Characterization of the gastro-entero-pancreatic system of Osteoglossomorpha: an immunohistochemical, immunocytochemical, and molecular study

Azza Abdul Fattah Al-Mahrouki

A thesis submitted in conformity with the requirements for the degree of Doctor of Philosophy in the Department of Zoology University of toronto

@ a Al-Mahrouki, 2001 National Library Bibliothèque nationale 1*1 of Canada du Canada Acquisitions and Acquisitions et Bibliographie Services services bibliographiques 395 Wellington Street 395. nie Wellington Ottawa ON KI A ON4 Ottawa ON KIA ON4 Canada Canada vow& Yom réienww

Our fi& Notre réftirence

The author has granted a non- L'auteur a accordé une licence non exclusive licence allowing the exclusive permettant à la National Lïbrary of Canada to Bibliothèque nationale du Canada de reproduce, loan, distri'bute or sell reproduire, prêter, distribuer ou copies of this thesis in microform, vendre des copies de cette thèse sous paper or electronic formats. la forme de microfiche/nlm, de reproduction sur papier ou sur format électronique.

The author retains ownership of the L'auteur conserve la propriété du copyright in thesis. Neither the droit d'auteur qui protège cette thèse. thesis nor substantial extracts fiom it Ni la thèse ni des extraits substantiels may be printed or otherwise de celle-ci ne doivent être imprimés reproduced without the author's ou autrement reproduits sans son permission. autorisation- Abstract

Characterization of the gastro-entero-pancreatic system of Osteoglossomorpha: An immunohistochemical, immunocytochemical and molecular study. Doctor of Philosophy,

2001- Azza Abdul Fattah Al-Mahrouki. Department of Zoology. University of Toronto.

The principal objectives of this study were twofold: fjrst, to compare the distrr'bution and structure of the GEP system of the osteoglossomorphs with that of the other teleosts and with vertebrates in general; second, to examine the possibility of using the primary structure of the GEP system peptides in a phylogenetic analysis. To fulnU these objectives several approaches were use& including histolo gy, immunohisto - and immunocyto chernistry, and mo lecular cloning. The results indicated that an islet organ (endocrine pancreas) with aelyscattered islets consists of four cell types (A, B, D and F); there are two D-ceU subtypes. Endocrine ce& which produce some of the same peptides as the islet cells were also observed within the intestinal epithelia; some variation occurs among the species. For exaziiple, insulin was detected in ceils wIthin the intestinal and stomach epithelia of Pantodon bucholzi, and Sequently in the intestinal epithelia of Osteoglossum biciwhosum and Gnarhonernuspetersii, but not in the other species. In the molecular biology studies, preproinsulin and the preprosomatostatin sequences were deduced fiom the corresponding cDNA sequences and were used in the phylogenetic aflSL].ysis.The phylogenetic analysis of the preproinsulin sequences confïrmed the monophyletic grouping of the osteoglossomorphs, and also indicated that the osteoglossomorphs are not the most basal living teleosts as previously thought, but rnay be more generalized teleosts. In addition, the phylogenetic analysis showed the elephantnose to be closely related to the feather fh kdefis4 and the arawana to be closely related to the butterfly fïsh. These data are consistent with previous reports (Lamier and Liem 1983; Li and Wilson 1996). The intragroup relationship was cobedby phylogenetic analysis using sequences of cDNA of preprosomatostatia These resdts indicated that the primary structure of the GEP system peptides is usefül for inferring phylogenetic analysis to some extent, but also that the sequence analysis of more than one peptide should be considered in such analyses. This study suggests that there is a generalized position of the GEP system in cornparison with that of the other teleosts. Acknowledgments

No matter how independent 1 like to think myseK 1couid not have completed the

Doctor of Philosophy's degree thesis without the help of mmy other people to whom 1owe at Ieast a debt of gratitude.

1would hrst like to mention my advisors, Professor John Youson, Professor David

Butler, and Associate Professor David Irwin. Through the years, Prof. Youson has provided sound advice at appropriate crossroads. His enthusiasm, scient& insight, and confidence has contriibuted immensely to my research, and to the completion my thesis. Thank you for giving me the research opportunity and for the research fieedom that promoted my self confidence and autonomy. 1thank Prof. Butler for bis advice, guidance, and his very fhitful discussions on the philosophicai views of the research; it is so important to bandy ideas with more experienced researchers. 1 am also thankful to Dr. Irwin for his advice and constructive feedback in the field of molecular biology: hi, help with the editing of the last two chapters of the thesis has ken invaluable. 1 am also grateful for his help with the production of the phylogentic trees in this thesis.

1 &O want to aclmowledge Dr. E. Plketskaya, Dr. P. Andrews, and Dr. C. Yip for providing me with some of the antisera used in this research. 1 am especialiy gratefùi to Dr.

L. Graham for providing me with the goldeye tissues, which otherwise have been very hard to obtain 1 also want to acknowledge Dr. M. Wilson for allowing me to use the cladograrn fkom his publication in the general introduction to this thesis.

1 would iike to thank Raymond Or for teaching me the use of the electron microscopy, his technical help is very much appreciated. The graphic expertise of Ken Jones helped in the final production of the photogmphs in the thesis, and is also very much appreciated.

My coileagues at the Division of Life Sciences are all acknowledged for their advice on Metent research issues and for helping me in many different ways. Thank you to everyone, 1am extremely gratefûl.

1 also owe special thanks to my family, to whom 1 would Like to dedicate Siis thesis - to, not only for their moral support, but for bearing with me throughout this journey and creating a very positive atmosphere, making it easier to withstand the difficulties that sometimes arose. To my parents, Nabilla and Abdul Fattah Al-Mahrouki, and to my children, Ahmad and Areag Osman: thanks for helping me in every way you could.

This study was supported by a grant ftom the Naturai Sciences and Engineering

Research Council of Canada to Dr. John H. Youson. Abbreviations

APUD amine precursor uptake and decarboxylation aPY anglerfish polypeptide tyrosine

CAS casein

CODEHOP Consensus-Degenerat e Hybrid Oligonucleo tide Primers

GEP gastro-entero-pancreatic system

GLP glucagon-like peptide

GLU glucagon

INS insulin mINS mammalianinsulin

WY neuropeptide tyrosine

0s04 osmium tetroxide

PAUP phylogenetic analysis using parsimony

PBS phosphate buffered saline

PCR polymerase chain reactiom

PYY peptide tyrosine tyrosine

RACE rapid amplification of cDN-4 ends

RNA riinucleic acid

SST somatostatin

SST-14 synthetic invariant somato statin- 14

SST-25 sahon somatostatin-25

SST-34 iamprey somatostatin-34 Table of Contents

Abstract ii

Aclmowledgements iii

Abbreviations v

List of figures ix

List of tabIes xiv

General Introduction 1

Chapter 1 Tmmunohistochemical studies of the endocrine cells within the gastro-entero- pancreatic system of Osteoglossomorpha

Introduction 15

Materials and Methods Anirnals Tissue preparation - Histology - Lrnmunohistocbemistry Results Morphology and histology + Immunohistochemistry - Islet cells - Gastro-intestinal cells

Discussion 34

Chapter 2 hunohistochemistry of the endocrine cells in the gastro-entero-pancreatic system of a generalized teleost (Catostomus comrnersoni) and a more derived teleost (Am bloplites rupestris). Introduction

Materials and Methods Animals 0 Tissue preparation - Histology - Immunohistochemistry

Results 0 General morphology Immunohistochemistry - Islet ceils - Gastro-intestinal cek

Discussion

Chapter 3 Ultrastructure and immunocytochernistq of the islet organ of Osteoglossomorpha

Introduction

Materials and Methods Animais 0 Tissue preparation - Routine electron microscopy - Immynocytochemistry

Results

Discussion

Chapter 4 Molecular cloning of preproinsulin cDNAs fiom several osteoglossomorphs and a cyprinid

Introduction

Materials and Methods 74 Animais a RNA extraction, first-Strand cDNA synthesis & RACE vii Isolation of the preproinsulin eDNAs Sequencing & identification of the cDNAs

Results

Discussion O Structure analysis O Phylogenetic anabsis

Chapter 5 Characterization of somatostatin cDNAs molecular identification and comparative analysis

Introduction

Materials and Methods Animais Extraction and quantification of total RNA Isolation of the preprosomatostatin cDNAs & sequence anaSsis O Northern analysis 93

Results

Discussion Structure heterogeneity Phylogenetic aaalysis

General Discussion 106

Summary 117

Literature Cited 120

Appendix 1 1

Appendix 2 II

Appendix 3 III:

Glossary List of Figures

Figure Page

L Diagrammatic representatim of the distrr'bution of the endocrine tissue 13

in fish relative t O the exocrine pancreatic tissue.

2 Phylogenetic relationships of major osteoglossomorph taxa.

3 Diagrammatic representation of a ventral view of the viscera in fke

species of osteoglossomorphs.

4 Photograph of a ventrai view of the viscera in H ahsoides.

5 Sections of pancreatic islets and exocrine acini (e) stained with

hematoqdin and eosin [A] and various antkra p-Dl.

6 Sections of islets of several osteoglossomorphs demonstrating

immuoreactivity to various antisera.

7 Sections of pancreatic islets of two O steoglossomorphs demonstrating 28

immunoreactivity to various antisera

8 Sections of pancreatic islets of H. alosoides demonstrating

immunoreactivity to various antisera. Figure Page

9 Sections of the stomach [A & Cl and intestine @3& Dl of several 30

osteoglossomorphs showing hmoreactivity to various antisera.

10 Sections through the intestine of various osteoglossomorphs

demonstrating immunoreactivity to various antisera

11 Sections of the stomach [A &BI and intestine CC & Dl in alosoides

showing irnmunoreactivity to various antisera

12 Sections tbrough islet tissue in a Brockmann body. [A] & [BI in

C. commersoni, [Cl in A. nrpestris.

13 Sections through the islet tissue of C. commersoni immunoreacting to

various antisera.

14 Sections through the islet tissue of A. rupestris immunoreacting to

various antisera. [A] To anti-m-INS, [BI & CC] to anti-SST.

15 Sections through the islet tissue of A. rupestris demonstrating

immunoreactivity to various antisera. [A] To anti-GLP, @3] to anti-

GLU, [Cl to anti-PYY and p]to anti-aPY. Figure Page

16 Two sections through the intestinal tissues of A. rupestris and 50

C. comrnersoni.

17 Light micrograph of a 0.5 pm section of a portion of an islet m the

pancreas of 0.biciwhosunz stained with toluidine blue.

18 Low maguification electron micrograph of a portion of an &let in

N. chitala-

19 Portion of a B-cell and several unlmown cell types in G. perersii.

20 B-cells (B) and a D 1-ceU@ 1) in 0. bicirrhosum.

21 F-cell (F) and B-cell (B) in G. petersii.

23 Portions of A (A) and F-cek (F) in 0.bicin-hosum.

24 Double Iabeling of D 1-ceU granules [a] in in. biciwhosum & [b] in

N. chitala with two Merent sizes of protein-A gold phcles. Figure Page

25 Protein-A gold labeled granules [a] and [b] of B-tek, following incubation 66

with ad-mINS serum. [cl and Id] ofD-ceh folIowing incubation with anti-

SST serum [el A-cell granules label with anti-GLU, [fl granules in a F-ceIl

label with anti-PW and [gJ Double incubation of two adjacent A-cell (A)

and F-cell (F).

26 Diagrammatic representation of the structure of the preproinsutin cDNA. 79

27 Preproinsulin cDNA sequence fkom the butterfiy kh. 80

28 Preproinsulin cDNA sequences A) fiom the arawana, B) fkom the goldeye, 81

C) fiom the feather fin Mekh, D) ficorn the sucker, and

E) fiom the elephantnose.

29 Alignment of deduced preproinsuiin amino acid sequences fiom

osteoglossomorph species.

30 Alignment of preproinsulin sequence fkom diverse vertebrate species.

31 Phylogeny of proinsulùi sequences. Sequences fiom

osteoglossomorphs and diverse vertebrate species were analyzed by

parsimony. Figure Page

A schematic representation of the preprosornatostatin cDNA.

The hybridization ofRNA samples with labeled preprosornatostatin

pro b fiom the arawana

Preprosornatostatin cDNA sequences A) fiom the arawana, B) fiom the

butterfly hh. C) fiom the elephantnose, D) fiom the feather fin knife fkh,

E) fiom the white sucker.

Alignment of preprosomatostatin sequences fiom different vertebrates

including the osteogiossomorph species.

Phylogeny of preprosomatostatin sequences. Sequences fiom

osteoglossomorphs and diverse vertebrate species were analyzed

by parsimony. List of Tables

Page

1 Relative abundance of cells irnmunostaining with antisera against 33

various peptides in ceh within the endocrine pancreas, intestine and

stomach of five representative species of osteogIossomorphs.

2 Relative abundance of cells immunostaining with antisera against

various peptides in cells within the endocrine pancreas, intestine and

stomach of Catostornus commersoni and Ambloplites rupestris. General Introduction

The gastro-entero-pancreatic sysfem (GEP) in vertebrates

The gastro-entero-pancreatic (GEP) system in vertebrates refers to the endocrine ce& in

the stomach and intestine (hence gastro-entero) and the endocrine tissues associated with the

exocrine pancreatic tissues, namely the idet tissue forming the endocrine pancreas or klet organ

(Youson and Al-Mahrouki, 1999)- In the stoniach and intestine the endocrine cek are present

within the mucosa of both organs, and produce a wide range of polypeptide hormones (Norris,

1997). The presence or absence of these peptides in the stomach andlor intestine Vary fiom one

species to another . Generally, these peptides include somatostatin, gIucagon family peptides,

peptide YY (PYY), and neuropeptide Y (NPY). In addition, severai other peptides are &O present in the stomach and/or intestine: secretin, cholecystokinin (CCK), gastrin, gastrïc- inhi'bitory peptide (GIP), and vasoactive intestinal polypeptide (VIP) . In the endocrine pancreas, the cek contain insulin (B cek), somatostatin @ cells), glucagon (A cek) and peptides of the pancreatic polypeptide fdy(F cells). As in the case of the stomach and intestine, the presence or absence of these islet celIs or the coIocalization of some of these peptides in one ceU type may

Vary fiom one species to another. These endocrine cek are believed to have an endodermal origin (Fher, l98Sa,b), and sometime are cded APUD cek (amine precursor uptake and decarboxylation) which relates them to a larger family of endocrine cells (Nomis, 1997).

Furthemore, it is believed that the vertebrate GEP endocrine ceh £ktoriginated in the nervous system then migrated to the gut and pancreas (Falkmer, 1999, suggesting a strong developmental relationslip between the alunentary canal and the endocrine pancreas in vertebrates (Faber, 1995; Maake et al,, 1998). Thus it is of value to relate the endocrine pancreas to stomach and intestine, and to study the whole GEP system. The importance of the gastro-enfero pancreatic system (GEP) iiî fnh

Several reports (Powers, 1989; Dickhoff et ai., 1990; Gorbman, 1990; Codon, 2000)

indicated that the use of fih systems as models in studies fkom various fields such as

developmental biology, neuro bio logy, and endocrinology proved to have made major

contnitions to these fields. Research on the gastro-entero-pancreat ic system (GEP) dates back

to the eady nineteenth century. This interest has encompassed investigations of the distribution,

structure, function, and peptides of the system. Studies ako were expanded to ident* the importance of the system to the individual fïsh species, how it developed, and how it relates to the GEP system of higher vertebrates (Faber, 1995). The investigation of the GEP system and its peptides in fish contriiuted to the advancement of our understanding of the biochemistry and physiology of the GEP system in vertebrates (Plisetskaya, 1990% b).

Plisetskaya (1 990% b) and Plisetskaya and Mommsen (1 996) reported fiom their studies on commercially valuable species that the hormones of the GEP system in fkh are essential for growth and maturity. Because of the importance of these hormones in &h and in higher vertebrates as well, researchers were interested in investigating the orïgin and the development of the cek (ontogenesis), and also the history (phylogenesis) of the appearance of the GEP system in vertebrates (Elonner-Weir and Weir, 1979; Epple and Brinn, 1986, 1987). It is believed that the ontogenic developrnent of the GEP system in bony £k4and the order in the sequence of appearance of its peptides in the developing endocrine pancreas, reflects the phylogenetic development of the endocrine pancreas in vertebrates (Berwert et al., 1995). In general, fïsh are an important mode1 when investigating the phylogeny of a vertebrate system because of the large number and diversity of its species. To better assess a phylogenetic pattern of the development of the GEP system in fish, many species should be investigated in this large vertebrate group (Youson and Al-Mahrouki, 1999). The importance of the GEP system in fish is not just restricted to its value as king a vital

system, but it was also considered in the treatment of some human diseases- For example, Yang

and Wright (1995) investigated the possibiIity of using the tilapia's Brockmam body region

"endocrine pancreas" for transplant in mammalian systems because of its compact and pure

nature. Furthemore, Sheridan et al. (2000) indicated that the characterization of the somatostatin

genes in fish might assist in developing new sornatostatin analogs that cmenhance the treatment

of some human diseases such as tumors and neurological disorders.

Actinopterygii and the GEP system

Actinopterygii, the ray-hed fishes, is a large class of fish of total number of nearly

24,000 species, wbich include the division Teleostei of the subclass Neopterygii. The division

Teleostei is monophyletic and includes the subdivisions Osteoglossomorpha., Elopomorpha,

Clupeomorpha, and Euteleostei (De Pinna, 1996). To date the GEP system has ken investigated

in only a fiaction of this large number of species. Studying the GEP system fkom representatives

of all the different subgroups is important because of the highly divergent habits and ancestral

origins of its members (Nelson, 1994). It is important to note that the studies to date have

favored the more derived teleosts over the generalized and the basal species (Youson and

Al-Mahrouki, 1999). Using the available data, and given the smd sample size that have ken investigated so far, is considered as one of the limitations when comparing the morphological features of the system among the different groups. Other iimitations would include the inconsistency among investigators in selecting suitable cornparison pararneters and the different interpretation of observations. Youson and Al-Mahrouki (1999) proposed a set of effective morphological parameters that can be used in the cornparison of the GEP system among several groups of fish. These parameters included: the pattern of islet distniution; thek size; degree of accumulation; relation with the exocrine tissue; absence or presence of principal islets or

"Brockmann bodies"; types of endocrine cek and the pattern of theÏr distn'bution, which can be

identified using immunoreactivity to antisera against hormones produced by these cells.

Electron microscopy and imrnunocytochemistry can also be a good tools to confüm cell identity

or to identify possible colocalization of peptides in certain cell types. However, Brinn (1973) reported that using the morphological parameters lead to inconsistency even among the closely

related species. As a result it is important to understand that to be able to make definitive phylogenetic conclusions regarding the GEP system in Actinopterygii, certain parameters should be considered including the sample size.

Ine origin of fhe GEP sysfemin- and its endocrine cell types

A diagrammatic summary of the developmental sequence of the GEP system cells in &h is presented in Figure 1. Disseminated endocrine cek are distri'buted within the gut epithelia of protochordates and ultimately cells of similar nature form an islet organ in bony tishes (Youson and Ai-Mahrouki, 1999). The endocrine cells of the GEP system are believed to have originated as neurons, then become disseminated ce& in the digestive mucosa, and ultimately ended up as a solid gland (Faikmer, 1995). Several investigators (Falkmer and Patent, 1972; Van Noorden,

1984; Fallaner, 1985 a, b, 1995) traced the origin of the GEP system in vertebrates as having a brain-gut axis. It is believed that the peptides of the vertebrate GEP system appeared fïrst in early evo lut ion in the invertebrate nervo us system (Van Noorden, 19 84) fo llo wed by it s appearance in the gut of a derived invertebrate (Faher, 1995). In protochordates dual occurrence of these peptides was reported in both the brain and the gut (Faher, 1995). Using imrnmohistochemistry, Reinicke (1981) was able to map the immunoreactivïiy to insulin, glucagon, somatostatin, and pancreatic polypeptide in the gut of Branchiostoma Zanceolatum

(protochordate). The budding of these gut endocrine ce& to form the islet organ was fïrst

reported in the larval hprey (basal vertebrates, jawless fish), where a one-hormone islet organ

is found to produce insulin (Barrington, 1945), with the ceils primarïly originating fiom the

intestinal epithelium and residing in the submucosal connective tissue (Youson, 2000). In

Wh,whîch is another basal vertebrate, a two-hormone islet organ was reported (Faher,

1985b), producing insulin and somatostatin, but originating Eom the extracornmon bile duct, with many of these peptides present in the intestine (Van Noorden, 1990). However, in adult lampreys the three-hormone islet organ was reported to contain B- and D-ceils (Ebott and

Youson, 1986; Yui et al., 1988), and F-cells (Cheung et al., 1991). The presence of the classical four-hormone islet tends to be the case in other fishes. In basal actinopterygian species,

Lepisosteus spp. (gars) and Amia culva (the bowfk), the use of immunohistochemistry revealed islet ceils producing the four classical hormones. GroE and Youson (1 998) correlated fine structural observations with immunocytochemistry, which revealed three cell types in the islet organ of lepisosteus, B-, D- tells, and a third A,ceil type, where peptides of the pancreatic polypeptide family colocalized with glucagon. However, the cells of the islet organ of the bowfin immunoreacted with antisera against insulin, somatostatin, glucagon, and pancreatic polypeptide families (Youson and Al-Mahrouki, 1999; Youson et al., 200 1). This implies the presence of four endocrine cell types in the klet organ of the bowfin, which is a characteristic feature of more derived ActinopterygK It is of interest that Gardiner et al. (1 996) used parsimony analysis of the primary structure of insulin and glucagons fiom the gar and the bowfïn and showed that the bowfùi is more closely related to teleosts tha.to the gars. In euteleosts the four principal ceil types (A, B, D, F) have been descn'bed in the klet

organ of many species, in sahonids (Wang et al., 1986; Nozaki et al., 1988 a, b), in anglerfish,

Lophius amerkanus (Johnson et al., 1976), in sea bass, Dicentrarchus Zubrax (Agulleiro et al.,

1993; Gomez-Viscus et al., 1996, 1998), and the gilt-head sea bream, Spmsaurata (Abad et al.,

1992). In addition, two D ce11 types were also descriid (Dl and D2) in the isiets of many euteleosts, with specific immunoreactivity to dinerent types of sornatostatin antisera and specïfïc localization in the islet (Nozaki et al., l988a; Abad et al., 1992). Maglio and Putti (1 998) reported the relevance of the D-cek distriibution to their modulatory effect on the secretions of A and B ce&.

The identification of the different ceil types in the islet organ of teleosts was achieved using Mmunocytochemistry and ultrastructure (L,ozano and Agdleiro, 1986; Beccaria et al.,

1990; Agulleiro et al., 1993; Gomez-Visus et al., 1996). This investigation was also extended to nonteleost actinopterygian (ray-hed hh), the gar (Groff and Youson, 1998) and the bowfin

(Youson et al., 200 1).

Exocrine-Endocrine tissue relation in the GEP s>sfem

At all intervals of the posternbryonic periods of the life cycles of lamprey and the ha@4 exocrine pancreatic cek are present within the epitheliurn of the digestive tract. In addition, endocrine aggregations (islets) reside in the submucosal co~ectivetissue. In adult lampreys where the islet aggregations outside the digestive tract are present with no exocrine tissues around it, the endocrine tissue is considered to be present as principal islets (Youson, 2000).

Later on in the development of the GEP system in Bhthe exocrine and the endocrine tissues of the pancreas are both weil established outside the digestive tract. However, the size of the islet organs and their distriiution within the exocrine pancreas varies in the different groups of

Actinopterygii. In the basal Actinopterygii (gar and bowfin) the endocrine tissue is found widely

dispersed throughout the exocrine acbi, no distinct concentrations of islets are reported (Epple

and Brinq 1975; Groff and Youson, 1997). In the eek (Elopomorpha), which is considered to be

a lower Teleostei, the general view is that the eel pancreas is compact with smd islets scatîered

throughout the exocrine pancreas, and with no principal (large) islets or aggregation of islets

cded Broclanann bodies (Kobayashi and Takahashi, 1970; Epple and Brinn, 1975). Recently,

Plisetskaya and Mommsen (1996) defhed Broclanann body as the largest accumulation of

endocrine tissue surrounded by a thin rim of exocrine tissue, whereas, Epple and Brï~(1986,

1987) defhed it as large accumulation of islets closely associated with exocrine pancreatic

tissue. However, L'Hermite et al. (1985) descnbed single islets in the middle of the pancreas of

the glas eel and they cded it a Brockmann body. In the more derived euteleosts a notable

aggregation of the islet tissues were reported formuig Brockmann bodies (Fig. 1) with no

associated exocrine or with very little exocrine tissues (Youson and Al-Mahroulü, 1 999). These arrangements (reviewed by Youson and Al-Mahrouki, 1999) are reported in derived orders

including Cyprinodontiformes, Lophiformes, and Scorpaeniformes (Boquist and Patent, 1971;

Johnson et al., 1976; Klein and Lange, 1 977; Patent et al., 1978; Klein and Van Noorden, 1980:

Stefan and Falkmer, 1980). Another distribution pattern of several large, intermediate and small

islets associated with iimited exocrine tissue was also reported in derived orders, including

Mugiliformes, Perciformes, and Tetradontiformes (Kobayashi et al., 1976; Abad et al., 1986;

Lozano et al., 199 1 a, b; Magiio and Putti, 1998).

Generally, the comparative analysis of the endocrine pancreas of the actinopterygian fishes have been of an essential significance in understanding how these endocrine cells originated withi.the GEP system of vertebrates (Epple and Brinn, 1987;

Fallaner, 1995).

In this study, the GEP system is investigated in the osteoglossomorphs and in two other teleosts, a generalized and a derived species (Chapters 1 and 2). The distriiution of the endocrine tissue in relation to the exocrine tissue withui the system of this teleost group is investigated and compared among the species under study to establish its position in cornparison to that of the other teleosts. me primary sttucture and the molecular evolution

The primary structures of the peptides in the GEP system were characterized fiom several fkh species (Codon, 1995,2000; Larhammar, 1996; Sheridan et al., 2000), and were important for the analyses used to explain the molecular evolution of the peptides. Insulin bas ken characterized fiom several actinopterygians including the gar, Lepisosteus spathula

(Pollock et al., 1987), the bowfîn, Amia calva (Conlon et al., 199l), the paddlefish, Polyodon spathda (Nguyen et al., 1994), and the carp, Cyprinus carpio (Hahn et al., 1983). However, to date the preprosomatostatin cDNAs have been characterized fiom only four teleost species, namely the trout, goldi%h and anglerfïsh (see Sheridan et al., 2000 for review), and catfish

(Minth et al., 1982), and f?om the sarcopterygian, the Iungfkh (Trabucchi et al., 1999).

The comparative dysisand the characterization of the sequence of the hormonal peptides in basal and more-derived organisms should give us an idea of how the peptides might have evolved fkom the ancestral molecules. For example, there is a 33 amino acid daerence between hagfish and rnammhn insulins (Peterson et al., 1974). Some researchers (Codon and

Hicks, 1990; Aguileiro, 1995; Dores et al., 1996) mentioned that the low molecular weight polypeptide hormones are not useful for constnicting a phylogenetic tree. However, Dores et al.

(1 996) constructed a simiificant phylogenetic tree using the variable spacer region of proinsulin (C-peptide), and found it useful for a phylogenetic analysis. Hahn et ai. (1983) used the proinsulin sequences to evaluate the evolutionary distance between fiesh water and marine teleosts. The proinsulin sequence was also used by Conlon et al. (1997a) to prove that lungfkh and amphiiians share a common mcestry, and by Conlon et al. (1998) to confïrm that the polypterids are the most ancient living actinopterygians. Other GEP peptides such as glucagon- iike peptide have been used to address several phylogenetic questions, such as clarifiing the phylogenetic relationship of amphibians (irwin and Sivarajab, 2000) and the antiquity of the peptide as seen through agnathan phylogeny (Irwin et al., 1999). The preprosomatostatin sequence confirmed the close relationship of the lmgfish and tetrapods (Trabucchi et al., 1999).

Furthemore, Conlon (2000) reported that pancreatic polypeptide (PP) is of value as a molecular marker in clari@ing the early tetrapod phylogenetic relationships. Nevertheless, Dores et al.

(1996) concluded that although the maximum parsimony analysis of hormone coding genes cm be usefûl to resolve phylogenetic related questions, the analysis of several hormone-coding genes is required in such cases.

Osteoglossomorpha

This study is focusing on the characterization of the GEP system of the

Osteoglossomorpha @ony-tongues). In particular, the morphology, the characterization of its peptides, and how they relate to those of other Actinopterygii are king examined.

Osteoglossomorpha is an interesting teleost group to study, with ancient geological origin that goes back to the late Jurassic era. The geographical distribution of the present members representing this group was uifluenced by tectonic plate movement (Li and WWilson, 1996).

Farnilies of this group are present in North Amenca (Hiodontidae), in fica(Mormyridae and Pantodontidae), in South America and Australia (O steoglossidae), and in Asia (Notopteridae).

Figure 2 provides a recent view of the taxonmrnic relationship of the osteoglossomorph families

(Li and Wilson, 1996). Several researchers have debated the taxonomy of osteoglossomorphs

and there is no clear conclusion as to whethex they are the most basal living teleosts (Nelson,

1973; Patterson and Rosen, 1977; Taverne, f1986) or are more general teleosts (Arratia, 1991;

OYNeiUet al., 1998; Van Le et al., 1993). Li and Wilson (1 996) and Zhang (1 998) indicated that

the Osteoglossomorpha is a monophyletic grmup. Li et al. (1997) reported a close relationship

between the Notopteridae and the Morrny-ridzie, and they aIso indicated that the two families are a

sister-group of the Hiodontidae. However, Ahes-Gomes and Hopkins (1 997) analyzed the

sequence of the mitochondrial12S rRNA and 16s rRNA and concluded that the Pantodontidae

departs fiom the Mormyridae. Recently, the molecular forms of gonadotropin-releasing

hormones were analysed (O'Neill et al., 199s) and showed a close relationship between the

Pantodon and the eel (elopomorph).

The goals of the study

Since Osteoglossomorpha was argued to be the most basal living teleosts, it was anticipated that its endocrine pancreas conskt of small islets scattered throughout the exocrine pancreas and contain the four major cell types (A, B, D and F). To test this hypothesis several approaches were planned including histology, Mmunohisto- and -cytochemistry. Furthemore, the characterization of the prirnary structure ofsome of the peptides produced by these endocrine cek and the comparative analysis of these structures among the osteoglossomorphs and to other teleosts could give better understanding of theposition of the osteoglossomorphs among teleosts.

In the present study the answee to two main questions were investigated. Fir~t,where does the GEP system of the osteoglossomorphs fit in comparison to tbat of other teleosts? This question wilI be annvered by considering the morphology of the system, types and distribution of the endocrine tek within the system, and their relation to the exocrine tissue. Second, is it possible to use the primary structure of some of the GEP peptides of this group to better understand the molecular evolution of these peptides in vertebrates, or to try to resolve some of the debate over the phylogenetic position of the osteoglossomorph fdesboth among themselves and within the teleosts?

To provide answers to these questions, several approaches were followed. These approaches include the identification of the GEP system in six osteoglossomorphs through general histology, and localization of endocrine ceils using immunohistochemical techniques

(Chapter 1). The same approach was used to descrii the GEP system in a generalized (white sucker), and in a derived (rockbass) teleost (Chapter 2) for comparison with that of the osteoglossomorphs.

The identity of the diflierent islet cell types descnid in Chapter 1 was confïrmed through

Meranalysis using ultrastructural and immunocytochemical techniques (Chapter 3). The anaiysis of the distri'bution and immunoreactivity of ceil types (Chapters 1,2 and 3) formed the basis for the molecular studies that foilowed in Chapters 4 and 5. The identification of the primary structure of some of the GEP peptides involved the use of the recombinant DNA technology. This technology ailowed the characterization of the cDNAs of the preprohormones of the peptides of interest, and fiom which amino acid sequences were deduced. The focus of this study was to characterize the primary structure of both insulin and somatostatin because of some curious morphoIogical differences that were observed among the different species (Chapters 1 and 2). The primary structure of insuh was characterized fiom five osteoglossomorphs and a generalized teleost (cyprinid) (Chapter 4). The prirnary structure of somatostatin was characterized fkom four osteoglossomorphs and also a cyprinid (Chapter 5). Furthemore, the preprohomones of both insulin and somatostatin were used dong with previously published data fkom O ther fish species to construct phylogenetic trees; the results of these analyses are discussed in detail in Chapters four and five.

The general discussion follows the Iast Chapter of this thesis to discuss the overd results and to ilimate how these results can be interpreted towards answering the main questions proposed at the commencement of the study. Fig. 1 Diagrammatic representation of the distri'bution of the endocrine tissue (idet organ) in

fishes relative to the exocrine pancreatic tissue, esophagus (O), stomach (S), intestine

(unlabeled portion of gut), liver CL), gall bladder (G), and bile duct (tube leading fiom the

gall bladder). The epithelid cek of the digestive tube are enIarged relative to the cells of

other structures to denote the presence of endocrine (dark) and exocrine (hatched) cek.

Represented are a protochordate (a), a larval lamprey: Northern Hemisphere (b), a larval

iamprey: Southern Hemisphere (c), an adult lamprey: Northern Hemisphere (d), an adult

lamprey: Southem Hemisphere (e), a hagfïsh (0, a holocephalian (g), an elasmobranch

(h), a basai actinopterygian (i), and a derived euteleost 0). The protochordate has no

pancreas but enteroendocrine cek and ha1and adult lampreys and hagfish have

exocrine and endocrine cek in the intestine. The islet organ is seen as scattered follicles

at the esophagus-intestinal junction in larval lampreys and around the bile duct-intestinal

junction in the hagfish. Islet organs are one (e) or two (d) principal islets in addt

lampreys. The exocrine tissue and islet organs are intermingled in the remaining species,

with a compact organ in the two cartilaginous fishes (g, h) and a more mseorgan in

the two bony fishes (i, j); islets are sdand difEUse in (0 but more concentrated (BroclanaM body) and with a cranial principal islet in a). (This figure and the

description are duplicated fiom Youson and Al-Mahrouki, 1999).

Fig. 2 Phylogenetic relationships of the major osteoglossomorph taxa. A) with other basal

TeIeosts (Elopomorpha and Clupeomorpha) and with the generalized and derived

teleosts (Euteleostei). B) The intragroup relationship of the osteoglossomorphs show

Notopteridae (e-g. feather fin knife fish) closely related to Mormyridae (e.g.

elephantnose), Pantodontidae (e-g. butterfly fkh) closely related to Osteoglossidae (e-g.

silver arawana), and Hiodontidae (e.g. goldeye) a sister group to the other groups

(modifïed fiom Li and Wilson, 1996). Osteoglossomorpha

Elopomorpha

Teleostei Clupeomorpha

Euteleostei Hiodontidae

Osteoglossomorpha :, 8' ' ,',\ . 1. .

Pantodontidae

Mormyridae

Notopteridae Chapter 1

Immunohistochemical studies of the endocrine cells within the gastro-entero-pancreatic system of Osteoglossomorpha

From

Al-Mahrouki, A.A. and Youson, J.H. (1998). Immunohistochemical studies of the endocrine cells within the gastro-entero-pancreatic system of Osteoglossomorpha, an ancient teleostean group. Gen. Comp. EndocrinoL 110,125-139.(with permission of the Academic Press) Together with the immunohistochemical study of Hiodsn alosoides. Introduction

Osteoglossomorpha are considered to be one of the most ancient superorders of the teleosts, the most abundant of the actinopterygian fishes (Patterson and Rosen, 1977).

Osteoglossomorpha is an interesting group to study because of the anatomy, physiology, geographic distribution, and ancient fossil record of its species. These fieshwater fish display interesting biogeographic distniutions including examples of endemism: Hiodontidae in North

Amenca, Mormyridae and Pantodontidae in aca,Osteoglossidae in South Amerka, and

Notopteridae in Asia. Potentially, they represent an interesting case to study the role of continental movements in the distri'bution of fieshwater fishes, and of the importance of fossil taxa in building phylogenetic reiationships of extant species. Osteoglossomorphs are a geologicaiiy ancient group of fieshwater fkh and their ancestral origins go back to the late Jurassic. This history suggests that their recent disiribution should have been iduenced by plate tectonic events

(Li and Wilson, 1996).

Comparative analyses of the endocrine pancreas of the actinopterygian fïshes have been of fundamental signifïcance in understanding the nature of hormonal peptides and the ongin of cells within the gastro-entero-pancreatic (GEP) system of vertebrates (Epple and Brinn, 1987;

Falkmer, 1995). Brinn and Epple (1 990) reported fkom cytoarchitectural, innervation, and vascular perfusion data that the actinopterygian pancreatic islets are highly specialized, sim.to those in the birds and mamMalS, and can be called islet organs. Falkmer (1995) proposed that the endocrine cek of the pancreas start as neurons, then become disseminated cells of open or closed type in digestive mucosa, and uitiniately end up as a soïid gland. Thus there is a strong developmental relationship between the organs of the GEP system, and it is of value to study the related endocrine ce& of the endocrine pancreas, the stomach, and the intestine concomitantly. 16 Since the superorder Osteo~glossomorphaholds a pivotal position in the evolution of the

teleostean group, it presents an interesthg case to study with regard to the development of the

GEP system. At the very least, knoiwledge of the presence and distri'bution of several peptides in

the endocrine ceb of the GEP sy&-ern of osteoglossomorphs could be significant in the

understanding of the phyiogenetic development of the endocrine pancreas in actinopterygian

hhes. To date, only passing attentfion has ken given to the pancreas of members of the

Osteog~ossomorpha(Mc Cormick, 1925; Epple and Brinn, 1975).

Several studies of the endocrine pancreas of the teleosts using immunohistochemical techniques (Van Noorden and Patemt, 1978; Stefan and Fakmer, 1980; Rombut and Taverne-

Thiele, 1982; Abad et ai., 1986; Mommsen and Plisetskaya, 1991) have reveded the existence of four endocrine ceil types immunoreactive to antisera raised against insulin (B-ceiis), glucagon (A- cells), somatostatin @-cek), and members of the pancreatic polypeptide family (F-ceiis) . The present study examines the presence of cells types imnimoreactive to antisera raised agaimt inçulin, somatostath, glucagon family peptides, and peptides of the pancreatic polypeptide family within the GEP system of five species of osteoglossomorphs.

Materials and Methods Animals: Adult specimens of Ostel-oglossumbiciwhosum (siiver arawana), Scleropages jardini

(Australian arawana), Pantodon buehholzi (butterfly hh), Notopterus chitala (feather fh Me

&h), and Gnathonemuspetersii (eleephantnose), were obtained fkom several hhaquaria within the Toronto area. A total of seventeen anin& (a minimum of three for each species except only oIie S. jardini) were used in this study. Animals were sacrificed by an overdose of 0.05% trkaine methanesulfonate (MS-222) in the+ holding water. Adult Hiodon alosoides (goldeye) were fished fiom the Assinboine River ne= Winnipeg, and the viscera was obtained on site and ked in

Bouin's fluid for 24 hr, then stored m 70% ethanol. Tissue preparation: HiSfoIogy: A large portion of the viscera containing the stomach, pyloric caeca, gall bladder, intestine and connecting mesentery were excised and fixed intact in Bouin's fhid for Light micro scopy.

Routine light microscopic techniques consisted of fixation for 24 hr, storage in 70% ethanol and, before tissue processing, a detailed drawing ofthe general vicera was carried out

(Fig. 3). The vicera of the H dosoides was photographed (Fig. 4). The tissues were eventually embedded in parfiwax, followed by sectioning at 7 pmthickness and staining with hematoxyh and eosin (H&E).

Immunohistochemisfry: Excised tissues were also fïxed in Bouin's fluid for a period of 20 hr, dehydrated in a graded series of ethano ls, embedded in paraffin, and adjacent, 7 pm sections placed on glas slides.

The site of immunolabehg active was revealed by a commercial rabbit histostain-SP kit

(Zymed-lab-SA System, San Francisco, CA). This kit uses a biotinylated second antiidy, a horseradish peroxidase-streptavidin conjugate, and a substrate-chromogen mixture to demonstrate antigen in ce& or tissues. Positive staining is indicated by red staining and there is a hematoxylin counterstain. The foilowing primary antisera were used in this study: (1) guinea pig anti-bovine insulin (anti-mINS; a kind gift of Dr.C.Yip, University of Toronto) diluted 1: 1000; (2) rabbit anti- somatostatin IgG fiaction (anti-somatostatin-14; LINCO Research, Inc., Missouri) diluted

1: 1000; (3) rabbit anti-salmon somatostatin-25 (SST-25) and (4) rabbit anti-salmon polypeptide tyrosine tyrosine (PYY) anti-SST-25 and anti-PYY (a kind gifî of Dr. E. M. Plisetskaya,

University of Washington) diluted 1: 1000; (5) rabbit anti-synthetic glycine-extended angIerfish polypeptide Y (anti-aPY, couilresy of P. C. Andrews, University of Michigan) diluted 1: 1000; (6) rabbit anti-human neuropeptide Y (mti-NPY, P. C. Andrews) diluted 1: 1000; (7) rabbit anti- 18 porcine glucagon (anti-GLU, Zymed Laboratories Inc., San Francisco, CA) diluted 1: 1000; and

(8) rabbit anti-saimon glucagon-like peptide (anti-GLP, E. M. Pktskaya) diluted 1: 1000.

Positive immunoreactivity was detected in comparison with that of rat pancreas, and negative controls consisted of replacement of prirnary antisera with phosphate-buffered saline (PBS), pH

7.6, and of antisera previously absorbed with excess antigen.

Results Morphology and Histology The overail anatomy of the whole viscera varied among the six species (Figs. 3 and 4), resulting in slight variations in the distribution of the pancreatic tissue. In ail species, the pyloric caeca arose fiom the most cranid portion of the anterior intestine (duodenum). Both O. biciwhosum and S. jardini, the two arawana species, had the same anatomical features, where the two, fïnger-Wce pyloric caeca seems to be a family-specific feature (Fig. 3A). The shape and number of the pylonc caeca in G. petersii was similar to that of O. bicirrhosum and S. jmdini, but foLlowed the curved anterior intestine (Fig. 3D). The two pyloric caeca of N. chitaln were shorter and wider than the forenamed species but also curved with the intestine (Fig. 3C). P. buchholzi had one sac-like pylonc caecum (Fig. 3B). H alosoides ako had one pyloric caecum but with a kger-like shape similar to that of the arawana (Fig. 4). In all species, the caeca appeared to leave the intestine as a common tube before they branched.

The exocrine achar cells had numerous, intensely stained, apical acidophilic granules and a basophilic cytoplasm and nucleus, whereas the smaller endocrine islet cells had a slightly acidophilic cytoplasm, with less basophilic nuclei (Fig. 5A). Each islet was enveloped by a delicate connective tissue capsule, which separated it fiom the surrounding exocrine tissue (Figs. 5-8).

In all species, the exocrine pancreas was diffusely distriuted within the mesentery connecting the outer surface of the gastrointestinal tract, the pyloric caeca, the gali bladder and the her (Figs. 3 and 4). As for the endocrine tissue, the islets of various sizes were widespread amongst most of the designated pancreas region and surrounded by exocrine pancreas (Fig. 5A).

In both O. biciwhosum and S jardini the islets, especially the large ones, were concentrated at the region where the pyloric caeca connected to the anterior intestine. In contrast, the rest of the pancreatic tissue contained small, widely dispersed islets. The mode of islet distribution within the exocrine pancreas was similar in chitala, G. petersii, and H. alosoides where both large and srnall islets were well spread thtoughout most of the area penetrated by exocrine tissue. The Sets were most widely dispersed throughout the whole exocrine tissues in P. buchhoZzï.

The stomach and intestinal walls in the six species consisted of several tunics, characterized by the inward folded mucosal layer, lined by a simple columnar epithelium Pigs. 9-

11). The surface epithelium of the stomach was comprised of mucous cek and gastric pits leading into gastnc glands within the lamina propria (Fig. 9A and C). However, the epithelium of the intestinal mucosa had enterocytes with characteristic bmh border, and a large number of goblet cells (Fig. 9B and D). A similar structure to the intestine was observed in the pyloric caeca, but there were fewer goblet cells.

Immunohistochemistry

A summary of the immunohistochemical results to different antisera in the three organs of the six species is presented in Table 1, where the relative number of cek staining with the antisera are compared. The evaluation of relative imrnunostaining intensity is based on the degree of staining with the various antisera and is provided in the text. The immu11oreactivity within the pancreatic tissue indicates that the osteoglossomorphs

have the four major types of islet celis immunoreactive to antisera raised against kuh@-cells),

somatostatin (D-cells), glucagon (A-cells), and members of pancreatic polypeptide (F-cells).

The central portion of the islets in the six species showed an abundant, intense

immunoreactivity to the m-INS antisem In 0.bicirrhosum, S. jardini, G. petersii and H

alosoides the imrnunoreactivity was centrai within the islets and was surrounded by a moderately

thick rim of non-immunoreactive ce& (Fig. 5B and 8A). In the case of P. buchholzi and AL

chitala, the larger central core of immunoreactive cells left only a thin rim of non-

immunoreactive cek at the periphery (Fig. 6A).

The cells that did not immunostain for insulin in the central core of the islets were found

in adjacent sections to be mainly D cek, but this cell type also extended to the periphery of the

islet. Weak immunoreactivity with SST-14 antiserum was observed in many &let ce& of O. bicirrhosum, N. chitala and G. petersii (Fig. 6B), whereas a moderately abundant and more intense immunoreactivity with this antiserum was observed in the islet ceils at the central core of the islets of bo th S. jardini and P. buchholzi In contrast, in H. alosoides the immunoreactivity was restricted to a thick rim towards the periphery of the islets Vig. 8B). In addition, aggregated

cenaal ceUs of the islets showed abundant, intense immunoreactivïty with SST-25 antisenim in five species (Fig. 6C), in contrast with G. petersii, where the immunoreactivity was less conspicuous.

The intense immunoreactivity to aPY antisemm was only detected in ce& at the very periphery of the islets in ail the species (Fig. 7C). A si.immunoreactivity to this was foumi with the NPY and PWantisera in five of the species (Figs. 6D, 7 B and D and 8C), but in contrast, immunoreactivity in N. chitala, to both NPY and PW antisera was infkequent. 2 1 Immunostaining with both NPY and PW antisera appeared in a few cek scattered throughout the islets in some of the species (Fig. 7B)-

In O. bicin-hosum and G. petersii the islet penpheries demonstrated a thick layer of cek intensely immunoreactive to both GLU and GLP antisera (Fig. 5C and D). These A cek partidy accounted for the thick unstained ri.of cek surrounding the central core of B cek (Fig. 5B) in these two species. However, in S. jardini, P. buchhoZzÏ, N. chitala, and H. dosoides, the intense immunoreactivity to both antisera was also in cells scattered throughout the islet, in addition to the peripheral cells (Figs. 7A and 8D).

Scattered ceUs that were immunoreactive to the different antisera were observed among the exocrine acini in the six species. Although some ce& may have been part of small islets, others appeared to be isolated. Meanwh.de, no iriununoreactivity to any of the antisera was detected in ducts of the exocrine pancreas in any of the species examined.

Gastro-intestinal ceik

Due to the differences in tissue area behveen the pancreas and either the stomach or the intestine, the relative abundance of staining cek was oniy compared between the stomach and intestine.

Immunoreactivity to the dinerent antisera in the pyloric caeca and the anterior intestine was similar in all of the species, and was consistent with the structural simil;irity of the two organs. The immunoreactive cells in the intestine were randomly scattered among other epithelial cek of the mucosa of the antenor intestine. No immunoreactivity was detected in the posterior intestine of any of the six species.

hûequent immuoreactivity to m-INS antisem was detected in ceils scattered within the intestine of 0. bicirrhosum, P. buchholri and G. petersii, and no immunoreactivity to this antiserum was observed in the intestinal cells of the other species. Although a moderate INS 22 irnrnunoreactivity was only O bserved in some cells of the gastric ghdsof P. buchhoZzi (Fig, 9A), no immunoreactivity to this antisenrm was observed in any stomach cell of the other species.

Clurnps of adjacent cells in the intestine of O. biciwhosum, P-buchhohi, N. chitala and G. petersii demcinstrated a moderate immunoreactivity with the SST-14 antiserum (Fig.gB), but no immunoreactivity was observed in S. jardini. Although the stomach cek of four species did not immunoreact with the SST-14 antisenim, an equal abundance of anti-SST-14 immunoreactive cells was observed in both the intestine and in the stomach glands of G. petersii. In contrast, in H. alosoides the immmoreactivity to both antisera was only observed in the stomach (Fig. 1la).

The SST-25 antiserum showed numerous, intensely immunoreactive glandular tek in the stomach in G. petersii (Fig. 9C) but fewer immunoreactive cells in the stomach of S. jardini and

H. alosoides (Fig. 1 1b). Anti-SST-25 staining ceils were present in the intestinal ce& of only

G. petersii (Fig. 9D). No immunoreactivity to this antiserum was observed in either the stomach or the intestine cells of the other species.

Intense immunoreactivity to NPY antisenun was observed in scattered intestinal ceils in four of the species (Fig. lOB), however, the intestinal cells of P. buchhohi and the H. alosoides demonstrated a rather infiequent and weak immunoreactivity to this antisenim, Similarly, scattered intestinal ce& of all the species demonstrated intense immunoreactivity to aPY antiserum (Figs. 1OC and 1 1D), but no immunoreactivity to this antiserum was detected in the stomach in any of the species. The irnxnunoreactivity to the PYY antiserum in the intestinal ce& was moderate in intensity and fiequency in aU the species (Fig. 10D). Irmnunoreactiviity to this latter antisenrm was only observed in the stomach ceUs of P. buchholzi.

For the most part, immunoreacti~to GLU and GLP antisera had similar distri'bution . It was intense in moderately abundant intestinal cells in four of the species (Figs. 9D, 10A and

1IC), and intense with infrequent cells in 0.bicirrhosum. P. buchholki stained only with 23 anti-GLP. No immmoreactivity to these antisera was observed in the stomach cek of any of the species. 24

Fig. 3. Diagrammatic representation of a ventrd view of the viscera in five species of

osteoglossornorphs. hcluded are the oesophagus (O), stomach (S), intestine (i),liver (L),

gall bladder (G),pylonc caeca (Py) and spleen (Sp) The endocrine pancreas is distri'buted

"thin the pancreas (P). [A] represents both 0.bicirrhosum and S. jardini,

@3]P. buchhohi, [Cl N. chitala and @3]G. petersii.

25

Fig. 4. Photograph of a ventral view of the viscera in alosoides. hcluded are the

oesophagus (O),stomach (S), intestine (I),liver (L), pyloric caeca (Py) and the pancreas

(P) which include the endocrine pancreas.

Fig. 5 Sections of pancreatic islets and exocrine acini (e) stained with hematoxylin and eosin [A]

and various antisera @3-Dl.p] Dernomation of immunoreactivity to anti-mNS in

G. petersii with a moderately thick rim of non-immunoreactive cek (arrowhead) at the

isiet periphery. [Cl hti-GLU in G. petersii showing a thick peripheral rim of

immu11oreactive cek (arrowhead). (PI Anti-GLP in G. petersii &O showing a thick

peripheral rim of immunoreactive cells (arrowhead). X560.

Fig. 6. Sections of islets of several osteoglossomorphs demonstrating immtmoreactivis. to

various antisera. [A] Anti-m-INS in N. chitala showing a thin peripherd rim of

nonimmtmoreactive ceh (arrowhead) at the islet penphery. p]Anti-S ST- 14 in

G. petersii shows weakly immunoreactive D ce&. [Cl Anti-SST-25 in N. chitala showing

a wide butio ion of D cek fiom the center to the outer edge (arrowhead) of the islet.

Anti-NPY in O. bicirrhosum showing two adjacent islets with F cells at the periphery

(arrowhead) but a few in more central locations (arrow). X560.

Fig. 7. Sections of pancreatic islets of two osteoglossomorphs demonstrating

irnmunoreactivity to various antisera. [A] Anti-GLU in P. buchhohi shows

scattered immunoreactive cek in a large islet (arrowhead and arrow). @3-D]Adjacent

sections fiom the endocrine pancreas in G. petersii. p]Anti-NPY, [Cl Anti-aPY and

p] Anti-PYY is reactive in F-cek at the blet periphery. X560.

Fig. 8 Sections of pancreatic islets of H almides demonstrating immunoreactivity to various

antisera. [A] Anti-m-INS shows immunoreactivity in the central core of the islet wiih a

peripheral thick rim of nonimmunoreactive cek (arrowhead). @3] Anti-S ST- 14 shows

imrnunoreactivity confhed to the periphery of the isiet, unlike in the other species. [Cl

Anti-NPY shows immunorectivity at the periphery of the islet. PD]Anti-GLU shows

imrnunoreactiviis. both at the outer rim (arrowhead) and the inside (arrow) of the islet.

X560.

Fig. 9. Sections of the stomach [A and C] and intestine (B and D] of several osteogIossomorphs

showing immunoreactivity to various antisera- [A] Anti-m-INS is localized within a single

ceU in the gastric glands of P. buchhoZzi. @] Anti-SST-14 in adjacent cek of the

intestinal mucosa of 0.biciwhosm. [Cl Anti-SST-25 intensely stained cek in gastric

glands in G. petersii. p]Anti-GLP stained cell in P. buchholzi intestine. X900.

Fig. 10. Sections through the intestine of various osteoglossomorphs demonstrating

irnmunoreactivity to various antisera. [A] Anti-GLU-stained ceii in P. buchholzi.

1131 Anti-NPY-stained cek in 0.bicirrhosurn. [Cl Anti-aPY-stained cell N, chitala

] Anti-PYY-stained ceii in G. petersii. X900.

Fig. 11. Sections of the stornach [A and BI and intestine [C and Dl in H. dosoides showing

imrnunoreactivity to [A] Anti-SST-14; p]Anti-SST-25; [Cl Anti-GLU;

p] Anti-@Y. X900.

Table 1 Relative abundance of cells immunostainhg with antisera against various peptides in cells within the endocrine pancreas, intestine and stornach of six representative species of osteoglossomorphs. S~ecies Tissue mINS* SST-14 SST-25 NPY aPY PYY GLU GLP -A- ,,.,. .. -,-..- -...... -. ...A.. ..-.**- .. --. . - ...... Ost eoglosstrm bicirrhosum Pm. +++ +tt ttt tt tt tt +tt $-t-t Int.** + tt - ++ tt tt + t Stom ------Scleropages jardini Panc. -kt+ tt ttt tt tt tt ttt -t++ Int.** - - - tt tt t+ tt 4-1- Stom tt " " ------* -. -- - - -.- -A.------L Paniodon buchholzi Pm. +tt u ttt u tt u ttt +U Int.** t tt - t +t u t Stom tt- - - - - St- - - Notopterus chit ala Panc. ttl- +tt ttt t t+ + ++t 3-tt- Int** - tt - tt ++ 3-3- ++ -tt - - - ...... - ...... ,.-..--Stom .... "*- ...... - - - - -...... Gnafhonemus pe f ersii Pm. t++ +t+ ++ t+ ++ -t+ ttt ~t+

Stom - u ++t - - - - - Hiodon alosoides Pm, t+t tt-t t+ ttt ttt tt+ tt+ ut

Stom - tt ti------

(ttt)Abundant ceus; (tt)Moderately abundant cells; (+) Infiequent ; (-) No cells. Int ., anterior intestine; Panc., endocrine pancreas; Stom., stomach.

* See materials and methods for these antisera.

** The pyloric caeca had sirnilar results to the anterior intestine. 3 4 Discussion Epple and Brinn (1975) idenaed different rnorphological types of exocrine-endocrine

pancreatic relationships in fishes including the actinopterygian type, in which the exocrine

pancreas is scattered dong the bile ducts, abdominal blood vessels, outer gastrointestind tract,

gail bladder, and liver. This study confjrmed that the pancreas of six species of

osteoglossomorphs is of the actinopterygian type. Specifïcally, the islet tissue making up the

endocrine pancreas of osteoglossomorphs is principally found in the mesentery connecting the

anterior dintestine, pyloric caeca, gall bladder and siornach.

EppIe and Bh(1986) identified Broc- bodies as consisting of one or a few

large accumulations of islets (principal islets) and their cbsely associated exocrine acini. Falkmer

(1985a) uses the terms Brockmann bodies and principal islets synonymously. This concentration

of large principal islets is pMyfound in more derived teleosts, namely the Ctenosqmmata

(Fdkmer, 1985a). Among other teleosts, there is a ''difhse or disseminated" pancreas where there

are msny small islets scattered among pancreatic acini. The mode of islet distniution in osteoglossomorphs did not quite fit the term 'Brockmnnn bodies", but rather scattered islets within the exocrine pancreas. However, in both 0. bicirrhosum and S. jardini the islets, especidy the large ones, were concentrated at the region of the pyloric caeca close to the junction of the anterior intestine with the stomach, while the rest of the pancreatic tissue contained smaller more dispersed islets. This distriiution of islets in the two arawana species may be considered as

Brockmann-body-like islet accumulations (Epple and Brinn, 1975). This earlier report showed larger islet aggregation in 0.biccirhosum, however, Osteoglossumferreirai had a diBise islet distriiution (Epple and Brinn, 1975). In contrast to the Osteoglossum genus, N. chitalcz and

G.petersii had both large and srnail islets well spread throughout the whole pancreatic region, while P. buchholzi showed wide islet dispersal to the greatest exireme. These resuIts may 35 have some phylogenetic signifïcance, since the Notopteridae (e.g. N chitala) and Mormyridae

(e.g. G. petersio are closely related (Nelson, 1994). Furthemore, there was consistency in

dlstnïiution of the islets in two arawana species (0.biciwhosum and S. jardini). The Panîodon

(e.g.. P. buccholn'), with the most exireme variation in islet distrï'bution, is most distant in the

phylogeny of the Osteoglossomorpha (Li and Wilson, 1996). A tendency for larger islet

aggregations was reported for the osteoglossomorph, Hiodon tergisus (Mc Cormick, 1925) but

Epple and Brim (1975) found variable sized Sets in another hiodontid, Amphiodon alosoides.

Sim.ilarly, H. alosoides of the present study showed large and sdislets scatîered througliout the

pancreatic tissues.

The islet tissue of teleosts contains four hormone-producing cells and, for the most part,

this feature is consistent for all islets within each species (Fakmer, 1%Sa; Epple and Br& 1986).

However, there are some reports of peptides of the pancreatic polypeptide family king absent in

islets near the pyloric region (Norris, 1997) or spleen (Faber, 1985a). The

immunohistochemical results of the present study reveded four cell types in ceIl islets throughout the pancreas in six species of osteogbssomorphs.

In aU the osteoglossomorphs, immunoreactivity to anti-mlNS was observed in the central core of the islets; this feature has been reported in other teleosts and in manmals (El- Naggar et al., 1993; Rombout and Taverne-Thiele, 1982; Abad et al., 1987; Jonsson, 1991).

In 0. biciwhosum, S. jardini, G. petersii, and H. alosoides there was a thicker rim of non- immunoreactive cells at the penphery, compared with islets fiom the other two species. Thus, although the general distri'bution of B cek in the osteoglossomorphs islets appears to be consistent with that of other species of vertebrates, this study shows that some variation occurs even among closely related species. The intense immunoreactivity of B celi granules to m;imm;ilian antisera indicates a sufficient structural similarity between the insulin of 3 6

osteoglossomorphs and that of mammals. This same mammalian antihdy reacted weakly to B

cells of a lâmprey (Youson and Potter, 1993) and this species was eventuaUy shown to harbor

only proinsulin in the islet tissue (Codon et al., 1995b).

It has been accepted that insulin cells leave the epithehofthe digestive tract early in

both ontogeny and phylogeny (Fakmer, 1995). Among the vertebrates, this is well illustrated in

cycIostomes (Faher et al., 1984; Youson and Cheung, 1990). However, cells immunoreactive

for insuiin are present in the digestive tract of tuales (Gapp and Polak, 1WO), sea bass (Hemandez

et al., 1994), a southem hemisphere lamprey (Youson and Potter, 1993), and the gar (Groff and

Youson, 1997). In osteoglossomorphs, two of the species under study showed infrequent

irnmunoreactivity in the intestine, but P. buccholzi reveded immunoreactivity to mINS in

epithelial cells of both the intestine and stomach. The other species showed no inimunoreactivity

in either region. Since P. buccholzi is considered to be distant in taxonomie tenns fiom the

other species (Li and Wilson, 1W6), this result might be of phylogenetic si~cance.This may be an interesting species to study with respective to the ontogeny of the islet tissues.

The anti-SST-immunoreactive ceils are scattered throughout the central portion of the islets in the six species, and this is consistent with the observations descriid in several teleost species (Abad et al., 1986; Lozano and Agulleiro, 1986; Nozaki et al-,1 988a) and an ancient actinopterygian, the gar (Groff and Youson, 1 997). In contrast, D cells are usually located at the islet periphery in mammals, toads, fiogs, and some species of fish (EL-Salhy et al,1 982; Abad et al., 1986, EL-Naggar et al. , 1993). There is strong immunoreactivity to anti-SST-25 in five species but weak stahhg with this ant~bdyin G. petersii. Whereas moderate immunoreactivity to SST-14 was observed in S. jardini and P. buccholzi, G. petersii and N. chitala showed weak immunoreactivity. In contrast, strong immunoreactivity to SST- 14 was observed in H alosoides 3 7 at the islet periphery. The intestine and stomach showed an ever-greater variation in

immunostaining with these two SST-antisera between and within species (Table 1). For example,

immunoreactivity to oniy anti-SST-14 is observed in O. biciwhosum, P. buccholzi and N chitala

intestinal epithelial cells and only anf-SST-25 stained stomach cek in S. jardini. At the other

extreme, immunoreactivity to both antisera is present in both the intestine and stomach cek of

G. petersii. This latter situation is similar to ~~~IMIs(Baldissera et al., 1985). A Mer

variation is seen in H. alosoides where both antisera irnmmoreacted only in the stomach. These

results do not provide a clear phylogenetic picture among the species of osteoglossomorphs.

The difference in this immunostaining may be due to the nature (e-g. titer, number of binding

sites) of the antibodies, but positive controls for anti-SST-14 showed strong specificity.

Tmmunostaining with dinerent SST antibodies (Youson and Potter, 1993) has provided initial

insight into the structure of the peptide, which was later codkmed foliowing its isolation (Conlon

et al., 1995a and b).

Anti-SST-25 used in this study bas been characterized in previous studies, and it appears

to possess two binding sites, one directed against the 1I N-terminal amino acids of salmon

SST-25 and one directed against the 14 C-terminal amino acids (Plisetskaya et al., 1986). In the

present case, the weaker affinity of the D cell granules to anti-SST-14, suggests that

osteoglossomorphs may contai.a SST molecule with a stronger affinity to the binding site of the

N-terminal amino acids of anti-S ST-25. The osteoglossomorph S ST may have some resemblance

to SST-25 or be a prosomatostatin molecule. Some interspecific variations in prosomatostatin processing has been noted in lampreys (Andrews et al., 1988; Codon et al., L995a and b).

Therefore, it wiU be of interest to examine SSTs in severai of the osteoglossomorphs. The variable immunoreactivity to anti-SST-14 and anti-SST-25 may indicate some variations in processing of the prosomatostatin molecule in different orggans between and within species. 38

Larhammar (1996) reported that the structurai similanties between the aPY, NPY and

PYY suggest that all have a single ancestral gene. Furthemore, Codon (1 995) proposed tbat the ancestrai gene underwent a series of duplications resulting in separate genes encoding for each of these three peptides. Tmmunoreactivity for these peptides is observed in some of the peripheral cells of islets in the six species of osteoglossomorphs. It is &ely that the immunoreactivity is within the same ceil type and that the same cytoplasmic peptide had an epitope which was recognized by the different antisera. These redts are consistent with that of the gar (Groff and

Youson, 1997). Multiple colocalizations of P W immunoreactivity have also kendescribed in fis4 where several CO-storedpeptides (aPY, NPY, PYY) have ken found in the same cellular type (Cheung et al., 199 1).

In salmon, both GLU and GLP have been isolated and sequenced, and they are processed in A celis of the endocrine pancreas (Nozaki et al., 1988b). Osteoglossomorphs showed immunoreacti~to both GLU and GLP antisera in the pancreatic islet ceils in ali the six species stuclied. Ceils of the intestine, but not those of the stornach, were immunoreactive to GLP in the six species. However, intestinal cells irnmunoreacted with anti-GLU sera in S. jardini,

N chirala and G. petersii, 0. bicirrhosum, and N. alosoides but not with P. buchholzi.

Conversely, immunoreactivity to GLU and GLP are demonstrated m the intestinal mucosa of the lamprey and not in the idet organ (Elliott and Youson, 1988; Cheung et al., 1991).

In the present study, the immunoreactivity in both pancreas and intestine to both GLU and

GLP antisera suggests either similar molecular fonns of these peptides in both organs, or an epitope in a single moiecule which is recognized by both antisem On the other hand, the absence of the immunoreactivity to anti-GLU in the intestine and its presence in the islet organs of some of the species studied for the same peptide, suggests that there may be a different processing of the 39 proglucagon molecule in the two organs. It is perhaps noteworthy that P. buchholzi, the most distant among the species studied, once again showed the greatest variance in immunoreacti~to glucagon family antisera (Table 1). Studies are presently underway to determine the nature of some of the glucagon-family peptides in the GEP systern of osteoglossomorphs.

Within the islets, there iç a strong immunoreactivity to GLU and GLP antisera in the peripheral cek of osteoglossomorphs, but they are also found scattered through the central portion as well. Since this distn'bution is consistent with that of the pancreatic polypeptide My it might suggest a CO-locaiizationof peptides between the latter and the glucagon family in some of the species, as is the case in the gar (Groff and Youson 1997, 1998). In generai, CO-existence of peptides belonging to the pancreatic polypeptide family and glucagon fdyhave ken reported previously in the teleosts (Abad et al., 1987 and Agulleiro er al., 1993) and in Iamprey

(Cheung et al., 199 1). However, since there are more numerous GLU- and GLP-immunoreactive cells than PP-fdy cells at the periphery, it is most likely that some cells at this site are exclusively A ceh. We will attempt to resolve this question of coexistence of peptides f?om the pmcreatic polypeptide and glucagon fdesthrough immunoc ytochemistry and electron microscopy.

In conclusion, the present study has shown rhat the GEP systern of the O steoglossomorphs contains the various endocrine cek that have kenrecognized in other teleosts. However, there are some subtle dinerences among these species that may reflect either their separate evolutionary histories, or merences in the ontogeny of the cells of their GEP system with variations in po sttransiational processing of pro hormones among some of the regdatory peptides. Chapter 2

Immunohistochemistry of the endocrine cells in the gastro- entero-pancreatic system of a generalized teleost (Cutosfomus commersoni) and a more derived teleost (Ambloplifes rupestris). Introduction

The importance of the gastro-entero-pancreatic (GEP) system to the individual fkh

species and how its tissue and peptides relate to that of the other higher vertebrates have attracted

the interest of many investigatoe (for review see Youson and Ai-Mahrouki, 1999). The

hormones produced by this system play an important role in metaboh that in turns effects hh

growth (Plisetskaya and Mommsen, 1996). Moreover, the bcalization ofthe endocrine celis

within the GEP system, their origin, how they developed, and when they first appeared are

important features that give us a better understanding of the developmental history of the GEP

system in vertebrates. Comparative analysis of this endocrine system can be helpfül in further

clarincation of how this system and its peptides functioa

Several studies have identïfied the structure of the GEP system in euteleosts

(McCormick, 1925; Cardo et al., 1986; Abad et al., 1992; Gomez-Viscus et al., 1996, 1998).

However, in a recent review Youson and Al-Mahrouki (1 999) pointed out that our picture of the phylogenetic development of the GEP system in teleosts has been derived hmthe investigation of SMnumbers of teleost species. In the previous chapter, the detailed identification of the

GEP system morphology, histo logy, and the localization of its peptides in the osteoglossomorphs

(which are believed to be a basal teleost group) was presented. In this study, the white sucker

(Catostomus cornmersuni) has been chosen as a representative of a generaiized teleost. As far as

1know, McCormick (1925) and Youson and Al-Mahrouki (1999) provide the only reports of the

GEP system in the sucker, and both are only brief descriptions. Furthemore, the rock bas

(Ambloplites rupeshis) has ken chosen as a more derived teleost, and to my lmowledge its GEP system has not been previously investigated. It is also planned to study the GEP peptides of these two species using a molecular approach. For cornparison with the peptides f?om the osteoglossomorphs; these wiu be used in the phylogenetic analysis. The histology of the system, and the localkation and Immmoreactivity of its peptides would be a usefiil adjunct to the

molecular analyses. In this study, the GEP system in the two species is identifie4 and the results

are compared with those from the osteoglossomorpbs.

Materials and methods

Animais: Adult specimens of Catostomus cornrnersoni (white sucker) were trap-netted in the

Lake Nipissing watershed (North Bay, Ontario, Canada) and live specimens were obtained fiom a local hhsupplier. Adult samples of ArnblopZites nrpestris were O btained fÏom Diiffins Creek in Southern Ontario. Live samples were sacrifïced by an overdose of 0.05% tricaine methanesulfonate (MS-222) in their holding water. The viscera were obtained and ked in

Bouin's fluid for 20-24 hr then stored in 70% ethanol-

Tissue preparation:

Histology: A large portion of the viscera containhg the stornach, pyloric caeca, gall bladder, intestine and connecting mesentery were excised and fked intact for 20 hr in Bouin's fluid.

Foliowing fixation and storage in 70% ethanol for a day, the tissues were eventually embedded in paralEu wax, serially sectioned at 7 pthichess, and stained with hematoxyiin and eosin (H&E).

ImmunohisfochemCsîry: Tissues were prepared as above, and adjacent 7 pm sections were p!aced on different glas slides. The site of immunolabelling activity was revealed by a commercial rabbit histostain-SP kit (Zymed-lab-SA System, San Francisco, CA). This kit uses a bio tinylated second antibo dy, a horseradish peroxidase-strept avidin conjugate, and a substrate- chromogen mixture to demonstrate antigen in celk or tissues. Positive stWIiTig is indicated by red staining and there is a hematoxylin counterstain. The primary antisera used in this study were the same as those presented in Chapter 1. Positive Mmunoreactivity was detected in cornparison 42 with that of rat pancreas, and negative controls consisted of the replacement of prirnary

antisera with phosphate-buBered saline (PBS) pH 7.6, and of antisera previously absorbed with

excess antigen (Appendix 1).

Results

General morphology

The general anatomy of the viscera in both species was similar. A large number of

pyloric caeca were observed arising fiom the most caudal portion of the stomac. unlike in the

osteoglossomorphs where only one or two pyloric caeca were O bserved. Principal islets were

visible to the naked eye in the connective tissue connecting the pylonc caeca, the gall bladder,

the Liver and the antenor intestine. C. commersoni had one or two principal islets surrounded by

a thin rirn of exocrine acini (Fig. 12 A & B); one principal islet was located within a concavity of

the liver. A few scattered small islets are also found among the exocrine acini in this species.

In A. rupestrïs there were two or three principal islets associated with some exocrine tissue

(Fig. 12C), as well as large and small islets fUy embedded and surrounded by the exocrine

acini. In both species, the epithelial cells of the islets stain lightly with hematoxylin and eosin

compared with the acinar cells of the exocrine pancreas (Fig. 12C). The term Brockmann body is commonly used to refer to the largest accumuIations of endocrine tissue in the fihpancreas, which is associated with some or little exocrine tissue. Therefore, Brockmann bodies are present

in the two species.

Immunohistochemistry

A summary of the immunohistochemical results of the different antisera in the three organs of the two species is presented in Table 2, where the relative number of celis staining with the antisera are compared. B-ceUs in the islets of C. commersoni were moderately immunostained with anti-m-INS, the reaction king more pronounced in smaller islets than in larger ones (Fig. 13A and B).

However, in A. mpestns the immunoreaction to anti-m-INS was throughout the center of the islet and was the same in both smd and large islets (Fig. 14A). It seems that the same D-ceils immunoreacted with anti-SST-14 and anti-SST-25 in C. comrnersoni, and the cek were distriiuted throughout the idet (Fig. l3C). In A. rupesîris, anti-SST- 14 immmo-stained ce& were scattered throughout the islet (Fig. 14B), but anti-SST-25 resulted in immunoreactive cells

&y located at the penphery of the islet with very few in the central region (Fig. 14C).

A-cefi irnmmoreacted with anti-GLU and anti-GLP in both species, and were primarily localized in the periphery of the islets (Fig. 15A); a few ce& were scattered tbroughout more central regions at the core ofthe islet (Fig. 13D and 15B).

F-ce& immunoreacted with anti-aPY, ami-NPY, and anti-PYY was sllnilar in both species and was located at the very periphery of the islets (Fig. 15C and D). Some polypeptide cells were also observed among the exocrine acini (Fig. 15C). The observations of A- and F-ceils localized at the periphery of the islets may irnply a colocalization that should be investigated

Mer.

As in Chapter 1, the relative abundance of stainûig cells was compared only in the stomach and intestine (Table 2).

Tmmunoreactivity to the different antisera in the pyloric caeca and the anterior intestine was similar in the two species, and was consistent with the structural sbdarity of the two organs. The immunoreactive cells in the intestine were randomly scattered among other epithelial ce& of the mucosa of the anterior intestine. No imrnunoreactivity was detected in the 44 posterior intestine of the two species.

No immmoreactivity to anti-m-INS was detected in either the intestine or the stomach of

the two species. In both species, ceils immunoreactive to anti-SST-14 and anti-SST-25 were

observed in the intestine. Only a few cells of A. ncpestris stomach immunoreacted with anti-

SST-25. Anti-GLU and anti-GLP immunoreactivity was detected in the intestine of both

species but not in the stomach (Fig. 16B). The use of pancreatic polypeptide antisera showed rnoderate immunoreactivity in the intestine of the two species. In A. rupestris no immunoreactivity was detected in the intestinal ceils with anti-NPY. Similarlly, no irnmunoreactMty was detected to any of these aatisera in the stomach of the two species. Fig. 12. Sections through islet tissue in a Brockmann body. [A] A Brockmann body in

C. commersoni, is surrounded by a narrow rim of exocrine acini (arrow) and is located

within a concavity of the liver (L). Hematoxylin-eosin. X15O. PIA higher magdïcation

of (A) showing the liver (L), and exocrine (arrow) and endocrine cells (arrowhead)

X 3 50. [Cl A portion of a principal islet in A. rupestris, the endocrine tissue (En) is

surrounded by the exocrine tissue @m.Hematoxylin-eosin. X2 00.

Table 2 Relative abundance of cells Unmunostaining with antisera against various peptides in cells within the endocrine pancreas, intestine and stomach of Catostomus commersoni and Am bloplites rupestris.

Ant isera

Stom. - - - " - - " - A. rupestris Pan. ttf tt+ +++ tff 3-t-t- +U -kt+ +tt

Int.* - +++ 4+t +t t-t tt - t

(t+t)Abundant cells; (tt)Moderately abundant ceiis; (+) Idequent; (-) No cells. Antisera against the following peptides: dNS, marnrnalian insulin; SST- 14, marnrnalian sornatostatin-14; SST-25, salmon sornatostatin-25; GLU, glucagons; GLP, glucagons-like-peptide; aPY, anglerfish polypeptide tyrosine; NPY, neuropeptide tyrosine; PYY, polypeptide tyrosine tyrosine. Int., anterior intestine; Pan., endocrine pancreas; Stom., stomach.

* The pyloric caeca had similar results to the anterior intestine. Fig. 13. Sections through the islet tissue of C. commersoni immunoreacting with various

antisera [A and BI demonstrate the immunoreactivity to anti-m-INS throughout the

parenchyma in a sdislet [A] and in a large islet @3]. [Cl D-ceUs immunoreacting with

anti-SST-25; the principal islet is within a concavity of the liver (L). p]Portion of a

large islet showing Mmunoreactivity of A-cells to anti-GLU sera X350.

Fig. 14. Sections through the islet tissue of A. rwpestris. [A] Immunoreactivity to anti-m-INS,

note blood vesse1 (bv). @ & C] show immunoreactive D-cells in principal islets. @3]

Immunoreactivity to auti-SST-14, is observed in ceIls scattered throughout the islet. [Cl

Anti-SST-25 immunoreactivity is rnainly localized in cells at the islet periphery. X350.

Fig. 15. Sections through the islet tissue of A. rupesîris demonstrating Mmmoreactivity to

various antisera. [A] Anti-GLP imnzunoreactivity and [BI anti-GLU imrnunoreactivity in

A-cek mainly at the isIet periphery with few scattered towards the central core region in

the midde of an islet. [Cl Anti-PYY immunoreactive F-cek Iocated at the edge of a

portion of a large islet, with few scattered immunostained ce& among the exocrine acini

(arrowhead). @3]Anti-aPY imrnunoreactivity is similar to that in Fig. 1SC. X350.

Fig. 16. Two sections through the intestinal tissues of A. rupesms and C. commersom'. [A]

Anti-SST- 14 immunoreactivity in an epithelial celI (arrowhead) of A. rupestris. p] Anti-

GLU immunoreactivity in some intestinal ce& (arrowheads) in C,commersoni. X900.

Discussion

The largest aggregation of islets (endocrine tissue) in the fish pancreas associated with

some or little exocrine tissue is called a Brockmanu body/principal islet, as defined by many

investigators (Epple and Brinn, 1986; Faber, 1985 a, b, 1995; Yang and Wright, 1995; Mkglio

and Putti, 1998). A principal islet is best defined as a large islet which may or may not be

associated with other islets in a Broclmÿuui body (Youson and Al-Mahrouki, 1999). The

principal islet/Brockmmn bodies are mainly associated with some exocrine tissue and located

near the gall bladder and liver. These cases are mainly found in more derived euteleosts that

were referred to as Ctenosquamata (Epple and Brinn, 1975). Phylogenetic development tends to

move toward hhgone or few large principal islets (Falkmer, 1995; Youson and Al-Mahrouki,

1999). The presence of a large concentration of pancreatic islets among the viscera of fis4

referred to as Broclanann bodies, is reported in several representatives of different euteleost

groups: for example, fiom the orders Cyprinodontiformes, Lophiformes, Scorpaeniformes, and

Perciformes (Boquist and Patent, 1971 ; Johnson et al., 1976; Klein and Lange, 1977; Patent et

al., 1978; Klein and Van Noorden, 1980; Stefan and Falkmer, 1980; Maglio and Putti 1998).

In this study, one or two principal islets were observed in C. commersoni, which belongs

to the order cyp~odontiformes.The principal islet was surrounded by a thin rim of exocrine tissue, in addition to small islets scattered throughout the exocrine pancreas. This pattern of islet distri'bution is similar to that reported in Barbus conchonitlî (Rombout et al., 1979), which is &O a cyprin6dontiforme. A single principal islet (l3roclanann body) with very Little exocrine tissue around it was descriid in other cyrinids (Klein and Langer, 1977; Klein and Van Noorden,

1980). However in the goldfish, Carassius (ano ther weIl-investigated cyprinid) islets of varying sizes are scatîered throughout the exocrine pancreas, in addition to large visMe islets found in the mesentery and these are considered as principal islets (Kobayashi and TakahaShi, 1970). In 52 A. rupes~rz3(Perciformes) two principal islets associated with more exocrine tissues were

observed, as weil as large and srnail islets fully ernbedded and surrounded by the exocrine acini

This fïnding varies fkorn that descrkd in other Perciformes (Maglio and Putt& 1998), where a

limited exocrine tissue was associated with a single principal islet. However, in other

Perciformes such as the sea beam and sea bass, single principal islets were reported,

accompanied by other variable-sized islets and exocrine tissues (Abad et al., 1986; Lozano et al.,

1991 a, b). The pattern of distriiution of islets within the exocrine tissue in these two species fits

in with the term Broclmiann body (Youson and Al-Mahroukj, 1999). The presence of

Brockmann bodies is suggested to be an advanced form of the endocrine pancreas in its phylogenetic development. In contrast, Br0ckman.n bodies were not observed in the osteoglossomorphs (Chapter 1), where both large and small islets were found scattered throughout the exocrine pancreas. In temis of phylogenetic developrnent this may irnply that the pattern of endocrine tissue distribution within the exocrine tissues in osteoglossomorphs is less advanced than that in the white sucker or the rock bas.

The presence of four principal cell types within the endocrine pancreas has ken described in many teleost species (Klein and Van Noorden, 1980, Maglio and Putti, 1998), and in Chapter 1. A-cells immunoreact with anti-glucagon fdysera, B-cek immunoreact with anti-insulin sera, D-cells immunoreact with different anti-somatostatin sera, and F-cek immunoreact with anti-polypeptides family sera. These cek have a consistent pattern of localization through the islet organ, particularly, the D-cells, which have an influence on peptide secretion fiom A and B-ceils (Maglio and Putti, 1998). In both C. cornmersuni and

A. rupestris, it seems that the four principal celI Spes are present. B-cells were mainly localized in the central core of the idets. However, less abundant cek were observed in C. cornmersuni, especially in larger islets, which may account for what seemed to be the presence of more D- and 53 A-cells. D- and A-cek were present in both species throughout the kIets and also at the periphery. The pattern of D-cell distn'bution redting fkom the use of different anti-SST sera may indicate the presence of two D-celI types sirnilar to those reported in many euteleosts

(Nozaki et ai., 1988a; Abad et al., 1992) and in the osteoglossomorphs (Al-Mahrouki and

Youson, 1999). However, the suggestion of the presence of more than one D-ceil in any of these two species needs to be investigated further. Either double immunostaining andior electron microscopie imrnu11ocytochemistry is required.

In both species, F-celis were dyfound at the very periphery of the islet. The localization of F-cek and some A-cek may suggest a co-existence between the two families of peptides (the gIucagon and the pancreatic polypeptide families). This co-locaiization was previously reported in teleosts (Abad et al., 1987; Agdeiro et al., 1993) and has been more recently demonstrated in the gar (Groff and Youson, 1997; 1998) and bowfin (Youson et al.,

200 1). To investigate this co-localization merstudies are required, as descnibed above for somatostatin and D-cell types. However, as will be descriid in Chapter 3, this co-existence between the two fdesof peptides in the osteoglossomorphs were no t found using immuno cytochemistry.

Generaily, irnmunoreactivity to the various anti-sera was iimited to the stomach in both species. Only a few cells in the stomach of A. rzrpestris immunoreacted with mi-SST-25.

Insulin immunoreactive cek were reported in the stomach of P. buchholzi (Chapter 1).

Tmmunoreactivity to anti-glucagon sera was reported in the sea bass stomach and intestine

(Gomez-Visus et al., 1998). Cells immunoreactive to anti-S ST sera were also reported in various species includmg the turbot stomach (Reinecke et al., 1997) and also in the osteoglossomorphs

(Chapter 1). Since the immunoreactivity to various antisejra in the digestive tract does not 54 represent a developmental pattern among the less and more derived species, it does not seem to reflect any clear phylogenetic pattern-

In conclusion, the present study has shown the presence of Brockman.bodies in both the white sucker and the rock bass. These bodies represent a more derived form of the islet organ, Yi cornparison with the islet organ in the osteoglossomorphs. This result implies that the endocrine pancreas of the osteoglossomorphs is less advanced than that of these two species.

This study also reported the presence of various endocrine ceils within the GEP system, and aithough the identities of the cek are not absolutely codkmed, it opens the door to fbture ultrastructural and imrnunocytochemical studies at the cellular level, or for the usage of double labeiing technique at the light microscope level- Most importantly, this study has helped to idente the location of the endocrine pancreas and the peptides produced by the GEP system, as a necessary step for tissue isolation in the molecular biology approach that will be covered in

Chapters four and five. Chapter 3

Ultrastructure and immunocytochemistry of the islet organ of Osteoglossomorpha (Teleostei).

From

Al-Mahrouki, A.A. and Youson, J.H. (1999). Ultrastructure and immunocytochemistry of the isïet organ of Osteoglossomorpha (Teleostei). Gen. Comp. Endocrinol. 116,409-421. (with permission of the Academic Press) Introduction

The ancient origin of members of subdivision (superorder) Osteoglo ssomorpha among

the teleosts (Patterson and Rosen, 1977), their pivotal position in teleost evolution, and the

unsettled debate over their taxonomy, make osteoglossomorphs an interesthg case for

investigating the phylogenetic development of organ systerns. hcluded arnong these systems is

the endocrine pancreas (islet organ) and its peptides.

Until recently, only passing attention has been given to the pancreas of members of the

order OsteogIossiformes (McCormick, 1925; EppIe and Brinn, 1975; Langer et al., 1979).

However, in a detailed immunohistochemical study of the gastro-entero-pancreatic (GEP) system

(Chapter 1) it was reported that various species of osteoglossomorphs showed a structural diver sity that reflected their distinctive taxonomic groupings. This study also suggested that meranalysis of the endocrine cell types of the islet tissue and their immunoreactivity might provide additional data useful for taxonomic comparisons.

The different ceil types in the islet organ of teleosts have been identif5ed in several immunocytochemical and ultrastructural studies (Lozano and Agdeiro, 198 6; Beccaria et al.,

1990; Agulleiro et al., 1993; Gomez-Visus et al., 1996). Recently this type of investigation was extended to include the gar (Grog and Youson, 1W8), a nonteieost actinopterygian (ray-hed fish) of ancient lineage (Gardiner et al., 1996).

This study aims to provide further identification of the cell types within the islet organ of the Osteoglossomorpha, to extend our knowledge of the structural diversity or conformity of this organ among actinopterygian khes in gened, and among the taxonomic groups of osteoglossomorphs in particular. Materials and methods

Anid

Adult specirnens of Osteoglossum bicimhosum (silver arawana), Pantodon buchholzi

(butterfly fish), Notopterus chitala (feather fin Metkh), and Gnathonemuspetersii

(elephantnose) were obtained fiom several commercial fish aquaria within the Toronto area A total of 21 anunaIs (minimum of three for each species) were sacrificed by an overdose of

0 -05% tricaine methanedonate (MS-222) in their holding water.

Tksue preparuîion

Routine electron microscopy

The regions of the pancreas containing the islet organ (Chapter 1) were excised fiom the viscera, diced, and fked in a mixture of O. 1% glutaraldehyde and 4% paraformaldehyde in O. 1

M phosphate buffer, pH 7.4 (Millonig 1961), ovemight at 4OC. Mer this primary fixation, the tissues were washed in bufEer and subsequently postfïxed in 1% phosphate-buEered osmium tetroxide (Os04) for 1 hr at room temperature. The tissues were then washed in buffer, dehydrated in a graded series of ethanols, and embedded in Spurr's low-viscosity embedding medium (Poly-sciences Inc., Warrington, Pa.). Thick sections (0.5 p)were mounted on glas slides, stained with 1% toluidine blue in saturated tetraborate, and were used to locate islet tissue using light microscopy. Ultrath sections were cut with glas knives, mounted on uncoated copper grids, and stained with uranyl acetate and lead citrate. Sections were examined in a

Siemens Elmiskop- 102 electron microscope. Immunocytochemistry

Tissues were primary ked as descnid above for 2 hr. Unosmicated tisswes were

dehydrated and embedded in Spurr's low-viscosity embedding medium. Ultrathin sections were

mounted on nickel grids. The nonspecific bbiding sites were blocked by incubating the grids on

CAS-block (Zymed Laboratories Inc. San Francisco, CA) for 10 min. The tissues were then

incubated for 2 hr at room temperature in one of the foilowing antisera: (1) guinea pig anti-

bovine insulin (anti-mINS, a kind gifi of Dr.C.Yip, University of Toronto) diluted 1:XO; (2)

rabbit anti-lamprey somatostatin-34, diiuted 1500 (anti-SST-34); (3) rabbiî anti-salmon

somatostatin-25 (anti-SST-25) diluted 1500; (4) rabbit anti-somatostatin-14 IgG *action (anti-

SST- 14, LINCO Research, Inc., St. Charles, Mo) diluted 1: 1000; (5) rabbit anti-saIrnon

polypeptide tyrosine tyrosine (anti-PYY) diluted 1500; and (6) rabbit anti-porcine glucagon

(anti-GLU; Zymed Laboratories Inc., San Francisco, CA) diluted 11500 (anti-S ST-25, anti-SST-

34, and anti-PYY; are kind gifts of Dr. E. M. Plisetskaya, University of Washingtmn). The

incubation in the plimary antiserum was followed by several washes in phosphate-buffered

saline (PBS) for 20 min, and incubation for 10 min in CAS-block solution. Grids were then

blotted gently and transferred to 10 nm protein A-gold (Sigma Chemical Co. St. Lais., Mo., diluted 1:20 with the blocking solution) for 1 hr at room temperature. Several wasnies in PBS followed and then a final wash in distilled water was applied. The tissue was then connasted with uranyl acetate and lead citrate. For double immunogold iabehg, the same prmcedures as above were followed, except that the grids were dried before the ha1distiiied-watcer wash and the contrast staining. The grids were turned over, incubated in the CAS-blocking salution for 10 min, and then floated on a drop of the second antïibody for 2 hr. A different size (20 nm) of protein A-gold (Sigma Chemical Co.) was used for the second antibody. Upon the completion of 58 the second Mmunogold reaction, the grids were contrasted with uranyl acetate and lead citrate.

The specifïcity of the reactions O btained was tested using the following controls: (1) by replacing

the primary antisenun with PBS or (2) with antiserum absorbed with excess comphentary

antigens, or other antigens (if available) to test for cross reactivity.

Resdts

The findings in the ultrastructural study support the earlier view (Chapter 1) that there are

four primary islet ceIl types (A, B, D and F cells) among the four species examuied

(0. bicinhosum, P. buchholzi, N. chilnla, and G. petersii).

The above conclusion was based on distrt'bution patterns of cells previously descriid in

the immunohistochemical results, on what appeared to be cell-specifk grandes, and on the

imrnunocytochemical results of this study. In addition, fine structural observations revealed

scattered ce& among the exocrine aci.which were mainly B- and D-cek.

The endocrine cells could easily be distinguished £iom the exocrine cells due to the large secretory granules of the latter (Fig. 17). Furthemore, the endocrine cek were in close association with blood vessels and nerve terminais (Figs 17, 18 and 19). The islet ceils had numerous cornmon features, including a large oval nucleus and large vacuoles (presumably only containing lipid); the latter were particularly pronounced in B-cells (Fig. 19). Weil developed cisternae of rough endoplasmic reticulum and a Golgi apparatus were also observed (Fig. 19).

However, no table Merences were O bserved in the morphology of the secretory granules and mitochondria among the cells (Figs. 19-23) and their immunocytochemistry (Figs 24 and 25).

In three of the species examine& the B-cells contained granules 93 -3 37 nm in diameter

(meanirement made at the longest axis of the granule) each with a loose-fitting membrane surroundhg a hexagonal or sphencal-shaped rnatrix (Figs. 18, 19 and 2 1). However, in O. bicirrhosum, the diameter of B-cell granules had a range of 154-338 nm and the shape of the granule matrix varied greatly between round, oval, rectangular, trhgular, rod, and stacked rods

(Fig. 20). The immunocytochemical techniques confïrmed the fine structural results, where

B-cell granule cores immunoreacted with ad-mINS serum in all four species, indicated by the labeling of protein-A gold particles (compare Figs. 25a and b). The immunoreactivity was comparatively weaker in B-ceil granules of O. bicin-hosum.

The islet organs of the osteoglossomorphs seemed to have two types of

D-cells scattered throughout the islets (Figs. 18,22, 24% b). D 1-cells had similar sphericai granules of mked electron density in the three species, but the granules were less electron dense in O. biciwhosum (Figs. 20 and 24a). Generaily, the granules' diarneter ranged fiom 72-320 nm in aii the species. In contrast, granules in other D ce&, hereafter referred to as DX-ceils O;igs.

22,24b), were highly electron dense, elongated or tear-drop shaped with a tight limiting membrane, and tended to bave a more limited diarneter range of 69-240 nm in all examined species. The anti-SST immunoreactivity confirmed the presence of two D-cell types. Dbcell grande matrices immunoreacted with ad-SST-25 and anti-SST-14 (Fig 2%); this feature was also revealed by the double 1abeling technique (Fig. 24a). DX-ceils had differently shaped grande cores compared with D 1-cek, and they immunoreacted with anti-SST-34 (Fig. 25d).

Furthemore, the absorption of anti-SST-34 with SST-14 antigen prior to application did not abolish the labeling of DX-cell granules (Fig. 24b).

A-cells were located at the periphery of the islets in all four species, some A-cells were scattered towards the middle of the islets in P. buchhohi and in N. chitala. The A-ceil grandes had a moderate electron density and the diameter ranged between 93-3 85 nm (Fig. 23), with a tight-fitting, Iïmiting membrane and oval or circular matrix profiles. lcmmunoreactivity to anti-GLU serum cohedthe identity of the A-cek (Fig. 25e). 60 F-cek were found at the very penphery of the islets and they possessed much smaller granules ( 48-200 nm in diameter). The electron density of the granule cores was iess than that of the A-ceUs (Figs. 21 and 23). The features of the F-cell were similar in the different species examined F-cell grande cores immunoreacted with anti-P W sem(Fig. 25f) but not with anti-GLU serum. The double labeling technique indicated no coexistence between the glucagon family peptides and the pancreatic fimily peptides in all of these species (Fig. 25g).

These immunocytochemical results with A and F cek were Merverined by absorbing the antiserum with excess complimentary antigens, or with other antigens to test for the cross react ivity. 6 1 Fig. 17 Light micrograph of a 0.5 psection of a portion of an islet in the pancreas of

O. bicirrhosum staiued with toluidme blue. The relationship between the exocrine (EX)

and the endoc~etissue (en) is illustrated .Note the large oval nucleus (arrows), large

lipid inclusions (arrowheads) and blood vessels (bv). X 1000.

Fig. 18 Lo w magnification electron micrograph of a portion of an islet in N. chitala, showing

B-ceik (B), and D 1-cek (Dl). bv, bIood vessel. X 10 000.

Fig. 19 Portion of a B-cell and several unknown cell types in G. petersii. In the B-ceil are a

Golgi apparatus (GA), mitochondria (m), nucleus 0,a Iipid body (lb), and cisteme of

rough endopiasmïc reticulum (RER). Note the nerve fibers (nt) among the cek.

X 20 000.

Fig. 20 B-cek (B) and a D 1-ceU @ 1) in 0.bicirrhosum. Note the polymorphism of granule

cores in B-cells (arrows). The Dl grandes have a homogeneous granule core with

moderate electron density. Nucleus (N). X 25 000.

Fig. 21 F-cell (F) and B-ceil @) in G. petersii. Compare the homogeneous, more sphencd,

granule cores in the B-ceIls (arrowheads) with those in Fig. 20. The B-ceil is shown with

rough endoplasmic reticulum (RER). Note the unusual shaped mitochondria (m) and the

smaller granules in the adjacent F-ceil (F). X 20 000.

Fig. 22 DifEerent D-cells in P. buchholzi. Note D 1-ceU type @ 1) wÏth round larger granule

cores, compared to the DX-ceIl type (DX) with rnany rod-like granule cores. Also shown

are mitochondria (m), nerve termkk 0,rough endoplasrnic reticulum (RER),

nudeus (N) and nucIeolus (Nu). X 25 000,

Fig. 23 Portions of A (A) and F-cells in 0. bicirrhosum. Note the large round and oval

granule cores with tight fitting, linziting membrane in the A-cells. Grande cores of F-

ceil are much smaller. X 25 000.

Fig. 24 a) Double labeling of D 1 -ceU granules in 0.bicrn.hosum. The 1 0 nm particles

(arrowheads) label the immunoreactivity to anti-SST-14 and the 20 nm particles (arrows)

label the immunoreactivity to anti-SST-25. Note the iuilabeled grande cores of an

adjacent A-cell (A). X 48 000. Inset. Higher magnincation of the granules indicated in

the box. X 66 000.

b) The &hg of DX-ceU granules in N. chitala with 20 nm protein-A gold particles

(arrowheads) after incubation in anti-SST-34 pre-absorbed with SST-14 antigen,

The adjacent B-cell (B) grandes did not react with the anti-senim. X 48 000.

66 Fig. 25 a) Protein-A gold Iabeled granules of B-cell (arrowheads) in G. petersii following

anti-mINS incubation. Compare the granule cores with those in Fig. 2 1. X 63 000.

b) Following incubation with anti-mINS serum, goid particles are present over the

polymorphic cores of grandes (arrowheads) of a B-cell in 0. biciwhosum, indicating the

presence of insulin (compare with Fig. 20). Note the smaller numbers of particles over

the granules compared to Figure 25a. X 50 000.

c) D 1-cell type in 0. bicirrhasum labeled with protein-A gold particles, foiiowing an

incubation with SST-25 antisenun (compare with Fig. 20). Note the unlabeled granule

cores fiom a DX-cell. X 50 000.

d) Following incubation with SST-34 antisenim, granules of DX-cek in N. chitala label

with gold particles indicating the presence of SST over hi& electron-dense granules

(arrowheads) w-ith some rod-shaped profiles (Compare with Fig. 23). X 90 000.

e) Following double incubation, A-cell granules in 0.biciwhosum label with anti-GLU

(20 nm particles- arrowhead) but not with anti-PYY (10 nm particles). Compare with

Fig. 22. X 63 000.

f) Following double incubation, granules in a F-cell fiom G. petersii label with anti-PYY

(10 nrn particles, arrowheads) but not with anti-GLU (20 nm particles). Compare with

Fig. 22. X 63 000. Inset. Higher magdïcation of the granules indicated in the box.

X 90 000.

g) Two adjacent A-cell (A) and F-cell (F) in 0. bicitrhosum. Following double

incubation, the 20 nm particles (A-cell granules, arro wheads) label the immrinoreactivity

to anti-GLU and the 10 nmparticles (F-cell granules, arrows) label the immunoreactivity

to anti-PYY. X 50,000.

67 Discussion

The present data indicate the presence of the four major cell types specificaliy localizing only one of the following peptides: insuiin (B-cell), somatostatin (D-cell), glucagon (A-cell), and a peptide (PYY) of the pancreatic polypeptides fkdy (F-ceil). Severai studies of the islet organs of teleost fish using immunohistochemical (Langer et al., 1979; Stefan and Falkmer,

1980; Rombout and Taverne-Thiele, 1982; Abad et aL, 1986; Wang et aL, 1986) and immunocytochemical (Camiio et al., 1986; Abad et aL, 1988; Beccaria et al., 1990; Agulleiro et al., 1993) techniques have revealed the existence of four general endoc~ecell types.

The present data support the routine light microscopie, and immunohistochemical study which recognized putative A, B, D and F cek in osteoglossomorphs (Chapter 1).

In earlier reviews it was emphasized that the B-granules of £ish islets are round and do not have the substructure often found in grandes of higher vertebrate B-ce& (Bruin, 1973). This opinion has been connmied through subsequent observations of several teleost species (Brinn,

1975; Klein and Lange, 1977; Lozano and Agulleiro, 1986, Carrillo et al., 1986; Agulleiro et al.,

1 993), although some rod-like crystdoid content was found in a few granules (Klein and Lange,

1977; Carrillo et al., 1986; Agulleiro et al., 1993). In at lest the sea bas (Dicentrarchus labm), the crystaIloid content was immunoreactive with anti-insulin (Agulleiro et ai., 1993). Fibrillar granules were typical of B-cells in Scorpaena scopha (E3oquist and Patent, 1971). In Carassius carassius longsdofi mo st B-cell granules contain the needle-like or bar-shaped crystdoids

(Kobayashi and TakahaShi, 1970). Brinn (1973) concluded that variation in the appearance of B- cell granules in fishes could reflect Merences in the molecular configuration of their insulins.

For instance, differences in zinc content of insulin molecules has long ken suspected to idluence polymorphism of B-ceil granules in hh. Nakamura and Yokote (1971) and Wagner et al. (1 981) stressed the high incorporaticn of this metal in B-celis of a teleost. Baker et al. (1988) 68 reported that the insulin molecule forms a hexamere upon binding the zinc, and this gives the insulin grande core its characteristic hexagonal or round shape, while the zinc-free insulin forms dimers or monomers. There was sida& in shape of the insulin granule cores in three of the osteoglossomorph species studied here (P. buccholzi, N. chitala and G. petersii). Since their granules were hexagonal or round shape, it suggests that their insulins have similar conformation and that they are Zn-binding.However, in 0.bickrhosm the conformation of the innilin molecde may be ciifferent fiom that of the other three species resulting in a greater diversity of shapes in the granule matrices. In addition, it was no ted that the immunoreactivity of B granules to mammalian insulin antisenim in 0. bicirrhosum was weaker than that of the other species.

There nnay be some taxonomie signifïcance to clifferences in insulin granules among osteoglossomorphs, for genus Osteoglossum is clearly distant fkom at least two of the genera studied (Notop~msand Gnathonemus) and its relationship to the other members of suborder

OsteoglossoideÏ (Pantadon) is unclear (Li and Wilson, 1996). Nelson (1994) has Osteoglossum and Pantadon genera in merent families. The Merence in structure of B-granules among the osteoglossomorphs studied may be explained folIowing an analysis of the primary and secondary structures of their insulin molecules.

Several studies identifïed different forms of somatostatin in teleosts (Hobart et al., l98Oa;

Plisetskaya et aL, 1986; Conlon et al., 1987, 1988) including SST-14 1 and II, SST-25 and

SST-28 involving two distinct genes (Conlon et al., 199%; Kittilson et al., 1999). Gene 1 encodes for a preprosomatostatin that will be processed to give SST-28 and /or SST 14 1, whereas gene II encodes for a preprosomatostatin that wiu be processed to give either a 28-residue peptide or a 25-residue peptide including SST- 14 II (~~r',~1~"). Many teleosts (anglerfjsh, coho salmon, raihbow trout, sea bass and tubot) have two distinct D-ceIl types immunoreactive to either SST-14 or SST-25 antisera (McDonaid et al., 1987; Nozaki et al., l988a; Lozano et al. 69 199la; Abad et al., 1992; Garcia-Heniandez and AguIleiro, 1992; Reinecke et al., 1997).

Furthermore, the separate sub-populations of islet cek of these teleosts have different

anatomical relationships with other cell types in the islet, IIILUE1J.y SST-14 containing D2-cells

near central B-cells and SST-25 contaking Dl cells at the periphery (Agulleiro et al., 1993;

Reinecke et al., 1997). In contrast, the more ancient, non-teleost actinopterygian, the gar, showed

only a single type of D-celI (Groff and Youson, 1997, 1998).

Another fonn of sornatostatin, SST-34, was isolated fkom lamprey (Andrews et aL,

1988). The use of antisera to SST-34 in osteoglossomorphs yielded a D-ce11 not

immunoreactive to anti-saIrnon SST-25 or to anti-SST-14. This cell, terrned a DX-ce& also

had granules of different shape than those of the Dl-cell which were immunoreactive with

anti-S ST-25 and anti-SST- 14. This result implies that the SST-34 antisenim has been directed

against a portion of lamprey SST-34 that cannot be recognized by the SST(s) in the Dl-cells, and

that DX-cek have a SST in a diEerent conformation, or perhaps a different SST than that

present in D 1-celis. Thus the osteogIossomorphs share the common feature of other teleosts of

having two types of D-cells in the pancreatic islets. The immunoreactivity to both anti-SST-14

and anti-SST-25 in D 1-cek of the osteoglossomorph idet organ is consistent with that reported

for D 1-cells in the developing and mature islet organ of sea bass (Lozano et al., 199 la; Garcia-

Hernandez and Agulleiro 1992). Moreover, similar CO-existenceof SST- 14 and -25 was reported

in the stomach of Spms auratu (Elbal et al. , 199 1). However, D 1-ce& in osteoglossomorphs

are distributed throughout the &let tissue (chapter 1) as compared with the peripheral distri'bution

in teleosts (e.g. Reinecke et al,, 1997). In contrast, the DX-cells in osteoglossomorphs are spread

throughout the islet, like DZ-cells of other teleosts. The reason for two sub-populations of D-cells

in teleosts is not known, but each type may have a different paracrine action on the other endocrine ceiis (Abad et a1.,1992; Reinecke et al., 1997) and/or there could be independent 70 fiinctions for the different SSTs released fiom ceils of the islet organ of teleosts (Nozaki et al.

1988a; Plisetskaya, 1990% b).

The pancreatic polypeptide-like (PP) peptides pancreatic polypeptide (PP),

peptide YY (PYY) and neuropeptide Y (NPY) have a similar molecular structure and seem to be

derived firom a single ancestral gene (Tatemoto, 1982% b; Tatemoto et al., 1982; Codon, 1995;

Larhmmar, 1996). AU of these 36-amino acid polypeptides show cross-reactivity (Laburthe et

al., 1986; Inui et ai., 1988). Furthermore, in many vertebrate species, including £kh, these peptides are often CO-localizedwith peptides of the glucagon fdy( Lozano et al., 1991b;

Agulleiro et al., 1993). Co-existence between glucagon and neuropeptide Y was also reported in the gar islet organ (Groff and Youson, 1998).

Herrera et al. (199 1) reported tbat during early development of the murine pancreas most glucagon-immunoreacting cek contain PP irnmunoreactivity, whereas in the later stages the fiequency of ody one of the two hormones in distinct celis increases. However, in turbot (Scophïhalmus maximus) both peptides appeared in distinct islet ceils throughout the various developmental stages (Berwert et al., 1995). In osteoglossomorphs two distinct cells immunoreacted with either anti-GLU (A-cell) or anti-PYY ( F-ceii). The size and shape of the cytoplasmic granules in the F-cells of the osteoglossomorphs were similar to those of the

NPY-positive cells in the embryonic pancreas (54 mm stage) of the dogfish (Chiba et al., 1995).

The CO-existenceof peptides of the glucagon and polypeptide fdesin the GEP system of bony fkh and other vertebrates is still a curiosity. Present results show species-specifc dBerences among teleosts that may be related to the existence or absence of an ancestral ceU type, or a molecular-processing mechanism similar to glucagon and pancreatic polypeptides

(Putti et al., 1991).

In summary, the present observations indicate species clifferences among 71 the Osteoglossomorpha in the nature of B-cell grdes, but all species share the feature of two sub-populations of D-ceils similar to more derived teleosts; the Iatter is thus a consistent teleost character. However, distribution and immunoreactivity of the two D-cell types is different between osteoglossomorphs and other teleosts. The number of islet cell types has some phylogenetic relevance among the vertebrates (Falkmer, 1995). In this context, the existence of four cell types (A, B, D, F) in the islet organ of osteoglossomorphs suggests a more derived position for this group over Semionotiformes (e.g., gar) among the Actinopterygii (Groff and

Youson, 1998).

In conclusion, the fïndings of the study are of some taxonomie significance to

O steoglossomorph speciation and to the Osteoglossomorpha among the Teleostei and other

Actinopterygü. Chapter 4

Molecular cloning of preproinsulin cDNAs from several osteoglossomorphs and a cyprinid.

From

Al-Mahrouki, A.A., Irwin, D. M., Graham, L. C., and Youson, J.H. (2001). Molecular cloning of preproinsulin cDNAs from several osteoglossomorphs and a cyprinid. Mol. CelL Endocrino& 174(1-2), 51-58 (with permission of Elsevier Science).

Together with the addition of the cDNAs figures. Introduction

Osteoglossomorpha is an ancient subdivision of Mgteleost fïshes, whose ancestors

originated at the late Jurassic era about 200 million years ago before the separation of the major

continental areas (Li and Wilson, 1996). The wide geographicd disûi'bution of the extant

osteoglossornorpbs make them a potentially interesthg group of fishes to study. Patterns of

endemisrn were reported for the different taxa (Osteoglossidae, in South America; Notopteridae,

in East Asia and Afiica; Pantodontidae and Mormyridae in West Afiica and Hiodontidae, in

North Amerka). This interesthg geographical distribution illustrates the influence of the tectonic

plate movements on the distn'bution of Eesh-water organisms (Li and Wion, 1996).

Furthemore, their relationships to other teleosts are not weil resolved. Nelson (1973), Patterson

and Rosen (1977) and Taverne (1986) indicated that the osteoglossomorphs are the most ancient

LiWig teleosts. This view was challenged by Arratia (1991) who suggested that there is an

elopomorph-osteoglossomorph sister group relationship, supported by a phylogenetic analysis

using 28s rRNA sequences (Van Le et al., 1993). Recently, 07NeiUet al. (1998) examhed the

phylogeny in ancient teleosts using gonadotropin-releasing hormones (GnRH) rnolecular forms

and found a close relationship between the eel (an elopomorph) and the butterfiy fkh (an

osteogIossomorph). Several researchers agree that the Osteoglossomorpha is a monophyletic

group (Li and Wilson, 1996; Zhang, 1998) and there have been several recent analyses on the

inter-relationship among the osteoglossomorphs. Li et al. (1997) reported that the Notopteridae

(e-g. feather fin knife fkh) and Mormyridae (e.g. elephantnose) are most closely related and

together they are a sister group to the Hiodontidae (e.g., goldeye). Sequence analysis of the mitochondrial12S rRNA and 16s rRNA (Alves-Gomes and Hopkins, 1997) indicated that the

Pantodon (e.g. butterfIy fish) departs considerably fiom the Mormyriformes (e-g. elephantnose).

The characterization and comparative analysis of hormonal peptides in more- and 73 less-derived organisms shodd give us an idea of how the peptides might have evolved fiom

the ancestral molecdes. For example, insulin of the -4 the oldest extant vertebrate, was

found to be identical to mammaliaTi insulin in only 20 amino acids out of 53 (Peterson et al.,

1974). However, subsequently Chan et al. (1984) reported highly divergent regions of the guinea pig preproinsulin in cornparison with other mammak especially in the regions encoding the B-

and A-chains. This indicates that insulin structure may not dways follow phylogenetic trends.

Insulin has been isolated and characterized fiom several phylogenetically ancient actjnopterygians (ray-ked fïshes) including the gar, Lepisosteus spathda (Pollock et al., 1987), the bowfïn, Amia calva (Conlon et al., 199l), and the paddefish, PoZyodon spathula

(Nguyen et al., 1994). In addition, insulin cDNAs have been reported fiom carp, Cyprinus carpio

(Hahn et al., 1983).

The above mentioned studies on insulin led to a view that Iow molecular weight polypeptide hormones are not usefid for constructing a phylogenetic tree (Conlon and Hicks,

1990, AgulIeiro, 1995, Dores et al., 1996). Nevertheless, Dores et al. (1996) were able to constnict a usefùl phylogenetic tree using proinsulin sequences based on the variable spacer region of proinsulin (C-peptide), which proved to have informative sites for phylogenetic analysis. Proinsulin sequences have been used to address several phylogenetic questions, such as the evolutionary distance between fiesh water and marine teleosts @ah et al., 1983) or for the confirmation of the hypothesis that lungfish and amphihians share a common ancestor (Codon et al., 1997a). Also, proinsulin sequences dlowed the co~tionof the phylogenetic position of the polypterids as king the most ancient living actinopterygians (Conlon et al., 2998).

In this study, the cDNAs of preproinsulin have ken cloned and sequenced f?om four

Osteoglossomorpha species: Osteoglossum bicirt-hosurn (arawana), Pantudon buchhoki

(buttedy fish), Notopterus chMa (feather fin knife fish), Hiudon alosoides (goldeye), and proinsulin fiom Gnathonemtrs petersii (elephantnose). In addition, 1have cloned and

sequenced a preproinsuh cDNA f?om Catostomus commersoni (white sucker, as a

representative of a more generalized teleost order, Cypriniformes).

The deduced amino acid sequences fiom the cDNAs of the different species are

compared with each other as well as with those eom other vertebrate species, to study the degree

of conservation of the various domainR and to provide a better understanding of the phylogenetic

relationship among the osteoglossomorphs and between this group and other vertebrate species.

MateriaIs and Methods

Anirnals

Addt Usteoglossum bicirrhosurn (arawana), Puntodon buchhola- (butterfly Gh),

Gnarhonemus petersii (elephantnose), and Notopterus chitalu (feather fin knife kh) were

purchased fiom a local commercial aquarium in Toronto. Adult Hiodon alornides (goldeye) were

fished &om the Assinboine River near w'dpeg, MN, and the viscera were obtained and frozen

in liquid nitrogen. Cutosromus commersoni (white sucker) were trap-netted in the Lake Nipissing

watershed (North Bay, ON, Canada). Live specirnens were also obtained kom local fish supplier.

The live specimens were sacficed by an overdose of 0.05% tricaine methanesulfonate (MS-

222) in their holding water. Pancreatic tissues were excised, fiozen in liquid nitrogen and kept at

- 80°C until used.

RNA exiraction,flrsf-*and cDNA synthesis and RACE

RNA was isolated fiom the fiozen pancreatic tissues fiom the six species studied using

TRTzo 1 reagent (Life Technologies, Grand Ishd, NY,USA) according to the x~nufacturer~s directions. The samples were homogenized in TRIzol that contain phenol and guanidine isothiocyanate, and the RNA was then recovered by precipitation with isopropyl aicohol. The

RNA was reverse transcrïi'bed into cDNA using the Omnkcript Reverse Transcriptase kit 75 (QIAGEN, Mississauga, ONyCanada) according to the maaufàcturer's protocol, using the reverse transcriptase enzyrne, deoxynucleotide (dNïP) mix, and deoxythymidine [oligo(dt)] primer. The resulting first-strand cDNA was then used as a template for subsequent polymerase chain reactions (PCR). The partial insulin cDNAs were obtained using the following oligonucleotide primers: 11, 5' GYNSCNCAGCACCTGTGCGGATCCCA 3' corresponding to the N- terminal of the B-chain and I2,5' R?TRCARTA~~NRTYCAGGTCGAAGATGCA3' corresponding to the C-terminal of the A-chah (R=A+G, S=W,Y=C+T, N=A+CtG+T) (Fig. 26).

Another set of primers was designed following the alignment of the initial partial insulin cDNA hgments Grom each different species under study;

IR5,5' CGAAC;ATGTTGCAGGGNCGGTGGCARCA~~GYTC3' was used to obtain the 5' RACE

(rapid ampiification of cDNA ends) product and corresponded to the N-terminal of the

A-chah The other primer, which is used to obtain the 3' RACE product, was

IF3,S GCCCTGTACCTGGTGTGCGGCGAGAGRGGNTTYTTY3' and it corresponded to the

C-terminal of the B-cbain (Fig. 26). The primers were designed using the CODEHOP program

(Rose et al., 1998, www.blocks.fhcrc.org/blocks/he1p/CODEHOP)anci were synthesized by

Sigma Genosys (Oaksdle, ON, Canada). The UCEreaction was perfonned using the SMART

RACE cDNA amplification kit (CLONTECH, Palo Alto,CA, USA) according to the manufacturer's instructions, using the universal primers provided by the kit and the previously mentioned primets in the RACE PCR reactions.

IsoIafion of fhe preproinsufin cDNAs

The 1" set of oligonucleotide primers were used in PCR reactions to obtain partial insulin cDNAs. The PCR product was cloned using pCR 2.1-TOPO vector (Invitrogen, Carlsbad, CA,

USA) according to the ~aflufhcturer's guidelines. The plasmid vector (pCR 2.1 -TOPO) has a 3 ' thymidine (T) overhang for TA cloning and topoisornerase I that ligates a PCR product 76 containing 3' A-overhangs very efficiently. The cloned reaction was then transformed into

competent cek. PIasrnid DNA fiom positive clones was then isolated using a QIAprep Miniprep

kit (QMGEN?Mississauga, ON, Canada) according to the manufacturer's guidelines. The

plasmid preparatims were treated with aikaline Iysis foliowed by adsorption of DNA onto silica-

gel membranes, which was then washed and eIuted fiom the membrane. The isolated pIasmid

DNA was Meranalyzed using a restriction digest with EcoRI. Only the plasmids containing

the insert were sequenced. The PCR RACE products were cloned and isolated using the same approach.

Sequencing and identiïcafiun of the cDNAs

Two sequencing methods were used; automated and manual sequencing. For the

automated sequencing, samples were sequenced either at the Core Molecular Bio logy Facility,

York University (Toronto, ON, Canada) or at Cortec DNA Service Laboratories, Queen's

University (Kingston, ON, Canada). The manual sequencing was performed using a T7 sequencing kit (Pharmach Biotech, Baie d'Urfe, QC, Canada), and the radioisotope used in sequencing ( 35~)was purchased fkom Amersham (Oakviue, ON, Canada). The sequenced sampies were nui on 6% polyacrylamide gels using a Mode1 S2 Sequencing Gel Electrophoresis

Apparatus (Canadian LXe Technologies, GTBCO BRL, Burlington , ON, Canada). Manual sequencing was pe&ormed in two directions using an Ml 3-reverse primer purchased &om

Pharmacia Bio tech (Baie d9Urfe,Quebec), and the Ml 3-fornard primer (provided by the kit).

The nucleotide sequences were independently aligned fiom several partial clones as well as fiom the RACE clones fkom each of the species studied, then were translateci to the correspondhg amino acid sequences, which were used in the phylogenetic analysis applying the PAUP 4.0b2

@hylogenetic analysis using par simony; Swo fford, 1999) program. Totd pancreatic RNA was reverse trmcriid to synthesize the cDNA, which was used as a template for the PCR reactions. The first set of primers (Tl and 12) were designed towards a conserved regions of the B- and A-chains (Fig. 26); these gave a PCR product of - 264 base pairs encoding the B-chain, C-peptide and A-chah regioa The sequence was confirmed by sequencing at least three clomes fkom each species, and the clones containing the PCR product were identified using a restriction digest of the plasmid DNA with EcoRI. The partial cDNAs were then aligned and a set of gene specific primers were designed (IR5 and IF3). The IR5 primer was used for 5' RACE reaction and produced a PCR product of - 435 base pairs, which included the 5' untranslated region, and the coding region for the signal peptide, the B-chain, the

C-peptide and the N-tenninm of the A-ch& The IF3 primer was used for 3'RACE and a PCR fiagment of 369 base pairs was produced £tom the buttedy fkh cDNA which included the coding region for the C-ternainus of the B-chah, the C- peptide, the A-chah and the

3' untranslated region ending with a poly A-tail. To confirm the sequences, five clones were sequenced for each PCR fiagment fiom each species. The overlapping S'and 3' RACE products yielded a sequence of 458 base pairs fiom the buttedy fish, as an example for the osteoglossomorphs (Fig. 27)- As for the other species (excludhg elephantnose ), the overlapping region of the partial cDNAs sequences and the S'products sequences codhmed the identity of the sequence (Figs. 28A-D; GenBank # AF199585-AF199589 and AF 282408). The open reading fiame consisted of 110 amino acids in both the butterfiy fkh and the goldeye, 111 amino acids in both the arawana and the feather fin, and 108 amino acids in the white sucker. The partial sequence obtained fiom the elephantnose consisted of 87 amino acids (Fig. 28E). The cornparison of the sequences with those fkom other species was used to ident~the possible processing sites of the deduced prohormonal sequence. This analyses revealed that the 78 preproùisulin of the osteog1ossomorphs consists of a signal peptide located in the 5' region of

the molecule with a leagth of either 23 amino acids in butterfly &h, feather fin, and goldeye or

24 amino acids in arawana, a B-chah with a length of 30 amino acids in three species and 3 1

amino acids in feather fin, and foilowed by the 32 amho acids C-peptide, then by the 2 1 amino

acids A-chah in ail the species under study. However, the preproinsulin fiom white sucker

consisted of a signal peptide with a length of 23 amino acids, a B-chah of 28 amino acids,

fo ilo wed by a C-peptide with 32 amino acids, and ending with A-chah a 2 1 amino acids long.

Cornparison of the aligned sequences among the osteoglossomorphs indicated that the

A-chah was the most conserved followed by the B-chab and then the C-peptide region

(Fig. 29). The sequence cornparison of the osteoglossomorphs and other vertebrates including

fishes, indicated the closer similarity to fishes in general, specfically to the anglerfkh, Nile

tilapia and chum salmon, whereas the white sucker sequence was similar to that of the common

carp and the zebrafish (Figs. 30 and 3 1).

Phylogenetic analysis of the preproinsulin sequence using parsimony resulted in two

equal parsimony trees (Fig. 3 1). The two trees differed only in the interrelationship of zebmfish,

white sucker and common carp. In both trees the Osteoglossomorpha was monophyletic, in

which the elephantnose and the feather fin were more related to each other than to the other

species, and the butterfly fis4 goldeye, and arawana were more related (Fig. 3 1). Based on these types of analysis, the Osteoglossomorpha was not the most basal group among the living teleosts

rather, it is more generalized within the teleosts (Hg. 3 1). Interestingly, the white sucker was

shown to be closely related to the common carp, which is also a cyprinid (Fig. 3 1). Fig. 26 Structure of the preproinsulin cDNA. The open box indicates the predicted open

reading fiame, which includes the codmg region for the signal peptide (S), and the

insuiin B-chain (B), C-peptide (C) and A-chah (A). Thick iines indicate the 5' and the

3' untrmlated regions. The thick bars with arrow indicate the position and the orientation

of the oligonucleotide primers, whereas thin lines indicate the cDNA clones used for

sequencing.

Fig. 27 Preproinsuh cDNA sequences fiorn the butterfly fish. The nucleotide sequence is

presented with the deduced amino acid sequence indicated by single letters- Potential

processing sites are shown in bold. The amino acid sequence of the predicted B- and

A- chsinn is underhed. The cDNA sequence is numbered Çom the 5' end, while the

amino acid sequence is numbered positively fiom the predicted N-terminus of

proinsulin, The signal peptide is indicated by negative numbers. The polyadenylation

signal is double underlined. CCAATAACCCCTCCATTCTCACCTCCACCATCTCCACAGCTCCCCCTCACC

-23 Signal peptide MALWLQAFTLLVLLV ATG GCC CTC TGG CTG CAG GCG TTT ACC TTG CTG GTG CTG CTG GTA

LSSPGAQSASSQHLC TTG TCC TCA CCA GGC GCT CAG TCC GCC TCC AGC CAA CAC CTG TGT

GSHLVDALYMVCGEK GGC TCC CAT CTG GTG GAT GCC CTC TAT ATG GTG TGC GGC GAG AAG

30 33 GFFYQPKTKRDVDPL GGC TTC TTC TAC CAA CCC AAAACC AAAAGA GAC GTG GAT CCC CTG

C peptide L G FLSPKSAQENEAD CTA GGT TTC CTC TCT CCA AAA TCA GCA CAG GAG AAC GAA GCG GAC

64 E Y PYKDQGDLKVKRG GAA TAT CCC TAC AAA GAC CAG GGT GAT CTG AAA GTG AAG AGG &

EQCCHHPCNIFDL ATC GTG GAG CAG TGC TGT CAT CAC CCC TGC AAC ATC TTC GAC CTG

Q N Y C N STOP CAG -4AC TAC TGT AAC TGA GATGTGCGCCTCCCTGCCTTCTTCCTTTTAAGTGC ' Fig. 28 Prepromsulin cDNA sequences. The nucleotide sequence is presented

with the deduced amino acid sequence indicated by single letters. Potential processing

sites are shown in bold. The amui0 acid sequence of the predicted B- and A-chahs is

underhed. The cDNA sequence is numbered fiom the 5' end, while the amino acid

sequence is numbered positively fiom the predicted N-terminus of proinsulin. The signal

peptide is numbered negatively.

A) fiom the arawana, B) fkom the goldeye, C) fiom the feather fin knife fïsh,

D) fiom the sucker and E) fiom the elephantnose. GCCCAGTCGCATCGTCCTACGCTATTTCTCTGTTACCTCTCTGCCAGCAGCACCAACATC

-24 Signal peptide MAIWLQAFPLLV TCCACAGCACC ATG GCG ATT TGG CTC CAG GCG TTC CCT CTG CTG GTC

-1 +l LLVLSSSPGAESSSS CTG CTG GTG CTT TCC TCC TCC CCT GGG GCA GAA TCC AGC TCA AGC

B-chah QRLCGSHLVDALYMV CAG CGC TTG TGT GGC TCC CAT CTG GTG GAC GCG CTC TAT ATG GTT

CGDRGFFYSPKSRRE TGC GGG GAC CGC GGC TTC TTC TAT AGC CCC AAA AGT AGG AGG GAG

C-peptide AEPLLGFLSPKSGQE GCA GAG CCG CTG CTG GGC TTC TTG TCT CCA AAA TCA GGC CAA GAG

NEVDEYPYKEQGELK AAC GAA GTG GAC GAA TAC CCT TAC AAG GAG CAG GGC GAG CTG AAA

VKRGIVEQCCHRPCN GTT AAG AGA GGC ATT GTG GAG CAG TGC TGT CAT CGC CCC TGC AAC

IFDLQNYCN ATC TTC GAC CTG CAG AAC TAT TGT AAC B) GCATCCTTTGGCACTAT

TCAACTCCTCCATAACCCCTCCATTCTCACCTCCACCIITCTCCACIIGCTCCCCCTCACC

-23 Signal peptide M ALWLQAFTLLVLLV ATG GCC CTC TGG CTG CAG GCG TTT ACC TTG CTG GTG CTG CTG GTA

-1 +I L SSPGAQSASSQHLC TTG TCC TCA CCA GGC GCT CAG TCC GCC TCC AGC CAA CAC CTG TGT

GSHLVDALYMVCGEK GGC TCC CAT CTG GTG GAT GCC CTC TAT ATG GTG TGC GGC GAG AAG

30 33 GFFYQPKTKRDVDPL GGC TTC TTC TAC CAA CCC AAA ACC AAA AGA GAC GTG GAT CCC CTG

C peptide LGFLSPKSAQENEAD CTA GGT TTC CTC TCT CCA AAA TCA GCA CAG GAG AAC GAA GCG GAC

64 EYPYKDQGDLKVKRG GAA TAT CCC TAC AAA GAC CAG GGT GAT CTG AAA GTG AAG AGG &

IVEQCCHRPCNIFDL ATT GTG GAA CAG TGT TGC CAC CGC CCC TGC AAC ATC TTC GAC CTG

NQYCN AAT CAA TAC TGT AAT -23 Signal peptide M A v W L RAFSLLVLLV ATG GCA GTT TGG CTC CGG GCT TTC TCT TTG TTG GTG CTG CTG GTG v S S P G ANAASNQHLC GTG TCC TCC CCT GGG GCAAAC GCC GCC TCC AAC CAG CAC CTG TGC

GSHL'+- v EALYLVCGER GGC TCC CAC CTG GTA GAG GCA CTC TAC CTG GTG TGT GGA W CGA

31 34 GFFYNPKMDKRDAEP GGC TTC TTC TAC CCC -4AA ATG GAT Am AGA GAT GCA GAG CCC

C peptide L L G F L S PKSGLENEV 52 TTA CTA GGC TTC CTA TCG CCAAAG TCA GGC CTG GAG AAT GAG GTT 268

65 D E Y P FK DQGDVKMKR 67 GAT GAA TAT CCC TTC AAG GAC CAG GGC GAC GTG AAG ATG AAG CGG 313

- GIEQ v C CHRPCNIFD 82 GGA ATA GTG GAA CAG TGT TGT CAC CGC CCC TGC AAC ATC TTC GAC 358

QNQYCN CAG AAC CAA TAC TGC AAC GATACATCTACATTTTCT 18

CAGCTCCACTACCATCTACCTCTGTATT'IETCTCGTCCTCTGCATTAAGAACAGTGTGACT 77

-23 Signal peptide MAVWLQAGALLFLLA -9 ATG GCA GTG TGG CTC CAG GCT GGT GCT TTG CTG TTC CTG TTG GCT 122

VSGVNANVAPQHLC 7 GTC TCT GGT GTC AAT GCA AAT GGA GTG GCC CCG CAG CAT CTG TGT 167

GSHLVDACYLVCGPT 22 GGA TCT CAT CTG GTT GAT GCA CTC TAC CTG GTC TGT GGT CC. ACA 212

28 GFFYNPKRDVDPLIG 37 GGC TTC TTT TAT AAC CCC AAG WGA GAT GTT GAC CCC CTC ATC GGT 257

C peptide FLPPKSGENEVADF 52 TTC CTT CCT CCT AAA TCT GGC CCG GAA AAT GAG GTG GCT GAC TTC 302

61 AFKDHAELIRKRGIV 67 GCA TTT AAA GAT CAT GCT GAG CTG ATA AGG AAG AGA GGC ATT GTA 347

EQCCHRPCNIFDLEK 82 GAG CAA TGC TGC CAC CGC CCC TSGC AAC ATC TTC GAC CTG GAG AAA 392

YCN TAT TGC AAC B-chain APAQHLCGSHLVEAL GCT CCG GCA CAG CAC CTG TGC GGA TCC CAC CTG GTA GAG GCG CTC

FLVCGERGFFFNPDT TTC CTG GTG TGT GGG CGC GGC TTC TTC TTC AAC CCC GAC ACT

33 c peptide K R D V D s LLG F L s e K AAA AGA GAT GTG GAC TCC CTA CTT GGC TTC CTA TCT CCA AAA

G PENE A DEY R Y K E Q A GGC CCA GAG AAT GAA GCT GAT GAG TAC CGC TAC AAA GAA CAG GCA

64 A-chain E VKVK R GIVE - - H GAA GTG AAG GTG AAG AGG GGG ATC GTG GAG CAG TGC TGT CAC CAC

PCNIFDLNQYCN CCC TGC AAC ATC TTC GAC CTG AAC CAG TAC TGT AAC 82

Fig. 29 Alignment of deduced preproinsulin amho acid sequences fiom osteoglossomorph

species. Potential processing sites (paired basic) are bold and in brackets at the end of the

B-chah and the C-peptide. The signal peptide sequence was not determined from the

elephantnose and was denoted by space. Dashes (-) indicate amino acid gaps in the

different species. Signal peptide

Butterfly f ish Goldeye Arawana Feather fin Elephantnose

1 5 10 15 20 25 30 Butterfly fish AS SQHLCGSHLVDALYMVCrJEKGFFYQPKT- (KR) Goldeye ASSQHLCGSHLVDALI'MVCGEKGFFYQPKT-(KR) Arawana SSSQRLCGSKLVDALYMVCGDRGFEYSPKS-(RR) Feather fin ASNQHLCGSHLVEALYLVCGERGFFYNPKMD (KR) Elephantnose APAQHLCGSHLVEALFLVCGERGFFFNPDT-(KR)

C peptide

1 5 10 15 20 25 30 Butterfly fish DVDPLLGFLSPKSAQENEADEYPYKDQGDLKV (KR) Goldeye DVDPLLGFLSPKSAQENEADEYPYKDQGDLKV(KR) Arawana EAEPLLGFLSPKSGQENEVDEYPYKEQGELKV(KR) Feather fin DAEPLLGFLSPKSGLENEVDEYPFKDQGDVKM (KR) Elephantnose DVDSLLGFLSPKSGPENEADEYRYKEQAEVKV (KR)

1 5 10 15 20 Butterf ly fish GIVEQCCHHPCNIFDLQNYCN Goldeye GIVEQCCHRPCNIFDLNQYCN Arawana GIVEQCCHRPCNIFDLQNYCN Feather fin GIVEQCCHRPCNIFDQNQYCN Elephantnose GIVEQCCHHPCNIFDLNQYCN Fig. 30 ALignment of prepromsulin sequences fiom diverse vertebrate species. The amino acid

sequence is numbered starting at the begiiining of the signal peptide. Dashes (-) indicate

amino acid gaps in the different species, except in the elephantnose where the sequence

encoding for the signai peptide was not detemined- References of the sequences fiom

diverse vertebrate species: common carp, Hahn et aL, 1983; zebrdïsh, Milewski et al.,

1998; anglerfïsh, Hobart et al., 198Ob; tilapia, Mansour et al., 1998; salmon, Kavsan et

al., 1983; monkey, Wetekam et aL, 1982; rabbit, Devaskar et al., 1994; sheep, Ohlsen et

al., 1994; chicken, Perler et al., 1980; Xenopus, Shuldiner et al., 1989). Signal Peptide B chah b

Common Carp Zebrafish AnglerEish Nile tilapia

Butte- fish Goldeve Arawana Feather fin Elephantnose

Chum salmon Monkey Rabbit Sheep Chicken xenopus

C Peptide 4

White sucker

Conmon carp Zebrafish Anqlerfish Nile tilapia

Butterflv fish Goldeve Arawana Feather fin Elephantnose Fig. 31 Phylogeny of proinsulin sequences. Sequences fiom osteogIossomorphs and

diverse vertebrate species (see Fig. 30) were analyzed by parsimony. Two similnr trees

were obtained, the ody Merence was in the reiation of the white sucker to the (common

carp/zebrafkh). Branch length is proportional to the number of nucleotide replacement

(i. e., resulting amino acid substitution) inferred by parsimony (S wofford, 1999). I

Sheep-

Rabbit

Chicken

I Xenopus

-Elephant fish Lr-Chum salmon uAnglerf ish r Common carp fl-, White sucker

- 5 amino acid changes Discussion

Structure anaiysis

Crystal structure determination of insulinn as well as a1ani.e-scanning mutagenesis has

identifïed a number of amino acid positions which are essential for its three-diniensional

structure and its effective binding to the insulin receptor (Baker et al., 1988, Kristensen et ai.,

1997). Cornparison of the deduced insulin amino acid sequences fiom the osteoglossomorphs

and the white sucker revealed conservation of the residues involved in the cysteine bridge

formation (A7-B7, A6-AI 1, A20-B 19). Most residues known to be involved in the hexamer

formation were conserved (B6, B 10, B 14, B 18), with exception of B 17, A13 and A14. At these three positions conserved substitutions were seen, with leucine at B 17 replaced by methionine in the butterfly fis4 the arawana, and the goldeye; leucine at A13 replaced by isoleucine; and tyrosine at A14 replaced by phenylalanine in all of the species. The residues involved in the hexamer formation are believed to interact with the 2n2+to form stable hexamers (Baker et al.,

198 8), and a substitution of one of these residues can affect the hexamer formation. In the case of the hagfïsh, a histidine was replaced by an aspartate at B 10, which prevented the formation of hexamers. The stabilization of the hexamer forms in the presence of a hi& concentration of zn2+ was verified in severai teleost insulins (Cutfield et al., 1986). This type of substitution was not observed in any of the osteoglossomorphs including the arawana, in which secretory granules of

B-cells have a polymorphic crystalline core and are different fiom those observed in the other osteoglossomorphs (Chapter 3)- The N-terminal of the B-chah may also play a role in the hexamer formation (Codon et al., 1991) and this domain is weakly conserved in osteoglossomorphs. It is noteworthy that in arawana the serine replaces aIanine at B 1 (Appendix

2); this may explain the structural diversity of the secretory granules of the B-ceils (Chapter 3).

The known receptor-binding region at residues B 12, B 16, B23-26, Al -AS, A1 9, 86 A21 were conserved in all the sequences, except for positions B16 and B26 in elephantnose-

At these two positions are conservative substitutions: phenylalanine for tyrosine, but these

substitutions might lead to reduced bhding dE&y (Codon et al., 199 1). Several sites are

involved in rnaintaining the receptor-binding conformation: glycine at B23, phenyIalanine at

B24, isoleuciue at A.2, valine at A3, and tyrosine at A19. These residues interact with the receptor and with other residues as well, for instance, leucine at B6, glycine at B8, leucine at

B 11, glutamate at BI 3 and phenylalanine at B25 (Codon, 2000). These residues are conserved in the present sequences except for B 1 3, where a conservative substitution (aspartate for glutamate) was observed in the arawana, goldeye, butterfly fkh and white sucker. Generally, carbohydrates are low in teleost diets, and it was reported that teleosts do not use glucose efficiently as a general source of energy (WiIson, 1994). Pankas et al. (1994) reported the presence of fùnctional insulin receptors in tilapia skeletal muscle, which supports the earlier reports (King and Kahn, 1981; Mommsen and Plisetskaya, 1991) of the important role of insulin in promoting somatic growth in teteosts. In general, however, the effects of insulin on metabolkm and growth in fish are less differentiated than they are in rnammals, perhaps due to the overlap in the function of insulin and insului-like growth factor (Planas et al., 2000).

Therefore, despite the conserved substitution of residues in the receptor binding region of osteoglossomorph insulins, one must be cautious in extrapolating these data in a fünctional connotation. On the other hand, Wright et al. (2000) have recently emphasized that insulin plays an important role in glucose homeostasis in tilapia, at least at the level of the islet tissue. Since species variability in glucose met abolism may exkt among the vat numbers of teleosts, bioassay of osteoglossomorph insulins are an essential fbture project.

An extension to the N-terminus of the B-chah was reported for insulins fiom some &h 87 Species mcludmg lungfish (Codon et aL, 1997a), tuna, carp, anglerfïsh and cod (Shulinder et

al., l989), and lamprey (Plisetskaya et aL, 1988). Amino acid extensions do not seem to affect

the biological activities (Conlon et al., 1997a), but their presence may be used to confïrm

monophyletic origins as in the case of holarctic and southem hemisphere lampreys (Conlon,

2000; Youson, 2000). In this study no extensions to the N-terminus of the B-chah was observed

in any of the species under study; in the case of the osteoglossomorphs, it suppoas theù

monophy1etic origins.

Pliylogenetic anaiysis

Hormonal peptides have been used by many to address phylogenetic questions in teleosts

and in other vertebrate species. As explained in the introduction to the present investigation,

insulin is one of these peptides (Conlon, 2000). Other GEP peptides, such as the sequences of

glucagon-like peptide, have ken used to chifying the phylogenetic relationship of arnphibians

(Irwin and Sivarajah, 2000) and other vertebrates (Irwin et al., 1999). Also the preprosomatostatin sequence was used to confirm the close relationship of the lungfish and tetrapod (Trabucchi et ai., 1999). Several studies have reported the usefulness of some polypeptide hormone precursors (Grober et al., 1995; Rubin and Dores; L994, 1995) in phylogenetic amlysis, where the spacer region was O ften the informative character. Cornparison of the deduced amino acid sequences of the preproinsulùis of the present study to those ftom other fish and vertebrate species using maximum parsimony, revealed a grouping of the

O steoglossomorphs, suggesting a monophyletic ongin (same clade). Several reports, using the hard and soft anatomy support this conclusion (Li and Wilson, 1996 and Zhang, 1998). Our comparative sequence analysis grouped the Notopteridae (e.g., feather fin Mefish) with the

Mormyroidae (e.g., elephantnose), which supports the view of Li et al. (1997). The resulting 88 relationship between the Pantodontidae (e.g., buttedy f&h)and the OsteogIossidae (e-g.,

arawana) is supported by several reports (Lauder and Liem, 1983; Van Le et al., 1993).

Greenwood (1973) grouped the Hiodontidae with the Notopteridae but the recent reassessment of the Hïodontidae concluded that it is a sister-group to all the osteoglossomorphs (Van Le et al.,

1993; Li et al., 1997; Zhang, 1998). In our anaiysis the Hiodontidae (e.g-, goIdeye) was not grouped with the Notopteridae, but with the Pantodontidae and the OsteogIossidae, which in part agrees with other recent assessments. It has been suggested that the osteoglossomorphs are the most ancient living teleost fish (Nelson, 1973, Patterson and Rosen, 1977, Taverne, 1986), however, our phylogenetic analysis using preproinsulin showed that osteoglossomorphs are more generalized teleosts, which support the views of Arratia (1 99 1) and O 'Neill et al. (1 998). As has kenpointed out, the use of hormone-coding genes in phylogenetic analysis probably requires the sequence anaiysis of severai polypeptides (Dores et al., 1996).

This study has made two primary contriiutions. The results provide some assistance in resolving the debate over the position of the osteoglossomorphs within the teleosts, confïrming their monophyletic origin, and providing a usefûl assessrnent of intragroup relationships. The second contribution is to comparative endocrinology dysis, where a description of the deduced peptide sequences attempts to idenMy the conservation of putative functional sites of insulin molecule in a key group of teleost with ancient lineage. Chapter 5

Characterization of somatostatin cDNAs molecular identification and comparative analysis. 89 Introduction

Somatostatin (SST) is a multi-fûnctional molecule which effects the secretion of growth hormone in the pituitary and of insulin and glucagon fiom the endocrine pancreas (Reichlin,

1983; Gerich, 1983; Eilertson and Sheridan, 1993), and of progesterone in the ovary (Holst et al.,

1995). SST is also associated wah the coordination of metabolism, growth and development

(Gerich, 1983; PateI, 1992), as a neurotransmitter/ neuromodulator in the central nervous system, CNS (Reisine and Bell, 1995), and in the control of rend function (Roca et aL, 1986).

In mammals, preprosomatostatin (PPSS) encodes one precursor, which can be processed either at a monobasic site or at a paired basic site to yield two different bioactive forms, SST-28 or invariant SST-14, respectively. However, in other vertebrates, including teleost fkh and fkogs, two genes have been found (Sheridan et al., 2000). Preprosomatostatin 1 (PPSS 1) encodes invariant SST- 14 and is similar to the mammalinn preprosomatostatin (PPS S).

Preprosomatostatin II (PPSS II) eencodes for a variant SST [~yr', G~~'~]-SST-14 at the C- terminus and for an extended amino acid peptide at the N-terminus with varying length in the different species (Le., SST-25 to -37). Generally, PPSS 1is more conserved through evolution tban PPSS II (Codon et al., 199%).

SST active forms, are encoded by PPSS 1and produced by both the CNS and the peripheral organs as weU, while the variant isoforms encoded by PPSS II are mainly produced by the gastro-intestinal cells and/or cek of pancreatic islets (Moore et al., 1995). Recently, a PPSS that codes for a variant [Ro2] SST-14 was identifed fiom the brain of the Anican Iungfish,

Protopterus annectens (Trabucchi et al., 1999) and the goldnsh (Lin et al., 1999), and a similar form was descriid fbm a pituitary extract of sturgeon (Nishi et al., 1995). An additional novel variant @?ro2, et 13] SST- 14 was characterized only fiorn the fiog brain but not the pancreatic islets (Vaudry et al., 1992; Tostivint et aL, 1996). 90 In teleost fkh, both PPSS 1 and PPSS II yield somatostatins that are important

modulators of lipid and carbohydrate metabo lism (Sheridan, 1994). DiEerent SST iso fonns

are derived fiom PPSS II and are msinly produced in the pancreatic islets (Moore et al., 1995) or

the intestine (Uesaka et al., 1995) of teleost fish. These isoforms include [~yr~,~l~"] SST-14 or

SST-25 in salmonids (Plisetskaya et al., 1986; Moore et al., 1999), a unique variant SST-22 in

catfkh (Magazin et al., 1982), and SST-28 in both anglerflsh (Goodman et al., 1980; Hobart et

al., l98Oa) and in sculpin (Codon et aL, 1987). Although a single PPSS may encode for the large

SSTs, 33 to 37, among lamprey species (Andrews et al., 1988; Codon et al., 19953, b), three

PPSSs were identifïed in tissues of &th bout (Moore et al., 1995; Sheridan et al., 1997; Kittilson

et al., 1999) and goldbh (Lin et al., 1999). In trout, a PPSS II' encodes [Tyr7, SST- 14 or

SST-28, a PPSS II" encodes [~~r~,~l~"]SST- 14 or SST-25, and a PPSS 1encodes SST-14 or

SST-26. In goldfish brai. a PPSS 1 encodes SST-14 or SST-26, a PPSS II encodes [~lu',Tyr7,

~l~'']SST- 14 or SST-28, and a PPSS III encodes pro2] SST-14 were reported. The functional simiificance of the dserences in these posttranslation modification is not known.

The use of different SST antrhdies in several immunohistochemical and immunocytochemical studies suggested the presence of difEerent forrns of SST in cells of the gastro-entero-pancreatic system of teleosts (McDonald et al., 1987; Lozano et al., 1991a; Abad et al., 1992; Reinecke et aL, 1997; Chapters 1,2,3). To date, characterization of PPSS cDNAs in teleosts is limited to a few of the known 23,000 species as descriid above. The Limited information demonstrates the need for more research on PPSSs in teleosts to reveal the extent of the variation of these precursors. These types of data may in turn help in understanding the evolution of SSTs and what implications molecular heterogeneity has on the hction of this peptide (Sheridan et al., 2000). 91 Usmg molecular clonhg techniques, we have characterized several somatostatins fiom

four osteoglossomorphs, an ancient teleost group and f?om white sucker, a generalized teleost.

The objectives of this study are to extend our knowledge of the diversity of SSTs arnong teleosts,

and to investigate whether any variations in Ïsoforms of somatostatin in osteoglossomorphs can

be used for interspecïfïc phylogenetic analysis between members of this group and between

general and more derived teleosîs-

Materials and Methods

Adult specimens were purchased fkom a local commercial aquarium in Toronto, ON,

Canada, including Osteoglossum bicirrhosum (arawana), Pantodon buchholn (buttedy fish),

Gnathonernuspetersii (elepbtnose), and Notopterus chitala (feather fin knife fïsh). Catostomus cornmersoni (white sucker) were trap-netted in the Lake Nipissing watershed (North Bay, ON,

Canada) and live specimens were obtained fiom a local fïsh supplier. The live fkhes were sacrificed in their holding water using an overdose of 0.05% (W/V) tricaine methanesulfonate

(MS-222). The pancreatic tissues were quickly excised, immediately fiozen in liquid nitrogen, and then were stored at - 80°C.

Extraction and quantification of total RNA

The fiozen tissues were weighed and the appropriate amount of TRIZOL reagent (Life

Technologies, Grand Island, NY, USA) was added (lmV50-100 mg of tissues). The isolation procedures were according to the manufacture's directions. The samples were homogenized in

TRIzol that contain phenol and guanidine isothiocyanate, and the RNA was then recovered by precipitation with isopropyl alcohol.

Isolation of preprosomatostatin cDNAs and sequence anaiysis Using the Odcript Reverse Transcriptase kit (QIAGEN? Mississauga, ON, Canada), 92 RNA samples were reversed transcnid to fïrst-strand cDNA which was used as a ternplate in the polymerase chab reaction (PCR). The preprosornatostatin identification strategy included two stages. In stage one, the fo llowing set of oligonucleotide primers were used:

S, 5' GCCGGACTGCTGWSNCARGARTGGWSNAA 3', corresponding to the N-terminus of a non- coding region of the preprosomatostatin and

Sc,5' GCAGGAGGTGAAKSYCTTCCARWARWARTT3 ', corresponding to the C-terminus of the region coding for SST-14 (R=A+G, S=CcC,Y=C+T, K=GtT, W=A+T, N=A+C+G+T)@g. 32). This stage resulted in the amplincation of partial cDNAs. In stage two, another set of prhers was used; SR, 5' GCAGGAGGTGMGGTW?TCCAGWAGAA 3'corresponding to the SST- 14 C-terminus and was used in the 5' RACE reactions (rapid amplification of cDNA ends);

S3', 5' CGGGAGCGGAAGGCCGGCTGYAARAA3' corresponding to the middle of SST-28 and the end of the N-terminus of SST-14 and was used in the 3' RACE reactions. The primers were designed using the CODEHOP program (Rose et al., 1998; www--.blocks.£hcrc.ore/blocksl hel~lC0DEH0P)and were synthesized by Sigma Genosys (Oakville, ON, Canada). The RACE reaction was performed using the SMART RACE cDNA arnpiilïcation kit (CLONTECH, Pa10

AIto, CA, USA) according to the mandacturer 's instructions, using the universal primers provided by the kit and the previously mentioned primers in the RACE PCR reactions. The resdting PCR fi-agments were cloned using the TOPO TA cloning kit (Invitrogen, Carlsbad, CA,

USA). The PCR fiapents were purified and cloned; five clones for each fragment were isolated and amplined in plasmids. Sequencing of the plasrnid DNA was carried out by manual sequencing using a T7 sequencing kit (Pharxnacia Biotech, Baie d7Urfe,QC, Canada), and "S purchased ftom Amersharn (Oakville, ON, Canada). Sequences were cobedby automated sequencing at Cortec DNA Service Laboratones Queen's University Wgston,ONy Canada). Northern Analysis

The isolated total RNA (10 - 15 pg) from the pancreatic tissues of the four

osteogIossomorphs were separated in a 15%formaldehyde-agarose gel, and transferred onto a

positively charged Nylon Membrane (Boehnnger Mannheim, Lavai, QC, Canada).

Preprosomatostatin cDNA fiom arawana was Iabeled with [a32~]d~~~using the Prime4

Random Primer labehg kit, according to the manufacturer's instructions (Stratagene, La lolla,

CA, USA). The Prime-It labeling kit uses the Klenow fiagment of DNA polymerase 1that

synthesizes new DNA by incorporating the labeled nucleotide. Hybridization and washes were

carried out under low stringency conditions @vin and Wong, 1995). The size of each of the

redting bands was determined usuig a 0-24-9.5 Kb RNA ladder (Life Technologies, Grand

Island, NY,USA)

Resuits

The reverse traflscriid est-strand cDNA was used as a template for the PCR reactions, together with primers (S and Sc) and produced a fiagrnent of a 222 base pairs (bp) fiom the

arawana and fiom the butterfly fish (Fig. 32). The sequence fiom several clones of the PCR product encoded the hormonal region of somatostatin and part of the non-hormonal region excluding the signal peptide. These sequences were then used to design primers for RACE.

The SR primer was used in the 5' RACE reaction (Fig. 32) resulting in a 38 1 bp fiagrnent for arawana, 450 bp fiagment for butterfly fish, 470 bp kagrnent for elephantnose, 471 bp fiagment for feather fin Mefkh, and 487 bp fiagment for white sucker. The S3' primer was used in the 3' RACE reaction (Fig. 32) resulting in 230 bp for arawana, 252 bp for butterfiy fish,

270 bp for elephantnose, 276 bp for feather fin knife fish, and 414 bp for white sucker. The PCR fragments were pufied and cloned; three clones for each fiagrnent were sequenced manually, the sequences were confïrmed upon the aligmnent and were also sequenced ushg the automated 94 services. The corresponding sequences fiom each individual species were aligned where the

overlapping regions were identical, and the cDNA sequences were determined (GenBank #

AF292650-AF292654).

To know the approramatte size of the preprosomatostatin message, Northern blot analysis

was used. The total RNA fiom the dflerent species were hybridized with a labeled cDNA probe

fiom the arawana and resulted in a major band (- 0.6 Kb) fiom the arawana, (- 0.675 Kb) fkom

the butterfly fish and (- 0.680 Kb) fiom the elephantnose, (- 0.730 Kb) fiom the feather £inMe

fkh and (- 0-8Kb) i?om the white sucker (Fig. 33).

The open reading fiame encodes for 1 15 amino acids in arawami (Fig. 34 A), 1 14 amino

acids in the other osteoglossomorphs (Figs. 34 B-D), and 120 amino acids in the white sucker

(Fig. 34 E). The encoded 115 or 1 14 amino acids represent the signal peptide of 24 amino acids

corresponding to the 5' end in all the species except in the buttedy &h, where there were 25

amino acids, and the white sucker, where a 26 amino acids signal is predicted. The si@ peptide

is followed by the spacer region or the non-hormonal region of the somatostatin gene, whicb is

63 amino acids in three of the species, but 61 in the butterfly eh, and 65 in the white sucker.

The remahhg sequence revealed a [T~T',~l~l~] SST-14 (Appenduc 3A), which is predicated to be cleaved at the paired basic site (RK) in all the species. However, the extended region, which is likely cleaved at the single basic site (R), is variant among the studied species.

The variation in location of the single basic site predicted SST-27 in the arawana and butterfly hh, and SST-26 in the elephantnose and the feather fin knife fih (Fig. 35). However, the white sucker SST-28 cm result fiom the processing at a single basic site, this resulting peptide cm potentially be cleaved resulting in nonvariant SST- 14 (Appendix 3B). In arawana and buttedy fis4 SST-27 varied fiom each derin only four residues at the N-terminal, while the elephantnose and feather fh knife f%hSST-26 varied fiom each other in only one residue at the 95 N-terminal. The N-terminai of white sucker SST-28 varied fiom the arawana SST-27 in five residues, and seven residues fioom the SSTs of the other species.

The sequences fiom the osteoglossomorphs, the white sucker, and O ther vertebrates were

aligned (Fig. 35) and used to construct a phylogenetic tree. The resulting single tree revealed two different groupings of the osteoglossomorph species. Arawana and butterfiy fkh were grouped together and elephantnose and feather fin Wefish grouped together, whereas the white sucker was grouped with the goldnsh (Fig. 36). Based on this analysis, the group (arawana and butterfly f%h) was closely related to teleost PPSS II, and the other group (elephantnose and feather fin k.efiçh) was closely related to teleost PPSS 1. Fig. 32 A schematic representation of the preprosomatostatin cDNA The open box uidicate the

predicted open reading fiame, mcluding the signal peptide, the spacer region and the

sornatostatin coding regioa The locations of the oligonucleotides are indicated with short

thick hes foliowed by thin hes of the cloned cDNAs. Somatostatin Signal peptide 5' S pacer coding motif 3' 97

Fig*33 Total RNA samples were hybndized with 32~-labeledpreprosomatostatin cDNA &om

arawana This resulted in bands of - 0.6 Kb £kom the arawana (A), - 0.675 Kb fiom the butterfly &h (B), and - 0.68 Kb fiom the elephantnose (E), and - 0.730 Kb fiom feather Mefish (F), and - 0-8Kb f?om the white sucker (S). The size of each band was determined using an RNA ladder. ABE FS Fig. 34 Preprosomatostatin cDNA sequences. The nucleotide sequence is presented with the

deduced amino acid sequence indicated by single letters. Potential processing sites are

shown in bold. The predicted signal peptide (single underhe) and the polyadenylation

signal (double underline) are hdicated. A) fiom îhe arawana , B) fFom the butterfly fïsfisq

C) fiom the elephantnose, D) f?om the feather fm kniffkh, and E) fkom the white sucker. CCCCCACCGCGTTCAGACCAGCACAGCTCCTTCATC

MKICQIHCTLVLLGL ATG AAG ATC TGC CAA ATC CAC TGC ACC CTG GTA CTC TTG GGG CTT

VLGLYCPSAASQPDL GTT CTG GGC CTG TAC TGC CCC AGC GCT GCT TCA CAA CCT GAC CTG

RYRSFLQRAHAAAMS CGC TAC CGA AGC TTC CTC CAG AGA GCG CAT GCT GCC GCC ATG AGC

PQDWSKQAVEELLÇR CCA CAG GAC TGG AGC AAG CAA GCG GTG GAA GAA CTC CTG TCC CGG

LAPAQGEVPQGAVSA CTG GCC CCC GCA CAG GGT GAG GTC CCC CAG GGA GCG GTG TCT GCG

ADEEEDVRVDLERSL GCA GAC GAA GAG GAG GAC GTC CGT GTG GAT TTG GAG CGC TCT TTG

ELNNLPPRERRAGCK GAA CTC AAC AAC CTG CCT CCG CGG GAG CGG AAG GCC GGC TGT AAG

NFYWKGFTSCSTOP AAC TTC TAC TGG AAG GGC TTC ACC TCC TGC TGA GATCTCGCCGAGGCT

TGAACCGTTTGTCGCGGAAGCACCACACAGATTTATCACATC

GTACAATAAAGCGAAAAGCTCCATTTGAAA GAGCCTTCCATTCAGCTGTCCACAGCATTCCCAGCCCTGTCTAATTCCACATCCACA

MKLCQVHCILALLGL ATG AAA CTC TGC CAA GTC CAC TGC ATA CTG GCT CTG CTG GGA CTT

VLGMCGSSSATQLDS GTG CTG GGC ATG TGC GGT TCA AGC AGT GCC ACC CAG CTG GAT TCT

RYRSLVQRARAASMG CGC TAC CGC AGC CTT GTG CAG AGA GCC CGC GCC GCT TCC ATG GGG

PQDGKLSVEDLSLL CCA CAG GAC TGG GGT AAA CTT TCT GTG GAPL GAC CTG TCC CTG CTG

AATEADMPFGDMSAA GCC GCT ACA GAA GCC GAC ATG CCA TTC GGG GAC ATG TCC GCT GCA

EESEGAHLDLERSVE GAG GAFL AGT GAG GGC GCT CAC CTG GAT CTC GAG CGC TCG GTG GAG

PGNVPPRERKAGCKN CCT GGC AAT GTG CCA CCA CGG GAG CGG AAG GCC GGC TGT AAG AAC

FYWKGFTSCSTOP TTC TAC TGG AAG GGC TTC ACT TCC TGC TGA GGTTGAGCCGGTTTCTGCC

CCCATCAGTGTTATATGCAACTAGGGCAGAGCTGTGATTGWZAAATTTATAAAATGTTG GACAGAAG

MLSSRI QCALALL TAGGAAGG ATG CTT TCC AGT CGT ATC CAA TGC GCT CTT GCG CTC CTC

SLALAVSSVSAAPSD TCT CTA GCG CTG GCG GTT AGC AGC GTG TCT GCA GCG CCG TCA GAC

LKLRQLLQ RSLLAPA CTG AAA CTG CGC CAG CTG CTC CAG AGA TCG CTG CTT GCA CCT GCA

SKQDLARN PLEELLS AGC AAA CAG GAT CTA GCC CGA AFL%C CCC CTA GAA GAA CTG CTG TCC

EMVRVENEALEPDDL GAA. ATG GTG CGG GTA GAG AAC GAA GCG CTG GAG CCT GAC GAC CTG

SRGADQEEVRLELER TCC CGC GGA GCG GAT CAG GAA GAA GTG CGT CTC GAA CTT GAA CGA

AAGPALAP RERKAGC GCC GCC GGT CCC GCT TTG GCT CC= CGG GAG CGG AAG GCC GGC TGC

KNFYWKGF TSCSTOP AAG AAT TTT TAC TGG AAA GGT TTT ACT TCC TGC TGA GATTCCTTCCA

ATCWATCMGTAATCCTAGTCTCTCACACCCCCTACCCTGATATATACTGAATAAAT

ATAAAGT GCAGACAAA CGGAGTCACACTTTAGCCTCGCGAGTTATCTCACGTTTA

MLSTRIQCALAL ACTTAGAAGAG ATG CTT TCC ACT CGT ATC CAA TGC GCC CTT GCG CTC

LSLALPVSSVYAAPS CTC TCT CTA GCG CTG CCG GTT AGC AGC GTC TAT GCA GCG CCG TCA

DLKLRQLLQRSIIAP GAC CTT AAA CTG CGC CAG CTG CTT CAA AGG TCC ATT ATT GCA CCC

ASKQELARYTLAELL GCA AGC AAA CAG GAG CTG GCA CGA TAC ACA CTA GCA GAA CTG CTG

SELAQVENEALESDD TCG GAA CTG GCG CAA GTA GAG AAC GAA GCG CTC GAA TCT GAC GAC

LSRGADQEEVRLELE CTG TCC CGT GGA GCA GAT CAG GAA GAA GTG CGT CTC GAG CTG GAG

RAAGPPLAPRERKAG CGA GCT GCC GGC CCT CCA CTG GCT CCT CGG GAG CGG AAG GCC GGC

CKNFYWKGFTSCSTOP TGC AAG AAT TTC TAC TGG AAG GGC TTC ACT TCC TGC TGA GGCCCCA

CCCAGGCCCCGCCCGCTMTACAATCGTGTCCTCCCTGCCACCCTTAACTGGGCTAAA

TTATGCAAGGAAAAACTTTTGCTT GCTTGT CAAAT CTATAATGTACATAATAATCTT CAAATA AGAAC

MRLCELHCYLAL AGCAGTTTGAAATG AGA CTG TGT GAG CTT CAT TGT TAT TTG GCC CTG

LGLSLVLCDRGADSQ TTG GGT CTG TCT CTT GTG CTA TGT GAC CGC GGT GCT GAC TCA CAA

LEPDMDFRHRRLLQR CTA GAA CCA GAC ATG GAC TTC CGC CAT CGT AGA CTT TTG CAG AGA

ARAIGLATQDWTKKD GCA CGT GCA ATT GGT TTG GCT ACA CAG GAC TGG ACT MGAAA GAC

IEELLSQLSLPEIEA ATA GAG GAG CTG CTT TCC CAG CTG TCT TTG CCT GAG ATA GAG GCT

RENGVSTTGGNDDLH CGT GAG AAT GGT GTT TCT ACG ACA GGT GGG MT GAT GAT CTG CAT

LELERSAENTNQLYP TTG GAG CTA GAG CGT TCC GCA GAG AAC ACG AAT CAG CTC TAC CCT

RERKAGCKNFFWKTF CGG GAG CGG AAG GCC GGC TGT AAA AAT TTC TTC TGG AAA ACT TTC

T S C STOP ACG TCG TGT TAA TTACTCTCCGCAAAGCGTTTGTTAGTTTGTTTGTTTGTTTGTA

ATTTACTTTTCTTATTCCTTCTGTTCTCCGCCTTTCACATCTGTATATAAGNTAAA

ATGGGTTATAGCCTATTTGTTGTTTGACGATAAAGATGATAGCTCTCATTGACTATTTA

AATAAAAT CTAT GTTT CAA?@A Fig. 35 Alignment of preprosomatostatin sequences fkom different vertebrates includuig the

osteoglossomorph species. The amho acid sequence is numbered starting at the

beginning of the signal peptide. The amino acid gaps are denoted by dashes (-).

The potential cleavage sites are in bold, and the resdting SST is indicated in the box on

top of the aligned sequence. The amino acid gaps in some species can redt in a shorter

form of SST as indicated in the other box on top of the aligned sequences.

References of the sequences fiom different vertebrate species: goldnsh, Lin et al., 1999;

rainbow trout, Moore et al., 1995; angIerfish I & LI, Hobart et al., 1980; Goodman et al.,

1980; catfish, Minth et al., 1982 ;lungfisk Trabucchi et al., 1999. Arawana Butterfiv fish Elephantnose Feather fin

Catfish AnglerfishI Lunsflsh

Ara- Butterflv fish Elephantnose Feather fin

Catfish AnglerfishI Lungfish

Butterflv fish Elephantnose Feather fin

Caffish AngleEfishI Lungfish 100

Fig. 36 Phylogeny of preprosornatostatm sequences. Sequences fiom osteogIossomorphs and

diverse vertebrate species (see Fig. 35) were analyzed by parsimony. One tree was

obtained, the group (arawana and buttedy fkh) was closely related to teleost PPSS II,

and the other group (elephantnose and feather fin knife fish) was closely related to teleost

PPSS 1. This indicate the presence of two PPSS genes. The names of some of the

species are abbreviated. Bootstrap probabilities of the relationships are shown on the

corresponding branches. Bootstrap

IElephantnose

Catfish

Anglerfish I

Arawana

-Butterfly fish 7Goldfish

Rainbow

Anglerfish II

African lungfish Discussion

Structure heterogeneity

Somatostatin (SST) peptides of varying lengths (14 to 37 amùio acids) have been isolated andor predicted from cDNAs fiom a variety of vertebrate species (Sheridan et ai., 2000). SST-

14 was fkst isolated fiom the sheep brain (Brazeau et ai., 1973), and since then many SST- 14 molecules of similar structure were characterized fiom representatives fiom all major vertebrate groups. This conserved form of SST is called the invariant form and the gene that codes for it is called PPSS or PPSS 1 gene. However, other forms of SST- 14 (variant SST- 14) have ken identifïed. For example, ratfkh [sers] SST- 14 (Codon, 1990) and the sturgeon pro2] SST- 14

(Nishi et al., 1995). hother variant form, [~yr', ~l~"]SST- 14 has been reported in many teleosts, and the gene that codes for it is lcnown as PPSS 11 (Hobart et al., 1980a; Moore et al.,

1995). Furthemore, three distinct somatostatin cDNAs were isolated Çom the trout: PPSS 1, encoding SST-14; PPSS II' and PPSS II", encoding [~y-r',~l~''] SST-14 (Moore et al., 1995,

1999; Sheridan et al., 1997; Kittilson et al., 1999). Three different cDNAs fkom the goldfbh brain (Lin et al., 1999) also hdicate PPSS 1, encoding SST-14, PPSS II, encoding lu', ~y',

~l~'']SST-14, and PPSS III, encoding [pro2] SST- 14.

In this study a preprosomatostatin have been cloned fiom each of four osteoglossomorphs that contains the variant teleost form of SST- 14, [~yr', ~l~'']SST- 14. This SST- 14 is separated by cleavage sites fiom an N-temiinal extended form of SST. These larger isoforms included

SST-27 fiom the arawana and the butterfly fis4 and SST-26 fiom the elephantnose and the feather fin knife fish. The arawana and butterfly f%h SST-27 varied fiom each other, and fiom salmon and trout SST-25 in four residues. However, arawana SST-27 varied fÏom goldnsh SST-

28 in five residues, and fiom anglerfish SST-28 in four residues. The butterfiy &h SST-27 102 varied fiom goldssh SST-28 in six residues, and fiom anglerfish SST-28 in i5ve residues. On

the other hand, the elephantnose and the feather £inHe f~h SST-26 varied fiom each other in

oniy one residue, fiom salmon and trout SST-25 in six residues, nom goldnsh SST-28 in nine

residues, and fiom anglerfïsh SST-28 in eight residues. SimiZarly, various SST-isoforms were

reported fiom other species, SST-28 fiom mammrils and m~nyother species (flounder, Codon et al., 1987; goldfïsh, Lin et al., 1999; trout, Moore et al., 1993, SST-27 f?om the lungkh

(Trabucchi et al., 1999), SST-26 @owfï.n, Wang et al., 1993; trout, Kittilson et ai., 1999), and

SST-25 (salmon, Plisetskaya et al-, 1986; trout, Sheridan et al-¶1997). In this study we also identified a PPSS cDNA fiom the white sucker that encodes invariant SST- 14 and potentidy a

larger form (SST-28). The white sucker SST-28 varies fiom both the goldfïsh SST-28 11, and anglerfish SST-28 I in seven residues, and fiom trout SST-26 1 in eight residues. The available cDNA information indicates that some of these isoforrns were encoded by either PPSS 1 or PPSS

II genes or by both as in the case of anglerfish SST-28 1 and SST-28 II (Goodman et al., 1980;

Hobart et al., 1980a).

In osteogIossomorphs, 1 previously reported a stronger immunoreactivity to anti-salmon

SST-25 in the pancreas than to anti-bovine SST-14 (Chapters 1 and 3). At that time it was suspected that the Werential immunoreactivity might be reflecting the nature of the sornatostatin molecule. Plisetskaya et al. (1986) reported that anti-salmon SST-25 possesses two binding sites, one is against the 11 N-te& amino acids and the other is against the 14 C- terminal Rmino acids. The comparative analysis of the deduced SST sequences with those of salmon SST-25 II revealed a hi& similarity between the two. The imrn~flohistochemicalstudies are usefiil preliminary indicators of potential molecular koforms withùi fish species (Youson and

Al-Mahrouki, 1999).

Raynor and Reisine (1992) reported that the disuEde bond (cYs3)is important for 103 bioactivity and it is bighly conserved in both variant md invariant forms of SST. Although

the bioactivity of SSTs in the osteog~ossomorphswas not examine4 the results show the

conservation of the sites involved in the disd£ïde bond formation in both the osteogIossomorphs

and the white sucker. Raynor and Reisine (1992) also reported that the sequence, FWKT,

corresponding to positions 7 to 10 constitute a p-tum and are very critical to the function of

SSTs. In the white sucker these residues are conserved, whereas in the four osteoglossomorph

species the variant form of SST is present with residues YWKG at positions 7 to 10. This latter

sequence of amino acids is not uncommon in teleosts whether they are basai, general or derived

species (see Sheridan et al., 2000). However, this amino acid sequence in teleosts is a product of

the gene PPSS II.

Phylogenetic analysis

The phylogenetic and comparative analyses of the osteoglossomorph PPSS sequences

with those of other teleost species resulted in two different groupings of the osteoglossomorph

species. In one group (arawana and butterfly fkh) the PPSS was closely related to teleost PPSS

II, and the other group PPSS (eiephantnose and feather fin kdefish) was closely related to teleost PPSS 1. This fïnding implies that despite the presence of [TJT', SST-14 in all the osteoglossomorph species they may be products of two preprosomatostatin genes. PPSS 1is believed to code for SST- 14 1, and this invariant SST is present in aIl studied vertebrate species.

As for the white sucker, SST-14 was encoded by a precursor that is most sirniIar to the PPSS II gene fiom goldfish, another cyprinid. These two sets of hdings fkom the sucker and fkom the osteoglossomorph species suggest that either the variant and the invariant forms of SST-14 in these species can be encoded by PPSS 1or PPSS II, respectively, or that their PPSS genes are markedly different fiom those previously reported in the trout and the goldfïsh.

The presence of at least two preprosomatostatin genes in teleosts in generai, and three 1O4 reported Eom the trout and the goldfïsh, iin addition to the many SST isoforms identified fiom various species such as fiog brai. pro2, M~~~~IssT-~~(Vaudty et ai., 1992; Tostivint et al.,

1996), sturgeon pro2]ss~-14 (Nishi et al.,1999, ratfkh [serS]ss~-14(Conlon, 1990), and lamprey [~er'~]~~~-14(Andrew et al., 1 988) indicate that SST arose fkom a gene that has been duplicated. The existence of an SST gene: duplication event in the ancestor of the species is suggested by rnany hvestigators (Conlon et al., 1997b; Sheridan et al., 2000). Having more than one gene in certain species but not others mdicates that the duplication probably took place after the divergence of the group to which these species belong. The lack of information (cDNA data) of PPSS in Agnatha does not ailow us to speculate on how eariy in vertebrates SST gene duplication occurred. Ho wever, the availaible pro tein seqüencing data of various S ST iso forms fiom lampreys (for review see Sheridan etd., 2000; Youson, 2000) might suggest that these isoforms may have arisen fiom two or mo-re PPSS, indicating that SST gene duplication may have predated the teleosts.

The phylogenetic anaiysis in the present study resulted in one tree, but the analysis did not help to cwthe phylogenetic positioan of the osteoglossomorphs within teleosts, because the resdts suggest the presence of two genes. However, arawana and butterfly fïsh were grouped together; elephantnose and feather fin Mefïsh were grouped together; and the white sucker was grouped with another cyprinid, goldfish. Ka is noteworthy that a similar grouping of

O steoglo ssomorph species appeared when their insulin cDNAs were used in parsimony analysis

(chapter 4). These latter data, in turn, supported earlier morphological taxonomie data (Alves-

Gomes and Hopkins, 1997; Li et al, 1997:; Zhang, 1998). The parsimony analysis with insulin also grouped the white sucker with the common carp, which is also a cyprinid.

In conclusion, the results of the present study suggest the possibility of having two PPSS genes in the osteoglossomorphs as is the case in other teleosts. This investigation also indicates 105 that fùture studies on Osteoglossomorpha and other teleost SSTs are required. For instance, the presence of another PPSS gene in the osteoglossomorphs should be considered. Furthemore, the study supports the view of a great deal of diversity amongst the teleostean vertebrate group.

In addition, the differential processing of PPSS genes is one of mmy interesting features of their biology that can be used to study molecular evolution among the vertebrates. General Discussion

In this study the gastro-entero-pancreatic (GEP) system in osteoglossomorphs is intensively investigated using various approaches. The attraction for studying the GEP system in this group followed an initial hterest in the development of the system in &h in general. The

Osteoglossomorpha are a very interesthg group of teleosts to study in their own right, and the reasons for using them as a mode1 were detded in the general introduction to the present study.

Briefly, their taxonomy arnong the teleost was debated, with no clear conclusion about whether they are the most basal living teleost, or a more generalized teleost (Nelson, 1973; Arratia,

1991). Certainly, osteogIossomorphs have an ancient origin, and it is generdy believed that the wide geographical distribution of its present rnembers is influenced by the tectonic plate movement (Li and Wilson, 1996). To investigate this system, five species representing the major taxa of Osteoglossomorpha were chosen; in addition, two more species, representing a generalized and a derived teleost, were studied for comparative purposes. The principal objectives of this study were fkt, to investigate the distn'iution and fom of the GEP system of the osteoglossomorphs relative to tbat of the other teleosts; and second, to examine the possibility of using the primary structure of the peptides of the GEP system in a phylogenetic analysis. To füEU the objectives of this investigation, several studies were designed. The results of those studies are discussed below.

In the past, only passing attention was made to the morphology of the pancreas of members of the osteoglossomorphs (Mc Cormick, 1925; Epple and Brinn, 1975; Langer et al.,

1979). The present study serves as the first detailed investigation of the GEP systern in

Osteoglossomorpha vouson and Ai-Mahrouki, 1999). GenetaUy, in actinopterygians the distribution of islet tissue seems to have a developmental-phylogenetic pattern, where the islet tissues are scattered throughout the exocrine tissues in the basal (non-teleo st) actinopterygians 107 and are in a more specialwd form in the endocrine pancreas of the more derived teleosts,

where they form principal islets (Epple and Brinn, 1975).

The histological and the imm~11ohistochemicalstudies Oerein characterize the GEP

system of the osteoglossomorphs as a well-developed system, where islets of variable size were

scattered tbroughout the exocrine pancreas. This pattern of islet distribution does not quite fit the

term Brockmam body as defined by Youson and Al-Mabrouki (1999). In the eels (Elopomorpha,

a basal teleost group), a scattered distriiution of the islets was descriid Kobayashi and

Takahashï, 1974). However, in the glass eel, L'Hermite et al. (1985) observed a single islet in the

middle of the pancreas in addition to the other scattered islets that are found in the head and tail

of the pancreas; the authos cded this single islet a 'Brockmann body". In contrast to the

absence of a Broclanann bodies in the osteoglossomorphs, Brockmann bodies were O bserved in two euteleosts; the white sucker (generalized teleost) and the rock bass (derived teleost). The presence of a Brockmann body is considered to be a more specialized form of the endocrine pancreas in teleosts (Epple and Brinn, 1975). In the rainbow trout and coho salmon

(Salmoniformes, generalized teleosts), large islets of the size of principal islets are surrounded by exocrine acini (Wang et al., 1986; Nozaki et al., 1988a,b). This pattern of distribution in sahonids was considered to be an intermediate step in the phylogenetic development of the principal islet (Youson and Al-Mahroluki, 1999). This pattern is also similar to that found in the osteoglossomorphs, where large islets of a size approaching those of principal islets were surrounded by exocrine acini.

The osteoglossomorpba was considered to be a basal teleost, and recently it was argued to be a sister group of the elopomorphs (eels) (Arratia, 1991; O'Neill et al., 1998), which have a more or less similar islet distri'bution pattem The scattered distribution of large and small islets wahin the exocrine pancreas in the osteoglossomorphs, and the absence of Brockmann bodies 108 (a highly concentrated region of isIets both small and large) may irnply that its endocrine

pancreas is of the basal type. However, the presence of large islets in its pancreas, similar to

those in salmonids, may suggest an intermediate stage of deveIopment. The scattered pattern of

isIet distriion in the pancreas was reported in many species includmg basal Actinopterygii,

and tetrapods. Examples of studies showing scattered islet distri'bution in basal Actinopterygii

included Po2yptem.s and Amia (Epple and Brinn, 1975); this distriiution in Amia has recently

kenconfirmed (Youson et al., 2001). Furthemore, the presence of Brockmann bodies/ principal

islets were only reported in teleosts but not in any of the tetrapod vertebrates. This may suggest

that the use of the endocrine islets distriiution pattern in inferring phylogenetic development in

vertebrates is not an optimal parameter. To Merinvestigate the position of the GEP system of

the osteoglossomorphs within the actinopterygians, another parameter was considered, where the

endocrine cell types were identifïed using immunohistochemical and immunocytochemical

techniques. When considering the presence and absence of classical ceil types in the endocrine

pancreas, Youson and Al-Mahrouki (1 999) indicated that these morphological parameter infers a

phylogenetic sequence in vertebrates.

In the osteoglossomorph species investigated herein, four endocrine cell types (A, B, D

and F) were identifïed in the endocrine pancreas. This conclusion was fïrst reached using various

antisera for glucagon fâmily peptides, insulin, several forms of sornatostatin, and the pancreatic polypeptide fàmily at the light microscopic level. Confirmation of these results was checked by

examining the immunoreactivïty at the cellular level. The presence of the four classical endocrine cells in the pancreas ofthe osteoglossomorphs can be considered as a tendency toward specialization. In vertebrates, the presence of the pancreatic endocrine cek outside the gut

started in the lamprey larva, with a one-hormone islet containing B-cells. The number of endocrine cek in the islet increased to two (B- and D-cells) in the hagfkh, and to three (B-, D- 109 and F-cek) in the addt lamprey or to three @-, D- and A/F-celis) in the gar, a non-teleost

actinopterygian (Groff and Youson, 1998) - Furthemore, the nurnber of the endocrine cek in the

islets reached four ceIl types (A-, B-, D- and F-cells) in most other f%h. This trend towards

increased cell types in fish is representative of the phylogenetic sequence in vertebrates in

general (Youson and Ai-Mabrouki, 1999).

In addition to the presence of the four classical endocrine cell types in the osteoglossomorph endocrine pancreas, immunocytochemistry revealed the presence of two

somatostatin cell types, and this is a trend towards more spec~tionamong fishes. The two D cell types (somatostatin cek) were characterized using anti-SST-14, anti-SST-25, and anti-SST-

34 sera, where SST-14-like immunoreactivity and SST-25-like immunoreactivity were detected in the same ceil type, and SST-34-like irnrnunoreactivity was detected in another celI type. These observations were mfEciently interesting to encourage Merinvestigation around the identification of somatostatin mo lecules, since the morpho logicd data implied the possible presence of somatostatin isoforms. Therefore, a molecuIar analysis of somatostatin in osteoglossornorphs was carried out; these findings wiU be discussed later in th& section. The presence of two D-cell types was reported in other teleosts (anglerfish, McDonald et al., 1987; coho salmon, Nozaki et al., 1988a; sea bas, Lozano et al., 1991a; turbot, Reinecke et al., 1997), and was correhted to somatostatin isoforms.

Although immunoreactivity to insuiin antihdy using light microscopy was more or less similar among the different species, an interesting morphological finduig was revealed using immunocytochemistry (electron microscopy). The observation of polymorphic shapes of the B- ceil granules in ody one of the studied osteoglossornorphs (0.bicirrhosum) indicated that there may be interspecinc differences in insulins. At this level of the study it was suspected that a modincation in a zinc binding site could be due to an amino acid substitution, with a resulting 110 morphoIogica1 variation of the granules. This assumption was based on earlier reports of

zinc binding effects on the rnorphoIogy of insulin grandes (Nakamura and Yokote, 1971;

Wagner et al., 1981; Baker et ai., 1988). This variation in rnorphology of insulin granules was

the stimulus to investigate the primary structure of insulin in O. bicirrhosum and other

osteoglossomorph species.

It is accepted that insulin cells leave the digestive tract eariy in both ontogeny and

phylogeny (Faber, 1995). In Osteoglossomolpha, insulin cells were observed in the digestive

tract of P. buccholzi @utter£ly hh), were Sequent in the intestine of O. bicirrhosum and G. petersii, and were not observed in the digestive tract of the other species. This finding may

assume a tendency towards an intermediate stage in the phylogeny of osteoglossmorphs, with

insulin ceUs present only in the endocrine pancreas. It may also imply a doser intragroup

relationship among the species in which insuün cells were observed. This finding was another

stimulus to characterize the insulin mokcule in the osteoglossomorphs. The presence of insulin

ceils in the digestive tracts has not only been found in some osteoglossomorphs, but has been

documented in other fish species as weLl, for example in lamprey (a basal vertebrate; Youson and

Potter, 1993) and in the gar (a basal actinopterygian; Groff and Youson, 1997). On the other

hand, the presence of insulin cek in the gut of both the sea bass (a derived teleost; Hernandez et

al., 1994), and the turtle (a derived vertebrate; Gapp and Polak, 1990), suggests that peptide

distri'bution within the GEP system should not be a measure of phylogenetic relatedness; this

idea has ken impiied in several other studies (Langer et al., 1979; Gomez-Visus et al., 1996).

In the present study, immunohistochemistry was used to identify the distniution and

types of the endocrine celIs in the GEP system in the white sucker and in the rock bass. This

analysis revealed the presence of the classical four cell types in the endocrine pancreas of both

species. In addition, insulin cells were not detected in the digestive tracts of either. To Mer 111 confiun the identity of the cell types additional studies are required. These studies would include ligl~tmicroscopy and double imrnunolabeling to check for any CO-localizationbetween the peptides, or with electron microscopy and immunocytochemistry using double immunolabeling with different sizes of protein A-gold, to check for myCO-localization of the peptides. These approaches have been used recently for the bowfïn (Youson et al., 2001), the gar

(Groff and Youson, 1998) and the osteoglossomorphs (Chapter 3).

Generdy, the presence of the classical four endocrine cell types in the osteogbssomorphs and the scattered distriiution of the islets do not contradict the aim of this study. However, the presence of two sub-D-cell types, and large islets of the size of principal islets shows a well specialized endocrine system which, compared with other actinopterygians, may indicate an intermediate developmental step towards more specihtion.

As descrïbed above, morphological variations in certain aspects of the osteoglossomorph

GEP system aroused a curiosity to pursue further studies of their insulin molecules. To characterize the primary structure of insulin, molecular biology techniques were used. The amino acid sequence of the insulin molecules was deduced fiom cDNAs. The fiill cDNA sequences were identifiecl fkom four osteoglo ssomorphs (0. bicirrhosum, P. buchholzi, iV chitala, and

H. alosoides) and a cyprinid (C.commersoni). In addition, a partial cDNA fiom another osteoglossomorph (G. petersii) was obtained. All the osteoglossomorph sequences were aligned with each other and with insulin sequences available fkom other vertebrates. This aiignment was subsequently used to perform a phylogenetic analysis through the construction of phylogenetic trees.

The use of polypeptide hormones, especially those of low molecular weight, to answer phylogenetic questions has ken approached with caution. This caution is due to the high conservation of certain regions of the peptides and their shoa sequences (Codon and Hicks, 112 1990; Aguileiro, 1995; Dores et al-, 1996). Another more successfûl approach has been to use the variable spacer regions fiom the polypeptide hormones, for example, the use of the C-peptide sequence (Dores et al., 1996). Nevertheless, the sequences of some polypeptide hormones were used in several studies, and proved to be efficient in answering some phylogenetic questions.

Examples of these studies included the use of proinsulin (Hahn et al., 1983; Codon et al., 1997%

1998), of glucagon-like peptide (Irwh et al., 1999; Ennn and Sivarajah, 2OOO), and of preprosomatostatin (Trabucchi et aL, 1999). However, Dores et al. (1996) pointed out that the use of polypeptide hormones in phyl~geneticanalysis probably requires the analysis of several polypeptide sequerces. For this reason a decision to investigate and to analyze the sequence of more than one polypeptide hormone was made. On the other hand, the use of non-hormonal polypeptides of large molecular weight to answer phylogenetic questions may not be an optimal approach. Recently, Kumazawa and Nishida (20 00) used the sequences of mitochondrial pro tein genes in the phylo genetic assessrnent of osteoglossomorphs. Their dysisindicated that the osteoglossomorphs are basai to the carp and trout (generalized teleosts), ho wever, the same analysis was no t successful in inferring the intragro up relationship among O steo glo ssids, notopterids, and morrnyrids.

The phylogenetic anaIysis of insuiin sequences revealed a monophyletic grouping ofthe

O steoglossomorphs, which is consistent with previous reports (Li and Wilson, 1996; Zhang,

1998; Kumazawa and Nihida, 2000). As for the intergroup relationship of the presently investigated osteoglossomorphs species, N. chitala (feather fin Wef%h) and G. petersii

(elephantnose) were grouped together, and O. bicirrhosum (arawana), P. buchhoIn- (buîterfly

Esh) and H. alosoides (go ldeye) were grouped together. The grouping of the feather fin knife f%h and the elephantnose together is consistent with eariier reports (Li et al., 1997). Alves-

Gomes and Hopkins (1997) reported that the Pantodon (e.g., butterfiy &h) departs fiom the 113 momyriform (e.g., elephantnose); this suggestion is supported by the present phylogenetic

analysis. In addition, earlier reports related the Osteoglossidae (e.g. arawana) to the

Pantodontidae (e.g. butterfly &h) (Lauder and Liem, 1983; Van Le et al., 1993), and thk relationship does not contradict the results of the present phylogenetic analysis. However, the

intergroup relationship of the Hiodontidae (e.g. goldeye) to the Other osteoglossomorphs was no t clear. Although Hiodontidae was previously grouped with the Notopteridae (e.g., feather fin

Hefish) (Greenwood, 1973), it was recently reported to be a sister group to ail the osteoglossomorphs (Zhang, 1998; Li et al., 1997; Van Le et al., 1993). In this çtudy using the

insulin sequences in the phylogentic analysis, the goldeye wodontidae) was sho wn to be closely related to the butterfiy fish.

The observation of insulin cells in the digestive tract in 0.biciwhosum (arawana) and in

P. buccholzi (butterfly fish), may indicate a close relationship between the two species. This relationship is supported by the phylogenetic analysis using the preproinsulin sequences of both

O. bicirrhosum (silver arawana) and P. buccholn' (butterfly bh), which indicated that both are not very distant fiom each other.

In this study the phylogenetic analysis revealed a close relationship between the white sucker (cyprinid) and the common carp, and both are cyprinids. The overd evduation of the osteoglossomorph position within Teleostei, based on the insulgi sequences, placed the bony- tongues at a more advanced position than originally thought. This conclusion is supported by the morphological data, which indicated the presence of a well specialized endocrine system in the osteoglossomorphs. It also confirmed its monophyletic grouping. In addition to the molecular phylogenetic anaIysis of insulin in this study, the comparative analysis of the insulin molecular structure is also a contriiution to the comparative endocrinology. The latter contribution is discussed below. The ultrastructural studies reveaIed morphological differences in the shape of insulin secretory granules in 0.bicirrhosurn (silver arawana), in cornparison to tho se fkom the other osteoglossomorphs. As mentioced earlier in this discussion, this finding was suffïciently interesthg to stimulate Merinvestigations into the primary structure of the insului molecule. The presence of a zinc molecule is essential for the hexamer conformation of the insulin molecule (Nakamura and Yokote, 1971; Wagner et al. 1981; Baker et al. 1988), and a substitution of an smino acid wodd prevent hexamer formation: for example, in the hagfïsh, where a histidine was replaced by an aspartate at position B-IO (Chan et aI., 198 1)- It was suspected that this might be the case in 0. biciwhosum, but after analyang the insului sequence of the arawana and comparing it with those fiom other species, the sites assumed to be nivolved in the Zn-binding and hexamer formation were found to be conserved. This result suggests that there may be other factors involved in the insulin molecular co~tion.Codon et al. (1 99 1) suggested that the N-terminal of the B-chah may also play a role in the hexarner formation. This domain is weakly conserved in osteoglossomorphs, and in the arawana serine replaces alanine at position B-1, which may explain the structural diversity of the secretory granules of the B-cells as descnibed in Chapter 3.

The use of immunohisto chemistry, ultrastructure, and immunocytochemistry in which vayhg somatostatin ant~hdieswere used, revealed the identzcation of more than one somatostatin cell me. This finding was the basis for further aaalytical studies of the somatostatin moIecule, and the characterimion of its primary structure. Another rationale behind this section of the study was the wed to analyze more than one polypeptide hormone, so that an effective phylogenetic analysis could be inferred. 115 The same molecular biology techniques, previously desmkd, were &O used to characterized the somatostatin molecules in osteoglossomorphs and in other advanced teleosts.

In this study the prepro somatostath was characterize fkom four O steoglossomorphs

(0.bicirrhosum, arawana; P. buchhoZiiybutterfly fish; N; chitala, feather fin knife fish; and G. petersiï, elephantnose) and fiom a cyprinid (C. commersonz]. Several somatostatin isoforms were predicted; variant somatostatin [~yr'~GI~~~]-SST-~~ fiom aU of the osteoglossomorph species studied, SST-27 fiom both arawana and butterfiy fïsh, and SST-26 fiom feather fin knife £ish and elephantnose. In the white sucker, invariant SST-14, and SST-28 was predicted. The presence of various SST isoforms was reported in other species including

SST-28 (goldfïsh; Lin et al., 1999), SST-27 f?om the lungfkh (Trabucchi et aL, 1999), and SST-

26 fiom trout (Kittilson et al., 1999). These isoforrns probably have varying functions in dBerent species so, in the fbture, uivestigating the function of these isoforms should be considered. The sequences of these isoforms are quite short, and do not yield meaningfid phylogenetic trees in phylogenetic analysis. As a result, the fidl sequences were used including the signal peptide, the spacer region and the somatostatin coding region

The phylogenetic analysis revealed the grouping of the white sucker with the gold£ïsh, and both are cyprinids. However, the arawana and the butterfly fish were also grouped together as were the elephantnose and the feather fin Mefkh. These intergroup relationships were supported by the results fiom the insulin phylogenetic analysis mentioned above, as weU as reports fiom other studies as previously mentioned in the ùisulin section of this discussion.

Nevertheless, the monophyletic grouping of the osteoglossomorphs was not confïrmed by this analysis because the results suggest the possibility of having two preprosomatostatin genes. This suggestion is supported by the earlier immunocytochernical studies (Chapter 3), where the absorption of anti-SST-25 with SST-14 antigen prior to application did not abolish the labeling. 116 This hding indicates the absence of any cross reactivity; however, the immunoiabehg was

weaker with anti-SST-14 than with anti-SST-25. Thus as indicated in Chapter 1, the variable

immunoreactivïty may imply that one SST form is more dominant than the other. The number of

SST genes in osteoglossomorphs will require fürther investigation in the fbture.

In conclusion, the GEP system of the osteoglossomorphs seerns to be a weil developed

system. Although no Brockrnann bodies were observed, a feature typical of more derived teleosts

(Youson and Al-Mahrouki, 1999), the presence of large islets of the size of principal islets

surrounded by exocrine tissues similar to those in some generalized teleosts (salmonids), in addition to the presence oftwo somatostatin ceii types, can be considered as an intermediate developmental stage toward more speciali7iition. The use of polypeptide hormone sequences for perfonning phylogenetic analyses in this study proved to be valuable to a certain extent. This approach cont'irmed the monophyletic grouping of the osteoglossomorphs, and showed an overd generalized position of the osteoglossornorphs within the teleosts. Tt also proved to have some significance in clmgthe intergroup relationship, where the elephantnose and the feather fin Wefkh are cioseiy rehted, and the arawana and the buttedy fkh are also closely reiated. However, for future studies, further investigations of the somatostatin genes are required, in addition to the molecular identification of the other peptides produced by the system (e-g., glucagon fàmily peptides, and pancreatic polypeptide My).Furthemore, molecular data

(cDNAs) are not available from representatives of either Elopomorpha or Clupeomorpha (basal teleosts). These data are important for comparative reasons to give a clearer view of the position of the osteoglossomorph group relative to other putative basai teleosts. Summary

The gastro-entero-pancreatic system (GEP) was investigated in severai osteoglossomorphs, and in two other teleost species. The osteoglossomorph species included:

OsteogIossum bicirrhosum (silver arawana), Scleropagesjurdini (Australian arawana), Puntodon buchholzi (buttedy Gh), Notopterus chitula (feather fin Mefish), and Gnathonernuspetersii

(elephantnose), and Hiodon alosoides (goldeye). The other teleost species included Catostomus commersoni (white sucker) and Ambloplites rupesh.s (rock bass). Several aspects of the system were studied using different approaches and are summarized as foliow:

The morphology and the histology of the GEP system was identified in the different

species using several morphological parameters. This was followed by the

Unmunohistochemical techniques to identify the presence or absence of endocrine ceU

types and their pattern of distri'bution within the system. The cell types were veri6ed

using Mmuflocytochemistry and electron microscopy.

The histological and the immunohistochemical studies characterized the GEP

system of the osteoglossomorphs as a weil developed system, where large and smaU

islets were seen to be scattered throughout the exocrine pancreas. This pattern of

islet distribution does not quite fit the term Brockmann body, as observed in the other two

teleosts: C. commersoni, the white sucker (a generalized teleoçt), and

A. rupesîris the rock bass (a derived teleost).

In the osteoglossomorph species, four endocrine cell types (A, B, D and F) were

identified in the endocrine pancreas. This observation was verified by examiniog the

immunoreactivity at the cellular level (electron microsco~v\.In addition- two D-cell 118 subtypes were also observed, and the CO-localizationbetween A-ceil peptides and F-

cell peptides was not detected.

4) Morphoiogical investigations revealed polymorphic shapes of the B-ceU granules in only

one of the studied osteoglossomorphs (0.bicirrhosum) and variation of insulin

irnrnunoreactivity in the digestive tract.

5) Using recombinant DNA technology, the preprohormones of insulin and somatostatin

were characterized fiom several species through the deduced arnino acid sequences. The

sequences fiom the O steoglossomorphs and diverse vertebrate species were analyzed by

parsimony to infer a phylogenetic relation .

6) The phylogenetic anaIysis of the preproinsulin sequences revealed a monophyletic

grouping of the osteoglossomorphs. These sequences also provided intragroup

relationships of the osteoglossomorph species: N. chMa(feather fia luSe fkh) and G.

petersii (elephantnose) were grouped together, and 0. bicirrhosum (arawana),P.

buchholzi (butterfly fkh) and W. alosoides (goldeye) were grouped together.

7) The preprosomatostatin phylogenetic analysis revealed the grouping of the white

sucker with the goldfish, and both are cyprinids. However, the arawana and the

butterfly fkh were grouped together, and the elephantnose and the feather fin Me

fish were grouped together. These intragroup relationships were paaially supported by

the results fiom the insulin phylogenetic analysis mentioned above. Furtherrnore, the

analysis suggested the presence of two sornatostatin genes in the osteogIossomorphs.

8) The somatostatin structure analysis revealed the presence of several somatostatins:

variant somatostatin [~yr~,G~~'O]-SST-~~ fkorn alI studied osteoglossomorph species, 119 SST-27 fbm both arawana and butterfly fkh, and SST-26 fiom feather fin Eaiife fkh

and elephantnose. In the white sucker, invariant SST-14 and SST-28 were identified.

9) From this study the GEP system of the osteoglossomorphs seems to be a weii

developed system, demonstrated by the presence of four classical cell types

in addition to the presence of two somatostatin cell types. The GEP system f?om the

other studied generalked and deiived teleosts may seem more specialized than that of

the osteoglossomorphs with regard to the presence ofBrockmann body, but the

identification of its ceil types require fbrther study. The use of polypeptide

hormone sequences for inferring phylogenetic anaiysis in this study proved to be

valuable to some extent. Literature cited

Abad, M. E., Agulleiro, B., and Rombout, J. H. W. M. (1986). An immunocytochemical and dtrastructural study of the endocrine pancreas of Sparus auratus L- (Teleostei). Gen. Cornp. Endocrinol. 64, 1-12.

Abad, M. E., Peeze Binkhorst, F. M., Elbal, M. T., and Rombout, J. H. W. M. (1987). A comparative immunocytochemical study of the gastro-entero-pancreatic (GEP) endocrine system in a stomachless and a stomach-containing teleost. Gen. Comp. Endocrinol. 66, 123-136.

Abad, M. E,, Taverne-Thiele, J. J., and Rombut, J. H. W. M. (1988). Tmmunocytochemical Characterization of coexistence of pancreatic polypeptide and glucagon-like immunoreactivity in the pancreatic endocrine cells of Sparus auratus L. (Teleostei). Gen. Cornp. Endocrinol. 70,9-19.

Abad, M.E., Garcia Ayala A, Lozano MT, and Agulleiro, B. (1992). Somatostatin 14- and somatostatin-25-like peptides in pancreatic endocrine cells of Sparus awafa(Teleost): a light and electron microscopic immunocytochemical study. Gen. Comp. Endocrinol. 86, 445-452.

Agulleiro, B. (1 995). Phyiogenetic approach to the vertebrate endocrine pancreas. Ne fierlands J. ZOO.45(1-2), 124-128.

Agulleiro, B., Lozano, M. T., Abad, M. E., and Garcia Hemandez, M. P. (1993). Electron- microscopic immunocytochemical study of the endocrine pancreas of sea bass Dicentrarchus lm.Ce12 Tissue Res. 274,3 03-3 14.

Al-Mahrouki, A.A. and Youson, J.H. (1998). Tmmunohistochemicai studies of the endocrine ce& wahin the gastro-entero-pancreatic system of Osteoglossomorpha, an ancient teleostean group. Gen. Cornp. Endocrinol. 110, 125- 139.

Al-Mahrouki, A-A. and Youson, J.H. (1 999). Ultrastructure and immunocytochemistry of the islet organ of Osteoglossomorpha (Teleostei). Gen. Comp. Endocrinol. 116,409-42 1.

Al-Mahrouki, A.A., Irwin, D. M., Graham, L. C., and Youson, J.H. (200 1). Molecular cloning of preproinsulin cDNAs fiom several osteoglossomorphs and a cyprinid. Mol. Cell. Endocrinol. l74(1-2), 5 1-58.

Alves-Gomes, J., and Hopkùis, C.D. (1997). Molecular insights into the phylogeney of Momyriform fishes and the evolution of thek electric organs. Brain Behav. Evol. 49,324-35 1.

Andrews, P.C., Pollock, H. G., ELliott, W. M., Youson, J. H., and Pketskaya, E. M. (1988). Isolation and characterization of a variant somatostatin- 14 and two related somatostatins of 34 and 37 residues fioom hmprey (Petromyzon marinus). J. Biol. Chem. 263, 15809-15814. 121 Arratia, G. (1991). The caudal skeleton of Jurassic teleosts: A phylogenetic analysis. In " Early vertebrates and reiated problerns of evolutionary biology" (Chang, M-M., Liu, Y.H., Zhang, GR), pp. 249-340. Science Press, Beijing.

Baker, E.N., Blundell, T.L., Cutfield, J.F., Cutfïeld, S.M., Dodson, E. J., Dodson, G.G., Hodgkin, D.M.C., Hubbard, RE-Jsaacs, N-W., Reynolds,C.D., Sakabe, K, Sakabe, N. and Vijayan, N.M. (1988). The structure of 2Zn pig insulin crystak at 1.5 A resolution. PhX Trans. R. Soc. Lond B 319,369-456.

Baldissera, F. G. A., Holst, J. J., Jensen, S. L., and Kraup, T. H- (1985). Distrr'bution and molecular forms of peptides containing somatostatin immunodeterminants in extracts fiom the entire gastrointestu-ialtract of man and pig. Biochem. Biophys. Acta 838, 132-143.

Barriflgton, E. J. W. (1945). The supposed pancreatic organs of Peb-ornyzonfluviatilis and Myxine glutinosa. Q. J, Microsc. Sci. 85,391-41 7.

Beccaria, C., Diaz, J.-P., Gabrion, J., and Connes, R (1990). Maturation of the endocrine pancreas in the sea bas, Dicentrarchus labrm L.(Teleostei): an immunocytochemical and ultrastructural study. 1. Glucagon-producing ce&. Gen. Comp. Endocrinol. 78, 80-92.

Berwert, L., Segner, H., and Reinecke, M. (1995).0ntogeny of IGF-I and the classical islet hormones in the turbot, Scophthalrnus marimus. Peptides 16, 113- 122.

Boquist, L. and Patent, G. (1 971). The pancreatic islets of the teleost Scolpaena scropha. An ultrastructural study with particular regard to fibrillar granules. Z Zellforsch. 115, 4 16-425.

Borner- Weir, S ., and Weir, G. C. (1 979). The organization of the endocrine pancreas: A hypothetical unifjrhg view of the phylogenetic differences. Gen Comp. Endocrinol. 38328-37.

Brazeau, P., Vaie, W., Burgus, R,Ling, N., Butcher, M., Rivier, J., and Guillemin, R (1973). Hypothalamic polypeptide that inhibits the secretion of immunoreactive pituitary growth hormone. Science 179, 77-79.

Brinn, J. E., Jr. (1973). The pancreatic islets of bony fishes. Am. Zool. 13,653-665.

Brinn, J. E. (1975). Pancreatic islet cytology of Ictaluridae (Teleostei). Cell Tissue Res. 162, 357-3 65,

Brinn E. J., and Epple A. (1990). Phylogenetic patterns of the endocrine pancreas. Progress in Comparative EndocrinoZogy, Wiley-Liss, Inc., 493-498.

Carrillo, M., Zanuy, S., Duve, H., and Thorpe, A- (1 986). Identification of hormone-producing cells of the endocrine pancreas of the sea bas, Dicentrachus labrm, by ultrastructural irnmunocytochemistry. Gen. Comp. Endocrinol. 61,287-3 0 1. Chan, S.J., Emdin, S.O., Kwok, S.C., ber,LM., Falkmeer, S., and Steiner, D.F. (1981). Messenger RNA sequence and primary structure of preproinsulin in a primitive vertebrate, the Atlantic hagfïsh, J- Biol. Chern. 256(14), 7595-7602.

Chan, S.J., Episkopou, V., Zeitlin, S., Karathanasis, S.K, MacKreil, A., Steiner, D.F., and Efktratiadis, A., (1 984). Guinea pig preproùlsulin gene: An evolutionary compromise?. Pro-Natl. Acad. Sei. 81,5046-5050.

Cheung, R,Andrews, P.C., Plisetskaya, E. M., and Youson, J. H. (1991). Tmmunoreactivity to peptides bebnging to the pancreatic polypeptide family (NPY, aPY, PP, PYY) and to glucagon-like peptide in the endocrine pancreas and anterior intestine of adult lampreys, Petromyzon marinus: An immmohistochemical study. Gen. Comp. EndocrinoL 81, 5 1-63.

Chiba, A., Honma, Y., and Oka, S. (1995). Ontogenetic development of neuropeptide-Y- me- immunoreactive cek in the gastroenteropancreatic endocrine system of the dogfïsh. CeZl Tissue Res. 282, 33-40

5 Codon, J.M. (1990). (Ser )-somatostatin- 14: Isolation fiom the pancreas of holocephalan fis& the pacific ratfkh (Wydrologus colliei). Gea Comp. Endocrinol. 80,314-320.

Codon, J. M. (1995). Peptide Tyrosine- Tyrosine (FYY)-An Evolutionary Prespective. Amer. Zool., 35,466-473 -

Conlon, J.M. (2000). Molecular evolution of insulin in non-mâmmalian vertebrates. Amer. Zool.40,200-212.

Codon, J.M., Davis, M.S., Fallaner, S., and Thim, L. (1987). Structural characterization of peptides derived fiom pro somatostatins 1 and II iso lated fiom the pancreatic islets of two species of teleostean fish: the daddy sculpin and the flounder. Eur. J. Biochem. 168, 647-652.

Conlon,. J. M., Deacon, C. F., Hazon, N., Henderson, 1. W., and Thim, L. (1988). Somatostatin -related and glucagon-reiated peptides with musual structural features &oom the European eel Anguilla anguiIIa. Gen. Comp. Endocrinol. 72, 18 1- 189.

Codon, J.M., and Hicks, J. W. (1990). Isolation and ultrastructural characterization of insulin, glucagon and somatostatin fiom the tuale, Pseudemys scripta. Peptides 11,461-466.

Codon, LM., Youson, J. H., and Whittaker, J. (1991). Structure and receptor-binding activity of insulin fiom a holostean fis& the bowfin (Arnia calva). Biochem. J. 276,26 1-264.

Conlon, J.M., Bondareva, V., Rusakov, Y., Piketskaya, E.M., Mynarcik, D.C., and Wttaker, J., (1 995a).Characterization of insulin, glucagon, somatostatin ftom the river lamprey, Lampefrafluviatili~~ Gen. Cornp. Endocrinol., 100, 96-1 05. 123 Codon, J. M., Nieken, P. F.,Youson, J. H., and Potter, 1, C. (1995b). Proinsulin and somatostatin fiom the islet organ of the southern hemisphere larnprey Geotria austraZiss Gen. Comp. EndocrinoZ. 100,41 3-422.

Codon, J.M., Platz, J.E., Nielsen, P.F., Vaudey, H., and VaIlarino, M. (1997a). Primary structure of insului fkom the Afiican Lungfish, Profopterus annectens. Gen. Comp. Endocrinol. 107,421-427.

Codon, J. M., Tostivint, W., and Vaudry, H- (199%). Somatostatin- and urotensin II-related peptides: mo lecular diver sity and evo lutionary perspectives. Regdatory Peptides 69, 95- 103.

Conlon, J-M-, Fan, H., and Fri+ach, B. (1998). Purification and structural characterization of insulin and glucagon fiom the Bichir Poljptertrs senegalis (Actinopterygii: Polypterifomes). Gen. Comp. Endocrinol. 109, 86-93.

Cutfield, J. F., Cuf5eld, S. M., Carne, A., Emdin, S. O., and Fallarier, S. (1986). The isolation, purification and amino-acid sequence of insulin fiom the teleost &h Cottus scorpius (daddy sculpin). Eur. J. Biochem. 158, 117- 123.

De Pinna, M. C. C. (1996). Teleostean monophyl. In 'Tnterrelationships of fïshes" (M. L. J. Stiassny, L. R Parenti, and G. D. Johnson, Eds.), pp. 147-262. Academic Press, San Diego.

Devaskar, S.U., Giddings, S. J., Rajakumar, P. A., Camagh~,L. R., Menon, R.K. and Zahm, D. S. (1994). l[nsulin gene expression and insulin synthesis in mammalian neuronal celis. J. Biol. Chem. 269 (1 l), 8445-8454.

Diclchoff, W. W., Brown, C. L., Sullivaq C. V., and Bern, H. A. (1990). Fish and amphi%ian mode1 for developmental endocrinology. J. Exp. Zool. Szppl. 4, 90-97.

Dores, KM., Rubin, D.A., and Quinn, T.W. (1996). 1s it possible to construct phylogenetic trees using polypeptide hormone sequences?. Gen. Comp. Endocrinol. 103, 1- 12.

Eilertson, C.D., and Sheridan, M.A. (1993). Differential effects of somatostatin-14 and somatostatin-25 on carbohydrate and lipid metabolism in rainbow trout, Oncorhpuchus mykiss. Gen. Comp. Endocrinol. 92, 62-70.

Elbal, M.T., Garcia, A.A., and Agulleiro, B-(199 1). Light and electron microscopic immu11ocytochernicaldemonstration of coexistence of somatostatin 14- and somatostatin 25-like peptides in endocrine ceils of the stomach of Sparus auratus L. Gen Comp. Endocrinol. 84, 3 6-43.

Eliiott, W. M., and Youson, J. H. (1986). Immu11ocytochernical localization of insulin and somatostatin in the endocrine pancreas of the sea larnprey, Petromyzon marinus L. at various stages of its Me cycle. Cell Tissue Res. 243,629-634. 124 ElIiott, W. M., and Youson, J. H. (1988). Fine structure and immunocytochemistry of cek within the endocrine pancreas of baland adult sea lampreys, Petromyzon marinus L. Am. J. Anat. 182, 73-83.

El-Naggar, M. M., Elayat, A. A., Ardawi, M. S. M., and Tahir, M. (1993). Isolated pancreatic islets of the rat: An immunohistochemical and morphmetric study. Anut. Rec. 237, 489-497.

El-Salhy, M., Wilander, E., and Abu-Sinna, G. (1 982). The endocrine pancreas of anuran amphiiians: A histological and imm~11ocytochemicalstudy. Biorned Res. 3, 579-5 89.

Epple, A., and Brinn, J. E. (1975). Islet histophysiology: evolutionary correlations. Gen. Comp. Endocrinol. 27,320-349.

Epple, A., and Brinn, J. E. (1986). Pancreatic islets. in Vertebrate Endocrinology: Fundamentals and Biomedical Interpretations, Vol. 1, cc Morphological Considerations" (P. K. T. Pang and M- T. Schreïbman, Eds.), pp. 279-3 17. Texas Tech. Univ. Press, Lubbock.

Epple A., and Brinn, J. E. (1 987). " The comparative physiology of the pancreatic islets". Heidelberg: Springer-Verlag, pp 223.

Falkmer, S. (1985a). Comparative morphology of pancreatic islets in anunals; in Volk BW, ArquiUa ER (eds): The diabetic Pancreas, ed 2. New York, Plenum Press, pp 17-52.

Falkmer, S. (1985b). Phylogenetic aspects of the brain-gut axk, with special reference to islet hormones in invertebrates and lower vertebrates. In cNeurosecretion and the Biology of Neuropeptidesyy(H. Ko bayashi, Ed.), pp.3 17-325. Japan Sci. Soc. Press, Tokyo. [Springer-Verlag, Berlin].

Falkmer, S. (1995). Origin of the parenchymal cells of the endocrine pancreas: some phylogenetic and ontogenetic aspects. In "Endocrine tumors of the Pancreas,"Vol.23," Frontiers in Gastrointestid Research" (M. Mignon and R T. Jansen, Eds.), pp. 2-29. Karger, Basel.

Falkmer, S., and Patent, G. J. (1 972). Comparative and embryological aspects of the pancreatic islets. In "Haandbook of Physiology, Vol. 1, The Endocrine Pancreas" @.F. Steiner and N. Freinkel Eds.), pp. 1-23. Williams & Wh,Baltimore.

Fallcmer, S., El-Salhy, M., and Titlbach, M. (1984). Evolution of the neuroendocrine system in vertebrates. A review witd particular reference to the phylogeny and postnatal maturation of the islet parenchyma. In cTvolutionand Tumor Pathology of the Neuroendocrine Systemyy(S. Fakmer, R. Hakanson, and F. Sundler, Eds.), pp. 59-87. Elsevier, Amsterdam..

Gapp, D. A., and Polak, J. M. (1990). Localization of insulin to gastro-entero-pancreatic cek in the turtle gastrointestinal tract. Gen. Comp. Endocrinol. 78,4845- Garcia- Hemandez, M.P., and Agulleiro, B. (1992). Ontogeny of the endocrine pancreas in sea bas (Dicen frmchus labrax) . Cell Tissue Res. 270, 33 9-3 52.

Garcia-Hernandez, M. P. G., Lozano, M. T., and Agulleiro, B. (1994). Ontogeny of some endocrine celis of the digestive tract in sea bass (Dicen~archtrslabrax): An immiuiocytochemical study. Cell Tissue Res. 277, 373 -3 83.

Gardiner, B. G., Maisey, J. G., and Littlewood, D. T. J. (1996). Interrelationships of basal neopterygians. In 'Tnterrelationships of Fishes" (M. L. J. Stiassny, L. R Parenti, and G. D. Johnson, Eds.), pp. 117-146. Academic Press, San Diego.

Gerich, J. E. (1 983). Somatostatin and analogues. In M. Eilenberg and & Rifkïn (eds.), Diabetes mellitus: theory and practice, 3" ed., pp. 225-254. Medical examinationç, New York.

Gomez-Visus, I., Abad, M. P., Hernandez, G., and Agulleiro, B. (1 996). Occurrence of somatostatin and uisulin immunoreactivities in the stomach of sea bass (Dicentrarchus labrmr L.) : Light and electron microscopie studies. Gen. Comp. Endocrinol- 102, 16-27.

Gomez-Visus, I., Garcia- Hernandez, M.P., Lozano, M. T., and Agulleiro, B. (1998). Glucagon- and NPY-reiated peptide-immunoreactive cells in the gut of sea bass (Dicentrarchus labrmc L.): A light and electron microscopy study. Gen. Comp. Endocrinol. 112, 26-37.

Goodman, RH., Jacobs, J. W., Chin, W. W., Lund, P.K, Dee, P.C., and Habener, J.F. (1 980). Nucleotide sequence of a cloned structural gene coding for a precursor of pancreatic somatostatin, Proc. Natl. Acad. Sei. 77, 5869-5873.

Gorbman, A. (1 990). Cyclo stome models for speculation on evolution of vertebrate endocrine systems. J. Exp. Zool. Suppl. 4, 137- 141.

Greenwood, P.H. (1973). Interrelationships of osteoglossomorphs, In '?nterrelationships of f%hes9' (Greenwood, P.H., Miles, R.S., Patterson, C., eds.). p. 307-332. Acadernic Press, London.

Grober, M.S., Myers, T.R, Marchaterre, M.A., Bass, A.H., and Myer, D.A. (1995). Structure, localization, and molecular phylogeny of a GnRH cDNA fiom a paracanthopterygian fish, the plainfin midshipman (Porichthys notatus). Gen. Comp. Endocrinol. 99, 85-99.

Groff, K., and Youson, J. H. (1997). An immunohistochemical study of the endocrine cells withthe pancreas, intestine, and stomach of the gar (Lepisosteus osseus L.). Gen. Comp. Endocrinol. 106, 1- 16.

Groff, K.E. and Youson, J. H. (1998). Fine structure and immunocytochemistry of cells within the endocrine pancreas of the gar (Lepisosteus osseus). CmJ. Zool. 76,6-18.

Hahn, V., Wmkler, J., Rapopoa, T.A., Liebscher, D.H., Coutelle, C., and Rosenthal, S. (1983). Carp preproixisulin cDNA sequence and evolution of insulin genes. Nucleic Acids Res. 11 (13), 4541-4552. Herrera, P-L., Huarte, J., Sanvito, F., Meda, P., Orci, L., and VassaIli, J-D (1991). Embryogenesis of the murine endocrine pancreas: early expression of pancreatic polypeptide gene. Development 113, 1257-L 265.

Hobart, P., Crawford, R, Shen, L.P., Pictet, R, and Rutter, W.J. (1980a). Clonhg and sequence of cDNAs encoding two distinct somatostatin precursors found in the endocrine pancreas of anglerfish. Nature 288, 137-241.

Hobart, P., Shen, L.P., Crawford, R, Pictet, R, and Rutter, W.J. (198Ob). Comparison of the nucleic acid sequence of anglerfish and mammalian insulin mRNAs fiom cloned cDNAs. Science 210 (4476), 1360- 1363.

Holst, N., Jacobsen, M. B., Haug, E,, Tanab, T., and Abyholm, T. (1995). Somatostatin in physio Io gical concentrations inhibits basal and enhances luteinizing hormone-stimulat ed progesterone release fiom human granulose-luted cek. Human Reprod. 10, 1363-1366.

Inui, A, Oya, M., Inoue, T., Sakatina, N., Morioka, H., Shü, K., Yokomo, K., Mizuno, N., and Baba, S. (1988). Peptide YY receptors in the brain. Biochem. Biophys. Res. Common. 150925-32-

Irwi~~,DM., and Wong, J. (1995). Trout and chicken proglucagon: alternative splicing generates mRNA transcripts encodiog glucagon-like peptide 2. Mol. Endocrinol. 9,267-277.

Innn, D-M-, Huner, O., and Youson, J-H. (1 999). Lamprey progIucagon and the origin of glucagons-like peptides. Mol. Biol. Evol. 16 (1 l), 1548- 1557.

Irwin, D.M-, and Sivarajah, P. (2000). Proglucagon cDNAs fiom the leopard fiog, Ranapipiem, encode two GLP-1 -mepeptides. Mol. Cell. Endocrinol- 162, 17-24.

Johnson, D. E., Torrence, J. L., Elde, R P., Bauer, G. E., Noe, B. D., and Fletcher, D. J. (1 976). Tmmuno histo chemical 10 ca1i;lation of somatostat in, insulin and glucagon in the principal islets of the anglerfïsh (Lophius americanus) and the channel catfish (Ictalums punctata). Am. .L Anat. 147, 1 19- 124.

Jonsson, A. C. (1 99 1). Regulatory peptides in the pancreas of two species of elasmobranchs and in the Brockmann bodies of four teleost species. Cell Tissue Res. 266, 163- 172.

Kavsan, V. M., Petrenko, O. I., Rynditch, A.V., Dobrovolskaya, G.N., Sova, V. V., Prokopenko, 1. V., Yuriu, V. L., Demjanova, N. G., Zlochevskij, M. L., Koztov, Y. I., Naroditskaya, V. A., Skoblov, Y. S., Rebentish, B. A., Sorokin, A. V. and Debabov, V. G. (1983). Synthesis, cloning and sequence determination of S1kria.n salmon preproinsulin cDNA. Mol. Biol. (Moscow) 17 (l), 42-53.

King, G.L., and Kahn, C.R. (1981). Non-pardel evolution of metabolic and growth-promothg fiuictions of insulin. Nature 292,644-646. Kittilson, J. D., Moore, C. A., and Sheridan, M. A. (1999). Polygenic expression of sornatostath in rainbow trout, Oncorhynchus mykiss: Evidence of a preprosomatostatin encoding somatostatin-14. Gen. Comp. Endocrinol. 114, 88-96.

Klein, C., and Lange, R H. (1977). Principal cell types in the pancreatic idet of a teleost fish, Xiphophorus helleri K CeZZ Tissue Res. 176,529-55 1.

Klein, C., and Van Noorden, S. (1980). Pancreatic polypeptide (PP) and glucagon cells in the pancreatic islet of Xiphophorus helleri H. (Teleostei). Correlative immunohistochemistry and electron microscopy. CelZ Tissue Res. 205, 187- 198.

Kobayashi, K., and Takahaski, Y. (2 970). Light and electron microscopic observations on the Ïslets of Langerhans m Carassius carussius longsdoflii Arch. Histol. Jpn. 3 1,433-454.

Kobayashi, K., and Takahashi, Y. (1 974). Fine structure oflangerhans7&let cek in a marine teleost Congerjaponicus Bleeker. Gen. Comp. Endocrinol. 23, 1- 18.

Kobayashi, EL, Sbi'basaki, S., and Takahashi, Y. (1976). Light and electron microscopic study on the endocrine cells of the pancreas in a marine teleost Fugu rubripes mripes. CelZ Tissue Res. 174, 161-182.

Kristensen, C., Kjeldsen, T., W~ïberg,F.C., Schaffier, L., Hach, M., Havelund, S., Bass, J., Steiner, D.F., and Andersen, A.S. (1997). Alanine scanning mutagenesis of insulul. J. Biol. Chem. 272, 12978-12983.

Kumazawa, Y ., and Nishida, M. (2000). Molecular phylogeny of Osteoglossoids: A new mode1 for gondwanian origin and plate tectonic transportation of the Asian arowana. Mol. Biol. Evol. 17 (12), 1569-1878,

Labuthe, M., Chenut, B., Rouyer-Fessard, C., Tatemoto, K., Couvineau, A., Servin, A., and .dLmiranoff, B. (1986). Interaction of peptide YY with rat intestinal epithelial plasma membranes: bindiag of the radioiodulated peptide. Endocrinolo~118, 1910-1917.

Langer, M., Van Noorden, S., POU,J.M., and Pearse, AGE. (1979). Peptide hormone-iike immunoreactivity in the gastrointestinal tract and endocrine pancreas of eleven teleost species. Cell Tissue Res. 199, 493-508.

Larhammar, D. (1996). Evolution of neuropeptide Y, peptide YY and pancreatic polypeptide. ReguL Pept. 62, 1- 11.

Lauder, G.V., and Liem, ICF. (1983). The evolution and interrelationships of the actinopterygian hhes. Bulletin of the Museum of Comparative Zoology, Harvard University f 50,9S-l97.

Li, G. Q., and Wilson, M. V. H. (1996). Phylogeny of Osteoglossornorpha. In" Interrelationships

of Fishes " (M. L. Stiassny, L. R Parenti and G. D. Johnson), pp. 163- 174. Academic Press, London. Li, G.Q., and Wilson, M.V.H., Grande, L. (1997). Review of Eohiodon (Teleostei: Osteo glo ssomorpha) fkom western North America, with a p hylo genetic reassessment of Hiodontidae- J. Puleont. 71(6), 1109-1 124.

Lin, X., Otto, C.J., and Peter, RE. (1999). Expression of tbree distinct somatostatin messenger nibonucleic acids (mRNAs) in go1dfish brain: Characterization of the complementary deoxyriinucleic acids, distri'bution and seasod variation of the mRNAs, and action of a somatostatm-14 variant. Endocrinology 140,2089-2099.

L'Hermite, A., Ferrano, R,Dubois, M. P., and Andersen, A. C. (1985). Detection of endocrine cells by immunoflourecence method in the gastroenteropancreatic system of the adult eeI, glass eel, and leptocephalic larva (Anguilla anguilla L.). Gen. Comp. Endocrinol. 58, 347-359.

Lozano, M. T., and Agdeiro, B. (1986). Tmmunocytochemical and ultrastructural study of the endocrine pancreas of MugiZ auratus and Mu@ saliens L. (Teleostei). J. Submicrosc. Cytol. 18, 85-98.

Lozano, M.T., Garcia AyahA., Abad, M-E., and Agdeiro, B. (1 99 la). Pancreatic endocrine ceIls in sea bass (Dicentrarchus Zabrax L.). II. Tmmunocytochemical study of insulin and somatostatin peptides. Gen. Comp. Endocrinol. 81, 198-206.

Lozano, M.T., Garcia Ayala, A, Abad, M.E., and Agulleiro, B. (199 1b). Pancreatic endocrine cells in sea bass (Dicentrarchus labrm L.). 1. Tmmunocytochenical characterization of glucagon- and PP-related peptides. Gen. Comp. Endocrinol. 81, 187- 197.

Maake, C., Hanke, W., and Reinecke, M. (1998). An immunohistochemical and morphometric analysis of insulin, insulin-like growth factor 1, glucagon, somatostatin, and PP in the development of the gastro-entero-pancreatic system of Xenopus laevis. Gen. Comp. Endocrinol, 110, 182- 1%.

Magazin, M., Minth, C.D., Funckes, C.L., Deschenes, Tavianini, M.A., and Dixon, J.E. (1 982). Sequences of cDNA encoding pancreatic preprosomatostatin-22. Pro. Natl. Acad Sci. 79,5 152-5156.

Maglio, M., and Putti, R (1998). MorphoIogical basis of the interaction between endocrine ceil types in the endocrine cell types in the pancreatic islets of the teleost, Blennius gattomggrgrne.Tissue CeZl 30, 672-683.

Mansour, M., Wright, J. R, and Pohajdak, B. (1 998). Cloning, sequencing and characterization of the tilapia insulin gene. Comp. Biochem. Physiol. 121 (3), 29 1-297.

McConnick, N. A. (1925). The distribution and structure of the ishds of Langerhans in certain fiesh water and marine fïshes. Trans. Roy. Can. Inst. 15, 57-8 1. 129 McDonald, JK, Greiner, F., Bauer, G.E., Elde, RP.,and Noe B.D. (198'7)- Separate cell types that express two different forms of somatostatin in anglerfish islets can be irnmunohistochemicaly Herentiated. J. Histochem. Cytochem., 35, 155-162.

Milewski, W. M., Duguay, S. J., Chan, S. J., and Steiner, D. F. (1998). Conservation of PDX-1 structure, function, and expression in zebrafish. Endocrinology 139 (3), 1440-1449.

Millonig, G. (1961). Advantages of a phosphate bufEer for 0s04 solutions in fixation. J Appl. Phys. 32, Vol. 16, 1637. (Abstr.).

Minth, CD., Taylor, W.L., Magazin, M., Tavia- M.M.A., Collier, Ki, Weith, H.L., and Dixon, J.E. (1982). The structure of cloned DNA complementary to cat6ish pancreatic somatostatin-14 messenger RNA, J. Biol. Chem. 257, 10372- 10377.

Mommsen, T. P. and Phetskaya E. M. (199 1). Insulin in flshes and agnathans: bjstory, structure and metabolic reguiatioa Rev. Aquat Sci 4,225-259.

Moore, CA, Kittilson, J.D., Dahl, S.K-, and Sheridaa, M.A. (1995). Isolation and characterization of a cDNA encoding for preprosomatostatin containing [~yr',~l~"] -somatostatin-14 f?om the endocrine pancreas of rainbow trout, (Oncorhynchus mykiss). Gen. Cornp. Endocrinol. 98,243-26 1.

Moore, C.A., Kittilson, J.D., Ebnnan, M.M., and Sheridan, M.A. (1999). Rainbow trout, Oncorhynchus mykiss possess two somatostatin mRN& that are differentialiy expressed. Am. J. Physiol. 246, R1553-R1561.

Nakamura, M., and Yokote, M- (1 97 1). Ultrastructural studies on the islets of Langerhans of the carp. 2. Anat. EntwickL -Gesch. 134,61 -72.

Nelson, G.J. (1973). Notes on the structure and relationships of certain Cretaceous and Ecocene teleostean fishes. American Museum Novitates 2524, 3 1P.

Nelson, J. S. (1994). Tishes of the World", 3* edition. Wdey, New York.

Nguyen, T., Mommsen, T.P., Mirns, S.D., and Conlon, J.M. (1994). Characterization of insulins and pro-glucagon-derived peptides fiom a p hylogenetically ancient fis4 PoZyodon spathula. Biochern. J. 300,33 9-345.

Nishi, M., Moverus, B., Bukovskaya, O. Takahashi, A., and Kawauchi, H. (1995). Isolation and characterization of [pro2]-somatostatin-14 and melanotropins from Russian sturgeon, Acipenser gueldenstaedti. Gen. Comp. Endocrinol. 99, 6-12.

Norris, D. 0. (1997). Vertebrate Endoc~ology3d edition. Acadernic Press, New York.

Nozaki, M., mata, K, Oota, Y., Gorbman, A, and Plisetskaya, E. M. (1988a). DifEerent cellular distri'butions of two somatostatins in brain and pancreas of salmonids, and their associations with msuün- and glucagon-secreting cells. Gen. Cornp. Endocrinol. 69, 267-280. No* M., mata, K., Oota, Y., Gorbman, A., and Plisetskaya, E. M. (1988b). Colocalization of glucagons-like peptide and glucagons immunoreactivities in pancreatic islets and intestine of salmonids. CelZ Tissue Res. 253,372-375.

O'Neill, D.F., Powell, J.F.F., Standen, E.M., Youson, J.H., Warby, CM., Sherwood, N.M. (1998). Gondotropin-releasing homone (GnRH) in Ancient teleo sts, the bo nyto ngue fshes: putative ongin of salmon GnRH. Gen. Cornp. Endocrïnol. 112,415-425.

Ohlsen, S. M-, Lugenbeel KA., and Wong, E.A. (1994). Characterization of the linked ovine insulin and insulin-like growth factor-II genes. DNA CelZ BioL 13 (4), 377-3 88.

Parrizas, M., Plauas, J., Plisetskaya, E. M., and Gutierrez, J. (1 994). Insuiin binding and receptor tyrosine kinase activity in skeletal muscle of carnivorous and omnivorous fish, Am. J. Physiol. 266, R l9U-Rl9SO.

Patel, Y.C. (1992). General aspects of the biology and fùnction of somatostatin. In C. Weil, E. E. Muller, M.O. Thonier (eds.), Basic and clinical aspects of neuroscience, Vol. 4, pp. 1 - 16. Springer-Verlag, Berlin.

Patent, G. J., Kechele, P. O., and Tomichek Carranu, V. (1978). Nonconventional innervation of the pancreatic islets of the teIeost &h, GiZZichrhys mirabdis. CeZZ Tissue Res. 191, 305-3 15.

Patterson, C., and Rosen, D. E. (1977). Review of ichthyodectiform and other Mesozoic teleost fshes and the theory and practice of classiQmg fossils. Bull- Am. Mus. Nat. hTistt 158, 81-172.

Perler, F., Efstratiadis, A., Lomedico, P., Gilbert, W., KoIodner, R, and Dodgson, J. (1980). The evolution of genes: the chicken preproinsulin gene. Ce11 20 (2), 555-566.

Peterson, J.D., Coulter, C.I., Steiner, D.F., Edmin, S.O., and Falkmer, S. (1 974). Structural and crystdographic observations on hagfish insulin. Nature Lond. 251,239-240.

Planas, J.V., Mendez, E., Banos, N., Capilla, E., Castilio, J., Navarro, I., and Gutierrez, J. (2000). Fish insulin, IGF-1, and IGF-II receptors: A phylogenetic approach. Amer. Zool., 40,223-233.

Plisetskaya, E.M. (1990a). Recent studies of fish pancreatic hormones:selected topics. Zool. Sci- 7,335-353.

Plisetskaya, E.M. (199Ob). Endocrine pancreas of teleost hh: A mode1 for interaction of islet hormones. J. Exp. Zool. S@. 4,53-57.

Plisetskaya, E. M., Pollock, H. G., Rouse, J. B., Hamilton, J. W-, KimmeI, J. R., Andrews, P. C., and Gorbman, A. (1986). Characterization of coho salmon (Oncorhynchus kisutch) islet somatostatins. Gen. Comp. Endocrinol. 63,252-263. 131 Plisetskaya, E.M., Pollock, H.G., Elliott, W.M., Youson, J.H., and Andrews, P.C. (1988). Isolation and structure of lamprey (Petromyzon Marinus) insullli. Gen Cornp. Endocrinol. 69 (l), 46-55.

Plisetskaya, E.M, and Mommsen, T. P. (1996). Glucagon and glucagons-like peptides in fishes. Int. Rev. Cyfol. 168, 187-257.

Pollock, H.G., Kirnmel J.R., Hamilton, J.W., Rouse, J.B., Ebner, K.E., Lance, V., and Rawitch, A.B. (1987). Isolation and structures of alligator gar (Lepisosteus spatula) insului and pancreatic polypeptide. Gen. Comp. Endocrinal. 67,375-3 82.

Powers, D. A. (1989). Fish as mode1 systerns. Science 246, 352-357.

Putti, R., Varano, L., Cavagnuolo, A., and Laforgia, V. (199 1). The endocrine pancreas of Podarcis S. sida RAF: an immunocytochemical study at light and electron microscopic levels. Eur. J. Basic Appl. Histochem. 35, 145- 159.

Raynor, K., and Reisine, T. (1992). Somatostatin receptors. Neurobiology 16,273-289.

Reichllli, S. (1983). Sornatostatin. In D. T. Krieger, N. Brownstein, and J. B. Martin (eds.), Brain peptide, pp. 712-752. John Wdey and Sons, New York.

Reinecke, M. (198 1). Tmm~110histochemicallocalization of polypeptide hormones in endocrine cells of the digestive tract of Branchios~alanceolatum. CeZZ Tissue Res. 219, 445-456.

Reinecke, M., Muller, C., and Segner, H. (1 997). An immunohistochemical analysis of the ontogeny, distriiution and coexistence of 12 regdatory peptides and serotonin in endocrine ceils and nerve fibers of the digestive tract of the turbot, ScophthaZmus maximus (Teleostei). Anat. Embryol. 195, 87- 102.

Reisine, T., and Bell, G. 1. (1995). Molecular biology of somatostath receptors. Endocr. Rev. 16, 427-442.

Roca, B ., Arilla, E., and Preito, J-C. (1986). Evidence for somatostatin binding sites in rabbit kidney. ReguZ. Pept. 13,273-28 1.

Rombout, J. H. W. H., Radernakers, L. HI. P. M., and van Hess, J. P. (1979). Pancreatic endocrine cells of Barbus conchonius (Teleostei, Cyprinidae). CelZ Tissue Res. 203, 9-23.

Rombout, J. H. W. H., and Taverne-Thiele, J. J. (1982). An immunocytochemical and electron-microscopical study of endocrine cek in the gut and pancreas of stomachless teleo st fis4 Barbus conchonius (Cyprlnidae). CeZZ Tissue Res. 227, 5 77-593.

Rose, TM. Schultz, ER,Henikoe J.G., Pietrokovski, S.,McCailum, CM., Henikoe S. (1998). Consensus-degenerate hybrid O ligonucleotide prirners for amplification of distantly related sequences. Nucleic Acids Res. 26(7), 1628- 163 5. Rubin, DA, and Dores, RM. (1994). CIoning of a growth hormone fkom a primitive bony &h and its phylogenetic relationships. Gen. Comp. Endocrinol. 95, 7 1-83.

Rubin, D.A., and Dores, RM. (1995). Obtaining a more resolute teleost growth hormone phylogeny by the introduction of gaps in sequence aiigmnent. Mol. Phylogenet. Evol. 4, 129-138.

Sheridan, M.A. (1994). Regulation of lipid rnetabolism in poikilotherrnic vertebrates. Comp. Biochern. Physiol. 107b, 495-508.

Sheridan, M-A,, Kittilson, J.D., Ehnnan, M.M-, and Moore, C.A. (1997). Polygenic expression of somatostatin in rainbow trout. In S. Kawashima and S. Kikuyama (eds.), Advances in comparative endocrinology, pp, 291-294. Moduzzi Editore, Bologna.

Sheridan, M.A. ,Kittilson, ID., and Slagte, B. J. (2000). Structure-hction relationships of the signahg system for the somatostatin peptide hormone family- Amer. Zool. 40,269-286-

Shulinder, A.R, Bennett, C-,Robinson, E.A., and Roth, J. (1989). Isolation and characterization of two different insuIins fiom an amphibian, Xenopus laevis. Endocrinologv 125, 469-477.

Stefan, Y., and Falluner, S. (1 980). Identification of four endocrine cell types in the pancreas of Cortus scorpius (Teleosteî) by immunofluorescence and electron microscopy. Gen. Comp. Endocrinol. 42, 17 1- 178.

Swo fford, D.L. (1 999). PAUP* .Phylogenetic Analysis Using Parsimony (* and Other Methods). Sinauer Associates, Sunderland, MA Version 4.0b2.

Tatemoto, K. (1982a). Isolation and characterization of peptide YY(PYY), a candidate gut hormone that inhibits pancreatic exocrine secretion. Proc. Natl. Acad. Sei. USA 79, 2514-2518.

Tatemoto, K. (1982b). Neuropeptide Y: the cornplete sequence of the brain peptide. Proc. NdAcad Sci. USA 79,5485-5489.

Tatemoto, K., Carlquist, M., and Mutt, V. (1982). Neuropeptide Y: a novel brain peptide with structural similarities to peptide YY and pancreatic polypeptide. Nature (Sondon) 285, 417-418.

Taverne, L. (1986). L'evolution de l'antorbitaire et son incidence sur la phylogenie des teleosteens prjmitifi. Biol-Jb. Dodonaea 54: 142-160.

Tostivint, H., Lihrman, I., Bucharles, C., Didier, V., Coulouarn, Y., Fournier, A., Codon, J.M., and Vaudry, H. (1 996). Occurance of two somatostatin variants in the f?og brain: Characterization of the cDNAs, distniution of the mEWAs, and receptor-binding atkities of the peptides. Pro. Natl. Acad Sei. 93, 12605-126 10. 133 Trabucchi, M., Tostivint, H., Lihrmann, I., Jegou, S., Vallaniio, M., and Vaudry, H. (1999). MolecuIar cloning of the cDNAs and distribution of the mRNAs encoding two somatostatin precursors in the ecanlungfkh Protopterus annectem. J. Comp. Neurology 410,643-652,

Uesaka, T., Yano, K, Yamasaki, M., and Ando, M. (1995). Somatostatin vasoactive intestinal peptide and granului-like peptides iso lated fiom intestinal tracts of go Idf%h (Carassius auratus). Gen. Comp. Endocrinul. 99,298-306.

Van Le, H.L., Lecointre, G., and Perasso, R (1993). A 28s rRNA-based phylogeny of the Gnathostomes: füst steps in the analysis of conflict and congruence with morphologïcdy based cladograms. Molecular Phylogeny and Evolution. 2(1), 3 1 -5 1.

Van Noorden, S. (1984). The neuroendocrine system in Protostomian and deuterostomian invertebrates and lower vertebrates. lii 'Tivolution and Tumour Pathology of the Neuroendocrine system"(S. Famer, R Hakamon, and F. Sundler, Eds.), pp. 8-3 8. Elsevier, Amsterdam.

Van Noorden, S. (1990). Gut hormones in cyclostomes. Fish Physiol. Biochem. 5, 399-408.

Van Noorden, S., and Patent, G. J. (1 978). Localization of pancreatic polypeptide (PP)-iike irnmunoreactivity in the pancreatic islets of some teleost fishes. CeZZ Tissue Res. 188, 521-525.

Vaudry, H.,Chatrel N., and Codon, LM. (1992). Isolation of [pro2, et L3]sornatostatin-14 and sornatostatin-14 from the f?og brai.reveals the existence of a somatostatin gene family in a tetrapod. Biochem. Biophys. Res. Commun. 188,477-482.

Wagner, G. F., Mc Keown, B. A., and Popham, D. J. (1981). The autoradiographic locaiization of zinc within the pancreatic islets of the rainbow trout, Saho gairdneri. Nislochernistry. 72, 113-121.

Wang, Y-Q., Plisetskaya, E. M., Baskin, D. G., and Gorbman, A. (1986). Tmmunocytochemical study of the pancreatic islets of the Pacific salmon, Oncorhynchus kisutch. Zool. Sci. 3, 123-129.

Wang, Y., Youson, J.H., and Codon, J.M. (1993). Prosomatostatin-1 is processed to somatostatin-26 and somatostatin- 14 in the pancreas of the boaAmia calva. Regul. Peptides 47,33-40.

Wetekam, W., Groneberg, J., Leineweber, M., Wengenmayer, F., and Wmcker, E. L. (1982). The nucleotide sequence of cDNA coding for preproinsulin fiom the primate Macaca fascicularis. Gene 19, 179- 183.

Wilson, R P. (1994). Utilization of dietary carbohydrate by fkh. Aquaculture 124, 67-80. Wright, J.R, Bonen, A., Codon, M.J., and Pohajdak, B. (2000). Glucose homeostasis in the teleost fkh tilapia: Znsights fiom Brolanann body xenotransplanting studies. Amer. 2001, 40,234-245,

Yang, H., and Wright, J. R?Jr. (1995)- A method for mass harvesting islets (Brockrnann bodies) fiom teleost fïsh Ce22 Transpl. 4,62 1-628.

Youson, J. H. (2000). The agnatthan enteropancreatic endocrine system: Phylogenetic and ontogenetic histories, structure, and fiinction. Amer- Zool. 40, 179- 199.

Youson, J. H., and Cheung, R. (1990). Morphogenesis of somatostatin-and -insulin-secreting cells in the lamprey endocrine pancreas. Fish Physiol. Biochern. 8,389-397-

Youson, J. H., and Po tter, 1. C. (1993). An imm~flohistochemicalstudy of enteropancreatic endocrine cells in larvae and juveniles of the southern hernisphere lampreys Geotria austraZis and Mordacia rnordux. Gen. Comp. Endocrinol. 92, 15 1-167.

Youson, J. H., and AL-Mahrouki, A. A. (1999). Ontogenetic and phylogenetic development of the endocrine pancreas (islet organ) in fishes. Gen. Comp. Endocrinol. 116,303-33 5.

Youson, J. H., Al-Mahrouki, A. A., Naumovski, D., and Conlon, M. J. (2001). The endocrine cek in the gastro-entero-pancreatic endocrine system of the boaAmia calva L.: An immunohistochemical, ultrastructural and immunocytochemical analysis. J. Moph. (in press).

Yui, R, Nagata, Y., and Fujita, T. (1988). hunohistochemical studies on the &let and the gut of the Arctic iamprey, Lampeîra japonica . Arch. Histol. Cytol. 5 1, 1 09- 1 19.

Zhang, J.Y. (1 99 8). Morphology and phylogenetic relationships of dagger Kuntulunia (Teleostei: Osteoglossomorpha). J. Vert. Paleon. 18(2), 280-300. Appendix 1

A section of a pancreatic islet and exocrine acini in A. mpestns showing the absence of immmoreacti~in the islet tissue (arro w) following the substhtion of the primary antibody wah phosphate-bufEered saline (PB S). X560.

Appendix 2

Chromatogram of the preproinsulin original sequence fiom the arawana. Show the sequence ofthe S'-RACE cDNA in the reverse direction and is complement to the preproinsulin sequence. The initiation codon of the signal peptide is underlined with a single line, and the N-terminal at the beginning of the B-chah is underlined with double line (refer to Figs. 28A, 29 and 30). Insulin fkom arawana varied f?om those of the other osteoglossomorphs, where, position 1 at the N-terminai of the B-chain in arawana has serine replaces alanine.

Appendix 3

Detailed preprosomatostatin sequences using manual sequencing. The nucleotide sequence is identified at the top of each Iane (Guanosine, G; Adenosine, A; Thymidine, T;

Cytidine, C).

A) From the elephantnose. The numbers refer to the position of the nucleotides in the

cDNA (refer to Fig. 34C). The nucleotide sequence encoding (~yr~,~1~'') SST-14

(variant SST-14) is present within the indicated bracket.

B) From the white sucker. The numbers refer to the position of the nucteotides in the

cDNA (refer to Fig. 34E). The nucleotide sequence encoding the

non-variant S ST- 14 is present within the indicated bracket.

Glossary

Acinar cells: pancreatic ce& that secret the digestive enzymes (zymogen granules).

Actinopterygii: rayhedfishes a major class of bony fishes.

An tisera: plural of antisem that is blood senun that contains antihdies to a specific antigen.

Basal and generalized: reflect close to the starting place of a character or a group (Youson and Al-Mahrouki, 1999).

Bootstrap: a technique has been frequently used as a means to estimate the confidence level of phylogenetic hypotheses.

Brockmann body: a large accumulation of islet tissue closely associated with exocrine pancreatic tissue; the Brockmann body can contain one or more islets.

Derived: refer to a character or a group that has been taken from another (earlier) source (Youson and Al-Mahrouki, 1999).

Endocrine: refers to hormonal secretion produced in cek and generaiiy passed into the blood.

Exocrine: extemal secretion into a duct (e.g. secretion of digestive enzymes into the lumina of the pancreatic ducts).

Imm uno histochemistry : a histochemical technique using antisera to detect specific antigens.

Islet organ: the entire endocrine pancreatic tissue within an organism.

Gastroenteropancreatic (GEP): a system of endocrine cells involving the stomach, mtestine, and pancreas.

Maximum parsimony: where character states (e.g., the nucleotide or amino acid at a site) are used, and the shortest pathway leading to these character states is chosen as the best tree.

Monop hyletic: a group of taxa is said to be monophyletic if they are derived fkom a single common ancestor.

Ontogeny: is development of a single individii;iL or a system within the individual, fiom fertikation to death (Youson and Al-Mahrodci, 1999). Phylogenetic tree: the evolutionary reiationships among a group of organisms are illustrated by means of a phylogenetic tree. 1t is a graph composed of nodes and branches, in which only one branch connects any two adjacent nodes. The nodes represent the taxonomie units, and the branches dekethe relationships among the units in terms of descent and ancestry-

Poiymerase chain reaction (PCR):a technique that enables the production of enormous numbers of copies of a specifïed DNA sequence.

PoIypeptide: a molecule made of many amino acids joined by peptide bonds.

Primer: a short nucleotide sequence that is used in PCR reactions or sequencing reactions to initiate the synthesis of DNA.

Principal islet: a large body of pure islet tissue devoid oÇ or nearly devoid of, strands of exocrine pancreas.

Phylogeny: is developrnent that involves modification of a species or a group of species, Le., the evolutionary history of a lineage (Youson and Al-Mahrouki, 1999). Generar and Comparative Endocrinology 110. 125-1 39 (1998) Article No. GC987070

~mmunohistochemicalStudies of the Endocrine Cells within the Gastro-entero-pancreaticSystem of Osteoglossomorpha, an Ancient Teleostean Group

Azza A. Al-Mahrouki and John H. Youson' Deparunent ofZoology and Division of Life Sciences. University of Tomnro at Scarborough. Scarbomugh. Onrario MI C 1A.I. Canada

Accepted February 4. 1998

The identification and distribution of endocrine cells the PP farnily antisera in al1 five species. However. within the gastro-entero-pancreatic (GEP) system of five immunoreactivities to GLU, GLP. SST. and m-INS anti- species of the Osteoglossomorpha (Osteoglossum bicimho- sera were variable in intestinal cells of the species- sum. Sclempages jardini, Pantodon buchholzi, Notopterus Immunoreactivity with sera raised against m-INS and chitala and Gnathonemus petersii) were analyzed by PYY was also observed in the stomach of f? buchholzi, irnmunohistochemistry. Four immunoreactive ce11 types The significance of these fmdings is discussed in both were identifïed within the pancreatic islets (A, B. D. and ontogenetic and phylogenetic contexts with respect to the F cells). using antisera directed against mammalian GEP system in actinopterygian fshes and with respect to the insulin (m-INS) , somatostatins (SST-14.SST-25). and possibility of variable processing of prohorrnones in the members of the pancreatic polypeptide (aPY. NPY' PYY) ciiffint organs of these osteoglossomorphs. 1998 hdcrnic P- and glucagon (GLU, GLP) families. The B cells were Key Words: gastro-entero-pancreaticsystem; Gnathone- located throughout the center of the islets in the five mus petersii; immuno histochemis try; No top terus chitala; species and, in general, D cells had a similar distribution. osteoglossomorphs; Osteoglossum bicirrhosum; Pantodon However, immunoreactivity to anti-somatostatins varied buccholzi; Scleropagesjardini. between four of the species and G- petersii. which showed less intensely stained D cells in the islets. but OsteogIossnrnonha are considereci to be one cf the greater SST immunoreactivity in both the intestinal ana most ancient superorders of the teleosts (Patterson and the stomach epithelia than in comparable epithelia of Rosen, 1977). the most abundant of the actinopterygian other species. For peptides of both the pancreatic poly- fishes. Osteoglossornorphs are interesting to both ich- peptide and the glucagon families, the immunoreactivity thyologists and paleontologists because of their was detected at the periphery of the islets, and there was anatomy. physiology. geographic distribution. and an- a suggestion of an interfamily colocalization of peptides cient fossil record. Exclusively freshwater fishes. at in some ceus. In addition, glucagon family peptides least in the modern fauna, they display interesting showed a scattered immunoreactivity throughout the biogeographic distributions including apparent ex- central portion of the islets. A moderately abundant nurnber of cells in the intestine were immunoreactive to amples of endemism of extant suprageneric taxa (Hi- odontidae in North America, and Mormyroidae and Pantodontidae in Africa), as well as circumtropical or To whom reprint requests should be addressed. Fax: (416)287-7642. old-world tropical distributions (Osteoglossidae and E-mail: youson@~.utoronto.ca;or almahrouki@~car.utoronto.ca Notopteroidae), Potentially, osteoglossomorphs repre-

0016-6480/98 $25.00 Copyright 1998 by Acadcrnic Press AI1 rlghu of reproduction in any rom nscrvcd. GenecaI and Comparative Endocrinology 116.409-42 1 (1999) 03 Article ID gcen. 1999.7380. available online at httpi//www.idealibrary.comon 10 E

Ultrastructure and Immunocytochemistry of the Islet Organ of Osteoglossomorpha (Teleostei)

Azza A. Al-Mahrouki and John H. Yousonl Depanment of Zoology and Division of Life Sciences. University of Tbmnto at Scarborough. Scarborough. Ontario MlC lA4. Canada

Accepted September 7. 1999

Both routine electron microscopy and immunocytochem- des of different shape. The cells of the islet organs of istry with protein A-gold were used to identie the ce11 these osteoglossomorphs are more similar to those in types within the islet organs of four species of teleosts more derived teleosts than they are to those of nonteleost (Osteoglossum bicirrhosum. Pantodon buchholzi, Notop- actinopterygians. c 1999hdcrnic P- terus chitala, and Gnathonemus petersii) within Osteoglos- Key Words: Gnathcnemus petersii, islet organ; irnmuno- somorpha, a subdivision with an ancient lineage. Four cytochemistry; Notopterus chitala; osteoglossomorphs; primary endocrule ce11 types. A, B. D. and E were Osteoglossum bicirrhosum; Pantodon buccholzï; ultrastruc- identified within the islets of the four species examined, ture. The B- and D-cells were located mainly in the central core of the islet in the four species. In general, the A-cells were located at the islet periphery in ail of the four The ancient origin of members of the subdivision species but in P. buchholu' and N. chitala they were also (superorder) Osteoglossomorpha among the teleosts differently distributed toward the islet core- F-cells were (Patterson and Rosen, 1977). their pivota1 position in present only at the islet periphery Granules of B-cells in teleost evolution. and the unsettled debate over their three species had a relatively homogeneous shape of the taxonomy make osteoglossomorphs an interesting case matrix core. but in 0,bicirrhosum, the shape varied for investigating the phylogenetic development of greatly. Variation in matrix shape of B-ce11 granules may organ systerns. Included among these systems is the indicate a different conformation of insulin molecules endocrine pancreas (islet organ) and its peptides. among at least some species of osteoglossomorphs, and Until recentiy, only passing attention had been made this observation may have some taxonornic significance. to the pancreas of members of the order Osteoglossifor- Two sornatostatin-containing (SST) D-ce11 types (Dl and mes (McConnick. 1925; Epple and Brinn. 1975: Langer DX) with granules of different shape were observed in al1 et al.. 1979). However, a detailed immunohistochemical four species of osteoglossomorphs. The granules of the study of the gastro-entero-pancreatic (GEP) system two D-cells immunostained either with anti-SST-25 and (Al-Mahrouki and Youson. 1998) reported that various anti-SST-14 @ 1-cells) or with anti-SST-34 @X-cells) . species of osteoglossomorphs showed structural diver- ~mmunocytochemistryconfirmed that A-cells, contain- sity which reflected their distinctive taxonomic group- ing glucagon-family peptides, and F-ceus, containing ings. This study also suggested that further analysis of peptides of the pancreatic polypeptide family. had gran- the endocrine ce11 types of the islet tissue and their immunoreactivity might provide additional data use- 'To whom reprint requests should be addressed. Fax: (416) ful for taxonornic comparisons. 287-7676. E-mail: [email protected]. The different ce11 types in the islet organs of teleosts

0016-6480/99 530.00 Copyright O 1999 by Acadcrnic Prcss Al1 rlghts of rcpmduction in any rom nscrved. Ceneral and Comparative Endocrinology 116.303-335 (1999) Artide ID gcen. L999.?376. available online ar h~p~/~~ww.idealibrary.comon IO E kE

REVIEW Ontogenetic and Phylogenetic Development of the Endocrine Pancreas (Islet Organ) in Fishes

John H. Youson and Azza A. Al-Mahrouki Department of Zoology and Division of Life Sciences. University of Toronco at Scarborough, Scarborough. Ontario Ml C 1A4 Canada

Accepted August 25. 1999

The morphology of the gastroenteropancreatic (CEP) relatively small isIets in the generalized euteleosts and system of fishes was reviewed witk the objective of the tendency for the concentration into Brockrnann providing the phylogenetic and ontogenetic development bodies of large (principal) islets (with or without second- of the system in this vertebrate group, which includes ary islets) in the more derived forrns. The holostean agnathans and gnathostome cartilaginous, actinoptyeryg- actinopterygians (Amiiformes and Semiontiformes) share iai. and sarcopterygian fishes- Particular emphasis is with the basal teleosts (osteoglossomorphs, elopo- placed on the fish homolog of the endocrine pancreas of morphs) the diffuse arrangement of the components of other vertebrates, which is referred to as the islet organ. the islet organ that is seen in generalized euteleosts, The one-hormone islet organ (B cells) of larval lampreys Since principal islets are also present in adult larnpreys is the most basic pattern seen among a free-living the question anses whether principal islets are a derived vertebrate, with the two-hormone islet organ (B and D or a generalized feature arnong teleosts. There is a cells) of hagfkh and the three-hormone islet organ (B. D. paucity of studies on the ontogeny of the GEP system in and F cells) of adult lampreys implying a phylogenetic fish but it has been noted that the timing of the trend toward the classic four-hormone islet tissue (B. D, appearance of the islet ce11 types parallels the tirne that E and A cells) in most other fishes. An earlier stage in the they appear during phylogeny; the theory of recapitula- development of this phylogenetic sequence in verte- tion has been revisiced- it is stressed that the lamprey life brates may have been the restriction of islet-type hor- cycle provides a good opportunity for studying the mones to the aiimentary canal. Iike that seen in protochor- development of the GEP system. There are now several dates- The relationship of the islet organ to exocrine markers of ce11 differentiation in the mammalian endo- pancreatic tissue, or its equivalent, is variable among crine pancreas which would be useful for investigating bony, cartilaginous, and agnathan fishes and is likely a the development of the islet organ and cells of the manifestation of the early divergence of these piscine remaining GEP system in fish. o 1999 Acadernic P- groups. Variations in pancreatic morphology between individuals of subgroups within both the Imprey and Scientific interest in the distribution. structure, and chondrichthyan taxa are consistent with their evolution- function of endocrine tissues associated with the aIi- ary distance. A cornparison of the distribution and mentary canal (gastroenteropancreatic, CEP. system) degree of concentration of the components of the islet of fishes can be traced back at Ieast to the beginning of organ arnong teleosts indicates a diffuse distribution of the 19th century. This interest has heightened as we

00 I6-6480/99S 30.00 Copyright o 1999 by Acadcrnic Pnu AI1 rights o~rcpmductionin any rom rcservcd.