Biotechnol. Prog. 1991, 7, 462-467
402
Production of Recombinant Proteins by Baculovirus-Infected Gypsy Moth Cells Michael J. Betenbaugh,' Laura Balog, and Po-Shun Lee Department of Chemical Engineering, Johns Hopkins University, Baltimore, Maryland 21218 An experimental study was undertaken to evaluate alternative insect cell lines to Sf9 [from Spodoptera frugiperda (fall armyworm)] for the production of recombinant proteins. Insect cell lines from two different organisms were considered: IPLB-LdEIta (LdEIta) from Lymantria dispar (gypsy moth) and IPLB-HvT1 (HvT1) from Heliothis uirescens (tobacco budworm). Both LdEIta and H v T l produced higher total activity levels of recombinant P-galactosidase in monolayer culture than Sf9 after infection with the Autographa californica nuclear polyhedrosis virus (AcMNPV). However, only LdEIta generated a product yield (activity per milligram of total protein) which exceeded that of Sf9 (by 25%), so its growth and production characteristics were investigated in depth. LdEIta generated production levels and yields of a recombinant rotaviral protein, VP4, which exceeded those of Sf9 by 84 and 3872,respectively. In suspension culture, the LdEIta cells grew as aggregates with a doubling time several hours longer than Sf9, but the recombinant product yields of LdEIta were still higher than Sf9 by 38% in this culture environment. ,&Galactosidase expression rates and cell death rates suggested that the difference in productivity between the two hosts was due to the ability of LdEIta to survive the baculovirus infection and produce recombinant proteins longer than Sf9. The presence of LdEIta aggregates in suspension culture may be used as a method to separate live cells from dead cells, labile product, and spent medium in recombinant protein production processes.
Introduction Recombinant DNA technology has revolutionized the biotechnology industry. Due to the molecular biology advances made over the past 15 years, it is now possible to genetically engineer a number of different organisms to manufacture valuable proteins, vaccines, and other therapeutics. Insect cells offer a number of advantages for expression of genetically engineered proteins. They are capable of performing many posttranslational modifications including glycosylation, phosphorylation, and secretion during the expression of recombinant proteins (1). Recombinant viral antigens synthesized by insect cells have been shown to induce neutralizing antibodies in both in vitro and in vivo studies, making these proteins valuable as geneticallyengineered vaccines (1-3). Also, the common insect cell expression vector, baculovirus,will not replicate in mammals and thus does not represent a significant health hazard (4, 5 ) . Given the potential of insect cells, it will be important to develop a host-vector system which is efficient at recombinant gene expression and can be adapted to largescale protein production. Several investigators have developed vectors which have improved expression propand others have investigated the importance erties (6,7), of bioreactor operation conditions (8-1 1 ) . Another factor which will contribute to the efficiencyof recombinant gene expression is the host choice. The vast majority of all previous insect cell cloning and expression work has utilized a single host, the Sf9 cell h e , from Spodoptera frugiperda (fall armyworm) (12). However, 17 insects and more than 50 cell lines of Lepidoptera can be infected by the baculovirus cloning vector, Autographa californica nuclear polyhedrosis virus (AcM-
* Corresponding author. 8758-7938/9 1/3007-0462$02.50/0
NPV) (13). Therefore, each is a potential host for the production of genetically engineered proteins and biopesticides using AcMNPV, and different insect cell lines may have properties which may be favorable for a particular application. Several recent studies have elucidated the potential of alternative insect cell lines for the expression of different recombinant proteins. Shuler et al. (9) noted the potential for the Trichoplusia ni (TN368, cabbage looper) cell line, and Hinket al. (15)observed expression of recombinant proteins using 23 different cell lines. In this study, several alternative cell lines which have the greatest potential for replacing Sf-9 for certain applications have been evaluated. Two alternative insect cell hosta, IPLB-LdEIta (LdEIta) from Lymantria dispar (gypsy moth) (16)and IPLB-HvT1 (HvT1) from Heliothis uirescens (tobaccobudworm) (16),were considered and compared to Sf-9 on the basis of recombinant &galactosidase productivity. These two alternative hosts were chosen on the basis of two criteria: (1) the ability to replicate the AcMNPV vector and generate high yields of wild-type AcMNPV virus (Dwight E. Lynn, personal communication, 1990) and (2) the capacity to grow in the serum-free medium. The reasons for applying this second criterion were severalfold. First of all, the use of serumfree medium provides several advantages including lower culture costs, easier purification of the product, and elimination of medium variability associated with the serum. These advantages will be significant when insect cells are used in large-scale production studies. Second, the productivities of the different cell lines were compared on the basis of total protein content. This is much easier inserum-free medium due to the absence of serum proteins that interfere with the assay measurement. The insect cell lines considered and their sources are included in Table I.
0 199 1 American Chemlcal Society and Amerlcan Instkute of Chemlcal Engineers
463
Bktechnd. Rog., 1991, Vol. 7, No. 5
Table I. Insect Cell Lines and Species and Tissue Sources
species Lymantria dispar (gypsy moth) Heliothis virescem (tobacco budworm) Spodoptera frugiperda (fall armyworm)
Table 11. Average Recombinant &Galactosidase Activity and Yield for Different Insect Cell Hosta Infected in Four Duplicate Monolayer Cultures with AcMNFV-@-gal
tissue embryos
designation IPLB-LdEIta
ref 16
larval testes
IPLB-HvT1
16
cell line HvTl LdEIta
pupal ovaries
Sf9
I2,24
SEI
The gypsy moth cell line LdEIta proved especially promising, so its growth and production characteristics were investigated in depth. The production of another recombinant protein, rotaviral VP4, was compared between LdEIta and Sf9, and the behavior of LdEIta in suspension culture was evaluated to determine its potential for large-scale protein production.
Materials and Methods Insect Cell Hosts. The cells lines IPLB-LdEIta and IPLB-HVT1 were provided by Dwight Lynn of USDA (16) in monolayer cultures. Sf9 cells were obtained from the American Type Culture Collection. Recombinant AcMNPV Vectors. The AcMNPV vector containing the @-galactosidasegene (AcMNPV-@gal) was provided by Max Summers of Texas A&M (12). The recombinant AcMNPV vector containing the gene for the rotavirus VP4 protein (AcMNPV-VP4) was obtained from Mario Gorziglia of NIH (3). The VP4 protein is a nonglycosylated 85 000 MW protein which is one of the two outer capsid proteins of the porcine OSU rotavirus. Viruses were produced in Sf9-infected cultures and maintained in serum-free Excell 400 (J. H. Biosciences) medium at 4 O C with a virus titer of between lo7 and lo8 PFU/mL as measured by the end-point dilution method ( 1 7, 12). Monolayer Experiments. Insects cells were maintained in serum-free Excell 400 in 25-cm2monolayer flasks (Falcon 3013 from VWR) in a humidified 27 "C incubator (Fisher Scientific). The cells were infected in these flasks by adding 1mL of medium containing recombinant AcMNPV to 4 mL of the uninfected medium. Cells were infected 1day after passing to ensure that the cells were infected at a low cell density and in fresh medium, Cells infected with wild-type baculovirus at high cell densities and in spent medium have shown reduced infectivities and lowered production rates of polyhedra (18,19). The infections were allowed to proceed for 7-10 days, over which time complete infection and cell lysis was achieved as observed microscopically (Nikon TMS inverted microscope) by the absence of cell-surface attachment and loss of cell-membrane continuity. The lysed cells and broth were collected and centrifuged at 120 rpm for 10min (IEC HN-SII), and then the supernatant and pellet were collected as separate fractions. The lysed pellet was resuspended in 2 mL of PBS (phosphate-buffered saline, Gibco Inc.) and sonicated for three 15-s cycles (Tekmar). Recombinant protein activity and total protein levels were measured in the supernatant and pellet. A control flask was also maintained in which 1 mL of virus-containing medium was added to 4 mL of medium without cells in order to determine recombinant protein activity level added to the flask initially. Recombinant Protein Assays. Recombinant /3-galactosidase activity was measured in the supernatant and pellet using the assay method of Pardee et al. (20). This activity level can be converted to Miller activity units by multiplication by 7.5 (21). Recombinant levels of antigenic VP4 protein were measured with an ELISA assay using monoclonal antibody against the OSU rotavirus and
&gal activity/mL (1.7 f 0.2) X l(r (1.8 f 0.4) X l@ (1.2 f 0.1) x 104
@-galactivity/mg of protein (3.5 0.3) x 103 (9.4 f 0.8) X 103 (7.5 f 0.2) x 103
correlated against astandard dilution curve generated from OSU rotavirus (3, 22). Total Protein Assay. Total protein levels were measured in the supernatant and the pellet using the Bradford method (Bio-Rad Co.) and correlated against a standard of bovine serum albumin. Spinner Flask Experiments. Cellswere cultured with 100mL of medium in a 125-mLsuspension culture spinner flask (Corning) in a humidified 27 "C incubator. The agitation rate was maintained at 70 rpm. Growth and infection experimentswere performed by seeding the flasks from an inoculation suspension culture. These inoculation flasks were seeded from 4-6 monolayer flasks of the particular insect cell line. Infection of these cultures was accomplished by adding 5-10 mL of a recombinant AcMNPV in order to achieve an equivalent multiplicity of infection (MOI) of between 1 and 10 for each flask containing the different hosts. Cells were typically infected in the early to middle exponential growth phase and at a low cell density. Infection at low cell density will eliminate product yield reductions that occur if the insect cells are infected in suspension cultures at high cell densities (11, 23). All infections were carried out with a Sf9 viable cell density above 90% and LdEIta viability above 80%. Samples were withdrawn at regular 24-h intervals using a sterile l-mL pipet and delivered into sterile 1.5-mL micro test tubes. Separate samples were taken for determining viable cell counts and for analysis of recombinant 0-galactosidase activity and total protein content in the pellet and supernatant. Cell Counts. Cell concentrations in the spinner flask were determined by microscopic counting with a hemacytometer. Viable cells were identified by trypan blue exclusion test (12). Mild vortexing of the LdEIta samples was required to reduce aggregate size prior to cell counting.
Results Monolayer Experiments. Production of Recombinant galactosidase. In the first set of monolayer infection experiments, several flasks with each type of insect cell line were infected with the recombinant AcMNPV-&gal virus. After complete infection and cell lysis, the dead cells and broth were collected and the recombinant 8-galactosidaseactivity was measured. The average &galactosidase activity levels obtained from four duplicate cultures are shown in Table 11. Both HvTl and LdEIta generated activity levels of recombinant 8-galactosidase which exceeded those of Sf9 by more than 40 7%. Such a difference in total activity would be significant for large-scale production of recombinant proteins. However, since each monolayer cell culture possessed slightly different cell densities, a more accurate comparison of productivity is established by comparing recombinant product yields. The yield is calculated by dividing total recombinant protein activity by cell number or total cell protein. In this study, recombinant product yields were based on total protein content rather than cell number because the three cell lines differ in size. As a result, the recombinant protein productivity or yield was calculated by dividing recombinant 8-galactosidase activity by the
Bbtechnol. Rog., 1991, Vol. 7, No. 5
464
Table 111. Average Recombinant VP4 Production and Yields As Measured by ELISA for LdEIta and Sf9 after Infection of Duplicate Monolayer Cultures with AcMNPV-VP4 cell line MEIta Sf9
VP4 activity/mL 0.94 f 0.07 0.51 f 0.08
VP4 activity/mg of protein
18.2 f 0.9 13.2 f 0.1
. z
2.0 lo6
0
0,
I
0
1.5
lo6 -
1.0
lo6 -
> ul c
I
I
I
1
LdElta Cell Counts
Doubling time = 30. hours
0
I
0
0
0
amount of total cell protein (milligrams) in the culture. The product yields for this same set of monolayer experiments are also shown in Table 11. Though HvTl generated higher recombinant /%galactosidase activity levels than Sf9, part of this difference was caused by the ability of HvTl to grow to much higher densities in monolayer cultures. As a result, recombinant productivities of the HvTl cells were lower than those achieved in Sf9. LdEIta also achieved higher protein densities in culture than Sf9, yet LdEIta was still 25% more productive than Sf9. Production ofRecombinant W 4Protein. Since the gypsy moth cell line LdEIta was more productive than Sf9 for expression of @galactosidase, the production of another recombinant protein, VP4 from porcine rotavirus, was compared between these two hosts. In this way, it could be determined whether the difference in productivity was due to the host cell or to the particular recombinant protein considered. The results of this comparisonare illustrated in Table 111. Total production levels of VP4 by LdEIta were almost double the levels of Sf9,and the recombinant VP4 product yields achieved by LdEIta were also considerably higher (38%)than those in Sf9 monolayer cultures. One difference between the two cell lines was in the relative level of VP4 accumulated in the supernatant and pellet a t the completion of this monolayer experiment. While only 5 % of the VP4 protein was present in the pellet of the Sf9cells, fully 12 76 of the VP4 was present in the pellet of the LdEIta monolayers. This difference may be due to the relative retention of the protein in the cells or linked to the time for cell lysis in the two cell lines. Suspension Culture Experiments: GrowthStudies. We investigated the growth of LdEIta in a suspension culture system in order to evaluate its potential for largescale cell culture. The growth of LdEIta cells was examined in 100-mLvolume spinner flasks with Excell 400 medium. The viable cell counts for a growth experiment with LdEIta are shown in Figure 1along with the exponential growth curve. Viable cell densities reached a maximum of about 1.5 X lo6cells/mL for LdEIta, which was half the maximum cell density achieved using Sf9 cells. However, the LdEIta cells are several times larger than the Sf9 cells and actual final protein levels were higher for the LdEIta cultures (results not shown). The doubling time for the LdEIta cells, between 28 and 35 h, was about 5 h slower than the doubling lines observed for Sf9 cells in Excell 400. Another difference between the LdEIta and Sf9 cells grown in suspension culture was the formation of aggregates by the LdEIta cells. The LdEIta cells typically formed aggregates of between 50 and 200 cells in suspension culture and these aggregates increased in size as the growth experiment proceeded. The difference in aggregate formation between LdEItaand Sf9cells in suspensionculture is illustrated in Figures 2 and 3. The LdEIta cells exist almost exclusively in aggregates while the Sf9 cells are found mostly as individual cells but occasionally in small aggregates. Production ofRecombinant @-Galactosidase.Given the aggregation of the LdEIta in suspension culture, it
0
B
5
5.0 10’
0
-
t
I
Figure 2. Micrograph (75X) of LdEIta aggregate from 100-mL spinner-flask culture.
Figure 3. Micrograph (75X) of Sf9 cells from 100-mL spinnerflask culture.
was important to evaluate its production capabilities in this environment as compared to Sf9. Experiments were performed in which the two cell lines were infected in suspension cultures with the AcMNPV-@-galvirus. Each of the two cultures was infected at low cell density by adding virus into the medium a t an equivalent MOI. The LdEIta cells were infected at a slightly lower cell density (1.5 X lo5cells/mL) than Sf9 (4.2 x lo5cells/mL) because of the larger size and lower maximum cell density achieved by LdEIta cells. Following infection, samples were collected daily and the recombinant &galactosidaseactivity, total protein content, and cell viability monitored. The recombinant @-galactosidase production results for a typical experiment are shown in Figure 4. The final product yields (top panel) reflect the same behavior observed in the monolayer experiments, namely, that the
Bbtechnol. h g . , 1991, Vol. 7, No. 5
-g -2
1.2 lo5
.-2 > .c 2
6.0 10'
.-
E 0
1.0
los
m
. e
8.0 10'
-m Z
4.0
104
2.0 104 0
0 0
1
2
3
4
5
6
7
8
Time (days post infection)
Time (days post infection)
Figure 4. (Top) Product yield and (bottom) total production of @-galactosidaseby LdEIta and Sfs after infection of 100-mL spinner-flask cultures with AcMNPV-@-gal.
LdEIta cells were more productive (38%) than Sf9 cells. In fact, the final product yield values for both cell lines in suspension culture were similar to those achieved in monolayer culture. Total recombinant @-galactosidase production levels (bottom panel) were nearly equal for the first 2 days but production in the LdEIta culture increased over the next 6 days to a final value which was more than double that in the Sf9 culture. The total production levels of the Sf9 culture increased only 75% after 3 days while the production levels in LdEIta culture nearly tripled from 3 to 8days postinfection. This suggests that the difference in productivity between the two cultures may be linked to the ability of the LdEIta cells to remain viable and productive for a longer time period than Sf9 cells. In order to test this hypothesis, the viable cells in both the Sf9 and LdEIta cultures were examined following infection with the recombinant baculovirus. The viable cell counts for each of the cultures are plotted in Figure 5. After 3 days, the rate of cell death was much higher for the Sf9 cells and this corresponded to the time at which production of @-galactosidaseslowed. The rate of cell death was much slower and more gradual for the LdEIta cells. Even though these cellswere infectedat a cell density that was 3 times lower, the viable LdEIta cell density actually exceeded the Sf9 viable cell density after 6 days, suggesting that the gypsy moth cells can indeed survive a baculovirus infection longer. To determine if the product yield differences in suspension culture were due to the particular infection conditions, several additional suspension infection experiments were performed. Since all previous experiments were performed with seed virus from Sf9cultures infected with AcMNPV, another experiment was performed with
1
2
3
4
5
6
7
8
Time (days post infection)
Figure 5. Viable cell density of LdEIta and Sf9 in 100-mL spinner-flask cultures after infection with AcMNPV-8-gal.
seed virus obtained from LdEIta cultures. Product yields for the LdEIta cultures (1.4 X lo5 activity units/mg of protein) were higher than Sf9 (8.3 X lo4 units/mg) regardless of the viral source and yields were in fact higher when sourced from the LdEIta cultures. Also, the infection experiments were repeated after resuspending the cells in fresh medium and similar results to those shown in Figure 4 were observed. Aggregate Behavior after Infection witb Recombinant AcMNPV. The LdEIta aggregates remained and often became slightly larger in size for several days following infection with the recombinant baculovirus.Also, the aggregatestended to settle out of the suspension rapidly if the agitation was stopped. By the end of the infection period approximately a week later, however, the clumps had essentially disintegrated and the suspension was uniform in composition. The aggregate breakup was thought to be linked to cell death and lysis caused by the AcMNPV infection, so a series of selectiveviable cell counts following infection of the LdEIta spinner flasks were performed to test this proposition. Once every 24 h, a sample was withdrawn from the spinner flask and allowed to settle for 5 min in a sterile microtube. Then the top half volumetric portion of this sample was removed to a separate microtube. Separate viable cell counts were performed on the top and bottom fractions. The viable and total cell counts in the top and bottom fractions for one spinner flask experiment are shown in Figure 6. Essentially all of the viable cells were found in the bottom fraction initially (top panel). The cells in the bottom fraction were most likely in aggregates since these settled out of suspension rapidly. As the infection progressed and the cells died, the number of viable cells in the bottom fraction declined until viable cell densities were essentially zero in both fractions. In Figure 6, bottom panel, the total (viable plus nonviable) LdEIta cells in the bottom and top fractions are plotted versus days postinfection. Initially, the great majority of cells were in the bottom fraction and most of these were viable. However, after 4 days the number of total cells in the top fraction became significant and reached a level approximately equal to the bottom fraction by the end of the infection period. Most of the cells in the top fraction were not aggregated since they did not settle out of suspension, and they also were not viable (from Figure 6, top panel). Combined, the results of Figure 6 indicate that the viable cells in the culture were found mostly in large aggregates while nonviable cells were present chiefly as single cells or in small aggregates. Subsequent microscopic examination of the cultures confirmed these findings.
Biotechnol. Rog., 1991, Vol. 7, No. 5
400 q 1 2
.zc ZI
I
I
I
6
0,
Time (days post infection)
E
,. a,
0
2
4
6
8
10
Time (days post infection) Figure 6. (Top) Viable and (bottom) total (viable nonviable) cell density of LdEIta in top and bottom fractions of settled sample from a 100-mL spinner-flask culture infected with AcMNPV-8-gal.
+
Discussion In this study, the gypsy moth cell line, LdEIta, was found to generate higher production levels and yields of recombinant @-galactosidaseand VP4 than Sf9 when infected with baculovirus AcMNPV. Also, it was shown that LdEIta can be grown and infected to achieve high product yields in both suspension and monolayer cultures. In an earlier study with monolayer cultures, Hink et al. (15) observed higher production levels of recombinant 0-galactosidase by Sf9cells, but the two studies are not directly comparable for several reasons. The previous researchers observed activity levels after 48 and 72 h rather than following complete cell lysis, used a cell number basis rather than total protein for yield comparisons, and compared the productivities in different media. Since neither j3-galactosidase nor VP4 is glycosylated, subsequent studies should compare differences in productivity between Sf9and LdEIta for these types of protein products. Hink et al. showed that insect cell lines glycosylate proteins differently, and this may be an important factor in choosing the appropriate cell line. If LdEIta shows as much promise in the production of glycosylated proteins as nonglycosylated ones, experiments should be undertaken to develop a single clonal isolate from LdEIta much as Sf9 was derived from IPLB-Sal-AE (12,24). A possible reason for the improved productivity of LdEIta over Sf9was suggested from the suspension culture data. The LdEIta cells maintained viability and retained recombinant protein synthesis for a longer time period than Sf9 cells. Further evidence of LdEIta cell survival was the accumulation of higher final protein levels in LdEIta cultures, the presence of larger relative levels of recombinant VP4 protein in the pellets of LdEIta cultures,
and the microscopic observation of large viable infected LdEIta cells (unpublished data). It was especially interesting to note that the LdEIta cells were equally productive in monolayer and suspension cultures even though the LdEIta cells grew as aggregates in suspension culture. Such a finding suggests that cell aggregation may not always be a critical factor limiting the recombinant productivity. Previously, Wood et al. (25) observed that cell-to-cell contact was an important factor limiting infection and baculovirus production in wild-type infections of monolayer cultures of 2'. ni (TN368). In fact, the cell aggregation characteristic of gypsy moth cells presents certain advantages that may be used in the development of an insect cell bioreactor system. The rapid settling of live LdEIta cell clumps could be used as a means for separating these live infectious cells from the spent medium, dead insect cells (present in single-cell form), and unstable recombinant proteins. Replacement of the spent medium and dead cells with live cells and fresh medium would enable the operation of a bioreactor in a continuous or semicontinuous mode. With Sf9 cells, the spent medium must be separated from cells by centrifugation or filtration, but neither of these processes can selectively eliminate dead cells. Unfortunately, the formation of aggregates also may present mass transfer limitations, and this factor must be considered in largescale bioreactors. Subsequent studies should examine chemical and mechanical methods for altering the size of LdEIta aggregates in suspension culture. This study has shown that an alternative insect;cell line to Sf9,LdEIta from the gypsy moth, can efficientlyproduce recombinant proteins in several different culture environments. Ultimately, a number of insect cell hosta may be applied to the production of recombinant proteins much like many different procaryotic and eucaryotic organisms are currently used in the recombinant DNA industry.
Acknowledgment We are grateful to Dwight Lynn, Ed Dougherty, and James Vaughn of the Agricultural Research Service of USDA for supply of insect cell lines and much advice on this project. We thank Mario Gorziglia of the Laboratory of Infectious Diseases of NIH for provision of the AcMNPV-VP4 baculovirus. This project was supported in part by National Science Foundation Grant BCS9007762. Literature Cited (1) Luckow, V. A,; Summers, M. D. Trends in the Development of Baculovirus Expression Vectors. Biol Technology 1988,6, 47-55. (2) Coelingh, L. L. V. W.; Murphey, B. R.; Collins, P. L.; LebacqVerheyden,A. M.; Battey, J. F. Expression of biologically active and antigenically authentic parainfluenza type 3 hemagglutinin-neuraminidase glycoprotein by a recombinant baculovirus. Virology 1987, 160, 465-472. (3) Nishikawa, K.; Fukuhara, N.; Liprandi, F.; Green, K.; Kapikian, A. 2.;Chanock, R. M.; Gorziglia, M. VP4 Protein of Porcine Rotavirus Strain OSU Expressed By a Baculovirus Recombinant Induces Neutralizing Antibodies. Virology 1989, 173,631-637. (4) Volkman, L. E.; Goldsmith, P. A. In vitro Survey of Autographa californica Nuclear Polyhedrosis Virus interaction with Nontarget Vertebrate Host Cells. Appl. Enuiron. Microbiol. 1983, 45, 1085-1093. ( 5 ) Tija, S. T.; Meyer zu Altenschildesche, G.; Doerfler, W. Autographa californica nuclear polyhedrosis virus (AcNPV) DNA does not persist in mass cultures of mammalian cells. Virology 1983, 125, 107-117.
Blotechnol. Rog., 1991, Vol. 7, No. 5
(6) Fraser, M. J. Expression of Eucaryotic Genes in Insect Cell Cultures. In Vitro Cell. Dev. Biol. 1989,25, 225-235. (7) Jarvis, D. L.; Fleming, J.-A. G. W.; Kovacs, G. R.; Summers, M. D.; Guarino, L. A. Use of Early Baculovirus Promoters for Continuous Expression and Efficient Processing of Foreign Gene Products in Stably Transformed Lepidopteran Cells. BiolTechnology 1990,8,950-955. (8) Murhammer, D. W.; Goochee, C. F. Scaleup of Insect Cell Cultures: Protective Effects of Pluronic F-68.Biol Technology 1988,6,1411-1418. (9) Shuler, M. L.;Cho, T.; Wickham, T.; Ogonah, 0.;Kool, M.; Hammer, D. A,; Granados, R. R.; Wood, H. A. Bioreactor Development for Production of Viral Pesticides or Heterologous Proteins in Insect Cell Cultures. Ann. N.Y. Acad. Sci. 1990,589,399-422. (10) 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. (11) Lindsay, D. A,; Betenbaugh, M. J. Identification and Quantification of Cell Culture Factors Affecting Recombinant Protein Productivity in Baculovirus-Infected Insect Cells. Biotechnol. Bioeng. 1991,in press. (12) Summers, M. D.;Smith, G. E. A Manual of Methods for Baculovirus Vectors and Insect Cell Culture Procedures; Texas Agricultural Experiment Station Bulletin No. 1555; Texas A&M University: College Station, TX, 1987. (13) Groner, A. Specificity and Safety of Baculoviruses. In The Biology of Baculoviruses. Vol. I. Biological Properties and Molecular Biology; Granados, R. R., Federici, B. A., Eds.; CRC Press, Inc.: Boca Raton, FL, 1986;pp 177-202. (14) Deleted in press. (15) Hink, W. F.; Thomsen, D. R.; Meyer, A. L.; Davidson, D. J.; Castellino, F. J. Expression of Three Recombinant Proteins Using Baculovirus Vectors in 23 Insect Cell Lines. Biotechnol. Prog. 1991,7,9-14. (16) Lynn, D. E.;Dougherty, E. M.; McClintock, J. T.; Loeb, M. Development of Cell Lines from Various Tissues of Lepidoptera. In Invertebrate and Fish Tissue Culture; Kuroda,
467
Y., Kurstak, E., Maramorosch, K., Eds.; Japan Scientific Societies Press and Springer-Verlag: Tokyo and Berlin, 1988; pp 239-242. (17) Hughes, P. R.; Wood, H. A. In The Biology of Baculouiruses. Vol.II. Practical Application for Insect Control; Granados, R. R., Federici,B. A., Eds.; CRC Press, Inc.: Boca Raton, FL, 1986;pp 1-30. (18) Vaughn, J. L. The Production of Nuclear Polyhedrosis Viruses in Large-Volume Cell Cultures. J.Inuertebr. Pathol. 1976,28,233-237. (19)Stockdale, H.; Gardiner, G. R. The Influence ofthe Condition of Cells and Medium on Production of Polyhedra of Autographa californica Nuclear Polyhedrosis Virus in Cell Culture: Methods for Infecting Cells. J.Inuertebr. Pathol. 1977,30, 330-336. (20) Pardee, A.B.; Jacob, F.; Monod, J. The genetic control and cytoplasmic expression of "inducibility" in the synthesis of @-galactosidaseby E . coli. J . Mol. Biol. 1959,1, 165-178. (21)Miller, J. H.1972. Assay
[email protected] Experiments in Molecular Genetics;Cold Spring Harbor Laboratory: Cold Spring Harbor, NY, 1972;pp 352-355. (22)Juarbe-Osorio, L. G.; Gorziglia, M.; Betenbaugh, M. J. Recovery and Characterization of a Rotavirus Outer Capsid Protein Expressed in a Recombinant Insect Cell System. Protein Expression Purif. 1991 (submitted for publication). (23) Caron, A. W.; Archambault, J.; Massie, B. High-Level Recombinant Protein Production in Bioreactors Using the Baculovirus-Insect Cell Expression System. Biotechnol. Bioeng. 1990,36,1133-1140. (24) Vaughn, J. L.; Goodwin, R. H.; Tompkins, G. J.; McCawley, P. Establishment of Two Cell Lines from the Insect Spodoptera frugiperda (Lepidoptera; Noctuidae). In Vitro 1977,13,213217. (25) Wood, H.A.; Johnston, L. B.; Burand, J. P. Inhibition of Autographa californica Nuclear Polyhedrosis Replication in High-Density Trichoplusia ni Cell Cultures. Virology 1982, 119,245-254. Accepted August 9,1991. Registry No. @-Galactosidase,9031-11-2.
