Biotechnol. Prog. 1993, 9, 146-152
146
Insect Cell Hosts for Baculovirus Expression Vectors Contain Endogenous Exoglycosidase Activity Peter J. Licari,+JDonald L. Jarvis,§ and James E. Bailey**+ Department of Chemical Engineering, California Institute of Technology, Pasadena, California 91125, and Department of Entomology, Texas A&M University, College Station, Texas 77843-2475
Four different insect cell lines that can be used as hosts for baculovirus infection were assayed for the presence of endogenous exoglycosidases. All four cell lines, derived from Spodoptera frugiperda, Trichoplusia ni, Bombyx mori, or Malacosoma disstria, contained N-acetyl-p-glucosaminidase, N-acetyl-P-galactosaminidase,P-galactosidase, and sialidase activities. Exoglycosidase activities were found in cell lysates as well as cell-free supernatants from uninfected and wild-type baculovirus infected cells. Oligosaccharide analysis of cellular glycoproteins using lectins recognizing GalP1, 3GalNAc, Galpl,4GlcNAc, and NeuAca2,6Gal demonstrated that only Gal@l,3GalNAc was present. The demonstration that these cells contain exoglycosidases raises the possibility that the oligosaccharides of baculovirus-expressed glycoproteins are subject to enzymatic degradation.
Introduction The insect cell-baculovirus expression system is a popular means of expressing recombinant proteins (Luckow, 1991; Luckow and Summers, 1988; Miller, 1988). By replacing the baculovirus polyhedrin structural gene with the gene of interest, recombinant proteins can be produced in high yields relative to other eukaryotic expression systems. Synthesis of heterologous proteins begins at approximately 24 h post-infection and continues until approximately 4-5 days post-infection, when the cells lyse. A number of Lepidopteran insects and cell lines derived from these insects are susceptible to baculovirus infection and can be used as hosts for heterologous protein expression. Four examples are insect cell lines derived from Spodoptera frugiperda (fall army worm), Trichoplusia ni (cabbage looper), Bombyx mori (silkworm), and M a lacosoma disstria (forest tent caterpillar). The insect cell-baculovirus system is a promising tool for the expression of heterologous glycoproteins because insect cells are capable of both N- and 0-linked glycosylation. However,glycoproteins produced in mammalian systems are usually slightly larger than the same proteins expressed using the baculovirus expression system. This size difference has been attributed to differences in the structures of the oligosaccharide side chains on these proteins. Most evidence suggests that insect cells can add oligomannosidic N-linked oligosaccharides to newly synthesized glycoproteins and convert them to endo-8-Nacetyl-D-glucosaminidase H (endo H)resistant structures, but they cannot convert them to the complex structures found in higher eukaryotes (Butters et al., 1981; Hsieh and Robbins, 1984; Jarvis and Summers, 1989;Kuroda et al., 1990). Similarly,0-linked oligosaccharidessynthesized in insect cells appear to lack the terminal sialic acid residues that are often found in higher organisms (Thomsen et al.,
* Author to whom reprint requests should be addressed. Current address: Institute fur Biotechnologie, ETH-Honggerberg, CH-8093, Zurich, Switzerland. California Institute of Technology. Current address: Merck & Co., Inc., P.O. Box 2000, RY810-116, Rahway, NJ 07065-0900. 5 Texas A&M University. +
*
8756-7938/93/3009-0146$04.00/0
1990). On the basis of these findings, most investigators believe that insect cells lack the enzymes required for the latter steps in N- and 0-linked oligosaccharide processing. However, several recent reports from one laboratory have contradicted this conclusionand suggested that insect cells are capable of terminal N-glycosylation events, including the addition of sialic acid (Davidson et al., 1991;Davidson and Castellino, 1991; Davidson et al., 1990). Presently, there is no explanation for the discrepancy in these results. Exoglycosidases are enzymes responsible for the hydrolysis of an 0-glycosidic bond in a nonreducing terminal monosaccharide. N-Acetyl-@-glucosaminidaseand N-acetyl-8-galactosaminidase hydrolyze the glycosidic bond of the corresponding N-acetyl-@-hexosaminewhen it occupies a terminal position on an oligosaccharideside chain. These two different activities have been attributed to a single enzyme originally found in ram testes extract (Woolen et al., 1961a). This lack of specificity is reflected in the preferred name, N-acetyl-@hexosaminidase. Common sources of N-acetyl-@-hexosaminidaseinclude ram testes (Woolen et al., 1961b),jack bean meal (Li and Li, 1970), and Streptococcus pneumoniae (Hughes and Jeanloz, 1964a,b). Other exoglycosidasesare @-galactosidase,which hydrolyzes the glycosidic bond of terminal @-galactoside residues, and sialidase, which hydrolyzes the glycosidic bond of terminal sialic acids. The specificity of exoglycosidases depends on the cell type from which they are isolated. For example, 8-galactosidase isolated from jack bean meal cleaves P1,4 linkages faster than P1,3 linkages (Arakawa et al., 1974), whereas @-galactosidaseisolated from bovine testes preferably cleaves 81,3 linkages (Li et al., 1975). Sialidase isolated from Clostridiumperfringens favors the hydrolysis of a2,3linkages (Corfieldet al., 1981), whereas Arthrobacter ureafaciens sialidase prefers a2,6 linkages (Uchidaet al., 1979). In spite of thesepreferences, both sialidases can hydrolyze a2,3, cu2,6,and a2,8 linkages. Besides their monosaccharide and glycosidic bond specificities, the activity of exoglycosidases depends on the precise structure of the oligosaccharideor the glycoprotein to which it is attached ("aglycon specificity": Kobata (1979)). It was recently reported that Chinese hamster ovary cells contain a stable sialidase activity that can
0 1993 American Chemical Society and American Institute of Chemical Engineers
Bktechnd.
Rw.,1993, Vol. 9, NO. 2
accumulate in the extracellular environment and release sialic acid from a glycoprotein at mammalian cell culture pH (Goochee and Gramer, 1992). The purpose of this study was to test four different insect cell lines for the presence of endogenous exoglycosidases, includingN-acetyl-8-glucoeaminidase,N-acetyl8-galactosaminidase, ,!?-galactosidase,and sialidase. The results of these studies showed that all four of these cell lines contain substantial levels of N-acetyl-p-hexosaminidase and sialidase activities, as well as low levels of &galactosidase activity.
Materials and Methods Cell Lines, Culture Conditions, and Harvesting Procedure. S.frugiperda (Sf9)cells were obtained from the ATCC (No. CRL-1711); the T. ni (TN-368), B. mori, andM. disstria (MD108)cell lines were originally obtained from Dr. Max Summers. The Sf9 and TN-368 cell lines are derived from ovarian tissue, the B. mori cell line from embryos, and MDlO8 from larval hemocytes. All cells were cultured at 28 "C in stationary T-25 flasks containing 4.5 mL of TNM-FH medium (Hink, 1970) supplemented with 10% fetal bovine serum (Hyclone, Logan, UT), 5000 unita/L penicillin, and 5 mg/L streptomycin (Gibco BRL, Grand Island, NY). All cell lines were also adapted to a protein-free medium, Insect-Xpress (Whittaker Bioproducta, Walkersville, MD). Unless otherwise noted, experiments were performed using cells grown in TNM-FH. Insect cell culture media has aslightly acidic pH (6.0-6.4). Cells in the late exponential growth phase (=lo6 cells/ mL, >95% viable) were harvested by centrifugation at lOOOg for 5 min. The supernatant was saved and the cell pellet was washed twice with 1mL of phosphate-buffered saline (PBS). Cells were resuspended in PBS and subjected to three rounds of freeze-thaw and mechanical homogenization. Homogenate5 were then used directly in exoglycosidase assays. The total protein content of samples was determined using the Bio-Rad Protein Assay (Bio-Rad, Richmond, CA),an assay based on the Bradford colorimetric procedure (Bradford, 1976). Due to the high protein concentrations of these samples (7-10 g/L), exoglycosidase substrate concentrations in the assays described below were not significantly affected by the addition of cell lysate. Controls for all experiments consisted of an equal volume of PBS (without cell lysate) added to the exoglycosidase substrates. Infection Conditions. Cells in the exponential growth phase were infected with wild-type Autographa californica multicapsid nuclear polyhedrosis virus (AcMNPV) at a multiplicity of infection of 10 plaque-forming units/cell. Infected culture samples were centrifuged in a microfuge to remove cellular debris, and the supernatant was subsequently stored at -20 "C. N-Acetyl-&hexosaminidase Activity Assay. p-Nitrophenyl N-acetyl-8-D-glucosaminideand p-nitrophenyl N-acetyl-8-D-galactosaminide (Sigma, St. Louis, MO; Borooah, 1961) were dissolved in HzO at a concentration of 1g/L. Further dilutions were made in H20 as required. The substrate was preincubated at the appropriate temperature prior to the beginning of the assay. Cell lysate or supernatant and 1mL of substrate solution were added to microcuvettes. Quantities of cell lysate or supernatant used are indicated in the appropriate figure and table legends. The absorbance at 495 nm was read in 2-min intervals for 20 min at constant temperature. pH values were monitored before and after absorbance readings to assure a constant pH (5.5). Water was used instead of buffer because the high protein content of samples tended
147
to cause precipitation; activity was also observed using McIlvaine buffer (0.05M sodium phosphate, 0.025 M citric acid, pH 5.5; Montreuilet al., 1987). The rate of hydrolysis is expressed in terms of micromoles of p-nitrophenyl released per minute. @-GalactosidaseActivity Assay. @-Galactosidase activity was determined using 0-nitrophenyl &galactoside (Sigma) and resorufin &galactoside (Boehringer Mannheim, Indianapolis, IN; Hofman and Sernetz, 1984). The assay employing 0-nitrophenyl &galactoside has been described elsewhere (Licari and Bailey, 1991). Resorufin &galactoside was dissolved in dimethyl sulfoxide (25 mM). This solution was diluted in PBS to a concentration of 0.05 mM (pH 6.0). For spectrophotometric studies, cell lysate was added to 1 mL of substrate solution. For spectrofluorometric studies, cell lysate was added to 3 mL of substrate solution. The release of resorufiin (7-hydroxy2-phenoxazone)was monitored via spectrophotometry and spectrofluorometry as indicated (absorbance maximum at 572 nm and emission maximum at 583 nm; extinction coefficientat 572 nm = 66 L mmol-' cm-l). ,&Galactosidase purified from bovine testes was used as a standard to allow the conversion of fluorometric readings to activity units. One activity unit is defined as the amount of 8-galactosidase required to hydrolyze 1pmol of resorufii &galactoside within 1 min at pH 4.3 and 37 "C. Sialidase Activity Assay. Cell lysate or supernatant was added to 5 p L of 50 mM 2'-(4-methylumbelliferyl)a-D-N-acetylneuraminic acid (Boehringer Mannheim) and 25 pL of H20 (Meyers et al., 1980). The resulting pH was 5.5-6.0. Following incubation at 37 "C for 1 h, 20 pL of the reaction mixture was diluted with 1mL of H20. This dilution in HzO followed by immediate fluorometric readings was sufficient to slow the reaction to a rate that allowed accurate measurements. The sample was then excited at 365 nm, and the emission was read at 448 nm. Emission readings were converted to activity units by comparison with sialidase from Arthrobacter ureafaciens (Oxford Glycosystems, Rosedale, NY). One activity unit is defined as the amount of sialidase required to hydrolyze 1pmol of N-acetylneuraminosyl-D-lactose per min at pH 5 and 25 "C. Controls consisting of substrate with PBS and without cell lysate were used: the hydrolysis rate of these samples served as zero activity for rate calculations. Sialidase activity also was determined using fetuin (Boehringer Mannheim) as a substrate. A confluent monolayer of S. frugiperda cells (2 X 106 cells/mL) from a T-25 flask was harvested. Cells were washed twice with 1mL of PBS, resuspended in 300 r L of PBS, and sonicated, resulting in a total protein concentration of approximately 2.5 mg/mL. This sample was aliquoted into two tubes (150pLhube). Fetuin (100 pL) in PBS (1g/L) was added to one tube, and 100 pL of PBS without fetuin was added to the second tube, serving as a control. A second control consisting of 100 pL of fetuin in PBS (1g/L) and 150 pL of PBS without cell lysate was also prepared. All samples were vortexed briefly and then incubated at 37 "C for 18 h. N-Acetylneuraminic acid (NeuAc) content was then determined by high-pH anion exchange chromatography (HPAEC) as described below. High-pH Anion Exchange Chromatography. The HPAEC system consisted of a Dionex BIO LC liquid chromatography unit, a Carbopac PA1 column (4 X 250 mm), and a pulsed electrochemical detector operating in the pulsed amperometric detection mode (Dionex Corp., Sunnyvale, CAI. Samples were prepared as described in the text and centrifuged for 1 min in a microcentrifuge immediately prior to analysis. Then 20 pL of sample was
Biotechnol. Rog., 1993, Vol. 9, No. 2
148 7
9
6
I
I
I
,
8
5 h
.-E 5
7 4 h
.E 2 b
3
P - 6
z
2
3
>
"7
5
1
4
0
3
-1
0
40
80
120
160
200
Total protein (pg)
Figure 1. Rate of p-nitrophenyl hydrolysis (wmol of p-nitrophenyl released per min) from p-nitrophenyl N-acetyl-6-Dglucosaminide and p-nitrophenylN-acetyl-0-D-galactosaminide as a function of total cellular protein (pg) from S. frugiperda cells at 28 and 37 "C. S.frugiperda cell lysate contains substantial levels of N-acetyl-0-glucosaminidase, but not N-acetyl-p-galactosaminidase activity. Symbols correspond to p-nitrophenyl N-acetyl-fl-D-glucosaminidehydrolysis at 37 "C ( 0 )and at 28 "C (0) and p-nitrophenylN-acetyl-8-D-galactosaminide hydrolysis at 37 "C (I and ) at 28 "C (0). injected into the column and eluted with a constant concentration of NaOH (100 mM) and a linear gradient of sodium acetate (50-180 mM) over 20 min with a flow rate of 1 mL/min at room temperature. NeuAc was identified by its coelution with N-acetylneuraminic acid standard (Dionex). Lectin Analysis. Total cellular proteins were fractionated on an 8.7 7% SDS-polyacrylamide gel and transferred to nitrocellulose in a Hoefer blotting apparatus. Oligosaccharide moieties were detected by using a Glycan Differentiation Kit (Boehringer Mannheim). This method involves binding digoxigenin conjugated lectin to glycoprotein, followed by immunodetection of digoxigenin.
Rssults N-Acetyl-@-hexosaminidaseActivity. The hydrolysis of p-nitrophenyl N-acetyl-8-D-glucosaminide and p nitrophenyl N-acetyl-0-D-galactosaminidein the presence of increasing amounts of S. frugiperda cell lysate were investigated at 28 and 37 "C (Figure 1).Under the specified conditions, the p-nitrophenyl group was detectably hydrolyzed from p-nitrophenyl N-acetyl-8-D-glucosaminide, but not from p-nitrophenyl N-acetyl-P-D-galactosaminide. The rate of p-nitrophenyl N-acetyl-0-D-glucosaminide hydrolysis increased with a corresponding increase in total cellular protein. Thep-nitrophenylgroup was hydrolyzed more slowly at 28 "C than at 37 "C, but even at the lower temperature there was significant activity. The negative rate observed for the p-nitrophenyl N-acetyl-8-D-galactosaminide substrate is due to cell suspensions settling in the cuvette. (These experiments were carried out with cell lysate, resulting in a turbid solution that settled with time. After the soluble and insoluble fractions were separated by centrifugation, exoglycosidase activity was predominantly located in the soluble lysate fraction.) Hydrolysis of p-nitrophenyl N-acetyl-@-D-galactosaminide was observed with prolonged incubation at 37 "C (data not shown). The activity of N-acetyl-P-glucosaminidasefrom S. frugiperda cells followed Michaelis-Menten kinetics as indicated by a Lineweaver-Burk plot (Figure 2). The VMAXand K, for hydrolysis of p-nitrophenyl N-acetyl-
0
2
4
6
8
10
12
MSI
Figure 2. Lineweaver-Burk plot for hydrolysis of p-nitrophenyl N-acetyl-0-D-glucosaminideusing protein from s. frugiperda cells. Inverse rate, I/V (min/pmol), is plotted as a function of the inverse substrate concentration,1/[Sl(mL/pmol). Approximately 58 pg/mL was incubated with substrate at 37 "C. 8-D-glucosaminidewere calculated to be 2.7 X lo4 pmoll min and 0.13 pM, respectively, under the specified conditions (37 "C, 58 pg of S.frugiperda protein). Similar behavior was observed with extracts from T. ni, B. mori, and M.disstria cells. Table I shows the results obtained for the hydrolysis of p-nitrophenyl N-acetyl-P-D-glucosaminide at 37 "C with the different cell lines. To allow direct comparisons, specific activities based on total cellular protein are presented; the T. ni cell lysate demonstrated the greatest N-acetyl-8-glucosaminidase specific activity. Fresh TNM-FH with 10% serum, which had never been exposed to insect cells, had some endogenous N-acetylP-glucosaminidase activity (Table 11). However, this activity was increased significantly by incubation with any one of the four different insect cell lines. The cell-specific nature of the observed exoglycosidase activity was demonstrated by adapting each of these cell lines to InsectXpress medium, a protein-free medium that contained no detectable N-acetyl-8-glucosaminidaseactivity (Table 11). Cell extracts (data not shown) and the cell-freesupernatant fraction (Table 11)from all four cell lines grown in InsectXpress medium contained significant levels of N-acetyl(3-glucosaminidase activity. N-Acetyl-8-glucosaminidaseactivity in the supernatant of S. frugiperda cells infected with wild-type AcMNPV is presented in Figure 3 as a function of time post-infection. Throughout the infection, the level of extracellular N-acetyl-p-glucosaminidaseactivity was always considerably greater than the activity present in fresh TNM-FH medium containing 10% fetal bovine serum. A dramatic rise in extracellular N-acetyl-0-glucosaminidase activity was observed during the first 30 h of infection. The elevated levels of extracellular N-acetyl-8-glucosaminidase activity compared to those of fresh medium indicated a release of the exoglycosidase from the cells into the medium. &Galactosidase Activity. The presence of 8-galactosidase activity was investigated by using two different artificial substrates, 0-nitrophenyl @-galactosideand resorufin @-galactoside. No &galactosidase activity could be detected with 0-nitrophenyl /3-galactoside as the substrate. However, resorufin P-galactoside provided different results. In the presence of @-galactosidase, resorufin @-galactosideis cleaved to yield @-galactoside and resorufin. Resorufin may be detected by spectro-
Bbtect~nol.hog., 1993, Vol. 9, No. 2
140
Table I. Rate of Hydrolysis of pNitrophenyl from pNitrophenyl N-Acetyl-8-D-glucosaminide (pNP-GlcNAc) and pNitrophenyl N-Acetyl-8-~-galactosaminide (DNP-GalNAc) at 37 O c a specific activity PNP-GlcNAc pNP-GalNAc cell line (pmol/min.pg x 10-5) (pmol/min.pa) nd S. frugiperda 0.7 f 0.1 nd T.ni 1.3 f 0.1 nd B. mori 0.9 f 0.1 nd M.disstria 0.5 f 0.1 Specific activity is expressed in units of pmol of p-nitrophenyl released per min per pg of total cellular protein. Approximately equal quantities of total protein were used for each cell line (30 pg/ mL) with a substrate concentration of 1 mg/mL. nd indicates hydrolysis was not detected for the concentrations indicated. Table 11. Rate of Hydrolysis of gNitrophenyl from pNitrophenyl N-Acetyl-8-D-glucosaminide in the Presence of Cell Culture S u w r n a t a n t s supernatant derived from cell line medium TNM-FH S. frugiperda T. ni TNM-FH B. mori TNM-FH TNM-FH M.disstria TNM-FH no cells Insect-Xpress S. frugiperda T.ni Insect-Xpress B. mori Insect-Xpress Insect-Xpress M.disstria no cells Insect-Xpress
activity (pmol/min.pL x 10-6) 3.8 f 0.2 1.5 f 0.2 5.9 i 0.2 3.0 f 0.2
0.9 f 0.2 0.9 i 0.2 0.5 f 0.2 2.8 f 0.2 4.9 f 0.2 0.0 f 0.2
100 pL of clarified supernatant was added to 1 mL of 1 mg/mL p-nitrophenyl N-acetyl-B-D-glucosainide,and the rate of hydrolysis was monitored at 37 "C. For reference, N-acetyl-8-glucosaminidase activitiesof fresh TNM-FH medium and fresh Insect-Xpress (proteinfree medium) are also listed. (I
2
I
I
I
I
TNM-FH medium
0.8
1
Table 111. 8-Galactosidase Activity As Measured Using Resorufin 8-Galactoside as a Substrate a t 28 and 37
cell line S. frugiperda T.ni B. mori M.disstria
spectrophotometry response 28 O C 37 o c nd nd nd nd nd nd nd nd
fluorometry (munitslpg x 10-3) 28 "C 37 o c nd 1.5 f 0.2 0.8 f 0.2 1.3 f 0.2 1.4 f 0.2 2.8 f 0.2 1.5 i 0.2 2.9 f 0.2 ~
The rate of substrate hydrolysis was detected by spectrophotometry and spectrofluorometry as indicated. For the spectrophotometric response, total protein concentrations equal to those of Table I (30 pg/mL) were used to allow comparison. nd indicates hydrolysis was not detected.
fluorometry and increasing the protein concentration to 170 pg/mL, a characteristic color response was detected, indicating a relatively low, but measurable 8-galactosidase activity. By the same fluorescence assay, &galactosidase activity was detected in all cell lines at 37 "C;activity was also detected a t 28 O C for all cell lines except S.frugiperda (Table 111). A very high @-galactosidase activity was observed in fresh TNM-FH medium that was not exposed to cells (Table IV). Supernatants from the four cell lines grown in TNM-FH had a slightly lower activity than the fresh medium, perhaps due to 8-galactosidase degradation during the time of the cultivation. Similar supernatant analysis with cells cultivated in protein-free medium indicated a zero basal level of activity in fresh InsectXpress medium and an elevated level in medium removed from healthy cell cultures (Table IV). Furthermore, cells cultivated in Insect-Xpress hydrolyzed the resorufin 8-galactoside,indicating that the cellular activity presented in Table I11is not an artifact of residual serum-containing medium associated with the cells. SialidaseActivity. Varying quantities of S.frugiperda cell lysate were incubated with 2'-(4-methylumbelliferyl)a-D-N-acetylneuraminic acid as described in the Materials and Methods section. A linear response of sialidase activity was observed with increasing quantities of cell extract (Figure 4). Activity was not observed in fresh TNM-FH medium that was not exposed to cells, but was seen for TNM-FH that had been exposed to cells (Table V). In addition, sialidase activity was demonstrated to be present in the extracellular medium of wild-type AcMNPVinfected cells (data not shown). Although a pH optimum for this enzyme was not determined, sialidase activity was observed for pH values between 5.5 and 7.0 using substrate in phosphate-citrate buffer. The pH of insect cell culture media is slightly acidic: TNM-FH has a pH 6.0-6.4 and Insect-Xpress has a pH 6.0-6.1. T. ni, B. mori, and M . disstria cell lysates were also assayed for sialidase activity. Each had detectable sialidase activity at 37 "C when 2'-(4-methylumbelliferyl)a-D-N-acetylneuraminic acid was used as the substrate (Table V). T. ni, B. mori, and M.disstria cells contained more activity per milligram of total protein than S. frugiperda cells. There was little difference in the levels of sialidase activity detected in S. frugiperda cells at 28 and 37 O C . All cell lines also exhibited sialidase activity when cultured in protein-free medium (data not shown). Fetuin, a sialylated glycoprotein, was incubated with lysed, uninfected S. frugiperda cells for 18 h along with the appropriate controls as described in the Materials and Methods section. After incubation, the samples were analyzed by HPAEC for free N-acetylneuraminic acid (Figure 5). The peak area corresponding to NeuAc was 4 times greater for the sample containing fetuin (panel C) than the control sample containing an equal volume of
i 80
0
20
40
60
100
120
Time post-infection (h)
Figure 3. N-Acetyl-8-glucosaminidasein the supernatant of wild-type baculovirus infected S. frugiperda cells a t different times post-infection (h). Clarified supernatant (100 pL) was added to 1 mL of 1 mg/mL p-nitrophenyl N-acetyl-8-D-glucosaminide, and the rate of hydrolysis was monitored a t 37 O C . For reference, the N-acetyl-8-glucosaminidaseactivity of fresh TNM-FH medium containing 10%fetal bovine serum is indicated by the horizontal line.
photometric or spectrofluorometric techniques, the latter being more sensitive. Using conditions similar to those used in the N-acetyl-8-glucosaminidase studies, no 8-galactosidase activity was detected spectrophotometrically at 28 or 37 OC (Table 111). These data can be compared to those in Table I, as equal protein concentrations (30 pg/mL) were used and the response is based on a change in concentration over the same time scale. By using
~~
Blotechnol. Rog., 1993, Vol. 9, No. 2
150 12 1 10
I
I
I
I
I
I
i
“ I .d .
5 E
8
v
~-
i
2/L----J 0 0
20
60
40
80
100
120
140
Total Protein (pg)
Figure 4. Sialidase activity in S.frugiperda cell lysate. Sialidase activity (munita) was assayed by incubation with the substrate, 2’-(4-methylumbelliferyl)-a-~-N-acetylneuinicacid, and plotted as a function of total cellular protein (rg). The cell lysate was prepared as described in the Materials and Methods section prior to incubation with the substrate for 1 h at 37 O C . Table IV. @-GalactosidaseActivity in Cell Culture Supernatantss
supernatant derived from cell line medium
activity (munits/pL x 10-3) TNM-FH 1.5f 0.1 S. frugiperda TNM-FH 2.3 f 0.1 T. ni 1.4 f 0.1 TNM-FH B . mori 2.0 f 0.1 TNM-FH M . disstria 2.5 f 0.1 TNM-FH no cells 0.07 f 0.05 S. frugiperda Insect-Xpress 0.11 f 0.05 Insect-Xpress T . ni 0.07 f 0.05 Insect-Xpress B . mori 0.25 f 0.05 Insect-Xpress M . disstria 0.00 f 0.05 Insect-Xpress no cells Clarified supernatant was added to 3 mL of 0.05 mM resorufin /+galactoside, and the rate of hydrolysis was monitored at 37 O C . For reference, &galactosidase activities of fresh TNM-FH medium and fresh Insect-Xpress (protein-free medium) are also listed.
Table V. Sialidase Activity As Measured Using 2~-(4-Methylumbelliferyl)-a-~-N-acetylneuraminic Acide cell line T (“C) activity (munitsipg) S. frugiperda 28 0.07 f 0.02 S. frugiperda 37 0.08 f 0.01 37 0.14 f 0.01 T.ni B. mori 37 0.12 f 0.01 M . disstria 37 0.14 f 0.01 TNM-FH 37 0.00 f 0.01 TNM-FHISf9 37 0.01 f 0.01 0 Specific activity is expressed in munits per pg of total protein. Emission at 448 nm was detected after 1 h of incubation at the indicated temperature. TNM-FH/Sf9is medium harvested from a culture of S.frugiperda cells (lo6cells/mL, >90 viability) and clarified
prior to assay.
PBS and cells (panel B). No NeuAc was observed for the control consisting of fetuin in PBS (data not shown). Chromatogram A represents 1 nmol of NeuAc standard. From a standard curve, the peak area corresponding to NeuAc in chromatogram B is approximately 1 nmol and the peak area from chromatogram C is approximately 4 nmol. This data suggests that 3 nmol of NeuAc is released from fetuin during the incubation with cell lysate. Based on the quantity of fetuin added, approximately 10% of available NeuAc was hydrolyzed. The control containing cell lysate without fetuin has a peak which indicates that sialic acid exists in the insect cell lysate. A recent report demonstrates the existence of polymeric sialic acid in
Drodsophila melanogaster (Roth et al., 1992). However, the peak observed in chromatogram B is believed to be free NeuAc existing prior to cell lysis and is not due to degradation of polymeric sialic acid because similar results are obtained if cells are lysed, clarified, and immediately analyzed by HPAEC (data not shown). Lectin Analysis. The glycosylation of native cellular proteins was investigated by using lectin binding assays (Figure 6). Peanut agglutinin (PNA),a lectin specific for the GalP1,3GalNAc linkage, bound selectively to several large proteins in lysates from each of the four different cell lines. By contrast, Datura stramonium agglutinin (DSA), a lectin specific for Gal81,4GlcNAc, showed little or no binding, although a faint band was observed at approximately 30 kDa. Sambucus nigra agglutinin (SNA), with a specificity for NeuAca2,6Gal, provided similar results. This suggests that the 30-kDa protein contains terminal NeuAca2,6Gal~1,4GlcNAc,which is not thought to exist in insect cells. The molecular weight of this protein produced in T . ni cultures differs from that obtained with the other cell lines. Nonspecific binding of both DSA and SNA to this 30-kDa protein has not been ruled out. Figure 6 is from an experiment using protein from cultures grown in medium containing serum. However, analogous results were obtained for cells grown in protein-free medium, including binding of SNA to a 30-kDa protein. This indicates that the 30-kDa protein is indeed an insect cell derived protein and not a serum-derived protein. The GalB1,3GalNAc moiety detected by PNA likely belongs to an 0-linked oligosaccharide, whereas carbohydrates recognized by DSA and SNA are present on N-linked oligosaccharides. Studies with all four cell lines infected with wild-type AcMNPV provided similar results with these three lectins (data not shown).
Discussion This study has demonstrated that four different insect cell lines, derived from Spodoptera frugiperda, Trichoplusia ni, Bombyx mori, and Malacosoma disstria, contain significant levels of endogenous N-acetyl-b-hexosaminidase and sialidase activities and low levels of 8-galactosidase activity. By comparison with fresh TNM-FH meand dium, elevated levels of N-acetyl-P-hexosaminidase sialidase activity were observed in the supernatant from insect cell cultures. In accord with previous literature, serum-containing medium that was not exposed to insect cells contained all exoglycosidases studied except for sialidase (Goi et al., 1989;Ramage and Cunningham, 1975; Schauer et al., 1985). However, the cell-associated nature of the exoglycosidaseactivity was documented by growing the cells in protein-free medium, which lacked detectable exoglycosidase activity. Insect cell exoglycosidases were detected both in cell extracts and in the extracellular medium. The extracellular activity probably results from specific or nonspecific secretion of exoglycosidases. Baculovirus infection induced a dramatic increase in the levels of extracellular exoglycosidaseactivity. Presumably, this reflects a loss of membrane integrity and cell lysis due to the viral infection. Alternatively, it may result from a virus-induced effect on the expression and secretion of these enzymes. Although the pH optima of these enzymes were not determined, studies in this laboratory as well as published pH optima for similar exoglycosidases suggest that these enzymes are active in the slightly acidic insect cell culture medium. It is important to recognize the limitations of artificial substrates in the analysis of exoglycosidases. Although it provides the researcher with a product that is relatively
151
Bbtectmol. Rug., 1993, Vol. 9, No. 2
0.2
A
NeuAc
0.0
0.4
0.2
C
0.0
-0.2
1 1 1 1
1 1 1 1
0
2
l l l r
1 1 1 1
4
6
1 1 1 1
8
1 1 1 1
1 1 1 1
10
Minutes
12
14
1 1 1 1
1 1 1 1
16
18
20
Figure 5. Removal of NeuAc from fetuin incubated with cell lysate. Pulsed amperometric response (pC) is plotted as a function of time (min). N-Acetylneuraminic acid standard (1nmol) elutes a t approximately 15min in chromatogram A. Chromatogram B is from total uninfected cell lysate and indicates a small peak eluting a t the same time as the NeuAc standard. Chromatogram C is from fetuin incubated with an equal protein concentration of cell lysate as in chromatogram B. Exact sample preparation is described in the Materials and Methods section. The area of the peak corresponding to NeuAc in chromatogram C is approximately 4 times that of the peak area for chromatogram B. Elution conditions were constant 100 mM NaOH and a gradient of sodium acetate from 50 to 180 mM over 20 min with a flow rate of 1 mL/min.
m-
PNA
DSA
--SNA -
43-
Figure 6. Lectin analysis of uninfected cellular protein. Total protein was fractionated on an 8.7% SDS-polyacrylamide gel, transferred to nitrocellulose, and probed with the indicated lectins. The lectins are noted a t the top of each panel. The letter designations a t the bottom of each panel refer to S. frugiperda (S), 7’.ni (T),B. mori (B), and M. disstria (M). A molecular weight ladder is shown on the left.
easy to assay, the hydrolysis of artificial substrates is not identical to that of natural substrates. For most artificial substrates, the rate of hydrolysis is greater than the rate of hydrolysiswith natural oligosaccharidesor glycoproteins (Montreuil et al., 1987). However, even with natural substrates, rates are dependent on the nature and structure of the substrate. The usefulness of 2’-(4-methylumbellifery1)-a-D-N-acetylneuraminicacid for predicting behavior with natural substrates with S. frugiperda cells has been demonstrated here in experiments employing fetuin as the substrate. The hydrolysis of resorufin @-galactosideand lectin studies demonstrating the existence of the Gal@1,3GalNAcmoiety are not contradictory. Lectin studies with PNA indicating the presence of the Gal@1,3GalNAclinkage should not be construed as evi-
dence for the lack of @-galactosidase,since this exoglycosidase isolated from a variety of sources has demonstrated a dramatic linkagespecificity (Arakawa et al., 1974; Li et al., 1975). With regard to the p-nitrophenyl substrates, the results for the two different substrates presented in Figure 1provide an interesting control. The only structural difference between these substrates is the position of the hydroxyl group a t C-5. The difference in the hydrolysis rate of the two compounds suggests enzymatic recognition of the N-acetyl-@-hexosamineand not the p-nitrophenyl moiety. Similar behavior has been observed for other N-acetyl-@-hexosaminidases. Ram testes N-acetyl-@-hexosaminidasehydrolyzed p-nitrophenyl N-acetyl-@-Dglucosaminide approximately 7 times faster than p-nitrophenyl N-acetyl-@-D-gdactosaminide (Woolen et al., 1961a). The presence of endogenous exoglycosidases in baculovirus-infected insect cells may be important for several reasons. Since the infection is lytic, glycoproteins produced in this system are likely to come in contact with these exoglycosidases, regardless of where the relative populations are sequestered in healthy cells. During the production and isolation of recombinant glycoproteins, endogenous exoglycosidasescould degrade oligosaccharide side chains, possibly affecting protein properties and certainly complicating the subsequent analysis of these glycoforms. In addition, baculovirus expression vectors have been employed to produce heterologous glycosyltransferases, e.g., a2,6-sialyltransferase. If these enzymes are to be used for the in vitro synthesis of oligosaccharides, it is necessary to recognize that endogenous glycosidases exist in this system and to ensure purified recombinant glycosyltransferases. In conclusion, the presence of exoglycosidase activities must be taken into consideration when using baculovirus expression systems, and measures should be taken to minimize the potential effects of endogenous exoglycosidases in studies on insect cellderived oligosaccharide structures.
Acknowledgment We are grateful to Dr. Charles Goochee and Michael Gramer for their advice and critical review of this
B i o t ~ m l Rug., . 1993, Vol. 9, No. 2
152
manuscript. This research was supported by the National Science Foundation (Grant No. BCS-87219731,by a grant for Predoctoral Training in Biotechnology from the National Institute of General Medical Sciences, by National Research Service Award 1T32 GM 08346-01 from the Pharmacology Sciences Program, and by a Kelco Graduate Fellowship.
Literature Cited Arakawa, M.; Ogato, S.; Muramatau, T.; Kobato, A. @-galactosidases from jack bean meal and almond emulsion. J . Biochem. 1974, 75, 707-714. Borooah, J.; Leaback, D. H.; Walker, P. G. Studies on glucosaminidase. Biochem. J. 1961, 78, 106-1 10. Bradford, M. A rapid and sensitive method for the quantitation of microgram quantities of protein utilizing the principle of protein-dye binding. Anal. Biochem. 1976, 72, 248-254. Butters, T. D.; Hughes, R. C.; Vischer, P. Steps in the biosynthesis of mosquito cell glycoproteins and the effects of tunicamycin. Biochim. Biophys. Acta 1981,640, 672-686. Corfield,A. P.;Veh,R. W.; Wember,M.;Michalski, J. C.;Schauer, R. The release of N-acetyl- and N-glycolloyl-neuraminic acid from soluble complex carbohydrates and erythrocytes by bacterial, viral and mammalian sialidases. Biochem. J. 1981, 197,293-299. Davidson, D. J.; Castellino, F. J. Asparagine-linked oligosaccharide processing in Lepidopteran insect cells. Temporal dependence of the nature of the oligosaccharides assembled on asparagine-289 of recombinant human plasminogen produced in baculovirus vector infected Spodoptera frugiperda cells. Biochemistry 1991, 30, 6167-6174. Davidson, D. J.; Fraser, M. J.; Castellino, F. J. Oligosaccharide processing in the expression of human plasminogen cDNA by Lepidopteran insect Spodoptera frugiperda cells. Biochemistry 1990, 29, 5584-5590. Davidson, D. J.; Bretthauer, R. K.; Castellino, F. J. a-mannosidase catalyzed trimming of high mannose glycans in noninfected and baculovirus-infected Spodoptera frugiperda cells IPLBSF-21AE. A possible contributing regulatory mechanism for assembly of complex-type oligosaccharides in infected cells. Biochemistry 1991, 30, 9811-9815. DiCioccio, R. A.; Klock, P. J.; Barlow, J. J.; Matta, K. L. Rapid procedures for determination of endo-N-acetyl-a-D-galactosaminidase in Clostridiumperfringens,and of the substrate specificity of exo-@-D-galactosidases. Carbohydr. Res. 1980, 81, 315-322. Goi, G.; Fabi, A,; Lombardo, A.; Bairati, C.; Bovati, L.; Burlina, A. B.; Agosti, S.; Serio, C.; Tettamanti, G. The lysosomal @-DN-acetylglucosaminidase isozymes in human plasma during pregnancy: separation and quantification by a simple automated procedure. Clin. Chim. Acta 1989, 179, 327-340. Goochee, C. F.; Gramer, M. J. GlycobiologyKeystone Symposium. J . Cell Biochem. 1992, Suppl. No. 16D, 154. Hink, W. F. Established insect cell line from the cabbage looper Trichoplusia ni. Nature 1970, 226, 466-467. Hofmann, J.; Sernetz, M. Immobilized enzyme kinetics analyzed by flow-through microfluorimetry. Anal. Chim. Acta 1984, 163, 67-72. Hsieh, P.; Robbins, P. W. Regulation of asparagine-linked oligosaccharide processing: oligosaccharide processing in Aedes albopictus mosquito cells. J . B i d . Chem. 1984, 259, 2375-2382.
Hughes, R. C.; Jeanloz, R. W. The extracellular glycosidases of Diplococcus pneumoniae. 1. Purification and properties of a @-N-acetylglucosaminidase.Action on the al-acid glycoprotein of human plasma. Biochemistry 1964a, 3,1535-1543. Hughes, R. C.; Jeanloz, R. W. The extracellular glycosidases of Diplococcus pneumoniae. 2. Purification and properties of a @-N-acetylglucosaminidase.Action on a derivative of the al-acid glycoprotein of human plasma. Biochemistry 1964b, 3, 1543-1548. Jarvis, D. L.; Summers, M. D. Glycosylation and secretion of human tissue plasminogen activator in recombinant baculovirus-infected insect cells. Mol. Cell. Biol. 1989,9,214-223. Kobata, A. Use of endo- and exoglycosidases for structural studies of glycoconjugates. Anal. Biochem. 1979,100, 1-14. Kuroda, K.; Geyer, H.; Geyer, R.; Doerfler, W.; Klenk, H. D. The oligosaccharides of influenza virus hemagglutin expressed in insect cells by a baculovirus vector. Virology 1990,174,418429. Li, S.; Li, Y. Studies on the glycosidases of jack bean meal. J. Biol. Chem. 1970,245, 5153-5160. Li, S.; Mazzotta, M. Y.; Chien, S.: Li, Y. Isolation and characterization of jack bean @-galactosidase. J. Biol. Chem. 1975, 250,6786-6791. Licari, P.; Bailey, J. E. Factors influencing recombinant protein yields in an insect cell-baculovirus expression system: multiplicity of infection and intracellular protein degradation. Biotechnol. Bioeng. 1991, 37, 238-246. Luckow, V. A. In Recombinant D N A Technology and Applications; Prokop, A., Bajpai, R. K., Ho, C. S., Eds.; McGrawHill, Inc.: New York, 1991; pp 97-152. Luckow, V. A.; Summers, M. D. Trends in the development of baculovirus expression vectors. Biotechnology 1988, 6, 4755. Meyers, R. W.; Lee, R. T.; Lee, Y. C.; Reynolds, L. W.; Uchida, Y. The synthesis of 4-methylumbelliferyl a-ketoside of Nacetylneuraminic acid and ita use in a fluorometric assay for neuraminidase. Anal. Biochem. 1980, 101, 166-174. Miller, L. K. Baculoviruses as gene expression vectors. Annu. Rev. Microbiol. 1988, 42, 177-199. Montreuil, J.; Bouquelet, S.; Debray, H.; Fournet, B.; Spik, G.; Strecker, G. In Carbohydrate Analysis; Chaplin, M. F., Kennedy, J. F., Eds.; IRL Press: Oxford, UK, 1987; pp 143204. Ramage, P.; Cunningham, W. L. The occurrence of low a - ~ fucosidase activities in normal human serum. Biochim. Biophys. Acta 1975,403, 473-476. Roth, J.; Kempf, A.; Reuter, G.; Schauer, R.; Gehring, W. J. Occurrence of sialic acids in Drosophila melanogaster. Science 1992,256, 673-675. Schauer, R.; Sander-Wewer, M.; Gutachker-Gdaniec, G.; Roggentin, P.; Randow, E. A.; Hobrecht, R. Sialidase activity in the sera of patients and rabbits with clostridial myonecrosis. Clin. Chim. Acta 1985, 146, 119-127. Thomsen, D. R.; Post, L. E.; Elhammer, A. P. Structure of 0-glycosidically-linked oligosaccharides synthesized by the insect cell line Sf9. J. Cell. Biochem. 1990, 43, 67-79. Uchida, Y.; Tsukada, Y.; Sugimori, T. Enzymatic properties of neuraminidases from Arthrobacter ureafaciens. J.Biochem. 1979,86, 1573-1585. Woolen, J. W.; Heyworth, R.; Walker, P. G. Studies on glucosaminidase. Biochem. J . 1961a, 78, 111-116. Woolen, J. W.; Walker, P. G.; Heyworth, R. Studies on glucosaminidase. Biochem. J. 1961b, 79, 294-298. Accepted January 7, 1993.
