The Distribution of Mitochondria-Rich Cells in the Gills of Air-Breathing Fishes

Transcription

1 215 The Distribution of Mitochondria-Rich Cells in the Gills of Air-Breathing Fishes Hui-Chen Lin* Wen-Ting Sung Department of Biology, Tunghai University, Taichung 407, Taiwan, Republic of China Accepted 10/21/02 ABSTRACT Respiration and ion regulation are the two principal functions of teleostean gills. Mainly found in the gill filaments of fish, mitochondria-rich cells (MRCs) proliferate to increase the ionoregulatory capacity of the gill in response to osmotic challenges. Gill lamellae consist mostly of pavement cells, which are the major site of gas exchange. Although lamellar MRCs have been reported in some fish species, there has been little discussion of which fish species are likely to have lamellar MRCs. In this study, we first compared the number of filament and lamellar MRCs in air-breathing and non-air-breathing fish species acclimated to freshwater and 5 g NaCl L 1 conditions. An increase in filament MRCs was found in both air-breathing and non-air-breathing fish acclimated to freshwater. Lamellar MRCs were found only in air-breathing species, but the number of lamellar MRCs did not change significantly with water conditions, except in Periophthalmus cantonensis. Next, we surveyed the distribution of MRCs in the gills of 66 fish species (including 29 species from the previous literature) from 12 orders, 28 families, and 56 genera. Our hypothesis that lamellar MRCs are more likely to be found in air-breathing fishes was supported by a significant association between the presence of lamellar MRCs and the mode of breathing at three levels of systematic categories (species, genus, and family). Based on this integrative view of the multiple functions of fish gills, we should reexamine the role of MRCs in freshwater fish. Introduction Fish gills are organs with multiple functions, including gas exchange, ion regulation, and acid-base balance (Laurent and * Corresponding author; hclin@mail.thu.edu.tw. Physiological and Biochemical Zoology 76(2): by The University of Chicago. All rights reserved /2003/ $15.00 Perry 1991; Jurss and Bastrop 1995; Marshall and Bryson 1998; Randall and Brauner 1998; Evans et al. 1999). The four pairs of branchial arches in the teleost consist of filaments and lamellae, both covered with epithelial cells, which can be further characterized into pavement cells, mitochondria-rich cells (MRCs, formerly chloride cells), mucous cells, and undifferentiated cells (Laurent 1984; Perry 1998; Evans et al. 1999). Pavement cells account for more than 90% of the gill surface (mostly in the lamellae) and are considered to be the site of gas exchange (Laurent and Dunel 1980; Perry 1998), although recent evidence suggests that pavement cells may also play an important role in ion and acid-base regulation (Perry 1998; Evans et al. 1999). MRCs are mainly found in the gill filaments of teleosts. In general, MRCs are believed to be the site of ion extrusion by fish in seawater and ion uptake by fish in freshwater (Laurent and Perry 1991; Perry 1998; Evans et al. 1999; Piermarini and Evans 2000). Nevertheless, several studies have reported the presence of lamellar MRCs in various teleost species, and lamellar MRCs are inducible in certain species acclimated in fresh or deionized water. For example, proliferation of lamellar MRCs was found in Anguilla anguilla after 1 wk acclimation in deionized water (Laurent and Dunel 1980) and also was found in the freshwateracclimated Atlantic stingray (Dasyatis sabina; Piermarini and Evans 2000). Most of the conclusions on the role of lamellar MRCs have been drawn on a species-specific basis, and thus far there has been no systematic investigation of their functions. Because of their rather large and oval shape, lamellar MRCs increase diffusion distance and therefore impede gas exchange (Greco et al. 1995; Perry 1998; Piermarini and Evans 2000). If a fish can exchange gases from other parts of the body, the gills may play a more important role in alternative physiological functions such as osmoregulation and acid-base balance. This directed our attention to air-breathing fishes, which have the ability to exchange gases directly with the aerial environment (Graham 1997). Depending on their behavioral and habitat conditions, air-breathing fishes are categorized into two modes, amphibious and aquatic (Graham 1997). The air-breathing organs include the skin, the lungs and respiratory gas bladders of the more primitive teleosts, and structures derived from buccal, pharyngeal, and branchial cavities and also from different parts of digestive tracts of the more advanced teleosts (Graham 1997). Since air breathing has evolved independently many times among the fishes (Graham 1997) and this trait can be found among species from various habitats, it is necessary to consider

2 216 H.-C. Lin and W.-T. Sung the divisions of fish according to their life history and habitats. Freshwater (FW) fishes are generally classified into primary, secondary, and peripheral FW fishes (Myers 1938 [cited in Helfman et al. 1997]). While primary FW fishes spend their entire life cycles in FW, secondary FW fishes are generally restricted to FW but may occasionally enter saltwater (Helfman et al. 1997). Peripheral FW fish species have taken up more or less permanent residence in FW or spend part of their life cycle in FW and another part in marine habitats (Helfman et al. 1997). In this study, we conducted two independent experiments to elucidate whether the likelihood of possessing lamellar MRCs is correlated with the life history and air-breathing capability of the fish. First, we compared the relative changes in the numbers of filament and lamellar MRCs in air-breathing and nonair-breathing fishes when they were maintained in FW and brackish water (5 g NaCl L 1 ; hereafter, this artificial seawater will be referred to only as 5 g L 1 ). Second, the distribution of both filament and lamellar MRCs in 12 orders, 28 families, 56 genera, and 66 species (including 29 species from the literature) was examined by scanning electron microscopy, zinc-iodide osmium staining (ZIO), or immunofluorescent staining. We proposed that air-breathing fish are more likely to have lamellar MRCs while non-air-breathing fish are less likely to have them. The association between air-breathing capability and the presence of lamellar MRCs was tested. We also examined the association between the distribution of lamellar MRCs and the divisions of FW fishes. Material and Methods Fish-Rearing and Acclimation Conditions Most FW fishes cannot tolerate either full seawater or deionized water (Helfman et al. 1997). For FW species, fish were first acclimated in FW or 2 g L 1 artificial saltwater for at least 1 wk (but not longer than 1 mo) and then transferred to FW and 5 g L 1, respectively, for 1 wk. For peripheral FW species, fish were first acclimated in 7 g L 1 artificial saltwater for at least 1 wk (but not longer than 1 mo) and then further transferredtofwand5gl 1, respectively, for 1 wk. Water compositionsforfwand5gl 1 were as follows (in mm): Na 1.30, 85.22; K 0.17, 4.10; Ca , 2.95; Mg , 4.17; Cl , 76.25; and ph 7.39, 7.20, respectively. Fish were kept in aerated water at 28 1 C on a photoperiod of 14 L : 10 D and were fed daily with commercial pellets. Artificial saltwater was prepared by adding commercial salt (Coralife Scientific Grade Marine Salt, Carson, Calif.) to local tap water and aerating the mixture for at least 1 mo. About one-half of the tank water was changed every 3 d. Sample Preparation Gill filaments from one side of the second branchial arch were removed and prefixed in phosphate-buffered, 4% paraformaldehyde (Sigma) and 5% glutaldehyde (Wako) for 8 12 h at 4 C, ph 7.4. They were rinsed three times with 0.1 M phosphate buffer (PB; [0.2 M NaH 2 PO 4 7 2H 2 O 0.2 M Na 2 HPO 4 7 2H 2 O] : deionized water p 1:1, php 7.2) before being post- fixed with phosphate-buffered 1% osmium tetroxide for 1 2 h. The tissues were dehydrated in seven ascending concentrations of ethanol (30% 100%) and in 100% acetone and were then dried using a Hitachi HCP-2 critical-point drier. After sputter-coating for 3 min with a gold-palladium complex using an Eiko ib-2 vacuum evaporator, the filaments were examined using a scanning electron microscope (SEM; Hitachi, S-2300). Gill filaments from the other side of the second branchial arches were collected for either ZIO or immunofluorescent staining. For ZIO fixation, we followed the procedures described by Watrin and Mayer-Gostan (1996). In particular, the zinc iodide (ZnI 2 ) solution must be prepared immediately before use (0.24 g ZnI 2 powder [Aldrich] in redeionized water to 8 ml total volume). Tissue samples were fixed in the dark for h, depending on the size of the gills, and rinsed with deionized water for 4 6 h before serial alcohol/xylene dehydration and paraffin embedding. Six-micrometer sections (Leica, RM2025) were glued on glass slides coated with polyl-lysine (Sigma). After paraffin removal (with 100% xylene, 2 min # 3 times), sections were mounted with Permount mounting medium. According to Greco et al. (1995) and Watrin and Mayer-Gostan (1996), MRCs with numerous plasma membrane infoldings will give a black color with ZIO staining. For immunofluorescent staining, fish gills were fixed with a 2% paraformaldehyde (Sigma) and 0.5% glutaldehyde (Wako) mixture at 4 C for 10 min and rinsed three times with 0.1 M PB. Samples were placed in 30% sucrose/deionized water for 1 h before embedding with tissue-freezing medium (Shandon). Sections 8 10 mm thick (Leica, Jung frigocut 2800N) were glued on glass slides coated with poly-l-lysine (Sigma) and were rinsed three times with 0.1 M PB. Sections were placed in 10% normal goat serum at room temperature for 1 h and then incubated at 4 C for h with a 1 : 100 dilution of a-5 monoclonal antibody against the a-subunit of chicken Na,K - ATPase (Developmental Studies Hybridoma Bank, Iowa City, Iowa). After washing (three times) in 0.1 M PB, the sections were incubated for 1 h at room temperature in a 1 : 200 dilution of fluorescein isothiocyanate conjugated goat anti-mouse IgG (Jackson). The Relative Numbers of MRCs in Filaments and Lamellae Our goal was to compare the numbers of filament and lamellar MRCs in both air-breathing and non-air-breathing fishes with salinity change. Six fish species were included in this experi-

3 Mitochondria-Rich Cells in the Gills of Air-Breathing Fishes 217 ment: Colisa lalia, Trichigaster trichopterus, Periophthalmus cantonensis, Oreochromis mossambicus, Monodactylus argenteus, and Toxotes jaculator. The first three species are air breathers while the latter three are not. They were acclimated and transferred, respectively, to FW and 5 g L 1 following the procedure described previously. For every fish species, at least four individuals from each water condition were collected for MRC determination. Within the same individual, at least four different regions were examined and averaged for filament MRCs. Ten lamellae per fish gill were randomly chosen for quantification. The afferent side of a gill filament from the second branchial arch was examined using SEM (Hitachi, S-2300). Because the size of the filament varied among species, the samples were examined under different magnifications. Therefore, the actual areas for counting MRCs on the SEM screen were not the same, ranging from 400 to 1,080 mm 2. For comparison, we later standardized the results to become number of MRCs per 1,000 mm 2. Lamellar MRCs were examined from the sagittal sections after ZIO fixation preparation. To avoid MRCs located on the base of a lamella, only those cells that were at least 5 mm away from the base were included. The lengths of the lamellae were determined by the use of Image-Pro Plus 4.0 data-analysis software (Media Cybernetics, Silversprings, Md.), and the numbers of lamellar MRCs were later standardized as number of MRCs per 100 mm of lamella in length. The number of MRCs in FW divided by the number of those in 5 g L 1 gave an indication of the degree of variation with salinity change. The Association between Lamellar MRCs and Air-Breathing Capability in 66 Fish Species To examine the likelihood of having lamellar MRCs in airbreathing fish, we examined lamellar MRCs of the FW and peripheral species under FW and 5 g L 1 conditions and of marine species under seawater conditions (35 g L 1 ). At least five individuals per species were examined for both filament and lamellar MRCs. We followed Nelson (1994) for classification and nomenclature of the fish. For analysis purposes, fish were classified as (1) air-breathing fish, (2) fish species that have never been described as air breathers but belong to families with known air-breathing species, and (3) non-air-breathing fish according to Graham (1997). Air-breathing fish are generally further classified as those that have specific air-breathing organs (amphibious mode) and those that have the ability of aquatic surface respiration (aquatic mode) species (Graham 1997). We pooled these two groups of air-breathing species in this analysis. The three species (Epiplatys fasciolatus, Fundulus heteroclitus, and Glossogobius olivaceus) that have never been described as air breathers but belong to the families with known air-breathing species were not included in the genus- and species-level statistical analyses. There were four categories in the analysis of number of lamellar MRCs (Tables 1, 2): (1) a definite presence of lamellar MRCs, (2) no lamellar MRC found, (3) few if any lamellar MRCs, and (4) medium-dependent lamellar MRCs. When more than one species was examined in the same genus, we determined which category the genus belonged to depending on whether these species were unanimous in terms of the presence of lamellar MRCs. The same criteria were applied to determine the category at the family level, except that those genera that contain species with no known air breathers but that belong to air-breathing families were pooled into the air-breathing families. Species in the category of few lamellar MRCs had less than three incidences of a single MRC among the 10 or more lamellae examined under SEM. We assumed that these few isolated MRCs should not have too much of an effect on either impeding gas exchange or enhancing ion regulation, and, therefore, those species were pooled into the category of no lamellar MRCs for further statistical analysis. The association between air-breathing capability and the distribution of lamellar MRCs was tested using a 2 # 2 Fisher exact statistical analysis. The experiments and handling of the animals comply with the current laws of Taiwan, Republic of China. Results An increase in filament MRCs was found in both air-breathing and non-air-breathing fish acclimated to FW, except Monodactylus argenteus. Lamellar MRCs were only found in airbreathing species, but the numbers of lamellar MRCs did not change significantly with water conditions, except in Periophthalmus cantonensis (Fig. 1). Both Colisa lalia and Trichigaster trichopterus are primary FW fish with labyrinths in their gill chambers to enhance air breathing. For both species, a significantly higher number of filament MRCs were found in those from FW than from 5 g L 1 (Fig. 2). However, no difference was found in the number of lamellar MRCs between the two media in either species. Periophthalmus cantonensis is an amphibious, peripheral FW fish. The number of MRCs in filaments was significantly higher in FW. On the contrary, lamellar MRCs were significantly higher in 5 g L 1 than in FW (Fig. 2). The secondary FW fish Oreochromis mossambicus had no MRCs in the lamellae, and the number of filament MRCs was significantly higher in FW than in 5 g L 1. In the two peripheral FW fish M. argenteus and Toxotes jaculator, very few lamellar MRCs could be found at the base of the lamellae. A significantly higher number of filament MRCs were found in FW for T. jaculator, while there was no significant difference in the number of filament MRCs between the two water conditions for M. argenteus (Fig. 2). We examined 37 fish species for the presence of lamellar MRCs. Among them, 19 species were examined in 5 g L 1 and FW conditions; two in both saltwater and FW; 13 in only FW;

4 218 Table 1: Examination of the presence of lamellar mitochondria-rich cells (MRCs) in the gills of 66 fish species, including 29 species reported in the previous literature Order, Family, and Species Division Medium SW 5gL 1 FW Methodology References Air-breathing fish with specific air-breathing organs: Perciformes: Belontiidae: Betta splendens 1 Colisa lalia 1 Macropodus opercularis 1 Trichogaster leeri 1 Trichogaster trichopterus 1 Gobiidae: Periophthalmus cantonensis p Periophthalmodon schlosseri p a Na,K -ATPase immunohistochemistry Wilson et al Boleophthalmus pectinirostris p f Scartelaos virides p Cypriniformes: Cobitididae: Cobitis taenia 1 LM and TEM Pisam et al Misgurnus anguillicaudatus 1 LM and TEM Kikuchi 1977 Siluriformes: Loricariidae: Hypostomus plecostomus 1 LM, TEM, and SEM Fernandes and Perna-Martins 2001 Callichthyidae: Corydoras aeneus 1 Acipenseriformes: Lepisosteidae: Lepisosteus osseus 2 Anguilliformes: Anguillidae: Anguilla anguilla p LM and TEM Laurent and Dunel 1980 Anguilla japonica p LM and TEM Shirai and Utida 1970 Na,K -ATPase immunohistochemistry Sasai et al and quantified Anguilla rostrata p f SEM and TEM Perry et al. 1992a, 1992b Anguilla marmorata p f Cyprinodontiformes: Aplocheilidae:

5 219 Rivulus marmoratus p LM and TEM King et al Air-breathing fish with aquatic surface respiration capability: Cypriniformes: Cyprinidae: Balantiocheilus melanopterus 1 f Barbus tertrazona 1 Labeo erythrurus 1 f Carassius auratus 1 LM and TEM Kikuchi 1977 SEM and quantified Lee et al. 1996a Carassius carassius 1 LM and TEM Kikuchi 1977 Cyprinus carpio 1 LM and TEM Kikuchi 1977 LM and TEM Hwang 1988 SEM and quantified Lee et al. 1996a Gobio gobio 1 LM and TEM Pisam et al Paracheilognathus himantegus 1 Fish not described as air breathers in Graham (1997) but belonging to families with known air-breathing species: Cyprinodontiformes: Cyprinodontidae: Epiplatys fasciolatus 2 Fundulus heteroclitus 2 SEM and quantified Hossler et al Na,K -ATPase immunohistochemistry Katoh et al Perciformes: Gobiidae: Glossogobius olivaceus p Non-air-breathing fish: Characiformes: Characidae: Pristella maxillaries 1 Prionobrama filigera 1 Colossoma bidens 1 Serrasalmus nattereri 1 Siluriformes: Ictaluridae: Ictalurus nebulosus 1 f SEM and TEM Perry et al. 1992a, 1992b Perciformes: Cichlidae: Astrenotus ocellatus 2 Haplochromis ahli 2

6 220 Table 1 (Continued) Order, Family, and Species Division Medium SW 5gL 1 FW Methodology References Melanochromis auratus 2 Oreochromis aureus 2 LM and TEM Avella et al Oreochromis mossambicus 2 Oreochromis niloticus 2 LM and TEM Cioni et al LM and TEM Avella et al Pelvicachromis pulcher 2 Pterophyllum scalare 2 Symphysodon aequifasciatus 2 Tropheus duboisi 2 Cichlasoma biocellatum 2 LM Mattheli and Stroband 1971 Cyprinodontiformes: Oryziidae: Oryzias latipes 2 SEM and quantified Lee et al. 1996b Poecilidae: Lebistes reticulatus 2 SEM Pisam et al Poecilia reticulata 2 LM Shikano et al Poecilia latipinna 2 Perciformes: Lobotidae: Datnioides quadrifasciatus p Monodactylidae: Monodactylus sebae p f Monodactylus argenteus p f f Percichthyidae: Lateolabrax japonicus p Na,K -ATPase immunohistochemistry Hirai et al Toxotidae: Toxotes jaculator p f f Salmoniformes: Salmonidae: Oncorhynchus mykiss p f SEM and quantified Laurent and Hebibi 1989 Treated with cortisol; SEM Perry et al. 1992a, 1992b (also known as Salmo gairdneri) SEM Laurent and Perry 1990 Oncorhynchus keta p Chum salmon fry Na,K -ATPase immunohistochemistry Uchida and Kaneko 1996; Uchida et al Mature chum salmon Na,K -ATPase immunohistochemistry Uchida et al. 1997

7 221 Salmo trutta p SEM and quantified Brown 1992 Na,K -ATPase immunohistochemistry Seidelin et al Osmeriformes: Plecoglossidae: Plecoglossus altivelis p LM and TEM Hwang 1988 Perciformes: Centropomidae: Ambassis gymnocephalus m Leiognathidae: Leiognathus nuchalis m Percichthyidae: Morone saxatilis m LM, TEM, and SEM King and Hossler 1991 Teraponidae: Terapon jarbua m Mugiliformes: Mugilidae: Mugil cephalus m SEM Hossler et al Pleuronectiformes: Soleidae: Solea solea m LM and TEM Dunel-Erb and Laurent 1980 Scophthalmidae: Scophthalmus maximus m LM and TEM Pisam et al ZIO Watrin and Mayer-Gostan 1996 Note. Fish were classified into three groups: (1) air-breathing fish, (2) non-air-breathing fish, and (3) fish species that have never been described as air breathers but belong to the families with known air-breathing species, according to Graham (1997). The divisions of the fish families used are primary freshwater (FW) fish (1 ), secondary FW fish (2 ), peripheral FW fish (p), and marine fish (m), following Helfman et al. (1997). The 37 species we examined were acclimated in both 5 g L 1 artificial saltwater and FW. In the Medium columns, we present the results as a definite presence of lamellar MRCs ( ), no lamellar MRCs ( ), or few if any lamellar MRCs (f). SW p saltwater. a Acclimated at 15 g L 1.

8 222 H.-C. Lin and W.-T. Sung Table 2: Summary of the number of species with lamellar mitochondria-rich cells in the gills Division Air-Breathing Fish Non-Air-Breathing Fish Non Air Breather but in Air- Breathing Family f / f / p m Note. The notations are the same as in Table 1, and those that changed among the three media (saltwater, 5 g L 1 artificial saltwater, and freshwater) are indicated as /. one in saltwater; one in 5 g L 1 ; and one in saltwater, 5 g L 1, and FW conditions. For the 29 species reported in the previous literature, there were two in saltwater; one in 5 g L 1 ;10in FW;threeinboth5gL 1 and FW; 12 in both saltwater and FW; and one in saltwater, 5 g L 1, and FW conditions. There were 27 air-breathing fish, 36 non-air-breathing fish, and three species that have never been described as air breathers but belong to families with known air-breathing species according to Graham (1997). The results on the presence of lamellar MRCs of fish in different water conditions are presented in Table 1. Lamellar MRCs were found in 20 species, among which 18 were air breathers. There were 39 species that have zero or few lamellar MRCs, and 32 of them were non air breathers. The number of species in each category is summarized in Table 2. There was a significant association between the presence of lamellar MRCs and the mode of breathing on all three levels of systematic categories (species, genus, and family; Table 3). Seven species did not have a consistent observation among different media: Boleophthalmus pectinirostris, Anguilla japonica, Monodactylus sebae, Lateolabrax japonicus, Oncorhynchus mykiss (previously known as Salmo gairdneri), Oncorhynchus keta, and Salmo trutta. These species were not included in the statistical analysis but will be discussed separately later. According to the division to which the fish family belongs (Helfman et al. 1997), there were 22 primary FW species, 18 secondary FW species, 19 peripheral FW species, and seven marine species included in this study (Table 2). Among them, 15 primary FW species had lamellar MRCs, and 15 secondary FW species had no lamellar MRCs. There was a lot of variation in the peripheral FW species, and no generalization can be made currently. Most of the variation in the presence of lamellar MRCs was found in peripheral FW species from the orders Perciformes, Anguilliformes, Cyprinodontiformes, Salomoniformes, and Osmeriformes. None of the seven marine fish species had lamellar MRCs. When the division of the fish and air-breathing capability were considered simultaneously, we found that those primary FW fish species that are also air breathers are more likely to have lamellar MRCs. Because almost all secondary FW species are non-air-breathing species (Graham 1997; Helfman et al. 1997), the confounded relationship was not resolved until we examined Lepisosteus osseus, an air-breathing and secondary FW species. However, it was not possible to conduct a statistical analysis with both air-breathing capability and the division of fish family considered at the same time due to the low number of observations (less than five) in certain categories in Table 2. Discussion Although many authors have reported that lamellar MRCs can be found in several fish species acclimated in FW or soft water (see Table 1 for references) or can be induced by hormones (Perry et al. 1992a, 1992b; Bindon et al. 1994; Perry 1998; Dang et al. 2000), there has been little discussion as to which kinds of fish are more likely to possess lamellar MRCs. Nevertheless, it is suggested that lamellar MRCs may play a role in ion uptake in FW species, such as FW trout Salmo gairdneri (Laurent and Perry 1990), sea trout Salmo trutta (Brown 1992), rainbow trout Oncorhynchus mykiss (Greco et al. 1995), chum salmon Oncorhynchus keta (Uchida et al. 1997), Japanese eel Anguilla japonica (Sasai et al. 1998), Japanese sea bass Lateolabrax japonicus (Hirai et al. 1999), and armored catfish Hypostomus plecostomus (Fernandes and Perna-Martins 2001) and in ammonia elimination in the mudskipper Periophthalmodon schlosseri (Wilson et al. 2000). These additional roles played by gill lamellae are not without a cost. According to a review by Perry (1998), proliferation of lamellar MRCs would lead to an increase in blood-to-water diffusion distance and, therefore, an impairment of gas transfer. In addition, several compensatory physiological responses, such as hyperventilation and a higher

9 Mitochondria-Rich Cells in the Gills of Air-Breathing Fishes 223 Figure 1. Mitochondria-rich cells (MRCs) in gill lamellae. A, Immunofluorescent staining and (B) zinc-iodide osmium (ZIO) staining indicating a definite presence of lamellar MRCs in Periophthalmus cantonensis. C, Immunofluorescent staining and (D) ZIO staining indicating filament MRCs but no lamellar MRCs in Monodactylus argenteus. E, Negative control for immunofluorescent staining. F p gill filament; I p interlamellar region; L p gill lamella. Arrowheads indicate the MRCs. affinity of hemoglobin to oxygen, were observed following lamellar MRC proliferation (Perry 1998). Apparently, there is a trade-off between gas exchange and ion regulation in gill lamellae. Whether there is a functional differentiation between gill filaments and lamellae is constantly under debate. Because of concurrent increases in MRCs on both filament and lamellar epithelia of S. gairdneri in FW, Laurent and Perry (1990) suggested that the function of MRCs is the same in filaments and lamellae. However, Uchida et al. (1996), in their study of anadromous chum salmon (O. keta) fry, first proposed the hypothesis that MRCs in the lamellae may be the site of ion uptake in hypoosmotic conditions and those in the filament are for salt secretion in hyperosmotic conditions. This hypothesis of a functional differentiation in fish gills was later supported by Sasai et al. (1998), examining the catadromous Japanese eel (A. japonica), and by Hirai et al. (1999) in Japanese sea bass L. japonicus, a euryhaline marine fish. Nevertheless, the hypothesis proposed by Uchida et al. (1996) is limited to those species that have lamellar MRCs and is not applicable to those that have no lamellar MRCs. In this study, it appeared that unequal changes in the numbers of filament and lamellar MRCs were found in the three species of air-breathing fish acclimated to FW and 5 g L 1.

10 224 H.-C. Lin and W.-T. Sung Figure 2. Comparison of filament and lamellar mitochondria-rich (MR) cells in six species of fish acclimated to freshwater and 5 g L 1 saltwater conditions. A C, Three species of air-breathing fish. D F, Three species of non-air-breathing species. One asterisk and three asterisks indicate significant difference according to Student s t-test ( P! 0.05 and P! , respectively). There were 1.99 to 3.86 times more filament MRCs but only 0.69 to 1.08 times more lamellar MRCs when they were in FW as compared to when they were in 5 g L 1. However, when the multicellular complex of filament MRCs found in saltwater was taken into account (Hwang 1988; Pisam et al. 1990), we may have underestimated the number of filament MRCs in the 5 g L 1 condition, and the difference between the relative changes of MRCs in filaments and lamellae could become insignificant. Therefore, our results should support the discussion by Laurent and Perry (1990) that filament and lamellar MRCs have similar functions due to their concurrent change in number. The reason to include aquatic surface respiration breathers in the category of air breathers was that these species can obtain oxygen through a gulp of air or a breath of oxygen-saturated water. Because air contains about 30 to 40 times more oxygen than water (depending on temperature, salinity, etc.), the physiological demand for oxygen should be relieved. Therefore, it is not surprising to find lamellar MRCs in these species. The comparative methods for explaining adaptations (Harvey and Pagel 1991; Harvey and Purvis 1991) address the importance of seeking independent evolutionary origins of characteristics rather than simply counting species as independent data points. The two most popular phylogenetically based statistical methods are the independent contrast methods (Fel-

11 Mitochondria-Rich Cells in the Gills of Air-Breathing Fishes 225 Table 3: Association between the presence of lamellar MRCs and air-breathing capability at species, genus, and family levels Presence of Lamellar MRCs Yes No/Few Species: Air-breathing species # 10 Non-air-breathing species 1 30 Species in the families with known air-breathing species 1 2 Genera: Air-breathing genus # 10 8 Non-air-breathing genus 1 26 Genus in the families with known air-breathing species 1 2 Families: Air-breathing family Non-air-breathing family 1 13 Note. The three species that have never been described as air breathers but belong to families with known airbreathing species according to Graham (1997) were not included in the genus- and species-level statistical analyses. Species that fell into the category of few lamellar MRCs were pooled into the category of zero lamellar MRCs for statistical analysis. Significance was calculated using 2 # 2 Fisher exact probability. P senstein 1985) and the generalized least squares models (Garland and Ives 2000). However, these two approaches are for continuous variables rather than for discrete ones to which our characters (i.e., presence of lamellar MRCs and air-breathing ability) belonged. Instead, Ridley and Grafen (1996) and Grafen and Ridley (1996) made an in-depth discussion on the discrete comparative methods. In our analysis, we did not include any phylogenetic tree due to the fact that the branch lengths of the phylogeny or even the phylogeny itself are basically not known. Instead, variation in the presence of lamellar MRCs was found at the genus level in this study. Therefore, an intergeneric analysis should be considered as an independent comparison. Nonetheless, a significant association between air-breathing capability and the presence of lamellar MRCs was found at the species, genus, and family level (Table 3). Primary FW fishes that are also air breathers usually have lamellar MRCs, while those that are not air breathers (e.g., the four species from Characiformes in Table 1) usually do not have lamellar MRCs (Table 3). Most of the secondary FW species are not air breathers (Graham 1997), and no lamellar MRCs were found. The only air-breathing secondary FW species examined in this study was Lepisosteus osseus, an ancient airbreathing fish according to Graham (1997), and it has MRCs in the gill lamellae. However, until more air-breathing secondary species are examined, no suggestion can be made as to the association between the presence of lamellar MRCs and airbreathing capability in the secondary FW fish species. The situation became complicated in the peripheral FW fishes, which primarily include the anadromous salmonids, catadromous anguillids, and certain Perciforme families (Table 1). Members from the two former groups have very complicated life histories and are susceptible to hormonal regulation (such as cortisol and prolactin) during their salinity adaptation (Laurent and Perry 1990; Perry et al. 1992a; McCormick 1994; Uchida et al. 1998). Therefore, it should not be too surprising to find inconsistent observations from different studies of A. japonica (Shirai and Utida 1970; Sasai et al. 1998), O. mykiss (Laurent and Hebibi 1989; Perry et al. 1992a, 1992b; Greco et al. 1995), and O. keta (Uchida and Kaneko 1996; Uchida et al. 1996, 1997). The other two non-air breathers with lamellar MRCs are the anadromous Plecoalossus altivelis (Hwang 1988) and the amphydromous L. japonicus (Hirai et al. 1999). The order Perciformes is the most advanced euteleostean clade and contains at least 148 families and around 9,300 species (Helfman et al. 1997). While most species in this order are marine fish, there are eight families that are primary FW fish, one family that is secondary FW fish, and nine families that are peripheral FW. Together, these three divisions of FW fishes account for about 2,100 species (Berra 1981). There are also various types of air-breathing fishes within this order that deserve further discussion as soon as the phylogenetic systematics becomes conclusive. According to Graham (1997), air breathing is an ancient trait that evolved independently in many fishes and there is great diversity in this bimodal respiratory specialization. Of course, the degree of air breathing should not be taken as a dichotomous trait but rather a continuous spectrum. But, in this study, for the convenience of statistical analysis and sample size, we only classified the species into air breathers and non-air breathers. Gills, as a multifunction organ, should constantly have

12 226 H.-C. Lin and W.-T. Sung functional differentiation. That is, extrabranchial gas exchange may lead to an increase in the role of ion regulation in gills. Therefore, lamellar MRCs, which would greatly impede respiration, were found. However, since MRCs can also be found in the gills of the hagfish Myxine glutinosa (Choe et al. 1999) and in the gill lamellae of the FW stingray Dasyatis sabina (Piermarini and Evans 2000), it is also considered as a rather ancient trait. According to this rationale, the filament MRCs are universally found in all fishes, and if air breathing should evolve, distribution of lamellar MRCs may follow. Hirai et al. (1999) observed the migration of MRCs from filaments to lamellae, which may explain the origin of lamellar MRCs. In this study, only those species that have been documented as air breathers were included in the analysis. For species that have never been reported to be air breathers but belong to the air-breathing families, a better description on their life histories is needed. More studies on the physiology of the air-breathing fishes from an integrative approach would allow a comparison of the relative increase in both filament and lamellar MRCs with salinity change in air-breathing and non-air-breathing fishes. Acknowledgments We are grateful to Dr. Tsung-Han Lee and Dr. Pung-Pung Hwang for their comments during the preparation of the manuscript. This study was supported by the National Science Council (NSC B ) and Academia Sinica (Major Group-Research Project), Taiwan, R.O.C. to H.C.L. Literature Cited Avella M., J. Berhaut, and M. Bornancin Salinity tolerance of two tropical fishes Oreochromis aureus and O. niloticus. I. Biochemical and morphological changes in the gill epithelium. J Fish Biol 42: Berra T.M An Atlas of Distribution of the Freshwater Fish Families of the World. University of Nebraska Press, Lincoln. Bindon S.D., K.M. Gilmour, J.C. Fenwick, and S.F. Perry The effects of branchial chloride cell proliferation on respiratory function in the rainbow trout, Oncorhynchus mykiss. J Exp Biol 197: Brown P Gill chloride cell surface-area is greater in freshwater-adapted adult sea trout (Salmo trutta, L.) than those adapted to sea water. J Fish Biol 40: Choe K.P., S. Edwards, A.I. Morrison-Shetlar, T. Toop, and J.B. Claiborne Immunolocalization of Na /K -ATPase in mitochondrion-rich cells of the Atlantic hagfish (Myxine glutinosa) gill. Comp Biochem Physiol 124A: Cioni C., D. DeMerich, E. Cataldi, and S. Cataudella Fine structure of chloride cells in freshwater- and seawateradapted Oreochromis niloticus (Linnaeus) and Oreochromis mossambicus (Peters). J Fish Biol 39: Dang Z., P.H.M Balm, G. Flik, S.E. Wendelaar Bonga, and R.A.C. Lock Cortisol increases Na /K -ATPase density in plasma membranes of gill chloride cells in the freshwater tilapia Oreochromis mossambicus. J Exp Biol 203: Dunel-Erb S. and P. Laurent Ultrastructure of marine teleost gill epithelia: SEM and TEM study of the chloride cell apical membrane. J Morphol 165: Evans D.H., P.M. Piermarini, and W.T.W. Potts Ionic transport in the fish epithelium. J Exp Zool 283: Felsenstein J Phylogenies and the comparative method. Am Nat 125:1 15. Fernandes M.N. and S.A. Perna-Martins Epithelial gill cells in the armored catfish, Hypostomus CF. plecostomus (Loricariidae). Rev Bras Biol 61: Garland T., Jr., and A.R. Ives Using the past to predict the present: confidence intervals for regression equations in phylogenetic comparative methods. Am Nat 155: Grafen A. and M. Ridley Statistical tests for discrete crossspecies data. J Theor Biol 183: Graham J.B Air-Breathing Fishes: Evolution, Diversity, and Adaptation. Academic Press, San Diego, Calif. Greco A.M., K.M. Gilmour, J.C. Fenwick, and S.F. Perry The effects of softwater acclimation on respiratory gas transfer in the rainbow trout Oncorhynchus mykiss. J Exp Biol 198: Harvey P.H. and M.D. Pagel The Comparative Method in Evolutionary Biology. Oxford University Press, New York. Harvey P.H. and A. Purvis Comparative methods for explaining adaptation. Nature 351: Helfman G.S., B.B. Collette, and D.E. Facey The Diversity of Fishes. Blackwell Science, Malden, Mass. Hirai N., M. Tagawa, T. Kaneko, T. Seikai, and M. Tanaka Distributional changes in branchial chloride cells during freshwater adaptation in Japanese sea bass Lateolabrax japonicus. Zool Sci 16: Hossler F.E., G. Musil, K.J. Karnaky, Jr., and F.H. Epstein Surface ultrastructure of the gill arch of the killifish, Fundulus heteroclitus, from seawater and freshwater, with special reference to the morphology of apical crypts of chloride cells. J Morphol 185: Hossler F.E., J.R. Ruby, and T.D. McIlwain The gill arch of the mullet, Mugil cephalus. I. Surface ultrastructure. J Exp Zool 208: Hwang P.P Multicellular complex of chloride cells in the gills of freshwater teleosts. J Morphol 196: Jurss K. and R. Bastrop The function of mitochondriarich cells (chloride cells) in teleost gills. Rev Fish Biol Fish 5: Katoh F., A. Shimizu, K. Uchida, and T. Kaneko Shift

13 Mitochondria-Rich Cells in the Gills of Air-Breathing Fishes 227 of chloride cell distribution during early life stages in seawater-adapted killifish, Fundulus heteroclitus. Zool Sci 17: Kikuchi S Mitochondria-rich (chloride) cells in the gill epithelia from four species of stenohaline fresh water teleosts. Cell Tissue Res 180: King J.A.C., D.C. Abel, and D.R. DiBona Effects of salinity on chloride cells in the euryhaline cyprinodontid fish Rivulus marmoratus. Cell Tissue Res 257: King J.A.C. and F.E. Hossler The gill arch of the striped bass (Morone saxatilis). IV. Alterations in the ultrastructure of chloride cell apical crypts and chloride efflux following exposure to seawater. J Morphol 209: Laurent P Gill internal morphology. Pp in W.S. Hoar and D.J. Randall, eds. Fish Physiology. Vol. 10A. Academic Press, New York. Laurent P. and S. Dunel Morphology of gill epithelia in fish. Am J Physiol 238:R147 R159. Laurent P. and N. Hebibi Gill morphometry and fish osmoregulation. Can J Zool 67: Laurent P. and S.F. Perry The effects of cortisol on gill chloride cell morphology and ionic uptake on the freshwater trout, Salmo gairdneri. Cell Tissue Res 259: Environmental effects on fish gill morphology. Physiol Zool 64:4 25. Lee T.H., P.P. Hwang, and S.H. Feng. 1996a. Morphological studies of gill and mitochondria-rich cells in the stenohaline cyprinid teleosts, Cyprinus carpio and Carassius auratus, adapted to various hypotonic environments. Zool Stud 35: Lee T.H., P.P. Hwang, S.F. Feng, and F.L. Huang. 1996b. The gill structure and branchial mitochondria-rich cells of the medaka, Oryzias latipes. Acta Zool Taiwanica 7: Marshall W.S. and S.E. Bryson Transport mechanisms of seawater teleost chloride cells: an inclusive model of a multifunctional cell. Comp Biochem Physiol 119A: Mattheli J.A.M. and H.W.J. Stroband The effects of osmotic experiments and prolactin on the mucous cells in the skin and the ionocytes in the gills of the teleost Cichlasoma biocellatum. Z Zellforsch Mikrosk Anat 121: McCormick S.D Ontogeny and evolution of salinity tolerance in anadromous salmonids: hormones and heterochrony. Estuaries 17: Myers G.S Fresh-water fishes and West Indian zoogeography. Smithson Inst Annu Rep 1937, pp Nelson J.S Fishes of the World. 3d ed. Wiley, New York. Perry S.F Relationships between branchial chloride cells and gas transfer in freshwater fish. Comp Biochem Physiol 119A:9 16. Perry S.F., G.G. Goss, and J.C. Fenwick. 1992a. Interrelationships between gill chloride cell morphology and calcium uptake in freshwater teleosts. Fish Physiol Biochem 10: Perry S.F., G.G. Goss, and P. Laurent. 1992b. The interrelationships between gill chloride cell morphology and ionic uptake in four freshwater teleosts. Can J Zool 70: Piermarini P.M. and D.H. Evans Effects of environmental salinity on Na /K -ATPase in the gills and rectal gland of a euryhaline elasmobranch (Dasyatis sabina). J Exp Biol 203: Pisam M., G. Boeuf, P. Prunet, and A. Rambourg Ultrastructural features of mitochondria-rich cells in stenohaline freshwater and seawater fishes. Am J Anat 187: Pisam M., A. Caroff, and A.S. Rambourg Two types of chloride cells in the gill epithelium of a freshwater-adapted euryhaline fish: Lebistees reticulatus, their modifications during adaptation to seawater. Am J Anat 179: Randall D.J. and C. Brauner Interactions between ion and gas transfer in freshwater teleost fish. Comp Biochem Physiol 119A:3 8. Ridley M. and A. Grafen How to study discrete comparative methods. Pp in E.P. Martins, ed. Phylogenies and the Comparative Method in Animal Behavior. Oxford University Press, Oxford. Sasai S., T. Kaneko, S. Hasegawa, and K. Tsukamoto Morphological alteration in two types of gill chloride cells in Japanese eels (Anguilla japonica) during catadromous migration. Can J Zool 76: Seidelin M., S.S. Madsen, H. Blenstrup, and C.K. Tipsmark Time-course changes in the expression of Na,K - ATPase in gills and pyloric caeca of brown trout (Salmo trutta) during acclimation to seawater. Physiol Biochem Zool 73: Shikano T., E. Arai, and Y. Fujio Seawater adaptability, osmoregulatory function, and branchial chloride cells in the strain selected for high salinity tolerance of the guppy Poecilia reticulata. Fish Sci 64: Shirai N. and S. Utida Development and degeneration of the chloride cell during seawater and freshwater adaptation of the Japanese eel, Anguilla japonica. Z Zellforsch Mikrosk Anat 3: Uchida K. and T. Kaneko Enhanced chloride cell turnover in the gills of chum salmon fry in seawater. Zool Sci 13: Uchida K., T. Kaneko, M. Tagawa, and T. Hirano Localization of cortisol receptor in branchial chloride cells in chum salmon fry. Gen Comp Endocrinol 109: Uchida K., T. Kaneko, A. Yamaguchi, T. Ogasawara, and T. Hirano Reduced hypoosmoregulatory ability and alteration in gill chloride cell distribution in mature chum salmon (Oncorhynchus keta) migrating upstream for spawning. Mar Biol 129: Uchida K., T. Kaneko, K. Yamauchi, and T. Hirano Morphometrical analysis of chloride cell activity in the gill filaments and lamellae and changes in Na,K -ATPase activity during seawater adaptation in chum salmon fry. J Exp Zool 276:

14 228 H.-C. Lin and W.-T. Sung Watrin A. and N. Mayer-Gostan Simultaneous recognition of ionocytes and mucous cells in the gill epithelium of turbot and in the rat stomach. J Exp Zool 276: Wilson J.M., D.J. Randall, M. Donowitz, A.W. Vogl, and A.K.- Y. Ip Immunolocalization of ion-transport proteins to branchial epithelium mitochondria-rich cells in the mudskipper (Periophthalmodon schlosseri). J Exp Biol 203: