Release of Atrial Natriuretic Factor Prohormone Peptides 1-30, 31-67 And 99-126 From Freshwater- and Seawater-Acclimated Perfused Trout (Oncorhynchus Mykiss) Hearts

Atrial natriuretic factor (ANF) is a hormone produced and secreted from the heart (DeBold et al. 1981; Dietz, 1984). In mammals, ANF and three other peptide hormones are derived from a common prohormone (Flynn et al. 1985; Seidman et al. 1985; Vesely et al. 1989). In humans, ANF is the carboxy-terminal peptide resulting from protease cleavage of the 126-amino-acid prohormone at position arginine–serine (98–99) (Imada et al. 1988). Three other peptides are derived from the amino-terminal portion of the prohormone and are biologically active (Vesely et al. 1987). Furthermore, their effects are both more potent and longer-lasting than that of ANF (Martin et al. 1990; Vesely et al. 1994). These three proANF peptide hormones, consisting of amino acids 1–30, 31–67 and 79–98, have been tentatively named long-acting natriuretic peptide, vessel dilator and kaliuretic peptide, respectively, based upon their most potent biological properties (Vesely et al. 1994). Like ANF, these hormones circulate in the blood until they act at various receptor sites to cause natriuretic (increased Na⁺ excretion), diuretic (increased urine production) and hypotensive (low blood pressure) effects (Vesely et al. 1992, 1994; Martin et al. 1990). Thus, natriuretic peptides are intimately involved in blood volume regulation.

An extensive literature exists regarding natriuretic peptides and their release from the heart (see Vesely, 1992, for a review), but the majority of this is based on mammalian systems. ANF appears to be a highly conserved peptide for blood volume regulation, being present in several taxa, invertebrate and vertebrate alike. For instance, both ANF and proANFs have been found in the oyster (Crassostrea virginica; Vesely et al. 1991; Palmer et al. 1994) and the red-ear turtle (Pseudemys scripta; Baeyens et al. 1989), while Baranowski and Westenfelder (1989) demonstrated that ANF was released from isolated ventricular myocytes obtained from a fish heart. In mammals, the most effective and best-documented stimulus for ANF release is atrial distension (Dietz, 1984; Lang et al. 1985; Ledsome et al. 1985; Anderson et al. 1986; Edwards et al. 1988; Agnoletti et al. 1989), and the same appears to be true for a non-mammalian system since ANF release increases during atrial distension of a perfused rainbow trout heart (Cousins and Farrell, 1996). ANF has been identified in the plasma of many fish species (Evans et al. 1989; Uemura et al. 1990; Takei et al. 1990; Westenfelder et al. 1988; Uva et al. 1993), but the release of proANF peptides has not been previously documented.

Although it appears that the mechanism for ANF release may be highly conserved among the vertebrates, such that atrial distension results in ANF release in both mammals (Dietz, 1984) and fish (Cousins and Farrell, 1996), it is not clear whether an ionic mediator, such as Na⁺, can influence ANF production and release. A decrease in dietary Na⁺ in rats results in a decrease both in atrial ANF content (DeBold, 1979) and in levels of ANF and proANF peptides in the circulation (Dietz et al. 1992). Increases in dietary salt levels, however, do not result in an increase in atrial ANF content (Takayanagi et al. 1985; Dietz, 1987) even though the level of plasma ANF rises (Tanaka et al. 1984; Sagnella et al. 1985; Weidmann et al. 1986). This suggests that higher salt levels increase ANF release without simultaneously stimulating ANF synthesis. When rainbow trout are fed high-salt diets, their blood and pulse pressures increase (Eddy et al. 1990), but it is not known whether the levels of plasma ANF also change. Although several studies have shown that seawater fish tend to have higher ANF plasma concentrations than freshwater fish (Westenfelder et al. 1988; Evans et al. 1989; Smith et al. 1991), it is not known whether ANF production and release of ANF differ between animals in these two environments. Interestingly, when a euryhaline species is transferred to sea water, as shown by Eddy and Chang (1993) for Atlantic salmon parr (Salmo gairdineri), plasma ANF levels increase.

While atrial stretch, which is associated with hypervolemia, is the main stimulus for ANF release from the perfused heart of freshwater trout (Cousins and Farrell, 1996), it is not known how an osmotic challenge would affect this secretion. We have, therefore, investigated the differences in the regulation of ANF secretion in the perfused hearts of freshwater-and seawater-acclimated rainbow trout. In addition, we also investigated the possibility that fish hearts release other ANF-related cardiac peptides and whether the acclimation of freshwater trout to sea water affected the release of these ANF-related peptides.

All rainbow trout [Oncorhynchus mykiss (Walbaum)] in this study were obtained from a local supplier (West Creek Trout Farms, Aldergrove, British Columbia) and weighed 300–600 g. The experimental animals were fed a maintenance diet (1 % body mass) of commercial trout chow (Silver Cup Fish Food, Murray, UT) on a daily basis. The freshwater trout were held indoors in 2000 l fiberglass tanks supplied with flowing, dechlorinated fresh water (13 °C; 12 h:12 h L:D photoperiod). A group of 12 fish was transferred to a tank supplied with recirculating sea water (salinity 28 ‰). These trout were allowed to acclimate to the sea water for at least 6 weeks and showed no signs of distress during this time. There were no mortalities during acclimation.

Perfused hearts from rainbow trout were used in situ to investigate how changes in filling pressure (PI) affected ANF release. By subjecting both seawater-acclimated trout (SW; N=6) and freshwater trout (FW; N=6) to an identical protocol, the effects of salinity on ANF release could be examined. The perfused trout heart preparation utilized in this study was similar to that described by Farrell et al. (1986, 1988) and Cousins and Farrell (1996). In brief, an animal was randomly selected and anesthetized using a buffered MS222 solution (1:5000 w/v). The animal was then measured, weighed and transferred to an operating sling. During surgery, the gills were irrigated with chilled and aerated buffered MS222 solution (1:10 000 w/v). At this time, plasma samples were obtained from the blood drawn from the caudal vein. An input cannula was inserted into the sinus venosus through a hepatic vein, and the pericardium was secured around the cannula. An output cannula was inserted into the bulbus arteriosus through the ventral aorta and, finally, the ducti Cuvier were ligated to prevent backflow and to ensure that the heart received input only from the cannula. The total time for the surgery was approximately 20 min.

During the experiment, the circulated perfusate was collected from the output tubing at 1 min intervals and the fractions were then frozen by immersion in an ethanol and dry ice mixture. The perfusion saline for the FW trout was identical to the saline used by Cousins and Farrell (1996) and contained (in mmol l⁻¹): NaCl, 124.1; KCl, 3.1; CaCl₂, 2.5; MgSO₄, 0.9: dextrose, 0.5. The perfusion saline for the SW trout consisted of (in mmol l⁻¹): NaCl, 150; KCl, 5; CaCl₂, 2.3; MgSO₄, 2.0; dextrose, 0.5 (modified from Farrell et al. 1985). Both perfusate recipes were buffered with 10 mmol l⁻¹ Tes with a pH of 7.85 at 13 °C. The osmolality of the freshwater perfusion solution was 281 mosmol kg⁻¹ and that of the seawater solution was 328 mosmol kg⁻¹, which simulated their ambient osmolalities. For all preparations, adrenaline (5 nmol l⁻¹; bitartrate salt, Sigma Chemical, St Louis, MO) was added at a constant rate (50 μl min⁻¹) to the perfusate solution to provide a resting tonus to heart rate and to reduce deterioration of the preparations (Farrell and Milligan, 1986).

Once the surgery was completed, the heart was left to pump under control conditions for 30 min to ameliorate any effects associated with surgical trauma. We have previously demonstrated that ANF release, which is initially high after surgery, falls rapidly to a stable, basal rate that can be maintained for several hours if perfusion conditions are constant (Cousins and Farrell, 1996). Control conditions consisted of (1) a filling pressure (PI) set to maintain a cardiac output (Q̇) of 17 ml min⁻¹ kg⁻¹ body mass and (2) an afterload or output pressure (PO) set at 4.91 kPa, which resemble conditions established for resting trout in vivo (Graham and Farrell, 1989; Kiceniuk and Jones, 1977). Since changes in heart rate are typically small, Q̇ and stroke volume were adjusted by changes in PI We measured ANF release under four conditions, namely, Q̇ (the value for Q̇measured under resting conditions), 30 % Q̇_max, 60 % Q̇_max and Q̇_max (maximum cardiac output), where Q̇_max was estimated from previous studies (e.g. Cousins and Farrell, 1996) and from the body mass of the animal. The order of testing the elevated Q̇ conditions was randomize d to control for possible cross-over effects. Each level of Q̇ was maintained for 10 min with intervening periods when Q̇ was maintained at control levels for 30 min. At the end of these test periods, fractions of heart perfusate were collected for a period of 1 min and immediately frozen, with the exception of the initial equilibration period when 1 min fractions were collected every 10 min. Once completely frozen, the samples were freeze-dried for approximately 24 h and then stored at room temperature (25 °C) in a powdered form until reconstituted with 1 ml of the radioimmunoassay buffer and assayed for immunoreactive ANF (ir-ANF).

Natriuretic peptides utilized for the radioimmunoassays described below were obtained from Peninsula Laboratories, Inc. (Belmont, CA, USA). The measurement of ir-ANF in the FW-and SW-acclimated trout plasma and reconstituted perfusate samples was accomplished using a radioimmunoassay (RIA) specific for human α-ANF-(99-126). Samples of FW and SW-acclimated trout plasma were obtained from blood drawn from the caudal vein and were collected into chilled tubes containing EDTA (1 mg ml⁻¹ plasma) to prevent proteolytic breakdown of any peptides that might be present. These samples were transported on ice and immediately centrifuged at 3000 gat 4 °C for 15 min. After centrifugation, each sample was extracted with 100 % ethanol (1:2 dilution), vortexed, and allowed to stand at 4 °C for 30 min (Cousins and Farrell, 1996). Radioimmunoassays to measure peptides from the N terminus of the ANF 126-amino-acid prohormone were developed to amino acids 1–30 (proANF 1-30) and 31–67 (proANF 31-67) of this prohormone as previously described by Winters et al. (1989) and Vesely et al. (1989, 1994). For each RIA, the extracted plasma and/or perfusate sample was first reconstituted in 100 μl of 0.1 mol l⁻¹ phosphate buffer (pH 7.4) containing 0.05 mol l⁻¹ NaCl, 0.1 % bovine serum albumin, 0.1 % Triton X-100 and 0.01 % NaN₃. To the redissolved sample, 100 μl (0.03 mg) of rabbit IgG plus 100 μl of the respective antisera were added and the sample was incubated for 24 h. Then, 100 μl of ¹²⁵I-labeled peptide (10 000 cts min⁻¹) was added and the mixture was vortexed and incubated for 18 h at 4 °C. The antibody-bound tracer was then precipitated by adding 100 μl of goat anti-rabbit globulin and incubating this mixture for 2 h at room temperature (25 °C). Each tube was then centrifuged at 3000 g for 20 min. The supernatant was carefully aspirated off the remaining pellet. Each of these assays was performed in triplicate. The inter-assay coefficient of variation for the RIAs was 4.8 % for proANF 1-30, 5.3 % for proANF 31-67 and 5.7 % for ANF. The intra-assay coefficient of variation was 8.0 % for proANFs 1-30 and 31-67; ANF intra-assay variation was 6.9 %. Recovery was examined by adding synthetic unlabeled proANFs 1-30 and 31-67 and ANF at 100, 200 and 400 pg ml⁻¹ to pooled plasma. Recovery for proANF 1-30 was 83.5±13.2 %, for proANF 31-67 100.9±8.9 % and for ANF 92.0±11.0 % (means ± S.E.M., N=10).

The molecular forms of the immunoreactive proANF peptides and ANF in trout plasma were determined using high-performance gel-permeation chromatography (HP-GPC). After centrifugation (3000 g) for 30 min, 10 ml of trout plasma was extracted with 100 % ethanol (1:2 dilution) and then dried in a Speed-Vac A160 concentrator (Savant Instruments, Inc., Farmingdale, NY). The plasma extracts were then resuspended in 100 μl of HP-GPC column mobile phase (10 mmol l⁻¹ trifluoroacetic acid containing 0.3 mol l⁻¹ sodium chloride and 30 % acetonitrile) for HP-GPC assay (as described previously in Vesely et al. 1994). These 100 μl samples were applied to a TSK-Gel G2000SW column [7.5 mm×600 mm equipped with a 7.5 mm×75 mm guard column (Toyo Soda, Tokyo, Japan)] and eluted isocratically with HP-GPC column mobile phase at a flow rate of 0.3 ml min⁻¹. Fractions (0.3 ml) were collected, dried and then assayed for proANFs 1-30 and 31-67 and ANF by their respective RIAs described above. Blue Dextran (2×10⁶ Da) was used to determine the void volume and p-aminohippuric acid (216 Da) the total volume of the column. To calibrate the column, the following molecular markers were used: carbonic anhydrase (29 000 Da), myoglobin (16 900 Da), cytochrome c (12 384 Da), vasoactive intestinal peptide (3300 Da) and (try¹)-somatostatin (1700 Da). The elution positions of proANF 1-30, proANF 31-67 and ANF were determined using ¹²⁵I-labeled synthetic peptides. Recoveries of these labeled peptides were as follows: 77 % for ANF, 74 % for proANF 1-30 and 82 % for proANF 31-67. High-performance gel-permeation chromatography was repeated three times for each of the RIAs.

If there were no changes in ANF release between the FW-and SW-acclimated groups, then step increases in filling pressure would result in similar increases in ANF release. Statistical comparisons were, therefore, made between the rates of ANF release from FW-and SW-acclimated groups. A repeated-measures three-way analysis of variance (ANOVA) was used to compare the mean values for the rate of ANF release at each PI level (Winer et al. 1991) with significant differences set at the 95 % confidence level (P<0.05). The test trials were run from a double Latin square design such that each treatment was preceded by every other treatment. This design controlled for the possibility of a cross-over effect (Jones and Kenward, 1989). All cardiovascular and ANF release mean values were subjected to multiple comparisons using a single-factor ANOVA with the Student–Newman– Keuls method (SigmaStat Statistical Analysis System V.1, Jandel Corporation). In addition, statistical comparisons of morphometric mean values between the freshwater and seawater groups were carried out using a Student’s t-test. Significant differences between the mean values were tested at the 95 % confidence level.

The morphometric comparisons between the FW-and SW-acclimated rainbow trout are presented in Table 1. There was a significant difference in the mean values between the groups for body mass (P=0.032). However, there were no significant differences found between the group means for fork length (P=0.350), atrial mass (P=0.294) or ventricular mass (P=0.930). In addition, there were no significant differences noted between the treatment groups for the percentage of relative ventricular (P=0.066) or relative atrial (P=0.795) mass. In each treatment group, the sex was predominantly female. The freshwater group contained five females and one male, while the seawater group contained four females and two males.

ProANF 1-30 and proANF 31-67 were present in plasma (Fig. 1) and in the perfusate ejected from the perfused trout heart (Table 2). The ANF plasma content in the SW-acclimated group was slightly higher (37.20±2.80 pg ml⁻¹), but not significantly different (P=0.164) from that of the FW group (30.63±3.19 pg ml⁻¹). Both proANF 1-30 and 31-67 were found at plasma concentrations that were significantly higher than that for ANF (P=0.005; Fig. 1). In the FW group, the plasma contents of proANFs 1-30 and 31-67 were not significantly different (P=0.563); however, in the SW-acclimated group, the plasma content of proANF 1-30 was significantly higher than that of proANF 31-67 (P=0.001).

At the control filling pressure, the rates of ANF release in the FW-and SW-acclimated groups were similar (Table 2). These values are similar to the control value for freshwater trout reported by Cousins and Farrell (1996) at 30.68±4.76 pg min⁻¹ g⁻¹ wet heart mass. ProANFs 1-30 and 31-67 were released at similar basal rates in both the FW-and SW-acclimated groups (Table 2). However, the basal release rate of proANF 1-30 was approximately 10 times higher (P<0.05) than that of ANF, and the basal release rate of proANF 31-67 was approximately 20 times higher (P<0.001) than the basal release rate of ANF (Table 2).

In the FW group, when the filling pressure to the perfused hearts was increased to elicit Q̇_max, the release rates of proANFs 1-30 and 31-67 and ANF all increased by approximately fivefold compared with their basal rates of release (Table 2). Furthermore, randomized step increases in filling pressure, lasting 10 min each, resulted in proportional increases in both the stroke volume and the rate of ANF release (Table 3; Fig. 2). For each step increase in filling pressure, the rate of ANF release immediately increased to a new and higher level. At the highest filling pressure, the perfused hearts from the SW-acclimated group released proANF 1-30, proANF 31-67 and ANF at significantly higher rates (P=0.006) than those of the FW group (Table 2). Thus, the release of cardiac hormones occurred over the full range of filling pressures and was proportional to Q̇.

Similar findings were made for the SW-acclimated group. However, increasing the filling pressure to elicit Q̇_max caused an even larger release (6-to 9.5 times their basal release rate) of proANFs 1-30 and 31-67 and ANF (Table 2) than in the FW group. For instance, at the highest filling pressure, the amount of ANF secreted from a SW-acclimated trout heart in 1min was 4.20 pg ml^-1 perfusate, whereas the amount of ANF secreted from a FW-acclimated trout heart was 3.67pgml^{_ 1} perfusate. In addition to a higher maximum ANF release rate, the perfused hearts from the SW-acclimated group had significantly higher cardiac outputs (P=0.016) and stroke volumes (P=0.022) than those found for the FW group (Table 3). In the FW group, an increase in the filling pressure from the control level (0.03±0.04kPa) to the highest value (0.58±0.03kPa) significantly increased stroke volume, Q̇ and power output by approximately threefold (P<0.05; Table 3). In the SW-acclimated group, a similar increase in filling pressure (from 0.02±0.02 to 0.55±0.02kPa) also significantly increased stroke volume, Q̇ and power output (P<0.05; Table 3), but to levels that were 10-15 % higher than those found for the FW group.

A typical standard curve for the proANF 1-30 RIA and the good parallelism of the circulating immunoreactive N-terminal ANF prohormone-like peptide(s) recognized by this assay in trout plasma, as well as in human plasma, is shown in Fig. 3.

Evaluation of this assay revealed that it is sensitive, with an EC₅₀ of 180 fmol per tube. The proANF 1-30 RIA is also specific, showing less than 0.5 % cross-reactivity with either proANF 31-67 or ANF (Fig. 3). The antiserum to proANF 1-30 had no detectable cross-reactivity with oxytocin, vasopressin, somatostatin, insulin or adrenocorticotrophic hormone.

The excellent parallelism of circulatory fluid concentrations of proANF 31-67 in trout and human plasma with a typical proANF 31-67 RIA standard curve is illustrated in Fig. 4. The proANF 31-67 RIA can detect as little as 35 fmol of synthetic proANF 31-67, and its EC₅₀ of 120 fmol per tube shows that it is a sensitive assay. It is also specific, as shown by the dilution curves with proANF 1-30 and ANF (Fig. 4), showing that the cross-reactivity with these two peptides is less than 0.5 %. In addition, the proANF 31-67 antiserum, like the proANF 1-30 antiserum, had no detectable cross-reactivity with oxytocin, vasopressin, somatostatin, insulin or adrenocorticotropic hormone.

The addition of 100, 200, 400 and 800 μl of trout and human plasma also revealed very good parallelism for circulating ANF-like peptides with the ANF RIA standard curve, as observed in Fig. 5. The ANF RIA had less than 0.5 % cross-reactivity with either proANF 1-30 or proANF 31-67 (Fig. 5) and is a sensitive assay able to detect as little as 1.4 fmol of ANF and with an EC₅₀ of 11 fmol per tube.

Human synthetic proANF 1-30 was found in 0.3 ml fractions 70–74, with fraction 71 having the predominant peak when examined by HP-GPC (Fig. 6). The human synthetic form of proANF 31-67 was found in 0.3 ml fractions 67–73, with the major peak in fraction 69 (Fig. 6). The human synthetic form of ANF was in 0.3 ml fractions 74–79, with the peak in fraction 76 (Fig. 6). The elution profiles of [¹²⁵I]proANF 1-30 (peak of fraction 71), [¹²⁵I]proANF 31-67 (peak at fraction 70) and [¹²⁵I]ANF (peak of fraction 76) were nearly identical to those of the unlabeled peptides.

Analyses of trout plasma by HP-GPC followed by RIAs to amino acids 1–30, 31–67 and 99–126 (i.e. ANF) of the ANF prohormone indicated that proANFs 1-30, 31-67 and 1-98 each circulate in the blood (Fig. 7). The RIA for proANF 1-30 revealed that this assay recognized a peptide with a molecular mass of 9800 Da. This finding is consistent with the 98-amino-acid N terminus of the prohormone without an attached C terminus (i.e. ANF). This assay also had a peak where the human synthetic form of proANF 1-30 elutes (Fig. 7A), suggesting that proANF 1-30 exists in trout plasma as a separate entity from proANF 1-98. From the peaks shown in Fig. 7A, one can observe that approximately 50 % of the material recognized by the proANF 1-30 assay actually is proANF 1-30, while the other 50 % of the material recognized by the probe is circulating proANF 1-98. Similar HP-GPC evaluation of trout plasma followed by proANF 31-67 RIA of 0.3 ml fractions revealed only one peak where the human sequence of proANF 31-67 elutes (Fig. 7B). Neither proANF 1-98 nor the ANF prohormone (i.e. proANF 1-126) was immunologically recognized in trout plasma by the proANF 31-67 assay. Evaluation of trout plasma by HP-GPC followed by ANF RIA also revealed only one peak, which eluted identically with the pure synthetic form of ANF (Fig. 7C).

This is the first study to compare the effects of atrial stretch on ANF release rates in freshwater-and seawater-acclimated rainbow trout, and is only the second study to report ANF release rates in fish. The basal release rate for ANF from FW-acclimated perfused hearts is comparable to that reported earlier for the same type of preparation (Cousins and Farrell, 1996). We found a higher rate of release in the SW-acclimated group than in the FW group for the same level of cardiac filling pressure at the maximum cardiac output and, furthermore, the difference was significant relative to the overall changes that are possible in the ANF release rate. In the FW group, ANF release increased 5.8-fold (from 33.9 to 195.8 pg min⁻¹ g⁻¹ wet heart mass), whereas in the SW-acclimated group ANF release increased 9.5-fold (from 30.0 to 285.4 pg min⁻¹ g⁻¹ wet heart mass). Part of this difference can be attributed to the fact that the SW-acclimated group had a significantly greater Q̇_max (Table 3). Even though the ventricular mass (Table 1) and filling pressures were not significantly different between the FW-and SW-acclimated groups (P=0.930 and P=0.384, respectively), the SW-acclimated hearts still had greater stroke volumes (Table 3). Nevertheless, the SW-acclimated group released more ANF per milliliter of perfusate passing through the heart relative to the FW group (4.20 pg ml⁻¹versus 3.67 pg ml⁻¹). The consequence of this higher release rate is that the trout hearts in the SW-acclimated group were more sensitive to filling pressure in terms of both ANF release and the Starling response (Fig. 2). In fact, for a 0.1 kPa change in filling pressure, the change in ANF release for the SW-acclimated group (12.78 pg min⁻¹ g⁻¹ wet heart mass) is almost twice the value for the FW group (7.23 pg min⁻¹ g⁻¹ wet heart mass). We have no definitive explanation for these differences. We do know, however, that Thorarensen et al. (1996) reported higher in vivo maximum Q̇ measured during swimming in SW than in FW trout. Our Q̇_max va lues for FW trout are virtually identical to the maximum Q̇ reported by Kiceniuk and Jones (1977) (53.4 ml min⁻¹ kg⁻¹ body mass) during prolonged swim challenges. Thus, it appears that the acclimation to sea water somehow caused the hearts to have an increased sensitivity to the higher filling pressures, which in turn resulted in an increase in ANF release from the heart when the myocytes were stretched to the peak of the Frank–Starling curve. This finding also suggests that ANF release in the perfused trout heart was driven more by volume changes than by pressure changes.

By comparing the osmotic and volume challenges faced by FW-and SW-acclimated trout, we can assess the significance of the greater release rate of ANF and the heightened sensitivity of release to the major release stimulus (atrial stretch). Marine fish are in hyperosmotic environments and are in danger of hypovolemia (i.e. low blood volume). Freshwater fish have serum Na⁺ and Cl⁻ contents above that of their environment, subjecting them to the danger of hypervolemia (i.e. high blood volume). Euryhaline fish, such as the adult rainbow trout used in this study, can inhabit both freshwater and seawater environments while maintaining relatively constant blood volumes (Olson, 1992) and arterial blood pressures. The ventral aortic blood pressures of freshwater rainbow trout are reported to be only slightly higher (5.33 kPa, Stevens and Randall, 1967) than those of seawater-acclimated trout (4.55 kPa; Thorarensen, 1994). Since ANF is known to be released from the trout heart in response to elevated cardiac filling pressures (Cousins and Farrell, 1996) and given its known vasoactive, natriuretic and diuretic activities (Duff and Olson, 1986; Olson and Meisheri, 1989; Olson and Duff, 1992, 1993; Jensen and Olson, 1994), it seems reasonable to assume that ANF is intimately involved in the osmotic, ionic and blood volume regulation of the animal. As soon as ANF is released from the trout heart, it circulates through the primary osmotic and ionic regulating organ of the animal, namely, the gills. In freshwater trout, ANF-clearance gill receptors have been shown to remove at least 60 % of an arterial injection of synthetic ANF (Olson and Duff, 1992, 1993). Since SW-acclimated trout are continually faced with water loss and excess salts, their higher rates of ANF release may be related more to a natriuretic effect than to a diuretic effect. Whether the higher ANF release rate is also related to the slightly lower arterial blood pressures through vasoactive effects will require further studies to resolve.

The high-performance gel-permeation chromatography (HP-GPC) evaluation of fish plasma followed by ANF radioimmunoassay revealed that an ANF-like peptide in fish plasma had identical elution characteristics to those of the pure human sequence of ANF (Fig. 7A), indicating that ANF in trout is very similar to the human form of ANF. The amino acid sequence of ANF is known to be highly conserved, showing a 93 % homology when five species were compared (Vesely, 1992). This HP-GPC evaluation further revealed that the ANF assay recognized ANF in fish plasma and was not measuring ventricular natriuretic peptide (VNP). Since trout VNP is a 35-amino-acid peptide with a molecular mass of 3860 Da (Takei et al. 1994) and no peptide eluted at this molecular mass when the trout plasma was subjected to HP-GPC followed by the ANF radioimmunoassay, it appears that the human ANF assay does not immunologically recognize VNP in the trout plasma. It should also be noted that, using this evaluation, the ANF prohormone, which elutes at a molecular mass of 12 900 Da, does not circulate in the trout plasma (see Fig. 7). A similar finding has been reported in humans (Vesely et al. 1994) where ANF, proANF 1-30 and proANF 31-67 circulate, but the prohormone form from which they are derived does not.

In the present investigation, levels of circulating proANFs 1-30 and 31-67 in the plasma of a fish were measured for the first time. To assess whether the trout peptides identified by the proANF 1-30 and 31-67 RIAs were similar to the proANF 1-30 and 31-67 peptides in human plasma, the fish plasma was diluted and compared with human plasma using standard binding curves for these peptides. The excellent parallelism of trout plasma with the proANF 1-30 (Fig. 3) and proANF 31-67 (Fig. 4) binding curves suggests that the assays are specific for these peptides in trout plasma and that the amino acid sequences of the peptides in trout are similar to the sequences in humans. If the fish and human amino acid sequences were dissimilar, these peptides would not be recognized by antisera raised to the human sequences of proANF 1-30 and proANF 31-67. This is also true for ANF in trout plasma, which closely parallels the human ANF standard curve (Fig. 5). To characterize further proANF 1-30 and proANF 31-67 in fish plasma, HP-GPC of the plasma was followed by proANF 1-30 and proANF 31-67 RIAs. This technique revealed that trout plasma was contained a peptide that had identical elution characteristics to those of the pure human amino acid sequence of proANF 1-30. The HP-GPC evaluation of the trout plasma followed by a proANF 31-67 RIA also showed that only one peptide showed an immunological reaction and this co-eluted with the human form of proANF 31-67 (Fig. 7B). The proANF 31-67 assay does not immunologically recognize the N terminus of the ANF prohormone in trout plasma, and neither the proANF 1-30 nor the proANF 31-67 assay recognizes the ANF prohormone in the fish plasma, strongly suggesting that the ANF prohormone does not circulate in trout.

ProANFs 1-30 and 31-67 and ANF were found to be released proportionately with increasing filling pressure in perfused trout hearts from both the FW-and SW-acclimated groups, and the size of the increase at maximum filling pressure was similar for ANF, proANF 1-30 and proANF 31-67. Thus, it appears that an increase in circulating blood volume is important for the release of these cardiac peptide hormones in both freshwater and seawater fish. This parallels events in humans, where an increase in central blood volume simultaneously releases proANFs 1-30 and 31-67 and ANF (Vesely et al. 1989) and the relative increase in each is similar to that found in the FW trout. This also occurs in other mammals, where it has been found that an increase in atrial pressure of only 0.27–0.53 kPa in isolated rat hearts simultaneously releases proANFs 1-30 and 31-67 and ANF (Dietz et al. 1991). Thus, it appears that similar mechanisms for the cardiac release of proANFs and ANF may operate for trout and mammals.

The release of higher levels of ANF peptides in the SW-acclimated group could have been caused by an increase in plasma [Na⁺] and [Cl⁻] in those animals, which somehow improved their myocardial performance at the higher filling pressures. The enhanced release of these cardiac peptides in the SW-acclimated trout exposed to high salt concentrations does have a theoretical advantage in that proANFs 1-30 and 31-67 and ANF are all potent stimulators of enhanced Na⁺ excretion in mammals (Martin et al. 1990; Vesely et al. 1994) and may also have this role in fish. In this context, it has previously been shown in other euryhaline animals, such as the oyster (Crassostrea virginica), that the amount of salt in the external environment directly regulates the release of proANFs 1-30 and 31-67 and ANF, with increasing concentrations of peptides being released as Na⁺ levels in their environment rise (Palmer et al. 1994).

Our findings for the effect of filling pressure on proANF release in trout are consistent with previous findings in mammals where the cardiac release of proANFs and ANF is enhanced by increasing central venous volume. Furthermore, the proportional increase in release of proANF and ANF is the same, even though basal release rates are quite different. These proANF peptide hormones, which cause prolonged natriuresis and diuresis compared with ANF in humans (Vesely et al. 1994), were found at threefold higher concentrations than ANF. A similar ratio of proANF to ANF concentration has been found in the hemolymph circulation of American oysters and blue crabs (Callinectes sapidus; Poulos et al. 1995), where proANF 31-67 levels are 2.5-fold (oyster) and fourfold (crab) higher than ANF levels. In mammals, the [proANF]:[ANF] ratio is even higher; the concentration of proANF 31-67 in the circulation is approximately seven times higher than the concentration of ANF in humans (Winters et al. 1989), dogs (Ngo et al. 1989) and rats (Martin et al. 1990). Whether variations in the [proANF]:[ANF] ratio are adaptations to living in aqueous as opposed to terrestrial habitats is not yet clear but, obviously, both terrestrial and seawater vertebrates face a similar danger of dehydration.

In summary, this study compares measurements of ANF release rates from the hearts of freshwater-and seawater-acclimated rainbow trout and demonstrates that ANF-related cardiac peptides (proANF 1-30 and proANF 31-67) are released from the fish heart and circulate in the blood at higher concentrations than ANF. The release of ANF from the perfused trout heart in situ was found to be a graded response to step increases in filling pressure. In both FW-and SW-acclimated trout, the ANF release rate changed by several orders of magnitude in direct proportion to increases in filling pressure, but this response was greatly enhanced in the SW-acclimated animals. The rate of release of both the proANFs 1-30 and 31-67 also increased when the filling pressure was increased to the maximum cardiac output. The sensitivity of ANF release in SW fish to filling pressure, however, was heightened compared with that of FW trout. It appears, therefore, that the secretion of proANFs and ANF in the trout heart may be regulated more by changes in volume than by changes in pressure.

We wish to thank Francois Bellavance for his statistical advice. This work was supported by an operating grant from the Natural Sciences and Engineering Research Council of Canada to A.P.F. and by a research award (Simon Fraser University Graduate Fellowship) to K.L.C.