- Email: [email protected]
Expression of grass carp growth hormone by baculovirus in silkworm larvae
Biochimica et Biophysica Acta 1381 (1998) 331^339
Expression of grass carp growth hormone by baculovirus in silkworm larvae W.K.K. Ho a; *, Z.Q. Meng b , H.R. Lin c , C.T. Poon a , Y.K. Leung a , K.T. Yan a , N. Dias a , A.P.K. Che a , J. Liu b , W.M. Zheng c , Y. Sun c , A.O.L. Wong d a
Department of Biochemistry, Chinese University of Hong Kong, Shatin, N.T., Hong Kong, China b Zhejiang Academy of Agricultural Sciences, Hangzhou, Zhejiang, China c Department of Biology, Zhongshan University, Guangzhou, Guangdong, China d Department of Zoology, University of Hong Kong, Hong Kong, China Received 10 February 1998; revised 25 March 1998; accepted 2 April 1998
Abstract A total of five recombinant Bombyx mori nuclear polyhedrosis viruses (BMNPV) carrying the grass carp (Ctenopharyngodon idellus) growth hormone (GH) cDNA were constructed in this study. Two of them were able to express the hormone up to a level of 12 Wg/ml medium when cultured B. mori cells were infected for 4 days. Inoculation of the viruses into silkworm (B. mori) host significantly increased the level of GH achievable. The amount of hormone produced per larva was estimated to be around 1 mg. The recombinant grass carp GH had immunological and biological activities similar to the native hormone. The N-terminal sequence of the recombinant hormone was the same as the native one, indicating that the fish signal peptide was correctly processed by the insect cells. Silkworm powder prepared from larvae infected with the recombinant virus was used as food supplement for fish. Compared with the control, this dietary supplement was effective in increasing the growth rate of juvenile carp. ß 1998 Elsevier Science B.V. All rights reserved. Keywords: Baculovirus; Fish; Growth hormone; (Silkworm); (Bombyx mori)
1. Introduction Growth hormone (GH) is a polypeptide hormone secreted from the anterior pituitary gland for inducing growth in vertebrates. Its amino acid sequence from a number of species has been determined by cDNA cloning in recent years. The analogy of these sequences has established the role of GH in species in the animal kingdom ranging from ¢sh to man. Fish GH has been extensively studied in the past 10 years
* Corresponding author. Fax: +852 26035123; E-mail: [email protected]
and the GH sequences of over 20 species of ¢sh have been determined. Other than in comparative studies, the interest in ¢sh GH is due to its potential application in aquaculture. For example, the addition of GH to the food of ¢sh was shown to be able to enhance growth rate [1^3]. In order to utilize GH as a food supplement, the hormone has to be manufactured in a cost e¡ective way. Fish GH has been successfully expressed in Escherichia coli in a number of studies [4^10]. Unfortunately, in every case the hormone was expressed as inclusion body and required to be refolded before it could be biologically active. This additional manipulation makes the process ine¤cient and uneconomi-
0304-4165 / 98 / $19.00 ß 1998 Elsevier Science B.V. All rights reserved. PII: S 0 3 0 4 - 4 1 6 5 ( 9 8 ) 0 0 0 4 4 - 0
BBAGEN 24641 14-8-98
332
W.K.K. Ho et al. / Biochimica et Biophysica Acta 1381 (1998) 331^339
cal. To overcome this problem, others have tried using yeast [11] and baculovirus [12] to produce the hormone in a soluble form. However, both of these systems have drawbacks. For yeast the plasmid transformant may be unstable in the host, and for baculovirus the cost of producing the hormone in cell culture may be too ine¤cient and expensive. In the present study, we investigated the feasibility of using silkworm (Bombyx mori) larvae to produce recombinant grass carp (Ctenopharyngodon idellus) GH by infecting them with a recombinant baculovirus. Our results show that high levels of biologically active GH can be produced in this manner. The hormone produced had in vitro and in vivo properties similar to the native hormone. 2. Materials and methods 2.1. DNA manipulation The cDNA clone carrying the grass carp GH was isolated and characterized as described by Ho et al. [13]. The parent plasmids, pBM030 [14] and pAc373 [15], used for the construction of the transfer vectors, were kindly provided by Dr. S. Maeda of the University of California (Davis, CA, USA) and Prof. Y.K. Sun of the Shanghai Institute of Biotechnology, respectively. Wild-type BMNPV and B. mori cells were also provided by Dr. S. Maeda. For the construction of pSD61, pSD78 and P6, the parent plasmid pAc373 was used and for pEE and pEX, pBM030 was used. Grass carp GH cDNAs with different 5P sequences to the ATG initiation site were prepared by standard procedures using either modi¢ed clones of GH cDNA we have at hand (pSD61, pSD78, p6, pEE) or PCR ampli¢ed fragments with speci¢c cloning sites inserted at the 5P and 3P ends (pEX). The proper alignment of the £anking sequences (Fig. 1) in the di¡erent vectors were con¢rmed by dideoxynucleotide sequencing [16]. In pSD61 and pSD78, the ¢rst ten codons were E. coli preferred codons and not the ones found in the ¢sh cDNA [13]. 2.2. Construction of recombinant BMNPV BMNPV DNA was prepared from wild-type virus
particles as described by Summers and Smith [17]. One microgram viral DNA was mixed with 2 Wg transfer vector in a ¢nal volume of 100 Wl HBS bu¡er (20 mM HEPES, 150 mM NaCl, pH 7.4). The DNA was then added into a 250 Wl HBS solution containing 70 Wg DOTAP (Boehringer Mannheim Biochemica, Germany). The mixture was incubated at room temperature for 10 min and then transferred dropwise into a 100 mm dish in which B. mori cells were growing in log phase. After incubating for 24 h, the cell medium was changed. Cells were allowed to grow at 26³C until occlusion bodies began to appear. The cell medium was harvested at this point and used as the primary stock for the screening of recombinant virus. The transfection was considered not successful if no occlusion body was observed after 2 weeks. Recombinant virus containing the grass carp GH cDNA was screened by cycles of plaque assay essentially as described by Maeda [14]. A fragment of 32 P labeled GH cDNA was used as the probe. After the initial identi¢cation of the recombinant virus, further puri¢cation was carried out by limiting dilution. The ¢nal preparation of the virus was considered to be pure when no occlusion body could be seen after an extensive period of infection. All ¢ve recombinant viruses prepared in this study were con¢rmed to carry the GH cDNA by Southern blot hybridization, even though some of them did not express the hormone at a level high enough for detection by Western blotting. 2.3. Infection of silkworm Silkworms in the early 5th instar stage were used for infection. Five microliters viral solution containing approx. 2U105 particles were injected into the body cavity of the silkworm larvae. To assess the expression of GH, hemolymph samples were collected from the larvae on di¡erent days by cutting one of the cuticles in the abdomen area. Normally, no symptoms of infection could be observed on the ¢rst 3 days of infection. By the 4th day, the larvae began to lose appetite and most of them would die on the 5th day. For the preparation of silkworm powder, larvae were normally harvested on either the 4th or the 5th day after inoculation. Silkworms were frozen in liquid nitrogen and stored
BBAGEN 24641 14-8-98
W.K.K. Ho et al. / Biochimica et Biophysica Acta 1381 (1998) 331^339
at 370³C before they were ground up and freezedried. 2.4. Puri¢cation of GH from hemolymph Silkworm larvae were infected with the recombinant virus P6 for 4 days and their hemolymph were collected as the starting material for puri¢cation. The hemolymph from six larvae were ¢rst pooled and diluted 10 times with PBS (0.01 M phosphate bu¡er, pH 7.4, containing 0.15 M NaCl). The diluted material was then passed through a Seppak C-18 cartridge (Waters, MA) 3 times. The cartridge was then washed 3 times with 10 ml PBS followed by 1 ml 20% acetonitrile in 1 mM trichloroacetic acid 5 times. The GH was eluted by ¢ve 1 ml fractions of 60% acetonitrile in 1 mM trichloroacetic acid. The organic solvent was immediately removed from the semipuri¢ed GH by passing through a Sephadex G25 column and the peak containing the hormone was lyophilized. The lyophilized material was then dissolved in 2 ml PBS and the GH in it was immunoaf¢nity puri¢ed by passing through a column containing an anti-grass carp GH monoclonal antibody immobilized on Sepharose beads. The bound hormone was eluted by 0.1 M glycine bu¡er (pH 2.7). At this point, the purity of the recombinant GH was around 90% as determined by SDS-PAGE. For Nterminal sequencing, the GH was further puri¢ed in a 12.5% gel by SDS-PAGE and then transferred to the PVDF membrane for sequencing.
333
versity of Alberta. The hormone was labeled with I using the iodogen method [20]. Radioreceptor assay was performed essentially as described by Fryer [21] using liver membrane prepared from common carp. Immunohistochemical studies on grass carp pituitary sections were carried out according to Ge and Peter [22]. The antiserum used was the same as the one used for radioimmunoassay. The binding of the antibody was visualized by staining with the Vectastain ABC Kit (Vector Lab., CA, USA). 125
2.6. Fish experiment Juvenile grass carp with an average weight of 25.9 g were used in the feeding study. The ¢sh were randomly divided into two groups and were given diets with and without GH as described in Table 1. The basic food contained, by weight, 40% casein, 30% dextrin, 2% peanut oil, 1% vitamins, 17.7% cellulose and 9.3% minerals and salts. Food was given in the form of semi-dried pellet twice a day. The groups of ¢sh were housed in separate aquariums with water temperature maintained at around 20^25³C. At weekly intervals, the weight and length of the ¢sh were measured. The relative somatic and linear growth rates (RSGR and RLGR) were calculated according to the following equations: RSGR = (Wt 3W0 )/W0 U100%; RLGR = (Lt 3L0 )/L0 U100%, where t = 6 weeks.
2.5. Analytical methods
3. Results
SDS-PAGE was performed according to Laemmli [18], and the gels were visualized by either Coomassie blue or silver staining. Western blotting was done in a BioRad (Richmond, CA, USA) semi-dry transfer apparatus according to the manufacturer's instruction. The primary antibody used was a mouse monoclonal antibody against the grass carp GH. This antibody cross-reacts with common carp and gold ¢sh GH but not with human and bovine GH. An alkaline phosphatase labeled anti-mouse IgG was used to visualize the location of the primary antibody. Radioimmunoassay was performed by a procedure described by Marchant et al. [19] using an anti-grass carp GH serum supplied by R.E. Peter of the Uni-
3.1. Construction of recombinant BMNPV A total of ¢ve transfer vectors were constructed to produce the recombinant BMNPV for the expression of the grass carp GH. All of them were designed to insert the GH cDNA downstream from the polyhedrin promoter in the viral genome by recombination. The major di¡erences between these constructs are the length and base composition of the 5P untranslated ends. Details of the 5P £anking sequences of each of the transfer vectors are summarized in Fig. 1. With the ¢ve transfer vectors, we were able to generate ¢ve di¡erent recombinant viruses. By dot blot and Southern hybridization, each one of them
BBAGEN 24641 14-8-98
334
W.K.K. Ho et al. / Biochimica et Biophysica Acta 1381 (1998) 331^339
Fig. 1. Features of the 5P untranslated regions of the transfer vectors used in this study. For the pAc373 derived vectors, the cloning site was BamHI and for the pBM030 derived ones, the cloning site was between EcoRI and XbaI. Bases marked with an asterisk indicate the position of a stem loop structure. Spacing between the polyhedrin promoter and the ATG start codon is as indicated. Codon usage by the di¡erent vectors is as indicated.
was shown to carry the grass carp GH cDNA (results not shown). When B. mori cells were infected with these viruses for 4 days, only strains P6 and EX were able to express GH in quantities high enough to be detected by Western blotting. Strain EE showed a much lower level of expression while strains SD78 and SD61 did not appear to show any expression at all. The concentration of hormone secreted into the medium over a 4 day period by the di¡erent recombinant viruses as determined by radioimmunoassay is summarized in the right column of Fig. 1. 3.2. Expression of GH in silkworm larvae Silkworm larvae at the 5th instar stage were infected with P6 and EX virus at a dose of approx.
2U105 particles per worm in a volume of 5 Wl. Worms were collected and bled on di¡erent days to ascertain the progress of infection and the expression of GH. As indicated in Fig. 2, between days 0 and 3, the level of GH in the hemolymph was undetectable. On day 4, the concentration of GH in both P6 and EX infected larvae increased rapidly from none to over 700 Wg per larvae. By day 5, all the infected silkworms became sick and began to die. The level of GH in the hemolymph also began to decrease, presumably due to proteolytic degradation by enzymes released from the dying cells. An SDSPAGE/Western blot analysis on P6 infected silkworm hemolymph samples is shown in the insert of Fig. 2. An immunoreactive band with a size of approx. 21^25 kDa was clearly noticed in samples col-
BBAGEN 24641 14-8-98
W.K.K. Ho et al. / Biochimica et Biophysica Acta 1381 (1998) 331^339
335
Fig. 2. Time course in the expression of GH in silkworm larvae infected with P6 and EX baculovirus. Inset is an SDS-PAGE and Western blot analysis of hemolymph samples collected on di¡erent days after P6 inoculation. Lanes: 1, bovine serum albumin; 2, molecular weight markers; 3, uninfected larvae; 4, 24 h sample; 5, 48 h sample; 6, 72 h sample; 7, 96 h sample; 8, 120 h sample. The immunoreactive band associated with the 30 000 molecular weight marker (carbonic anhydrase) was presumably due to cross-reactivity with the anti-grass carp GH monoclonal antibody.
lected on days 4 and 5. In contrast, in the Coomassie blue stained gel, only a faint band was noticed between the 20.1 and 30 kDa markers, indicating that the amount of recombinant GH in these samples was relatively low compared with that of other proteins in the hemolymph. Nonetheless, when GH in the hemolymph was puri¢ed by immunoa¤nity chromatography, its molecular mass (21 kDa) was similar to that of the native hormone (Fig. 3). 3.3. Characterization of GH expressed by P6
Fig. 3. SDS-PAGE analysis of recombinant GH puri¢ed by immunoa¤nity chromatography. Lanes: 1, molecular weight markers; 2, semi-puri¢ed GH from P6 infected silkworm; 3, a semi-puri¢ed GH from grass carp pituitary glands. The gel was stained with silver reagent and overloaded to visualize contaminating components.
Since the expression e¤ciencies of GH by P6 and EX are similar, we decided to select only one of them for further characterization. In both radioimmunoassay and radioreceptor assay, crude preparations of P6 infected silkworms were able to displace 125 I labeled GH in a dose dependent manner (Fig. 4). The binding displacement curves generated by the
BBAGEN 24641 14-8-98
336
W.K.K. Ho et al. / Biochimica et Biophysica Acta 1381 (1998) 331^339
Fig. 4. Radioimmunoassay and radioreceptor assay of GH from P6 infected silkworm larvae. The weight of the silkworm powder and hemolymph was the dried weight of the lyophilized material. Puri¢ed native hormone was used as the standard.
BMNPV expressed GH were essentially parallel to those generated by the native hormone. When uninfected preparations were used in either the radioreceptor or radioimmunoassay, no apparent displacement of radioactively labeled GH was observed, even at concentrations 10^100-fold higher. To further assess the similarity of the P6 expressed GH with the native one, we tested the ability of the P6 infected hemolymph to block the immunospeci¢c staining of somatotrophs in the grass carp pituitary gland. The antiserum used in this study was raised against natural grass carp GH and as indicated in Fig. 5A, it speci¢cally stained the intermediate region of the pituitary gland. Pretreatment with either GH or hemolymph collected from P6 infected larvae
abolished this staining (Fig. 5B,D). In contrast, hemolymph from silkworm infected with P6 on day 1 was not able to block the staining by the GH-speci¢c antiserum (Fig. 5C). This is to be expected, as we have previously shown that no GH could be detected at this time. We have made a number of attempts using conventional chromatography to purify the recombinant grass carp GH from silkworm. Unfortunately, these methods proved to be rather ine¤cient and we could not get the hormone puri¢ed from contaminants in the hemolymph. In another study, we raised a highly speci¢c monoclonal antibody against the grass carp GH. Using this antibody as an immunoa¤nity ligand, we were able to purify the P6 produced hormone to approx. 90% purity (Fig. 3). After further puri¢cation by SDS-PAGE, the hormone was taken for N-terminal sequencing. The sequence obtained was Ser-Glu-Asn-Gln-Arg-Leu-Phe-Asn-Asn-Ala-; it was exactly identical to the N-terminal sequence of the matured native hormone [13]. 3.4. Induction of growth in ¢sh by feeding with P6 infected silkworm To evaluate the biological activity of the GH produced in the P6 infected silkworm, a feeding experiment was performed. Juvenile grass carp with an average weight and length of 25.9 g and 12.6 cm, respectively, were randomly divided into two groups of 15 ¢sh each. The ¢rst group was fed with a diet containing 20 mg uninfected silkworm powder per gram diet and the second group with a diet containing 20 mg P6 infected silkworm powder per gram diet. The two groups of ¢sh were housed in separate aquariums and fed twice a day for a total of 6 weeks. The weight and length of the ¢sh were measured weekly. At the end of the period, the RSGR and RLGR were calculated, and the results are presented in Table 1. Compared with the control, ¢sh fed with
Table 1 E¡ect of feeding P6 infected silkworm powder on growth of juvenile grass carps Treatment
n
RSGR (%) (6 weeks)
RLGR (%) (6 weeks)
Uninfected silkworm (20 mg/g diet) P6 infected silkworm (20 mg/g diet)a
15 15
67.26 þ 4.41* 86.20 þ 11.7*
18.47 þ 1.76 20.08 þ 3.06
*Values are signi¢cantly di¡erent (P 6 0.001) from each other by the t-test. This amount of silkworm powder roughly contained 20 Wg GH as estimated by radioimmunoassay.
a
BBAGEN 24641 14-8-98
W.K.K. Ho et al. / Biochimica et Biophysica Acta 1381 (1998) 331^339
337
Fig. 5. Cross-section of grass carp pituitary gland stained with an anti-grass carp GH serum. (A) No blocking; (B) blocked with 0.5 mg/ml grass carp GH; (C) blocked with P6 infected silkworm hemolymph collected 24 h after inoculation; (D) same as C except hemolymph was collected 120 h after inoculation. The hemolymph was diluted 10U with bu¡ered saline before immunostaining.
the diet containing the P6 infected silkworm had a 28% and a 8.7% higher RSGR and RLGR, respectively. The well-being of the ¢sh fed with the infected silkworm appeared to be similar to that of the control group; there was no sign of viral infection. In a separate study, the process of freeze-drying was determined to be e¡ective in killing the recombinant BMNPV in the silkworm larvae. 4. Discussion GH from a number of species of ¢sh has been successfully expressed in E. coli by di¡erent research groups [4^10]. A major interest in recombinant ¢sh GH is to utilize it to enhance growth in ¢sh. To achieve this goal, an e¤cient and economical method in producing GH must be established. E. coli is one of the most e¤cient micro-organisms for producing recombinant proteins. In the case of ¢sh GH, an expression level up to 20^30% total cell protein has been reported [6^10]. Unfortunately, in all of these cases, the hormones were expressed in inclusion body form and they had to be solubilized and renatured before they could become biologically active. These
extra steps can signi¢cantly reduce yield and increase the production cost of the hormone. In the present study, we have attempted to overcome the problems associated with E. coli expression using a baculovirus. This system, although less e¤cient, can, however, produce a properly folded soluble GH with biological activity. Since our primary goal was to utilize the recombinant hormone for feeding ¢sh, we reasoned that as long as the hormone can be produced in a non-toxic biologically active form, it could be used directly without any puri¢cation. The use of silkworm larvae as the host for the recombinant BMNPV was a logical choice because (1) up to milligram quantities of recombinant protein can be produced in one larva, (2) the larva is non-toxic and can be used as food for the ¢sh, and (3) BMNPV has a very narrow host range, and as a result, it should be relatively safe to use. We have prepared a total of ¢ve recombinant BMNPV to express the grass carp GH in this study. Only two of them, P6 and EX, were able to express the hormone in high quantity. The major di¡erence in these constructs was the spacing between the polyhedrin promoter and the ATG translation initiation site. For P6 this spacing was 73 bp and for EX it was
BBAGEN 24641 14-8-98
338
W.K.K. Ho et al. / Biochimica et Biophysica Acta 1381 (1998) 331^339
28 bp. Compared with the other three low-expressing recombinants, viz., EE, SD61 and SD78, these spacings were similar in length. Thus, the higher level of expression by P6 and EX cannot be merely accounted for by a di¡erence in spacing between the polyhedrin promoter and the ATG translation initiation site. The present results are in contrast to the ¢ndings of Luckow and Summers [23] which showed that a shorter spacing is more e¤cient for expression. Another factor that may a¡ect the e¤ciency of mRNA translation is the presence of a stem loop structure in the leading sequence of the translation initiation site [24,25]. When the leading sequences of the ¢ve recombinants were analyzed, EX and EE did not have any stem loop structure while P6, SD61 and SD78 had one (see Table 1). Since the spacings in EX, SD61 and SD78 were relatively short, ranging from 15 to 28 bp, it is possible that the presence of a stem loop may have had a detrimental e¡ect on expression. On the other hand, P6 has one stem loop in the leading sequence but its position was quite far away from the start codon; hence it might have less negative in£uence on message translation. In the case of EE, there was no stem loop and yet it could not express GH e¤ciently. In this respect, the untranslated region's sequence or composition is also known to play a critical role in expression [25]. The grass carp GH expressed by P6 has an expected size and N-terminal sequence similar to the native hormone. This indicates that silkworm cells could recognize the signal peptide of the ¢sh and processed it correctly. In a similar study on human GH, Kadono-Okuda et al. [26] also observed the correct processing of the human signal peptide. The level of GH expressed in the present study was around 1 mg/ml hemolymph. This level of expression is manyfold higher than that achievable in cell culture. Since silkworms can be raised at much lower cost than cell culture, this approach is probably one of the most economical ways to make a soluble recombinant protein. In spite of the aforementioned merits, the silkworm expression system has two disadvantages. First, the recombinant protein may be more di¤cult to purify. Second, the process may be di¤cult to validate because parameters in culturing silkworm larvae may be subject to more variables
than in in vitro methods. As far as our application in using the recombinant hormone as food supplement is concerned, these two problems do not appear to be critical as we do not need a puri¢ed product for our application. A key consideration in using BMNPV infected silkworm as food is the potential environmental concerns it may raise. We have performed some preliminary experiments on the viability of the virus after freeze-drying the infected silkworms. Using cultured B. mori cells, we were unable to detect any sign of viral activity from the silkworm powder. Presumably, the process of drying may be enough to destroy the virus. Further tests are presently being carried out to ascertain the e¡ect of P6 in infecting other species, including other ¢sh species, during its application as food. One of the major advantages of using baculovirus to express eukaryotic genes is its ability to produce a recombinant protein closely resembling the native one. The product is normally expressed in soluble form and in higher yield. Although insect cells can perform glycosylation, the carbohydrate chain added may not be identical to that made by mammalian cells [27]. In this study, we have examined the integrity of the grass carp GH expressed in silkworm by four criteria, i.e., molecular weight, immunoreactivity, receptor binding and N-terminal sequence. Compared with the native hormone, the recombinant GH was shown to have similar properties. Unfortunately, due to the lack of a good and consistent biological assay for ¢sh GH we were not able to compare the speci¢c activity of the recombinant hormone with the native one. Nonetheless, the result of our feeding experiment did provide partial evidence to indicate that the GH produced in the silkworm was biologically active. Acknowledgements This research was supported by grants from the Croucher Foundation, Industrial Support Fund and the Research Grants Council of Hong Kong. We are most grateful to Prof. S. Maeda and Prof. R.E. Peter for their gifts of research materials. We also would like to thank Dr. T.A. Marchant for performing the radioreceptor assays.
BBAGEN 24641 14-8-98
W.K.K. Ho et al. / Biochimica et Biophysica Acta 1381 (1998) 331^339
References [1] E. McLean, E.M. Donaldson, H.M. Dye, L.M. Souza, Aquaculture 91 (1990) 197^203. [2] S. Moriyama, H. Yamamoto, S. Sugimoto, A. Toshio, T. Hirano, H. Kawauchi, Aquaculture 112 (1993) 99^106. [3] H.J. Tsai, J.C. Kuo, S.W. Lou, T.T. Kuo, Prog. Fish Culturist 56 (1994) 7^12. [4] S. Sekine, T. Mizukami, T. Nishi, Y. Kuwana, A. Saito, M. Sata, S. Itoh, H. Kawauchi, Proc. Natl. Acad. Sci. USA 82 (1985) 4306^4310. [5] L.B. Agellon, T.T. Chen, DNA 5 (1986) 463^471. [6] A. Saito, S. Sekine, Y. Komatsu, M. Sato, T. Hirano, S. Itoh, Gene 73 (1988) 545^551. [7] F. Rentier-Deirue, D. Swennen, J.C. Philippart, C. L'Hoir, M. Lion, O. Benrubi, J.A. Martial, DNA 8 (1989) 271^278. [8] N. Sato, M. Sakaguchi, M. Nonaka, K. Murato, K. Watanabe, T. Hayamie, Y. Kariya, A. Kimura, S. Kimura, Appl. Microbiol. Biotechnol. 30 (1989) 153^159. [9] H.J. Tsai, K.L. Lin, T.T. Chen, Comp. Biochem. Physiol. 104B (1993) 803^810. [10] M. Fine, E. Sakal, D. Vashdi, V. Daniel, A. Levanon, O. Lipshitz, A. Gertler, Gen. Comp. Endocrinol. 89 (1993) 51^ 61. [11] H.J. Tsai, C.F. Tseng, T.T. Kuo, Bull. Inst. Zool. Acad. Sin. 32 (1993) 162^170. [12] H.J. Tsai, H.M. Chen, C.F. Lo, Can. J. Fish. Aquatic Sci. 51 (1994) 1^7. [13] W.K.K. Ho, W.H. Tsang, N.P. Dias, Biochem. Biophys. Res. Commun. 161 (1989) 1239^1243.
339
[14] S. Maeda, in: J. Mitsuhashi (Ed.), Invertebrate Cell System Applications, Vol. I, CRC Press, Boca Raton, FL, 1989, pp. 167^181. [15] V.A. Luckow, M.D. Summers, Biotechnology 6 (1988) 47^ 55. [16] F. Sanger, S. Nicklen, A.R. Coulson, Proc. Natl. Acad. Sci. USA 74 (1977) 5463^5467. [17] M.D. Summers, G.E. Smith, A Manual of Methods for Baculovirus Vectors and Insect Cell Culture Procedures, Texas Agricultural Experiment Station Bulletin No. 1555, 1987. [18] U.K. Laemmli, Nature 227 (1970) 680^685. [19] T.A. Marchant, R.A. Fraser, P.C. Andrews, R.E. Peter, Regul. Pept. 17 (1987) 41^52. [20] P.R.P. Salacinski, C. McLean, J.E.C. Sykes, V. ClementJones, P.J. Lowry, Anal. Biochem. 172 (1981) 136^146. [21] J.N. Fryer, Gen. Comp. Endocrinol. 39 (1979) 123^130. [22] W. Ge, R.E. Peter, Gen. Comp. Endocrinol. 95 (1994) 213^ 221. [23] V.A. Luckow, M.D. Summers, Virology 170 (1989) 31^39. [24] E. Jay, A.K. Seth, J. Romments, A. Sood, G. Jay, Nucleic Acid Res. 10 (1982) 6319^6329. [25] H. Dalboge, S. Carlsen, E.B. Jensen, T. Christensen, H.M. Dahl, DNA 7 (1988) 399^405. [26] K. Kadono-Okuda, M. Ymamoto, Y. Higashino, K. Taniai, Y. Kato, S. Chowdhury, J.H. Xu, S.K. Choi, M. Sugiyama, K. Nakashima, S. Maeda, M. Ymakawa, Biochem. Biophys. Res. Commun. 213 (1995) 389^396. [27] D.M. Wojchowski, S.H. Orkin, A.J. Sytkowski, Biochim. Biophys. Acta 910 (1987) 224^232.
BBAGEN 24641 14-8-98
Expression of grass carp growth hormone by baculovirus in silkworm larvae W.K.K. Ho a; *, Z.Q. Meng b , H.R. Lin c , C.T. Poon a , Y.K. Leung a , K.T. Yan a , N. Dias a , A.P.K. Che a , J. Liu b , W.M. Zheng c , Y. Sun c , A.O.L. Wong d a
Department of Biochemistry, Chinese University of Hong Kong, Shatin, N.T., Hong Kong, China b Zhejiang Academy of Agricultural Sciences, Hangzhou, Zhejiang, China c Department of Biology, Zhongshan University, Guangzhou, Guangdong, China d Department of Zoology, University of Hong Kong, Hong Kong, China Received 10 February 1998; revised 25 March 1998; accepted 2 April 1998
Abstract A total of five recombinant Bombyx mori nuclear polyhedrosis viruses (BMNPV) carrying the grass carp (Ctenopharyngodon idellus) growth hormone (GH) cDNA were constructed in this study. Two of them were able to express the hormone up to a level of 12 Wg/ml medium when cultured B. mori cells were infected for 4 days. Inoculation of the viruses into silkworm (B. mori) host significantly increased the level of GH achievable. The amount of hormone produced per larva was estimated to be around 1 mg. The recombinant grass carp GH had immunological and biological activities similar to the native hormone. The N-terminal sequence of the recombinant hormone was the same as the native one, indicating that the fish signal peptide was correctly processed by the insect cells. Silkworm powder prepared from larvae infected with the recombinant virus was used as food supplement for fish. Compared with the control, this dietary supplement was effective in increasing the growth rate of juvenile carp. ß 1998 Elsevier Science B.V. All rights reserved. Keywords: Baculovirus; Fish; Growth hormone; (Silkworm); (Bombyx mori)
1. Introduction Growth hormone (GH) is a polypeptide hormone secreted from the anterior pituitary gland for inducing growth in vertebrates. Its amino acid sequence from a number of species has been determined by cDNA cloning in recent years. The analogy of these sequences has established the role of GH in species in the animal kingdom ranging from ¢sh to man. Fish GH has been extensively studied in the past 10 years
* Corresponding author. Fax: +852 26035123; E-mail: [email protected]
and the GH sequences of over 20 species of ¢sh have been determined. Other than in comparative studies, the interest in ¢sh GH is due to its potential application in aquaculture. For example, the addition of GH to the food of ¢sh was shown to be able to enhance growth rate [1^3]. In order to utilize GH as a food supplement, the hormone has to be manufactured in a cost e¡ective way. Fish GH has been successfully expressed in Escherichia coli in a number of studies [4^10]. Unfortunately, in every case the hormone was expressed as inclusion body and required to be refolded before it could be biologically active. This additional manipulation makes the process ine¤cient and uneconomi-
0304-4165 / 98 / $19.00 ß 1998 Elsevier Science B.V. All rights reserved. PII: S 0 3 0 4 - 4 1 6 5 ( 9 8 ) 0 0 0 4 4 - 0
BBAGEN 24641 14-8-98
332
W.K.K. Ho et al. / Biochimica et Biophysica Acta 1381 (1998) 331^339
cal. To overcome this problem, others have tried using yeast [11] and baculovirus [12] to produce the hormone in a soluble form. However, both of these systems have drawbacks. For yeast the plasmid transformant may be unstable in the host, and for baculovirus the cost of producing the hormone in cell culture may be too ine¤cient and expensive. In the present study, we investigated the feasibility of using silkworm (Bombyx mori) larvae to produce recombinant grass carp (Ctenopharyngodon idellus) GH by infecting them with a recombinant baculovirus. Our results show that high levels of biologically active GH can be produced in this manner. The hormone produced had in vitro and in vivo properties similar to the native hormone. 2. Materials and methods 2.1. DNA manipulation The cDNA clone carrying the grass carp GH was isolated and characterized as described by Ho et al. [13]. The parent plasmids, pBM030 [14] and pAc373 [15], used for the construction of the transfer vectors, were kindly provided by Dr. S. Maeda of the University of California (Davis, CA, USA) and Prof. Y.K. Sun of the Shanghai Institute of Biotechnology, respectively. Wild-type BMNPV and B. mori cells were also provided by Dr. S. Maeda. For the construction of pSD61, pSD78 and P6, the parent plasmid pAc373 was used and for pEE and pEX, pBM030 was used. Grass carp GH cDNAs with different 5P sequences to the ATG initiation site were prepared by standard procedures using either modi¢ed clones of GH cDNA we have at hand (pSD61, pSD78, p6, pEE) or PCR ampli¢ed fragments with speci¢c cloning sites inserted at the 5P and 3P ends (pEX). The proper alignment of the £anking sequences (Fig. 1) in the di¡erent vectors were con¢rmed by dideoxynucleotide sequencing [16]. In pSD61 and pSD78, the ¢rst ten codons were E. coli preferred codons and not the ones found in the ¢sh cDNA [13]. 2.2. Construction of recombinant BMNPV BMNPV DNA was prepared from wild-type virus
particles as described by Summers and Smith [17]. One microgram viral DNA was mixed with 2 Wg transfer vector in a ¢nal volume of 100 Wl HBS bu¡er (20 mM HEPES, 150 mM NaCl, pH 7.4). The DNA was then added into a 250 Wl HBS solution containing 70 Wg DOTAP (Boehringer Mannheim Biochemica, Germany). The mixture was incubated at room temperature for 10 min and then transferred dropwise into a 100 mm dish in which B. mori cells were growing in log phase. After incubating for 24 h, the cell medium was changed. Cells were allowed to grow at 26³C until occlusion bodies began to appear. The cell medium was harvested at this point and used as the primary stock for the screening of recombinant virus. The transfection was considered not successful if no occlusion body was observed after 2 weeks. Recombinant virus containing the grass carp GH cDNA was screened by cycles of plaque assay essentially as described by Maeda [14]. A fragment of 32 P labeled GH cDNA was used as the probe. After the initial identi¢cation of the recombinant virus, further puri¢cation was carried out by limiting dilution. The ¢nal preparation of the virus was considered to be pure when no occlusion body could be seen after an extensive period of infection. All ¢ve recombinant viruses prepared in this study were con¢rmed to carry the GH cDNA by Southern blot hybridization, even though some of them did not express the hormone at a level high enough for detection by Western blotting. 2.3. Infection of silkworm Silkworms in the early 5th instar stage were used for infection. Five microliters viral solution containing approx. 2U105 particles were injected into the body cavity of the silkworm larvae. To assess the expression of GH, hemolymph samples were collected from the larvae on di¡erent days by cutting one of the cuticles in the abdomen area. Normally, no symptoms of infection could be observed on the ¢rst 3 days of infection. By the 4th day, the larvae began to lose appetite and most of them would die on the 5th day. For the preparation of silkworm powder, larvae were normally harvested on either the 4th or the 5th day after inoculation. Silkworms were frozen in liquid nitrogen and stored
BBAGEN 24641 14-8-98
W.K.K. Ho et al. / Biochimica et Biophysica Acta 1381 (1998) 331^339
at 370³C before they were ground up and freezedried. 2.4. Puri¢cation of GH from hemolymph Silkworm larvae were infected with the recombinant virus P6 for 4 days and their hemolymph were collected as the starting material for puri¢cation. The hemolymph from six larvae were ¢rst pooled and diluted 10 times with PBS (0.01 M phosphate bu¡er, pH 7.4, containing 0.15 M NaCl). The diluted material was then passed through a Seppak C-18 cartridge (Waters, MA) 3 times. The cartridge was then washed 3 times with 10 ml PBS followed by 1 ml 20% acetonitrile in 1 mM trichloroacetic acid 5 times. The GH was eluted by ¢ve 1 ml fractions of 60% acetonitrile in 1 mM trichloroacetic acid. The organic solvent was immediately removed from the semipuri¢ed GH by passing through a Sephadex G25 column and the peak containing the hormone was lyophilized. The lyophilized material was then dissolved in 2 ml PBS and the GH in it was immunoaf¢nity puri¢ed by passing through a column containing an anti-grass carp GH monoclonal antibody immobilized on Sepharose beads. The bound hormone was eluted by 0.1 M glycine bu¡er (pH 2.7). At this point, the purity of the recombinant GH was around 90% as determined by SDS-PAGE. For Nterminal sequencing, the GH was further puri¢ed in a 12.5% gel by SDS-PAGE and then transferred to the PVDF membrane for sequencing.
333
versity of Alberta. The hormone was labeled with I using the iodogen method [20]. Radioreceptor assay was performed essentially as described by Fryer [21] using liver membrane prepared from common carp. Immunohistochemical studies on grass carp pituitary sections were carried out according to Ge and Peter [22]. The antiserum used was the same as the one used for radioimmunoassay. The binding of the antibody was visualized by staining with the Vectastain ABC Kit (Vector Lab., CA, USA). 125
2.6. Fish experiment Juvenile grass carp with an average weight of 25.9 g were used in the feeding study. The ¢sh were randomly divided into two groups and were given diets with and without GH as described in Table 1. The basic food contained, by weight, 40% casein, 30% dextrin, 2% peanut oil, 1% vitamins, 17.7% cellulose and 9.3% minerals and salts. Food was given in the form of semi-dried pellet twice a day. The groups of ¢sh were housed in separate aquariums with water temperature maintained at around 20^25³C. At weekly intervals, the weight and length of the ¢sh were measured. The relative somatic and linear growth rates (RSGR and RLGR) were calculated according to the following equations: RSGR = (Wt 3W0 )/W0 U100%; RLGR = (Lt 3L0 )/L0 U100%, where t = 6 weeks.
2.5. Analytical methods
3. Results
SDS-PAGE was performed according to Laemmli [18], and the gels were visualized by either Coomassie blue or silver staining. Western blotting was done in a BioRad (Richmond, CA, USA) semi-dry transfer apparatus according to the manufacturer's instruction. The primary antibody used was a mouse monoclonal antibody against the grass carp GH. This antibody cross-reacts with common carp and gold ¢sh GH but not with human and bovine GH. An alkaline phosphatase labeled anti-mouse IgG was used to visualize the location of the primary antibody. Radioimmunoassay was performed by a procedure described by Marchant et al. [19] using an anti-grass carp GH serum supplied by R.E. Peter of the Uni-
3.1. Construction of recombinant BMNPV A total of ¢ve transfer vectors were constructed to produce the recombinant BMNPV for the expression of the grass carp GH. All of them were designed to insert the GH cDNA downstream from the polyhedrin promoter in the viral genome by recombination. The major di¡erences between these constructs are the length and base composition of the 5P untranslated ends. Details of the 5P £anking sequences of each of the transfer vectors are summarized in Fig. 1. With the ¢ve transfer vectors, we were able to generate ¢ve di¡erent recombinant viruses. By dot blot and Southern hybridization, each one of them
BBAGEN 24641 14-8-98
334
W.K.K. Ho et al. / Biochimica et Biophysica Acta 1381 (1998) 331^339
Fig. 1. Features of the 5P untranslated regions of the transfer vectors used in this study. For the pAc373 derived vectors, the cloning site was BamHI and for the pBM030 derived ones, the cloning site was between EcoRI and XbaI. Bases marked with an asterisk indicate the position of a stem loop structure. Spacing between the polyhedrin promoter and the ATG start codon is as indicated. Codon usage by the di¡erent vectors is as indicated.
was shown to carry the grass carp GH cDNA (results not shown). When B. mori cells were infected with these viruses for 4 days, only strains P6 and EX were able to express GH in quantities high enough to be detected by Western blotting. Strain EE showed a much lower level of expression while strains SD78 and SD61 did not appear to show any expression at all. The concentration of hormone secreted into the medium over a 4 day period by the di¡erent recombinant viruses as determined by radioimmunoassay is summarized in the right column of Fig. 1. 3.2. Expression of GH in silkworm larvae Silkworm larvae at the 5th instar stage were infected with P6 and EX virus at a dose of approx.
2U105 particles per worm in a volume of 5 Wl. Worms were collected and bled on di¡erent days to ascertain the progress of infection and the expression of GH. As indicated in Fig. 2, between days 0 and 3, the level of GH in the hemolymph was undetectable. On day 4, the concentration of GH in both P6 and EX infected larvae increased rapidly from none to over 700 Wg per larvae. By day 5, all the infected silkworms became sick and began to die. The level of GH in the hemolymph also began to decrease, presumably due to proteolytic degradation by enzymes released from the dying cells. An SDSPAGE/Western blot analysis on P6 infected silkworm hemolymph samples is shown in the insert of Fig. 2. An immunoreactive band with a size of approx. 21^25 kDa was clearly noticed in samples col-
BBAGEN 24641 14-8-98
W.K.K. Ho et al. / Biochimica et Biophysica Acta 1381 (1998) 331^339
335
Fig. 2. Time course in the expression of GH in silkworm larvae infected with P6 and EX baculovirus. Inset is an SDS-PAGE and Western blot analysis of hemolymph samples collected on di¡erent days after P6 inoculation. Lanes: 1, bovine serum albumin; 2, molecular weight markers; 3, uninfected larvae; 4, 24 h sample; 5, 48 h sample; 6, 72 h sample; 7, 96 h sample; 8, 120 h sample. The immunoreactive band associated with the 30 000 molecular weight marker (carbonic anhydrase) was presumably due to cross-reactivity with the anti-grass carp GH monoclonal antibody.
lected on days 4 and 5. In contrast, in the Coomassie blue stained gel, only a faint band was noticed between the 20.1 and 30 kDa markers, indicating that the amount of recombinant GH in these samples was relatively low compared with that of other proteins in the hemolymph. Nonetheless, when GH in the hemolymph was puri¢ed by immunoa¤nity chromatography, its molecular mass (21 kDa) was similar to that of the native hormone (Fig. 3). 3.3. Characterization of GH expressed by P6
Fig. 3. SDS-PAGE analysis of recombinant GH puri¢ed by immunoa¤nity chromatography. Lanes: 1, molecular weight markers; 2, semi-puri¢ed GH from P6 infected silkworm; 3, a semi-puri¢ed GH from grass carp pituitary glands. The gel was stained with silver reagent and overloaded to visualize contaminating components.
Since the expression e¤ciencies of GH by P6 and EX are similar, we decided to select only one of them for further characterization. In both radioimmunoassay and radioreceptor assay, crude preparations of P6 infected silkworms were able to displace 125 I labeled GH in a dose dependent manner (Fig. 4). The binding displacement curves generated by the
BBAGEN 24641 14-8-98
336
W.K.K. Ho et al. / Biochimica et Biophysica Acta 1381 (1998) 331^339
Fig. 4. Radioimmunoassay and radioreceptor assay of GH from P6 infected silkworm larvae. The weight of the silkworm powder and hemolymph was the dried weight of the lyophilized material. Puri¢ed native hormone was used as the standard.
BMNPV expressed GH were essentially parallel to those generated by the native hormone. When uninfected preparations were used in either the radioreceptor or radioimmunoassay, no apparent displacement of radioactively labeled GH was observed, even at concentrations 10^100-fold higher. To further assess the similarity of the P6 expressed GH with the native one, we tested the ability of the P6 infected hemolymph to block the immunospeci¢c staining of somatotrophs in the grass carp pituitary gland. The antiserum used in this study was raised against natural grass carp GH and as indicated in Fig. 5A, it speci¢cally stained the intermediate region of the pituitary gland. Pretreatment with either GH or hemolymph collected from P6 infected larvae
abolished this staining (Fig. 5B,D). In contrast, hemolymph from silkworm infected with P6 on day 1 was not able to block the staining by the GH-speci¢c antiserum (Fig. 5C). This is to be expected, as we have previously shown that no GH could be detected at this time. We have made a number of attempts using conventional chromatography to purify the recombinant grass carp GH from silkworm. Unfortunately, these methods proved to be rather ine¤cient and we could not get the hormone puri¢ed from contaminants in the hemolymph. In another study, we raised a highly speci¢c monoclonal antibody against the grass carp GH. Using this antibody as an immunoa¤nity ligand, we were able to purify the P6 produced hormone to approx. 90% purity (Fig. 3). After further puri¢cation by SDS-PAGE, the hormone was taken for N-terminal sequencing. The sequence obtained was Ser-Glu-Asn-Gln-Arg-Leu-Phe-Asn-Asn-Ala-; it was exactly identical to the N-terminal sequence of the matured native hormone [13]. 3.4. Induction of growth in ¢sh by feeding with P6 infected silkworm To evaluate the biological activity of the GH produced in the P6 infected silkworm, a feeding experiment was performed. Juvenile grass carp with an average weight and length of 25.9 g and 12.6 cm, respectively, were randomly divided into two groups of 15 ¢sh each. The ¢rst group was fed with a diet containing 20 mg uninfected silkworm powder per gram diet and the second group with a diet containing 20 mg P6 infected silkworm powder per gram diet. The two groups of ¢sh were housed in separate aquariums and fed twice a day for a total of 6 weeks. The weight and length of the ¢sh were measured weekly. At the end of the period, the RSGR and RLGR were calculated, and the results are presented in Table 1. Compared with the control, ¢sh fed with
Table 1 E¡ect of feeding P6 infected silkworm powder on growth of juvenile grass carps Treatment
n
RSGR (%) (6 weeks)
RLGR (%) (6 weeks)
Uninfected silkworm (20 mg/g diet) P6 infected silkworm (20 mg/g diet)a
15 15
67.26 þ 4.41* 86.20 þ 11.7*
18.47 þ 1.76 20.08 þ 3.06
*Values are signi¢cantly di¡erent (P 6 0.001) from each other by the t-test. This amount of silkworm powder roughly contained 20 Wg GH as estimated by radioimmunoassay.
a
BBAGEN 24641 14-8-98
W.K.K. Ho et al. / Biochimica et Biophysica Acta 1381 (1998) 331^339
337
Fig. 5. Cross-section of grass carp pituitary gland stained with an anti-grass carp GH serum. (A) No blocking; (B) blocked with 0.5 mg/ml grass carp GH; (C) blocked with P6 infected silkworm hemolymph collected 24 h after inoculation; (D) same as C except hemolymph was collected 120 h after inoculation. The hemolymph was diluted 10U with bu¡ered saline before immunostaining.
the diet containing the P6 infected silkworm had a 28% and a 8.7% higher RSGR and RLGR, respectively. The well-being of the ¢sh fed with the infected silkworm appeared to be similar to that of the control group; there was no sign of viral infection. In a separate study, the process of freeze-drying was determined to be e¡ective in killing the recombinant BMNPV in the silkworm larvae. 4. Discussion GH from a number of species of ¢sh has been successfully expressed in E. coli by di¡erent research groups [4^10]. A major interest in recombinant ¢sh GH is to utilize it to enhance growth in ¢sh. To achieve this goal, an e¤cient and economical method in producing GH must be established. E. coli is one of the most e¤cient micro-organisms for producing recombinant proteins. In the case of ¢sh GH, an expression level up to 20^30% total cell protein has been reported [6^10]. Unfortunately, in all of these cases, the hormones were expressed in inclusion body form and they had to be solubilized and renatured before they could become biologically active. These
extra steps can signi¢cantly reduce yield and increase the production cost of the hormone. In the present study, we have attempted to overcome the problems associated with E. coli expression using a baculovirus. This system, although less e¤cient, can, however, produce a properly folded soluble GH with biological activity. Since our primary goal was to utilize the recombinant hormone for feeding ¢sh, we reasoned that as long as the hormone can be produced in a non-toxic biologically active form, it could be used directly without any puri¢cation. The use of silkworm larvae as the host for the recombinant BMNPV was a logical choice because (1) up to milligram quantities of recombinant protein can be produced in one larva, (2) the larva is non-toxic and can be used as food for the ¢sh, and (3) BMNPV has a very narrow host range, and as a result, it should be relatively safe to use. We have prepared a total of ¢ve recombinant BMNPV to express the grass carp GH in this study. Only two of them, P6 and EX, were able to express the hormone in high quantity. The major di¡erence in these constructs was the spacing between the polyhedrin promoter and the ATG translation initiation site. For P6 this spacing was 73 bp and for EX it was
BBAGEN 24641 14-8-98
338
W.K.K. Ho et al. / Biochimica et Biophysica Acta 1381 (1998) 331^339
28 bp. Compared with the other three low-expressing recombinants, viz., EE, SD61 and SD78, these spacings were similar in length. Thus, the higher level of expression by P6 and EX cannot be merely accounted for by a di¡erence in spacing between the polyhedrin promoter and the ATG translation initiation site. The present results are in contrast to the ¢ndings of Luckow and Summers [23] which showed that a shorter spacing is more e¤cient for expression. Another factor that may a¡ect the e¤ciency of mRNA translation is the presence of a stem loop structure in the leading sequence of the translation initiation site [24,25]. When the leading sequences of the ¢ve recombinants were analyzed, EX and EE did not have any stem loop structure while P6, SD61 and SD78 had one (see Table 1). Since the spacings in EX, SD61 and SD78 were relatively short, ranging from 15 to 28 bp, it is possible that the presence of a stem loop may have had a detrimental e¡ect on expression. On the other hand, P6 has one stem loop in the leading sequence but its position was quite far away from the start codon; hence it might have less negative in£uence on message translation. In the case of EE, there was no stem loop and yet it could not express GH e¤ciently. In this respect, the untranslated region's sequence or composition is also known to play a critical role in expression [25]. The grass carp GH expressed by P6 has an expected size and N-terminal sequence similar to the native hormone. This indicates that silkworm cells could recognize the signal peptide of the ¢sh and processed it correctly. In a similar study on human GH, Kadono-Okuda et al. [26] also observed the correct processing of the human signal peptide. The level of GH expressed in the present study was around 1 mg/ml hemolymph. This level of expression is manyfold higher than that achievable in cell culture. Since silkworms can be raised at much lower cost than cell culture, this approach is probably one of the most economical ways to make a soluble recombinant protein. In spite of the aforementioned merits, the silkworm expression system has two disadvantages. First, the recombinant protein may be more di¤cult to purify. Second, the process may be di¤cult to validate because parameters in culturing silkworm larvae may be subject to more variables
than in in vitro methods. As far as our application in using the recombinant hormone as food supplement is concerned, these two problems do not appear to be critical as we do not need a puri¢ed product for our application. A key consideration in using BMNPV infected silkworm as food is the potential environmental concerns it may raise. We have performed some preliminary experiments on the viability of the virus after freeze-drying the infected silkworms. Using cultured B. mori cells, we were unable to detect any sign of viral activity from the silkworm powder. Presumably, the process of drying may be enough to destroy the virus. Further tests are presently being carried out to ascertain the e¡ect of P6 in infecting other species, including other ¢sh species, during its application as food. One of the major advantages of using baculovirus to express eukaryotic genes is its ability to produce a recombinant protein closely resembling the native one. The product is normally expressed in soluble form and in higher yield. Although insect cells can perform glycosylation, the carbohydrate chain added may not be identical to that made by mammalian cells [27]. In this study, we have examined the integrity of the grass carp GH expressed in silkworm by four criteria, i.e., molecular weight, immunoreactivity, receptor binding and N-terminal sequence. Compared with the native hormone, the recombinant GH was shown to have similar properties. Unfortunately, due to the lack of a good and consistent biological assay for ¢sh GH we were not able to compare the speci¢c activity of the recombinant hormone with the native one. Nonetheless, the result of our feeding experiment did provide partial evidence to indicate that the GH produced in the silkworm was biologically active. Acknowledgements This research was supported by grants from the Croucher Foundation, Industrial Support Fund and the Research Grants Council of Hong Kong. We are most grateful to Prof. S. Maeda and Prof. R.E. Peter for their gifts of research materials. We also would like to thank Dr. T.A. Marchant for performing the radioreceptor assays.
BBAGEN 24641 14-8-98
W.K.K. Ho et al. / Biochimica et Biophysica Acta 1381 (1998) 331^339
References [1] E. McLean, E.M. Donaldson, H.M. Dye, L.M. Souza, Aquaculture 91 (1990) 197^203. [2] S. Moriyama, H. Yamamoto, S. Sugimoto, A. Toshio, T. Hirano, H. Kawauchi, Aquaculture 112 (1993) 99^106. [3] H.J. Tsai, J.C. Kuo, S.W. Lou, T.T. Kuo, Prog. Fish Culturist 56 (1994) 7^12. [4] S. Sekine, T. Mizukami, T. Nishi, Y. Kuwana, A. Saito, M. Sata, S. Itoh, H. Kawauchi, Proc. Natl. Acad. Sci. USA 82 (1985) 4306^4310. [5] L.B. Agellon, T.T. Chen, DNA 5 (1986) 463^471. [6] A. Saito, S. Sekine, Y. Komatsu, M. Sato, T. Hirano, S. Itoh, Gene 73 (1988) 545^551. [7] F. Rentier-Deirue, D. Swennen, J.C. Philippart, C. L'Hoir, M. Lion, O. Benrubi, J.A. Martial, DNA 8 (1989) 271^278. [8] N. Sato, M. Sakaguchi, M. Nonaka, K. Murato, K. Watanabe, T. Hayamie, Y. Kariya, A. Kimura, S. Kimura, Appl. Microbiol. Biotechnol. 30 (1989) 153^159. [9] H.J. Tsai, K.L. Lin, T.T. Chen, Comp. Biochem. Physiol. 104B (1993) 803^810. [10] M. Fine, E. Sakal, D. Vashdi, V. Daniel, A. Levanon, O. Lipshitz, A. Gertler, Gen. Comp. Endocrinol. 89 (1993) 51^ 61. [11] H.J. Tsai, C.F. Tseng, T.T. Kuo, Bull. Inst. Zool. Acad. Sin. 32 (1993) 162^170. [12] H.J. Tsai, H.M. Chen, C.F. Lo, Can. J. Fish. Aquatic Sci. 51 (1994) 1^7. [13] W.K.K. Ho, W.H. Tsang, N.P. Dias, Biochem. Biophys. Res. Commun. 161 (1989) 1239^1243.
339
[14] S. Maeda, in: J. Mitsuhashi (Ed.), Invertebrate Cell System Applications, Vol. I, CRC Press, Boca Raton, FL, 1989, pp. 167^181. [15] V.A. Luckow, M.D. Summers, Biotechnology 6 (1988) 47^ 55. [16] F. Sanger, S. Nicklen, A.R. Coulson, Proc. Natl. Acad. Sci. USA 74 (1977) 5463^5467. [17] M.D. Summers, G.E. Smith, A Manual of Methods for Baculovirus Vectors and Insect Cell Culture Procedures, Texas Agricultural Experiment Station Bulletin No. 1555, 1987. [18] U.K. Laemmli, Nature 227 (1970) 680^685. [19] T.A. Marchant, R.A. Fraser, P.C. Andrews, R.E. Peter, Regul. Pept. 17 (1987) 41^52. [20] P.R.P. Salacinski, C. McLean, J.E.C. Sykes, V. ClementJones, P.J. Lowry, Anal. Biochem. 172 (1981) 136^146. [21] J.N. Fryer, Gen. Comp. Endocrinol. 39 (1979) 123^130. [22] W. Ge, R.E. Peter, Gen. Comp. Endocrinol. 95 (1994) 213^ 221. [23] V.A. Luckow, M.D. Summers, Virology 170 (1989) 31^39. [24] E. Jay, A.K. Seth, J. Romments, A. Sood, G. Jay, Nucleic Acid Res. 10 (1982) 6319^6329. [25] H. Dalboge, S. Carlsen, E.B. Jensen, T. Christensen, H.M. Dahl, DNA 7 (1988) 399^405. [26] K. Kadono-Okuda, M. Ymamoto, Y. Higashino, K. Taniai, Y. Kato, S. Chowdhury, J.H. Xu, S.K. Choi, M. Sugiyama, K. Nakashima, S. Maeda, M. Ymakawa, Biochem. Biophys. Res. Commun. 213 (1995) 389^396. [27] D.M. Wojchowski, S.H. Orkin, A.J. Sytkowski, Biochim. Biophys. Acta 910 (1987) 224^232.
BBAGEN 24641 14-8-98