Organization and Expression of the Immunoresponsive Lysozyme Gene in the Giant Silk Moth, Hyalophora cecropia*

Lysozyme is one of the antibacterial proteins that are produced by the giant silk moth Hyalophora cecropia in response to bacterial infection or injury. As an essential step toward the understanding of the mechanisms involved in the immune response, we have isolated and characterized the lysozyme gene from Cecropia. The complete nucleotide sequence of the gene as well as the immediate flanking sequences have been determined. The gene includes three exons. Its first intron contains a repetitive sequence. In the evolution- ary aspect, the Cecropia lysozyme gene and two vertebrate lysozyme genes have been found to maintain a similar organization pattern of exons. The lysozyme gene has been found to be strongly induced by lipopoly- saccharides and a phorbol ester as well as bacteria. In the induction by bacteria, the lysozyme transcript ap- pears at about 2 h, reaches to the maximum level at about 24 h, and then declines. Comparison of the 5’- flanking sequences with several other genes involved in the immune response of H. cecropia and Drosophila melanogaster revealed a &-like consensus sequence. This sequence is specifically recognized by a nuclear protein from the induced pupa.

publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "aduertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
The nucleotide seguence(s) reported in thispaper has been submitted to the GenBankTM/EMBL Data Bank with accession numbeds) 505 754. studied insect lysozyme is the one from the giant silk moth Hyalophora cecropia (Engstrom et al., 1985).
Lysozyme, together with cecropins and attacins, are the three main types of antibacterial proteins induced in the diapausing pupae of Cecropia by bacterial infection or injury.
They constitute the major antibacterial components of the humoral immune system in Cecropia (Boman and Hultmark, 1987). As lysozyme is bactericidal only to a few Gram-positive bacteria that are also sensitive to cecropins, it is believed that the main function of lysozyme is to remove the murein sacculus left over after the action of cecropins and attacins (Hultmark et al., 1980;Boman and Hultmark, 1987). The Cecropia lysozyme has been isolated (Hultmark et al., 1980), and the protein sequence as well as the nucleotide sequence of its cDNA have been determined (Engstrom et al., 1985). It shows great sequence similarity to the chicken-type lysozymes.
We are interested in the regulation of the expression of the immune genes in Cecropia. As an essential step toward this goal, we have analyzed the structures of a number of immune genes from Cecropia. Here, we report the isolation, sequence characterization, and expression of the Cecropia lysozyme gene. A tentative promoter element will be discussed.

MATERIALS AND METHODS
Insects and Injection-Diapausing pupae of Cecropia were either purchased from American dealers or reared on a synthetic diet (Hultmark et al., 1980) and stored at 8 "C. The pupae were injected with either Enterobacter cloacae 612 (Hultmark et al., 1980), 100 pg of lipopolysaccharide (LPS)' from Escherichia coli D21 in Lepidopteran Ringer or 300 ng of phorbol12-myristate 13-acetate (PMA) (dissolved in dimethyl sulfoxide and diluted with Lepidoteran Ringer just before use) in a volume of 50 p1. As a control, 50 p1 of Lepidopteran Ringer containing the same amount of dimethyl sulfoxide as the PMA solution was injected.
Isolation of Cecropia Lysozyme Genomic Clones-A Cecropia genomic bank in the XEMBL 3 vector was constructed from a single pupa.' It was screened with the Cecropia lysozyme cDNA, pCP701 (Engstrom et al., 1985), 3ZP-labeled to high specific activity by random priming using an oligolabeling kit (Pharmacia LKB Biotechnology Inc.). Membrane transfer and hybridization, using Du Pont-New England Nuclear's NEF filters, were carried out according to the manufacturer's instructions (Du Pont). Phage DNA was isolated essentially according to Garber and Kuroiwa (1983), using E. coli Q359 as host bacteria. Restriction mapping and Southern blot hybridizations (Southern, 1975) were used to analyze the cloned DNA.
Cecropia DNA Preparation-The tissue of Cecropia pupae, mainly containing the fat body, was prepared as described (Xanthopoulos et al., 1988) and ground in a liquid nitrogen-containing mortar. The preparation of nuclei and subsequent proteinase K digestion were carried out essentially according to Fritton et al. (1983). After phenol extraction, the aqueous phase was treated with RNase A (100 pg/ml) and RNase T1 (100 units/ml) and then extracted again with phenol. The DNA was precipitated from the aqueous phase by the addition of 2 volumes of ice-cold ethanol.
Southern Blot Analysis of Cecropia Genomic DNA-Cecropia chromosomal DNA was digested with restriction enzymes of choice a t 37 "C overnight. The restricted DNA samples (-8 pg) were electrophoresed in a 0.7% agarose gel and blotted overnight onto a Hybond-N membrane (Amersham Corp.). The membrane was hybridized with the "'P-labeled probes according to McGinnis et al. (1984).
DNA Sequence Analysis-The EcoRI fragment from the recombinant phage XCP701, containing the lysozyme gene, was cloned into pTZ19R. The subsequent subcloning was carried out either by the "shot gun" method using the multifunctional vectors pTZ19R or pTZ18R or by the deletion method (double-stranded nested deletion kit, Pharmacia). The dideoxy chain termination sequencing (Sanger et al., 1977) was carried out using a Sequenase kit (U. S. Biochemical Corp.). Sequence analysis was performed with MacGene Plus software (Applied Genetic Technology, Inc.).
RNA Analysis-Pupal tissue was prepared in the same way as for DNA isolation. The RNA was isolated according to Klemenz et al. (1985). The RNA samples (10 pg) were electrophoresed in a formaldehyde-denaturing 1.5% agarose gel (Maniatis et al., 1982), blotted onto a Hybond-N membrane, and hybridized with '*P-labeled lysozyme cDNA according to the manufacturer's instructions. RNase protection was carried out essentially according to Gilman et al. (1987). For the preparation of the RNA probe, a DNA fragment covering 74 bases of the exon I and 492 bases of the 5'-flanking sequence was cloned into pTZ19R. The recombinant plasmid was linearized with EcoRI and used as template, and the radioactive antisense RNA was synthesized using T7 RNA polymerase (Stratagene). Total cellular RNA was hybridized with the probe ( 5 X lo5 cpm) at 45 "C overnight. After digesting with RNase T1 and RNase A at 30 "C for 60 min, the protected fragment was separated in a 5% sequencing gel. Sequencing reactions of known DNA were used as size markers.
Preparation of Nuclear Extracts-Nuclear extracts were prepared from the total tissues of the Cecropia pupa according to Gorski et al. (1986) with the exception that protease inhibitors (leupeptin, 0.7 pg/ ml; pepstatin A, 0.7 pg/ml; and benzamidine, 0.1 mM) were added to the homogenization buffer.
DNA Mobility Shift Assay-For the binding assay, DNA fragments were created by restriction cleavages from two deletion subclones covering the upstream region of the lysozyme gene and purified by electroelution (International Biotechnologies, Inc., New Haven, CT). The DNA fragments were labeled with T4 polynucleotide kinase with [-y-"'P]ATP. Binding reactions were carried out in 15 pl of a buffer containing 20 mM Hepes, pH 7.5, 70 mM NaCI, 1 mM EDTA, 1 mM dithiothreitol, 9% glycerol, 2 pg of poly(dl-dC) .poly(dl-dC), 100,000 cpm (1 ng) of end-labeled DNA and 10 pg of nuclear extract at room temperature for 20 min. For competition experiments, competitor DNA were added to the mixture prior to addition of the nuclear extract. The binding mixture was electrophoresed through a 5% native polyacrylamide gel. The gel was pre-run at 11 V/cm for 60 min at room temperature in 22.5 mM Tris borate, pH 8.0, 22.5 mM boric acid, 0.5 mM EDTA. Electrophoresis was carried out under the same conditions for 90 min. After fixation for 10 min in 5% glycerol, the gel was dried and exposed to x-ray film at -70 "C.

Isolation and Characterization of Lysozyme Genomic
Clones-The Cecropia genomic bank was screened using the lysozyme cDNA as probe. Eight positive clones were selected and purified. They were named XCP701-XCP708. Restriction mapping (Fig. 1B) and Southern blot analyses (data not shown) illustrated that they comprised two classes of overlapping clones. One group covered the complete lysozyme gene, 3.8 kb of upstream, and 7 kb of downstream sequences. The other group contained part of the gene and 8.5 kb of downstream sequences. These two groups of clones covered about 15 kb of the chromosome (Fig. lA). To determine the number of lysozyme genes, genomic Southern blot analyses were carried out with Cecropia chromosomal DNA. XCP701 and genomic DNA from three individual pupae were digested with the restriction enzyme EcoRI and probed with "P-labeled Cecropia lysozyme cDNA. Fig. 2A shows two independent digestions of DNA from the three individual pupae (A-C and A'-C'). All three pupae show a main hybridizing EcoRI fragment in the same size range as the hybridizingEcoR1 fragment from the genomic clone XCP701 (lane G ) . The weak band seen in the DNA digests from pupae C is most likely due to partial digestions since its strength, as compared with the main band, is different in the two digests. We then conclude that the Cecropia genome contain a single lysozyme gene. The small size discrepancies of the EcoRI fragment from the individual pupae suggested a restriction site polymorphism. This was confirmed by additional genomic Southern blot analysis (data not shown), and the size difference was localized to the XhoI-EcoRI fragment that contains the 3'flanking region (Fig. lA).
Sequence Characteristics of the Lysozyme Gene-Since XCP701 contains the complete lysozyme gene, this clone was selected for further analysis. As the first step of subcloning, the larger EcoRI fragment from clone XCP701 was inserted . Further subcloning and sequencing were carried out as described under "Materials and Methods." The sequencing strategy is shown in Fig. 1C. In most cases singlestranded templates were used, and the sequencing was done on both strands. When only one strand was sequenced, the sequencing experiment was repeated at least twice or until no uncertainties were left. The sequence of the Cecropia lysozyme gene, together with the 5"flanking region, is shown in Fig. 3. The whole transcriptional unit of the gene (from cap site to the second polyadenylation site) is 2875 bp long. The gene contains three exons interrupted by two introns with the boundary sequences of 5' ACTgtaagt-cagGG 3' (intron 1) and 5' ATCgtaagt-tagAG 3' (intron 2), which are in agreement with the consensus intronlexon border sequences of eukaryotic genes (Breathnach and Chambon, 1981). The first intron is much larger than the second one, 1644 as compared with 668 bp. A further analysis of the first intron is discussed below.
An insect specific cap site with the consensus sequence ATCA TPy (Snyder et al., 1982;Hultmark et al., 1986) has earlier been found in the cecropin B gene from Cecropia (Xanthopoulos et al., 1988). In the Cecropia lysozyme gene, we found a similar heptamer sequence, ATCATAC, 49 bases prior to the ATG codon for the initiation of translation. About 30 bases preceding the cap site is an AT-rich region containing the sequence (T)(A)TATAAA, identical to the consensus sequence of the TATA box (Breathnach and Chambon, 1981).
No sequences showing homology to the so-called CAAT box, frequently found in many eukaryotic promoters (Efstratiadis et al., 1980), were identified. We found, however, another plausible promoter element, GGAGGATTCCCC, at position -99 by comparing the 5"flanking region with the attacin genes (Sun et al., 1991). The homologous area in the attacin genes have been found to share a striking sequence homology with the immunoglobulin K chain enhancer, KB (Sen and Baltimore, 1986). The fact that this site occurs also in the lysozyme gene makes it even more likely as a common promoter element for the simultaneously induced immune genes in Cecropia. Two polyadenylation signals, AATAAA (Proudfoot and Brownlee, 1980), are found in the 3' end of the gene. T G The Cecropia Lysozyme Gene Contains a Repetitive Sequence in its First Intron-Repetitive sequences have been found in some vertebrate lysozyme genes (Li et al., 1988;Baldacci et al., 1981) and lysozyme-related genes (Hall et al., 1987). A transposable element has also been identified from the intron of the cecropin A gene in Ce~ropia.~ We investigated the possibility of a transposable element in the Cecropia lysozyme gene. The DNA filters used for the genomic Southern blots were re-probed with various subclones of XCP701. One of the subclones covering 1056 bases of the first intron and 74 bases of the second exon was shown to give a repetitive pattern. Deletion of the exon part and another 190 bases of the intron sequence from this subclone slightly changed its hybridization pattern but did not destroy its repetitive property (Fig. ZB), whereas further deletion of about 380 bases from the same direction totally destroyed its ability to generate a repetitive hybridization pattern. These results indicate that the repetitive sequence is localized in a region of 570 bases at the 3' end of the intron (Fig. 3, sequence in brackets).
Determination of the Transcriptional Znitiation Site in Znjured and Immunized Pupae-As mentioned above, the Cecropia lysozyme gene at its 5"flanking region has a heptamer sequence homologous to the cap site consensus sequence of the insect genes. To see if this sequence is used as an initiation site in the transcription of the lysozyme gene, we carried out a RNase protection assay. A 566-bp DNA fragment was transcribed in vitro to generate a radioactive RNA probe which would protect a 122-bp nucleotide fragment counting from the heptamer sequence. RNA from immune (16 h after bacterial injection), injured (16 h after Lepidopteran Ringer injection), or noninjected pupae were used. As shown in Fig.  4, two bands (marked by arrows) resulted from the protection for the immune RNA (lane 2) and the injured RNA (lane 3). The major band, with the size of about 122 bases, is in good agreement with the expected size, indicating the presence of a lysozyme transcript initiated from the heptamer. The RNase protection also showed that the injury response was much weaker than the immune response and that the same sites for the initiation of transcription were used.
The The induction kinetics of lysozyme gene expression was carried out using Cecropia diapausing pupae. The pupae were injected with E. cloacae Dl2 and used for RNA preparation a t different time intervals after injection. The induction of the lysozyme RNA was detected by both Northern blot (data not shown) and RNase protection analyses (Fig. 5A). Immediately after injection (0 h) of bacteria, no detectable lysozyme transcript was found. However, 2 h after injection a small amount of lysozyme RNA appeared. Like in the first RNase protection assay (Fig. 4), we could see two transcripts. They showed different kinetics in the bacterial induction. The larger transcript reached the maximum level at about 24 h and then declined. The smaller one reached the maximum level as early as 4 h and declined to a very low level a t 16 h (see also Fig. 4). This band declined and almost disappeared 24 h after injection. From this result we conclude that there are two alternative initiation sites for transcription (Fig. 6).
The Cecropia Lysozyme mRNA Can Re Induced by LPS and a Phorbol Ester (PMAI-It has been reported that bacterial LPS and phorbol esters activate a number of mammalian genes containing a KB-like promoter element via the activation of a transcription factor, NF-KB (Lenardo and Baltimore, 1989). Since the Cecropia lysozyme gene also contains a KBlike sequence element at the promoter region, we investigated the effect of LPS and PMA on the expression of this gene.
The Cecropia pupae were injected with LPS, PMA, or Lepidopteran Ringer as described under "Materials and Methods." The induction of lysozyme mRNA was followed by RNase protection assay. As shown in Fig. 5B, a  PMA or LPS, the levels of lysozyme mRNA were clearly enhanced.
A Nuclear Protein Specifically Binds to the &-like Sequence of the Lysozyme Gene-We investigated the possible function of the &like element in the regulation of the lysozyme gene using the DNA mobility shift assay. By RsaI restriction cleavages and exonuclease I11 deletions, we created two probes, P 2 0 0 and PIS,, from the promoter region of the lysozyme gene. P200 contains the complete &-like sequence, whereas PI83 has the major part of this site deleted (Fig. 6A). In the binding assay, nuclear extracts from both normal and bacteria-induced pupae were used. As shown in Fig. 6B, P200 could form a complex with the nuclear extract from bacteria-induced pupa (lane 3) but not with that from the normal pupa (lane 2). Furthermore, this binding could be abolished by using 30fold of the unlabeled Pz, (lane 4 ) . However, PI83 with the KBlike sequence deleted was ineffective as a competitor even a t a 50-fold molar excess (lane 5 ) . The inability of PIa3 for this binding was further confirmed by using the end-labeled PI83 directly in the binding assay (lane 6 ) .

DISCUSSION
Amino acid heterogeneities in the Cecropia lysozyme were earlier found and explained as caused by allelic variants of the gene (Engstrom et al., 1985). Sequence comparison of the Cecropia lysozyme cDNA (Engstrom et al., 1985) with the Cecropia lysozyme gene reveals 2 base pair substitutions in the internal coding sequence. Both substitutions cause amino acid changes, Arg-15 (number concerning the mature protein) to  Ser-66, that correspond to the discrepancies found previously.
The cDNA study demonstrated the similarity in the primary structure of the Cecropia and the chicken lysozymes (Engstrom et al., 1985). It is now possible to make a comparison of all the functional domains in the Cecropia lysozyme with those in two vertebrate lysozymes. In Fig. 7, we have aligned the full amino acid sequences deduced from the lysozyme genes of Cecropia, chicken (Jung et al., 1980), and mouse (Cross et al., 1988). The following features are revealed 1) the residues around the splicing sites (exon boundaries) are highly conserved. 2) The second and the third exons, which are known to code for the catalytic center and the specificity site of the enzyme (Jung et al., 1980), are considerably conserved, indicating a selection pressure during evolution. 3) The signal sequence of the Cecropia lysozyme shows a high degree of similarity but low degree of identity to those of the vertebrate lysozymes. The key features shared by these three signal peptides are their length and hydrophobicity. Furthermore, they all have a charged amino acid in the N-terminal. 4) Even though there are a few gaps when the sequences are best aligned, the positions of the two introns in the Cecropia lysozyme gene are identical to those of the first two introns in the vertebrate lysozyme genes. The splicings occur within the codons for Try-28 and Ala-83 of the mouse lysozyme M, corresponding to Try-25 and Gln-78 for the Cecropia lysozyme. The third intron of the vertebrate lysozyme genes does not exist in the Cecropia lysozyme gene, but the amino acid (Trp) at that splicing site is conserved. Despite the divergence of about 600 million years (Boman and Hultmark, 1987), the Cecropia lysozyme gene and the two vertebrate lysozyme genes maintain a conserved exon pattern. This observation supports the hypothesis that individual exons tend to maintain a degree of autonomy through evolution (Gilbert, 1978).
For the study of lysozyme gene expression, RNase protection assay was used. This technique has the advantage of being able to resolve RNA transcripts with only 1 base pair mismatch. This enabled us to identify two RNA transcripts with different induction patterns. As shown in Fig. 5A, apart from the major band, a smaller band was protected at the early stage (4 and 6 h) of induction. This small band declined at 16 h (Fig. 4, lane 2) and disappeared at 24 h after bacterial injection. Several lines of evidence indicate that the two bands result from two alternative sites of transcriptional initiation. First, apart from the heptanucleotide indicated in Fig. 3, the lysozyme gene contains another cap site-homologous sequence, ATCGTCG, 17 bp downstream of the first one. The predicted sizes of the transcripts initiated from these two tentative cap sites are in agreement with those of the two bands in the RNase protection. Second, the Cecropia genome contains only one lysozyme gene. That rules out the possibility of RNA transcripts from two different genes. Third, since the induction pattern of the smaller band (or transcript) is different from that of the larger one, the result cannot be explained by a mismatch in an allelic gene.
A decamer sequence, GGGGATTCCT, has been found previously in the upstream regions of both the acidic and the basic attacin genes (Sun et al., 1991). This sequence element appeared to share a striking homology to the consensus sequence of the mammalian nuclear factor KB binding site, GGGR TYYCC (Lenardo and Baltimore, 1989). Analyses of the published sequences for the cecropin genes in Cecropia and Drosophila (Xanthopoulos et al., 1988;Kylsten et al., 1990) also revealed this type of element (Sun et al., 1991). Since the lysozyme is induced simultaneously with the attacins and the cecropins, we looked for this element in the upstream region of the lysozyme gene. We found a palindromic sequence, GGAGGATTCCCC, which is also homologous to the NF-KB binding site. DNA mobility shift assay clearly showed that this &-like sequence could be specifically bound by a nuclear protein from the bacteria-induced Cecropia pupa. Deletion of the KB-like sequence from the probe abolished this binding. Nuclear extract from the uninduced pupa did not show any binding activity. This result is consistent with the in vivo expression of the gene.
The KB site was first found in the immunoglobulin K gene and later in a variety of other genes. These genes can be induced by LPS, phorbol esters, and a number of other inducers via the activation of the NF-KB, which is normally present in the cytoplasm as an inactive form (reviewed in Lenardo and Baltimore, 1989). It has been shown earlier that LPS strongly induces the immune system in Cecropia (Trenczek and Faye, 1988). We have now investigated the effect of LPS and a phorbol ester (PMA) on the expression of the lysozyme gene as well as the attacin genes that also contain the KB-like promoter element. We found that these genes were strongly induced by both LPS and PMA (Fig. 5B) (Sun et al., 1991). Taken together with the earlier findings that the immune genes in insects are co-induced (Boman and Hultmark, 1987;Kylsten et al., 1990), these data indicate that a common trans-acting factor is involved in the induction of these immune genes. This factor shares certain similarities with NF-KB probably in the way of activation and DNA binding specificity. Recently, NF-KB was found to have the identical DNA binding subunit as KBF1, a factor recognizing the KB-homologous sequence in major histocompatibility complex class 1 gene. Furthermore, these two factors were suggested to belong to the family of proteins that includes re1 oncoproteins and the Drosophila dorsal gene product (Kieran et al., 1990). The fact that the nuclear factor we found from Cecropia binds to a target sequence homologous to KB (and also to the target of KBF1) raises the question whether it A Y