Digestive physiology in marine fish larvae

Transcription

1 Digestive physiology in marine fish larvae the regulatory loop between cholecystokinin and tryptic enzyme activity with implications for microdiets Dissertation zur Erlangung des Doktorgrades der Mathematisch-Naturwissenschaftlichen Fakultät der Christian-Albrechts-Universität zu Kiel vorgelegt von Robert Tillner Kiel 2013

2 Erste/r Gutachter/in: Zweite/r Gutachter/in: Prof. Dr. Thorsten Reusch Prof. Dr. Carsten Schulz Tag der mündlichen Prüfung: Zum Druck genehmigt: gez. Prof. Dr. Wolfgang Duschl, Dekan

3 Wissen kann durchaus als Denkhindernis wirken Die Kontrollfrage ist immer: was weiß ich, wenn ich das weiß? Harald Welzer, Selbst denken: Eine Anleitung zum Widerstand

4

5 Content Summary 3 Zusammenfassung 5 Introduction Fisheries and aquaculture 7 Larviculture and larval nutrition 8 Larval feeding and digestion 17 Methodical considerations 24 Thesis outline 26 Chapters I II III Hormonal control of tryptic enzyme activity in Atlantic cod larvae (Gadus morhua): involvement of cholecystokinin during ontogeny and diurnal rhythm Published in Aquaculture , (2013) 27 Evidence for a regulatory loop between cholecystokinin (CCK) and tryptic enzyme activity in Atlantic cod larvae (Gadus morhua) Published in Comparative Biochemistry and Physiology A 166 (3), (2013) 49 The regulatory loop between gut cholecystokinin and tryptic enzyme activity in sea bass (Dicentrarchus labrax) larvae is influenced by different feeding regimes and trigger substances Submitted to Aquaculture 65 General discussion and outlook CCK-trypsin-axis 87 Implications for microdiets and feeding regimes 90 Outlook 92 References 97 List of tables 110 List of figures 111 Description of author contributions 113 Danksagung 115 Curriculum vitae 117 Eidesstattliche Erklärung 121 1

6 2

7 Summary Summary The reliable production of high quality offspring is a main bottleneck in marine aquaculture. One reason is the lack of knowledge of fundamental processes in digestive physiology during the larval stages, preventing the implementation of effective feeds and feeding regimes. Low digestive capacity during critical periods and an immature endocrine control of digestive enzyme synthesis and release might contribute to the overall high mortality during the early larval stages of marine species as well as the inability to efficiently utilize microdiets from first-feeding. In this study, I investigated the regulatory loop between the gut hormone cholecystokinin (CCK) and the main proteolytic enzyme trypsin in marine fish larvae. Another focus was on the possibility to enhance digestive capacity in the early larval stages by incorporating a trigger substance for the CCK-trypsin-axis into a commercial microdiet. The ontogenetic development of CCK and tryptic enzyme activity was described in Atlantic cod larvae (Gadus morhua) in chapter I. Increasing levels of CCK may indicate its maturing role in controlling pancreatic enzyme secretion during ontogeny but when this endocrine control becomes functional could not be explained with the results in this study. I examined the dynamic feedback mechanism between CCK and tryptic enzyme activity during a 12 hours feeding experiment which revealed a negative feedback control between both factors. Low tryptic enzyme activity at the end of this feeding period also suggests a limited digestive capacity in cod larvae to handle several meals per day. I further investigated the exact nature of the regulatory loop between CCK and tryptic enzyme activity in a tube-feeding study in chapter II. In this experiment single cod larvae were tube-fed different solutions containing purified standards, inhibitors or antagonists to evaluate the proposed regulatory loop from different sides. The obtained results further support the hypothesis of a negative feedback control, although the administration of a CCK antagonist and a trypsin inhibitor did not reveal conclusive results. However, untreated larvae showed a highly fluctuating pattern both in CCK and tryptic enzyme activity possibly caused by an underlying daily rhythm. On the other hand, the administration of the plant protein phytohemagglutinin (PHA) triggered a response in tryptic enzyme activity similar to the administration of CCK, indicating a stimulatory effect of PHA on the proteolytic enzyme capacity of cod larvae. 3

8 Summary The influence of different dietary treatments on tryptic enzyme activity during ontogeny in sea bass larvae (Dicentrarchus labrax) revealed a higher activity in larvae fed a microdiet compared to larvae fed Artemia in chapter III, possibly caused by a higher content in complex protein in the microdiet. The microdiet was additionally supplemented with PHA as a stimulatory trigger substance in two concentrations which resulted in reduced growth and tryptic enzyme activity, probably due to an overdose. A one-day observation revealed a similar trend in CCK independent of the diet or food deprivation which indicates a natural endogenous rhythm and an immature meal-responsiveness of CCK. Moreover, a reduction in tryptic enzyme activity during the day may point towards a limit in digestive capacity to deal with a certain number of meals. Remarkably, sea bass larvae seem to be able to synchronize tryptic enzyme activity with the feeding schedule as found in adult fish. 4

9 Zusammenfassung Zusammenfassung Die zuverlässige und kostengünstige Produktion von qualitativ hochwertigen Setzlingen ist einer der wesentlichen Limitationen in der marinen Aquakultur. Die Ursachen hierfür sind in dem fehlenden Wissen über grundlegende Prozesse der Verdauungsphysiologie während des Larvalstadiums zu suchen, was die Formulierung von optimiertem Futter und angepasster Fütterung erschwert. Eine gering entwickelte Verdauungskapazität während der kritischen Phase nach Beginn der Futteraufnahme kann dabei als mitentscheidender Grund für die allgemein hohe Sterblichkeit in den frühen Lebensstadien von marinen Fischarten angesehen werden. Dies dient zudem als Anhaltspunkt für die geringe Verwertung von mikropartikulärem Futter unmittelbar nach Fressbeginn. Im Fokus dieser Arbeit stand daher der Regulationsmechanismus zwischen dem Peptidhormon Cholecystokinin (CCK) und dem proteolytischen Verdauungsenzym Trypsin, welcher als Hauptbaustein in den Verdauungsprozessen mariner Fischlarven angesehen wird. Ein weiteres Ziel war die Erprobung eines Zusatzstoffs für mikropartikuläre Futtermittel, welcher die CCK-Trypsin-Achse anregt und damit die Verdauungskapazität in der kritischen Phase mariner Fischlarven erhöht. Die ontogenetische Entwicklung von CCK und Trypsin wurde am Beispiel für Dorschlarven (Gadus morhua) in Kapitel I beschrieben. Ansteigende CCK-Konzentrationen deuten dabei auf eine zunehmende, regulatorische Rolle von CCK in der Enzymsekretion hin, wobei jedoch der genaue, zeitliche Beginn dieser Regulation während der Ontogenese aus den Ergebnissen nicht abzuleiten war. Weiterhin konnte ich einen negativen Rückkopplungsmechanismus zwischen CCK und Trypsin in einem zwölfstündigen Fütterungsexperiment aufzeigen. Geringe Trypsinaktivitäten als Antwort auf mehrere Fütterungen wurden als Limitierung bei der Verdauungskapazität bei Dorschlarven im Tagesverlauf interpretiert. Der exakte Zusammenhang zwischen CCK und Trypsin wurde mit Hilfe von puren Standards, Inhibitoren und Antagonisten in Kapitel II untersucht, welche direkt in den Darm einzelner Larven appliziert wurden. Hierbei konnte die Hypothese einer negativen Rückkopplung bestätigt werden, jedoch brachten die Applikation eines CCK-Antagonisten und eines Trypsin-Inhibitors keine eindeutigen Resultate. Starke Schwankungen beider Faktoren in der unbehandelten Kontrollgruppe deuten zudem auf natürliche Schwankungen im Laufe eines Tages hin. Die Applikation des 5

10 Zusammenfassung Pflanzenproteins Phytohämagglutinin (PHA) stimulierte die Trypsinaktivität ähnlich wie verabreichtes, reines CCK. Der Einfluss von Trocken- oder Lebendfutter (Artemiennauplien) auf die Trypsinaktivität von Wolfsbarschlarven (Dicentrarchus labrax) in Kapitel III zeigte eine erhöhte Aktivität in der mit mikropartikulärem Futter gefütterten Gruppe, wahrscheinlich durch die Anpassung an einen höheren Proteingehalt dieses Futters. Die Zugabe von PHA als Stimulanz für die CCK-Trypsin-Achse zum kommerziellen Trockenfutter resultierte in einer reduzierten Trypsinaktivität, wahrscheinlich begründet durch negative Auswirkungen einer Überdosis PHA. Ich konnte einen ähnlichen Tagesrhythmus für CCK in allen Futter- und auch Hungergruppen feststellen, was auf eine noch nicht funktionsfähige, endokrine Kontrolle von CCK nach Futteraufnahme hindeutet. Aus dem Rückgang der Trypsinaktivität im Laufe des Tages kann auch hier auf eine begrenzte Verdauungskapazität geschlossen werden, die für eine begrenzte Fütterung am Nachmittag spricht. Zudem konnte eine Synchronisation der Trypsinaktivität mit den Fütterungszeiten in nicht gefütterten Larven festgestellt werden. 6

11 Introduction Introduction Fisheries and aquaculture Fish and fishery products represent a valuable source of nutrients of fundamental importance for diversified and healthy diets (FAO, 2012). This is specifically true for products from marine fish; two to three meals of fish per week are considered to reduce the risk of e.g. heart attack, stroke or Alzheimer disease by 50 to 60% (FAO/WHO, 2011). Consequently, the demand for fish has risen continuously during the last decades, reaching a mean world-wide peak in fish consumption of 18,6 kg per capita in 2010 (Germany: 15,7 kg). Concurrently, the growth of world fish supply has even outpaced the increase in the world s population during the last decades. However, population and economic growth, combined with technical improvements and overcapacity of fishing fleets, has resulted in an unsustainable fishing of roughly 90% of European fish stocks (FROESE et al., 2010). Despite scientific warnings, the predictive outcome has been the reduction in fish landings from eight to five million metric tons between 1995 and 2010 in the European Union. On a global scale, one third of marine fish stocks remain overexploited and the total fish catch has only minimal potential to increase in the future. Sustainable aquaculture has the potential to close the increasing gap between demand and supply in fish for human consumption. In this respect, aquaculture may not be regarded as a substitute for capture fisheries but as an additional contributor to world fish supply. Despite being virtually non-existent on a commercial scale a few decades ago, aquaculture now contributes 40% to the global fish production and represents the fastest growing animal food sector with annual growth rates of 9% (FAO, 2012). This growth is, however, mainly driven by increasing production in Asian countries, with China having the biggest share of 60% in world aquaculture production (crustaceans, molluscs and other aquatic animals included). By contrast, the contribution of the European sector accounts for only 4% of the global production with Norway holding 40% contribution as the main European producer. As a consequence, the EU needs to import 60% of its consumed fish to meet the increasing demand (Germany: 88%). Although more than 300 fish species are produced under captive conditions worldwide, the production is largely dominated by cyprinids in Asian freshwater environments (Fig. Intro-I). By contrast, the increase in production in industrialized 7

12 Introduction countries is mainly dominated by a few species, according to the preference of consumers for higher trophic level fish. Figure Intro-I: World aquaculture production composition of finfish by culture environment and major species (FAO, 2012). Among the difficulties in rearing these species, the availability of broodstock for controlled and independent reproduction and the reliable production of high quality offspring is paramount for successful aquaculture. This is true for both a shift from quantity to quality in established species as well as for closing the lifecycle of new candidate species. Larviculture and larval nutrition The desired increase in production of high-value fish is hampered by the fact, that the supply of high quality larvae and juveniles in stable quantities is still a bottleneck in the commercialization of many marine fish species for aquaculture. At present, survival of most marine fish species in commercial hatcheries might be as low as 30% one month after hatching, even under state-of-the-art conditions. Mortality rates of 100% can occur if abiotic and biotic conditions are not within tightly framed optima, leaving some researchers to conclude that fish larvae have a wish to die (DR. H. BROWMAN, pers. comm.), which is a consequence of the reproductive strategy of many teleost fish species. However, the most favourable conditions have yet to be defined for many species, given the extreme diversity of ontogenetic responses, especially during the larval stages (PITTMAN et al., 2013) and advances in zootechnical parameters have been following a trial-and-error approach so far. Even in 8

13 Introduction established species, such as European sea bass, Dicentrarchus labrax, large differences in abiotic conditions can be found among rearing protocols from commercial hatcheries with a trade-off between growth, mortality, malformations and economic considerations. For example, although there is increasing knowledge of the importance of light during larval development (VILLAMIZAR et al., 2011), hatcheries apply constant light vs. light/dark cycles and high vs. low light intensity. Nevertheless, the plasticity of rearing conditions varies largely between species and larvae of more sensitive species, like Atlantic halibut, Hippoglossus hippoglossus, require much more effort and specialized conditions to obtain a viable batch of larvae (HARBOE et al., 1998). Apart from the challenge to establish optimized abiotic conditions, the successful rearing of marine fish larvae heavily relies on feeding live feed for at least a few days after first-feeding (Fig. Intro-II, CONCEIÇÃO et al., 2010b). The live feed most commonly used in commercial hatcheries, rotifers, Brachionus spp., and brine shrimp, Artemia spp., were discovered a few decades ago as suitable prey organisms and are used around the world due to the existence of standardized and costeffective protocols (LAVENS et al., 1996). In fact, the discovery and extensive use of both organisms may have been the main driving force behind the tremendous growth in aquaculture production so far, although they do not represent the natural food of marine fish larvae. Consequently, their biology and use have extensively been studied and reviewed (CONCEIÇÃO et al., 2010b; LAVENS et al., 1996; ØIE et al., 2011; OLSEN, 2004; OLSEN et al., 2004; STØTTRUP et al., 2003). Figure Intro-II: Typical feeding schedules for Atlantic cod (Gadus morhua) and Altantic halibut (Hippoglossus hippoglossus) larvae. Bars indicate the period of feeding the respective organisms and faded ends a transition from one feed to another. Modified after OLSEN et al. (2004). 9

14 Introduction Rotifers of various species and strains can be cultured in high densities under a wide range of culture conditions, sustain high population growth and ingest particles by filtration, including microalgae, yeast and bacteria. Thus, rotifers represent a renewable resource and the production on-site reduces the dependencies on external suppliers. Culture systems for rotifers can be in the range of continuous to batch cultures, where rotifers are harvested daily or the whole production is harvested at the end of the exponential growth phase of the population, respectively. The application of recirculation technology and the consequent removal of ammonia and other waste products may enable culture densities of up to rotifers ml -1. The brine shrimp Artemia are able to form dormant cysts which are harvested from natural sources and can be easily transported, stored and kept viable for several years. When rehydrated, Artemia hatch as nauplii within 24 hours under favourable conditions, which makes them a convenient off-the-shelf product. On the other hand, the increasing demand and the dependency on unpredictable natural harvests have resulted in sky-rocketing prices. Newly hatched nauplii ( instar I ) rely on their endogenous nutrient sources and non-selectively filtrate particles of appropriate size upon the transition to the next naupliar stage ( instar II ) after about 8 hours. The movement patterns of rotifers and Artemia make them constantly available to fish larvae and consequently both organisms are easily detected and ingested, since fish larvae are believed to be mainly visual feeders (YÚFERA, 2011). From a functional perspective, the application of rotifers and Artemia as feed organisms is determined by the gape size of the larvae, since fish larvae are able to ingest and swallow particles between 30 and 60% the size of their mouth gape (YÚFERA, 2011). Depending on the species, the larvae may start feeding on rotifers ( µm body length, depending on species and strain) being replaced by newly hatched Artemia nauplii ( µm body length, depending on strain) and enriched (see below) instar II nauplii ( µm). Fish larvae with a bigger gape size, such as European sea bass, may directly start feeding on small Artemia strains. A third class of live feed in larval fish cultures are calanoid, harpacticoid or cyclopoid copepods, in decreasing order of importance. Since they represent the natural food of marine fish larvae, it is no surprise that larvae fed on copepods show considerably higher survival, growth and larval quality than larvae fed rotifers and Artemia (CONCEIÇÃO et al., 2010b). Copepods provide a wide range of prey sizes ( µm, depending on developmental stage and species), given their development 10

15 Introduction through different naupliar and copepodite stages which allows fish larvae to feed on a variety of appropriate prey sizes. Consequently, they are also well suited as prey for extremely small larvae with a small mouth gape. They can be cultured extensively in outdoor ponds or enclosed lagoons and harvested with plankton nets. Alternatively, fish larvae ready for first-feeding might be stocked in a plankton-rich water body to feed on copepods. The intensive cultivation has been proven difficult as fecundity is reduced at higher densities (e.g. >100 ind. L -1 for Acartia tonsa) and by low quality in dietary lipids (depending on the algae species provided as feed). The latter point seems to be less important in harpacticoid species since they are able to elongate fatty acids and are thus not dependent on fresh algae. From a nutritional perspective, copepods are discussed to be well suited to cover the dietary requirements of marine fish larvae (ØIE et al., 2011). In fact, the composition of copepods is explored to serve as a benchmark for the improvement of enrichments for rotifers and Artemia as well as for formulated feeds (VAN DER MEEREN et al., 2008). However, the cost-effective production remains a challenge on a commercial scale given the high volume of prey necessary for rearing fish larvae. Considering the limited availability, copepods may be fed to larvae only for a few days to improve survival and first-feeding success. Microalgae are a widely used fraction of live feeds which covers 16 genera of commercial importance. They are commonly produced on-site as batch cultures in plastic bags or as continuous cultures in sophisticated photo-bioreactors in ascending order of technological complexity. Their nutritional composition largely depends on the species and culture conditions, but many species became popular for their high content in highly unsaturated fatty acids, most importantly DHA (docosahexaenoic acid, 22:6ω3) and EPA (eicosapentaenoic acid, 20:5ω3). Microalgae may serve as a feed component for plankton cultures (rotifers and copepods) or as nutritional enrichment of rotifers and Artemia. Thereby mixes of different algae species may compensate one another for deficiencies in given nutrients (CONCEIÇÃO et al., 2010b). Some species may also be used directly in the larval rearing tanks according to the green water technique, a strategy that is established as state-of-the-art. In the first days after first-feeding the application of algae has been proven beneficial for fish larvae in terms of growth and survival (ØIE et al., 2011). They are passively taken up by the larvae (VAN DER MEEREN, 1991) and might serve as an early food source of essential nutrients (ØIE et al., 2011; REITAN et al., 1993) or might aid in the onset of the digestive processes (see chapter I). Largely underestimated in the past, microalgae are also thought to serve as microbiological 11

16 Introduction stabilizers both in the rearing water and the larval intestine preventing the bloom and harmful effects of opportunistic bacteria (OLSEN et al., 2004; VADSTEIN et al., 2004). In the larval tanks, the presence of microalgae also stabilizes the nutritional value of rotifers and Artemia, since both organisms continuously ingest the algal cells. A more subtle effect is the increase of visual contrast for the larvae which improves feeding ability and ingestion of prey (ROCHA et al., 2008). It has been found relatively early that rotifers and Artemia might be nutritional inadequate but that their nutritional value can be manipulated (KITAJIMA et al., 1979). Consequently, nutrition of the early life stages has received most attention in larval fish research in order to combat poor growth and low survival as well as the occurrence of deformities and abnormalities (PITTMAN et al., 2013). In order to improve their nutritional value, rotifers and Artemia (instar II nauplii) are enriched using either a mixture of different microalgae or, more commonly, highly concentrated commercial emulsions and formulated particles. Thereby, droplets or small particles are taken up and their nutrients are partly incorporated into body tissue or stored in the gut of rotifers and Artemia. The unselective filtration by rotifers and Artemia and short-term storage in the gut makes them carriers of virtually everything of appropriate size, e.g. also for pro- or antibiotics (TOURAKI et al., 2010; ZACARIAS-SOTO et al., 2011). It must be pointed out, however, that this shortterm bioloading is limited by the gut volume (max. 15% of body volume in rotifers, ROMERO-ROMERO et al., 2012) and the desirable long-term enrichment is restricted by the organisms own metabolism (see below). The nutritional requirements for several nutrients have been investigated in many studies (reviewed by CONCEIÇÃO et al., 2011; HAMRE et al., 2013; IZQUIERDO et al., 2011; KJØRSVIK et al., 2011; MOREN et al., 2011; OLSEN et al., 2004). One main obstacle is the difficulty of obtaining quantitative requirements of different nutrients for fish larvae. To date, quantitative requirements are often extrapolated from juvenile or adult fish and qualitative requirements can at best be narrowed down to optimum ranges (HAMRE et al., 2013). Nevertheless, the requirements may differ greatly from those of juvenile and adult fish as growing larvae experience dramatic changes in morphology and physiology, including metamorphosis in flatfish (HAMRE et al., 2013). The nature of this problem lies in the dependency of most marine fish larvae on live feed as food in the early stages and the selective catabolism of different nutrients has largely prevented the application of reliable dose-response studies as in juvenile and adult fish (CONCEIÇÃO et al., 2010b). 12

17 Introduction Due to the limited ability to synthesize fatty acids (FA) from shorter-chain precursors, much emphasis has been put on the requirements of highly unsaturated fatty acids for marine fish larvae, most importantly DHA, EPA and ARA (arachidonic acid, 20:4 ω6) (IZQUIERDO et al., 2011). These FA not only serve as metabolic energy, but also as structural components of phospholipids in cell membranes and as precursors for physiologically active substances (e.g. eicosanoids). Specifically DHA increases membrane fluidity and is also found in the retina of the eyes as well as in neural tissue and brain. In addition, eicosanoids are important for hormone release as well as for cardiovascular and immune function. Due to disproportionally big eyes and brain during the larval stages, it is not surprising that many studies found highly unsaturated FA to promote growth and survival (e.g. KOVEN et al., 1992). Since these FA may physiologically compete in the body, the relative proportion is highly relevant with optimal ratios of 2:1 for DHA/EPA and 3.5 5:1 for EPA/ARA. For instance, an optimal EPA/ARA ratio will lead to a normal pigmentation pattern in early flatfish (KJØRSVIK et al., 2011). While it is relatively easy to control the FA profile in rotifers, since it mirrors that of the diet/enrichment, this is more challenging in Artemia as they selectively catabolise DHA to EPA. Therefore, enrichments with a very high DHA/EPA ratio try to compensate for this imbalance which also implies the necessity to feed newly enriched Artemia as they gradually decrease in nutritional value in terms of FA after enrichment. The issue of highly unsaturated FA continues to be a problem for relatively large larvae, e.g. Atlantic halibut, which need newly hatched Artemia nauplii as first food which contain virtually no DHA and cannot be enriched (OLSEN et al., 2004). Another important factor is the distribution of FA among the lipid classes, since FA presented as phospholipids have been proven largely beneficial in the diet for marine fish larvae (KOVEN et al., 1993). It has to be noted that copepods contain high levels of phospholipids rich in highly unsaturated FA, whereas the lipid fraction in rotifers and Artemia largely consists of neutral lipids. Nevertheless, the enrichment procedure may lead to excess lipid contents and high levels of highly unsaturated FA transferred to the larvae increases the formation of free radicals due to oxidation and the risk of oxidative stress. In the case of amino acids (AA) differences exist in terms of essentiality for the organism. Essential AA, such as methionine, must be provided by the diet, whereas nonessential AA can be synthesized, preferably from glutamate as a precursor or amino group donor. Comparing the AA profile of rotifers, Artemia and whole-body fish larvae have historically served as a starting point to study AA in larval fish which provide rough estimates of their requirements (CONCEIÇÃO et al., 2011). It has been 13

18 Introduction concluded that both rotifers and Artemia show imbalances in their AA profile with histidine being the first limiting one. While the amino acids profile of rotifers is genetically hardwired, only the total protein content can be modified, whereas the AA profile of Artemia is closer to the presumed requirements of fish larvae. Recently, the positive effect of AA that don t structure proteins, such as the cysteine-derivate taurine, on growth, survival and metamorphosis has started to be explored (PINTO et al., 2010b). However, the functions of many of these AA remain largely unknown. It has been concluded that the content of vitamins or their precursors in live feed is adequate after enrichment to meet larval fish demands (MOREN et al., 2011; OLSEN et al., 2004) although the specific requirements are largely unknown for most vitamins (HAMRE et al., 2013). However, rotifers may be deficient in the lipid-soluble vitamins A and E as well as in most of the water-soluble forms. Since fish larvae have only limited storage capacity and water-soluble vitamins are poorly stored in organisms in general, they might be more vulnerable to deficiencies during the early stages (MOREN et al., 2011). On the other hand, hypervitaminosis, especially of the fatsoluble fraction (vitamins A, D, E & K), will induce malformations and deformities (e.g. FERNÁNDEZ et al., 2011). Most research on the influence of vitamins on larval fish has focused on vitamins C and E due to their unstable nature and functioning as antioxidants. Excess contents of both vitamins have been found to increase immunity and stress resistance (HAMRE, 2011), where vitamin E might in fact be needed as an anti-oxidant to alleviate oxidative stress caused by high levels of DHA (BETANCOR et al., 2011). The importance of minerals and trace elements for fish larvae has only recently started to be investigated which might be explained by the interference of sources from diets and the surrounding water (MOREN et al., 2011). Rotifers contain relatively low levels of minerals compared to copepods, but enrichment with selenium and iodine led to higher survival e.g. in cod larvae (HAMRE et al., 2008). While selenium has been found to play an important role in the formation of anti-oxidant enzymes (PENGLASE et al., 2010), sufficient levels of iodine might be essential for successful metamorphosis in flatfish (POWER et al., 2001). Iodine may also have bactericide properties and may reduce the bacterial load during the enrichment of live feed leading to lower transfer to and higher survival of the fish larvae (HAMRE et al., 2013). However, care must be taken in recirculation facilities using ozone and protein skimmers, since it dramatically reduces the bioavailability of iodine leading to deficiency syndromes such as goitre (RIBEIRO et al., 2012). The same effect might be 14

19 Introduction induced by high nitrate levels in the water by blocking the uptake mechanisms for iodine (MORRIS et al., 2011). The levels of phosphorus are not limiting in all live feeds and the composition of trace elements like zinc, copper, manganese and iron is relatively easy to control in rotifers. Nevertheless, high levels of these elements in copepods as target levels may point towards a higher demand in fish larvae compared to adults due to the lower retention efficiency (MOREN et al., 2011). In summary, Artemia nauplii can be considered more stable and predictive in their nutritional composition compared to rotifers (OLSEN et al., 2004). Nevertheless, it appears obvious that the use and dependency on live feed for marine fish larvae has several drawbacks. All organisms require large effort in terms of labour and facilities which may account for at least 50% of all hatchery costs (STØTTRUP et al., 2003). The production of microalgae and rotifers may result in culture crashes which also imposes the need to keep stock cultures as backups. However, grown microalgae onsite may be substituted with concentrated, frozen or freeze-dried off-the-shelf products to date. While Artemia can be used as instant feed, their quality is highly dependent on the natural harvests and they may also be vectors of pathogens such as Vibrio spp. (RITAR et al., 2004). In addition, the enrichment procedures of both rotifers and Artemia result in a high organic load in the incubation water and the proliferation of opportunistic bacteria which may establish a harmful microflora in the larval tanks (VADSTEIN et al., 2013). A major constraint is the limited variety in prey sizes when using rotifers and Artemia which results in suboptimal prey as the larvae grow. This is particularly important for very small larvae needing small prey at first-feeding, but also for the transition from rotifers to Artemia. Above all, the nutritional composition of rotifers and Artemia is highly variable and imbalanced with only limited opportunities for manipulation. Considering the substantial downsides of live feeds it is no surprise that the substitution of live feed with formulated microdiets has been of major interest for several decades (e.g. BARAHONA-FERNANDES et al., 1976) and still is one of the major challenges in marine larval fish production (LANGDON et al., 2011; ZAMBONINO INFANTE et al., 2007). Ideally, these microdiets would be steadily available, of known and uniform nutritional composition as well as of appropriate sizes to cover the requirements of marine fish larvae during ontogeny from first-feeding. Consequently, this would supersede the infrastructure and labour required for cultivating live feeds, opening large room for cost reduction. In fact, the large success of salmon, Salmo salar, aquaculture may be explained by the possibility to feed the offspring with microparticulate diets immediately after hatching. 15

20 Introduction Unfortunately, the starter feeds used for salmon and trout are not applicable to larvae of marine species for reasons of different nutritional requirements (see above) and technical constraints (see below). To date, technological progress allows the production of different types of microdiets as small as 25 µm with different properties (reviewed by LANGDON et al., 2011). Hand in hand with the progress in knowledge of nutritional requirements this has resulted in increasing success in replacing live feed evidenced in a wealth of studies for several species (e.g. CAHU et al., 2001; CURNOW et al., 2006a; KOVEN et al., 2001; LAZO et al., 2000b; YÚFERA et al., 2005). Nevertheless, the larvae of only very few marine species can be reared exclusively on microdiets, such as European sea bass (CAHU et al., 1998a). At present, feeding microdiets as the only feed is typically accompanied with low growth and survival. Consequently, many studies have focused on shortening the live feed period after first-feeding with a subsequent transition to microdiets ( weaning ) without compromising growth and survival compared to a live feed control treatment (e.g. CALLAN et al., 2003; CURNOW et al., 2006a; FLETCHER et al., 2007; NGUYEN et al., 2011). Another strategy is to provide microdiets along with live feed from first-feeding ( co-feeding ) to condition the larvae to the particles which results in enhanced performance at weaning (ENGROLA et al., 2009; MAI et al., 2009; NHU et al., 2010; ROSENLUND et al., 1997). The application of microdiets is, however, limited by several constraints and the most obvious limitations can be termed as ingestion. Since fish larvae have only limited visual potential and locomotion they rely on high encounter rates with suitable prey for successful capture. While rotifers and Artemia naturally stay and disperse in the water column, microdiet particles are characterized by sinking rates as high as 1 cm s -1 which makes them rapidly unavailable as prey (LANGDON et al., 2011). As a consequence, microdiets need to be provided in excess in highly frequent doses, probably every 10 to 30 minutes, which also implies higher cleaning efforts of the tank bottom to remove feed wastes and higher water turnover to prevent fouling processes. From a functional perspective, microdiet particles should ideally move fast enough in the water current to draw attention but sink slow enough for successful capture (HOLT et al., 2011). It is suggested that visual stimulation by moving live prey might not represent the only factor for ingestion, whereas chemical stimuli through the release of free amino acids from live prey have been found to play a vital role (KOLKOVSKI et al., 1997a; KOLKOVSKI et al., 1997c; KOVEN et al., 2001). However, other physical properties such as texture, shape and colour might contribute to the attractiveness of microdiet particles. 16

21 Introduction The second complex of limitations of microdiets can be summarized as nutrients and digestibility. The high surface-to-volume ratio of microdiets makes the particles prone to nutrient loss in the water. This leaching problem is particularly important for low-molecular-weight and water-soluble nutrients such as beneficial free amino acids and vitamins which may result in a substantial loss of this fraction within seconds to minutes after the immersion in water (KVÅLE et al., 2006; NORDGREEN et al., 2008). Consequently, these important ingredients might have been lost until the larvae will capture the particles, whereas leaching of specific ingredients would be desirable if they stimulate ingestion (LANGDON et al., 2011; YÚFERA et al., 2002). Another problem might be a lower bioavailability of minerals and trace elements, since they are often incorporated in microdiets in an inorganic form (MOREN et al., 2011). Encapsulation techniques and the use of specific binders in microdiets result in a higher retention of water-soluble components at the expense of lower digestibility in the immature gut of marine fish larvae (see below). In addition, although a microdiet particle may provide energy and nutrients several times higher than live feed organisms of the same size (ROSENLUND et al., 1997), the low water content in microdiets and their compact nature may contribute to the low digestibility and might even cause osmotic stress in the gut of larval fish (HOLT et al., 2011). The development of complex particles incorporating encapsulated microingredients may circumvent the problem of delivering bulk nutrients (proteins, lipids) to the larvae without losing nutrients susceptible to leaching (LANGDON et al., 2007). To date, however, establishing an equilibrium between nutrient digestibility and retention remains one of the major challenges in microdiets technology (HAMRE et al., 2013; LANGDON et al., 2011). Larval feeding and digestion The transition from endogenous nutrient reserves deriving from the depleting yolk sac to exogenous feeding and nutrient acquisition is the most critical period in developing fish larvae (KJØRSVIK et al., 2004; YÚFERA et al., 2007). Under controlled conditions, feed should be offered before the yolk sac is depleted to avoid starvation and subsequent mortality as well as to commence feeding behaviour (JAROSZEWSKA et al., 2011). However, the ability to capture prey is limited by an immature visual capacity and locomotion which implies the importance to offer prey of appropriate size and in appropriate densities (YÚFERA, 2011; YÚFERA et al., 2007). Once firstfeeding is initiated, fish larvae increase their attack and feeding success as they 17

22 Introduction develop and may turn into feeding machines revealing daily ingestion rates above their own weight (PARRA et al., 2000), e.g. up to several hundred rotifers per day in larval turbot, Scophthalmus maximus (REITAN et al., 1993). Therefore, marine fish larvae may show very high growth rates if natural zooplankton is offered as food, e.g. 20% daily growth in larval cod, Gadus morhua (VAN DER MEEREN et al., 1993). The potential for very fast growth and the consequent high nutrient requirements are contradicted by the very underdeveloped state of most marine fish larvae at first-feeding (CONCEIÇÃO et al., 2011). This is particularly true for the digestive system of altricial fish larvae which don t possess a stomach at first-feeding until the metamorphosis into a juvenile is complete. In fact, the acquisition of a fully developed and functional digestive system can be considered as a main feature of the end of the transition from larva to juvenile in fish (KJØRSVIK et al., 2004; LAZO et al., 2011). Upon the uptake of food, the ability to enzymatically breakdown large molecules into more simple, absorbable compounds is the next logical prerequisite for growth. Consequently, the ontogenetic development of the digestive system in several marine fish species has been the subject of numerous studies using a variety of methods (reviewed by KJØRSVIK et al., 2004; LAZO et al., 2011; RØNNESTAD et al., 2008; ZAMBONINO INFANTE et al., 2001). Although there are temporal differences in the developmental sequence, the general pattern in altricial larvae is conserved among species. The gut, as the main site of digestion and absorption, appears as an undifferentiated straight tube at first-feeding with no stomach but with developed accessory organs, like the liver, gallbladder and pancreas. During development, the folding and differentiation of the gut epithelium increases, thereby enhancing the digestive and absorptive capacity. Small polymers such as dietary oligopeptides or amino acids are absorbed and digested intracellularly in the epithal cells, called enterocytes, and by a variety of amino acid transporters and one peptide transporter (CONCEIÇÃO et al., 2011). As the gut develops, a shift from intracellular to extracellular digestion in the brush border membrane marks the maturation of the enterocytes and the gut epithelium (LAZO et al., 2011; ZAMBONINO INFANTE et al., 2001). In order to facilitate optimal conditions for digestion in the gut, a slightly alkaline environment is established by biliary secretions from the liver which have been stored in the gallbladder. The main digestive enzymes in altricial fish larvae are synthesized in the exocrine part of the pancreas just before the onset of first-feeding and their activity appears to be relatively low. These enzymes, like trypsin, lipases and amylase, are responsible for the luminal breakdown of macromolecules in the gut lumen which are then further digested by enzymes in the brush-border region of 18

23 Introduction the gut epithelium into absorbable monomers (see above, KJØRSVIK et al., 2004; ZAMBONINO INFANTE et al., 2001). Since growth in fish larvae means mainly protein deposition in the body (CONCEIÇÃO et al., 2011; KJØRSVIK et al., 2004), the effective hydrolysis and digestion of food protein is of major importance to acquire the building blocks for growth, namely amino acids. Since marine fish larvae lack a stomach with a functional pre-hydrolysis of protein by the enzyme pepsin in an acidic environment, they rely on proteases produced and released from acinar cells in the pancreas. This is specifically important for dietary proteins since they are digested at a relatively slow rate by proteases in an alkaline environment, whereas this is much more efficient if proteins are pre-conditioned by pepsin digestion in the stomach (KJØRSVIK et al., 2004). Among the pancreatic enzymes, trypsin is the most significant protease in the early larval stages and is considered to be a key enzyme in the digestive process (LIDDLE, 2006; UEBERSCHÄR, 1995; ZAMBONINO INFANTE et al., 2001). Following the ingestion of feed, trypsin is secreted as its inactive precursor trypsinogen from the acinar cells in the pancreas into the gut lumen and either auto-activated or activated by the enzyme enterokinase in the brush border membrane. Trypsin is an endopeptidase and specifically cleaves peptide bonds between the amino acids arginine and lysine. Due to its metabolic importance, sensitivity and the fast reaction time towards feeding and starvation, the activity level of trypsin is a well suited biochemical indicator for the nutritional condition of fish larvae (LAZO et al., 2011; UEBERSCHÄR, 1995; ZAMBONINO INFANTE et al., 2001). In addition, the activity of trypsin has been demonstrated to be a function of ingestion, gut filling and nutrient composition (RØNNESTAD et al., 2008; UEBERSCHÄR, 1995; ZAMBONINO INFANTE et al., 2007). Controversy exists to what extent digestive capacity in early larval stages is sufficient to deal with the food under intensive rearing conditions (LAZO et al., 2011; ZAMBONINO INFANTE et al., 2001; 2007). Although the main digestive enzymes, including trypsin, are present before the onset of first-feeding and show an increasing trend during ontogeny (reviewed by KJØRSVIK et al., 2004), stagnating or decreasing activities have been frequently observed in several species (Fig. Intro-III, PEREZ-CASANOVA et al., 2006; RIBEIRO et al., 1999; SUZER et al., 2007b; UEBERSCHÄR, 1995). Consequently, these low enzyme levels and the subsequent low digestive capacity might contribute to the overall high mortality and represents a critical phase in the early larval stages of marine fish (KJØRSVIK et al., 2004). 19

24 Introduction Figure Intro-III: Ontogeny of tryptic enzyme activity in larval halibut (black circles) and sea bass (red circles). White circles indicate starved sea bass larvae sampled on consecutive days. Circles and bars depict mean ± S.D. (n = 5-10, own unpublished data). As a result, it has been observed in halibut larvae that ingested Artemia are excreted undigested or even alive (Dr. T. HARBOE, Institute of Marine Research Austevoll, Norway, pers. comm.). This might be present in situations when ingestion of Artemia is too high and gut transit time too fast, overwhelming the relatively low digestive capacity during certain developmental stages. In this respect, debates exist whether a continuous (ZAMBONINO INFANTE et al., 2007) or meal-based feeding regime (HARBOE et al., 2009) is most suitable in terms of digestive capacity. Therefore, the observation of tryptic enzyme activity during feeding periods is a suitable tool to optimize and adapt the feeding regimes to the digestive capacity. In addition, the natural feeding behaviour of larvae may differ during a day and endogenic, physiological rhythms might be present as in mammals (RØNNESTAD et al., 2013; YÚFERA, 2011). With regard to the different feeds, it has long been hypothesized that the contribution of exogenous enzymes from ingested live prey considerably supplements the larva s digestive enzymes (DABROWSKI et al., 1978). However, repeated evidence from several studies revealed that this contribution appears to be negligible (KJØRSVIK et al., 2004; LAZO et al., 2011; RØNNESTAD et al., 2013). 20

25 Introduction While a considerable knowledge of the ontogenetic development of the digestive system and enzymes has been gathered, very little is known about how and when the digestive processes are controlled, either by neural or hormonal factors (RØNNESTAD et al., 2013; WEBB et al., 2011). This remains to be a key area in larval fish research and increasing knowledge will not only help to understand fundamental processes in marine fish larvae, but will also translate into applied aspects such as feeding regimes and feed formulations (LAZO et al., 2011). For instance, gut transit time of ingested food has been shown to be an important aspect for digestive efficiency (RØNNESTAD et al., 2008; RØNNESTAD et al., 2013). Continuous food supply under intensive rearing conditions may result in a continuous ingestion of prey, rapidly reducing the residence time for digestion, the efficiency of nutrient absorption and increasing the potential nutrient loss in faeces (RØNNESTAD et al., 2013). In addition to the limited digestive capacity, this might be due to a lack of a satiation signal from the digestive system or energy stores to the brain, since fish larvae may never encounter continuous food supply in nature and may have evolved for high ingestion rates when food is abundant (RØNNESTAD et al., 2008; RØNNESTAD et al., 2013). With respect to the hormonal control of the digestive system, the gut hormone cholecystokinin (CCK) is known to play a key role in the delay of gastric emptying, contraction of the gallbladder and peristalsis in the intestine (EINARSSON et al., 1997; SILVER et al., 1991). In addition, it is considered as one of the most important regulators of pancreatic secretion and also acts as a satiation signal in the brain (LIDDLE, 2006; VOLKOFF et al., 2005). CCK has long been studied in mammals and growing evidence exists that similar mechanisms are present in adult and juvenile fish (EINARSSON et al., 1997; GÉLINEAU et al., 2001; JÖNSSON et al., 2006; KOFUJI et al., 2007; LÕHMUS et al., 2008; MURASHITA et al., 2008; NGUYEN et al., 2013; OLSSON et al., 1999; PETERSON et al., 2012; VALEN et al., 2011). Consequently, models derived from mammals and adult fish serve as a starting point to study the same mechanisms in the larval stages. Upon the presence of nutrients in the gut, CCK is released from enteroendocrine cells in the gut epithelium into the body fluids and acts on target cells in the pancreas to release its secretions, including trypsinogen, into the gut lumen (Fig. Intro-IV). High tryptic enzyme activity in the gut lumen then acts as a negative feedback control for the release of CCK, suggesting a regulatory loop between these two factors in humans and adult fish (LIDDLE, 2006; MURASHITA et al., 2008). Although the distribution of CCK-reactive cells along the digestive tract as well as the 21

26 Introduction ontogenetic development of CCK has been described in the larval stages of several species (CAHU et al., 2004; GARCÍA HERNÁNDEZ et al., 1994; HARTVIKSEN et al., 2009; KAMISAKA et al., 2013; KAMISAKA et al., 2001; KAMISAKA et al., 2005; KAMISAKA et al., 2003; KUROKAWA et al., 2000; MICALE et al., 2010; ROJAS-GARCÍA et al., 2011; WEBB et al., 2010), understanding of its regulatory role is still in its infancy and remains to be demonstrated in the early stages (RØNNESTAD et al., 2007; WEBB et al., 2011). In mammals, different food components such as certain amino acids have been found to specifically stimulate the release of CCK and consequently pancreatic enzyme secretion (LIDDLE, 1995) and growing evidence suggests that similar mechanisms are present in larval fish (CAHU et al., 2004; KOVEN et al., 2002; NAZ et al., 2009; ROJAS- GARCÍA et al., 2002). Figure Intro-IV: Physiological actions of CCK released from endocrine cells in the gut wall (red triangles) on the digestive system upon the presence of food in the gut (left side) as well as the regulatory loop between CCK and trypsin in the gut lumen (right side). See text for details. One reason for the difficulty why most fish larvae cannot utilize microdiets efficiently from first-feeding might be that pancreatic enzyme secretion is not sufficiently stimulated, either mechanically or biochemically, to deal with the ingredients in artificial food (KOLKOVSKI, 2001; KOVEN et al., 2001; WEBB et al., 2011; YÚFERA et al., 2000). Additionally, live prey such as Artemia is considered to be more digestible due to their higher content in water-soluble proteins and the release of small, stimulating monomers such as amino acids and peptides caused by autolysis in the larval gut (CONCEIÇÃO et al., 2011). By contrast, the modification of the composition of microdiets is still limited by several constraints (see above). Nevertheless, the substitution of complex protein in microdiets with a moderate level of hydrolyzed protein to increase the water-soluble protein fraction has been proven beneficial (KVÅLE et al., 2002; TONHEIM et al., 2005; ZAMBONINO INFANTE et al., 2007). Results of tracer studies also suggest that hydrolyzed protein and amino acids are absorbed and processed very fast in contrast to complex protein (APPLEBAUM et 22

27 Introduction al., 2004; CONCEIÇÃO et al., 2011; CONCEIÇÃO et al., 2007; RØNNESTAD et al., 2012; RØNNESTAD et al., 2007; TONHEIM et al., 2005), underlining the fact that processing of protein is limited by proteolytic rather than by absorptive capacity (RØNNESTAD et al., 2007). In order to overcome these physiological deficiencies in digestive capacity it has been of interest to identify factors and substances which elevate pancreatic enzyme secretion and digestive capacity to circumvent the associated problems resulting in high mortality and reduced growth. The most straightforward approach has been the supplementation of microdiets with digestive enzyme extracts which resulted in higher assimilation of nutrients from a microdiet in larval gilthead seabream, Sparus aurata (KOLKOVSKI et al., 1993) but was not conclusive in other species (KOLKOVSKI, 2001; KOLKOVSKI et al., 1997b). In another approach, the addition of spermine, a natural polyamine occurring also in Artemia, to a microdiet induced a faster maturation of the gut and a higher digestive enzyme activity in larval sea bass (PÉRES et al., 1997). This effect was confirmed by the incorporation of active yeast into the diet, producing polyamines in situ in the larval gut (TOVAR-RAMÍREZ et al., 2004). In this context, the plant lectin phytohemagglutinin (PHA) has been proven a suitable trigger substance to modulate pancreatic enzyme secretion. Lectins in general are carbohydrate binding proteins and members of the Leguminosae whose seeds are especially rich in these ubiquitous plant constituents (HERZIG et al., 1997). PHA is extracted from the red kidney bean (Phaseolus vulgaris) and is routinely used to initiate mitosis in cell cultures in vitro and exerts its effects also in vivo (PENDÁS et al., 1993). PHA was found to stimulate the growth of the pancreas, intestine as well as the release of CCK and pancreatic enzymes in rats and piglets in several studies (BARDÓCZ et al., 1990; EVILEVITCH et al., 2005; HERZIG et al., 1997; JORDINSON et al., 1997; KICIAK et al., 2010; KORDÁS et al., 2000; LINDEROTH et al., 2005; OTTE et al., 2001; RADBERG et al., 2001; THOMSSON et al., 2007). The mode of action is attributed to the binding capacity of PHA to gut epithelial cells which results in the stimulation of cell proliferation in the gut and pancreas. Regarding marine fish larvae, the low digestive capacity and especially tryptic enzyme activity could be connected to low levels of and low stimulation by CCK during the critical larval phases. Following the results from mammalian studies, PHA might also have a stimulatory effect on CCK and consequently on tryptic enzyme activity in the gut of larval fish to overcome the described ontogenetic deficiencies in digestive capacity in early marine fish larvae (DROSSOU et al., 2006). 23

28 Introduction Methodical considerations Several experimental and methodical approaches are limited by the tiny and fragile nature of marine fish larvae as well as by the general difficulties in rearing these animals (CONCEIÇÃO et al., 2010a). To evaluate the physiological response of a specific nutrient or, in this case, a trigger substance, accurate knowledge of how much and when a larva has taken up this substance is paramount, but remains an ambitious mission in fish larvae (HOLT et al., 2011). Therefore, the tube-feeding method developed by RUST et al. (1993) and modified by RØNNESTAD et al. (2000) was applied which allows the administration of small volumes of test solutions into the gut of anaesthetised individual larvae. Figure Intro-V: Set-up of the tube-feeding procedure including a dissecting microscope, micromanipulator and -injector. Single larvae are placed in a water drop on a Petri dish prior to the injection. In brief, anaesthetised individuals were placed in a water drop on a Petri dish under a dissecting microscope. The capillary with the microinjector was attached to the manipulator and used to deliver the test solution in one single injection (Fig. Intro-V) which is easy to monitor in younger larvae due to their almost transparent nature. Upon injection, single larvae were transferred to incubation chambers for recovery and sampled after pre-defined incubation periods. The analyses of CCK and tryptic enzyme activity followed a combined protocol which allows the determination of both factors in the same larva (ROJAS-GARCÍA et al., 2001; UEBERSCHÄR, 1995). CCK was analyzed using a competitive radioimmunoassay and 24

29 Introduction tryptic enzyme activity was assessed using a highly sensitive fluorescence molecule binding to a specific substrate. Thereby, the increase in fluorescence due to the release of the fluorescence molecule is proportional to the amount of trypsin in the sample. The assessment of tryptic enzyme activity in individual larvae represents a major advancement to evaluate the individual response and variability compared to the low sensitivity of chromatographic methods which rely on a large number of larvae in pooled samples but are still applied to date (e.g. GISBERT et al., 2009; NGUYEN et al., 2011; SUZER et al., 2013; TONG et al., 2012). Previous studies have shown that a substantial fraction of CCK is present in the central nervous system (ROJAS-GARCÍA et al., 2002; ROJAS-GARCÍA et al., 2011) whose role has yet to be elucidated (RØNNESTAD et al., 2007) and which may mask changes in CCK in the gastrointestinal tract. Consequently, the heads of all larvae have been dissected prior to analyses of CCK and tryptic enzyme activity. 25

30 Thesis outline Thesis outline One reason for the overall unstable and unpredictable production of offspring in most marine fish species is the lack of knowledge of digestive physiology in the larval stages which prevents the formulation of adapted feeds and feeding regimes. In this respect, considerable evidence points towards an important role of the gut hormone cholecystokinin and the proteolytic enzyme trypsin in the digestive processes of marine fish larvae. Chapter I deals with the ontogenetic development of CCK and tryptic enzyme activity in Atlantic cod larvae. I provide first insights into digestive capacity during ontogeny including the presumed growing role of CCK. A short-term feeding experiment over one day was conducted to evaluate the interaction between CCK and tryptic enzyme activity under practical feeding conditions. Therefore, experimental groups were fed a different number of meals (rotifers) and were sampled every full hour to follow the digestive response in order to assess the diurnal digestive capacity in larval cod and the short-term feedback mechanism between CCK and tryptic enzyme activity. In chapter II I examine the regulation and feedback mechanism between CCK and tryptic enzyme activity in a tube-feeding study using larval cod. Individual larvae were tube-fed a single dose of a CCK standard, CCK antagonist or trypsin inhibitor, respectively, in order to determine whether tryptic enzyme activity in the gut is mediated by CCK and whether trypsin acts as a negative feedback control for CCK release. Another group was injected with PHA as a presumed stimulator to induce CCK and pancreatic enzyme secretion. The larvae were sampled in a time series between 0.5 and 8 hours after injection to investigate a possible time-dependent response of CCK and tryptic enzyme activity in the regulatory loop. The influence of different dietary treatments on the CCK-trypsin-axis was investigated in chapter III as well as the application of PHA as a stimulatory feed additive. Therefore, sea bass larvae were fed a commercial microdiet from firstfeeding onwards enriched with two different concentrations of PHA and compared with a control group, fed with the same diet without PHA. A group fed Artemia served as a control for the long-term effect of live feed or microdiets on growth and tryptic enzyme activity. In addition, the diurnal rhythm as well as the feedback of CCK and tryptic enzyme activity depending on the dietary treatments was explored over one day under practical feeding conditions. 26

31 Chapter I CHAPTER I 3 days old cod larva with remaining yolk-sac ready to feed 27

32 Chapter I Hormonal control of tryptic enzyme activity in Atlantic cod larvae (Gadus morhua): involvement of cholecystokinin during ontogeny and diurnal rhythm Abstract The ontogenetic development of the gut hormone cholecystokinin (CCK) and the key proteolytic enzyme trypsin was described in Atlantic cod larvae (Gadus morhua) from first-feeding until 38 days post first-feeding (dpff). CCK is known to play a major role in the endocrine control of digestive processes in mammals and adult fish, but its regulatory role in the larval stages of marine fish is largely unknown. Only small amounts of CCK were found in the body (excluding head) in cod larvae at firstfeeding, but CCK levels increased exponentially with development, suggesting a more pronounced role of CCK during ontogeny. Tryptic enzyme activity increased slightly until a standard length of ca. 8 mm (approx. 33 days dpff) with a significant increase in larvae larger than 8 mm standard length, indicating limited digestive capacity in the early stages. To entangle the short-term feedback mechanism between CCK and tryptic enzyme activity, we conducted a 12 hour feeding experiment at 21 dpff. Cod larvae receiving only algae revealed a noticeable response in tryptic enzyme activity within two hours in the morning, whereas larvae fed algae and rotifers at the same time showed a slightly delayed response up to four hours. Tryptic enzyme activity remained low in the group receiving only algae as well as the two fed groups in the afternoon. No reaction in tryptic enzyme activity was observed in larvae that received a second meal of rotifers in the afternoon, indicating limited regulatory and digestive capacity to handle several meals in a short period. CCK levels remained relatively constant throughout the day but increased in the afternoon in all three groups when tryptic enzyme activity was low, suggesting that a negative feedback mechanism between CCK and tryptic enzyme activity is present in larval cod at least from 21 dpff. Keywords: Atlantic cod larvae, trypsin, CCK, digestion, ontogeny, endocrine control 28

33 Ontogeny and diurnal rhythm Introduction The unstable and unpredictive production of juveniles of many marine fish species for aquaculture still prevents commercialization of many candidate species. One reason is the lack of knowledge of the function and efficiency regarding digestive physiology in the early stages that hampers formulation of proper feeds and feeding regimes. There is a good understanding of how the tissues and organs of the digestive system develop during larval fish ontogeny. Accordingly, the developmental gene expression and secretion patterns of the digestive enzymes have been well described (reviewed by LAZO et al., 2011). However, research on the endocrine control of digestive functions in fish larvae is still in its infancy (WEBB et al., 2011) and also lags behind that of adult fish (e.g. MURASHITA et al., 2008). Consequently, models derived from mammals and adult fish in most cases serve as starting points to discover similar mechanisms in developing fish larvae. Until altricial fish larvae acquire an adult-like mode of digestion, characterized by a fully functional stomach including gastric glands and acidic digestion, they rely mainly on serine proteases with trypsin-like enzymes as the most significant proteolytic enzymes in the early larval stages. These alkaline proteases are synthesized in the pancreas and secreted into the gut, following the ingestion of feed. Among these enzymes, trypsin is considered to be a key enzyme in the digestive process (ZAMBONINO INFANTE et al., 2001). Trypsin is secreted as its inactive precursor trypsinogen from the acinar cells of the pancreas into the gut lumen and either auto-activated or activated by the enzyme enteropeptidase. In marine fish larvae, the amount of tryptic enzyme activity in the gut has been demonstrated to be a function of feed ingestion, gut filling and the composition of nutrients (RØNNESTAD et al., 2007; UEBERSCHÄR, 1995). The gastrointestinal hormone cholecystokinin (CCK) is known to play a key role in contraction of the gallbladder, peristalsis in the intestine, delay of gastric emptying and pancreatic enzyme secretion in mammals (SILVER et al., 1991) and adult fish (EINARSSON et al., 1997). In addition, it acts as a satiation signal in the fish brain (VOLKOFF et al., 2005). In mammals, CCK is considered one of the most important stimulators of pancreatic enzyme secretion (LIDDLE, 2006) and is therefore an obvious candidate to study the same functions in larval fish. Upon the presence of nutrients in the gut, CCK is released from enteroendocrine cells in the gut epithelium into the body fluids and acts on target cells in the pancreas to release secretions into the gut lumen. The stimulation in the gut might be of mechanical and/or biochemical 29

34 Chapter I nature. High tryptic enzyme activity in the gut acts as a negative feedback control for the release of CCK in humans, suggesting a regulatory loop between these two factors (LIDDLE, 2006) and the same mechanism has been described in adult fish (MURASHITA et al., 2008). The spatial distribution of the CCK-producing cells in the larval gut seems to vary between fish species (e.g. KAMISAKA et al., 2005; WEBB et al., 2010), and knowledge of the regulatory mechanism between CCK and trypsin still remains limited in developing fish larvae. Moreover, the differences in the spatial and temporal appearance of these cells indicate species-specific differences in controlling digestive processes (ROJAS-GARCÍA et al., 2011). Additionally, it has been shown in mammals that certain food components and digestive end products, like intact protein or certain amino acids, stimulate CCK and consequently pancreatic enzyme secretion, more than other nutrients (LIDDLE, 1995). Results of controlled tube-feeding studies (KOVEN et al., 2002) as well as standard feeding trials (CAHU et al., 2004; NAZ et al., 2009) suggest that similar mechanisms are present in early larval stages of fish. Low amounts of these stimulatory components in commercial microdiets may contribute to the inability of most marine fish larvae to utilize these diets efficiently from first-feeding (YÚFERA et al., 2000). Although functional studies on daily rhythms in marine fish larvae exist (e.g. feed uptake; KOTANI et al., 2011; PEDRO CAÑAVATE et al., 2006), physiological studies on digestive processes are not widespread in the literature and focused mostly on the response of proteolytic enzyme activity in relation to feeding schedules (APPLEBAUM et al., 2003; MACKENZIE et al., 1999; UEBERSCHÄR, 1995). Nevertheless, knowledge of diurnal cycles of physiological aspects, including digestive processes, has gained some attention in recent studies (FUJII et al., 2007; HARBOE et al., 2009; ROJAS-GARCÍA et al., 2011; YÚFERA, 2011). They may provide insight, for instance, on feeding times, feeding amounts and number of meals in larviculture practices in relation to digestive capacities and endogenic rhythms of the larvae. Aquaculture of Atlantic cod, Gadus morhua, is a relatively young industry with many challenges, including larviculture. Moreover, as wild cod stocks are being highly exploited in the North Atlantic and there are many unknown factors involved in the recruitment and mortality of year classes, this species serves as a model organism to tackle basic questions in larval digestive physiology. Here, we describe the ontogenetic development of CCK and tryptic enzyme activity in Atlantic cod larvae to provide insights into the capacity to regulate digestive processes in early cod. We conducted a short-term feeding experiment to evaluate the interaction between 30

35 Ontogeny and diurnal rhythm CCK and tryptic enzyme activity over 12 hours following different numbers of meals. Previous studies have shown that there is a relatively large amount of CCK found in the head part (central nervous system mainly) which may mask changes of CCK in the gastrointestinal tract (ROJAS-GARCÍA et al., 2011). Therefore, all analyses in the present study were done on dissected larvae excluding the head. Apart from disclosing the postulated mechanisms of CCK and tryptic enzyme activity, this provides valuable information on the diurnal digestive capacity in early cod related to practical feeding conditions. Materials and Methods Larval rearing Fertilized cod eggs were incubated in a 75 L hatching incubator for 17 days with fullstrength seawater at C. Gentle bubbling from the bottom kept the eggs in suspension and the water was exchanged at 4 L min -1 to maintain optimal water quality. Newly hatched cod larvae (3 days post-hatch) were counted using density estimates of three tube samples and larvae were transferred to a firstfeeding tank (450 L) to establish a density of around 110 larvae L -1. The black feeding tank was equipped with a two-directional water inlet immediately below the water surface at 50% of the tank radius and was aerated with fine bubbling using an aeration ring in the middle bottom of the tank. Water flow was gradually increased from 0.6 L min -1 on day 1 of the experiment (1 day post first-feeding, dpff) to 3.0 L min -1 at the end of the experimental period (38 dpff). Oxygen remained between 93-99% saturation throughout the experiment. Water temperature was gradually increased from 6 C to 11 C. Light was provided 24 hours a day applying indirect illumination of the rearing room and a weak light bulb (100 lux) above the tank. Dead larvae and debris were removed daily by siphoning the tank bottom and by using an automatic and rotating cleaner arm later in the experiment. A surface skimmer was installed to keep the water surface clean. Microalgae Nannochloropsis sp. paste (Nanno 3600, Reed Mariculture, USA) was pre-mixed with seawater and added daily in the morning (10 ml, 1-16 dpff; 15 ml, dpff). Enriched rotifers (Brachionus plicatilis; LARVIVA Multigain, BioMar, Denmark) were administered twice a day in the morning (10:00) and afternoon (15:00) with increasing rotifer densities in the tank throughout the experiment (5 rotifers ml -1 to 30 rotifers ml -1 ). In addition, algae paste and rotifers were provided 31

36 Chapter I continuously after the second feeding using a separate storage tank and a peristaltic pump. Enriched Artemia instar II nauplii (EG Artemia, INVE, Belgium) were co-fed with rotifers at densities of 1 ml -1 from 32 dpff until the end of the experiment. The larval rearing followed the best larviculture practices at the Austevoll Research Station larvae were sampled randomly each sampling day prior to feeding in the morning using a pipette with a large opening. The larvae were transferred with minimum of seawater into an Eppendorf vial and then immediately frozen on dry ice and subsequently stored at -80 C until analysis. Short-term diurnal rhythm experiment A 12 hour experiment was conducted in seven, green 40 L tanks on 21 dpff. Each tank was equipped with a water inlet right below the water surface and was aerated the same way as the main black tank described above. Water flow was set to 0.3 L min -1 and water parameters (Temperature, O 2 %) were not different from the main black tank. One day in advance, 150 larvae were transferred to each tank after the second meal in the afternoon for acclimatization. On the day of the experiment, 1.25 ml algae paste was added to each tank at 9:30 and 15:15, respectively. Three tanks were fed enriched rotifers once at 10:30 ( one meal ) at a density of 22 rotifers ml -1 and another three tanks were fed twice at 10:30 and 15:30 ( two meals ) at densities of 22 rotifers ml -1 and 11 rotifers ml -1, respectively. One tank was only given algae paste (considered as the control group) larvae were sampled every full hour as quickly as possible from each tank between 8:00-20:00 as described above and stored at -80 C until analysis. Sample preparation Individual samples were analyzed for CCK and tryptic enzyme activity according to ROJAS-GARCÍA et al. (2001) and UEBERSCHÄR (1993). Described briefly, frozen samples were allowed to thaw on ice, rinsed with distilled water and the standard length (mm, tip of upper jaw to end of notochord) was measured on an ice-cold Petri dish under a microscope. Gut fullness was evaluated, using a simple gut fullness index after ROJAS-GARCÍA et al. (2011): 0% (empty), <25%, 25-50%, 50-75%, % (full). The head was dissected from each larva and excluded from the analyses. Each larva was then transferred to an individual Eppendorf vial and homogenized in 50 µl ice-cold distilled water using a motorized pestle. For extraction of CCK, 750 µl 32

37 Ontogeny and diurnal rhythm Methanol were added, the sample was vortex-mixed thoroughly and incubated on ice for 30 min. After centrifugation (15 min., 1700 g, 4 C) each sample was split in two by transferring the supernatant to a new Eppendorf vial. Both, the remaining pellet (methanol-insoluble fish precipitate) and the supernatant (CCK methanol extract) were evaporated to dryness using a vacuum desiccator attached to a waterjet pump and stored at -20 C until analysis. Analysis of CCK and tryptic enzyme activity The individual CCK extracts were assayed by a competitive radioimmunoassay (RIA) using CCK-RIA kits (RB302, Euro-Diagnostika, Sweden) according to the supplier s instructions and ROJAS-GARCÍA et al. (2001). CCK levels were interpolated from a standard curve ( pmol CCK L -1 ) and concentrations are expressed as fmol larva -1. Recovery of known amounts of CCK added to samples throughout the extraction procedure was 71%. Tryptic enzyme activity in individual pellets was measured using a highly specific fluorescence substrate (Nα-benzoyl-L-arginine-methyl-coumarinyl-7-amide-HCl) according to UEBERSCHÄR (1993). Values for tryptic enzyme activity are expressed as hydrolysed fluorescence products MCA (methyl-coumarinyl-7-amide, nmol MCA min -1 larva -1 ). The coefficient of variation between triplicate measurements of samples was 1.6% (n = 4 samples). Statistical analysis CCK concentrations and tryptic enzyme activity levels during ontogenetic development were averaged for length classes in steps of 0.5 mm ( , etc.). Data of the diurnal rhythm experiment were tested for normality and homogeneity of variance using the Shapiro-Wilks-test and Levene s test, respectively. Significant differences in standard length, gut fullness, CCK concentration and tryptic enzyme activity were analyzed using a nested One-way ANOVA with measurements of individuals in each tank nested in treatment groups for each sampling point. Upon significance, differences between groups were assessed with Student-Newman-Keul s test. Significant differences between sampling points within each fed group were analyzed using a nested One-way ANOVA followed by a post hoc Duncan's multiple-comparison test. Values of gut fullness were transformed using the formula gut fullness =arcsin gut fullness. Statistics were performed with SPSS 19.0 for Windows and the level of significance was set to p < Data are presented as mean ± S.D. 33

38 Chapter I Results Ontogeny The mean standard length of cod larvae was 4.59 ± 0.24 mm on 1 dpff and increased to 8.14 ± 0.41 mm at the end of the experimental period (38 dpff, Fig. I-I). Growth was gradual during the first 3 weeks but tended to stagnate between 23 and 30 dpff. Figure I-I: Standard length (mm) of Atlantic cod larvae during ontogenetic development 2-38 days post first-feeding (dpff). Data are presented as mean + S.D. (n = at each sampling point). CCK was present in the dissected body at the earliest developmental stage and increased exponentially with increasing larval standard length (Fig. I-II). Tryptic enzyme activity increased slightly until the larvae reached a standard length of approx. 8 mm with a sharp increase afterwards and ranged between 0.04 and hydrolysed MCA, nmol min -1 with increasing variability over time among individuals of comparable size (Fig. I-III). 34

39 Ontogeny and diurnal rhythm Figure I-II: CCK content (fmol larva -1 ) for length classes (standard length, 0.5 mm precision, , etc.) of larval cod (excluding head). Larvae showing signs of the Distended Gut Syndrome (DGS) were excluded. Data are presented as mean + S.D. and numbers represent the number of larvae in each length class. 35

40 Chapter I Figure I-III: Tryptic activity (hydrolysed substrate, nmol MCA min -1 larva -1 ) for length classes (standard length, 0.5 mm precision, , etc.) of larval cod (excluding head). Larvae showing signs of the Distended Gut Syndrome (DGS) were excluded. Data are presented as mean + S.D. and numbers represent the number of larvae in each length class. Short-term diurnal rhythm experiment The observations of the gut fullness over the 12 hour feeding period are given in Fig. I-IV for all three groups. Ingestion of rotifers is indicated by an immediate increase in gut fullness both in the one meal and two meals group as short as 30 min. after feeding in the morning. Mean levels of gut fullness remained relatively stable over the whole day in both fed groups with the values in the two meals group being slightly higher compared to the one meal group. Gut fullness in the control group refers to aggregation of algae trapped in the hindgut in some of the individuals and was generally lower compared to larvae of the two fed groups with significant differences (except 12:00 and 17:00) compared to the two fed groups after feeding in the morning. Highly significant differences in gut fullness between samplings points within groups were found before (10:00) and immediately after feeding (11:00) in the one-meal group (F 12,26 = 5.187; p < ) as well as in the two-meals group (F 12,26 = ; p < ). The standard length was 7.09 ± 0.46 mm, 7.34 ± 0.57 mm and 7.32 ± 0.52 mm for the control, one meal and two 36

41 Ontogeny and diurnal rhythm meals group, respectively. The larvae were not significantly different in standard length between all three groups at any sampling point (data not shown). Figure I-IV: Relative gut fullness in the three feeding groups between 8:00 and 20:00. Algae were provided at 9:30 and 15:15 (all groups), rotifers at 10:30 ( one meal, two meals) and 15:30 ( two meals ). Data are presented as mean + S.D. ( control n = 5 individuals; one meal, two meals n = 3 tanks). A grey arrow indicates the administration of algae and a black arrow indicates the administration of rotifers. Different letters indicate significant differences between the three groups at a specific sampling point, n.s. indicates non-significant differences (nested One-way-ANOVA, Student-Newman-Keuls test, p < 0.05). Values of gut fullness were transformed using the formula gut fullness =arcsin gut fullness. CCK levels increased between 9:00 and 14:00 in all groups with fluctuating levels afterwards until the end of the sampling period (20:00). The highest levels were recorded one hour earlier at 19:00 with 2.69 ± 1.52 fmol larva -1, 1.89 ± 0.16 fmol larva -1 and 1.94 ± 0.22 fmol larva -1 in the control, one meal and two meals group, respectively (Fig. I-V). This time point represents 8.5 hours after the addition of rotifers in the one meal group and 3.5 hours after the second addition in the two meal group. The lowest levels were recorded one hour later at the end of the sampling period at 20:00 with 0.70 ± 0.33 fmol larva -1, 0.75 ± 0.15 fmol larva -1 and 0.58 ± 0.27 fmol larva -1 in the control, one meal and two meals group, respectively. CCK levels were not significantly different between groups at all sampling points. 37

42 Chapter I Tryptic activity in the control group increased gradually with a marked drop at 11:00 which was 1.5 hours after the addition of algae (1.89 ± 1.41 hydrolysed MCA nmol min -1 larva -1, Fig. I-V), reaching a peak at 12:00 (7.51 ± 3.70 nmol min -1 larva -1 ) and decreased successively between 12:00 and 19:00, irrespective of the second addition of algae at 15:15. The development of tryptic activity showed almost similar patterns in the one meal and the two meals groups. After an initial increase in the morning, tryptic enzyme activity dropped at 11:00 in both groups, 30 min. after the addition of rotifers (3.55 ± 0.56 nmol min -1 larva -1 for one meal and 4.07 ± 1.64 nmol min -1 larva -1 for two meal larvae). 4.5 hours after the administration of rotifers tryptic activity increased to reach a peak at 15:00 (7.51 ± 1.76 nmol min -1 larva -1 for one meal and 7.15 ± 2.73 nmol min -1 larva -1 for two meal larvae). Tryptic enzyme activity levelled off afterwards in both fed groups reaching lowest levels at 20: hours after the addition of rotifers (2.10 ± 0.81 nmol min -1 larva -1 ) in the one meal group and 19: hours after the second addition of rotifers (2.60 ± 1.24 nmol min -1 larva -1 ) in the two meal group (Fig. I-V). No significant differences in tryptic enzyme activity were found between all groups at any sampling point. 38

43 Ontogeny and diurnal rhythm Figure I-V: Daily pattern of CCK (fmol larva -1 ) and tryptic enzyme activity (nmol MCA min -1 larva -1 ) for cod larvae (excluding head) at 21 days post first-feeding: A = control (n = 5 individuals), B = one meal (n = 3 tanks), C = two meals (n = 3 tanks). Data are presented as mean ± S.D. A grey arrow indicates the administration of algae and a black arrow indicates the administration of rotifers. No significant differences in CCK and tryptic enzyme activity were found between all groups at each sampling point. 39

44 Chapter I Discussion Ontogeny We conducted the present study to examine the ontogenetic development of CCK and tryptic enzyme activity in cod larvae between 1-38 dpff. An additional aim was to examine the dynamic dependency of both factors based on a 12 hour monitoring of CCK and tryptic enzyme activity related to a different number of meals. Growth expressed as standard length showed a gradual increase in the experimental period and was comparable to growth reported in recent literature on cod larvae (BUSCH et al., 2011; MEYER et al., 2012; PENGLASE et al., 2010). A noticeable growth depression was evident between dpff. In this period, a disease referred to as the Distended Gut Syndrome appeared, which symptoms have been described in several marine fish larvae, including cod, and causes cumulative mortality due to loss of appetite (KAMISAKA et al., 2010b). Around 20-30% of the larvae sampled in this period revealed signs of DGS and this might have been the reason for reduced growth. These larvae had generally lower gut fullness and tryptic enzyme activity values (data not shown). All larvae with signs of DGS were excluded from further statistical analysis. As mentioned previously, it has been shown that CCK is quantitatively dominant in the brain of larval fish, e.g. as described for herring, Clupea harengus (ROJAS-GARCÍA et al., 2011), halibut, Hippoglossus hippoglossus (ROJAS-GARCÍA et al., 2002) and sea bass, Dicentrarchus labrax (TILLNER et al., unpubl. data). Besides its regulatory role of the digestive processes, CCK also acts as a satiation signal in the brain of humans (SMITH, 2009) and adult fish (VOLKOFF et al., 2005). However, the latter role has not been experimentally explored in larval fish. In the present study, our aim was to investigate the role of CCK in relation to the regulatory functionality on tryptic enzyme activity in the gut. Consequently, the head of all cod larvae was separated before the analyses in order to exclude neural sources of CCK. However, it must be pointed out, that no differentiation between synthesized CCK in cells in the gut epithelium and released CCK into the body fluids was made in the present study. CCK was found in larval cod immediately after hatching, although these concentrations were close to the detection limit of the RIA, proposing a limited regulatory function of CCK at this developmental stage. The following exponential increase of CCK in larval cod over standard length suggests an increasing importance of CCK in the digestive system which is supported by data that demonstrated an 40

45 Ontogeny and diurnal rhythm increasing number of CCK-producing cells during ontogeny using immunohistochemical staining from 6 dpff onwards (HARTVIKSEN et al., 2009). These cells were mainly found in the anterior midgut where contact with stimulatory nutrients most likely permits a regulation of gallbladder and pancreas secretions but also in the hindgut later in development where the physiological role is less clear. A pronounced individual variability in the number of CCK-producing cells within comparable developmental stages was revealed, although the authors pointed out that different postprandial states or the low sensitivity of the staining method might have contributed to this variability (HARTVIKSEN et al., 2009). Nevertheless, individual differences in CCK-producing cells might explain the variability of CCK levels in the present study. In general, data on the ontogenetic development of CCK content in fish larvae are rather scarce and most studies focused on the description of CCKproducing cells (HARTVIKSEN et al., 2009; KAMISAKA et al., 2001; KAMISAKA et al., 2005; KAMISAKA et al., 2003; MICALE et al., 2010; WEBB et al., 2010), gene expression of CCK (KORTNER et al., 2011a; KORTNER et al., 2011b) or both (KUROKAWA et al., 2000; KUROKAWA et al., 2004). KORTNER et al. (2011b) revealed a moderate but consistent decrease in CCK expression in whole larval cod between 3-60 days post hatch. Given the assumption that CCK is mainly expressed in the brain, the authors contribute the decreased expression to a decreasing proportion of brain tissue compared to the whole body. This statement is supported by a study from CAHU et al. (2004) on larval sea bass in which a clear allometric relationship between dry weight specific CCK content and dry weight was evident. In contrast, a study on larval sea bream, Sparus aurata, using the same analytical method for CCK as in the present study, revealed almost no increase in whole body CCK until 40 days post-hatch in pooled samples (NAZ et al., 2009). To reveal the ontogenetic development of CCK in more detail, ROJAS-GARCÍA et al. (2002) described the compartmental distribution of CCK in first-feeding halibut larvae, between 7-26 dpff, where the proportion of CCK in the gut compared to whole body CCK increased from 2-62%. By contrast, a different distribution of CCK was found in larval herring where whole body CCK increased 15-fold until 40 days post-hatch, but CCK in the gut remained constant at relatively low levels (ROJAS- GARCÍA et al., 2011). This difference between halibut and herring could be contributed to morphological and consequently physiological differences of the digestive tract with herring having a straight and halibut having a rotated gut (ROJAS- GARCÍA et al., 2011). However, it remains to be proven if these low amounts are able to elicit a regulatory response in first-feeding fish larvae (ROJAS-GARCÍA et al., 2002) 41

46 Chapter I which also counts for the CCK concentrations found in the present study in larval cod. With regard to tryptic enzyme activity, only a slight increase could be observed in our study until the larvae reached a standard length of 8 mm (approx. 33 dpff), which is the same pattern found for trypsin by KORTNER et al. (2011b). On the transcriptional level, an increased mrna expression for trypsin until 17 days posthatch was observed by these authors, followed by a continuous decline thereafter. This discrepancy between gene expression and enzyme activity could be contributed to a post-transcriptional hormonal control or to a developing acidic digestion due to a developing stomach towards metamorphosis (KORTNER et al., 2011b). In the earlier case, CCK levels between fmol larva -1 found for larval cod of comparable age (to 17 dph in KORTNER et al., 2011b) in the present study might be the threshold for the regulatory role of CCK in larval cod. In general, the developmental pattern of different digestive enzymes has been described in several species, including California halibut, Paralichthys californicus (ALVAREZ-GONZÁLEZ et al., 2005), sea bass (CAHU et al., 1994), cobia, Rachycentron canadum (FAULK et al., 2007), red drum, Sciaenops ocellatus (LAZO et al., 2000a) and Atlantic cod (WOLD et al., 2007) and has been found to be species-specific. It has been proposed, that digestive enzyme activities in early stages of marine fish larvae are under transcriptional control but can be triggered by the nutritional composition of the diet (ZAMBONINO INFANTE et al., 2001). In our study, a decrease in tryptic enzyme activity is obvious between mm standard length (approx dpff) and might indicate only limited digestive capacity during this developmental period in larval cod. This intermediate decrease has been observed in several species and has been proposed as ontogenetic deficiencies in digestive capacity and the inability to digest food properly at certain developmental stages (UEBERSCHÄR, 1995). This assumption is strengthened by the fact that trypsin is by far the most important enzyme for protein digestion in the larval stages of altricial species (UEBERSCHÄR, 1995). For example, a decrease in digestive capacity has been observed in larval herring two weeks after first-feeding with copepods, independent of food density (PEDERSEN et al., 1990). Moreover, certain diets might be inadequate to provoke a digestive response via the CCKtrypsin-axis since the decrease in tryptic enzyme activity in our study coincides with the transitional feeding period from rotifers to Artemia. Species-specific differences and a more rapid development in digestive capacity may, at least in parts, contribute to the success in larviculture of different species (UEBERSCHÄR, 1995). Furthermore, individual variability in tryptic enzyme activity at comparable developmental stages 42

47 Ontogeny and diurnal rhythm of one species may divide individual larvae into winners and losers and might explain individual fates in survival and growth. Short-term diurnal rhythm experiment In order to evaluate the suggested feedback mechanism between CCK and tryptic enzyme activity in dependence on a different number of meals, we conducted a short-term experiment at 21 dpff. The immediate increase in gut fullness in the control group after the addition of algae in the morning suggests an active or passive uptake of algal cells which are trapped in the hindgut (KJØRSVIK et al., 1991). Addition of algae ( green water ) is routinely used in larviculture and is believed to aid in water quality, to enhance contrast for capturing prey (CONCEIÇÃO et al., 2010b; ROCHA et al., 2008; VAN DER MEEREN et al., 2007) and to reduce stress. Ingested algae were, however, not considered to represent a full meal in terms of energy and nutrients supplied. In the morning, live rotifers were immediately ingested by the larvae, indicated by increased mean gut fullness 30 min. after the addition in the two fed groups. The degree of gut fullness in those two groups was more or less constant over the sampling period and was only slightly higher in the two meal group at almost all sampling points. However, there was no clear effect of a second feeding in the afternoon, indicated by the gut fullness of the two meal group. Although the rotifer density in the tanks was not evaluated, a complete water exchange every 2-3 hours and the consequent removal of rotifers indicates a gut retention time of ingested prey up to several hours in cod larvae at this age. CCK increased slightly in all groups until 14:00 without clear differences in all groups. A marked increase in all groups was revealed at 19:00 reaching highest levels in all groups followed by a drop one hour later reaching lowest levels in all groups which points to a delayed response to a nutrient stimulus of newly ingested algae or algae and rotifers. The only study on diurnal rhythm of CCK in fish larvae was a recent work on Atlantic herring done by ROJAS-GARCÍA et al. (2011). In their investigations, no clear response of CCK in the body (excluding head) was observed after feeding over three days, although CCK levels were higher in fed larvae compared to starved larvae. An increase in the carcass/gut ratio of CCK 30 min. after feeding can be contributed to a release and re-synthesis of CCK (ROJAS-GARCÍA et al., 2011). Similarly, an increase in the carcass/gut ratio of CCK was also found in tube-fed halibut larvae four hours after the administration of protein (albumin) as a nutrient stimulus (ROJAS-GARCÍA et al., 2002). This underlying effect might be masked in the present study, since whole-body homogenates, excluding heads, were analysed. 43

48 Chapter I Tryptic enzyme activity started from relatively high levels in the morning prior to feeding in all three groups, although almost no larvae had food remainings in their guts at this time. This can be considered as the pre-feeding basal level of regularly fed larvae. Similarly, larvae of herring and Japanese eel (Anguilla japonica) have been shown to retain trypsin in their intestine up to several hours after a meal to be available for newly arriving prey (PEDERSEN et al., 1992; PEDERSEN et al., 2003). On the other hand, it has been shown that adult sea bream are able to synchronize their behaviour and increase enzyme secretion to a fixed feeding time to prepare for a forthcoming meal (MONTOYA et al., 2010). In addition, a diurnal rhythm in feed ingestion and tryptic enzyme activity over three days was observed by FUJII et al. (2007) for malabar grouper larvae (Epinephelus malabaricus, 3-5 days post-hatch) despite continuous light and stable prey densities in the rearing tanks. Similar reasons might be responsible for the relatively high tryptic enzyme activity in the morning in larval cod. Interestingly, a marked decrease in tryptic enzyme activity was evident at 11:00 in all groups. Since trypsin is retained in the gut and not reabsorbed for up to several hours after a meal (PEDERSEN et al., 1992), this marked decrease in tryptic enzyme activity might be the consequence of immediate binding of trypsin to algal cells and rotifers acting as a substrate. There was no real starving group represented in the present study but the increase in tryptic enzyme activity at 12:00 in the control group is most likely the consequence of ingested algae that trigger the digestion process, as proposed by other authors (CAHU et al., 1998b; REITAN et al., 1997). An additional group without algae and rotifers could have shed light on the role of algae in triggering digestive processes and also on graded responses in the digestive system to the size and type of ingested particles/prey (HJELMELAND et al., 1988). On the long term, the addition of algae might facilitate the maturation of the digestive system by triggering digestive processes and might contribute to the overall higher success of larviculture in green water described in the literature (e.g. CAHU et al., 1998b; FAULK et al., 2005; LAZO et al., 2000b). Consequently, rearing marine fish larvae in green water might be an essential part of best larviculture practices. The peak in the control group was followed by a gradual decrease towards the end of the day, indicating an immediate release of trypsinogen from the pancreas after nutritional stimulation followed by empty stores afterwards and basal tryptic enzyme activity levels in the gut. The increase in tryptic enzyme activity was delayed up to three hours in the two fed groups, reaching high levels at 15:00, respectively. This might be due to slow and 44

49 Ontogeny and diurnal rhythm gradual release of stimulatory components from the ingested rotifers although based on visual inspection of cod in the present study, the rotifers were rapidly degraded. The fact that maximum tryptic enzyme activity levels were not different between the two fed groups and the group receiving only algae (control) may imply a dominant role of a chemical stimulus over a mechanical stimulus of ingested prey. Alternatively, only slight mechanical stimulation by the algae might be sufficient to elicit a digestive response. By contrast, the content of trypsin in the gut is found to be a function of ingested prey items in larval herring (PEDERSEN et al., 1987). Similarly to the control group, tryptic enzyme activity tended to drop in the afternoon in both fed groups with no apparent effect of a second meal in the 2 meals group. This observation strengthens the indication, that pancreatic resources for trypsinogen in larval cod are exhausted after an initial stimulation with no response after subsequent ingestion of algae or algae and rotifers. This also emphasizes the importance of retaining and re-using trypsin in the intestinal lumen. By contrast, tryptic enzyme activity was persistently high in larval herring after two consecutive sequences of feeding and digestion, pointing to the ability of larval herring to digest several meals within a short period after an initial release of enzymes (PEDERSEN et al., 1988). Similar results were shown for larvae of African catfish, Clarias gariepinus, 7 days post first-feeding, where, given several meals a day, tryptic enzyme activity fluctuated around a stable level over 24 hours, indicating a higher capacity to handle several meals a day in this species (GARCÍA-ORTEGA et al., 2000). Similar to the present study, a decrease in tryptic enzyme activity 1-2 hours after food ingestion was found in the catfish, indicating an immediate utilization of trypsin in the gut for protein hydrolysis. Therefore, a response in tryptic enzyme activity to a food stimulus after ingestion might be reflected thereafter (GARCÍA- ORTEGA et al., 2000). A delayed postprandial response in tryptic enzyme activity has also been shown for larval turbot, Scophthalmus maximus, with a persistent postprandial increase in older larvae indicating an increasing digestive capacity with age in these species (UEBERSCHÄR, 1995). Taken together, CCK and tryptic enzyme activity revealed a reverse diurnal trend in all groups in the present study with a marked increase in CCK when tryptic enzyme activity levels were low. Given the high and stable gut filling over the entire sampling period in all groups which most likely results in a continuous and stable nutrient stimulus in the gut, the data suggest that CCK is synthesized and released as a stimulatory response when tryptic enzyme activity in the gut is low and vice versa. 45

50 Chapter I When this regulatory loop becomes functional still remains to be established. Such negative feedback control has been observed in humans (LIDDLE, 2000) and has been proposed to be also effective in fish larvae (CAHU et al., 2004; RØNNESTAD et al., 2007). In detail, the presently available data suggest that in the presence of trypsin a CCKreleasing peptide in the gut is degraded, whereas ingested protein competes as a substrate for trypsin and the intact CCK-releasing peptide stimulates the release of CCK (LIDDLE, 1995; MIYASAKA et al., 1989). In a controlled tube-feeding study on firstfeeding herring larvae, KOVEN et al. (2002) found an immediate increase in wholebody CCK after administration of solutions containing protein (albumin), protein plus free amino acid and free amino acid only (in descending order of CCK response). This was followed by an immediate increase in tryptic enzyme activity in all groups (in the same order of magnitude). Given the missing postprandial response in body-cck (excluding head) in the same species (ROJAS-GARCÍA et al., 2011) and in the present study, the increase in CCK in the study of KOVEN et al. (2002) is likely to be an increase in CCK in the brain and might act as a satiation signal instead of acting as a stimulation of pancreatic secretions. Similar conclusions were drawn for adult channel catfish, Ictalurus punctatus, and juvenile salmon, Salmo salar (PETERSON et al., 2012; VALEN et al., 2011). Finally, it has to be pointed out that standard length was chosen as the only proxy for growth due to several reasons. It reflects very well the developmental stage in larval cod (FINN et al., 2002) and is a conservative measure in starving larvae, which quickly lose protein and consequently body mass, which might have happened in the control group, but not in standard length. It has to be noted, that due to the analytical procedure, it was not feasible to assess body mass and/or protein content in individual larvae without any impact on the tryptic enzyme activity and CCK analytics. The focus on the analyses of tryptic enzyme activity and CCK in individual larvae in this study required to consider standard length as the reference. Conclusions Our study shows the individual development of CCK and the key enzyme trypsin in larval Atlantic cod and demonstrates a feedback mechanism among CCK and trypsin in regulating digestive processes in early larval stages of cod. The role of CCK as a trigger of pancreatic secretions is likely to mature as ontogeny proceeds, although spatial and temporal differences in CCK in different body compartments complicate the interpretation in developing fish larvae. Tryptic enzyme activity increased only slightly in early cod pointing towards limited digestive capacity early in 46

51 Ontogeny and diurnal rhythm development. Results of the 12 hour experiment revealed that tryptic enzyme activity increased immediately after a nutrient stimulus consisting of algae and rotifers even with the administration of an algae solution without rotifers, which supports earlier findings that algae may play a key role in the maturation process of the digestive system in marine fish larvae. A second meal of rotifers the same day did not result in increasing enzyme activity, suggesting a limited proteolytic capacity in cod larvae to handle several meals in a short time period. Therefore, feeding times, frequency and amounts should be matched to the digestive capacity of the larvae to maximize nutrient utilization and growth. A reverse trend between CCK and tryptic enzyme activity was evident in all groups, indicating a negative feedback control in cod larvae similar to that found in mammals. Acknowledgements This study was financially supported by the German Federal State of Schleswig- Holstein and the European Regional Development Fund (ERDF) through the NEMO project ( ) and was supported by COST Action FA0801 LARVANET (to R.Tillner), ISOS Integrated School of Ocean Sciences (to R. Tillner), Research Council of Norway (proj no ; GutFeeling to I. Rønnestad and T. Harboe; ; CODE to I. Rønnestad). The authors thank the technical staff at the Austevoll Research Station for technical assistance as well as Dr. Nils Roos and Frauke Repening from the Max Rubner-Institut Kiel for analytical support during the CCK analyses. 47

52 48

53 Chapter II CHAPTER II Set-up of the tube-feeding procedure 49

54 Chapter II Evidence for a regulatory loop between cholecystokinin (CCK) and tryptic enzyme activity in Atlantic cod larvae (Gadus morhua) Abstract In order to maximize protein digestion, the release of enzymes into the gut lumen is closely controlled by a regulatory loop. Cholecystokinin (CCK) is among the enteric hormones that play a key role in the control of digestive enzyme secretion, but its role in first-feeding larvae is still unclear and may differ between species. However, in all marine fish larvae that have not developed a stomach by first-feeding, trypsin is the most important proteolytic enzyme. In order to examine the regulation and feedback mechanisms in the gut of larval cod we therefore studied the interactions between cholecystokinin and tryptic enzyme activity following the administration of solutions containing test substances directly into the gut. We tube-fed a single dose of physiological saline solution containing either CCK, CCK antagonist, trypsin inhibitor, phytohemagglutinin (PHA; a possible trigger for the digestive response) or physiological saline alone, while a further control group was left untreated. We then followed the response in CCK and tryptic enzyme activity for 0.5 8h after the administration. We performed the experiment on larvae at 26 days post firstfeeding, which is before the stomach has evolved and the size of the larvae allows easier handling. Individual larvae were analyzed for CCK and tryptic enzyme activity using radioimmunoassay and fluorometric techniques, respectively. Both factors varied over time in the untreated control group, possibly due to an endogenous daily rhythm. The higher CCK levels at 4h and 8h in the saline-injected group may be caused by reflexes initiated by distension of the gut. An increase in tryptic enzyme activity after injection of CCK supports the hypothesis that this hormone plays a part in the release of pancreatic enzymes in larval cod at this developmental stage. However, administration of a CCK antagonist and a trypsin inhibitor did not reveal conclusive results, probably due to the relatively low concentrations used. The response in tryptic enzyme activity in the PHA group was similar to the administration of CCK, pointing towards a stimulatory effect of PHA on the proteolytic enzyme capacity of cod larvae. Keywords: CCK, trypsin, digestion, ontogeny, endocrine control 50

55 Regulatory loop Introduction The supply of high quality larvae and juveniles in stable quantities is still a bottleneck in the commercialization of many marine fish species for aquaculture. The reasons for this include the still very high mortality in many species and the variable quality of larvae. A major constraint is our lack of knowledge of larval digestive physiology, which apparently still hampers the provision of optimal feeds, particularly during critical periods after the onset of first-feeding, the period when the highest mortalities occur. Understanding larval digestive physiology is therefore a prerequisite to providing the nutritional standards required of both live prey and artificial food (CONCEIÇÃO et al., 2007; RØNNESTAD et al., 2013). The gastrointestinal tract is the largest endocrine organ in vertebrates (HOLST et al., 1996) and its hormonal factors interact at several levels with a complex enteric neural network and signalling pathways to and from the brain to maximize digestion and absorption. Of the many gastrointestinal hormones involved in digestion, cholecystokinin (CCK) is known to play a key role in the contraction of the gallbladder, peristalsis in the intestine, delay of gastric emptying and it is also one of the most important regulators of pancreatic enzyme secretion (LIDDLE, 2006). CCK also acts as a satiation signal in the brain (KONTUREK et al., 2004). Current data suggest that CCK is released from enteroendocrine cells in the gut epithelium into the systemic circulation upon the presence of nutrients in the gut lumen, where it acts on target cells in the pancreas to trigger the secretion of digestive enzymes into the gut lumen. Among the pancreatic enzymes, trypsin is believed to be the key proteolytic enzyme in the digestive process (LIDDLE, 2006). Trypsin is secreted as its inactive precursor trypsinogen from the pancreatic acinar cells into the gut lumen, where it is either auto-activated or activated by the enzyme enteropeptidase. In humans, high tryptic enzyme activity in the gut acts as a negative feedback control for the release of CCK, suggesting a regulatory loop between these two factors (LIDDLE, 2006). There is growing evidence that many digestive processes and related mechanisms such as feed uptake are conserved from fish to mammals (EINARSSON et al., 1997; GÉLINEAU et al., 2001; KOFUJI et al., 2007; MURASHITA et al., 2008; PETERSON et al., 2012; RØNNESTAD et al., 2010; RUBIO et al., 2008; VOLKOFF et al., 2005). For example, the synthesis and secretion of pancreatic enzymes is mediated by CCK in response to various nutrients in adult yellowtail, Seriola quinqueradiata (MURASHITA et al., 2008). The digestive tract in most marine fish larvae at first-feeding is simply organized and 51

56 Chapter II an adult-like mode of digestion, including a functional stomach, is acquired during metamorphosis that occurs later in development. We now have a good understanding of how the tissue and organs of the digestive system develop in marine fish larvae as well as of the ontogeny of digestive enzymes (reviewed by LAZO et al., 2011). In the absence of a stomach, protein hydrolysis is mainly dependent on the gut s alkaline enzymes, the most important being trypsin (UEBERSCHÄR, 1993; ZAMBONINO INFANTE et al., 2001). The enteric CCK-producing cells are located in various areas in the gut of fish larvae (HARTVIKSEN et al., 2009; KAMISAKA et al., 2001; KAMISAKA et al., 2005; MICALE et al., 2010), but they are consistently encountered in the anterior part of the midgut, which suggests that they play a regulatory role in digestion similar to that found in mammals (RØNNESTAD et al., 2007). However, little is known about the regulatory mechanism involving CCK and trypsin in developing fish larvae, compared to mammals and adult fish. Nevertheless, some studies have been focusing on the interaction between these two factors in the larval stages of marine fish and these studies have suggested that CCK plays an important role in the release of pancreatic enzymes (CAHU et al., 2004; KOVEN et al., 2002; ROJAS-GARCÍA et al., 2002). It has been hypothesized that marine fish larvae, lacking a functional stomach at first-feeding, possess intestinal and pancreatic enzymes in sufficient amounts to digest live food at first-feeding (reviewed by LAZO et al., 2011). However, it has been shown that in mammals, certain food components, like intact protein or some amino acids, stimulate CCK and consequently pancreatic enzyme secretion more than other nutrients (LIDDLE, 1995). Results of controlled tube-feeding studies (KOVEN et al., 2002), as well as standard feeding trials (CAHU et al., 2004; NAZ et al., 2009) suggest that similar mechanisms are present in early larval stages of fish. One possible explanation of the inability of most marine fish larvae to utilize microdiets efficiently from first-feeding is that pancreatic enzyme secretion is not sufficiently stimulated to deal with some of the ingredients in artificial foods (WEBB et al., 2011; YÚFERA et al., 2000). It is therefore of interest to identify factors that could enhance the production and release of trypsinogen in the early stages in sufficient amounts to increase digestive capacity. In experiments aimed at triggering CCK production, the plant protein phytohemagglutinin (PHA) was found to stimulate growth of the pancreas and small intestine as well as the release of CCK in rats and piglets (HERZIG et al., 1997; RADBERG et al., 2001). Following these results, it was supposed that PHA might have a stimulatory effect on CCK and consequently on 52

57 Regulatory loop trypsinogen synthesis, and its release in the larval fish gut (DROSSOU et al., 2006). While pancreatic enzyme secretion is mediated via CCK release (LIDDLE, 2006), the administration of the trypsin inhibitor camostat induced the release of CCK and pancreatic growth in mice, suggesting a negative feedback loop between CCK and trypsin (GUO et al., 2007). This study examines a possible interaction between CCK and trypsin in vivo in Atlantic cod larvae, applying the tube-feeding method developed by RUST et al. (1993) and RØNNESTAD et al. (2001), where small volumes of dissolved substances are administered directly into the gut. We administered CCK, a CCK antagonist, and a trypsin inhibitor to individual cod larvae in order to determine whether tryptic enzyme activity in the gut is mediated by CCK and whether trypsin acts as a negative feedback control for CCK release. PHA was tested as a potential feed additive to induce CCK and thus pancreatic enzyme secretion in larval cod. In order to determine whether there is a time-dependent response of CCK and tryptic enzyme activity, we followed the treated fish for between 0.5 and 8 hours after tubefeeding. In previous experiments, a significant concentration of CCK has been found in the brain of larval teleosts, including Atlantic halibut, Hippoglossus hippoglossus (ROJAS-GARCÍA et al., 2002) and herring, Clupea harengus (ROJAS-GARCÍA et al., 2011). Since the role of CCK in the brain has yet to be elucidated (RØNNESTAD et al., 2007), the heads of the larvae were removed and excluded from the analyses. CCK and tryptic enzyme activity were analyzed via a combination of highly sensitive methods (ROJAS-GARCÍA et al., 2001) which allow the analyses of both factors in individual larvae. Materials and Methods Larval rearing Fertilized cod eggs were incubated in a hatching tank (75 L) for 17 days in seawater (35 g L -1 ) at temperatures of C. Slight bubbling from the bottom kept the eggs suspended and the water was exchanged at 4 L min -1 in order to maintain optimal water quality. On hatching, the larvae were counted on the basis of density estimates of three tube samples, and larvae were transferred to a black firstfeeding tank (450 L) at a density of 110 larvae L -1. The tank was equipped with a twodirectional water inlet just below the surface at 50% of the tank radius and was aerated with fine bubbling through an aeration ring at the centre of the bottom of 53

58 Chapter II the tank. The water exchange rate was gradually increased from 0.6 L min -1 to 3.0 L min -1 during the experimental period (1-38 dpff; days post first-feeding). Oxygen was kept at between 93-99% saturation throughout the experiment and the water temperature was gradually increased from 6 C to 11 C. Light was provided 24 hours a day via indirect illumination of the rearing room and a 100 lux light source above the tank. Dead larvae and debris were removed daily by siphoning the tank bottom and by an automatic and rotating cleaner arm later in the experiment. A skimmer was installed to remove debris from the water surface. Microalgae paste (Nannochloropsis sp., Nanno 3600, Reed Mariculture, Campbell CA, USA) was pre-mixed with seawater and added daily in the morning (10 ml, 1-16 dpff; 15 ml, dpff). Cod larvae were fed enriched rotifers (Brachionus plicatilis; LARVIVA Multigain, BioMar, Aarhus, Denmark) twice a day in the morning (10:00) and afternoon (15:00) in order to raise rotifer densities in the tank in the course of the experiment (5 rotifers ml -1 to 30 rotifers ml -1 ). In addition, algae paste and rotifers were provided continuously after the second feeding using a separate storage tank and a peristaltic pump. Enriched Artemia instar II nauplii (EG Artemia, INVE, Ghent, Belgium enriched with LARVIVA Multigain) were co-fed with rotifers at densities of 1 ml -1 from 32 dpff until the end of the experiment. In summary, the cod larval rearing followed the best larviculture practices at the Austevoll Research Station (DR. T. VAN DER MEEREN, pers. comm.). Tube-feeding One day before tube-feeding, approx. 300 larvae were transferred into a small tank (40 L) and deprived of food for three hours after the second feeding in the afternoon, under the same hydrographical conditions as in the feeding tank. The larvae were thus fasted for at least 14 hours to ensure that their guts were empty. A further ten larvae were individually weighed on a micro-balance in order to adjust the concentrations of the experimental solutions according to the average wet weight. The tube-feeding experiment was carried out at 26 dpff. In the morning, the larvae were placed into a bucket with fresh and slightly aerated seawater and transferred to a temperature-controlled room whose temperature was similar to the rearing temperature. Individual larvae were anaesthetized with MS-222 (30µg/ml, Tricaine methanesulfonate, Sigma-Aldrich, St. Louis MO, USA) and tube-fed a volume of 9.2 nl with one of the following solutions: physiological saline as control, CCK (3 fmol 54

59 Regulatory loop mg -1 wet weight, Bachem, Torrance CA, USA), CCK antagonist Proglumide (0.5 pmol mg -1 wet weight, Sigma-Aldrich), Phytohemagglutinin (1 ng mg -1 wet weight, PHA, Sigma-Aldrich) and trypsin inhibitor (3 pmol mg -1 wet weight, Camostat, Tocris Bioscience, Bristol, UK). A summary of the experimental dosage protocol is shown in Tab. II-I. After injection, each larva was rinsed with clean seawater and transferred to an incubation well in a 6-well multidish, containing 5 ml temperature-controlled seawater. At least six injected larvae per treatment and sampling point were sampled after either 0.5, 1, 2, 4 or 8 hours incubation time. The survival of the tubefeeding procedures was 71%. Individual larvae were transferred with a droplet of seawater to 1.5 ml microcentrifuge tubes, immediately frozen on dry ice and stored at -80 C until analysis. Table II-I: Experimental tube-feeding protocol; concentrations of injected substances, incubation times and number of sampled larvae. Nr. of sampled larvae 9.2 nl injection volume per larva 0.5 h 1 h 2 h 4 h 8 h physiological saline CCK 3 fmol mg -1 wet weight CCK antagonist 0.5 pmol mg -1 wet weight phytohemagglutinin 1 ng mg -1 wet weight trypsin inhibitor 3 pmol mg -1 wet weight Sample preparation Individual samples were analyzed for CCK and tryptic enzyme activity according to ROJAS-GARCÍA et al. (2001) and UEBERSCHÄR (1993). In brief, frozen samples were thawed on ice, rinsed with distilled water and their standard length (mm, tip of upper jaw to end of notochord) was measured on an ice-cold Petri dish under a dissecting microscope. Heads were dissected and excluded from the analyses. Bodies were transferred to an individual 1.5 ml microcentrifuge tube and homogenized in 50 µl ice-cold distilled water using a motorized pestle. For extraction of CCK, 750 µl methanol was added, the sample was vortex-mixed thoroughly and incubated on ice for at least 30 min. After centrifugation (15 min., 1700 g, 4 C) the supernatant was transferred to a fresh 1.5 ml microcentrifuge tube. Both the remaining pellet (methanol-insoluble fish precipitate) and the supernatant (CCK methanol extract) were evaporated to dryness using a vacuum desiccator attached to a water-jet pump, and stored at -20 C until analysis. 55

60 Chapter II Analysis of CCK and tryptic enzyme activity The individual CCK extracts were assayed by competitive radioimmunoassay (RIA) using CCK-RIA kits (RB302, Euro-Diagnostika, Malmö, Sweden) according to the supplier s instructions and following the modifications of ROJAS-GARCÍA et al. (2001) for measurement in tissue of individual fish larvae. CCK levels were interpolated from a standard curve ( pmol CCK L -1 ) and concentrations are expressed as fmol larva -1. Recovery of known amounts of CCK added to samples throughout the extraction procedure was 71%. Tryptic enzyme activity in individual pellets was measured using a highly specific fluorescence substrate (Nα-benzoyl-L-arginine-methyl-coumarinyl-7-amide-HCl) according to UEBERSCHÄR (1993). Values for tryptic enzyme activity are expressed as fluorescence products MCA (methyl-coumarinyl-7-amide, nmol MCA min -1 larva -1 ). The coefficient of variation between triplicate measurements of samples was 1.6% (n = 4 samples). Statistical analysis CCK and tryptic enzyme activity were tested for normality and homogeneity of variance using the Shapiro-Wilk test and Levene s test, respectively. Significant differences between the experimental groups for each sampling point and between sampling points within each group were analyzed by One-way ANOVA for CCK and tryptic enzyme activity, respectively. Upon significance the Student-Newman-Keul s test was applied. Statistics were performed with SPSS 19.0 for Windows and the level of significance was set to p < Data are expressed as means ± Standard Error of the Mean (SEM). Results CCK levels (fmol larva -1 ) in the untreated group displayed the lowest fluctuation across sampling times between all groups (Fig. II-I). Tryptic enzyme activity (nmol MCA min -1 larva -1 ) increased between 0.5h (1.39 ± 0.64) and 1h (2.27 ± 1.28) and decreased afterwards, reaching its lowest levels after 8h (0.25 ± 0.06). A similar trend in both factors was evident in the group injected with physiological saline, although CCK levels were higher at 4h (1.96 ± 0.11) and 8h (2.51 ± 0.35) after injection, respectively. By contrast, tryptic enzyme activity was at a low level at 0.5h (0.61 ± 0.21) and 1h (0.31 ± 0.07) in the CCK-injected group, but substantially at 2h, 56

61 Regulatory loop 4h and 8h. Unlike both control groups, larvae treated with CCK antagonist had lower CCK values at 8h (1.27 ± 0.21) after injection and values of tryptic enzyme activity tended to be higher at 2h (1.94 ± 1.51) and 4h (1.87 ± 1.13). Larvae treated with the plant protein phytohemagglutinin (PHA) revealed continuously rising CCK values, with the highest levels observed 4h after injection (2.75 ± 0.59) and a subsequent decrease at 8h (1.85 ± 0.23), showing the same pattern as in the group injected with CCK. Notably, tryptic enzyme activity was higher than in both control groups, except for 1h after injection, reaching values in the same range as in the CCK-injected group. In the group injected with trypsin inhibitor, no differences in CCK were evident compared to the group injected with physiological saline. Tryptic enzyme activity was generally lower than in the two control groups, and displayed only slight variation. Larvae were not significantly different in standard length between treatment groups at all sampling points (data not shown). 57

62 Chapter II Figure II-I: CCK content (fmol larva -1 ) and tryptic enzyme activity (hydrolysed substrate, nmol MCA min -1 larva -1 ) in larval cod (excluding head). Data are presented as mean ± standard error of mean (n = 3-6 larvae per treatment and sampling point). Different letters indicate significant differences in CCK or tryptic enzyme activity between treatments at a specific sampling point and a different number of (+) significant differences between sampling points within each group. No letters or symbols indicate no significant differences (One-way-ANOVA, Student-Newman-Keuls test, p < 0.05). Discussion The present study intended to determine whether there is a feedback mechanism between CCK and tryptic enzyme activity in cod larvae similar to that found in mammals and adult fish. Due to their tiny and fragile nature, functional studies on endocrine mechanisms in marine fish larvae in vivo are difficult. High-resolution analyses of the compartmental distribution of hormones and enzymes, which are 58

63 Regulatory loop limited by both technical dissection skills and time-dependent measurements in the same individual in the smallest vertebrates on earth at hatching is therefore an ambitious mission. To complicate matters, natural and ontogenetic variations among individuals in fish larvae can be quite substantial and increase with age, making it difficult to draw precise conclusions from endocrinological studies. It is also possible that the anaesthesia had an influence on the organism as has previously been shown in a reduced digestive response in anesthetized zebrafish (HAMA et al., 2009) and rats (KORDÁS et al., 2000). Since CCK has been found predominantly in the head in larval herring (86%, ROJAS-GARCÍA et al., 2011) and sea bass, Dicentrarchus labrax (83%, TILLNER et al., unpubl. data) and its role in the brain of larval teleosts is largely unknown (RØNNESTAD et al., 2007), the heads of all larvae in our study were removed from the analyses. However, we must point out that no distinction between synthesized CCK in the gut and secreted CCK into the body fluids could be made, since whole-body homogenates (excluding head) were used for the analyses. Interestingly, both CCK and tryptic enzyme activity displayed a fluctuating pattern in untreated larvae over the sampling period, possibly due to an endogenous daily rhythm which has also been observed in mammals (LIDDLE, 2006). Slightly higher CCK values at 4h and 8h in the saline-injected (control) group may have been caused by a mechanical stimulation via stretch receptors in the gut due to the injected saline solution, since the 9.2 nl injection volume resulted in a slightly distended gut (visual observation). However, no clear response in terms of tryptic enzyme activity was observed in this group. Distension could also explain the results observed in onemonth-old larval herring, in which ad libitum ingested non-biodegradable polystyrene spheres triggered the release of trypsinogen from the pancreas, suggesting a mechanical stimulation (PEDERSEN et al., 1992). By contrast, KOVEN et al. (2002) found no stimulation of CCK and tryptic enzyme activity at an earlier developmental stage in the same species after tube-feeding physiological saline, which suggests a more mature role of a neural component in controlling digestive processes in older larval herring. Surprisingly, the administration of a presumed physiological dose of CCK resulted in low levels of tryptic enzyme activity within one hour of tube-feeding, neither could we detect an increase in CCK levels in these larvae. Since CCK is also responsible for gut motility (GRIDER, 1994), increased movements of the gut and thus the evacuation of intraluminal CCK and trypsin after stimulation by exogenous CCK might have led to low levels of both factors shortly after injection. This remains a speculative 59

64 Chapter II statement. Despite this, the clear rise in both factors from 2h after injection onwards (except for CCK at 8h) indicates that CCK plays a role in the release of pancreatic enzymes. In comparison with larval herring tube-fed liposomes containing bovine serum albumin as a protein source, a clear rise both in CCK and tryptic enzyme activity was visible as early as 15 min. after administration (KOVEN et al., 2002). This suggests differences between species with a straight (herring) and a coiled (cod) gut with regard to the reaction time of the digestive system in response to an injection. Nevertheless, since KOVEN et al. (2002) used whole-body homogenates including the heads, it cannot be ruled out that a significant fraction of CCK was of neural nature from the brain. Atlantic halibut larvae had a lower gut to body ratio (excluding the head) of CCK 4h after a nutrient stimulus, indicating that CCK had been released into the body fluids (ROJAS-GARCÍA et al., 2002). The same effect was induced in larval herring as early as 30 min. after feeding with Artemia (ROJAS-GARCÍA et al., 2011), an effect that may have been masked in our study. The fact that tryptic enzyme activity remained at relatively high levels up to eight hours after tube-feeding implies that trypsin has a long retention time in the gut. In addition, low levels of CCK at 8h suggest a negative feedback loop between these factors. In mammals, a CCKreleasing factor is secreted spontaneously into the gut lumen and is degraded by high levels of trypsin, preventing further release of CCK in this regulatory loop (LIDDLE, 2006) which could explain the lower values of CCK observed in our study. In comparison, plasma CCK levels were elevated 4-6h after a meal in adult rainbow trout, Oncorhynchus mykiss (JÖNSSON et al., 2006). In addition, CCK mrna was more expressed 1h after a nutrient stimulus in adult yellowtail, Seriola quinqueradiata (MURASHITA et al., 2008). It is likely that tryptic enzyme activity increases in a dosedependent manner with increasing concentrations of injected CCK, as was concluded from results in adult fish (EINARSSON et al., 1997), a result that needs to be confirmed for the larval stages of marine species. The administration of a CCK antagonist did not display a consistent pattern. The fact that tryptic enzyme activity was higher at 2h and 4h compared to both control groups did not correspond to low CCK levels, possibly bringing the postulated negative feedback control into question. There is a long tradition of studies of the feedback mechanism between CCK and trypsin by the use of CCK antagonists in mammals, but not in fish. However, studies of adult sea bass (RUBIO et al., 2008) and rainbow trout (GÉLINEAU et al., 2001) demonstrated that the administration of a CCK antagonist increased feed intake in these fish, supporting the same anorexigenic effect, e.g. that it reduces appetite and food consumption, as in mammals. In 60

65 Regulatory loop experiments on rats, the response of endogenous CCK after a nutrient stimulus was abolished by the presence of a CCK antagonist (EASTWOOD et al., 1998). The concentration administered in the present study may therefore have been too low to give a clear effect. However, future studies of endocrine regulation of digestive processes in fish larvae ought to take advantage of the highly selective receptor antagonists that are available for a range of hormones (e.g. RUBIO et al., 2008). Used in conjunction with a nutrient stimulus, these antagonists might confirm the involvement of various hormones in digestive processes. The plant protein phytohemagglutinin (PHA) was administered to stimulate CCK release and increased trypsinogen secretion into the gut, thus bringing about a rise in tryptic enzyme activity. Indeed, the development of CCK and tryptic enzyme activity was similar to that in the group injected with standard CCK, indicating that PHA has a stimulatory effect on the digestive system of cod larvae at 26 dpff. In our study, PHA extracted from red kidney beans, Phaseolus vulgaris, was applied in a very low concentration, equivalent to % of larval wet weight, which was at minimum ten times lower than concentrations in other studies. Since PHA can be part of a diet in livestock nutrition, e.g. for young piglets, most studies of its effects have been performed on these animals or other mammalian models. To the best of our knowledge, the only study on fish was that of DROSSOU et al. (2006) in which tilapia fry, Oreochromis niloticus, were fed a dry diet supplemented with PHA. The concentration used by DROSSOU et al. (2006) may have been too high, since tryptic enzyme activity was lower in the group supplemented with PHA. In suckling piglets, PHA was found to stimulate CCK and pancreatic secretion after a three-day exposure (EVILEVITCH et al., 2005) as well as after short-term exposure in adult rats (GRANT et al., 2000; KORDÁS et al., 2000). PHA might therefore be added in known concentrations as a trigger substance in practical diets to facilitate digestion in suckling piglets during weaning which suffer from diarrhoea due to an immature digestive system (LINDEROTH et al., 2005; THOMSSON et al., 2007). The effect is thus possibly mediated by PHA binding to the gut epithelial cells, leading to increased growth of the gut and pancreatic hypersecretion (HERZIG et al., 1997). However, given in excess, severe inflammation of the gut may appear (OTTE et al., 2001). The administration of a single dose of PHA initiated a response of the digestive system in larval cod in our study, which makes PHA a suitable candidate for a stimulatory feed additive in microdiets. Nonetheless, long-term effects and critical upper and lower limits should be investigated in future studies, possibly using additional response variables such as gut histology. 61

66 Chapter II The reaction of CCK in the group injected with a trypsin inhibitor was similar to that of the group injected with saline, while the development of tryptic enzyme activity was slightly lower. The falling trend in tryptic enzyme activity in this group indicates an inhibitory effect of the trypsin inhibitor in cod larvae. Nevertheless, the relatively low concentrations we used may not have been sufficient to reduce tryptic enzyme activity below a threshold to provoke a negative feedback response in CCK. In theory, these levels of tryptic enzyme activity might be sufficient to degrade a possible CCK-releasing factor which in return hampers the release of CCK. In halibut larvae, bovine serum albumin induced CCK release from the gut into the body fluids while a trypsin inhibitor did not, probably also due to the relatively low concentrations administered (ROJAS-GARCÍA et al., 2002). Moreover, tryptic enzyme activity was not measured in that experiment, making a comparison with the present study difficult. In a feeding study with larvae of spotted wolfish, Anarhichas minor, the addition of a soybean trypsin inhibitor clearly reduced survival, growth and digestive enzyme activity at 60 days post-hatching (SAVOIE et al., 2011). Interestingly, the group fed the trypsin inhibitor administered together with a moderate level of protein hydrolysate displayed better growth and survival than that given protein hydrolysate alone. This could be explained by an inhibition of trypsin in the gut, resulting in an overcompensation of pancreatic enzyme secretion and enhanced growth (SAVOIE et al., 2011). It therefore remains an open question to what extent gastrointestinal hormones, such as CCK and digestive enzymes other than trypsin are modulated in the long term by different levels of trypsin inhibitors in marine fish larvae. On the other hand, the effect of trypsin inhibitors via CCK in the regulatory loop resulting in increased pancreatic enzyme secretion and growth, is a well-understood mechanism in mammals (FRIESS et al., 1998; GUO et al., 2007; JORDINSON et al., 1996). In conclusion, this study supports the concept that a feedback mechanism exists between gut CCK and tryptic enzyme activity in Atlantic cod larvae. An endogenous rhythm of both factors under basal conditions appears to be present as has been found in mammals. Mechanical stimulation via distension, sensed by stretch receptors of the gut, could be a contributing signal to release trypsinogen but the administration of exogenous CCK clearly induces a response in tryptic enzyme activity which also indicates a role for CCK. The plant protein phytohemagglutinin acts as a trigger substance for facilitating proteolytic digestion in larval cod. The detailed temporal and spatial dynamics of the digestive process including minimum 62

67 Regulatory loop and maximum levels of CCK and tryptic enzyme activity in the regulatory loop remain unknown and should form the focus of further studies. Acknowledgements This study was financially supported by the German Federal State of Schleswig- Holstein and the European Regional Development Fund (ERDF) through the NEMO project ( ), and was supported by COST Action FA0801 LARVANET (to R.Tillner), ISOS Integrated School of Ocean Sciences (to R. Tillner), Research Council of Norway (proj no ; GutFeeling to I. Rønnestad and T. Harboe; ; CODE to I. Rønnestad). The authors thank the technical staff at the Austevoll Research Station for technical assistance as well as Dr. Nils Roos and Frauke Repening from the Max Rubner-Institut Kiel for analytical support during the CCK analyses. 63

68 Chapter II 64

69 Chapter III Chapter III 23 days old sea bass larvae fed Artemia (upper larva) or a microdiet (lower larva) 65

70 Chapter III The regulatory loop between gut cholecystokinin and tryptic enzyme activity in sea bass (Dicentrarchus labrax) larvae is influenced by different feeding regimes and trigger substances Abstract The inability of most marine fish larvae to efficiently utilize microdiets from firstfeeding can be explained by an immature digestive system including an undeveloped hormonal control of digestive functions. Improving the understanding of digestive physiology in first-feeding larvae is therefore a prerequisite to improve diet formulations and feeding protocols. In marine fish larvae that lack a stomach at firstfeeding trypsin represents the main proteolytic enzyme. CCK is one of the key regulators of digestive enzyme secretion in adult vertebrates, and there are data that suggest that it is also involved in early stages of teleosts, although this may vary between species. Here, we investigated the influence of Artemia and a commercial microdiet on the ontogenetic development of tryptic enzyme activity as an indicator for digestive capacity in first-feeding sea bass. In order to examine the regulation and feedback mechanisms in the digestive tract we followed the response of gut CCK and tryptic enzyme activity during a one-day observation depending on the feeding regime at 23 days post hatch. Larvae fed the microdiet showed a higher tryptic enzyme activity, probably as an adaptation to the higher content in complex protein in the diet. The plant protein phytohemagglutinin (PHA), added to the microdiet as a potential stimulator for the digestive system, did not induce elevated tryptic enzyme activity nor was it beneficial for growth. This was possibly due to adverse effects of too high doses. We observed an endogenous rhythm of CCK over the day, independent of the dietary treatment or short-term fasting. Higher tryptic enzyme activity in larvae fed Artemia during the day might indicate a better stimulation by live prey in the digestive tract or the superiority of a discontinuous feeding schedule in this group. We suggest that a reduction in tryptic enzyme activity after several feeding events indicates a limit in diurnal digestive capacity. Sea bass larvae are apparently able to adapt to the feeding schedule by synchronizing the tryptic enzyme activity like adult fish. Keywords: trypsin, CCK, digestion, ontogeny, endocrine control, microdiet 66

71 Feeding regimes Introduction The unstable and unpredictable production of juveniles of many marine fish species in aquaculture still prevents successful commercialization of several candidate species. One reason is the lack of knowledge of the function and efficiency of the digestive processes in the early stages hampering the formulation of suitable feeds and feeding regimes. Over the years an increasing knowledge has been gathered of how the tissue and organs of the digestive system develop during larval ontogeny. Accordingly, there is an ample literature on the ontogeny of the digestive enzymes in many fish species (both reviewed by LAZO et al., 2011). Nevertheless, there are still significant gaps in the understanding of how digestive processes are controlled in fish larvae compared to adult fish and mammals (RØNNESTAD et al., 2013; WEBB et al., 2011). Although there is growing information on fundamental processes in controlling digestive functions (e.g. KOVEN et al., 2002; ROJAS-GARCÍA et al., 2011) this rarely translates into practical implications for feed formulations or feeding regimes. However, the recent successful adaptation from a continuous to a meal based feeding regime in Atlantic halibut that significantly reduced malformations was hypothesized to be linked to gut transit time and lack of a fully functional gut-brainaxis signalling pathway for digestion and appetite (HARBOE et al., 2009). Current feeding of most marine fish larvae in aquaculture is still dependent on live feeds at least during the first-feeding stages (CONCEIÇÃO et al., 2010b). Nevertheless, commonly used organisms such as rotifers (Brachionus spp.) and brine shrimp (Artemia spp.) rely on time-, labour-consuming and costly rearing techniques, are unstable in quantity and nutritional quality and may represent vectors for diseases. Consequently, progress in food formulation, feeding techniques and a growing understanding of nutritional requirements in recent years have resulted in increasing substitution of live feeds with microdiets (HOLT et al., 2011). Nonetheless, only a few marine species can be reared exclusively on microdiets from first-feeding, such as European sea bass, Dicentrarchus labrax (CAHU et al., 1998a). Nevertheless, to date, the successful use of microdiets from first-feeding is still limited by constraints such as nutritional composition, nutrient leaching and digestibility, high sinking rates and low acceptability as well as tank hygiene. In addition, unadapted zootechnical parameters such as light regimes, water management, feeding strategy etc. may mask the unsuccessful application of microdiets from first-feeding. It has been proposed that the absence of a functional stomach and acidic digestion in marine fish larvae before metamorphosis sets a major hurdle that limits the use of 67

72 Chapter III microdiets at first-feeding (KOLKOVSKI, 2001). However, growing success in early weaning onto microdiets is apparently not only dependent on the presence of a functional stomach, meaning that marine fish larvae might have sufficient digestive capacity even early in development (MUGUET et al., 2011). Nevertheless, although fish larvae possess an effective battery of alkaline enzymes secreted from the pancreas for digestion in the intestine (LAZO et al., 2011), low levels of these enzymes have been observed in several species during ontogeny (RIBEIRO et al., 1999; SUZER et al., 2007a; UEBERSCHÄR, 1995). Consequently, low digestive capacity in the premature developmental stages may contribute to the overall high mortality usually observed in marine fish larvae after first-feeding (SAVOIE et al., 2006; UEBERSCHÄR, 1995). Among the pancreatic enzymes, trypsin is considered to be a key enzyme in the digestive process as it is the most significant protease in the early larval stages (ZAMBONINO INFANTE et al., 2001). Following the ingestion of feed, trypsin is secreted as its inactive precursor, trypsinogen, from the acinar cells of the pancreas into the gut lumen and either auto-activated or activated by the enzyme enteropeptidase. In general, in marine fish larvae the amount of tryptic enzyme activity in the gut has been demonstrated to be a function of feed ingestion, gut fullness and nutrient composition (RØNNESTAD et al., 2008; UEBERSCHÄR, 1995). Along the numerous hormones in the gastrointestinal tract, CCK is considered one of the most important stimulators of pancreatic enzyme secretion in mammals (LIDDLE, 2006). It has long been studied and is known to play a vital role in contraction of the gallbladder, peristalsis in the intestine, delay of gastric emptying and pancreatic enzyme secretion in mammals (CRAWLEY et al., 1994) and adult fish (EINARSSON et al., 1997). In addition, it acts as a satiation signal in the fish brain (VOLKOFF et al., 2005). Upon the presence of nutrients in the gut, CCK is released from enteroendocrine cells in the gut epithelium into the body fluids and acts on target cells in the pancreas to trigger the secretion of digestive precursors of digestive enzymes into the gut lumen. High tryptic enzyme activity in the gut acts as a negative feedback control for the release of CCK, suggesting a regulatory loop between these two factors in mammals (LIDDLE, 2006). Recently, a similar mechanism has been described in adult fish (MURASHITA et al., 2008). Although CCK-producing cells have been located in different areas in the larval gut of several fish species, e.g. in Atlantic cod, Gadus morhua (HARTVIKSEN et al., 2009), little is known about the regulatory mechanism between CCK and trypsin in developing 68

73 Feeding regimes fish larvae compared to mammals and adult fish. Nevertheless, the number of studies focusing on the interaction between these two factors in marine fish larvae has started to grow and these studies suggest that CCK might also play an important role in the release of pancreatic enzymes in developing fish larvae (ROJAS-GARCÍA et al., 2002) although there appear to be some differences between species when this regulation becomes fully functional (HAMA et al., 2009; ROJAS-GARCÍA et al., 2011; WEBB et al., 2011). Besides, it has been shown in mammals that certain food components and digestive end products, like intact protein or certain amino acids, stimulate CCK and consequently pancreatic enzyme secretion more than others (WANG et al., 2011). Results of tube-feeding studies (KOVEN et al., 2002) as well as standard feeding trials (CAHU et al., 2004; NAZ et al., 2009) suggest that similar mechanisms are present in early larval stages of fish. Low amounts of these stimulatory components in commercial microdiets as well as technical constraints in the manufacturing process to effectively release these components in the gut may contribute to the inability of most marine fish larvae to utilize these diets efficiently from first-feeding (YÚFERA et al., 2000). The plant protein phytohemagglutinin (PHA) was found to stimulate the growth of the pancreas, small intestine and consequently the release of CCK in rats and piglets in several studies (EVILEVITCH et al., 2005; HERZIG et al., 1997; JORDINSON et al., 1997; KORDÁS et al., 2000; LINDEROTH et al., 2005; OTTE et al., 2001; RADBERG et al., 2001; THOMSSON et al., 2007). Following these results it has been supposed that PHA might have a stimulatory effect on CCK in the gut and consequently on trypsinogen synthesis and release into the gut of larval fish (DROSSOU et al., 2006). In this study we examine the ontogenetic development of tryptic enzyme activity in larval sea bass fed either live prey (Artemia spp.), a commercial microdiet (Gemma Micro, Skretting, France) or co-fed a combination of both from first-feeding. Since tryptic enzyme activity is one of several indices for the maturation of the digestive system, we propose that this parameter differs during ontogeny depending on the diet. Two additional treatments were applied to larvae fed Gemma Micro supplemented with two concentrations of PHA to examine the long-term effect of PHA as a stimulator of the digestive system. In addition, we conducted a one-day observation at 23 days post hatch (dph) to evaluate the diurnal interaction between CCK and tryptic enzyme activity. It is suggested that both CCK and tryptic enzyme activity may be influenced by the dietary treatments. 69

74 Chapter III Larvae of European sea bass were used as a model in this study, since this species readily accepts microdiets from first-feeding which is an important prerequisite for the investigations on the effect of trigger substances from first-feeding. Apart from disclosing the postulated mechanisms of CCK and tryptic enzyme activity, this provides valuable information on the diurnal digestive capacity in sea bass larvae related to different diets and practical feeding conditions. Materials and Methods Larval rearing Sea bass larvae (3 days post hatch, dph) were transported from a commercial hatchery in France (Ecloserie Marine de Gravelines) to the experimental facilities in Kiel, Germany. The larvae were equally distributed to 15 green, conical tanks (65 L water volume each) to establish a density of around 150 larvae L -1 and kept in darkness until yolk sac absorption and mouth opening at 7 dph. Each tank was equipped with a peripheral water inlet to establish a gentle, circular current and a water outlet at the central tank bottom covered with mesh (100 µm). Coarse bubbling was provided close to the water outlet and surface skimmers were installed to facilitate swim bladder inflation. Water flow was gradually increased from 0.2 L min -1 to 0.5 L min -1 until the end of the experimental period (35 dph). Salinity was gradually reduced from 35 g L -1 to 26 g L -1 between 3 14 dph and increased again to full strength afterwards. Water temperature was gradually increased from 14 C to 17 C and oxygen was maintained above 80% saturation throughout the experiment. Toxic metabolites were below critical limits (ammonia: <0.02 mg L -1, nitrite: <0.01 mg L -1 ) over the experimental period. Light was provided 16 hours a day (7h00 23h00) and light intensity was increased from 50 to 700 lux at the water surface during the experiment. Dead larvae and debris were removed daily by siphoning the tank bottom. On 8 dph, three tanks were randomly assigned to each of the following feeding regimes: (1) Artemia, (2) Artemia + Gemma Micro (Skretting, France), (3) Gemma Micro, (4) Gemma Micro incl. 0.01% PHA, (5) Gemma Micro incl. 0.02% PHA. PHA (L8754, Sigma-Aldrich) was incorporated during the manufacturing process. Each tank received 1 ml microalgae Nannochloropsis sp. concentrate (BlueBioTech, Germany) at 8h00 and 16h00, respectively until 19 dph. Artemia instar I nauplii (AF 430 Artemia, INVE, Belgium) until 26 dph and enriched (S.presso, INVE, Belgium, 70

75 Feeding regimes according to the supplier) Artemia instar II nauplii from 25 dph were fed three times daily at 9h00, 14h00 and 19h00 (feeding ratio 2:1:1) in increasing densities between ml -1 during the experiment. Gemma Micro 75 (75 µm particle size) was provided hourly in excess between 8h00 22h00 by hand and automated feeders (Eheim, Germany) until 14 dph and gradually substituted by Gemma Micro 150 from 12 dph. Gemma Micro was provided 60 and 30 min. before the administration of Artemia in the Artemia + Gemma Micro group (Fig. III-I). Five larvae were sampled randomly each sampling day from each tank prior to feeding in the morning using a pipette with a large opening. The larvae were transferred with minimum of seawater into an Eppendorf vial and then immediately frozen on dry ice and subsequently stored at -80 C until analysis. Short-term diurnal rhythm experiment A 17 hour experiment was conducted at 23 dph. A sufficient number of larvae from the Artemia and Gemma Micro treatment, respectively were transferred to two tanks at 8h00 and left unfed for the whole day ( ArtS and GMS ). Five larvae were sampled from each tank including the main holding tanks every full hour between 6h00 and 23h00 as quickly as possible as described above and stored at -80 C until analysis. The feeding events in the holding tanks were delayed until the respective sampling was finished (e.g. 14h00 Artemia feeding after 14h00 sampling). Figure III-I: Feeding regime of the dietary treatments during ontogeny (left side) and during the diurnal rhythm at 23 days post hatch (right side) designated by bars for each treatment. Faded areas indicate the transition from one feed to another and increasing bars in the Artemia treatments indicate increasing feeding doses. Arrows indicate the feeding times of Artemia with bigger arrows indicating a higher feeding dose. GM 50, GM 150, instar I and instar II specify the particle size of Gemma Micro and the naupliar stage of the offered Artemia, respectively. 71

76 Chapter III Sample preparation Individual samples were analyzed for CCK and tryptic enzyme activity using a combination of highly sensitive methods according to ROJAS-GARCÍA et al. (2001) and UEBERSCHÄR (1993) which allows a resolution of both factors at the individual level. Described briefly, frozen samples were allowed to thaw on ice, rinsed with distilled water and the standard length (mm, tip of upper jaw to end of notochord) was measured on an ice-cold Petri dish under a microscope. Gut fullness was evaluated, using a simple gut fullness index after ROJAS-GARCÍA et al. (2011): 0% (empty), <25%, 25-50%, 50-75%, % (full). Previous studies have shown that a relatively large amount of CCK can be found in the head part (central nervous system mainly) which may mask changes of CCK in the gastrointestinal tract (ROJAS-GARCÍA et al., 2011). Therefore, all analyses in the present study were done on dissected larvae excluding the head. Each larva was then transferred to an individual Eppendorf vial and homogenized in 50 µl ice-cold distilled water using a motorized pestle. For extraction of CCK, 750 µl methanol were added, the sample was vortex-mixed thoroughly and incubated on ice for 30 min. After centrifugation (15 min., 1700 g, 4 C) each sample was split in two by transferring the supernatant to a new Eppendorf vial. Both, the remaining pellet (methanol-insoluble fish precipitate) and the supernatant (CCK methanol extract) were evaporated to dryness using a vacuum desiccator attached to a waterjet pump and stored at -20 C until analysis. Analysis of CCK and tryptic enzyme activity The individual CCK extracts were assayed by a competitive radioimmunoassay (RIA) using CCK-RIA kits (RB302, Euro-Diagnostika, Sweden) according to the supplier s instructions and ROJAS-GARCÍA et al. (2001). CCK levels were interpolated from a standard curve ( pmol CCK L -1 ) and concentrations are expressed as fmol larva -1. Recovery of known amounts of CCK added to samples throughout the extraction procedure was 71%. Tryptic enzyme activity in individual pellets was measured using a highly specific fluorescence substrate (Nα-benzoyl-L-arginine-methyl-coumarinyl-7-amide-HCl) according to UEBERSCHÄR (1993). Values are expressed as hydrolysed fluorescence products MCA (methyl-coumarinyl-7-amide, nmol MCA min -1 larva -1 ). The coefficient of variation between triplicate measurements of samples was 1.6% (n = 4 samples). 72

77 Feeding regimes Statistical analysis CCK concentrations and tryptic enzyme activity levels during ontogenetic development were averaged for length classes in steps of 0.5 mm ( , etc.). Data on standard length during ontogeny as well as CCK and tryptic enzyme activity values in the diurnal rhythm experiment were tested for normality and homogeneity of variance using the Shapiro-Wilks-test and Levene s test, respectively and log-transformed, if necessary. Values of gut fullness in the diurnal rhythm experiment were transformed using the formula gut fullness =arcsin gut fullness. Significant differences in standard length between feeding treatments during ontogeny were analyzed using a nested One-way ANOVA with measurements of individuals in each tank nested in treatment groups followed by Tukey-HSD post hoc tests. Significant differences in CCK concentration and tryptic enzyme activity between feeding treatments in the diurnal rhythm experiment were analyzed using a nested One-way ANCOVA with measurements of individuals in each tank nested in treatment groups for each sampling point and standard length as a covariate. Upon significance, differences between groups were evaluated using multi-comparison tests with Bonferroni correction. Statistics were performed with SPSS 19.0 for Windows and the level of significance was set to p < Data are presented as mean ± S.D. Results Ontogenetic development Significant differences in standard length during ontogeny were not evident before 15 dph (Tab. III-I). Both groups fed either Artemia ( Art ) or co-fed Artemia and Gemma Micro ( ArtGM ) grew bigger compared to the three groups fed solely on Gemma Micro. In addition, larvae fed solely on Artemia showed the highest standard length values among all groups after 19 dph until the end of the experimental period. Among the three Gemma Micro groups, the inclusion of 0.01% PHA led to higher growth at 13 and 15 dph, whereas the Gemma Micro group without PHA revealed a higher growth from 15 dph until the end of the experimental period. The inclusion of 0.02% PHA led to the lowest values in standard length after 13 dph among the three Gemma Micro groups, except at 21 dph (Tab. III-I). 73

78 Chapter III Table III-I: Standard length (mm) of sea bass larvae during ontogenetic development 9-31 days post-hatch (dph). Different superscripts within the same row represent significant differences between feeding treatments (nested One-way ANOVA, Tukey-HSD, p < 0.05). Data are presented as mean ± standard deviation (n = 3 tanks á 5 larvae). dph Art ArtGM GM PHA0.01% PHA0.02% ± 0.04 a 4.84 ± 0.08 a 4.70 ± 0.02 a 4.77 ± 0.05 a 4.75 ± 0.05 a ± 0.06 a 4.76 ± 0.03 a 4.74 ± 0.04 a 4.73 ± 0.15 a 4.81 ± 0.09 a ± 0.08 a 5.38 ± 0.41 a 5.11 ± 0.13 a 5.44 ± 0.13 a 5.14 ± 0.16 a ± 0.16 a 5.97 ± 0.35 a 5.46 ± 0.29 b 5.59 ± 0.05 ab 5.40 ± 0.32 b ± 0.65 a 6.63 ± 0.02 a 5.89 ± 0.16 b 5.90 ± 0.20 b 5.73 ± 0.14 b ± 0.22 a 7.37 ± 0.37 a 6.40 ± 0.38 b 6.26 ± 0.40 b 6.16 ± 0.22 b ± 0.14 a 7.45 ± 0.16 ab 7.03 ± 0.31 abc 6.48 ± 0.16 c 6.83 ± 0.15 bc ± 0.09 a 8.19 ± 0.52 ab 7.49 ± 0.15 bc 7.41 ± 0.35 bc 7.30 ± 0.41 c ± 0.58 a 9.09 ± 0.20 a 8.23 ± 0.54 b 7.63 ± 0.52 b 7.53 ± 0.13 b ± 0.30 a 9.08 ± 0.38 a 8.88 ± 0.36 a 8.50 ± 0.69 a 8.27 ± 0.47 a ± 0.59 a 9.63 ± 0.01 ab 9.08 ± 0.42 bc 8.87 ± 0.15 c 8.67 ± 0.28 c Feeding treatments: (Art) Artemia, (ArtGM) Artemia + Gemma Micro, (GM) Gemma Micro, (PHA0.01%) Gemma Micro incl. 0.01% phytohemagglutinin (PHA), (PHA0.02%) Gemma Micro incl. 0.02% PHA All groups revealed an increasing trend in tryptic enzyme activity during ontogeny with increasing individual variability with increasing size (Fig. III-II). The tryptic enzyme activity tended to be higher in larvae fed Artemia or Artemia and Gemma Micro in larvae smaller than 6.5 mm. Between 7.0 and 8.5 mm SL the tryptic enzyme activity was highest in the groups fed Gemma Micro among all groups followed by larvae fed Gemma Micro incl. 0.01% PHA. These differences levelled out in larger larvae. Among the three Gemma Micro groups, the inclusion of PHA led to a reduction in tryptic enzyme activity which was most evident in size classes between 7.0 and 8.5 mm. 74

79 Feeding regimes Figure III-II: Tryptic enzyme activity (hydrolysed substrate, nmol MCA min 1 larva 1 ) for length classes (standard length, 0.5 mm precision, , etc.) of larval sea bass (excluding head). Feeding treatments: (Art) Artemia, (ArtGM) Artemia + Gemma Micro, (GM) Gemma Micro, (PHA0.01%) Gemma Micro incl. 0.01% phytohemagglutinin (PHA), (PHA0.02%) Gemma Micro incl. 0.02% PHA. Data are presented as mean ± standard deviation (1 33 larvae per length class per treatment). Short-term diurnal rhythm experiment The administration of food was clearly reflected in an increasing mean gut fullness in all groups in the diurnal rhythm experiment (Fig. III-III). However, in both starving groups ( GMS & ArtS ) during the whole sampling period and in fed groups before first feed administration in the morning there were some remaining digesta in the gut lumen. The mean gut content in the ArtGM group (70 ± 28%) was lower at 10h00 compared to the Art group (93 ± 12%). In addition, a trend in reduced gut fullness in the ArtGM group after 18h00 was obvious and the administration of Gemma Micro in the morning at 8h00 was not clearly reflected in increased gut fullness at 9h00. Apparently, the filling of the gut was lower and increased progressively in time in all three Gemma Micro groups to reach a plateau at 17h00 with minimum 90% mean gut fullness in all three groups. 75

80 Chapter III Figure III-III: Relative gut fullness (%) in larval sea bass between 6h00 and 23h00 at 23 dph. Feeding treatments: (Art) Artemia, (ArtGM) Artemia + Gemma Micro, (GM) Gemma Micro, (PHA0.01%) Gemma Micro incl. 0.01% phytohemagglutinin (PHA), (PHA0.02%) Gemma Micro incl. 0.02% PHA. Starving treatments: (ArtS) deriving from Art, (GMS) deriving from GM. Grey bars represent the feeding period of Gemma Micro every full hour, grey arrows the administration of Artemia nauplii. The feeding events were delayed until the respective sampling was finished (e.g. 14h00 Artemia feeding after 14h00 sampling). Data are represented as mean + standard deviation ( ArtS and GMS n = 5 individuals, all feeding treatments n = 3 tanks). 76