A novel view of albumin-supported sperm capacitation: role of Lipid Transfer Protein-I*t*

Transcription

1 FERTILITY AND STERILITY Copyright e 1993 The American Fertility Society Vol. 59, No.3, March 1993 Printed on acid-free paper in U.S.A. A novel view of albumin-supported sperm capacitation: role of Lipid Transfer Protein-I*t* Stuart E. Ravnik, Ph.D. II~ John J. Albers, Ph.D.** Charles H. Muller, Ph.D. litt:j::j: University of Washington, Seattle, Washington Objective: To determine if one mechanism of albumin-mediated support of human sperm capacitation is lipid (cholesterol) transfer activity and contamination of albumin with Lipid Transfer Protein-I (LTP-I). Design and Main Outcome Measures: Measure lipid transfer activity in various bovine and human albumin preparations; relate this activity to albumin-supported capacitation (measured by zona-free hamster oocyte sperm penetration assay) and acrosome reactions; and attempt to detect LTP-I in active albumins. Remove LTP-I from albumin which supports capacitation and reassess this support. Reconstitute capacitation support by addition of purified LTP-I. Setting and Subjects: Healthy sperm donors with normal semen analyses were recruited by the Reproductive Biology-Andrology Laboratory in a university medical center. Results: Albumin preparations that effectively support capacitation have high levels of lipid transfer activity and of LTP-I, a protein responsible for lipid transfer activity. Preparations with lower levels of capacitation support have less lipid transfer activity. Removal of L TP -I from supportive albumin significantly reduces the capacitation support, and this is restored by purified LTP-I. Progesterone concentrations in these preparations are negligible. Conclusions: The variable abilities of albumin preparations to support in vitro sperm capacitation are largely dependent on the presence of contaminating LTP-I. The cholesterol transfer activity of this protein, which is present in human serum and follicular fluid, may be one mechanism in the process of capacitation. Fertil Steril 1993;59: Key Words: Human sperm capacitation, acrosome reaction, zona-free hamster oocyte sperm penetration assay, lipid transfer protein, albumin, cholesterol, serum Successful mammalian in vitro capacitation and fertilization was first achieved in the presence of various biological fluids: oviductal fluids (1), follicular fluid (FF) (2), or serum (3). Subsequently, it was discovered that albumin, as the only protein source, could support capacitation and fertilization in vitro (4,5). Received January 28, 1991; revised and accepted November 16,1992. * Supported by Population Research Center grant HD (to C.H.M.), National Heart, Lung, and Blood Institute grant HL (to J.J.A.), and National Research Service Award training grant T32 HD (to S.E.R.) from the National Institutes of Health, Bethesda, Maryland; and by the University of Washington Graduate School Research Fund (to C.H.M.), Seattle, Washington. t Presented in part at the 14th Annual Meeting of the American Society of Andrology, New Orleans, Louisiana, April 13 to 16, :j: This work was performed in partial fulfillment of Vol. 59, No.3, March 1993 requirements for the degree of Doctor of Philosophy, Department of Biological Structure (S.E.R.). Reproductive Biology-Andrology Laboratory, Department of Obstetrics and Gynecology. II Department of Biological Structure. '\[ Present address: Department of Genetics and Development, Columbia University College of Physicians and Surgeons, New York, New York. ** Northwest Lipid Research Laboratories, Departments of Medicine and Pathology. tt Population Center for Research in Reproduction. :j::j: Reprint requests: Charles H. Muller, Ph.D., Department of Obstetrics and Gynecology, RH-20, University of Washington, SeattIe, Washington Ravnik et al. Albumin, LTP-I, and sperm capacitation 629

2 Although commercially available, albumin has been used successfully for many years in capacitation media (6, 7); the exact mechanisms of its biological action in capacitation, acrosome reaction, and fertilization are not known. Many activities associated with albumin or its contaminants have been suggested as mechanisms for the in vitro support of capacitation. Some of these are hydrolysis of sperm membrane proteins (8); cation chelation, especially of zinc (9); sterol binding (10); and direct stimulation of the acrosome reaction (11). Oddly enough, different albumin preparations have been shown to have variable activities in capacitation or fertilization assays. Reported differences in zona-free hamster oocyte sperm penetration assay (SPA) scores have been attributed to differences in commercially available albumins (7). Our laboratory has been studying human FF lipid transfer activity as a stimulator of human sperm capacitation and acrosome reaction (12). The lipid transfer activity in FF is probably due to Lipid Transfer Protein-I (LTP-I) (13), previously isolated from human plasma (14). Our assays have been carried out in bovine serum albumin (BSA)-supplemented media. The BSA used had trace amounts of lipid transfer activity but did not elicit a dose-dependent response, either for lipid transfer activity or SPA activities (12). We wanted to know if this property was a characteristic of albumins in general, or if trace amounts of L TP (responsible for the lipid transfer activity) were present in commercially available albumins. Here we report a systematic study of different BSA and human serum albumin (HSA) preparations, measuring both lipid transfer activity and the ability of each preparation to support in vitro capacitation and acrosome reaction, determined by ultrashort SPA. Our findings indicate that some commercially available albumin preparations have L TP -I as a contaminant, which appears to be responsible for the variable abilities of albumins to support in vitro capacitation and acrosome reaction. Reagents MATERIALS AND METHODS Fourteen different albumin preparations from five commercial sources were used in this study. The most extensive testing was performed on the albumin preparations purchased from Sigma Chemical Co. (St. Louis, MO). The Sigma preparations used were fraction V-BSA no. 4503, fraction V-BSA 630 Ravnik et al. Albumin, LTP-I, and sperm capacitation no. 9647, fatty acid-free-bsa no. 6003, crystallized BSA no. 4378, fraction V-HSA no. 2386, fraction V-HSA no. 1653, fatty acid-free-hsa no. 1887, and crystallized-hsa no A single lot of each preparation was used throughout this study. Fatty acidfree-bsa (no. 6003) and crystallized-bsa (no. 4378) were produced from pooled lots of fraction V -BSA (no. 4503); fatty acid-free-hsa (no. 1887) and crystallized-hsa (no. 8763) were produced from pooled lots of fraction V-HSA (no. 1653). Albuminar-5 (Armour Pharmaceuticals, Kankakee, IL), Calbiochem crystallized-bsa (Calbiochem Corp., La Jolla, CA), Irvine albumins (fraction V-BSA, fatty acid-free-bsa, fraction V-HSA; Irvine Scientific, Santa Ana, CA), and Miles Pentex fraction V-BSA (Miles Inc., FBA Pharmaceuticals, West Haven, CT) were also tested. Human plasma was purchased from Seattle Plasma Corp. (Seattle, WA). Radioactive cholesterol (cholesterol, 4_14C) was obtained from New England Nuclear (Boston, MA). Activated Sepharose-4B (CnBr) was purchased from Pharmacia (Piscataway, NJ). Fluorescein isothiocyanateconjugated Pisum sativum agglutinin (FITC-PSA) was obtained from Vector Laboratories, Inc. (Burlingame, CA), and affinity purified goat-antihuman immunoglobulin (lg)g was purchased from Organon Teknika, Corp.-Capell (Cochranville, PA). Polyvinylpyrrolidone (average molecular weight [MW] 40,000) and Hoechst dye were obtained from Sigma Chemical Co. Lipid Transfer Assay The assay for lipid transfer activity was performed essentially as described by Ravnik et al. (12). Briefly, low-density lipoproteins (LDL) were isolated from fresh human plasma by differential ultracentrifugation and used as cholesterol ester (CE) acceptor lipoproteins. High-density lipoproteins (HDL) with plasma proteins, isolated by ultracentrifugation, were loaded with 14C-cholesterol and incubated overnight at 37 C, allowing the formation of CE by the lecithin-cholesterol acyltransferase reaction (14). After ultracentrifugation to separate HDL from plasma proteins, the resulting 14C-CE-HDL were used as donor lipoproteins. The specific activity of the donor lipoproteins was approximately 1,500 cpm/ij.g CE (assuming average MW of 650 for CE). All assays were performed at optimum donor:acceptor ratios (1:5) to yield maximum transfer activity with a given amount of lipid transfer activity partially purified from plasma as described by Albers et al. (14). Unless noted, 25 to 200 ij.l of a 10% solution of each albumin were added to 200 ij.l of do- Fertility and Sterility

3 nor:acceptor and made to a volume of 600 J.LL with 0.15 M NaCI, 20 mm N-(2-hydroxyethyl}piperazine-N'-(2-ethanesulfonic acid} (HEPES), ph 7.4 (HEPES saline). Controls had an equal volume of buffer. All assays were performed in triplicate. SampIes were incubated for 18 hours at 37 C in a shaking water bath. After incubation, LDL were removed by dextran sulfate:mgcl 2 precipitation of apo B-containing lipoproteins (15), and the supernatants were counted (an average of 96% ofthe added counts were recovered ifthe LDL precipitates were also counted). Lipid transfer activity is defined by the following equation (12): «Test cpm -;- control cpm) X 100} giving the percentage of labeled lipid transferred per time (% transfer/18 hours). Values for control samples were thus set to zero. Ultrashort SPA The SPA was performed as described by Ravnik et al. (12). In brief, superovulated Syrian hamsters (Simonsen, Sacramento, CA) were killed, and cumulus masses were teased out and washed through hyaluronidase to remove cumulus cells; oocytes were washed twice with Biggers, Whitten and Whittingham (BWW) medium containing 0.4% BSA (Sigma no. 4503), exposed to 0.1 % trypsin to remove zonae pellucidae (ZP) and finally washed twice with BWW before coincubation with sperm. Human sperm were obtained from healthy donors after liquefaction of semen at 37 C for 20 minutes and centrifuging three times (250 X g) with BWW containing one of the albumin preparations (0.4% concentration). The average sperm motility for all of the samples used was 75% (range, 60 to 90), and the sperm had normal rapidity and linearity. In addition, the motility of the sperm after different washing and treatment protocols (see Experiments 1 to 4 for details) was judged as good to excellent. After the washing procedure, each suspension of sperm was adjusted to approximately 10 X 10 6 sperm/ml with the appropriate albumin in BWW and incubated at 37 C under 1 % CO 2, After 2 to 2.5 hours, 95 J.LL of sperm suspension were transferred to a 5-J.LL drop of BWW under oil, and 25 to 30 zonafree hamster oocytes were added. Coincubation was for 2 hours after which the eggs were removed, washed, and compressed for visualization of sperm penetration using phase-contrast microscopy. Penetration was defined as the presence of a deconden sed sperm head with associated tail within the ooplasm. Results were recorded as percent penetra- tion (total number of penetrated eggs per total eggs) and as penetration index (total number of penetrated sperm per total eggs). Experiment 1 Human serum albumin and BSA preparations were compared for their effectiveness in support of sperm penetration ability. One ejaculate was used for each comparison of four different albumins (i.e., 2 different fraction V-BSAs, fatty acid-free-bsa, and crystallized-bsa in one comparison and the corresponding HSAs in another). Sperm were washed and incubated in each medium as described above. Each of these comparisons was repeated five times using four different donors for both BSA and HSA preparations. Experiment 2 The ability of a potent HSA preparation to continue to support sperm penetrating and acrosome reaction ability after removal of LTP-I was tested. In this experiment, repeated four times (a different donor was used for each replicate), a 40 mg/ml solution of Sigma crystallized-hsa no in Hepes saline was treated by affinity chromatography to remove LTP-I or control-treated (see Preparation of LTP-I-free HSA p. 632), dialyzed against BWW salt solution, and reconcentrated to 40 mg/ml. These preparations were diluted 1:10 to 4 mg/ml into BWW, used in the ultrashort SPA, and prepared for acrosome reaction assessment (see acrosome reaction assessment below) at 0, 1, 2, and 18 hours incubation. Fraction V -BSA no was used as a baseline control for this experiment. Experiment 3 The ability of Sigma crystallized-hsa no , treated to remove L TP -I, to give a dose response for sperm penetration and acrosome reaction was tested by incubating sperm in 0.4%,2.0%, or 3.5% albumin solution. A 40 mg/ml solution of treated or controltreated crystallized-hsa (see Preparation of LTP-I Free HSA p. 632) was diluted 1:10, 1:2, or 1:1.14 into BWW, and sperm were washed and incubated for 2 hours in each concentration of albumin-bww. After incubation, aliquots of sperm were coincubated with eggs and prepared for acrosome reaction assessment (see Acrosome Reaction Assessment p. 632). This experinlent was repeated three times using three different donors. Vol. 59, No.3, March 1993 Ravnik et al. Albumin, LTP-I, and sperm capacitation 631

4 Experiment 4 We wanted to determine the interaction of purified LTP-I with LTP-I-free HSA. To do this, sperm (a single ejaculate from each of 3 different donors) were washed and incubated for 1 hour with 0.4% LTP-I-free HSA in BWW. After the incubation, 200 ~L of the sperm suspension (10 X 10 6 sperm/ml) were treated for 15 minutes with 3.5 ~g/50 ~L of LTP-I purified by sequential column chromatography on phenyl sepharose, diethylaminoethyl sepharose, heparin sepharose, carboxymethyl cellulose, and hydroxyl apatite (14). This preparation ofltp-i had 11% transfer, which is within the linear range of the lipid transfer activity assay. The treated sperm were washed with LTP-I-free HSA BWW and incubated 1 hour further before eggsperm coincubation or acrosome reaction assessment. Control sperm suspensions had LTP-I-free HSA BWW medium alone added. Acrosome Reaction Assessment In some experiments, acrosomal status of spermatozoa was analyzed using a PSA staining method modified from Cross et al. (16). At the time eggs were added to sperm, 1 X 10 6 sperm from each treatment were incubated 10 minutes with 1 ~g Hoechst dye/ml and centrifuged (500 X g for 7 minutes) through 2% polyvinylpyrrolidone in Ham's F-lO medium (GIBCO, Grand Island, NY) without protein. Sperm were resuspended in 100 ~L cold 100% ethanol and kept at 4 C for at least 30 minutes. Aliquots of 20 ~L were dried on slides and hydrated with 20 ~L FITC-PSA (25 ~g/ml) in Hepes saline containing 0.1 mm CaC1 2 and 0.01 mm MnC1 2 After 30 minutes, the slides were washed in distilled water, mounted in 90% glycerol, and 100 sperm that had excluded the Hoechst dye were scored for acrosomal loss. Data are expressed as percent live acrosome reacted sperm. Gel Electrophoresis and Immunoblotting Polyacrylamide gel electrophoresis was carried out in the presence of sodium dodecyl sulfate (SDS) on discontinuous 5% stacking, 12% separating gels by standard methods. Gels were silver stained according to the method of Heukeshoven and Dernick (17). Western blotting of proteins was performed overnight, using a TE 42 Transphor Unit (Hoefer, San Francisco, CA). Alternatively, immunoblots were produced using a slot-blot apparatus. Nitrocellulose sheets (slot immunoblot or Western blot) were blocked in phosphate-buffered saline (PBS), containing 0.05% Tween 20 (PBST) for 5 hours. Overnight incubation was carried out at room temperature with affinity purified goat antibody to human LTP-I, produced as described (15), and diluted 1:300. The sheets were washed three times with PBST and incubated for 2 hours with peroxidase conjugated rabbit antigoat IgG, diluted 1:500. The sheets were then washed and reacted for 5 minutes with 0.5 mg/ml diamino-benzidine in 0.1 M Tris, ph 7.4 and 0.01% H 20 2 in water. Quantitative immunoblots were performed on the HSA preparations as follows. Serial dilutions of each 10% HSA solution (fraction V-HSA no and crystallized-hsa 8763,1:100 to 1:1600; fraction V-HSA no and fatty acid-free-hsa, 1:10 to 1:160) were slot-blotted and probed as described above. Slot blots were scanned with a densitometer, and the intensity of each reaction was determined by measuring the peak heights. Preparation of LTP-I-Free HSA Human serum albumin depleted of LTP-I was derived from crystallized-hsa (Sigma no. 8763) because this preparation had a high level of lipid transfer activity (fraction V-HSA no was no longer available). Removal of L TP -I was carried out by passage of crystallized-hsa over an anti-ltp-i sepharose antibody column (3.5 ml of gel), prepared with affinity purified goat antihuman-ltp-i antibody and cyanogen bromide-activated sepharose according to manufacturer's suggestions. A control antibody column was prepared in the same manner except affinity purified goat antihuman IgG was coupled to the gel. After equilibration in 0.15 M NaCl, 0.01 M Tris, ph 7.4 (Tris saline), 4 ml of a 40 mg/ml crystallized-hsa solution was recirculated through either the anti-ltp-i column or the anti-igg column for 1 hour at a flow rate of 30 ml/h. The columns were washed with 30 ml of Tris saline and eluted with 30 ml of 3 M NaSCN. Fractions of 3 ml were collected. Wash fractions were assayed for lipid transfer activity and slot immunoblotted for detection of LTP-1. Elution fractions were immediately dialyzed against Hepes saline and then assayed for lipid transfer activity and slot immunoblotted. Fractions containing lipid transfer activity from the anti-igg column and the corresponding fractions from the anti-ltp-i column were dialyzed against BWW salts and concentrated by ultrafiltration in Centricon -10 microconcentrators (Amicon, Danvers, MA) back to 40 mg/ml. These preparations were 632 Ravnik et al. Albumin, LTP-I, and sperm capacitation Fertility and Sterility

5 Table 1 Lipid Transfer Activity and Support of Human Sperm Penetration Activity by Different Albumin Preparations* Albumin preparation Lipid transfert Penetration:j: Penetration index:j: BSA-1 Fraction V no ± ± ± 0.01\[ BSA-2 Fraction V no ± ± 0.02 BSA-3 Fatty acid-free no ± ± 0.03 BSA-4 Crystallized no ± ± ± 0.03 HSA-1 Fraction V no ± 2.0** 15.2 ± 6.2tt 0.17 ± 0.07:j::j: HSA-2 Fraction V no ± 5.8 HSA-3 Fatty acid-free no ± 1.2 HSA-4 Crystallized no ± 2.6 % 28.8± ± ± ± ± ± 0.18 * Values are means ± SE. Analyses of variance followed by Fisher's protected least significant difference post hoc tests were applied to results for each group of four albumins. t Lipid transfer assays were performed in triplicate, three or four separate times. Albumins were assayed at 16 mg/ml because this concentration was consistently within the linear range of the assay. :j: Sperm penetration assays were performed in BWW medium containing 0.4 % of the indicated albumin, with five separate replications for each preparation. BSA-1lipid transfer activity is significantly greater than that of any other BSA (P < 0.05). II BSA-2 percent penetration is significantly lower than those of all other BSAs (P < 0.05). \[ BSA-2 penetration index is significantly lower than those of all other BSAs (P < 0.05). ** HSA-2 lipid transfer activity is significantly greater than that of any other HSA, and HSA-4 lipid transfer activity is significantly greater than those of HSA-1 or HSA-3 (P < 0.05). tt HSA-2 and HSA-4 percent penetrations are significantly greater than those of HSA-1 and HSA-3 (P < 0.05). :j::j: HSA-2 and HSA-4 penetration indexes are significantly greater than those of HSA-1 and HSA-3 (P < 0.05). assayed in the ultrashort SPA and for sperm acrosome reaction assessment. Other Methods Protein concentrations were determined by the micro-bradford protein assay. Progesterone (P) was analyzed by radioimmunoassay (RIA, Coat-A Count; Diagnostic Products Corp., Los Angeles, CA) in albumin samples used in experiments 2, 3, and 4. Interassay coefficient of variation was 10.3% at 7.8 ng/ml. The RIA was validated for BWW-HSA by addition of known concentrations (1 to 100 ng/ml) of P standards, resulting in an average recovery of 105% (95% at 1 ng/ml). Data were analyzed by ANOV A with Fisher's protected least significant differences test as a posthoc test, regression analysis, or Student's paired t-test, as appropriate. RESULTS Lipid Transfer Activity in Albumin Preparations We have previously shown that fraction V BSA no has trace amounts of lipid transfer activity with poor dose-response characteristics (12). In five experiments, we measured lipid transfer activity in eight albumin preparations obtained from Sigma Chemical Co: four of BSA and four of HSA. Of the BSA preparations, only fraction V no had de- tectable lipid transfer activity at all concentrations tested (4, 8, 16, or 32 mg/ml final protein concentration); another fraction V preparation, no. 9647, had no lipid transfer activity (Table 1). In contrast, each of the HSA preparations had lipid transfer activity. One fraction V-HSA preparation, no. 1653, had significantly less (P < 0.05) lipid transfer activity than did fraction V-HSA no (Table 1). Fraction V-HSA no and crystallized-hsa no had the highest lipid transfer activity amounts and were the only ones of these eight to exhibit a dose response (Fig. 1; BSA data not shown). At the highest dose tested (32 mg/ml), the Sigma BSA preparations exhibited only 0% to 13% transfer. In different experiments, albumin preparations from other companies also exhibited lipid transfer activity (assayed at 16 mg/ml), although in widely variable amounts (Miles Pentex fraction V -BSA: 7.6; Calbiochem crystallized-bsa: 48.4; Armour Albuminar-5: 3.6; Irvine fraction V-BSA: 3.1, fatty acidfree-bsa: 8.7, and fraction V -HSA: 24.3 % transfer/ 18 hours). These albumin preparations were not tested further. Sperm Penetration Activity in Albumin Preparations In experiment 1 each Sigma albumin preparation (BSA or HSA) was also tested for the ability to support or enhance in vitro sperm capacitation and acrosome reaction, measured using the ultrashort SPA. Similar to the lipid transfer activity results, BSA Vol. 59, No.3, March 1993 Ravnik et al. Albumin, LTP-I, and sperm capacitation 633

6 60.. c ! "0 a. 30 ::::i., FrVHSA FrVHSA 1 FAFHSA.,..._ CrystHSA Final protein concentration (mgjml) Figure 1 Lipid transfer activity in HSA preparations. Four different preparations of HSA purchased from Sigma were assayed for lipid transfer activity. Twenty-five to 200 JLL of a 10% solution of albumin was in the transfer assay, giving the final protein concentrations shown (4 to 32 mg/ml). Regression analysis tested the probability that the slopes were different from zero. Significant dose responses were found for both fraction V-HSA no (fraction V-HSA 2,. P < 0.01) and crystallized-hsa no (, P < 0.01). A much weaker response was produced by another fraction V-HSA, no. 1653, (fraction V-HSA 1, 6., P > 0.05) and by fatty acid-free-hsa (e, P > 0.05). Values given are means ± SE, n = 5. such as LTP-I. In silver-stained SDS-polyacrylamide gel electrophoresis (PAGE) gels, a high degree of contamination by other proteins in the albumin preparations is demonstrated (Fig. 2A). To test if L TP-I was a contaminant in the albumin preparations, two different preparations, fraction V-HSA no (high lipid transfer activity, high penetration index) and fatty acid-free-hsa no (low lipid transfer activity, low penetration index) were western blotted after SDS-PAGE and probed with anti-ltp-1 antibodies (Fig. 2B). This analysis revealed a reactive band at approximately 64 kd, the reported size of LTP-1 on SDS-PAGE, for the fraction V-HSA. In contrast, a similar band for fatty acid-free-hsa was only lightly stained. We also carried out quantitative immunoblot analysis of each HSA to determine the relative LTP-1 immunoreactivities of each. Albumins with the highest anti LTP-1 immunoreactivity, fraction V-HSA no and crystallized-hsa no. 8763, also had the highest lipid transfer activity and highest sperm penetration index. 158 Kd preparations supported only a low to moderate level of SPA activity (Table 1). Fraction V-BSA no was significantly better than fraction V-BSA no (n = 5, P < 0.05) in this respect. The HSA preparations overall did better than BSA in supporting the ultrashort SPA, and there were significant differences in the responses to different preparations (Table 1). As with lipid transfer activity, fraction V-HSA no stimulated higher SPA activity than fraction V-HSA no (n = 6, p < 0.05). These data appeared to show a relationship between the ability of an albumin preparation to support capacitation and the amount of lipid transfer activity present in that preparation. Regression analysis demonstrated that lipid transfer activity was significantly correlated with penetration index (r = 0.76, P < 0.001, y = 0.01X ). Silver Gel and Western Blot Analysis of Albumin Preparations The differences in lipid transfer activity and capacitation support between albumin preparations may be explained in part by compositional differences or by contamination with other proteins 68 Kd 44 Kd 29 Kd MW A 8 Figure 2 Western blot analysis and SDS-PAGE ofhsa preparations. Human serum albumin-preparations from Sigma were electrophoretically separated on SDS-polyacrylamide gels. (A), 0.1 JLg of each preparation was silver stained as described in Materials and Methods. Molecular weight markers are shown on the left. Lane 1: Fraction V-HSA no. 2386; lane 2: Fraction v.-hsa no. 1653; lane 3: Fatty acid-free-hsa no. 1887; lane 4: Crystailized HSA no (B), 10 JLg of fraction V-HSA no (lanes 1 and 3) and of fatty acid-free-hsa no (lanes 2 and 4) were either silver stained (lanes 1 and 2) or electroblotted onto nitrocellulose for Western blot analysis using anti-ltp-i (lanes 3 and4). 634 Ravnik et al. Albumin, LTP-l, and sperm capacitation Fertility and Sterility

7 Removal ofltp-i From Albumin Preparations In the next series of experiments, we studied the requirement of LTP -I contamination in an albumin preparation for support of capacitation. We employed anti-ltp-i affinity chromatography to remove LTP-I from crystallized-hsa no and assayed the resulting albumin for support of sperm SPA and acrosome reaction activities. Passage of a 4% albumin solution (lipid transfer activity = 6.8% transfer) over an anti-ltp-i sepharose column removed all of the lipid transfer activity, whereas 85% of lipid transfer activity (5.8% transfer) was recovered from a control column (antihuman IgG sepharose). Lipid transfer protein eluted from the anti LTP-Icolumn with 3 M NaSCN had 3.19% transfer for a 47% recovery, whereas 3 M NaSCN elution from the control column had no activity. These results were confirmed by Western blotting of the unbound fractions (Fig. 3A). The additional bands recognized by the affinity purified anti -LTP -I antibody are probably LTP-I breakdown products because similar bands are occasionally seen in purified LTP I preparations not using immunoaffinity purification (Albers JJ, unpublished observations). Progesterone concentrations in the 40 mgjml albumin preparations were very low (0.3 to 0.7 ngjml) and were undetectable «0.2 ngjml) when diluted to 4 mgjml final concentrations. Penetration Activity in LTP-I-Free Albumin Preparations Experiment 2 We next assayed the affinity column treated HSAs (unbound fractions) for their abilities to support human sperm capacitation and acrosome reaction. Sperm were washed in BWW containing fraction V-BSA no (baseline control), control HSA (anti-igg column recovered), or LTP-I-free HSA (anti-ltp-i column recovered) each at a final concentration of 0.4% and incubated for 2 hours before adding sperm to eggs to assess their penetration abilities. In four experiments, significantly (P < 0.01) more sperm penetrated eggs when incubated with control HSA than with LTP-I-free HSA or with BSA (Fig. 3B). In addition, at 0, 1, 2, and 18 hours an aliquot from each treatment was prep/ilred for acrosome reaction assessment. Figure 4A shows that significantly more live sperm were acrosome reacted at 1, 2, or 18 hours when incubated with control HSA than when sperm were incubated with LTP-I-free HSA (P < 0.05, n = 3). A B ~ =.3 I 1i.2 i.f 2 BSA Control HSA L TP free HSA Figure 3 Removal of LTP-I from human albumin decreases its ability to support human sperm capacitation (experiment 2). (A), The unbound fractions from either anti-igg or anti-ltp-i columns were separated by SDS-PAGE and electroblotted onto nitrocellulose for Western analysis using anti-ltp-i antibody. Ten micrograms of protein were loaded into each lane. Lane 1: Unbound crystallized-hsa no passed over anti-ltp-i column. Lane 2: Unbound crystallized-hsa no passed over anti-igg column. Molecular weights are shown on the left, and the arrow points to the major LTP-I band at 64K. (B), Sperm incubated for 2 hours in BWW with either 0.4% BSA, control HSA (anti-igg unbound fraction), or LTP-I-free HSA (anti-ltp-i unbound fraction) were tested in the ultrashort SPA (experiment 2). The mean penetration index is shown as bars with the lines representing the SD, n = 4. Control HSA produced significantly more penetrating sperm per oocyte than either BSA or LTP-Ifree HSA, P < 0.001, Student's paired t-test. The numbers above each bar are the mean percent penetration. Experiment 3 In this experiment, repeated three times, sperm were washed and incubated for 2 hours in BWW with 0.4%,2.0%, or 3.5% of either control HSA or LTP-I-free HSA, and then sperm and eggs were coincubated. In Figure 4B, control HSA exhibits a dose response for penetration index, but LTP-I -free HSA does not. Aliquots were also taken for acrosome reaction assessment. Control HSA had 11.3% ± 1.2% acrosome reaction, 21.7% ± 1.5% acrosome reaction, and 29.7% ± 3.1% acrosome reaction at 0.4%,2.0%, and 3.5% concentrations, respectively, whereas LTP-I-free HSA produced only 4% ± 1.7%, 3.3% ± 2.3%, and 3.7% ± 0.6% acrosome reaction, respectively. Experiment 4 In this experiment, sperm were washed with 0.4% LTP-I-free HSA and then half the sperm were treated for 15 minutes with 3.5 p,g of purified LTP-I (11.0% transfer). The sperm were coincubated with eggs, and an aliquot was taken for acrosome reaction Vol. 59, No.3, March 1993 Ravnik et al. Albumin, LTP-I, and sperm capacitation 635

8 E & 20 1 u I!! I u c '#. I IS 1 l Figure 4 0 II BSA ControlHSA II LTP-free HSA Tlme(hra) o Final Concenlrallon (%) Control HSA Time course and dose response of control versus LTP-Ifree HSA in stimulation of sperm. (A), Time course of sperm acrosome reaction after incubation with treated HSAs (experiment 2). Capacitation support was assessed by measuring the percent live acrosome-reacted human sperm using PSA lectin staining. Washed human sperm were incubated for 0,1,2, and 18 hours in BWW containing 4 mg of the appropriate albumin per ml. Bars are means, lines are SD, and asterisks indicate that % acrosome reaction response with control HSA is statistically highe~,than with LTP-I-free HSA at the same time (Student's paired t-test, P < 0.05, n = 3). (B), Dose response of sperm penetration stimulated by treated HSAs (experiment 3). Human sperm were washed and incubated for 2 hours in either control HSA or LTP-I-free HSA at concentrations of 0.4%, 2.0%, or 3.5% and tested in the ultrashort SPA. The figure shows the dose of albumin plotted against the penetration index. In the three experiments, control HSA C'-) gave a significant dose response (regression analysis revealed the slope was significantly different from 0, P < 0.05, r = 0.72, y = O.lx + 0.2), whereas LTP-I-free HSA (e) gave no dose response (P> 0.3, r = 0.4, y = 0.008x ). At each dose, control HSA gave significantly higher penetration indices than did LTP-I-free HSA (Student's paired t-test, P < 0.05, n = 3). assessment. As in the previous experiments, LTP-Ifree HSA did not support penetration or acrosome reaction, but the addition of LTP-I restored the ability of that albumin to support capacitation. Sperm treated with LTP-I-free HSA + LTP-I gave a significantly higher sperm penetration index than sperm treated with LTP-I-free HSA alone (0.32 ± 0.07 and 0.07 ± 0.02, respectively, Student's paired t-test, P < 0.05, n = 3). A similar result was seen for percent acrosome reaction (LTP-Ifree HSA + LTP-I: 21.7% ± 3.7%; LTP-I-free HSA alone: 4% ± 1%, Student's paired t-test, P < 0.05, n = 3). DISCUSSION The use of commercially available albumin preparations in media for in vitro capacitation and fertilization by mammalian sperm has proven to be successful. However, the mechanisms by which albumin is able to support in vitro capacitation are largely unknown. In this study, our data suggest that LTP-I, as an active contaminant, has a role in the ability of albumin preparations to support in vitro sperm capacitation. We have previously shown that there is a relationship between lipid transfer activity and stimulation of human sperm capacitation by human FF (12). In the present study, preparations of albumin that have active LTP-I are also able to support sperm capacitation and acrosome reaction, whereas those albumins with low or no detectable lipid transfer activity afford poor support to these sperm functions. In addition, if LTP-I is removed from an albumin that previously had lipid transfer activity, that albumin is no longer able to support as high a level of capacitation, unless L TP -I is added back. These data strongly suggest that LTP-I plays a significant role in the ability of some albumin preparations to effectively support sperm capacitation. These data also support the possibility that L TP-I may be important during in vivo capacitation because it is a serum protein (14) and is found in FF (12,13). Although many activities have been attributed to albumin in support of in vitro sperm capacitation, these effects could be due to the high capacity, low affinity binding of other molecules to albumin. Recent evidence suggests that P can stimulate human sperm acrosome reactions (18). To determine if P is present in the albumin preparations, we performed P RIAs on them. It is unlikely that P is playing a role in albumin-supported capacitation, because P concentrations are well below the reported optimal doses of 5 to 1,000 ng/ml and because P does not stimulate acrosome reactions in sperm incubated for as short a time as we have employed (18, 19). AI- 636 Ravnik et ai. Albumin, LTP-I, and sperm capacitation Fertility qnd Sterility

9 teration of sperm surface proteins by albumin (8, 20) could be due to proteases in the albumin preparation. It is possible that these and other albuminassociated activities, e.g., ability to chelate zinc (9), function in concert. Likewise, the reported removal of sperm membrane cholesterol by albumin (10, 21) could actually be due to the presence of LTP-I, with the albumin possibly acting as a lipid acceptor. In the absence of LTP-I, in vitro support of capacitation by albumin is not nullified but is greatly reduced, as evidenced by the results seen after 18 hours incubation (experiment 2, Fig. 4A). This may be evidence of other capacitation support activities in the albumin preparation. Alternatively, some sperm may acrosome react in the absence of stimulatory signals. There appears to be a population of sperm that acrosome react spontaneously (e.g., in the absence of the ZP) through an undetermined mechanism, in most studies of human sperm acrosome reactions. The biological significance of such sperm is unknown. It is also possible that lipid transfer activity is not an absolute requirement for support of capacitation by albumin but rather is responsible for accelerated rates of capacitation. If LTP -I is present throughout the female reproductive tract, this accelerated rate may in fact be the normal situation. Our laboratory is currently exploring this possibility. Also, L TP -I treatment of sperm to stimulate capacitation may produce a population of sperm that are synchronously capacitated and able to spontaneously acrosome react or respond to physiological acrosome reaction stimulators. In light of the results presented here, experimental evidence for this would require an in vitro capacitation system that did not use albumin in the medium. The albumin preparations we tested were produced by the Cohn fractionation technique, or modifications thereof (22-24). These methods of separation are crude at best, as shown by silver stains of SDS-PAGE of the preparations (Fig. 2A). Tests of purity in the original fractionation studies (22-24) were the determination of sedimentation rates or insensitive protein stains of electrophoretic gels. It is not surprising that the more sensitive silver stain would show more contaminants. In addition, the existence of serum molecules with physicochemical characteristics similar to albumin (e.g., LTP-I) (14) may have been unknown at the time. The method by Cohn et al. (22, 23), employing simple variations in ethanol concentration, temperature, and ph to precipitate albumin, could indeed produce a mixture of albumin, other proteins, and active LTP-I if these molecules share some characteristics. A major problem with the use of albumin is the variability of different albumin preparations, and indeed lots, in providing consistent in vitro sperm support (7). We examined two preparations each of fraction V HSA and BSA. There was almost a 10- fold difference in lipid transfer activity between the two HSA preparations, with approximately a twofold difference in SPA activities. The BSA preparations exhibited approximately a fivefold difference in SPA activities. Interestingly, further purification (crystallization, charcoal extraction) of the BSA appeared to reduce its lipid transfer activity, whereas the lipid transfer and SPA activities increased with crystallization of the HSA. This discrepancy could be due to differences between bovine and human lipid transfer proteins. These results point again to the need to test albumin preparations and lots for consistency. The standardization of a protein supplement for sperm capacitation media is an important issue because it is presently difficult to compare results from different laboratories. The choice of a protein is complicated by variations in production, expense, and medical concerns. Bovine serum albumin has gained popularity in human assisted reproductive technologies because of the fear of viruses in human serum preparations. Recently, the use of non-albumin proteins such as gelatin has been suggested (25). Gelatin preparations tested in our lab do not have lipid transfer activity. (Ravnik SE, Muller CH, unpublished observations). The use of homologous maternal serum addresses clinical concerns but introduces uncontrolled variation into the methods. Use of pooled sera (fetal or adult), tested for known pathogens, seems a compromise that completely addresses neither the clinical nor the variation problems. Perhaps the combination of an inert but clinically safe carrier macromolecule with a highly purified or synthetic capacitation-support molecule will eventually prove to be desirable. We previously have shown that LTP-I and LTP-Icontaining fluids are potent enhancers of capacitation (12, 13). In the present study, we have presented evidence that some commercial preparations of serum albumin have LTP-I as a contaminant. Albumins that effectively support sperm capacitation and acrosome reaction (as determined by SPA and acrosome reaction assessment) have more lipid transfer activity than those that have lower levels of support. In addition, if LTP-I is removed from Vol. 59, No.3, March 1993 Ravnik et al. Albumin, LTP-I, and sperm capacitation 637

10 albumin, the degree of support is significantly reduced. It is likely that in LTP-I-contaminated albumin, the albumin molecule itself has a role as a lipid acceptor in the absence of other appropriate molecules such as HDL. We are carrying out studies to determine the requirements of different lipid acceptors (i.e., albumin, HDL, artificialliposomes) for LTP-I-induced capacitation in protein-free media. Acknowledgments. We thank Janet Huson, B.A., for technical assistance with some of the SPAs and Timothy Ewers, B.A., for assistance with the SDS-PAGE and Western blotting. Bert Toivola, Ph.D., Department of Laboratory Medicine, University of Washington, Seattle, Washington, assisted with the P RIA. Irvine albumins were a gift from Paul Weathersbee, Ph.D., of Irvine Scientific. John Tollefson, B.A., (deceased January 1991) is fondly and gratefully remembered for his technical skills in producing LTP-I and LTP-I antibodies and for his thoughtful comments during the course of this work. REFERENCES 1. Yanagimachi R, Chang MC. Fertilization of hamster eggs in vitro. Nature 1963;200: Barros C, Austin CR. In vitro fertilization and sperm acrosome reaction in the hamster. J Exp ZooI1967;166: Yanagimachi R. In vitro capacitation of golden hamster spermatozoa by homologous and heterologous blood sera. BioI Reprod 1970;3: Toyoda Y, Yokoyama M, Hoshi T. Studies on the fertilization of mouse eggs in vitro. I. In vitro fertilization of eggs by fresh epididymal sperm. Jpn J Anim Reprod 1971;16: Bavister BD. Capacitation of golden hamster spermatozoa during incubation in culture medium. J Reprod FertiI1973;35: Rogers BJ. Mammalian sperm capacitation and fertilization in vitro: a critique or'methodology. Gamete Res 1978;1: Bronson RA, Rogers BJ. Pitfalls of the zona-free hamster egg penetration test: protein source as a major variable. Fertil Steril1988;50: Davis BK, Gergely AF. Studies on the mechanism of capacitation: changes in plasma membrane proteins of rat spermatozoa during incubation in vitro. Biochem Biophys Res Commun 1979;88: Aonuma S, Okabe M, Kishi Y, Kawaguchi M, Yamada H. Capacitation inducing activity of serum albumin in fertilization of mouse ova in vitro. J Pharmacobiodyn 1982;5: Go KJ, WolfDP. Albumin mediated changes in sperm sterol content during capacitation. BioI Reprod 1985;32: Lui CW, Cornett LE, Meizel S. Identification of the bovine follicular fluid protein involved in the in vitro induction of the hamster sperm acrosome reaction. BioI Reprod 1977;17: Ravnik SE, Zarutskie PW, Muller CH. Lipid transfer activity in human follicular fluid: relation to human sperm capacitation. J Androl 1990;11: Ravnik SE, Zarutskie PW, Muller CH. Purification and characterization of a human follicular fluid lipid transfer protein that stimulates human sperm capacitation. BioI Reprod 1992;47: Albers JJ, Tollefson JH, Chen C-H, Steinmetz A. Isolation and characterization of human plasma lipid transfer proteins. Arteriosclerosis 1984;4: Faust RA, Albers JJ. Synthesis and secretion of plasma cholesteryl ester transfer protein by human hepatocarcinoma cell line, HepG2. Arteriosclerosis 1987;7: Cross NL, Morales P, Overstreet JW, Hanson FW. Two simple methods for detecting acrosome-reacted human sperm. Gamete Res 1986;15: Heukeshoven S, Demick R. Improved silver staining procedure for fast staining in Phast System Development Unit I. Staining of sodium dodecyl sulfate gels. Electrophoresis 1988;9: Meizel S, Pillai MC, Diaz-Perez E, Thomas P. Initiation of the human sperm acrosome reaction by components of human follicular fluid and cumulus secretions including steroids. In: Bavister BD, Cummins J, Roldan ERS, editors. Fertilization in mammals. Norwell (MA): Serono Symposia USA, 1990: Margalioth EJ, Bronson RA, Cooper GW, Rosenfeld DL. Luteal phase sera and progesterone enhance sperm penetration in the hamster egg assay. Fertil Steril1988;50: Meizel S. Molecules that initiate or help stimulate the acrosome reaction by their interaction with the mammalian sperm surface. Am J Anat 1985;174: Davis BK. Uterine fluid proteins bind sperm cholesterol during capacitation in the rabbit. Experientia 1982;38: Cohn EJ, Strong LE, Hughes WL Jr., Mulford DJ, Ashworth IN, Melin M, et al. Preparation and properties of serum and plasma proteins. IV. A system for the separation into fractions of the proteins and lipoprotein components of biological tissues and fluids. JAm Chem Soc 1946;68: Cohn EJ, Hughes WL Jr, Weare JH. Preparation and properties of serum and plasma proteins. XIII. Crystallization of serum albumins from ethanol-water mixtures. J Am Chem Soc 1947;69: Chen RF. Removal of fatty acids from serum albumin by charcoal treatment. J BioI Chem 1967;242: Mack SO, Tash JS, Wolf DP. Human sperm capacitation in gelatin fortified medium. Fertil Steril1989;52: Ravnik et a!. Albumin, LTP-I, and sperm capacitation Fertility and Sterility