Purification and biological characterization of chinook salmon prolactin.

Prolactin from chinook salmon pituitaries was purified by acid acetone extraction, saline precipitation, chromatofocusing, and gel filtration. This procedure allowed us to recover highly purified prolactin as demonstrated by the presence of a single NH2-terminal amino acid and a single band in sodium dodecyl sulfate gel electrophoresis. Chinook salmon prolactin appeared to be a basic protein of 22,500 molecular weight. Throughout the purification, prolactin bioactivity was followed by radioreceptor assay for lactogenic hormones. The prolactin character of the purified protein was established by its lactogenic activity in the rabbit mammary gland in vitro and its sodium-retaining activity in hypophysectomized Fundulus heteroclitus.

It is now well established that the teleost pituitary contains a protein which is similar to mammalian prolactin and which is endowed with osmoregulatory activity (see reviews by Ensor, 1979;Clarke and Bern, 1980). Teleost prolactin is devoid 'of lactogenie activity as judged by the pigeon crop sac and mouse mammary gland bioassays (Nicoll ef al., 1966;Nicoll and Bern, 1968;Doneen, 1976;Farmer et al., 1977). However, Pnmet et al. (1979) and Houdebine et al. (1981) have recently observed that salmon and tilapia prolactins can induce casein synthesis in rabbit mammary gland in culture. These findings suggest that the lactogenic activity expressed in the rabbit and osmoregulatory activity expressed in various euryhaline teleosts may be two properties of teleostean prolactin. ,In a previous study (Prunet et al., 1979), partial purification of chinook salmon was carried out using the radioreceptor assay for lactogenic hormone as a test. In this communieation we discuss the purification and the characterizatibn of prolactin from' chinook salmon (Qncor&y&hus tschawyts:cha). To detect prolactin activity the radioreceptor assay was used throughout the purification" The final product exhibited the lactogenic activity previously observed in the partially purified fractions (Prunet et ai,, 1979). In addition, the purified hormone proved to have an osmoregulatory activity as judged by the Fundulus bioassay (see Crau et al., 1983).

Hormone Isdation
Salmon pituitary collection. Pituitary glands from female and male chinook salmon (0. tschawytscha) which had been in freshwater fur several weeks at the Spring Creek National Fish Hatchery (Underwood, Wash.) were collected following decapitation and were immediately frozen in liquid nitrogen and lyopbihzed.
Osmoregulatory activity. The sodium-retaining activity of this hormone was assessed in hypopbysectomized Fundulus heteroclitus transferred in freshwater according to the techniques of Grau er al. (1983).

Biochemical Analysis
Several analytical electrophoresis systems were used during. this purification. Anionic electrophoresis at pH 9.3 was performed in discontinuous polyacrylamide gel with a spacer gel of 3% acrylamide and a running gel of 7.5% (Maurer, 1968). Cationic electrophoresis at pH 4.5 was performed on 15% acrylamide gel according to the method of Riesfeld et aZ. (1962). Electrophoresis in the presence of sodium dodecyl sulfate (SDS) was done according to Weber and Osborn (1969). In all cases proteins were stained with 0.2% (w/v) Coomassie blue in 50% methanol solution and destained overnight in 7% acetic acid. Using these analytical techniques purified chinook salmon prolactin was compared with chum salmon prolactin purified by Kawauchi et al. (1983) and with homogenate of rostra1 pars distalis from coho salmon pituitaries.
To evaluate the purity of the final salmon prolactin preparation, 200 pg of this material was submitted to NH*-terminal analysis by the dansyl procedure (Gray 1967;Woods and Wang, 1967). The amino acid composition was determined by the method of Spackman et al. (1958) using an automatic amino acid analyzer. These analyses were performed by Professor H. Papkoff (Hormone Research Laboratory, University of California, San Francisco).
The electrofocusing experiment was monitored on a flat-bed FBE 1000 (Pharmacia). The agarose gel (1%) with a volume of 28 ml contained Biolyte ampholytes (2%, v/v) in the pH range 8-10. The cathode and anode solutions were, respectively, 1 N NaOH and 0.2 M histidine. Following 1 hr of prefocusing at 5 W, the material was focused for 3 h at 5 W.
First purification protocol tested (Idler et al., 1978;Ng et al., 1981). The alkaline extract from 10 g of lyophilized pituitary glands, using 0.05 M Tris-HCl, pH 9.5, 0.5 M NaCl buffer, was chromatographed on concanavalin A-Sepharose (Pharmacia) in two successive runs. The nonglycoprotein material flowing through was layered on an Ultrogel AcA 54 (IBF) column (Pharmacia K 50-100). The bioactive proteins corresponding to a moiecular weight of 22,000 were rechromatographed on the same gel. The resulting enriched prolactin fraction was subjected to ion-exchange chromatography (DEAE Bio-Gel A (Biorad)-equilibrated with 0.005 M Tris-HCl, pH 9.5). The nonabsorbed material which contained prolactin was submitted to a final gel filtration using Sephadex G-75 in 0.05 M NH4HC0, and the resulting fractions obtained were lyophilized.
Second purification protocol. This protocol followed initial acid acetone extraction indicated by Kawauchi et al. (1983). The acid acetone powder thus obtained was dissolved in water at pH 2.5 and NaCl was added to a final concentration of 10% NaCl saturation (Cole and Li, 1955;Kawauchi and Tubokama, 1979). The solution was stirred for 12 hr at 4" and centrifuged for 1 hr at 100,OOOg. The resulting supernatant was dialyzed against distilled water and lyophilized. This fraction was then layered on a Polybuffer exchanger gel, PBE 94 (Pharmacia), equilibrated with 0.025 M ethanolamine-HCl, pH 10.2. The elution was carried out with a linear pH gradient using 0.025 M ethanolamine-HCl buffer from pH 10.2 to pH 9.3. After dialysis and lyophilization the bioactive fraction was submitted to a final gel filtration on Sephadex G-50 in 0.05 M NH4HC03.
The fraction containing prolactin was then lyophilized.

Bioassays
Radioreceptor assay. The presence of salmon prolactin activity during purification was determined using a radioreceptor assay for lactogenic hormone initially proposed by Shiu et al. (1973) and also used for fish hormone (Prunet et al., 1977(Prunet et al., , 1979. The lack of parallelism between the displacement curves generated with ovine prolactin and with the salmon prolactin-containing fractions in the presence of 1z51-labeled ovine prolactin renders uneasy an accurate appreciation of the fish prolactin concentration. For the sake of convenience, the biological activity (kg of ovine prolactin/mg of protein) of each sample was estimated by finding the yield of the ovine prolactin concentration and the sample protein concentration, calculated by the displacement curves and both inhibiting 50% of the '2SI-labeled ovine prolactin binding.
Lactogenic activity. The lactogenic activity was evaluated by estimating its capacity to induce casein synthesis in rabbit mammary explants. The culture technique was similar to that presented in previous reports (Prunet et al., 1979;Houdebine et al., 1981). The ovine prolactin was the generous gift of the National Institutes of Health (NIH-P-S13). One major difference has been introduced in the estimation of casein synthesis. Instead of measuring the capacity of the tissue to incorporate labeled amino acids into casein, the content of the explant in p-casein was evaluated using a radioimmunoassay. Rabbit p-casein.was purified by two successive chromatographies on DEAEcellulose (Testud and Ribadeau-Dumas, 1973). Antirabbit @-casein was obtained in goat, and anti-goat immunoglobulin was obtained in rabbit. After a culture of 1 day in the presence of ovine prolactin or salmon prolactin, the explants (50 mg of tissue) were homogenized in 4 ml of 0.14 M NaCl, 0.01 M sodium phosphate @H 7.6), and 0.1% Triton X-100. To 50 ~1 of the SOOQg supernatant 30,000 cpm '251-labeled p-casein, labeled by a chloramine-T procedure, at a specific activity of 50-75 pS!i/ug and 200 ~1 of the goat anti-pcasein serum, diluted 2000-fold, were added. The final volume was adjusted to 600' pl with phosphate buffer.
After 24 hr ai room temperature, 200 p,l of a S-fold diluted rabbit serum containing anti-goat immunoglobulin antibodies was added to the incubate and the mixture was further incubated for 24 hr at room temperature. The immunoprecipitate was sedimented at 2200g and the radioactivity of the pellet was measured with a gamma counter.

Isolation of the Hormone
The first purification procedure led to a fractionation similar to that described previously by Idler et al. (1978). Chinook salmon prolactin, which was not retained on concanavalin A-Sepharose, eluted from gel filtration as a protein having a molecular weight of 22,000. Moreover, this protein was not retained on DFAE-ionic exchanger at pH 9 and low salt concentration (0.005 M Tris-HCl) . However, the prolactin preparation thus obtained did not seem to be pure as indicated by the asymetric peak obtained after Sephadex G-75 gel filtration (data not shown)., The low recovery of prolactin material (4 mg) after this gel filtration chromatography did not allow an additional chromatography step. The heterogeneity of the prolactin fraction obtained was confirmed by the presence of several N-terminal amino acids and by the presence of four bands in the electrophoresis trial performed at pH 4.5 (the main band having a Rf = 0.5) and that performed in the presence of SDS (Prunet, 1981). Moreover, this preparation did not give any clear band after an electrophoresis at pH 9.3. A preliminary .study of this material using electrofocusing techniques confirmed the basic pl of this protein (pZ > 8). These problems led us to question whether the saline an alkaline extraction procedure was suitable for this purification.
The second protocol based on acid acetone extraction gave rise to a recovery of 335 mg of powder from 5 g of lyophilized pituitaries. After saline precipitation at pH 2.5, 96% of the prolactin activity detected by the radiore.ceptor assay remained in the supernatant (Table 1). This crude prolactin preparation (80 mg) was layered on a chromatofocusing gel. The linear pM gradient eluted the, proteins having prs between 9 and 10. To obtain'the pR gradient described under Materials and Methods, the Polybuffer (P&E) 94 recommended by Pharmacia was replaced by a 0.025 M ethanolamine-HCl buffer. These conditions led TV a good recovery of proteins from the eluted fractions, a ,problem which proved to be difficult to solve in the presence of Polybuffer. About 87% of the prolactin activity was concentrated in the Fb fraction (Fig. 1). Chromatography on Sephadex G-5Q of this fraction led to separate the prolactin activity (first symetric peak> from sm!all,er contaminants , (Fig. 2)'. The prolactin preparation thus obtained wa.s further demonstrated, to be pure. Surprisingly, this chromatography step did not increase the biological activity measured:#in the radioreceptor asgay of the prolactin ~r,~~~~~~i~~ (Ttible 1):. This might be ,due to the ~~~'~~~~d of this gel filtration exp,eriment, Such a lo&v recovery of m&&d hasibeen ob.served,previously when very low quantities of pro- Note. The displacement curves obtained in the radioreceptor assay for each fish sample and for ovine prolactin are made linear by log-log& transformation. The biological activity of the different fractions is expressed in ovine prolactin equivalent per milligram of protein (for details see Material and Methods). Amount of ovine prolactin equivalent is obtained after multiplying the biological activity by the protein amount of the fraction.

53
phoresis of the rostra1 pars distalis of coho salmon pituitary glands revealed the existence of a major protein running at the same position as the purified chinook salmon prolactin (Fig. 3) and the prolactin fraction obtained using the first purification protocol: this band appeared to have sodiumretaining activity in hypophysectomized F. heteroclitus, a biological property which is specific to prolactin (Grau et al., 1983). Moreover, chum salmon prolactin kindly provided by Professor H. Kawauchi of Kitasato University also gave the same electrophoresis pattern as our chinook salmon prolactin (Fig. 3). When electrophoresis was performed at pH 4.5, essentially a single major band appeared (Rf = 0.5) with a second band of lower intensity running immediately behind (Fig. 3). The p1 of chinook prolactin determined by electrofocusing analysis confirmed its basic value (~1 = 9.4).
The comparison of the amino acid composition of chinook prolactin with tilapia prolactin indicated that the former has a lower content of glutamine and. glutamic acid and a higher content of lysine and arginine than the latter (Table 2). This finding is compatible with the high p1 of chinook salmon prolactin.
Except for alanine and valine which were higher in Tifiapia prolactin, the number of other residues was similar for both hormones.

Biological Properties
Lactogenic activity. Throughout the purification, prolactin bioactivity estimated by the radioreceptor assay progressively increased. The purified chinook salmon prolactin *as estimated to be ,24-fold mom potent than the crude pituitary extract (Table  1). However, considering the heterologous nature, of this assay, such biological @rantification should be viewed with some caution, It seemed necessary therefore to confIrm the lactogenic activity of the pure hormone in the mammary gland culture system. Sev- Electrophoresis of a-d was performed in the presence of SDS whereas e was carried out at pH 4.5 in the absence of the detergent. era1 concentrations of ovine or chinook salmon prolactin were added to the culture medium of rabbit mammary explants. Figure 4 indicates that the maximum response was obtained between 100 and 1000 rig/ml ovine prolactin with reduced responses at higher concentrations.
This result is in agreement with previous experiments (Houdebine et al., 1981;Djiane et al., 1982). Thus it may be concluded that evaluation of lactogenic activity of prolactin can be evaluated equally well by measuring either the biosynthesis of total casein (the previous method) or the accumulation of casein. The purified chinook salmon prolactin exhibited a clear lactogenie activity which was roughly 20 times lower than the activity of ovine prolactin.
In the Fundulus bioassay, depicted by Grau et al. (1983), chinook salmon prolactin showed maximum activity for injected dose as low as 4 rig/ml; thus our chinook prolactin preparation appeared to be approximately 100 times more potent than ovine prolactin.

DISCUSSION
Chinook salmon prolactin has been obtained in a highly purified form as judged by the presence of a single NH2-terminal amino acid and by its migration as a single band in electrophoresis in presence of SDS. The prolactin character of this purified material was established by its lactogenic activity in organ-cultured rabbit mammary gland and by its sodium-retaining activity in F. heteroclitus.
The first purification procedure adapted from Idler et al. (1978) and Ng et al. (1980) did not allow us to get pure chinook salmon prolactin. However, we obtained a preparation in which chinook salmon prolactin was the main protein as indicated by elec-  4. Induction of S-casein accumulation in rabbit mammary explants by ovine and salmon pro-la&in. In all cases insulin was present at the concentration of 5 ug/ml. The concentration of j3-casein estimated by a radioimmunoassay is expressed in nanograms of @casein per milligram of mammary explant. Each determination is the mean of three measurements. trophoresis and radioreceptor assay results m The second protocol seems particularly appropriate since only a few purification steps led to a highly purified hormone, One drawback remains the low yield of the final product (70 mg/kg). This is much lower than the yield obtained by Kawauchi et al. (1983) with chum prolactin (1 g/kg). This discrepancy could be explained by differences in osmoregulation physiology between these two species (Hoar, 1976;Weisbart, 1968). However, the development of another chromatofocusing system better adapted to a very basic pH is presently being developed in our laboratory, in order to hopefully obtain a better recovery of prolactin.
Chinook salmon prolactin appeared to be a basic protein according to the electrofocusing experiments, a property which is also shared by chum salmon prolactin (Idler, 1981;Kawauchi et al., 1983). This result is confirmed by the behavior of chi-nook salmon prolactin on a chromatofocusing gel equilibrated at a basic pH and. in electrophoresis performed at alkaline ~ p&; the absence of a stained band in this last technique is not associated with possible staining problem (Farmer et al., 19771, as this prolactin band can be clearly stained in the same electrophoresis system when polarity is reversed (Rf < 0.1).
The comparison between chinook salmon prolactin and chum salmon prolactin purified by Dr. H. Kawauchi shows @at these two prblactins have similar or nearli identical biochemical characteristics. They run the same .way in electrophoresis performed in presence of SDS, they are !both basic proteins and their amino acid compositions are essentially in agreement ~ (Kawauchi et al., 1983; see Table 2). This ilsnot surprising if #we consider the close phyllogenetic relationship between theses tw~o salmon species. As Tilapia prolactinj chknook 'salmon prolactin :has a molccul& weight smaller than that of mammalian bk& lactin and possesses four half-cystine resi-lished with rabbit mammary gland recepdues or two disulfide bridges. The simi-tors as a bioindicator for fish prolactins, larity between these two fish prolactins is and to quantify the bioactivity of the preppartially confirmed by the comparison of aration with the Fundulus bioassay. their amino acid compositions.
In electrophoresis carried out at pH 4.5 ACKNOWLEDGMENTS chinook salmon prolactin migrated as a We are grateful to Dr. E. G. Grau, R. S. Niskioka, major band preceded by a less intense and S. Steiny for their help and to Professor H. A. band, a picture strikingly reminiscent of the