasmodia at the filament tips. Finally, the myxospores of M. macrocapsularis differ in size, shape and number of sutural markings. Piscine gills are the most affected target tissue for infections by myxozoans. Many members of the Myxozoa specifically develop in common bream gill tissue (Shulman 1966), mostly belonging to the family Myxobolidae. Unfortunately, the development in local proximity can lead to confusion through postinfectional mergence of sporogonic stages or mature plasmodia. This can be misinterpreted as intraspecific morphological variation and may weaken epidemiological data and hamper taxonomic research. It was advantageous to collect non-fused, single pseudoplasmodia (accompanied by morphological analysis) to isolate a genotypically homogenous lineage throughout the subsequent transmission procedures. By separating the desired parasite via its

- 100 - -IV. Discussion- morphotypus, other myxozoan species contained in the initial infection material could be successfully suppressed by the selective use of phenotypically identical ‘clones’ for the next transmission step, that additionally were proven to be identical by PCR-RFLP. A species that would produce similar actinospores is M. macrocapsularis. In addition to the fact that a contamination was detected neither by PCR-RFLP nor by DNA sequencing, the myxospore mixture used for the first trial and all examined plasmodia from the transmission experiments contained no M. macrocapsularis myxospores, so a contamination with this species can be excluded. As both developmental stages were found in two consecutive transmission trials, which was supported by molecular identification, the developmental cycle of M. parviformis appears to be valid. The RFLP patterns of M. parviformis differed from those of other myxosporean species studied previously, including some common parasites of bream, such as Myxobolus hungaricus, M. bramae, M. macrocapsularis and M. impressus (Eszterbauer et al. 2002). The actinospore description presented largely refers to the guidelines given by Lom & Arthur (1989), but the suggestions of Paperna (1973) and Molnár (2002) were also considered to provide more exact data on the location of the developing plasmodia within the gill-arches. Nevertheless, it is unknown whether M. parviformis is polyxenous and restricted to the location within gill tissue where plasmodia were found in the present study. It is possible that Leuciscinae other than A. brama also may serve as suitable hosts, and even tissues other than gills could be target tissues for this parasite. These questions can only be answered by further transmission experiments. Facing the number of Myxobolus spp. already described (Eiras et al. 2005), one cannot exclude the existence of a senior synonym of the species. No taxonomist can ignore that these parasites vary in shape and size (the main species defining features in Myxozoa besides the host species) according to their location in the host, maybe also depending on host condition or parasite strain. Morphological variation is commonly observed in other myxozoan species (e.g. M. pseudodispar) and even in their actinosporean stages (Hallett et al. 2002, 2004). In most cases, transmission experiments to ascertain the full host range, which would give a valuable basis for a new species description, have not been performed. Janiszewska (1955) noted an ‘inner envelope’ as a principle structural component of the “Actinomyxidia”, and she designated this as the ‘endospore’ enclosed in a ‘sheath’ in contrast to the ‘epispore’ derived by the valve cells. According to her observations, the envelope is formed by one or two cells, but she did not consider this

- 101 - -IV. Discussion- sheath as a functional entity in transmission. The term was later found obsolete by Lom & Arthur (1989). The actinosporean sheath unit, once expelled, reveals a striking phenotype resembling an actinosporean genus referred to as ‘Neoactinomyxum Granata, 1922’ by some authors (Jirovec 1940, Oumouna et al. 2003), although in the original genus description (Granata 1922), Neoactinomyxum globosum possessed the unique flower-like shape that was mentioned by other authors later (Marquès 1984, Xiao & Desser 1998a, El-Mansy et al. 1998). Specimens without these appendices were either described from histological material, or specimens were gathered from field material filtrates, not from isolated hosts. The taxonomic validity of similarly shaped neoactinomyxum-type actinospores, e.g. N. globosum Granata, 1925 as described in Özer et al. (2002), cannot be questioned. The present study points out the possibility that some uncommonly found actinospores might be in fact sheath units without their original valve structure. One reason why this sheath structure has rarely been mentioned by previous authors might be that their studies focused on morphological descriptions rather than on the behaviour and the physiological functions involved in the transmission of actinospore stages. The function of the actinosporean sheath in transmission is unclear. That the polar capsules are widely counter-sunk and covered by the valve cells, suggests a different invasion strategy than the one used by actinospores with a spore architecture with acute protruding polar capsule tips. For instance, polar capsules of some aurantiactinomyxon-type spores are almost fully covered by the valve cells. According to the author’s observations, the polar capsules of M. cerebralis and H. nuesslini TAMs are much more protruding and a similarly prominent sheath could not yet be observed. In these species, the (by fish mucus) activated sporoplasm’s ability to migrate down the style to the junction of the caudal processes while leaving the polar capsules in an apical position can frequently be observed. On the other hand, in electron microscopical studies, El-Matbouli et al. (1999) described a filamentous structure left on the host’s epidermis after penetration of M. cerebralis sporoplasms into rainbow trout. This observation appears to be the pendant to the sheath of M. parviformis. The sheath unit could be the actual infective substage of the actinosporean form of M. parviformis. If in this species attachment does not occur by contact with the fish surface, the valve shells could open passively (perhaps upon stimulation by host cues) and release the sheath unit. This appanage could be supposed to sink to the bottom where it can readily be taken up by bream, which feed

- 102 - -IV. Discussion- by sediment ransacking through the gill openings. Alternatively, the sporoplasm is simply protected from precarious outer conditions (e.g. osmolarity change, disruption). The inner sheath could thus represent an adaptation to habitats of standing water, another mode of infection (e.g. oral) or it could be just an additional device of protection in freshwater during the moments after attachment to the host surface. Detailed ultrastructural data are needed to better understand the function of the inner sheath. Indeed, we do not know much about the invasion strategy and developmental pathways of any gill infecting myxozoan species. The infective cells may reach the gill lamellae via blood stages, when invasion is possible over the whole fish surface, or the actinospores may attach directly to the gills from the surrounding water and the sporoplasm cells directly reach the site of development.

2. Host invasion by actinospores

In parasite invasion, a sequence of responses to certain stimuli as seen in trematode cercariae can be differentiated (reviewed by Haas and Haberl 1997, Haas 2003). Host attachment and penetration by actinospores and migration of the amoeboid germ should be considered as separate steps, possibly requiring different cues. The results from this work indicate that M. cerebralis actinospores require a chemical, mucus-derived signal before they become mechanically sensitive for polar filament discharge. The combination of both triggers, formerly suggested by El- Matbouli et al. (1999), makes sense, as the length of the polar filament (about 30-40 µm) can mediate host surface anchorage only upon close contact. A response caused by chemical or mechanical signals individually would most certainly lead to inefficient discharge reactions. Examples for a combination of chemical and mechanical stimuli are numerous in coelenterates, but the chemoreceptors and their location have rarely been described (Kass-Simon & Hufnagel, 1992). No mechanoperceptive structure was described from myozoans yet. The maximum percentages of responding actinospores varied from about 40 to 80% in the bulk experiments. This variability could be caused by differences in spore age as shedding of the actinospore packets by oligochaetes cannot be controlled and therefore all experiments include spores of different age. For example, chemoreceptor function could be affected by spore age. It is not known whether non-viable spores may contribute to the discharge rates measured in the experiments. Actinospores with

- 103 - -IV. Discussion- non-viable polar capsule cells could lower the discharge rate. Markiw (1992) observed dead sporoplasm cells and intact polar capsules in the same spores when using the FDA/PI staining method. Thus, discharging spores with dead sporoplasms can give a wrong impression, if the discharge rate is put on a level with infectivity. The fact that the actinospores in this study showed intact amoeboid germs and very little nuisance on a cellular level (except for single discharged polar capsule cells) renders this factor negligible. Also, it became clear that the staining procedure causes discharge in several spores by itself, when the discharge rates caused by control substrates are compared with the observed cellular viability rate. Although the lack of host specificity of M. parviformis actinospores could be demonstrated, they showed very low reaction rates to any fish mucus isolates throughout several trials. In an attempt to apply chemical (bream mucus) and mechanical stimulation to single isolated spores directly by Ponader (2005), as conducted with M. cerebralis in this study, the actinospores did not react at all. This emphasises the theoretical considerations on a divergent host invasion mechanism, different stimuli or another infection pathway by this species. Another explanation may be that the concentration or availability of discharge triggers is different in bream mucus, which shows a much higher viscosity than trout mucus. The notably concealed polar capsule apertures suggest that M. parviformis actinospores require more protection than those of e.g. M. pseudodispar; whose polar capsules are very loosely attached (own observations). The question why several myxozoans show a narrow invertebrate host range but successfully develop in a rather wide choice of teleost host species (e.g. salmonids for M. cerebralis) is not yet understood. It was supposed, that actinospores might specifically recognise their appropriate hosts. Xiao & Desser (2000) observed differences in sporoplasm release ratio by actinospores to mucus of various fish. Also, T. hovorkai was reported to distinguish between cyprinid genera (Yokoyama 1997). These authors speculated that the spores did not penetrate non-susceptible fish, as no parasite developmental stages were found in tissues and myxospore production was not observed. Other observations by the latter authors revealed penetrated cells, but no reproduction of M. arcticus in non-susceptible salmonids. El-Matbouli et al. (1999) concluded that M. cerebralis actinospores were able to distinguish carp from rainbow trout from the amounts of attached actinospores in electron-microscopic images. The authors suggested that the spores did not penetrate non-susceptible fish from the fact

- 104 - -IV. Discussion- that no parasite stages were found in tissues and no myxospore production was observed. Differences in threshold concentrations triggering the actinospore’s receptors, varying proportions of cues present in mucus of different fish species and different mechanical sensitivity of the actinospores could be reasons for such varying discharge rates. On the other hand, Yokoyama et al. (1995a) found raabeia-type actinospores of M. cultus frequently reacting to various fish mucus isolates as well as to bovine submaxillary mucin. McGeorge et al. (1997) and Ozer & Wootten (2002) exposed several actinosporeans to mucus of salmon, trout, stickleback and bream and found all of them reacting to all mucus isolates by polar filament discharge and sporoplasm emission. Although these authors did not apply an additional mechanostimulus (except the use of pipettes and cover slips), the findings support the results of this study on the discharge triggering activity of mucus isolates from host and non-susceptible fish species. The attachment stimulating activity of carp and bream mucus at similar concentrations demonstrated, on the level of attachment, the failure of M. cerebralis actinospores to discern host fish species from species in which the parasite is unable to develop. This is underlined by the fact that the sporoplasms of this species actively left the spore in non-host mucus as numerous as in rainbow trout mucus (Borrelli, unpublished data). The two additionally included species H. nuesslini and M. parviformis did also react to mucus of non-susceptible fish. Ponader (2005) also confirmed a lack of host specificity during polar filament discharge of Myxobolus pseudodispar TAMs, a muscle-dwelling myxozoan of roach (Rutilus rutilus). This is in contrast to the results obtained by El-Matbouli et al. (1999) who did not observe a prospected amount of attached M. cerebralis actinospores on epithelial surfaces of exposed carp. However, data must be interpreted carefully, since an artificial stimulation of actinospores does not represent the true biophysical situation at the surface of a living fish. Furthermore, it is uncertain, if the same compounds in carp and trout mucus caused discharge. Muci of other vertebrates such as Rana epidermal homogenate triggered significant discharge rates. Whole frog stratum corneum patches contain high amounts of granular goblet- and Leydig-cell products, and therefore can serve as a comparable mucous substrate. As some myxozoan species parasitise amphibian hosts (reviewed in Browne et al. 2002), this result was not surprising. Together with the reactions to bovine submaxillary mucin, this alludes to a signal that is common in vertebrate mucus

- 105 - -IV. Discussion- or epithelia. Bovine mucin was a potent trigger for cnidae discharge in anemones (Watson & Hessinger 1987). Nevertheless, the results with bovine mucin remain ambiguous, as it did not show any triggering effect at 1 mg/ml (data not shown) and the discharge rates varied greatly when the concentration was increased. The observed speed of discharge and the length of the polar filament indicate that actinospores need a close contact of their apical region with host surface for successful attachment, which in turn may be very brief. As there was no optimum frequency for the discharge reaction, a thigmotropic mechanism can be assumed. The intrinsic retraction mechanism of the polar filament was always observed during discharge. Its purpose is to mediate ultimately close contact with the viscous fish surface. Uspenskaya (1982) assumed Ca2+ ions to be involved in the polar filament discharge mechanism and suggested that it is an active process due to contractile proteins. Exterior Ca2+ was not a major factor for the discharge mechanism in experiments with H. nuesslini actinospores, but osmolality of the surrounding medium obviously has to exceed a certain value for correct function as shown by the Ca2+- replacement experiments. When the effect of osmolality differences was neutralised and the response in control media was subtracted from discharge rates with stimulating mucus, Ca2+-ions indeed seemed to play a role. Discharge rates in Ca2+ containing water were doubled compared to those in Ca2+–deficient water of the same osmolality. However, in substrate without Ca2+ but with NaCl to restore the original osmolality, discharge rates were higher than substrates with Ca2+, indicating that eventually only an artificial effect of NaCl was observed. In sporoplasm emission experiments, most H. nuesslini sporoplasms left their valve shell after 5 to 6 min. This time course applies to mechanically absolutely undisturbed actinospores, whether discharged or not. It was observed that the apical opening does not occur passively by an intrinsic mechanism, as many discharged specimens did not open and sporoplasm release was impossible. Thereby, enhanced movement of the sporoplasms and even a rearrangement of the polar capsules was frequently observed during incubation of actinospores in fish mucus. Thus, although polar filament discharge surely facilitates sporoplasm emission, it is not an imperative prerequisite. Furthermore, it could be demonstrated that the sporoplasm primary cell is able to recognise mucus components even inside its undischarged actinospore. This opens a wide field for discussion whether these cells may be capable to actively manipulate or trigger polar filament discharge. M. cerebralis and H. nuesslini

- 106 - -IV. Discussion- actinospores also emitted their sporoplasms in mucus of non-susceptible fish species. Even though the bioactive fraction (F II from the LichroPrep RP 18 SPE) did also (weakly) stimulate sporoplasm emission, it cannot be stated at this point whether the chemical stimulus of sporoplasm emission and polar filament discharge is the same. It must be considered that increased emission could be due to mechanical facilitation after polar filament discharge. More experiments, including concentration series of muci of different fish, are needed to gain a deeper insight into this topic. However, a modified protocol of the bulk experiments must be used, i.e. a prolonged incubation period on the slides. The findings for the species investigated in this study allow the drawing of an updated picture of host invasion by actinospores: Waterborne actinospores receive a chemical cue from fish mucus and become mechanically excitable. Upon apical contact, a not yet known mechanoreceptor induces discharge of the polar filament. The retraction of the filament pulls the apical region of the actinospore tightly into the mucous surface. Forces generated by a parachute-like mechanism caused by water current on the floating appendices open the apical valve shell foldings and the actinosporean shell is eventually discarded. The sporoplasm (or sheath unit) is released and moves into deeper regions to penetrate the epidermal tissue.

3. Host signals for polar filament discharge

Considering transmission stages of non myxozoan parasites that invade freshwater teleost hosts, our knowledge on host cues and the specificity of their reactions and behavioural patterns is limited. The salmon sea louse Lepeophteirus salmonis performs host-specific rheotaxis towards its host, obviously by using toxic isophorone as a stimulant (Ingvarsdottir et al. 2002). One example for highly species-specific reactions in fish parasites is the oncomiracidium of Entobdella soleae (Kearn 1967). Diplostomum spathaceum cercariae, actively invading a wide range of host fish, also rely on host contact by chance. This species uses the increased CO2 pressure for their attachment response (Haas 1975), a very unspecific signal. The cercariae readily attach to various aquatic animals, but penetrate only, when specific macromolecular substances (sialic acid containing glycoproteins) and lipids are present (Haas et al. 2002). However, actinospores cannot leave the host after attachment and unspecific response to integuments of random aquatic organisms seems to be a futile strategy

- 107 - -IV. Discussion- for them. Responses to ubiquitous amino acids like amino acids, as shown for Ichtyophthirius multifiliis (Haas et al. 1999), would lead to arbitrary attachment responses to any organic matter in aquatic environments. The low discharge responses to trout muscle homogenate indicate that a mucus- specific compound is recognised by M. cerebralis actinospores. An attachment preference towards mucous cell openings as sites of entry gives an additional hint to link the discharge signal to the products of granular glands (El-Matbouli et al. 1999). Experiments with higher mucus concentrations to increase discharge rates by M. cerebralis actinospores suggest a threshold concentration between 0.1 and 1 mg/ml. Of course, when encountering the fish host, actinospores encounter a much denser (highly concentrated) mucus layer. The bulk experiments also show that the discharge reaction is independent from physical properties and the pH of the living fish surface The effect of pH on polar filament discharge was described by Smith (2001), indicating that extreme pH values favour spontaneous discharge. In the bulk experiments, the pH of media and test substrates did not differ, the detection of mucus via its pH is therefore unlikely. Osmolality is also a factor that influences polar filament discharge of M. cerebralis actinospores (Smith 2001). Lowered osmolarity was not responsible for the low discharge rates with deionised water, as buffered spore medium water was similarly ineffective. However, osmolarity gradients might be involved in the discharge reaction (see also Yokoyama et al. 1995a). Roberts & Powell (2004) determined the osmolality of “concentrated” trout mucus to be only 70 mosm/kg. Ponader (2005) found, that a D- (+)-mannite solution of 100 mosm/kg stimulated discharge similarly as untreated trout mucus homogenate (1 mg/ml). High concentrations of NaCl and glucose increased discharge rates, but also produced visible changes in sporoplasm cell integrity (collapse of internal pressure and sporoplasm shrinking). On the other hand, hyaluronic acid and DMSO did not cause discharge in M. cerebralis actinospores at osmolalities exceeding 100 mosm/kg (Ponader 2005). DMSO is known to pass biomembranes and hyaluronic acid is a highly negatively charged GAG, rendering ionic properties as less important. This suggests that only certain molecule species, i.e. those that may pass the parasite’s membranes or interfere with the homoeostasis of osmotic balance, artificially cause discharge. Another indication that osmolality is not the natural discharge trigger became apparent when a very low response to a myo-inositol solution with an osmolality similar to mucus substrates could be shown.

- 108 - -IV. Discussion-

Therefore, osmotically enforced (or blocked) discharge reactions always have to be considered during tests of substrates and have to be strictly separated from reactions caused by chemo-perception. Fish mucus contains up to 95% water and is comprised of highly glycosylated glycoproteins as main constituents (Reid & Clamp 1978). Carbohydrates are typical cell signalling structures and are manifest candidates for fish mucus recognition by parasites. The monosaccharide spectra of fish muci have repeatedly been investigated (Carlstedt 1985, Nakagawa et al. 1988, Stiegeler 1997). But carbohydrates are also abundant components of other aquatic organisms and their secretions. Tubificid glycoproteins contain several non N-acetylated hexoses that were also found in fish mucins (Rahemtulla & Løvtrup, 1974). Kalbe (1998) found galactose as the predominant hexose in L. stagnalis glycoproteins. In this study, the free hexoses glucose, mannose and N-acetyl-galactosamine were detected both in fish as well as in L. stagnalis mucus homogenate. The glucose content in trout mucus was discussed by Roberts and Powell (2005). However, this hexose did not stimulate discharge in different concentrations and TLC detected no hexose components in the bioactive fraction F II from LichroPrep RP 18 SPE (a concurrent GC/MS analysis is currently conducted). Sialic acids are vertebrate-characteristc sugars found in numerous conformations in teleost glycoproteins (Asakawa 1974) and are thus suitable host cues for recognition of fish in aquatic environments. Sialic acid containing glycoproteins are used by the fish-invading trematode cercariae of Acanthostomum brauni as stimulus for attachment to their host (Haas & de Núñez 1988) and stimulate penetration behaviour of D. spathaceum cercariae (Haas et al. 2002). Extended effort was undertaken to test sialic acids as host cues for actinospores. Neither neuraminidase digestion nor preparative sialic acid extraction gave an indication of an involvement of these sugars in polar filament discharge. Other mucus constituents that could be excluded to trigger polar filament discharge by M. cerebralis actinospores in this study were non-volatile inorganic electrolytes, all volatiles, free L-amino acids, glycosaminoglycans, bound and free hexoses and sialic acids, proteins, urea, inositols and mycosporine-like amino acids. Mycosporine-like amino acids show absorbance maxima at values > 310 nm, which was not the case within the bioactive fractions. Blocking of amino-groups did not affect the discharge triggering activity of mucus, which allows the exclusion of a wide variety of

- 109 - -IV. Discussion- substances. In experimental trials using the method described in this study, Ponader (2005) showed that substances that trigger nematocyst discharge, including L-proline, fatty acids and glutathione, did not stimulate discharge by M. cerebralis actinospores and that CO2 was not a discharge-triggering signal for this species (by resolubilisation of lyophilised trout mucus in CO2-free water). An UV-absorbance of the bioactive fractions (gradient chromatography with LichroPrep RP 18) was observed at 250 nm and at 295 nm in a second peak when the single main peak was split up during separation. Strong luminance was detected upon excitation at 312 nm, especially in the first signal. The signal could not be separated by lipophilic TLC, but hydrophilic TLC-separation demonstrated the presence of at least 5 luminescent compounds in the bioactive SPE-fraction. This indicates either a composition of different (related) substances, or represents products derived from degradation of one or two compounds. As early as 1959, Wright et al. found fluorescent substances in skin mucus of Lymnaea spp. after chromatographic separation and used them for taxonomic studies. This method was adapted by Barry & O’Rourke (1959) to distinguish marine fish species and strains by UV-detection of chromatograms obtained from skin mucus. Unfortunately, these authors could not identify the molecules responsible for this phenomenon. It is almost certain that the UV-active compounds detected in the bioactive fraction during this study belong to the same molecular class as the substances found by Barry & O’Rourke (1959). They seem to occur in a species-specific pattern, a factor that could be exploited by parasites. This suggests that the substances represent semiochemicals with kairomone activity and may play a role in intraspecific communication. Peptides could potentially be present in the bioactive fraction, but do not usually show similar luminance. Peptides still must be considered to co-elute at this volume, although this is not supported by its low content of amino groups. The carbohydrate signal is surely caused by nucleoside ribose. All nucleotides have absorbance maxima at 250-260 nm (depending on pH), but an increased UV absorbance at these wavelengths was not seen in full spectra of trout mucus homogenate < 3 kDa. The high absorbance values around 290 nm in mucus homogenate < 3 kDa may be due to luminance caused by excitation at these wavelengths, or could be caused by aromatic amino acids or peptides. Fluorescence/luminescence in the bioactive fraction is not likely a result of impurities from vial and tubing material as often observed in biochemical trace analysis. The luminescence signals co-elute directly with the peaks

- 110 - -IV. Discussion- that trigger polar filament discharge in gradient chromatography and could be either derived from tryptophan residues (λmax = 280 nm) or (cyclic) nucleotide derivatives (i.e. mono- and diphosphate nucleosides). NMR studies detected inosine and another nucleoside in the bioactive fractions from trout mucus. The reason for their abundance in mucus remains puzzling. Nucleotide derivatives inosine and its base, hypoxanthine, are well known components of fish skin (Hayashi & Saito 1968), that contains as major components. Inosine- monophosphate (IMP) and adenosine- monophosphate (AMP) were detected in salmon dermis and only IMP occurred both in rainbow trout and carp skin (Hayashi & Saito, 1968). IMP is the predominant nucleotide derivative in muscle tissue of different fish species, whereas AMP was most abundant among crustaceans (Mendes et al. 2001). In rainbow trout gill tissues, AMP was dephosphorylated and deaminated to inosine (Leray et al. 1979). The spectrum of ATP-catabolites in chill-stored rainbow trout was repeatedly analysed (Saito et al. 1959, Murray et al. 1985, Azan et al. 1989, 1990). These substances are thought to be metabolic intermediates resulting from cellular activity and have been extensively employed to assess fish freshness (Burt 1976, Ehira & Uchiyama 1987, Carsol & Mascini 1998, Mendes et al. 2001). The nucleotide bases hypoxanthine and guanine were found in trout skin (Hayashi 1972), but not yet in mucus. The above studies did not mention cyclic molecule species. Besides, their role in osmoregulation is widely accepted (Buck 2004, Hurley 2003) and they might play an important role in fish skin and mucus. Ferguson & Price (1973) have shown that cIMP is a deamination product of cAMP in toad bladder and affects the transport of sodium and water across tissues. It was only slightly less effective in stimulating cAMP-sensitive protein kinases than cAMP (Corbin & Krebs 1969). The designation of nucleosides or their derivatives as potential polar filament discharge triggers would explain many characteristics specified by mucus treatment and fractionation throughout this study. Nucleoside derivatives are heat stable and resistant to proteolytic activity. Activated charcoal treatment, routinely used for nucleotide removal, deleted all discharge triggering effect in mucus homogenate < 3 kDa. Purine-ribose derivatives react with resorcinol after acid hydrolysis to furfurals and thus presumably contributed to the values of neutral carbohydrate estimation. Only chromatography with LichroPrep RP 18 achieved separation, which shows the amphiphilic to hydrophobic character of the discharge triggers. This matches with the hydrophobic nature of the purine ring. It further explains their abundance in high

- 111 - -IV. Discussion- molecular and lipophilic fractions of the mucus homogenates. Otherwise, the effect of lipids described in this study cannot be explained. Attachment of the triggering substances to membrane phospholipids or proteins might be responsible for this phenomenon. Nucleosides indeed bind to transmembrane protein motifs, as they are involved in ion-channel activation (Hurley 2003). The fractionation into lipid classes could not isolate a stimulating lipophilic compound (Ponader 2005), which might be caused by an inability to release nucleosides from the aminopropyl-sidechains or reactions provoked by some of the various organic solvents during SPE. Unfortunately, nucleotide derivatives from fish also bind to Dowex 1 x 8 resins in HCOOH-form (Hayashi & Saito 1968) like sialic acids. Elution is achieved with acidic solvents such as 0.1 M formic acid + 0.1 M sodium formiate. In the Dowex 1 x 8 separation included in this work, bioactive substances were only located in the washing fraction of the three-step separation of trout mucus homogenate < 3 kDa. Considering all other ionic constituents (esp. electrolytes), it is suggested that especially the weakly binding nucleosides were immediately displaced and thus were mainly present in the first fraction. In contrast to pyrimidines, the glycosidic bond of purines is hydrolysed in strong acids and this may also account for the low discharge rates in the fractions F II and F III. A preliminary extraction procedure by activated charcoal (elution with ammoniacal ethanol) and the use of a less acidic solvent system would greatly improve the method. Separation by ionic interaction is problematic considering the stereochemical nature of some nucleoside derivatives. Substances like amino acids or cyclic inosine monophosphate (cIMP) are zwitterions (Sundaralingam et al. 1982) and thus are difficult to extract by ion exchange methods. Throughout the study, all ion exchange approaches were ineffective in binding and separation of the discharge triggering agents. Unfortunately, no final proof could be presented during this work that clearly identifies the molecular affiliation of a polar filament discharge trigger for actinospores. A discharge triggering effect of the nucleotide phosphate ester group is favoured by the observation of Ponader (2005) that M. pseudodispar responds very sensitively to 5 mM PBS with polar filament discharge. In contrast, Yokoyama (1995a) reported that sporoplasm emission of raabeia actinospores of M. cultus was inhibited in PBS solutions > 5 mM. The presence of phosphate in mucus homogenate < 3 kDa was suggested by HPIAC-MDD analysis but an identification of phosphate-bearing species among the compounds of the bioactive fractions was not yet confirmed by 31P NMR

- 112 - -IV. Discussion- studies. That inosine and its mono-phosphate did not elicit discharge suggests that possible nucleoside discharge triggers are present in an alternative conformation, e.g. as cyclic isoforms or in an N-oxide form. The first suggestion is strongly supported, as cyclic phosphate esters of nucleosides like inosine exhibit a strong fluorescence that might explain the luminescence in the bioactive fractions. Nevertheless, muscle homogenate was not an effective discharge triggering substrate, although it contains inosine and IMP suggesting that these compounds are not the sought discharge triggers, at least in their un-modified state. The discharge triggering effect of numerous nucleoside-related substances, including D-ribose, has to be tested yet. Although a volatile substance, the possible presence of the sea louse host cue isophorone in the bioactive fraction must also be examined. Extensive data are required to evaluate the discharge triggering effect of several inosine and adenosine derivatives along with consecutive detection (e.g. by TLC, NMR) in mucus of various fish species. Without question, it is not sure that the UV-active substances in the bioactive fraction are nucleosides and that these really are the polar filament discharge triggers for myxozoan actinospores. This question is now open to be solved using the experimental and biochemical methods presented in this study. The surface of myxozoan host cues has meanwhile barely been scratched.

4. Impacts on myxozoan transmission

Actinospores from all three myxozoan species that were examined in this study discharged their polar filaments when mucus of non-susceptible fish species was offered. It is therefore suggested, that unsusceptible fish species might serve as potent decoy organisms in waters enzootic for e.g. M. cerebralis due to developmental failure after penetration. The strategy of improving transmission by enormous numbers of transmission stages is not only found in myxozoan parasites and might be the consequence of a lack of species-specific invasion. One reason not to use a host- specific discharge mechanism could also be the speed at which the recognition process must take place between these parasites and their teleost hosts. Non-specific host recognition also creates an open backdoor for the acquisition of new host species and the development of new parasite species after transfer to a new environment. This approach certainly raises the question of how the pronounced host specificity of M. cerebralis is accomplished. To understand myxozoan host specificity, it is

- 113 - -IV. Discussion- imperative to clarify how a species or genus-specific myxozoan interacts with non- compatible fish species. In particular, the fate of the developmental stages after a possible entry into those non-susceptible fish species needs to be examined. During transmission, it is necessary for actinosporean stages to avoid accidental discharge at contact not only with dead organic matter, but also with aquatic invertebrate surfaces and secretions (most notably of their own oligochaete hosts) and other plankton organisms such as crustaceans and protozoans. This might be reflected by the choice of the discharge triggers, as the most obvious mucus components, carbohydrates, appear to be unsuitable to meet this problem. The discharge signal seems to be unapparent in L. stagnalis mucus, which indicates avoidance of attachment to invertebrates including the oligochaete host. Mollusc epithelial mucins possess great amounts of polysaccharides exhibiting amino sugars, glycoproteins, uronic acids and hexose sulphate (Denny 1983; Kalbe et al. 2000), analogous to vertebrate mucins. The efficacy of these glycoconjugates in stimulating discharge of actinospores can be neglected considering the results herein. A vertebrate-specific signal would enable actinospores to avoid attachment to aquatic invertebrates and explain the response to all effective test substrates. Nevertheless, this factor requires further investigation with other organisms (e.g. oligochaetes, hirudineans and molluscs). The presence of purine nucleosides in the bioactive subfractions was confirmed by NMR analysis in this study. These substances exist in manifold variations and play important roles in cellular signalling pathways as well as in ion channel activation (Hurley 2003). If the result that nucleotide derivatives act as polar filament discharge triggers for actinospores could be confirmed, questions are raised concerning the reasons why such components were chosen as host cues during evolution. IMP was also recognised to be a flavour enhancer (Kawai et al. 2002) and is recognised by gustatory tissues of turbot (Mitchell & Mackie 1983). N-oxide purines have indeed been shown to act as pheromones for fishes (Pfeiffer et al. 1985). It cannot be confirmed whether the nucleotide derivatives that may be responsible for polar filament discharge are ATP catabolites or (in the case of cyclic species) serve as hormones or second messengers like cAMP or cGMP. A pheromone function must also be considered. The fact that cyclic nucleotides act as second messengers only in animals provides a suitable explanation as to why parasite transmission stages like actinospores might choose such signals in order to avoid reactions to plant material.

- 114 - -IV. Discussion-

Bacterial and enzymatic activity is responsible for the conversion of metabolic nucleotides and all their derivatives (Howgate 2005). ATP follows the degradation scheme: ATP→ADP→AMP→IMP→Inosine→Hypoxanthine→Xanthine. The natural breakdown cascade might be the most important factor against a ubiquitous presence of nucleosides and their derivatives in aquatic environments. When recognised by actinospores (maybe exceeding a distinct threshold concentration), this ensures that the probability for a contact with a living fish is very high. By these means, myxozoan actinospores utilise recognition cues that cannot be simply abolished by fish. They seem to be reliable signals for differentiation between vertebrates and must be absent in integument or mucus secretions of other aquatic organisms. More experimental trials are needed to compare the discharge triggering activity of mucus isolates from more fish species. This, as a matter of course, is only reasonable when the molecular class of the chemicals causing discharge has been verified. The occurrence of discharge triggering molecules, once they are identified, could then be examined in muci of other animals such as aquatic snails or annelids. Furthermore, it can be estimated whether polluted or clean waters have beneficial effects on transmission of actinospores. Extensive effort should also be undertaken to find out more about the characteristics of the mechano-sensitive structures. Today, the main problem in research on myxozoan parasites is the low number of well established laboratory cycles, as field material is not suitable for most studies. Difficulties generating these problems are the parasite-free maintenance of both hosts, the long generation periods and a lack of experimental data on life cycles and infection modi of different species. Even in long known myxozoan species, our knowledge of developmental and transmission mechanisms remains fragmentary (Yokoyama 2003). The elucidation of these mechanisms is important for our understanding of transmission dynamics in the field. Additionally, premature or unspecifically induced discharge, actuated by biochemicals or decoy organisms, may provide an effective strategy to reduce infectivity of both actinospores and the myxosporean spore and reduce parasite burden in hatcheries and in the field. Reliable data on these topics, including the isolation of the triggering substances, can only be achieved through biophysical and behavioural experimental approaches using host material and laboratory cycles.

- 115 - -V. Summary-

V. Summary

Up to now, little information is available on the transmission mechanisms of the Myxozoa. Parasites of this huge phylum mostly infect teleost hosts via characteristic actinospores produced by secondary hosts. This study elucidated the life cycle of the myxozoan parasite Henneguya nuesslini in two salmonid host species. Naïve brown trout and brook trout were experimentally infected in two trials by triactinomyxon type actinospores from naturally infected Tubifex tubifex. In exposed common carp, no myxospore production was detected. Actinosporean and myxosporean stages are described by light microscopy and a 1417 bp fragment of the 18S rDNA gene was sequenced. This is the first record of a Henneguya sp. life cycle with a triactinomyxon- type actinospore, which suggests a close relationship to the Myxobolus group and the polyphyletic origin of the genus Henneguya. Myxobolus parviformis sp. n., a new species from the gills of common bream was also described. The development of both stages was achieved in two consecutive full transmission trials by isolating plasmodia of a common genotype for separation from sympatric Myxobolus spp. occurring in naturally infected gill lamellae. Isolated gill plasmodia representing individual ‘clones’ were used for subsequent infection of oligochaetes after molecular and morphological identification. The host cues which stimulate discharge of polar filaments of Myxobolus cerebralis actinospores were studied. Individual actinospores were locally immobilised in an experimental set-up and homogenised rainbow trout mucus as chemostimulus and tangency of the apical region of the spores as mechanical stimulation were applied subsequently. The actinospores showed discharged polar filaments exclusively when mucus substrate application was followed by mechanical stimulation of the polar capsule-bearing region, not when one stimulus was offered solely. Polar filament discharge rates were measured using mucus preparations in a microscopic assay applying supplementary vibration stimuli to ensure mechanical excitation. The actinospores responded similarly to different frequencies of vibration, which suggested a touch sensitive recognition mechanism. The discharge response of actinospores of three myxozoan species was not stimulated host-specifically: Mucus of non-susceptible fish species stimulated similar filament discharge rates as mucus from host species. Homogenised frog epidermis and bovine submaxillary mucin could also stimulate the attachment reaction but mucus of a pulmonate freshwater snail was

- 116 - -V. Summary- ineffective. Experimental results indicated that polar filament discharge of H. nuesslini actinospores is independent from external calcium ions. An extensive biochemical isolation of the discharge triggering chemostimuli from fish surface mucus was conducted. Non-volatile inorganic electrolytes, all volatiles, free L-amino acids, glycosaminoglycans, bound and free hexoses and sialic acids, proteins, urea and inositols could be excluded as triggers for polar filament discharge in M. cerebralis actinospores. UV-detectable substances, possibly nucleoside derivatives, were isolated as bioactive fraction. The myxozoan transmission by actinospores was further investigated by analysing parameters influencing sporoplasma emission. This reaction was found to be independent from polar filament discharge. An updated view on the sequence of the events during host invasion by actinospore transmission stages is presented. It reflects a perfect adaptation for the successful recognition and invasion of an aquatic host without energetically cost- intensive active swimming behaviour.

- 117 - -VI. Zusammenfassung-

VI. Zusammenfassung

Die Übertragung von Fischparasiten des Stammes der Myxozoa wurde untersucht, wobei die Lebenszyklen zweier Arten aufgeklärt werden konnten. Die Entwicklung von Henneguya nuesslini wurde in Bachforellen und Bachsaiblingen experimentell verfolgt. In zwei Transmissionsexperimenten gelang die Übertragung durch Actinosporen von natürlich infizierten Oligochaeten (Tubifex tubifex) auf die Salmoniden, nicht aber auf Karpfen. Actinosporenstadien und Myxosporenstadien wurden lichtmikroskopisch, sowie basierend auf amplifizierten Fragmenten des 18S rDNA Gens, umfassend beschrieben. Erstmals wurde als alternierendes Transmissionsstadium dieser Gattung der Morphotypus der Triactinomyxonspore gefunden, was die paraphyletische Natur des Genus sowie eine neue phylogenetische Einordnung untermauert. Desweiteren wurde der Entwicklungszyklus von Myxobolus parviformis sp. n. aufgeklärt, einer bislang unbeschriebenen Art, die in Kiemenlamellen von Brassen parasitiert. Die Entwicklung beider Transmissionsstadien gelang ausgehend von Myxosporen aus dem Gewebe natürlich infizierter Fische. Durch Isolierung reifer Plasmodien eines Genotyps von sympatrisch vorkommenden Arten und einer Erfassung der molekularen und morphologischen Merkmale der Transmissionsstadien konnte der Lebenszyklus reproduzierbar etabliert werden. Zur Anheftung an Fische nutzen die Actinosporen der Myxozoa ausschleuderbare Filamente, die in für die Gruppe charakteristischen Polkapseln enthalten sind. Erstmals wurde nachgewiesen, dass die Polfadenreaktion bei Myxobolus cerebralis Actinosporen durch eine Kombination chemischer und mechanischer Wirtssignale ausgelöst wird. Einzelnen immobilisierten Individuen wurde als chemisches Signal homogenisierter Regenbogenforellen-Mucus appliziert. Die Actinosporen schleuderten ihre Filamente nur dann aus, wenn ein mechanischer Reiz dem chemischen voranging. Zur Untersuchung der chemischen Wirtssignale wurde die Rate an Actinosporen erfasst, die ihre Filamente ausschleuderten. Hierfür wurde ein Verfahren entwickelt, daß nach Applikation der Testsubstrate und eines Vibrationsreizes zur mechanischen Stimulierung eine lichtmikroskopische Auswertung auf Objektträgern erlaubte. M. cerebralis Actinosporen reagierten auf Forellenmucus nach Stimulierung mit verschiedenen Vibrationsfrequenzen, was einen Mechanismus zur Thigmoperzeption nahelegt. Es konnte keine Wirtsspezifität bei der Auslösung der Polfadenreaktion aller drei Arten beobachtet werden. Lediglich der Mucus eines

- 118 - -VI. Zusammenfassung- aquatischen Pulmonaten war unwirksam. Experimente mit H. nuesslini Actinosporen zeigten, dass die Polfadenreaktion von Calcium-Ionen unabhängig ist. Eine umfangreiche biochemische Fraktionierung von Forellenmucushomogenisat diente der Isolierung von chemischen Komponenten, welche die Polfadenreaktion auslösen. Flüchtige Komponenten, anorganische Elektrolyte, freie Aminosäuren, Glycosaminoglycane, gebundene und freie Hexosen, Sialinsäuren, Proteine, Harnstoff und Inositolderivate aus dem Wirtsmucus konnten als Stimuli für die Polfadenreaktion bei M. cerebralis Actinosporen ausgeschlossen werden. UV-detektierbare Substanzen, möglicherweise Nucleoside, wurden mittels chromatographischer Auftrennung kleinmolekularer Mucusfraktionen als Stimuli für die Polfadenreaktion isoliert. Darüber hinaus wurden Faktoren analysiert, welche die aktive Emission der infektiösen Stadien (Sporoplasmen) bedingen. Es zeigte sich, dass die Auslösung dieser Reaktion unabhängig von der Polfadenreaktion ist. Anhand experimenteller Resultate wird eine aktuelle Sicht der Invasion von Fischen durch Actinosporen diskutiert. Die Sequenz der Ereignisse und spezifischen Reaktionen während der Transmission zeigen deren perfekte Anpassung an eine erfolgreiche Erkennung und Invasion aquatischer Wirte ohne energieaufwendiges, aktives Schwimmverhalten.

- 119 - -VII. Acknowledgements-

VII. Acknowledgements

First of all, my thanks go to my parents for the all patience and giving me the chance to study. Besides the whole parasitology group, I would like to thank Prof. Dr. Wilfried Haas for giving me the chance to work on this topic, his ultimate support in all aspects of scientific activities and being a great Doktorvater. Special thanks go to Prof. Dr. Mansour El-Matbouli for his trust and cooperation, manuscript correction and his open ear in problematic situations. I also wish to thank Christina Loy for her kind help, conversation and heaps of fun in the lab and Dr. Jan Hertel for his manifold help and revision of the manuscript. Additional thanks to Dr. Bernhard Haberl, Dr. Ralph Rübsam and Dr. Martin Kalbe for fruitful discussions and methodical advisory. Special appreciation goes to Dr. Edit Eszterbauer for all her kind help and being a great colleague. This study would not have been possible without the help and support of the staff of the Lehranstalt für Fischerei in Aufseß, especially Manfred Popp and his co-workers. Many thanks also to all people that practically helped me or contributed to this work: Dr. Martin Knaus, Dr. Stephan Adelt, Prof. Dr. Walter Bauer, Prof. Dr. Roland Schauer, Prof. Dr. Rudolf Geyer, Dr. Peter Kaese, Prof. Dr. Christer Erséus, Dipl. Biol. Jennifer Borrelli and Dipl. Biol. Sabine Ponader. Last but not least I thank my band and my friends for being so patient and keeping me alive and all the parasites that have taken over the non-parasitic world without public notice. This study was supported by the Deutsche Forschungsgemeinschaft. The work on the identification of the oligochaete hosts was supported by the Swedish Research Council, grant no. 621-2001-2788 (to Prof. Dr. Christer Erséus, while still at the Swedish Museum of Natural History, Stockholm); oligochaete sequencing was done by Ms. Bodil Cronholm. All other molecular biological work was supported by the Hungarian Scientific Research Fund (OTKA) grants No. F045908 and T042464.

- 120 - -VIII. References-

VIII. References

Adelt S, Plettenburg O, Stricker R, Reiser G, Altenbach HJ and Vogel G (1999) Enzyme-assisted total synthesis of the optical antipodes D-myo-inositol 3,4,5-trisphosphate and D-myo-inositol 1,5, 6- trisphosphate: aspects of their structure-activity relationship to biologically active inositol phosphates. J Med Chem 42, 1262-1273. Agapow PM, Bininda-Emonds OR, Crandall KA, Gittleman JL, Mace GM, Marshall JC and Purvis A (2004) The impact of species concept on biodiversity studies. Q Rev Biol 79,161-179. Anderson CL, Okamura B and Canning EU (1998) Hox genes confirm that Myxozoa are extremely reduced triploblasts. Nature 392, 346-347. Andree KB, Székely Cs Molnár K, Gresoviac SJ and Hedrick, RP (1999) Relationships among members of the genus Myxobolus (Myxozoa: Bivalvulida) based on small subunit ribosomal DNA sequences. J Parasitol 85, 68-74. Asakawa M (1974) Sialic acid containing glycoprotein in the external mucus of eel, Anguilla japonica Temminck et Schlegel-II Carbohydrate and amino acid compo-sition. Bull Jap Soc Sci Fish 40, 303- 308. Azevedo C and Matos E (2003) Fine structure of Henneguya pilosa sp. n. (Myxozoa: Myxosporea), parasite of Serrasalmus altuvei (Characidae), in Brazil. Folia Parasitol 50, 37-42. Barry JM and O’Rourke (1959) Species specifity of fish mucus. Nature 184, 2039. Bartholomew JL (1998) Host resistance to infection by the myxosporean parasite Ceratomyxa shasta: a review. J Aquat Anim Health 10, 112. Beauchamp KA, Kathman RD, McDowell TS and Hedrick RP (2001) Molecular phylogeny of tubificid oligochaetes with special emphasis on Tubifex tubifex (Tubificidae). Mol Phylogenet Evol 19, 216- 224. Bernt E and Bergmeyer HU (1970) Bestimmung von Harnstoff. In: Methoden der Enzymatischen Analyse, 2nd edn., vol. II (HU Bergmeyer, ed.), pp. 1659–1663, Verlag Chemie, Weinheim Bhavanandan VP and Sheykhnazari M (1993) Adaptation of the periodate/resorcinol method for determination of sialic acids to a microassay using microtiter plate reader. Anal Biochem 213, 438- 440. Bradford M (1976) A rapid and sensitive method for microgram quantities of protein utilizing the principle of protein - dye binding. Anal biochem 72, 245 - 248. Browne RK, Scheltinga DM, Pomering M and Mahony M (2002) Testicular myxosporidiasis in anurans, with a description of Myxobolus fallax n. sp. Systematic Parasitology 52, 97–110. Buck LT (2004) Adenosine as a signal for ion channel arrest in anoxia-tolerant organisms. Comp Biochem Physiol B 139, 401-414. Burt JR (1976) Hypoxanthine: a biochemical index of fish quality. Process Biochem 12, 32 -35. Canning EU and Okamura B (2004) Biodiversity and evolution of the Myxozoa. Adv Parasitol 56, 43- 131.

- 121 - -VIII. References-

Cannon Q and Wagner E (2003) Comparison of discharge mechanisms of cnidarian Cnidae and myxozoan polar capsules. Rev Fish Science 11, 185-219. Carisano A (1985) Rapid and sensitive method for the determination of proline by reversed-phase high- performance liquid chromatography with automated pre-column fluorescence derivatization. J Chromatogr 318, 132-138. Carlstedt I, Sheehan J, Corfield A and Galagher J (1985) Mucous glycoproteins: A gel of a problem. Essays Biochem 20, 41-69. Carsol MA and Mascini M (1998) Development of a system with enzyme reactors for the determination of fish freshness. Talanta, 47335-47342. Casal G, Matos E and Azevedo C (2003) Light and electron microscopic study of the myxosporean, Henneguya friderici n. sp. from the Amazonian teleostean fish, Leporinus friderici. Parasitology 126, 313-319. Castell JV, Cervera M and Marco R (1979) A convenient micromethod for the assay of primary amines and proteins with fluorescamine. A reexamination of the conditions of reaction. Anal Biochem 99, 379-391. Corbin JD, Krebs EG (1969) A cyclic AMP-stimulated protein. kinase in adipose tissue. Biochem Biophys Res Commun 36, 328-335. Denny M (1983) Molecular biomechanics of molluscan mucous secretions. In The Mollusca Vol.1 (Hochachka PW, ed), pp 431-465. Academic Press, New York. Edge AS (2003) Deglycosylation of glycoproteins with trifluoromethanesulphonic acid: elucidation of molecular structure and function. Biochem J 376, 339-350. Ehira S and Uchiyama H (1987) Determination of fish freshness using the K value and comments on some other biochemical changes in relation to freshness. In: Seafood quality determination (DE Kramer and J Liston (eds)). Pp. 185-207, Elsevier Science Publishers B.V., Amsterdam. Eiras JC (2002) Synopsis of the species of the genus Henneguya Thélohan, 1892 (Myxozoa: Myxosporea: Myxobolidae). Syst Parasitol 52, 43-54. Eiras JC, Molnar K and Lu YS (2005) Synopsis of the species of Myxobolus Butschli, 1882 (Myxozoa: Myxosporea: Myxobolidae). Syst Parasitol 61, 1-46. El-Mansy A, Szekely C, Molnár K (1998) Studies on the occurrence of actinosporean stages of fish myxosporeans in a fish farm of Hungary, with the description of triactinomyxon, raabeia, aurantiactinomyxon and neoactinomyxon types. Acta Vet Hung 46, 259-84. El-Matbouli M and Hoffmann RW (1991) Effect of freezing, aging and passage through the alimentary canal of predatory animals on the viability of Myxobolus cerebralis-spores. J Aquat Anim Health 3, 260-262. El-Matbouli M and Hoffmann RW (1998) Light and electron microscopic study on the chronological development of Myxobolus cerebralis in Tubifex tubifex to the actinosporean stage triactinomyxon. Int J Parasitol 28, 195-217. El-Matbouli M, Fischer-Scherl T and Hoffmann RW (1992) Present knowledge of the life cycle, taxonomy, pathology and therapy of some Myxosporea spp important for freshwater fish. Ann Rev Fish Dis 367-402.

- 122 - -VIII. References-

El-Matbouli M, Hoffmann RW and Mandok C (1995) Light and electron microscopic observations on the route of the triactinomyxon-sporoplasm of Myxobolus cerebralis from epidermis into the rainbow trout (Oncorhynchus mykiss) cartilage. J Fish Biol 46, 919-935. El-Matbouli M, Hoffmann RW, Schoel H, McDowell TS and Hedrick RP (1999) Whirling disease: host specificity and interaction between the actinosporean stage of Myxobolus cerebralis and rainbow trout Oncorhynchus mykiss. Dis Aquat Org 35, 1-12. Eszterbauer E (2004) Genetic relationship among gill-infecting Myxobolus species (Myxosporea) of cyprinids: molecular evidence of importance of tissue-specificity. Dis Aquat Org 58, 35-40 Eszterbauer E, Benkö M, Molnár K (2002) Differentiation of morphologically very similar gill parasite Myxobolus species (Myxosporea) by restriction fragment length polymorphism (PCR-RFLP) method. Magy Állatorv Lapja 124, 361-366. Ferguson DR and Price RH (1973) The production of cIMP by toad bladder, and its effects on transport of water and salt. FEBS Lett 34, 207-212. Ferguson HW, Morrison D, Ostland VE, Lumsden J and Byrne P (1992) Responses of mucus-producing cells in gill disease of rainbow trout (Oncorhynchus mykiss). J Comp Pathol 106, 255-265. Folch J, Lees M and Sloane Stanley GH (1957) A simple method for the isolation and purification of total lipids from animal tissues. J Biol Chem 226, 497-509. Fürst P, Pollack L, Graser TA, Godel H and Stehle P (1990) Appraisal of four pre-column derivatization methods for the high-performance liquid chromatographic determination of free amino acids in biological materials. J Chromatogr 499, 557-569. Gilbert MA and Granath WO (2003) Whirling disease of salmonid fish: life cycle, biology, and disease. J Parasitol 89, 658-67. Granata L (1922) Richerche sugli Attinomissidi. II. Neoactinomyxon globosum n. g., n. sp. Monit Zool Ital 33, 193-196.

Haas W (1975) Einfluss von CO2 und pH auf das Fixationsverhalten der Cercarie von Diplostomum spathaceum (Trematoda). Z Parasitenk 46, 53-60. Haas W (2003) Parasitic worms: Strategies for host-finding, recognition and invasion. Zoology 106, 349- 364. Haas W and Haberl B (1997) Host recognition by trematode miracidia and cercariae. In Advances in trematode biology (Fried B and Graczyk TK , eds.), pp.197-227. CRC Press, Boaca Raton. Haas W and Ostrowski de Núñez M (1988) Chemical signals of fish skin for the attachment response of Acanthostomum brauni cercariae. Parasitol Res 74, 552-557. Haas W, Haberl B, Hofmann M, Kerschensteiner M and Ketzer U (1999) Ichthyophtirius multifiliis invasive stages find their fish hosts with complex behaviour patterns and in response to different chemical signals. Eur J Protistol 35, 129-135. Haas W, Stiegeler P, Keating A, Kullmann B, Rabenau H, Schönamsgruber E and Haberl B (2002) Diplostomum spathaceum cercariae respond to a unique profile of cues during recognition of their fish host. Int J Parasitol 32, 1145-1154. Hallett SL, Atkinson SD and El-Matbouli M (2002) Molecular characterisation of two aurantiactinomyxon (Myxozoa) phenotypes reveals one genotype. J Fish Dis 25, 627-631.

- 123 - -VIII. References-

Hallett SL, Atkinson SD, Erséus C and El-Matbouli M (2004) Molecular methods clarify morphometric variation in triactinomyxon spores (Myxozoa) released from different oligochaete hosts. Syst Parasitol 57, 1-14. Hallett SL, Atkinson SD, Erséus C, El-Matbouli M (2005) Dissemination of triactinomyxons (Myxozoa) via oligochaetes used as live food for aquarium fish. Dis Aquat Org 65, 137-52. Handy RD (1989) The ionic composition of rainbow trout body mucus. Comp Biochem Physiol 93A, 571-576. Hayashi S (1972) Biochemical studies on the skin of fish-IV. Incorporation of 14C-formate into guanine and hypoxanthine in the skin of fish and accelelating effect of thyroxine on the incorporation. Bull Jap Soc Sci Fish 38, 279-284. Hayashi S and Saito T (1968) Biochemical studies on the skin of fish-I. Acid-soluble nucleotides in the dermis of some fishes. Bull Jap Soc Sci Fish 34, 816-826. Hedrick RP and El-Matbouli M (2002) Recent advances with taxonomy, life cycle, and development of Myxobolus cerebralis in the fish and oligochaete hosts. Am Fish So Symp 29, 45-53. Hedrick RP, El-Matbouli M, Adkison MA and MacConnell E (1998) Whirling disease: re-emergence among wild trout. Immunol Rev 166, 365-376. Hedrick RP, McDowell TS, Gay M, Marty GD, Georgiadis MP and MacConnell E (1999) Comparative susceptibility of rainbow trout Oncorhynchus mykiss and brown trout Salmo trutta to Myxobolus cerebralis, the cause of salmonid whirling disease. Dis Aquat Organ 37, 173-83. Hervio DML, Kent ML, Khattra J, Sakanari J, Yokoyama H and Devlin RH (1997) Taxonomy of Kudoa species (Myxosporea), using a small-subunit ribosomal DNA. Can J Zoolog 75, 2112-2119. Hillis DM, Dixon T (1991) Ribosomal DNA: Molecular evolution and phylogenetic inference. Quarterly Rev Biol 66, 411-453. Howgate P (2005) Kinetics of degradation of adenosine triphosphate in chill-stored rainbow trout (Oncorhynchus mykiss). Int J Food Sci Technol 40, 579-588. Hurley JH (2003) GAF domains: Cyclic nucleotides come full circle. Sci STKE, PE1. Ibragimov A (2001) Polar filaments of Myxobolus pseudodispar spores. Tsitologiia 43, 742-746. Ingvarsdottir A, Birkett MA, Duce I, Mordue W, Pickett JA, Wadhams LJ, and Mordue Luntz AJ (2002) Role of semiochemicals in mate location by parasitic sea louse, Lepeophtheirus salmonis. J Chem Ecol 28, 2107-2117. Janiszewska J (1955) Actinomyxidia. Morphology, ecology, history of investigations, systematics, development. Acta Parasitol Pol 2, 405–443. Jirovec O (1940) Zur Kenntnis einiger in Oligochaeten parasitierenden Protisten. II. Über Sphaeractinomyxon ilyodrili n. sp. und Neoactinomyxon globosum Granata 1922. Arch Protistenkd 94, 212-223 Jourdian GW, Dean L and Roseman S (1971) The sialic acids. XI. A periodate-resorcinol method for the quantitative estimation of free sialic acids and their glycosides. J Biol Chem 246, 430-435. Kalbe M (1998) Chemoorientierung von Miracidien zur Wirtsschnecke: Signale und Spezifität. PhD thesis, University of Erlangen-Nuremberg.

- 124 - -VIII. References-

Kalbe M, Haberl B and Haas W (2000) Finding of the snail host by Fasciola hepatica and Trichobilharzia ocellata: Compound analysis of ´Miracidia attracting glycoprotein´. Exp Parasitol 96, 231-242. Kallert DM, El-Matbouli M and Haas W (2005c) Polar filament discharge of Myxobolus cerebralis actinospores is triggered by combined non-specific mechanical and chemical cues. Parasitology 131, 609-616. Kallert DM, Eszterbauer E, El-Matbouli M, Erséus C and Haas W (2005a) The life cycle of Henneguya nuesslini Schuberg & Schröder 1905 (Myxozoa) involves a triactinomyxon-type actinospore. J Fish Dis 28, 71-79. Kallert DM, Eszterbauer E, Erséus C, El-Matbouli and Haas W (2005b) Life cycle studies of Myxobolus parviformis sp.n. from bream. Dis Aquat Org 66, 233-243. Kass-Simon G and Hufnagel LA (1992) Suspected chemoreceptors in coelenterates and ctenophores. Microsc Res Techniq 22, 265-284. Kawai M, Okiyama A and Ueda Y (2002) Taste enhancements between various amino acids and IMP. Chem Senses 27, 739-745. Kearn GC (1967) Experiments on host-finding and host specificity in the monogean skin parasite Entobdella soleae. Parasitology 57, 585-605. Kent ML, Andree KB, Bartholomew JL, El-Matbouli M, Desser SS, Devlin RH, Feist SW, Hedrick RP, Hoffmann RW, Khattra J, Hallett SL, Lester RJG, Longshaw M, Palenzeula O, Siddall ME and Xiao C (2001) Recent advances in our knowledge of the Myxozoa. J Eukaryot Microbiol 48, 395-413. Kent ML, Margolis L, Whitaker DJ, Hoskins GE and McDonald TE (1994) Review of Myxosporea of importance in salmonid fisheries and aquaculture in British Columbia. Folia Parasitol 41, 27-37. Leray C, Raffin JP, Winninger C. (1979) Aspects of purine metabolism in the gill epithelium of rainbow trout, Salmo gairdneri Richardson. Comp Biochem Physiol B. 62, 31-40. Lom J (1990) Phylum Myxozoa. In: Handbook of Protocotista (L Margulis, JO Corliss, M Melkonian and DJ Chapman, eds), pp. 36-52. Boston: Jones and Bartlett. Lom J and Arthur JR (1989) A guideline for the preparation of species descriptions in Myxosporea. J Fish Dis 12, 151-156 Lom J and Dyková I (1992) Protozoan parasites of fishes. Dev Aquacult Fish Sc 26, Elsevier, Amsterdam-London-New York-Tokyo. Markiw ME (1992) Experimentally induced whirling disease II. Determination of longevity of the infective triactinomyxon stage of Myxobolus cerebralis by vital staining. J Aquat Animal Health 4, 44-47. Marques A (1982) La spore des Actinomyxidies: Synactinomyxon longicauda n. sp., unnouveaue type de groupement sporal et adaptations planctoniques. Emission et structure du sporoplasms. J Protozool 29, 195–202. Mawhinney TP and Chance DL. (1994) Hydrolysis of sialic acids and O-acetylated sialic acids with propionic acid. Anal Biochem 223, 164-167. McAllister CT, Bursey CH, Upton SJ, Trauth SE and Conn DB (1995) Parasites of Desognathus brimleyorum (Caudata: Plethodontidae) from the Ouachita Mountains of Arkansas and Oklahoma. J Helminthol Soc W 62, 150-156.

- 125 - -VIII. References-

McGeorge J, Sommerville C and Wootten R (1997) Studies of actinosporean myxozoan stages parasitic in oligochaetes from the sediments of a hatchery where Atlantic salmon harbour Sphaerospora truttae infection. Dis Aquat Org 30, 107-119. Meier-Brook C (1978) Calcium uptake by Marisa cornuarietis (Gastropoda; Ampulariidae), a predator of schistosome-bearing snails. Arch Hydrobiol 82, 449–464. Mendes R, Quinta R, Nunes ML (2001) Changes in baseline levels of nucleotides during ice storage of fish and crustaceans from the Portuguese coast. European Food Research and Technology 202, 141–146. Mitchell AI and Mackie AM (1983) Binding of [14C]inosine to gustatory tissues of the turbot, Scophthalmus maximus. Comp Biochem Physiol A.75, 471-474. Molnár K (1994) Comments on the host, organ and tissue specificity of fish myxosporeans and on the types of their intrapiscine development. Parasitol Hung 27, 5-20. Molnár K (2002) Site preference of fish myxosporeans in the gill. Dis Aquat Org 48, 197-207. Molnár K, Eszterbauer E, Székely C, Dán Á, Harrach B (2002) Morphological and molecular biological studies on intramuscular Myxobolus spp. of cyprinid fish. J Fish Dis 25, 643-652. Moncada LI, Lopez MC, Murcia MI, Nicholls S, Leon F, Guio OL and Corredor A (2001) Myxobolus sp., another opportunistic parasite in immunosuppressed patients? J Clin Microbiol 39, 1938-40. Monsigny M, Petit C and Roche AC (1988) Colorimetric determination of neutral sugars by a resorcinol sulfuric acid micromethod. Anal Biochem Dec 175, 525-530. Müller J (1841) Krankhafter Hautausschlag mit specifisch organisirten Samenkörperchen (Psorospermien). Bericht der Preussischen Akademie für Wissenschaft 212, 246. Murray J, Thomson AB and McGill, AS (1985) Estimation of nucleotides and related products in fish muscle by reverse phase ion pair HPLC. J High Res Chrom 6, 209-210. Nakagawa H, Asakawa M and Enomoto N (1988) Diversity in the carbohydrate moieties of mucous glycoproteins of various fishes. Nippon Suisan Gakkaishi 54, 1653-1658. Nakagawa H, Hama Y, Sumi T, Li SC and Li YT. (2001) KDN-containing glycoprotein from loach skin mucus. Adv Exp Med Biol 491, 171-84. Nehring RB and Walker PG (1996) Whirling disease in the wild: The new reality in the intermountain west. Fisheries (Bertheseda), 21, 28-32. Nehring RB, Thompson KG, Taurman KA and Shuler DL (2002) Laboratory studies indicating that living brown trout Salmo trutta expel viable Myxobolus cerebralis myxospores. In Whirling disease: Reviews and current topics (Bartholomew JL and Wilson JC, eds), pp. 125-134. Am Fish Soci, Bethesda. Ogawa K, KP Delgahapitiya, T Furuta and Wakabayashi H (1992) Histological studies on the host response to Myxobolus artus Akhemerov, 1960 (Myxozoa: Myxobolidae) infection in the skeletal muscle of carp, Cyprinus carpio L. J Fish Biol 41, 363-371. Okamura B, Curry A, Wood TS and Canning EU (2002) Ultrastructure of Buddenbrockia identifies it as a myxozoan and verifies the bilaterian origin of the Myxozoa. Parasitology 124, 215-23. Oumouna M, Hallett SL, Hoffmann RW and El-Matbouli M (2003) Seasonal occurence of actinosporeans (Myxozoa) and oligochaetes (Annelida) at a trout hatchery in Bavaria, Germany. Parasit Res 89, 170-184.

- 126 - -VIII. References-

Overstreet RM (1976) Fabespora vermicola sp. n., the first myxosporidian in from a platyhelminth. J Parasitol 62, 680-684. Özer A and Wootten R (2002) Biological characteristics of some actinosporeans. J Nat Hist 36, 2199- 2209. Özer A, Wootten R, Shinn AP (2002) Survey of actinosporean types (Myxozoa) belonging to seven collective groups found in a freshwater salmon farm in Northern Scotland. Folia Parasit 49, 189- 210. Padrŏs F, Palenzuela O, Hispano C, Tosas O, Zarza C, Crespo S and Alvarez-Pellitero P. (2001) Myxidium leei (Myxozoa) infections in aquarium-reared Mediterranean fish species. Dis Aquat Organ. 47, 57-62. Paperna I (1973) Occurence of Cnidospora infections in freshwater fisches in Africa. Bull Inst Fondam Afr Noire, A 35, 509-521. Pfeiffer W, Riegelbauer G, Meier G and Scheibler B (1985) Effect of hypoxanthine-3(N)-oxide and hypoxanthine-1(N)-oxide on central nervous excitation of the black tetra Gymnocorymbus ternetzi (Characidae, Ostariophysi, Pisces) indicated by dorsal light response. J Chem Soc 11, 507-523. Ponader S (2005) Myxozoa: Chemische Wirtserkennung durch Actinosporen. Diploma thesis, University of Erlangen-Nuremberg. Rahemtulla F and Løvtrup S (1974) The comparative biochemistry of invertebrate mucopolysaccharides II. Nematoda; Annelida. Comp Biochem Physiol 49B, 639-646. Reed CC, Basson L and Van As LL (2003) Myxozoans infecting the sharptooth catfish, Clarias gariepinus in the Okavango River and Delta, Botswana, including descriptions of two new species, Henneguya samochimensis sp. n. and Myxobolus gariepinus sp. n. Folia Parasit 50, 183-189. Reid L and Clamp J (1978) The biochemical and histochemical nomenclature of mucus. Brit Med Bull 34, 5-8. Roberts SD and Powell MD (2005) The viscosity and glycoprotein biochemistry of salmonid mucus varies with species, salinity and the presence of amoebic gill disease. J Comp Physiol [B] 175, 1-11. Saito T, Arai KI, Yajima T (1959) Changes in purine nucleotides of red lateral muscle of rainbow trout. Nature 184 (Suppl 18),1415. Schlegel M, Lom J, Stechmann A, Bernhard D, Leipe D, Dyková I and Sogin M L (1996) Phylogenetic analysis of complete small subunit ribosomal RNA coding region of Myxidium lieberkuehni: evidence that Myxozoa are Metazoa related to the Bilateria. Arch Protistenkd 147, 1-9. Schuberg A and Schröder O (1905) Myxosporidien aus dem Nervensystem und der Haut der Bachforelle (Myxobolus neurobius n. sp., u. Henneguya nuesslini n. sp.). Arch Protistenkd 6, 47-60. Shulman SS (1966) The myxosporidian fauna of the USSR. Nauka, Moscow-Leningrad, 504 pp. Siau Y, Gasc C and Maillard C (1981) Premières obsevations ultrastructurales dùne myxosporidie appartenant au genre Fabespora, parasite de trématode. Protistologica 17, 131-137. Siddall M E, Martin DS, Bridge D, Desser SS and Cone DK (1995) The demise of a phylum of protists: phylogeny of myxozoa and other parasitic cnidaria. J Parasitol 8, 961-967. Siegenthaler U and Ritter W (1977) Mitteilungen aus dem Gebiete der Lebensmitteluntersuchung und Hygiene 68, 448-450.

- 127 - -VIII. References-

Sinha RP, Sinha JP, Groniger A and Häder DP (2002) Polychromatic action spectrum for the induction of a mycosporine-like amino acid in a rice-field cyanobacterium, Anabaena sp. J Photochem Photobiol B. 66, 47-53. Smith M. (2001) Extrusion of polar filaments in Myxobolus cerebralis actinospores. The Ichthyogram Vol. 11, 3-5. Starcher B (2001) A ninhydrin-based assay to quantitate the total protein content of tissue samples. Anal Biochem 292, 125-129. Stevens R, Kerans BL, Lemmon JC, and Rasmussen C (2001) The effects of Myxobolus cerebralis myxospore dose on triactinomyxon production and biology of Tubifex tubifex from two geographic regions. J Parasitol 87, 315-321. Stiegeler P (1997) Chemische Signale für das Verbleibeverhalten der Cercarien von Diplostomum pseudospathaceum am Fischwirt. Diploma thesis, University of Erlangen-Nuremberg. Štolc A (1899) Actinomyxidies, nouveau groupe de Mésozoaires parent des Myxosporidies. Bulletin International de l’Académemie des Sciences de Bohème 22, 1-12. Sundaralingam M, Haromy TP and Prusiner P (1982) The structure of the zwitterion inosine cyclic 3',5'- monophosphate (cIMP) monohydrate. Analysis of torsional flexibility in the furano-phosphate moiety. Acta Cryst B 38, 1536-1540. Székely C, Rácz O, Molnár K, Eszterbauer E (2002) Development of Myxobolus macrocapsularis (Myxosporea: Myxobolidae) in an oligochaete alternate host, Tubifex tubifex. Dis Aquat Org 48, 117-123. Taylor JW, Jacobson DJ, Kroken S, Kasuga T, Geiser DM, Hibbett DS and Fisher MC (2000) Phylogenetic species recognition and species concepts in fungi. Fungal Genet Biol. 31, 21-32. Thompson JD, Gibson TJ, Plewniak F, Jeanmougin F, Higgins DG (1997) The ClustalX windows interface: flexible strategies for multiple sequence alignment aided by quality analysis tools. Nuclei Acids Res 24, 4876-4882. Thompson K G, Nehring RB, Bowden DC and Wygant T (1999) Field exposure of seven species or subspecies of salmonids to Myxobolus cerebralis in the Colorado River, Middle Park, Colorado J Aquat Anim Health 11, 312. Timm T (1999) A guide to the Estonian Annelida. Estonian Academy Publishers, Tartu-Tallin. 208 pp. Uspenskaya, AV (1982) New data on the life cycle and biology of Myxosporidia. Arch Protistenk 126, 309-338. Uspenskaya AV (1995) Alteration of actinosporean and myxosporean phases in the life cycle of Zschokkella nova (myxozoa). J Eukaryot Microbiol 42, 665-668. Vital P, Corral L, Matos E and Azevedo C (2003) Ultrastructural aspects of the myxosporean Henneguya astyanax n. sp. (Myxozoa: Myxobolidae), a parasite of the Amazonian teleost Astyanax keithi (Characidae). Dis Aquat Organ 53, 55-60. Wagner E (2001a) Triactinomyxon polar filament discharge: Effects of mucus, magnetic fields, and MS- 222. The Ichtyogram 12 , 6-7. Wagner E (2001b) Effect of Calcium chelators, proline, and glutathione on polar filament extrusion of tractinomyxons of Myxobolus cerebralis. The Ichtyogram 12, 5-8.

- 128 - -VIII. References-

Wagner E (2002) Identification of Henneguya salmonicola in Mountain Whitefish in the Logan River Utah. Ichtyogram 13, 4-6. Watson GM and Hessinger DA (1987) Receptor-mediated endocytosis of a chemoreceptor involved in triggering the discharge of cnidae in a sea anemone tentacle. Tissue Cell 19, 747-755. Whitehead DL (1978) Application of a rapid method for determining glycosaminoglycans in mucus secreted by an aquatic pulmonate Biomphalaria glabrata. Comp Biochem Physiol 59A, 27-30. Wolf K and Markiw ME (1984) Biology contravenes taxonomy in the Myxozoa: new discoveries show alternation of invertebrate and vertebrate hosts. Science 225, 1449–1452. Wright CA, Harris RH and Claugher D (1959) The application of paper chromatography to a taxonomic study in the molluscan genus Lymnaea. J Linn Soc Lond 44, 222-237. Xiao C and Desser SS (1998a) Actinosporean stages of myxozoan parasites of oligochaetes from Lake Sasajewun, Algonquin Park, Ontario: New forms of echinactinomyxon, neoactinomyxum, aurantiactinomyxon, guyenotia, synactinomyxon and antonactinomyxon. J Parasitol 84, 1010–1019. Xiao C and Desser SS (1998b) Actinosporean stages of myxosporean parasites of oligochaetes from Lake Sasajewun, Algonquin Park, Ontario: New forms of triactinomyxon and raabeia. J Parasitol 84, 998–1009. Xiao C and Desser SS (2000) Molecular characterization of myxozoan parasites from Lake Sasajewun, Algonquin Park, Ontario, by riboprinting. J Eukaryot Microbiol 47, 85-89. Yokoyama H (1997) Transmission of Thelohanellus hovorkai Achmerov, 1960 (Myxosporea: Myxozoa) to common carp Cyprinus carpio through the alternate oligochaete host. Syst Parasitol 36, 79–84. Yokoyama H and Urawa S (1997) Fluorescent labelling of actinospores for determining the portals of entry into fish. Dis Aquat Org, 30, 165-169 Yokoyama H, Ogawa K and Wakabayashi H (1991) A new collection method of actinosporeans — a probable infective stage of myxosporeans to fishes — from tubificids and experimental infection of goldfish with the actinosporean, Raabeia sp. Fish Pathology (Gyobio Kenkyu) 26, 133–138. Yokoyama H, Ogawa K and Wakabayashi H (1993) Some biological characteristics of actinosporeans from the oligochaete Branchiura sowerbyi. Dis Aquat Organ 17, 223-228. Yokoyama H, Ogawa K and Wakabayashi H (1995a) Chemoresponse of actinosporean spores of Myxobolus cultus to skin mucus of goldfish Carassius auratus. Dis Aquat Organ 21, 7-11. Yokoyama H, Ogawa K and Wakabayashi H (1995b) Myxobolus cultus n. sp. (Myxosporea: Myxobolidae) in the goldfish Carassius auratus transformed from the actinosporean stage in the oligochaete Branchiura sowerbyi. J Parasitol 81, 446–451. Yokoyama, H (2003) A review: Gaps in our knowledge on Myxozoan parasites of fishes. Fish Pathol 38, 125-136. Zandt F (1923) Über ein Myxosporid des Seesaiblings (Salmo salvelinus L.). Zool Anz 57, 21-24.

- 129 - -IX. Appendix-

IX. Appendix

Table 1 Primers used for PCR and/or sequencing

Name Sequence Reference

18e 5’-CTG GTT GAT TCT GCC AGT-3’ Hillis & Dixon 1991 18g’ 5’-CGG TAC TAG CGA CGG GCG GTG TG-3’ Hillis & Dixon 1991 MX5 5’- CTG CGG ACG GCT CAG TAA ATC AGT-3’ Andree et al. 1999 MX3 5’- CCA GGA CAT CTT AGG GCA TCA CAG A-3’ Andree et al. 1999 MB5r 5’-ACC GCT CCT GTT AAT CAT CAC C-3’ Eszterbauer 2004 MB3f 5’-GAT GAT TAA CAG GAG CGG TTG G-3’ Eszterbauer 2004 MC5 5’-CCT GAG AAA CGG CTA CCA CAT CCA-3’ Molnár et al. 2002 MC3 5’-GAT TAG CCT GAC AGA TCA CTC CAC GA-3’ Molnár et al. 2002

- 130 - -IX. Appendix-

Fig. 1. Flow-chart diagram of the procedure employed for the isolation of a contamination-

free myxozoan lifecycle from naturally infected fish harbouring sympatric species of myxozoans. Myxospore isolates collected from host tissue are added to a mixture of long-term monitored oligochaetes. Produced actinospores are obtained from single infected oligochaetes

and classified by morphology and PCR- RFLP/18S rDNA sequencing. Infection of SPF host fish specimens is done with actinospores of the desired morphotype/genotype (Actinospore type 2 in the example). Myxosporean cysts from these fish are carefully

isolated and the individual morphology and PCR-RFLP pattern/18S rDNA sequence is recorded. If the data are consistent with those obtained from the actinospore morphotype, contents of matching cysts are pooled and used for infection of a SPF oligochaete culture yielding a monospecific myxozoan infection. RFLP of the 18S rDNA with multiple restriction enzymes alone often serves as a sufficient tool to monitor possible contaminations with sympatric species. Together with morphological congruity, repeated passage and comparison of 18S rDNA sequences of the parasite stages at each step provide adequate and reproducible results to establish a laboratory cycle. Parallel to these basic steps, infections of oligochaetes and fish with non-controlled (mixed) remainder parasite isolates may be performed to increase the total yield of parasite stages. These bulk infections are then stepwise purified as described. Spotted objects mark contaminative parasite stages. Dotted lines indicate the comparison of data obtained, shaded symbols designate contaminative myxozoan stages not included for further infection, (*) indicates the need to retain a similarly sized aliquot of animals as a control.

- 131 - -Erklärung-

Erklärung

Diese Arbeit wurde von mir selbständig und nur mit den angegebenen Quellen und Hilfsmitteln angefertigt.

………………………………. Dennis Kallert -Lebenslauf-

Name: Dennis Marc Kallert geboren am: 14.06.1975 Geburtsort: Erlangen Familienstand: ledig Staatsangehörigkeit: deutsch

Ausbildung

1981 - 1985 Grundschule (Adelsdorf )

1985 - 1994 Gymnasium (Höchstadt a. d. Aisch) Abschluß: Allgemeine Hochschulreife Zivildienst

1994 -1995 Pflegedienst beim Diakonischen Werk Erlangen

Akademische Ausbildung

1995 - 2001 Studium der Biologie an der FAU Erlangen-Nürnberg

2000 Diplomprüfung in Zoologie (Hauptfach), Genetik, Virologie und Biophysik

2000 - 2001 Diplomarbeit am Institut für Zoologie I, Parasitologie Abschluß Diplom-Biologe Univ. seit 2001 Promotion am Institut für Zoologie I der FAU Erlangen-Nürnberg, AG Parasitologie

Auslandserfahrung

2002 DAAD-Stipendium; praktische Teilnahme an DFG- Forschungsprojekt in Sulawesi, Indonesien

Berufstätigkeit

1995 -1996 Tätigkeit als Pflegehelfer beim Diakonischen Werk

1996 - 2001 Mitarbeiter im Stadtmuseum Erlangen

1998 - 2001 Tutor bei zoologischen Anfängerkursen (Tierphysiologisches Praktikum) seit 2001 Wissenschaftlicher Angestellter (Institut für Zoologie I der FAU Erlangen-Nürnberg), AG Haas Lehrauftrag: Biologische Übungen für Anfänger II F2/F 3 Praktika (Experimentelle Parasitologie)