THE JOURNAL OF EXPERIMENTAL ZOOLOGY 276403-414 (1996) Partial Purification and Characterization of Serum Embryotrophic Factor Required for Early Postimplantation Growth of Rat Embryos in Culture MAKOTO USAMI AND YASUO OHNO Division of Pharmacology, National Institute of Health Sciences, Kamiyoga, Setagaya, Tokyo 158, Japan ABSTRACT Serum embryotrophic factor (SEF) required for the growth of cultured postimplantation rat embryos was partially purified from rat serum. Rabbit serum was used as a basal medium for the embryo culture, and embryotrophic activity was measured as embryonic protein increase. For partial purification of SEF, the ra t serum globulin fraction obtained by ultracentrifugation and (NH4I2SO4precipitation was fractionated by gel filtration, diethylaminoethyl ion-exchange chromatography, and hydroxyapatite chromatography. Partially purified SEF was characterized by stability tests and affinity chromatography. SEF was inactivated by heat, acid, dithiothreitol reduction, or trypsin digestion. SEF bound to concanavalin A but not to heparin. These results indicated that SEF was a n acid-labile acidic glycoprotein with disulphide bonds and no affinity for heparin. The M , of SEF was estimated to be about 180 x lo3 by gel filtration. The specific activity (U/g protein) was increased about 25-fold with 9.4% recovery by the partial purification, when 1 U of SEF was defined as the amount giving 50% embryonic protein increase. By polyacrylamide gel electrophoresis, a protein most likely to be S E F was identified as a heterodimer composed of subunits of M, 116 x lo3 and 62 x lo3 linked by disulphide bonds, and was shown to be contained in the medium at micromolar concentrations. S E F appeared to be distinct from known protein embryotrophic factors, growth factors, or cytokines. o 1996 Wiley-Liss, Inc. Early postimplantation rodent embryos cultured in rat serum grow as well as in the uterus (for review, see New, 'go), and therefore their nutritional requirement can be investigated by examining the effects of serum factors in the embryo culture. So far, many known serum factors have been shown t o be effective for early postimplantation growth of rat embryos using elaborated culture media. Glucose and vitamins, such as pantothenic acid, riboflavin, inositol, folic acid, and niacinamide, were vital for embryos cultured in dialyzed rat serum (Cockroft, '79, '88; Gunberg, '76). High ,concentrations of rat transferrin ameliorated the anaemia of rat embryos cultured in human serum (Cumberland et al., '87; Cumberland and Pratten, '89). Methionine and iron promoted the growth of embryos cultured in dog serum (Flynn et al., '87). Methionine and haemoglobin improved the growth of embryos cultured in bovine serum (mug et al., '90). Methionine was also reported to be essential for neural tube closure in embryos cultured in bovine serum (Coelho and Klein, '90; Coelho et al., '89). Epidermal growth factor, insulin, and transferrin support the growth of embryos cultured in rat serum ex@ 1996 WILEY-LISS, INC. hausted by repeated use in the embryo culture (Pratten et al., '88). Insulin and insulin-like growth factors (IGFs) improved the growth of embryos cultured in guinea pig serum (Travers et al., '92). Besides these known serum trophic factors, there are yet unidentified embryotrophic factors required for early postimplantation growth of rat embryos, As shown in the studies using dialyzed rat serum, some serum macromolecule is indispensable for embryonic growth (Gunberg, '76). Cultures in heterologous sera, such as dog, human, monkey, and pig, supplemented with rat serum suggest the presence of some species-specific embryotrophic factor in rat serum (Gupta and Beck, '83; F'riscott, '83; Reti et al., '82; Steele, '85). Electrophoretic analyses of protein changes in rat serum as a medium after embryo culture showed depletion of some unknown proteins of M , 125 x lo3, 132 x lo3, and 214 x lo3 as well as a,-acid Received April 1, 1996; revision accepted August 6, 1996. Address reprint requests to Makoto Usami, Division of Pharmacology, National Institute of Health Sciences, 1-18-1, Kamiyoga, Setagaya, Tokyo 158, Japan. 404 M. USAMI AND Y. OHNO glycoprotein, a2-macroglobulin, and transferrin (Klein et al., '78; Priscott et al., ,831, but none of them has been isolated. We have previously shown that a growth factor, called serum embryotrophic factor (SEF), necessary for early postimplantation growth of rat embryos is present in the rat serum globulin fraction, using rabbit serum as a basal medium for the embryo culture (Usami et al., '92). When cultured in rabbit serum alone, rat embryos grew poorly. If rat serum was added, however, embryonic growth in rabbit serum was improved to an extent as in Eagle's miinimal essential medium containing the same amount of rat serum. Embryonic growth in rabbit serum was also improved by the addition of the ra.t serum globulin fraction prepared by (NH4),S04 precipitation. It was thus indicated that SEF was present in the rat serum globulin fraction. It wa3 also shown that rabbit serum was virtually non-toxic t o rat embryos, but lacked SEF. In addition, rabbit serum supplied unidentified nutrients common with rat serum and was easily preparable. Rabbit serum was therefore considered t o be suitable as a culture medium for the invesiigation of SEF. In the present study, we partially purified SEF by further fractionation of the rat serum globulin fraction, using rabbit seruni as a basal medium for the embryo culture. Subsequently we characterized and identified a protein most likely to be SEF. MATERIALS AND METHODS Embryo1 culture Postimplantation rat embryos were cultured by the roller bottle method as described previously (Usami et al., '92). Embiryos at head-fold to early somite stage were explanted from Wistar rats (Crj:Wis, Charles River Jiapan, Kanagawa, Japan) at day 9.5 of gestation. Tlhree embryos were placed in a 30-ml bottle containing 4 ml of medium, and rotated at 35 rpm, 37-38°C for 48 hours. Prior to the examination of emhryotrophic activity, fractionated rat sera were concentrated to an appropriately fixed volume in each assay by ultrafiltration using Diaflo membrane YM-10 (M, cut off 10 x lo3, Amicon, Danvers, MA) unless otherwise noted, and dialyzed using the SpectrdPore 7 dialysis tube (M, cut off 50 x lo3, Spectrum, Houston, TX) against Hanks' balanced salt solution (HBSS) containing 2 g/l glucose and 2 g/l NaHC03. The fractionated rat sera were added t o rabbit serum in the proportion of 25% (dv), i.e., one part of fractionated serum and three parts of rabbit serum were mixed, and the mixtures were used as culture media. The protein content in cultured embryos was determined by the method of Lowry et al. ('51). Preparation of sera and the rat serum globulin fraction Rat serum was prepared by immediate centrifugation (Steele and New, '74) of blood taken from the abdominal aorta of male Wistar rats fasted for 18 hours. Rabbit serum was prepared by immediate centrifugation of blood taken from the auricular artery of Japanese White rabbits (Std:JW/ CSK, Japan SLC, Shizuoka, Japan). These sera were heat-inactivated a t 56°C for 30 minutes and supplemented with 1 g/l of glucose. Sera were filter-sterilized and stored a t -30°C until use. The globulin fraction was prepared by (NH4I2SO4 precipitation after removal of lipids by ultracentrifugation. For ultracentrifugation, five parts of rat serum (110 ml per run) was mixed with four parts of 21% (WN) NaCl solution t o adjust the specific gravity to about 1.063 g/ml. The mixture was layered on 60% sucrose solution in a centrifugation tube to prevent the precipitation of the solute at the bottom. The tubes were placed in an angle rotor (RP50-2, Hitachi, Tokyo, Japan) and spun at 32,500 rpm (average centrifugal force 81,500g) at 14°C for 16 hours. From the non-lipid lower layer, the globulin fraction was prepared by (NH,),SO, precipitation as previously described (Usami et al., '92). Fractionation of the rat serum globulin fraction The globulin fraction (55 ml) was dialyzed against 0.05 M Tris-HC1, 0.1 M NaC1, pH 8.0 (5 liter). The dialyzed globulin fraction, 5 ml per run, was applied to a gel filtration column (Sephacryl S-300, Pharmacia, Piscataway, NJ; 2.6 x 96 cm) equilibrated with the dialysis buffer. The column was run at a flow rate of 1mumin at 4°C. For M , estimation, a calibration curve for the gel filtration column was made with calibration kit proteins (Pharmacia). The fractions with embryotrophic activity from three or four runs of the gel filtration were diluted twofold by distilled water and loaded on a diethylaminoethyl (DEAE) ion-exchange column (DEAE Sephacel, Pharmacia, 2.6 x 19 cm) equilibrated with 0.025 M Tris-HC1, 0.05 M NaC1, pH 7.0. After washing with three column volumes of the equilibration buffer, the column was run at a flow rate of 0.5 mumin a t 4°C and eluted with a PURIFICATION OF SERUM EMBRYOTROPHIC FACTOR 405 tridge was eluted by increasing the NaCl concentration to 2.0 M, and the bound fraction was collected. The unbound and bound fractions were concentrated to the same volume by ultrafiltration using centriprep-30 ( M , cut off 30 x lo3, Amicon). The presence of sugar in SEF was examined by concanavalin A (con A) chromatography. SEF was applied on a con A column (ConA Sepharose, Pharmacia, 1 x 8 cm) equilibrated with start buffer (0.025 M Tris-HC1, 0.5 M NaC1, pH 7.0) at a flow rate of 0.2 mum1 at 4"C, and the unbound fi-action was collected. After washing with five column volumes of the start buffer, the column was eluted with 0.3 M a-D-methylmannoside in the Characterization study same buffer, and the bound fraction was collected. The HA fractions with embryotrophic activity The unbound and bound fractions were concenwere pooled and used as SEF. SEF was concen- trated to the same volume by ultrafiltration ustrated to about 2 mg proteidml by ultrafiltration ing centriprep-30. before treatments. The treated SEF was dialyzed Protein determination and against HBSS prior to the assay of embryotrophic electrophoretic analysis activity. Protein content in the fractions was determined SEF was dialyzed against 0.025 M Tris-HC1,O.l M NaCl, pH 7.0, prior to the testing of heat and by a dye binding method using Protein Assay (Bioacid stability. To test heat stability, SEF was Rad). HA fractions were analyzed by one-dimenheated in boiling water for 3 minutes and centri- sional sodium dodecyl sulphate polyacrylamide gel fuged for the removal of the formed precipitate. electrophoresis (SDS-PAGE) as described by To test acid stability, SEF was incubated in the Laemmli ('70) with 7.5% resolution gel (7 x 8 cm) presence of 1M acetic acid (pH 2.3) at 25°C for 2 and 4% stacking gel of 0.75 mm thick. Samples hours, and then dialyzed against the start buffer were prepared by mixing HA fractions with sample buffer at 1:4 ratio with 2-mercaptoethanol at 4°C. SEF was dialyzed against 0.025 M Tris-HC1, 0.1 as a reducing agent, and were heated at 95°C for M NaCl (pH 8.0) prior to the testing of sensitivity 5 minutes. The sample mixtures (5 pl)?which conagainst reduction and protease digestion. To test tained 1.0, 2.7, 1.0, and 0.1 pg of protein for HA sensitivity against reduction, SEF was incubated fractions a, b, c, and d, respectively, were loaded in the presence of 0.05 M dithiothreitol (DTT) and on each lane of 5 mm width. Electrophoresis was 0,025 M NaHCO, at 25°C for 2 hours, and then run at a constant current of 15 mA for 15 mindialyzed against the start buffer at 4°C. To test utes and then at 20 mA. The gel was stained with sensitivity against protease digestion, SEF was silver (Silver Stain, Bio-Rad) and M , was estiincubated in the presence of 168 U/ml of insoluble mated from a calibration curve made with calitrypsin (from bovine pancreas, TPCK treated and bration kit proteins (Bio-Rad). The molar amount attached t o DITC glass, Sigma, St. Louis, MO) at of specified bands was calculated from the follow37°C for 2 hours. The digestion was stopped by the ing data: relative band intensity determined as addition of 1/32 volume of rabbit serum and the peak area by densitometry, total protein content insoluble trypsin was removed by centrifugation. in the sample, and the M , of the specified bands. The heparin binding of SEF was examined by The presence of disulphide bonds between subheparin chromatography. SEF was dialyzed against units was examined by two-dimensional diagonal start buffer (0.025 M Tris-HC1, 0.2 M NaC1, pH SDS-PAGE under non-reducing conditions in the 7.0). The dialyzed SEF was applied on a heparin first dimension and reducing conditions in the seccartridge column (Econo-Pac heparin cartridge, ond dimension (Wang and Richards, '74) with the Bio-Rad, 5 ml) equilibrated with the start buffer same gel as one-dimensional SDS-PAGE. Samples at a flow rate of 1mVmin at 4°C and the unbound were prepared and applied as described above exfraction was collected. After washing with ten col- cept that 2-mercaptoethanol was omitted t o leave umn volumes of the start buffer, the heparin car- disulphide bonds intact. After the run of the first combined linear gradient of NaCl concentration from 0.05 t o 0.5 M. The fractions with embryotrophic activity from two or three runs of the DEAE ion-exchange chromatography were pooled and dialyzed against start buffer for hydroxyapatite (HA) chromatography, 0.05 M potassium phosphate, pH 6.9. The dialyzed fraction was loaded on a HA column (BioGel HTP, Bio-Rad, Richmond, CA; 1.5 x 15 cm). After washing with three column volumes of the start buffer, the column was run at a flow rate of 0.5 or 1 mVmin at 4°C and eluted with a linear gradient of the potassium phosphate concentration from 0.01 to 0.2 M. M. USAMI AND Y.OHNO 406 dimension, the sample lane was excised in 1.2 mm width from the gel, and was laid on the gel for the second dimension. The second dimension was run after the equilibration of the gel lane with the sample buffer containing 2-mercaptoethanol for 10 minutes to cleave disulphide bonds. Electrophoresis was r u n and the gel was stained with silver i n the same way as one-dimensional SDS-PAGE. rat serum (% of DRS) in each run of assay to reduce the interassay variation, which was calculated as follows: (FracEmbPro- HbssEmbPro) Embryonic protein = x 100, increase (% of DRS) (DrsEmbPro- HbssEmbPro) where FracEmbPro is protein content in the embryo cultured in rabbit serum with the addition of fractionated rat serum of which embryotrophic acRESULTS tivity is to be determined, HbssEmbPro is protein Assay of embryotrophic activity content in the embryos concurrently cultured in For a n assay of embr!yotrophic activity at each rabbit serum with the addition of the carrier solupurification step, a quantitative and sensitive tion (HBSS) alone, and DrsEmbPro is protein conmeasure of the growth of cultured embryos had tent in the embryos concurrently cultured in rabbit to be determined. Since the embryonic protein con- serum with the addition of dialyzed rat serum. tent was considered most suitable among the six Embryotrophic activity of gel growth parameters that we had employed previfiltration fractions ously, i.e., presence of heartbeat and flexion, yolk The globulin fraction of rat serum showed five sac diameter, crown-rump length, somite number, and protein content (Usami et al., '921, we exam- major peaks when fractionated by gel filtration. ined its relationship with the amount of SEF, in At the initial rough division into five fractions, this case, dialyzed rat gerum expressed as pro- strong embryotrophic activity was observed in the tein content. As a resull,, there was a clear dose- fractions 3 and 4 (Fig. 2A). When these two acresponse relationship between the protein content tive fractions were further divided into eight fracin cultured embryos and the amount of dialyzed tions, embryotrophic activity was observed with rat serum in the 25% additives (Fig. 1);in this a peak around M , 180 x lo3 as estimated from case 1ml of dialyzed serum, equal t o 25% in 4 ml the calibration curve made with calibration kit of the medium, containled 68 mg of protein. On proteins (Fig. 2B). the basis of this finding, embryonic protein inEmbryotrophic activity of DEAE fractions crease compared to those by HBSS was used as a The gel filtration fractions d through g were measure for embryotrophic activity in this assay pooled and fractionated by DEAE ion-exchange system. In practice, embryonic protein increase chromatography. When the bound fraction was was expressed as percentage of those by dialyzed eluted with a steep gradient of NaCl concentration and divided into four fractions, strong embryotrophic activity was observed in the fractions 2 and 3 (Fig. 3A). On further division of the fractions 2 and 3 into six fractions by elution with a more gentle gradient of NaCl concentration, embryotrophic activity was observed in the fractions eluted around 0.12 M NaCl (Fig. 3B). The binding t o DEAE, a n anion exchanger, at the neutral pH indicated that SEF was acidic. 0 10 20 30 40 50 60 Dialyzed rat serum (mgproteinl4 mllbottle) 70 Fig. 1. A plot of the amount of dialyzed rat serum in the medium vs. embryonic protein content. Values are means for six embryos. Error bars indicate s.e.m. Embryotrophic activity of HA fractions The DEAE fractions b through e were pooled and fractionated by HA chromatography. When the eluate was divided into four fractions according to the major peaks in the chromatogram, strong embryotrophic activity was observed in the fraction 3 eluted around 0.13 M PO, (Fig. 4A). Further division of the fraction 3 showed that embryotrophic activity was proportional to the PURIFICATION OF SERUM EMBRYOTROPHIC FACTOR 0.5 0.4 I - 0.3 0.2 E 2 5 2 e $ 0.1 4 0 Elution volume (ml) ' 1 100 rn $ E .- I5 h cv) B& &; 0.5 0.4 1 0.3 9 m c? 50 o'2 0 .a 0E ' 2 ' 3 ' 4 ' 5 ' Gel filtration fraction 8 e 407 Characterization of SEF Table 1shows the stability of SEF against heat, acid, reduction, and protease digestion. Heat at 100°C or acid treatment with acetic acid at pH 2.3 completely inactivated SEF. Reduction with DTT or protease digestion with trypsin also inactivated SEF although to a lesser extent. These results indicate that SEF is an acid-labile protein with disulphide bonds. Table 2 shows the analysis of SEF by affinity chromatography. For a ligand, heparin was used, since heparin is known to bind several growth factors (Lobb et al., '86) and therefore it was expected that their relation to SEF could be examined. On heparin chromatography, almost all the embryotrophic activity was recovered in the unbound fraction. This means that SEF has no specific affinity for heparin. Concanavalin A, a lectin specific for a-D-mannopyranose or a-D-glucopyranose o r stereochemically related sugars (Goldstein et al., '651, was used to examine the presence of sugar in SEF. On con A chromatography, all the embryotrophic activity was recovered in the bound frattion. It was thus indicated that SEF had sugar. 5: 25 3 Summary of purification The dose-response relationships at each of the 0 0 purification steps are shown in Figure 6, The slope 350 150 200 250 300 of the dose-response curve became steeper as the Elution volume (ml) purification proceeded, indicating that the specific TizPn Gel filtration fraction activity increased. It was noted, however, that embryotrophic activity at the ultracentrifugation Fig. 2. Embryotrophic activity of gel filtration fractions. and (NH4),S04precipitation was lower than that Embryotrophic activity was expressed as embryonic protein at the start (dialyzed rat serum) and the doseincrease. Values are means for six embryos. A: Rough divi- response c u ~ e were s not linear exceptfor the sion of the rat serum globulin fraction. Arrows with figures start. Table 3 shows the summary of the partial indicate the elution volumes of calibration kit proteins with their M , x Vo means void volume of the column. The Purification by tentatively defining 1u of SEF as calibration kit consists of thyroglobulin, ferritin, catalase, al- the amount that gives 50% embryonic protein indolase, albumin, and ovalbumin in descending order of M,. crease; the amount of protein that contains 1 u B: Further division of the fractions 3 and 4 in A. An arrow of SEF can be read from the horizontal =is at the indicates the peak of embryotrophic activity with estimated point of intersection of the dose response-curve and M, x the dotted line at 50% embryonic protein increase in Figure 6, and its reciprocal is the specific activthird chromatographic peak (Fig. 4B),suggesting ity. The specific activities at each purification step that SEF is a major constituent of this peak. In were calculated as units per gram of protein from the subsequent characterization, therefore, the the results in Figure 6. From the specific activipooled HA fraction corresponding to the third chro- ties, it was indicated that SEF was purified about 25-fold. The recovery of SEF was 9.4%. matomaphic Desk was used. -~ Figure 5 shows the embryos cultured in rabbit Electrophoretic analysis of H A fractions serum supplemented with the pooled HA fraction as partially purified SEF. It was noted that also The HA fractions a through d contained two morphologically they grew well compared to the major protein bands of IW, 116 x lo3 (band A) and HBSS control. 62 x lo3 (band B), the intensities of both of which E W 0.1 408 M. USAMI AND Y.OHNO . 100, 0.6 -0.5 -0.4 ; I -0.3 G -0.2z" -0.1 -0 ' 1 Elution volume (ml) I 2 ' 3 ' 4 DEAE fraction i 100 el Fa p 75- 8 .C zEZ En 50- a% 2 g2 8$ wE 25- 0 SO 100 150 200 250 300 Elution volume (ml) 'a'b'c'd 'elf' DEAE fraction 350 400 450 Fig. 3. Embryotrophic acti.vity of DEAE fractions. Embryotrophic activity was expressed as embryonic protein increase. Values are means for six embryos. A Rough division of the gel filtration fractions d through g. B: Further division of the fractions 2 and 3 in A by more gentle gradient of NaCl concentration. were proportional t o the embryotrophic activity and the HA chromatog-am (Fig. 7A). In addition, only these two protein bands showed the intensities apparently proportiional to the embryotrophic activity at every purification step and characterization study (not shown). The bands A and B were thus most likely to be SEF. The molar amounts of the bands A and B in the HA fractions a through d determined by densitometry were plotted against embryotrophic activity in Figure 7B. Although the amounts of the bands A and B might be somewhat inaccurate because of possible overlapping by other faint bands, unknown sugar content, etc., the bands A and B showed similar dose-re sponse curves, suggesting that they were present in the same molar amount. From this and the M , of SEF by gel filtration, it was considered likely that the bands A and B composed heterodimeric SEF since their total M , (178 x lo3) was almost equal t o the M , of SEF by gel filtration (180 x lo3). The dose-response curves also show that the bands A and B were contained in the medium at micromolar concentrations, and that if the bands A and B compose heterodimeric SEF, 1U of SEF as defined above was equivalent to about 4 nmol(720 pg). The presence of disulphide bonds between the bands A and B was examined by the diagonal SDS-PAGE of the HA fraction b (Fig. 7C). The bands A and B appeared in the same shape on the same vertical line at M , > 200 x lo3 in the non-reducing first dimension, indicating that they were originally linked by disulphide bonds as a heterodimer. The inconsistency of the M , in PURIFICATION OF SERUM EMBRYOTROPHIC FACTOR 100 I 6 I jo 150 lbo 200 Elution volume (ml) 0.1 250 CIA fraction , 0.06 100 I E3 I" 0 A 160 200 300 400 500 Elution volume (ml) 'a b'c I d' HA fraction Fig. 4. Embryotrophic activity of HA fractions. Embryotrophic activity was expressed as embryonic protein increase. Values are means for six embryos. A: Rough division of the DEAE fractions b through e. B: Further division of the fractions 3 and 4 in A by more gentle gradient of PO4 concentration. the non-reducing first dimension with the total M , of the bands A and B is probably due to overestimation in the non-reducing first dimension; relative mobilities of protein can be lowered in SDS-PAGE under non-reducing conditions since SDS tends to bind to unreduced protein in less amount than to reduced protein (Pitt-Rivers and Impiombato, '68). DISCUSSION In the present study, we partially purified and characterized SEF in rat serum for the first time. The results can be summarized as follows. SEF was characterized as an acid-labile acidic glycoprotein with disulphide bonds and no affinity for 409 heparin. The M, of SEF was estimated to be about 180 x lo3 by gel filtration. Partial purification raised the specific activity about 25-fold with 9.4% recovery. By SDS-PAGE analysis, a protein most likely to be SEF was identified as a heterodimer composed of subunits of M, 116 x lo3 and 62 x lo3 linked by disulphide bonds, and was shown to be contained in the medium at micromolar concentrations. The estimated M, of SEF indicates that SEF is distinct from proteins known as embryotrophic factors. SEF is not transferrin, the only protein whose embryotrophic activity was directly shown in culture (Cumberland et al., '87; Cumberland and Pratten, '89; Pratten et al., '88). Rat transferrin, M , 68 x lo3 to 76.5 x lo3 (Charlwood, '63; Schreiber et al., '79) is smaller than SEF, and it was completely separated from SEF by gel filtration as pale red fractions in the fourth major peak in the present study (not shown). SEF also seems distinct from al-acid glycoprotein, M , 43.5 x lo3 (Nagashima et al., '80) and a2-macroglobulin, M , 760 x lo3 (Gordon, '761, which were depleted in rat serum used as a medium for the embryo culture (Priscott et al., '83). However, the relationship of SEF with unidentified proteins of M , 125 x lo3, 132 x lo3, and 214 x lo3 (Klein et al., '78; Priscott et al., '83) depleted in rat serum by the embryo culture is uncertain at present, since their M, are near t o the estimated M , of SEF and no other information is available. While there has been increasing evidence that growth factors and cytokines expressed in embryos or uteri play important roles in postimplantation growth (for review, see Giudice, '94; Mercola and Stiles, '88; Tabibzadeh, '941, none of them are considered identical to SEF. Among all the known growth factors or cytokines, only hepatocyte growth factor (HGF) can have M , comparable to that of SEF. The M , of HGF was initially reported to be about 150 x lo3 by gel filtration of serum from hepatectomized rats (Nakamura et al., '84). Unlike SEF, however, HGF binds t o heparin and is inactivated by incubation at 56"C, 30 minutes (Nakamura et al., '84, '86). No affinity of SEF for heparin also indicates that SEF is disparate from heparin binding growth factors, a group of growth factors that bind to heparin (Lobb et al., '86). Insulin-like growth factor-I (IGF-I), which showed embryotrophic activity in rat embryo culture with guinea pig serum (Pavers et al., '92), is present in rat serum as a large complex of M , comparable t o that of SEF. This IGF-I complex is of M , 150 x lo3 by gel filtration, and composed of an IGF-I M. USAMI AND Y.OHNO 410 Fig. 5. Rat embryos cultured in rabbit serum. Bars represent 1mm. A Culture witchthe addition of partially purified SEF (2.32 mg/4 ml/bottle). B: Culture with the addition of partially purified SEF (1.16 mg/4 ml/bottle). C: Culture with the addition of dialyzed rat serum. D: Culture with the addition of HBSS alone. binding component of glycoproteins of M , 29 x lo3 (IGFBP3) and 40 x lo3 associated with an acidlabile nonbinding subunit of M , 100 x lo3 (Yang et al., '89). It is, nevertheless, unlikely that SEF is identical t o this 1GF-I complex because the electrophoretic analysis shows no protein bands corresponding t o the components of this complex and because it has been suggested that decreased association of IGF-I with this complex increases the availability of IGF-I by the conceptus in pregnant rats (Bastian e t al., '93; Davenport et al., '90; Gargosky e t al., '90). Other growth factors or cytokines that appeared in recent reviews (Pimentel, '94) are too small in M, compared to SEF. The present results indicated that embryonic protein increase was an effectual measure for embryotrophic activity of SEF. In the past studies, embryonic growth was evaluated by combi- TABLE 1. Stability of SEF against various treatments' Treatment Acid (pH 2.3, 25"C, 2 hours) Heat (lOO°C, 3 minutes) Reduction (0.05 M D'M', 25"C, 2 hours) Protease digestion (168 U trypsidml, 37"C, 2: hours) Control (% of DRS) Treated (% of DRS) % of control 67.1 4.8 -7.2 63.8 1.0 1.6 78.8 25.9 32.9 72.5 11.7 16.1 'Values are mean embryonic protein increase for six embryos. PURIFICATION OF SERUM EMBRYOTROPHIC FACTOR 411 TABLE 2. Analysis of SEF by affinity chromatography' Ligand Unbound (% of DRS) Bound (% of DRS) Bound rate (%I2 Heparin Con A 74.6 -1.2 0.4 41.8 0.5 103.0 'Values are mean embryonic protein increase for six embryos. 2Calculated by dividing the bound activity by the total of the unbound and bound activities. nations of several parameters, such as viability, crown-rump length, morphological feature, and protein content. These combined parameters are of significance for precise evaluation of the growth, but were inadequate for the assay of SEF because of complexity in interpreting the results; for example, crown-rump length can not be measured without axial flexion even for the embryos with increased protein content. The use of embryonic protein increase as a measure made it possible to assay embryotrophic activity quantitatively and t o define the unit of SEF. This unit then enabled us to calculate the specific activity. If the band A and B protein is heterodimeric SEF, serum SEF level can be roughly estimated to be 10 pM (1.8 mg/ml) from the specific activity in dialyzed rat serum (37.8 U/g protein) and protein concentration in rat serum (66.4 mg/ml), because 1U of the heterodimeric SEF is considered equivalent t o 4 nmol(720 pg). In addition to these advantages, it should be noted that embryonic protein increase is accompanied with morphological growth as indicated by the appearance of cultured embryos. Besides SEF, there may be minor embryotrophic factors that are effective for the growth of rat embryos cultured in rabbit serum. The somewhat lowered specific activity in the globulin fraction and non-linear dose-response curves in Figure 6 suggest that minor embryotrophic factors were present in rat serum and lost during the purification procedures. These minor embryotrophic factors may include transferrin because it is present in dialysis serum and has species-specificity in its embryotrophic activity (Cumberland et al., '87; Cumberland and Pratten, '89). Purified SEF might be required at a higher rate for much better embryonic growth to compensate the absence of these minor embryotrophic factors. Rabbit serum served adequately as a basal medium for the investigation of SEF. The use of rabbit serum allowed a n assay almost specific for SEF. Sera of other species would not have worked for this purpose, since they can support the growth of cultured rat embryos well without SEF. Cultured rat embryos can grow without SEF in human, dog, cow, or guinea pig serum, if rat 4 Start (dialyzed rat semrn) +Ultracentrifugation and (NH4)2S04precipitation +Gel filtration +DEAE ion-exchange .s 2F w + J - chromatography HA chromatography 25 0 50 60 Protein content (mg/4 ml bottle) Fig. 6. Dose-response relationships between protein content in the fractions and embryotrophic activity at each step of the SEF purification. Embryotrophic activity was expressed as embryonic protein increase. Protein content represents the amount in the 1 ml additive t o the 3 ml medium (4 ml in total) per bottle. The purification steps are shown in order on the right. A dotted line indicates the 50% embryonic protein increase level for the calculation of the specific activity. Values are means for six embryos. M. USAMI AND Y. OHNO 412 TABLE 3. Summary of SEF purification' Purification step Specific activity (U/g protein)' Purification factor (foldj3 Total activity (U4 Recovery (%I5 37.8 1 251 100 0.73 2.86 73.7 70.8 Start (dialyzed rat serum) Ultracentrifugation and (NH4j2S04precipitation Gel filtration DEAE ion-exchange chromatography HA chromatography 27.5 108 451 952 11.9 40.2 25.2 23.5 29.4 28.2 16.0 9.4 'Values are from a representative run of purification. 'Obtained from the results in Figire 6 by defining 1U of SEF as the amount giving 50% embryonic protein increase. 3Calculated by dividing the specific activity in each step by that of the start. 4Calculated by multiplying the specific activity by total protein content in each step. 'Calculated by dividing the total iictivity in each step by that of the start. HA fraction A a b c iB d -3 MrXlO 200 + 116 + 97.4 CBandA + 66.2 + 45.0 + +Band B 0 2 4 6 Protein content (nmol/4 ml/bottle) 8 First dimension (non-reducing) -3 C Mrx10 b 116 66.2 HA fraction b -3 Mrx10 200 + 116 -b 4BandA + 4 Band B 97.4 66.2 + 4 Heterodimer Fig. 7. Electrophoretic analysis of the HA fractions. A One-dimensional SDS-PAGE of the HA fractions a through d. B A plot of the molar amount of bands A and B in the HA fractions a through d vs. embryotrophic activity. Embryotrophic activity was expressed as embryonic protein increase. C: Two-dimensional diagonal SDS-PAGE of the HA fraction b. The gel shows the run of the first dimension horizontally and the second dimension vertically. The calibration kit consists of myosin, P-galactosidase, phosphorylase b, bovine serum albumin, and ovalbumin in descending order of M,. PURIFICATION OF SERUM EMBRYOTROPHIC FACTOR 413 nutritional requirements of postimplantation rat embryos in culture. Teratology, 38~281-290. Coelho, C.N.D., and N.W. Klein (1990) Methionine and neural tube closure in cultured rat embryos: Morphological and biochemical analyses. Teratology, 42:437-45 1. Coelho, C.N.D., J.A. Weber, N.W. Klein, W.G. Daniels, and T.A. Hoagland (1989) Whole rat embryos require methionine for neural tube closure when cultured on cow serum. J. Nutr., 119:1716-1725. Cumberland, P.F.T., and M.K. Pratten (1989) Transferrin: a species-specific growth promoting factor for embryos. Biochem. SOC.Trans., 17:400-401. Cumberland, P.F.T., E.P.K. Mensah-Brown, T. Murakami, and M.K. Pratten (1987) Species-specificity of transferrin in embryonic growth. Biochem. SOC.Trans., 15:919-920. Davenport, M.L., D.R. Clemmons, M.V. Miles, C. CamachoHubner, A.J. DErcole, and L.E. Underwood (1990) Regulation of serum insulin-like growth factor-I (IGF-I) and IGF binding proteins during rat pregnancy. Endocrinology, 127:1278-1286. Flynn, T.J., L. Friedman, T.N. Black, and N.W. Klein (1987) Methionine and iron as growth factors for rat embryos cultured in canine serum. J . Exp. Zool., 244:319-324. Gargosky, S.E., P.E. Walton, P.C. Owens, J.C. Wallace, and F.J. Ballard (1990) Insulin-like growth factor-I (IGF-I) and IGF-binding proteins both decline in the rat during late pregnancy. J. Endocrinol., 127:383-390. Giudice, L.C. (1994) Growth factors and growth modulators in human uterine endometrium: Their potential relevance to reproductive medicine. Fertil. Steril., 61:l-17. Goldstein, I.J., C.E. Hollerman, and E.E. Smith (1965) Protein-carbohydrate interaction. 11. Inhibition studies on the interaction of concanavalin A with polysaccharides. Biochemistry, 4:876-883. Gordon, A.H. (1976) The cx macroglobulins of rat serum. Biochem. J., 1593343-650. Gressens, P., J.M. Hill, I. Gozes, M. Fridkin, and D.E. Brenneman (1993) Growth factor function of vasoactive intestinal peptide in whole cultured mouse embryos. Nature, 362:155-158. Gunberg, D.L. (1976) In vitro development of postimplantation rat embryos cultured on dialyzed rat serum. Teratology, 14:65-70. Gupta, M., and F. Beck (1983) Growth of 9.5-day rat embryos in human serum. J. Embryol. Exp. Morphol., 76:l-8. Klein, N.W., P.P. Minghetti, S.K. Jackson, and M.A. Vogler ACKNOWLEDGMENTS (1978) Serum protein depletion by cultured rat embryos (1). J. EXP. ZOO^., 203:313-318. The present study is supported in part by Grantin-Aid for Encouragement of Young Scientists mug, S., C. Lewandowski, L. Wildi, and D. Neubert (1990) Bovine serum: An alternative to rat serum as a culture me(08760286) from The Ministry of Education, Scidium for the rat whole embryo culture. Toxicol. In Vitro, ence, Sports and Culture, Japan. 4598-601. Koszalka, T.R., C.L. Andrew, R.L. Brent, D.A. Beckman, and LITERATURE CITED J.B. Lloyd (1994) Amino acid requirements in the early postimplantation rat conceptus. Placenta, 15:311-320. Bastian, S.E.P., P.E. Walton, J.C. Wallace, and F.J. Ballard Laemmli, U.K. (1970) Cleavage of structural proteins during (1993) Plasma clearance and tissue distribution of labelled the assembly of the head of bacteriophage T4. Nature, insulin-like growth factor-I (IGF-I) and an analogue LR31GF227:680-685. I in pregnant rats. J. Endocrinol., 138:327-336. Charlwood, P.A. (1963) Ultracentrifugal characteristics of hu- Lobb, R.R., J.W. Harper, and J.W. Fett (1986) Purification of heparin-binding growth factors. Anal. Biochem., 154:l-14. man, monkey and rat transfenins. Biochem. J.,88:394-398. Cockroft, D.L. (1979) Nutrient requirements of rat embryos Lowry, O.H., N.J. Rosebrough, A.L. Farr, and R.J. Randall (1951) Protein measurement with the folin phenol reagent. undergoing organogenesis i n vitro. J. Reprod. Fertil., J. Biol. Chem., 193:265-275. 57:505-510. Cockroft, D.L. (1988) Changes with gestational age in the Mercola, M., and C.D. Stiles (1988) Growth factor super- transferrin, methionine, iron, insulin, or IGF is added (Coelho and Klein, '90; Coelho et al., '89; Cumberland et al.,'87; Cumberland and Pratten, '89; Flynn et al., '87; Klug et al., '90; Travers et al., '92). This is probably because homologues of SEF in these species are effective for rat embryos. Fortunately, the rabbit homologue of SEF appears to be ineffective for rat embryos. Furthermore, SEF seems t o be a major embryotrophic factor required for the growth of rat embryos cultured in rabbit serum. The mechanism of action of SEF is not known a t present. SEF may act either as a growth regulator or as a carrier of small molecules or as a n amino acid source. If SEF act as a growth regulator, there must be specific receptors through which SEF transmits signals t o the cell. Without specific receptors, SEF may work a s a carrier of small molecules, such as peptides with embryotrophic activity. It is known that some circulatory peptides have embryotrophic activity. Vasoactive intestinal peptides promoted t h e growth of postimplantation mouse embryos cultured in human serum supplemented with rat serum (Gressens et al., '93). Otherwise, SEF may serve as a source of amino acids for embryonic growth. Similarity in the amino acid composition between serum and embryos in rats suggested the role of serum protein as an amino acid source for embryonic growth (Koszalka et al., '94). In adult animals, SEF may also play a physiological role, because it is present in both male and female sera. Further investigation of SEF, such as isolation, molecular cloning, and search for homologues in other species including humans, will lead t o a deeper understanding of the postimplantation growth and physiology of mammals. 414 M. USAMI AND Y.OHNO families and mammalian embryogenesis. Development, 10245 1-460. Nagashima, M., J. Urban, and G. Schreiber (1980) Intrahepatic precursor form of rat apacid glycoprotein. J. Biol. Chem., 255:4951-4956. Nakamura, T., K. Nawa, and A. Ichihara (1984) Partial purification and characterization of hepatocyte growth factor from serum of hepatectomized rats. Biochem. Biophys. Res. Commun., 122:1450-1459. Nakamura, T., H. Teramoto, and A. Ichihara (1986) Purification and characterization of 3 growth factor from rat platelets for mature parenchyinal hepatocytes in primary cultures. Proc. Natl. Acad. Sci. U.S.A., 83:64894493. New, D.A.T. (1990) Introduction. In: Postimplantation Mammalian Embryos. A Practical Approach. A.J. Copp, and D.L. Cockroft, eds. Oxford University Press, Oxford, pp. 1-14. Pimentel, E., ed. (1994) Handbook of Growth Factors, vols. 1-111. CRC Press, Boca Raton, FL. Pitt-Rivers, R., and F.S.A. Impiombato (1968) The binding of sodium dodecyl sulphate to various proteins. Biochem. J., 109:825-830. Pratten, M.K., A.M. Brooke, S.C. Broome, and F. Beck (1988) The effect of epidermal growth factor, insulin and transferrin on the growth-promoting properties of serum depleted by repeated culture of post implantation rat embryos. Development, 104~137-145. Priscott, P.K. (1983) Rat post-implantation embryo culture using heterologous serum. Aust. J. Exp. Biol. Med. Sci., 61:47-55. Priscott, P.K., P.G. Gough, and R.D. Barnes (1983) Serum protein depletion by cultured post-implantation rat embryos. Experientia, 39:1042-1043. Reti, L.L., F. Beck, and S. Bulman (1982) Culture of 9 1/2-day rat embryos i n human serum supplemented a n d unsupplemented with r a t serum. J. Exp. Zool., 223:197-199. Schreiber, G., H. Dryburgh, A. Millership, Y. Matsuda, A. Inglis, J. Phillips, K. Edwards, and J. Maggs (1979) The synthesis and secretion of rat transferrin. J. Biol. Chem., 254:12013-12019. Steele, C.E. (1985) Human serum as a culture medium for rat embryos. Experientia, 41:1601-1603. Steele, C.E., and D.A.T. New (1974) Serum variants causing the formation of double hearts and other abnormalities in explanted rat embryos. J. Embryol. Exp. Morphol., 31:707-719. Tabibzadeh, S. (1994) Role of cytokines in endometrium and at the fetomaternal interface. Reprod. Med. Rev., 3:ll-28. Travers, J.P., L. Exell, B. Huang, E. Town, M.J. Lammiman, M.K. Pratten, and F. Beck (1992) Insulin and insulinlike growth factors in embryonic development: Effects of a biologically inert insulin (guinea pig) on rat embryonic growth and development in vitro. Diabetes, 41:318-324. Usami, M., S. Nakaura, K. Kawashima, S. Tanaka, and A. Takanaka (1992) Culture of postimplantation rat embryos i n rabbit serum for t h e identification of t h e growth factor in fractionated rat serum. J. Exp. Zool., 264:2 14-2 18. Wang, K., and F.M. Richards (1974) An approach to nearest neighbor analysis of membrane proteins. J. Biol. Chem., 249:8005-8018. Yang, Y.W.-H., J.-F. Wang, C.C. Orlowski, S.P. Nissley, and M.M. Rechler (1989) Structure, specificity, and regulation of the insulin-like growth factor-binding proteins in adult rat serum. Endocrinology, 125:1540-1555.