A Baculovirus (Bombyx mori Nuclear Polyhedrosis Virus) Repeat Element Functions as a Powerful Constitutive Enhancer in Transfected Insect Cells*

It has been previously reported that baculovirus homologous regions, the regions of baculovirus genomes that contain the origins of DNA replication, can augment the expression of a small number of baculovirus genes in vitro. We are now reporting that a region of the genome of Bombyx mori nuclear polyhedrosis virus (BmNPV) containing the homologous region 3 (HR3) acts as an enhancer for the promoter of a nonviral gene, the cytoplasmic actin gene of the silkmoth B. mori. Incorporation of the HR3 sequences of BmNPV into an actin promoter-based expression cassette results in an augmentation of transgene expression in transfected cells by two orders of magnitude relative to the control recombinant expression cassette. This increase is due to a corresponding increase in the rate of transcription from the actin promoter and not to replication of the expression cassette and occurs only when the HR3 element is linked to the expression cassette in cis. A comparable degree of enhancement in the activity of the silkworm actin promoter occurs also in heterologous lepidopteran cells. Concomitant supplementation of transfected cells with the BmIE1 trans-activator, which was previously shown to be capable of functioning in vitro as a transcriptional co-activator of the cytoplasmic actin gene promoter, results in more than a 1,000-fold increase in the level of expression of recombinant proteins placed under the control of the actin gene promoter. These findings provide the foundation for the development of a nonlytic insect cell expression system for continuous high-level expression of recombinant proteins. Such a system should provide levels of expression of recombinant proteins comparable to those obtained from baculovirus expression systems and should also have the additional advantage of continuous production in a cellular environment that, in contrast to that generated by a baculovirus infection, supports continuously proper posttranslational modifications of recombinant proteins and the capability of expression of proteins from genomic as well as cDNA sequences.

It has been previously reported that baculovirus homologous regions, the regions of baculovirus genomes that contain the origins of DNA replication, can augment the expression of a small number of baculovirus genes in vitro. We are now reporting that a region of the genome of Bombyx mori nuclear polyhedrosis virus (Bm-NPV) containing the homologous region 3 (HR3) acts as an enhancer for the promoter of a nonviral gene, the cytoplasmic actin gene of the silkmoth B. mori. Incorporation of the HR3 sequences of BmNPV into an actin promoter-based expression cassette results in an augmentation of transgene expression in transfected cells by two orders of magnitude relative to the control recombinant expression cassette. This increase is due to a corresponding increase in the rate of transcription from the actin promoter and not to replication of the expression cassette and occurs only when the HR3 element is linked to the expression cassette in cis. A comparable degree of enhancement in the activity of the silkworm actin promoter occurs also in heterologous lepidopteran cells. Concomitant supplementation of transfected cells with the BmIE1 trans-activator, which was previously shown to be capable of functioning in vitro as a transcriptional co-activator of the cytoplasmic actin gene promoter, results in more than a 1,000-fold increase in the level of expression of recombinant proteins placed under the control of the actin gene promoter. These findings provide the foundation for the development of a nonlytic insect cell expression system for continuous high-level expression of recombinant proteins. Such a system should provide levels of expression of recombinant proteins comparable to those obtained from baculovirus expression systems and should also have the additional advantage of continuous production in a cellular environment that, in contrast to that generated by a baculovirus infection, supports continuously proper posttranslational modifications of recombinant proteins and the capability of expression of proteins from genomic as well as cDNA sequences.
In a previous communication (17), we reported on the development of an in vitro expression system that employed the promoter of the silkmoth A3 (cytoplasmic) actin gene (18,19) as a driver of foreign gene expression in lepidopteran cells. Furthermore, we also reported previously that a baculovirus transcriptional activator, BmIE1, that is encoded by an immediate early gene of BmNPV, functions as co-activator of the cytoplasmic actin gene promoter in transfected cells and increases the level of transcription from this promoter by two orders of magnitude (20).
In an attempt to increase the general utility of the actin gene promoter as a component of expression vectors for production of foreign proteins in insect cells, we have looked for additional genetic elements that can stimulate further the expression of the actin promoter. Because baculoviruses have evolved both to cooperate and compete with cellular genes and to take advantage of cellular structures by enhancing general cell maintenance functions, and in view of the previous reports regarding the in vitro effects of baculovirus HRs for a few viral genes, we have asked whether the enhancing properties of HR elements can also be seen when these elements are linked to gene promoters of the host cell and introduced into lepidopteran insect cells.
We are now reporting that, when linked to gene constructs expressing reporter proteins under the control of the cytoplasmic actin gene promoter of B. mori, a DNA fragment containing the HR3 of BmNPV enhances significantly the in vitro expres-sion of the reporter proteins. The observed stimulation is due to increases in the rate of transcription of the actin gene promoter. In accordance with the definition for enhancer elements, the BmNPV genomic fragment is capable of mediating its effect only in cis and in a position and orientation-independent manner. Finally, we show that linkage of the HR3 enhancer to the actin gene promoter combined with the presence of the BmIE1 protein in the transfected cells results in an augmentation in the level of actin promoter strength by three orders of magnitude relative to that of the basic actin gene promoter.
These results set the stage for the development of a stable insect cell transformation system for overexpression of recombinant proteins. Such a system should provide levels of expression similar to those achieved by recombinant baculoviruses but should be devoid of the limitations inherent to baculovirus expression systems. Well known limitations of baculovirus expression systems include first, a reduction in proper posttranslational modification of recombinant proteins during the high production phase (late stages of infection) due to damaging of the insect cell's protein modification machinery; second, the general inability of baculovirus-infected cells to process during the late stages of infection, complex, intron-containing transcripts and produce recombinant proteins from genomic sequences; and third, difficulties associated with the purification of recombinant proteins as a result of cell lysis at the end of the infection process, and release of proteases that cause degradation of the overexpressed products. None of these problems should occur in an insect cell expression system in which the cells remain intact for the duration of the production phase.

Cloning of the HR3 Element and Construction of Expression
Plasmids-A 6.6-kb ClaI fragment of BmNPV genomic DNA (map units 49.94 -54.94) containing genes p39, cg30, and p15 and a region with multiple EcoRI sites was previously cloned (Fig. 1A) (21,22). A 1.2-kb SspI subfragment of the cloned fragment (map unit 51.8 -52.7) containing multiple EcoRI sites (HR3 element of BmNPV) was subcloned into the SmaI site of vector pBluescript-SK ϩ (pBS; Stratagene). Two plasmids, p153 and p133, containing the 1.2-kb insert in opposite orientations were obtained and used subsequently for generating various expression cassettes. Plasmid pBmA.cat, which contains the Escherichia coli chloramphenicol acetyl transferase (CAT) open reading frame (ORF) under the control of the cytoplasmic actin A3 gene promoter of B. mori (17), was digested with SstI, and a 3.0-kb SstI fragment containing the actin.cat gene fusion was isolated and inserted in two orientations into the SstI sites of plasmids p133 and p153 to generate the recombinant expression cassettes p13314, p13315, p15316, and p15317 ( Fig.  2A). Plasmids pIE1/133 and pIE1/153, which contain the HR3 enhancer and the ie1 gene of BmNPV (20), were made by inserting a 3.8-kb ClaI fragment containing the ie1 gene into the ClaI site of plasmids p133 and p153, respectively, removing unwanted restriction sites in the polylinker of these plasmids by double digestion with SacII and BamHI, blunt-ending with T4 DNA polymerase and self-ligating the resultant plasmids. The double-enhanced expression cassettes pIE1/133A and pIE1/153A were generated by inserting a 2.2-kb SacI fragment containing the basic actin expression cassette from plasmid pBmA (17) into the unique SacI sites of plasmids pIE1/133 and pIE1/153, respectively. The recombinant expression vectors pIE1/133A.cat and pIE1/153A.cat were generated by inserting a 0.9-kb BamHI fragment, obtained from plasmid pBmA.cat and containing the ORF for CAT, into the unique BamHI sites of the expression cassettes pIE1/133A and pIE1/133A, respectively. The construction of plasmid pBmA.jhe(kk) containing the ORF of a modified version of the juvenile hormone esterase (JHE) cDNA of Heliothis virescens (23) under the control of the actin gene promoter was previously reported (20). Finally, the double-enhanced expression vector pIE1/153A.jhe(kk) was generated by first isolating a 1.8-kb EcoRI fragment containing the ORF for JHE (kk) from plasmid pAcUW21-KK (23), ligating NotI linker to its ends, digesting with NotI, and inserting it into the unique NotI site of the double-enhanced expression cassette pIE1/153A.
Cell Cultures and Transfections-Bm5 and Sf21 cells were maintained in IPL-41 medium with 10% fetal calf serum as described previously (24). They were transfected with various plasmid DNAs in 6-well microtiter plates (10 6 cells/well in 2 ml of medium) as described previously (20), except that cells in each well were transfected with 0.5 ml of transfection solution (20) containing 1 mg of pBmA.cat or equimolar amounts of each of the other expression plasmids described above and pBS DNA to a combined final DNA concentration of 5 mg/ml.
CAT Assays and Dot Blot Hybridizations-Transfected cells were pelleted 48 h posttransfection and washed three times with 1 ml of phosphate-buffered saline (10 mM KH 2 PO 4 , 2 mM NaH 2 PO 4 , 140 mM NaCl, 40 mM KCl). 90% of the cells were used to extract soluble protein, and 10-mg aliquots of the extracted protein were used for CAT assays as described (20). The remaining 10% of the transfected cells were used for Dot Blot DNA hybridization analysis. The cells were counted with a hemocytometer and Dot-blotted directly onto Hybond-N ϩ membranes (Amersham). The membranes were treated sequentially with 0.5 N NaOH and 0.5 M Tris-HCl, pH 7.5. Labeling of a 900-bp BamHI fragment of pBmA.cat containing the CAT ORF (17) with [␣-32 P]dCTP and hybridization of the immobilized DNA samples to the radioactive probe were as described (20).
JHE Assays-For JHE assays, cells were harvested as above, and their media were assayed as described previously (20,25). For quantitative measurements, samples were diluted to obtain juvenile hormone to juvenile hormone acid conversion rates between 10 and 50%, and activities were calculated on the basis of the dilution factors.
Nuclear Run-on Transcription and Hybridization Assays-Nuclei were isolated from pBmA.cat and p13315-transfected cells, and nuclear run-on reactions were carried out as described (20). Nuclear RNA labeled with [␣-32 P]CTP and [␣-32 P]GTP was used as probe against a dilution series of a 900-bp BamHI cat gene fragment isolated from plasmid pBmA.cat (17) and immobilized on Hybond-N ϩ membrane as described previously (20).

RESULTS
Cloning of the HR3 Region-The nucleotide sequence of the portion of the 1.2-kb SspI fragment (GenBank TM accession no. U77353) containing the HR3 element of BmNPV is shown in Fig. 1B. The HR3 element consists of a 72-bp sequence motif that is repeated completely four times and incompletely once. The fourth repeat unit is missing 45 bp of the repeat, whereas the first one contains two insertions, 14 and 56 bp (Fig. 1B, X and Y, respectively). Each repeat unit contains an EcoRI site located at the center of a 30-bp imperfect palindrome. The nucleotide sequence of HR3 presented here is somewhat different from that reported previously (GenBank TM accession no. L33180), especially in the rightward half of the repeat unit (data not shown). This difference is probably due to the fact that different BmNPV isolates were used.
HR3-mediated Enhancement of Gene Expression in Transfected Bombyx and Spodoptera Tissue Culture Cells-Derivatives of the expression vector pBmA.cat, which contains the CAT ORF under the control of the silkmoth cytoplasmic actin gene promoter, were constructed by inserting the 1.2-kb SspI fragment in four different combinations of position and orientation relative to the actin gene promoter (Fig. 2A). The ability of the HR3 fragment to enhance the actin promoter was assessed by transfecting Bm5 cells with pBmA.cat or each of the four HR3-containing plasmids. As shown in Fig. 2B, although some variability was observed between CAT activities obtained from constructs containing the HR3 element in various spatial arrangements relative to the actin promoter, the same fundamental result was obtained from all four HR3 configurations tested; cells transfected with actin expression cassettes containing the HR3 element resulted in expression of higher CAT activities than cells transfected with the basal expression vector. Quantitative analysis of the results shown in Fig. 2B  To ensure that the observed differences in CAT activity in transfected cells were not due to differences in transfection efficiencies between constructs or to plasmid DNA replication, we carried out hybridization-based determinations of plasmid copy numbers present in the cells. As shown in Fig. 2C, equivalent hybridization signals were obtained from cells transfected with each of the five plasmids indicating that they contained equimolar amounts of expression plasmids.
To exclude the possibility that an unknown gene product encoded by sequences present in the 1.2 kb sequence was responsible for the observed increases in actin promoter-driven cat gene expression, we carried out co-transfections of Bm5 cells with pBmA.cat and p133 or p153, each of which contains the 1.2-kb HR3 fragment. The CAT assays shown in Fig. 2D showed that neither p133 nor p153 was able to increase CAT expression from pBmA.cat in trans. Therefore, because the HR3 element is capable of stimulating the level of expression of the actin gene promoter only in cis and in a position-and orientation-independent manner, it can be defined as an enhancer of the cytoplasmic actin gene promoter.
Finally, to test whether the HR3 element is also functional in other lepidopteran cells, we transfected Spodoptera frugiperda (Sf21) tissue culture cells with plasmids pBmA.cat and p13315. As shown in Fig. 2E, the CAT assays revealed that the BmNPV HR3 element functions in these cells in a manner analogous to that seen in Bm5 cells.

HR3-mediated Activation Involves Increases in Transcription
Rates-To find out whether the observed increases in CAT activity obtained in the presence of the HR3 element were due to corresponding increases in the accumulation of CAT mRNA in the transfected cells, we carried out a semiquantitative RT-PCR analysis of CAT mRNA present in cells transfected with different plasmids. As shown in Fig. 3A (upper panel), cells transfected with vector p13315, which contains the enhancer upstream of the actin promoter (see Fig. 2A for its orientation), contained much higher quantities of CAT mRNA than cells transfected with vector pBmA.cat.
Nuclear run-on transcription assays were also performed to determine whether the observed increases in CAT mRNA accumulation observed in the presence of the HR3 element were due to corresponding changes in rates of transcription. As shown in Fig. 3B, an average of a 16-fold higher hybridization signal was detected with the run-on RNA probe obtained from nuclei of p13315-transfected cells than with the probe obtained from pBmA.cat-transfected cells. These results indicate that the HR3 element stimulates the expression of the actin.cat fusion gene by increasing the rate at which this gene is transcribed.
Super-activation of Transgene Expression by BmIE1 and HR3-The immediate early gene ie1 of baculoviruses is transcribed in infected lepidopteran cells by the host transcriptional machinery in the absence of any accessory viral products. We have previously shown that BmIE1, the protein encoded by the ie1 gene of BmNPV, functions as co-activator of the actin gene promoter in vitro and potentiates the level of expression of the actin gene promoter also by approximately two orders of magnitude (20). To determine whether BmIE1 and HR3 can act synergistically on the actin gene promoter, we co-transfected Bm5 cells with plasmid pBmIE1 and each of the four HR3/actin-cat constructs. As shown in Fig. 4, the CAT assays showed that CAT activity in the presence of both BmIE1 and HR3 was higher than that obtained in the absence of either BmIE1 or HR3, irrespective of whether the BmIE1 gene was present in the same plasmid as the HR3-enhanced actin promoter construct or supplied to the cells as a separate plasmid. Quantification of the observed enzyme activities (Table I) revealed that cells co-transfected with pBmIE1 and each of HR3/ actin-cat constructs contained more than 1,000-fold higher CAT activity than cells transfected with plasmid pBmA.cat. Primer extension experiments (not shown) revealed that the increases in actin-cat gene expression mediated by BmIE1 and HR3 resulted from increases in transcription rates and that the transcription start sites were the same as those observed with the basal actin expression cassette.
Finally, to demonstrate that the observed increases in transgene expression observed in the presence of the HR3 enhancer and the BmIE1 factor were not transgene specific, we tested the activity of two additional plasmids, pBmA.jhe(kk) and pIE1/153A.jhe(kk), in transfected Bm5 cells. These two plasmids contain the ORF for a modified version of the enzyme juvenile hormone esterase (JHE), a secreted glycoprotein of Heliothis virescens, under the control of the basal actin promoter or under the control of the double-enhanced (HR3 and BmIE1) actin promoter, respectively. As shown in Table I 133A.jhe(kk) was over 1,000-fold higher than that obtained from cells transfected with the basic recombinant vector pB-mA.jhe(kk). Therefore, the HR3-enhancing effect on transgene expression under the control of the silkmoth cytoplasmic gene promoter is promoter rather than transgene specific. Furthermore, considering that BmIE1 alone increases the activity of the actin promoter by approximately 100-fold (Table I) (20), these results also demonstrate that the BmIE1 protein and the HR3 enhancer are capable of increasing each other's effect on the in vitro transcriptional stimulation of the cytoplasmic actin gene promoter by an additional order of magnitude. DISCUSSION It was previously shown that AcNPV HRs can enhance the expression of some early baculovirus genes, such as 35k, 39k, and ie2, as well as that of the Rous sarcoma virus long terminal repeat promoter (3,4,15,16).
In the present study, we have demonstrated that a DNA fragment containing the HR3 sequences of BmNPV can function in vitro as an enhancer for the cytoplasmic actin promoter of B. mori. The definition of the HR3-containing fragment as an enhancer is based on three criteria: first, this fragment is capable of augmenting the level of foreign gene expression directed by the actin gene promoter in a position-and orientation-independent manner; second, it is unable to stimulate the expression of the actin gene promoter in trans; and, third, increased transgene expression from the HR3-enhanced actin promoter is due to significant increases in transcription rates.
Through the results presented in this paper, we have also demonstrated that a baculovirus transcription factor, BmIE1, that was previously shown to act as a transcriptional co-activator of the silkmoth actin gene promoter in transfected lepidopteran cells (20) acts cooperatively with HR3 and mediates a further increase in the transcriptional activity of the actin gene promoter. The resultant overall enhancement in transcriptional activity mediated by the combined action of the HR3 enhancer and the IE1 trans-activator is three orders of magnitude over the level of the basal actin gene promoter.
At present, neither the mode of function of the HR3 sequence-containing fragment nor the mechanism of the cooperative activation effected by the BmIE1 gene product are known. For the HR3 enhancer, however, it is likely that control is exerted through binding of a cellular trans-activator(s) or basal transcription factor(s) to the HR3 sequences. The case of the simian virus 40 (SV40) enhancer, which is composed of two 72-bp repeats and is capable of stimulating the expression of both early viral and cellular genes (26 -29), may be analogous to that of HR3. Mutagenesis and competition experiments have shown that the SV40 enhancer activity is mediated by multiple trans-activators, such as TET-1, HIP116, and Oct proteins (30 -32). Here, it is also worth noting that the cytoplasmic actin gene promoter and the baculovirus genes that are known to be enhanced by HR sequences are all transcribed by RNA polymerase II (33)(34)(35)(36) and also, as was shown by the studies presented here, that the enhancing effect also occurs in heterologous S. frugiperda cells. Therefore, although the mechanism by which the HR3 sequences exert their enhancing function is at present unknown, their enhancing capacity, at least in transfected cells, must be controlled by host factors that have been conserved during evolution.
Whether BmIE1, the viral trans-activator that stimulates the actin promoter in the absence of HR3 element (Fig. 4) (20) and augments the level of cis activation affected by the HR3 element (Fig. 4), is acting exclusively on the actin promoter sequences or if in the presence of the HR3 enhancer it also interacts with the latter remains to be investigated. Worth noting, however, is a previous finding that an AcNPV HR element-binding protein could be detected in insect cells following transfection with the AcNPV ie1 gene and that no obvious binding activity could be detected in control cells (37). Thus, it appears that in addition to binding to the actin promoter sequences, the BmIE1 protein may also bind to the HR3 enhancer and further augment its enhancing activity.
In conclusion, the work described in this paper has demonstrated that two baculovirus-specific genetic elements, the HR3 sequence-containing fragment and the ie1 gene product, are capable of enhancing cooperatively the activity of the silkmoth cytoplasmic actin gene promoter by three orders of magnitude in a virus-free environment. This finding provides an opportunity to develop a recombinant gene expression system in a baculovirus-free insect system. We anticipate that coupling of the enhanced actin promoter-based expression cassette with an appropriate antibiotic selection scheme for generating stably transformed cell lines containing chromosomally integrated copies of this recombinant expression cassette, should result in an insect cell expression system capable of yielding quantities of recombinant proteins comparable to those achieved through conventional infection with baculovirus expression vectors. Such a system should have the additional advantages of continuous production in a cellular environment that, in contrast to that generated by a baculovirus infection, supports continuously proper posttranslational modifications of recombinant proteins. Samples were diluted to obtain values of conversion lying within the linear range of the assays (from 10 to 50% conversion of substrate input), and relative enzyme activities were normalized for those obtained from cells transfected with the relevant genes under the control of the basal actin expression cassette. Units of activity are expressed in nanomoles of chloramphenicol acetylated per minute per mg of soluble cellular protein at 37°C (for CAT assays); and nanomoles of JH-III hydrolyzed per minute per ml of tissue culture medium at 25°C (for JHE assays). Numbers in parentheses indicate time of repeat transfections for each vector combination.