bacteria-binding protein from the silkworm, Bombyx mori - Europe PMC

Proc. Natl. Acad. Sci. USA

Vol. 93, pp. 7888-7893, July 1996 Immunology

Purification and molecular cloning of an inducible Gram-negative bacteria-binding protein from the silkworm, Bombyx mori (lipopolysaccharide/CD14/13-1,3 glucanase/silkworm/pattern recognition)

WON-JAE LEE*t, JIING-DWAN LEEt, VLADIMIR V. KRAVCHENKOt, RICHARD J. ULEVITCHt, AND PAUL T. BREY*§ Laboratoire de Biochimie et Biologie Moleculaire des Insectes, Institut Pasteur, 25 rue du Dr. Roux, 75724, Paris Cedex 15, France; and tDepartment of Immunology, The Scripps Research Institute, 10666 North Torrey Pines Road, La Jolla, CA 92037

Communicated by Fotis C. Kafatos, European Molecular Biology Laboratory, Heidelberg, Germany, April 11, 1996 (received for review November 20, 1995)

we discovered a 50-kDa pratein that demonstrates strong affinity to the cell wall of Gram(-) bacteria. We have called this protein GNBP. The strong affinity of GNBP to the surface of Gram(-) bacteria, -as well as its pattern of up-regulation following injury in the presence of bacteria, prompted us to establish its identity and to investigate its occurrence in healthy and bacteria-challenged silkworms.

A 50-kDa hemolymph protein, having strong ABSTRACT affinity to the cell wall of Gram(-) bacteria, was purified from the hemolymph of the silkworm, Bombyx mori. The cDNA encoding this Gram(-) bacteria-binding protein (GNBP) was isolated from an immunized silkworm fat body cDNA library and sequenced. Comparison of the deduced amino acid sequence with known sequences revealed that GNBP contained a region displaying significant homology to the putative catalytic region of a group of bacterial 18-1,3 glucanases and 13-1,3-1,4 glucanases. Silkworm GNBP was also shown to have amino acid sequence similarity to the vertebrate lipopolysaccharide receptor CD14 and was recognized specifically by a polyclonal anti-CD14 antibody. Northern blot analysis showed that GNBP was constitutively expressed in fat body, as well as in cuticular epithelial cells of naive silkworms. Intense transcription was, however, rapidly induced following a cuticular or hemoceolien bacterial challenge. An mRNA that hybridized with GNBP cDNA was also found in the 1(2)mbn immunocompetent Drosophila cell line. These observations suggest that GNBP is an inducible acute phase protein implicated in the immune response of the silkworm and perhaps other insects.

MATERIALS AND METHODS Organisms, Cells, Immunization, and Collection of Hemolymph. Silkworms, Bombyx mori, were reared under normal and axenic conditions to day 5 of the fifth instar. Silkworms were challenged by injecting 10 pk1 of late logarithmic-phase Enterobacter cloacae (Strain 57-9, Pasteur Institute Collection) or by procuticular abrasion (4). Hemolymph plasma was collected and stored as described (4). Gateffs 1(2)mbn immunocompetent cell line was used for Western and Northern blot analysis. These cells were maintained in Schneider's medium supplemented with 5% fetal calf serum at 220C (12). In Vitro Bacterial-Binding Assay. The in vitro bacterialbinding assay was carried out essentially according to Sun et al. (13). Briefly, 10 ml of late logarithmic-phase culture of E. cloacae or Bacillus licheniformis was centrifuged, the cells were washed and finally resuspended in 200 ,ul of 10 mM Tris HCl (pH 8.0). The bacterial suspension was added to 1 ml of immune hemolymph plasma and incubated at room temperature with mild agitation from 1-30 min. The mixture was centrifuged for 2 min at 10000 x g at 4°C and the supernatant was removed for a subsequent binding assay. The pellet was washed twice in 500 ,ul of 0.5 M NaCl solution and finally eluted with 200 ,ul of 0.5 M NaCl/0.1 M ammonium acetate (pH 4.0) or washed successively with 200 p.l of NaCl solution where the NaCl concentration was increased arithmetically after each wash (0.1, 0.2, 0.3, 0;4, and 0.5 M NaCl concentrations) for increased washing stringency. The proteins detached from the bacterial surface were separated by SDS/PAGE (14). In Situ Enzymatic Cleavage and Peptide Mapping for Generating the Internal Sequences. After electrophoresis, the gel was fixed by 50% methanol/10% acetic acid and stained with amido black for 2 min. The protein bands were cut and washed extensively in Milli Q water. The protein bands were cut into small pieces (1 mm x 1 mm) and dehydrated using a SpeedVac. The digestion of endopeptidase Lys-C (final concentra-

Insects, which inhabit almost all terrestrial and freshwater environments, must protect themselves from an array of commensal, pathogenic, and parasitic organisms that share the same ecological niches. Devoid of clonally selected lymphocytes, insects possess a rapidly responsive innate immune system, which is well adapted to their short life span (1). Insects are not directly exposed to the microbial environment because they are physically encased by a cuticular armor (2). This exoskeleton, which covers all external surfaces and epithelial invaginations of the insect, is the first line of defense between the insect and the environment (3). When this barrier is breached, microorganisms enter and an acute inflammatory response occurs resulting in the de novo synthesis of antibacterial effector molecules in the cuticular epidermal cells at the site of the integumental injury (4) and also in the fat body (4-6). Many inducible antibacterial molecules are found in insect hemolymph after integumental injury or injection of bacteria (7-9). Notwithstanding the ever increasing amount of information on inducible antibacterial effector molecules in insects, the soluble and membrane bound recognition molecules that lead to immune response signal transduction remain unidentified. On the contrary, in the mammalian immune system, two proteins have been identified for their involvement in the recognition of bacterial endotoxins, respectively lipopolysaccharide (LPS)-binding protein (10), and the membrane as well as soluble forms of CD14 (11). In our search for putative endotoxin-binding proteins in the hemolymph of silkworms,

Abbreviations: GNBP, Gram(-) bacteria-binding protein; MOPAC, mixed oligonucleotide-primed amplification of cDNA; LPS, lipopolysaccharide. Data deposition: The sequence reported in this paper has been deposited in the GenBank database (accession no. L38591). tPresent address: Laboratory of Immunology, Medical Research Center, College of Medicine, Yonsei University, CPO Box 8044, Seoul, Korea. §To whom reprint requests should be addressed.

The publication costs of this article were defrayed in part by page charge payment. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. §1734 solely to indicate this fact.

7888

Immunology: Lee et al. tion = 2 ,ug/ml) was carried out in 300 ,ul of 0.1 M Tris HCl (pH 8.8), containing 0.03% SDS at 30°C for 18 hr. After digestion, the whole reaction mixture was centrifuged and the supernatant was immediately injected onto the C-18 HPLC column (Vydac, 2.1 x 250 mm). The peptide fragments were separated on a linear gradient of acetonitrile (0-55%), containing 0.1% trifluoroacetic acid over 60 min. A flow rate of 200 ,ul/min was applied and absorbency was monitored at 214 and 280 nm. Each peak was collected manually and stored at -20°C for further analysis. Amnino Acid Sequencing Analysis. The immobilized intact protein on the poly(vinylidene difluoride) membrane or the enzymatically digested peptides separated by HPLC were subjected to automated Edman degradation with a 477A gas-phase microsequencer (Applied Biosystems), connected to an on-line model 120A phenylthiohydantoin analyzer. Mixed Oligonucleotide-Primed Amplification of cDNA (MOPAC) Analysis. Except when specifically mentioned, all DNA and RNA manipulations were carried out using standard techniques (15). Two pairs of degenerated oligonucleotide primers were synthesized, P1; 5'-AA(AIG)ATGAC(A/C/G/T) (C/T)T(A/C/G/T)TT(T/C)GC(A/C/G/T)TT-3'(sense) and its antisense P2, which included internal sequences complementary to all possible codons specifying amino acids Lys-MetThr-Leu-Phe-Ala-Phe and P3; 5'-GT(A/C/G/T)TA(T/

C)TA(T/C)AA(T/C)GC(A/C/G/T)GT(A/C/G/T)TT-3' (sense) and its antisense P4, which included all possible codons specifying internal peptide sequence Val-Tyr-Tyr-Asn-AlaVal-Phe. For the preparation of template RNA, one fifth instar (Day 5) silkworm was immunized with 10 ,ul of Escherichia coli 0127:B8 LPS (2 mg/ml). After 6 hr, the fat body was removed and total RNA was extracted (16). The single-stranded cDNA was synthesized from total RNA by using the antisense primers (P2 or P4) and murine leukemia virus reverse transcriptase (BRL). Single-stranded cDNA was amplified with 1 ,uM of each oligonucleotide in pairs (P1 and P4 or P2 and P3) in 10 mM Tris HCl (pH 8.0), 50 mM KCl, 1.5 mM MgCl2, 200 ,uM dNTP, and 2.5 units of Thermus aquaticus polymerase (Perkin-Elmer/Cetus) (17, 18). PCR was performed with a denaturing temperature of 94°C for 30 sec, annealing at 52°C for 1 min, and extension at 72°C for 2 min for 30 cycles using Tempcycler II, Model 110P (Coy Laboratory Products, Ann Arbor, MI). The aliquots of amplified mixtures were analyzed on 2% agarose gels, the ethidium bromide stained band was excised from the gel, and isolated DNA was cloned into pCR-Script SK(+) using a PCR cloning kit (Stratagene). Screening For Full-Length cDNA. A cDNA library made from fat body of challenged B. mori (10 hr after injection of heat-killed E. cloacae) was prepared using the Amersham AgtlO cDNA synthesis kit. The 160-bp PCR product from MOPAC analysis (combination of P1 and P4) was 32P-labeled by random priming (Mega Prime DNA Labeling System, Amersham) for screening the cDNA library. About 75,000 plaques were plated at a density of 2,500 plaques per 90-mm diameter plate and screened using Nylon filters (Amersham, Hybond-N). Hybridization was carried out in 6x SSC/2x Denhardt's solution/0.1% SDS/50% formamide/100 ,ug/ml of sonicated salmon sperm DNA at 42°C for 14 hr with the probe at _106 cpm/ml (lx SSC = 0.15 M NaCl/0.015 M sodium citrate, pH = 7; Denhardt's solution = 0.02% polyvinylpyrrolidone/0.02% Ficoll/0.02% bovine serum albumin). Filters were briefly washed in 2x SSC/0.1% SDS at room temperature, then washed twice at 50°C in 2x SSC/0.1% SDS for 1 hr and autoradiographed at -70°C using intensifying screens. Plaques that hybridized to the probe were rescreened and selected for sequencing. Subeloning and Sequencing Strategies. One of the positive AgtlO clones containing 2.3 kb was subcloned into a pBluescript vector. Two PstI digestion fragments, 0.8 kb and 0.3 kb, respectively, were again subcloned into pBluescript vectors.

Proc. Natl. Acad. Sci. USA 93 (1996)

7889

The remaining 1.2-kb fragment from the initial subcloning was religated. Sequencing on both strands was performed with the M13 universal primer, the T3 primer, and synthetic primers using a T7 sequencing kit (Pharmacia) or a Taq Track sequencing kit (Promega). Northern Blot Analysis. Total RNA from silkworm fat body, epidermal cells, and Drosophila 1(2)mbn cells was Northern blotted using random-primed 32P EcoRI/PstI-digested cDNA (corresponding to open reading frames of the 1.2-kb fragment encoding most of the GNBP protein sequence) (4). After autoradiography, the filter was dehybridized according to the manufacturer and rehybridized with a B. mori cecropin B2 probe (4, 19) and an a-tubulin probe as an internal standard (20). For detection of a putative GNBP homologue in 1(2)mbn cells, the blot was washed in 4x SSC/0.1% SDS at 22°C for 5 min and then twice at 42°C for 20 min. In Vitro Transcription, in 'Vitro Translation, and Fluorography. The synthesis of synthetic mRNA was performed using linearized pBluescript containing the GNBP insert by bacteriophage-encoded DNA-dependent RNA polymerase (T7 polymerase) using an RNA transcription kit (Stratagene). The rabbit reticulocyte lysate system (Promega) was used for translation of synthetic mRNA using [355]methionine. In vitro translation products were separated by SDS/PAGE. After electrophoresis, the gel was treated with En3Hance (DuPont) and fluorographed at -70°C. Computer Analysis For Sequence Homology. The FASTA, TFASTA, WORDSEARCH, and BLAST algorithm were used to determine nucleic acid and protein sequence homology between GNBP and other known proteins (21). DNA STRIDER 1.1 software was used for hydrophobicity plot analysis (22) Determination of 13-1,3 Glucanase Activity. f3-1,3 glucanase activity was determined by a colorimetric assay using laminarin as substrate (23) and a method using SDS/PAGE protein renaturation followed by a parchyman gel overlay (24). Both naive and immune hemolymph were assayed, as well as partially purified GNBP. Antiserum and Immunoblotting Experiences. Antiserum against rabbit CD14 was prepared using the following procedure: a goat polyclonal anti-rabbit CD14 antibody was raised against a recombinant polypeptide corresponding to amino acids +22 to +353 of rabbit CD14. The recombinant protein was prepared from bacterial lysate using the same procedure that was described for expression and purification of murine CD14 (25). Finally, microsequencing of the N terminus of recombinant CD14 revealed a molar equivalent of methionine followed by the expected rabbit CD14 amino acids (Met-AspThr-Pro-Glu-Pro-Cys-Glu-Leu-Asp ...). The resulting antiserum recognizes rabbit, human, and murine CD14 in Western blotting and also immunohistochemical analysis showed strong staining of CHO/Kl cells transfected rabbit CD14 expression construct, whereas untransfected CHO-Kl cells showed no

staining. B. mori hemolymph plasma, supernatant from a confluent culture of Drosophila 1(2)mbn cells and partially purified GNBP were subjected to SDS/PAGE and Western blotting (26). The goat anti-rabbit CD14 antiserum was diluted to 1:1500 and incubated for 2 hr. Subsequently, the blot was incubated with rabbit anti-goat IgG conjugated to horseradish peroxidase diluted 1:2500 for 1 hr.

RESULTS Purification and Amino Acid Sequencing of GNBP. To identify putative inducible bacterial-binding proteins, we incubated log phase E. cloacae with hemolymph plasma from naive silkworms and with hemolymph plasma from silkworms immunized with E. cloacae 18 hr prior to the binding assay. The bacteria were incubated for 1-30 min with hemolymph plasma followed by two successive 0.5 M NaCl washes to remove the

7890

,.._*8. S..i,X_.. .*. ,_s....xwci,;.*


Immunology: Lee et al.

After stringent washing, GNBP was the major protein that remained bound to the Gram(-) bacterial cell surface (Fig. A). The GNBP eluted from Gram(-) bacteria was separated )y SDS/PAGE and was subjected to in situ proteinase diges-ion to conduct an internal amino acid sequence analysis )ecause the intact GNBP protein was not susceptible to Edman degradation due to its blocked a-amino group. Several 3DS/PAGE separated bands corresponding to GNBP (-15 tg) were cut from the gel, pooled, and dehydrated prior to in ,itu endopeptidase Lys-C digestion. The resulting peptide Fragments were separated by reverse-phase HPLC (Fig. 2B) nd the two major peptide peaks were subjected to automatic mino acid sequencing for primary structure determination Fig. 2C). Mixed Oligonucleotide Primed Amplification of cDNA (MOPAC) Analysis and Isolation of the cDNA-Encoding GNBP. Using internal amino acid sequence data of GNBP, legenerated primers were designed and synthesized. PCR was -arried out as described in Materials and Methods using ,ingle-stranded cDNA from immunized fat body (combination f P1 and P4 or P3 and P2). Only one primer combination (P1 ind P4) gave a 160-bp product (data not shown). The 160-bp )roduct was cloned and sequenced. The deduced amino acid ,equence contained the three peptide fragments of GNBP as analyzed by Edman degradation (Fig. 2C). The cDNA library -onstructed from fat body of E. cloacae-immunized silkworms vas screened using the 32P random-primed 160-bp PCR prodict. Forty positive clones were isolated by screening approxmately 75,000 independent clones. Of 13 randomly chosen positive clones, all had an identical insert size (2.3 kb) except For one clone (1.8 kb). The nucleotide sequence of one of the 2.3-kb clones (ABP5001) was analyzed. Nucleotide and Deduced Amino Acids Sequence. Initially, he EcoRI fragment from clone ABP5001 was subcloned into )Bluescript. Further subcloning was performed as described in Materials and Methods. The nucleotide sequence and deduced amino acid sequence are shown in Fig. 3. The sequence of kBP5001 contained a 5' noncoding region, an open reading Frame of 1401 nucleotides corresponding to 467 amino acid residues. An untranslated flanking sequence of 837 nt was ?resent between the terminal codon and the poly(A) site,

1 2 34 5 6

W_

946743-

..

..

.4

3020 -

FIG. 1. SDS/PAGE analysis of hemolymph plasma proteins extracted from the surface of E. cloacae following the in vitro binding assay. The binding assay was performed using naive hemolymph (lane 3) or immunized hemolymph (lanes 1, 2, and 4-6) as described. The incubation times for the binding assay were 1 min (lane 1), 2.5 min (lanes 2 and 3), 5 min (lane 4), 15 min (lane 5), and 30 min (lane 6). Molecular size markers are given in kilodaltons (phosphorylase b, 94 kDa; bovine serum albumin, 67 kDa; ovalbumin, 43 kDa; carbonic anhydrase, 30 kDa, and soybean trypsin inhibitor, 20 kDa). The position of GNBP is indicated by a solid arrowhead and lysozyme (27) by an open arrowhead.

majority of bound plasma proteins. Subsequent to washing, proteins bound to the bacterial cell wall were extracted using 0.5 M NaCl/0.1 M ammonium acetate buffer (pH 4). The SDS/PAGE of hemolymph plasma proteins extracted from the surface of the Gram(-) bacteria showed two predominant inducible protein bands: a 15-kDa band was shown to be a lysozyme and an unknown 50-kDa protein called GNBP (Fig. 1). The amount of GNBP bound to the bacteria increased with time up to 15 min, leveling off thereafter (Fig. 1). The hemolymph-binding assay supernatant (after 10 min of incubation with bacteria), when re-incubated with a new lot of bacteria for the period of 10 min, revealed only insignificant amounts of GNBP, which suggests that the majority of GNBP was removed from solution by affinity to the first lot of bacteria (data not shown). In vitro binding assays conducted with Gram(+) bacteria (B. licheniformis) or Curdlan beads (,3-1,3 glucan polymer) showed only trace amounts of GNBP bound to these microbial cell walls following NaCl washing identical to those used for the Gram(-) bacteria (data not shown).

A

-ontaining two possible polyadenylylation signals

B -455

I_____

40

w~~~~

LiI7 ~~~'20 majr:DWYTlVFSI mkwo:MlTLFAFXGNLNX r

LDSTSVGTLSAEVLDPYNGRXVYEEPDLK

|

FIG. 2. SDS/PAGE analysis of hemolymph plasma proteins extracted from the surface of E. cloacae following the in vitro binding assay using stringent washing conditions as described. (A) Lane 1, a mixture of marker proteins from top to bottom (phosphorylase b, 94 kDa; bovine serum albumin, 67 kDa; ovalbumin, 43 kDa; carbonic anhydrase, 30 kDa); Lane 2, the control binding assay performed without hemolymph (bacteria alone); Lane 3, the binding assay performed with immunized hemolymph (immunization by injection of E. cloacae); Lane 4, the binding assay performed with immunized hemolymph (immunization by procuticular abrasion and topical application of E. cloacae). (B) Peptide mapping following the in situ digestion of GNBP by endopeptidase Lys-C. HPLC conditions are described in Materials and Methods. Peptide peaks indicated by solid and open arrowheads were subjected to amino acid sequencing analysis. (C) The amino acid sequence of the two peaks form peptide mapping. X denotes amino acids whose identity was questionable.

Immunology: Lee et al. -15

46

G

G

R

V

L

C

L

I

L

F

I

K

I

S

tacgctcaaatgcccgatgtgaaaatacaagcttttcgcccgaaaggacttcgaatatcc Y A Q M P D V K I Q A F R P K G L R I S gttcaagatgtccccaaaatgacactgttcgcgttccaaggcaacttgaaccataaactg V Q

166

D V

P

E

MY L

r

*

P

r

G

N

L N H

K

L

gacagcaccagcgtcgggacactgagtgcagaggtactggatccagtcaacggccgatg S T S V G T L S A E V L D P V NO aRWJ gtgtacgaagagcccgatcttaaactaaaagtcaaggacgtcgtctattataacgcggtc V YE E P DL K L K V K D V V T TN A V ttctcaatcaacaagaaaatatacgagaaaacaaaccaacagttcaccgtaacagagcta L S I N K K I Y E K T N Q Q F T V T E L gaagatcctaatgcaagcacagattctcagaaaccagaatgtaagccaacaaagacgaga E D P ® A S T D S Q K P E C K P T K T R gtgcgaggcggcaaagcgtgtgccggacaaacaatattcgaggagcaatttgattccctg V R G G K A C A G Q T I F E E Q F D S L gacgaaaacgtttggcaaat cgagcagtatataccgatttatcaccccgaataccccttc D E N V W Q I E 0 Y I P I Y H P E Y P F gtgtcc ctaccagcgtaataatttaacagtatctaccgcagatggaaacctacatattaac V S Y Q R N L T V S T A D G N L H I N gccaaact tcaacaacatatgcccggctttctggacgactctatatattctggcacactt A K L Q Q H M P G F L D D S I Y S G T L aatt tgttcagtgggtgta ttcgtcagcagaggcatgcatcaaacaggcttccggtgct 0

226 286

346

406 466 526

586 646

706 766

N L F S G C T S S A E A C I K Q A S G A gatattctaac :caatcgtcagcggcagaatcacatcaataggatt tgcatttacgtac D I L P P I V S G R I T S I G F A F T Y T

V

E

I

R

A

K

L

P

Q

G

D

W

L

Y

P

E

I

L

E

P

F

L

K

K

Y

G

S

M

N

Y

A

S

G

V

V

K

I

886

gcgtgtgcgcgcgggtaatgcagaactctattccggacctaatgactacagcaatacggtt

946

A C A R G N A E L Y S G P N D Y S N T V t tgtacggaggaccgatcatggatctggagtgccgcgagaac t ttctctccacaaaaaga L Y G G P I M .D L E C R E N F L S T K R

1006

1066 1126

1186

P

F

D

D

H

F

Y

I

T

L

G

V

A

A

G

G

I

T

E

1246

ttccgcgacgggtctataacctccgggggagtcaccaagccctggagagacagcgctcgg

1306

aaggcatccgtacatttctggcggcacatgtccgattggttcccacggtggagtcagcca

F

1366

R

D

G

S

I

T

S

G

G

V

T

K

294

Eci

554

DVMEARGRL

346 598 1.21 175 172

FHTYSVQTPDPFIALSVDG

135

GNBP Bci Fsu Cth

155

Barn

169 167 1/3 159

YHTYAFDWQPNSIKWYVDG FHITYAFD1WQPGYIKWYVDG YHTYAFDWQPNSIKNYVDG YHIYSVEWNSSIVKWFVNG

95

115

P

W

R

215 235 255

275 295 315

335 355 375 395

D

S

A

R

415

F

W

R

H

1

S

D

W

Q P tctttaatcgtggactttgtcaaagttatcgccttatgatctagtcatttagattt tgta

455

S

L

I

V

D

F

V

I

V

I

A

L

1

467

S

taattaagttattagtatagtttctgcaggcaaaattaattgcattctgagtttttattg

1486 1546 1606 1666 1726 1786 1846

tcaaactaca-tttaaatttaattcctaaaagattctctCtctttctaaatctgtaatt tagctaattttatttgtatcattttctttttttaaattttgatagagcggcgaatgacta gaagatgtgatctgcgaaamMactaagtgacatgcatgcaaaggcatacacacaaca

*

0

*0

*-

KIACARGNAELYSGPNDYSNVYGGPIMDLEC'RENFLSIKRRRDGI'SWGDS 345 ..... PGSVSGTIHFGG ...QWPVNQSSGGDYHFPEGQTFANI) 59 7

YHVYSVVWEEDNIKWYVDG FHTYGLEWTPNYVRWTVDG FHTYGFKWRPDYIDFYVDG YHTYAFDWQPNSIKWYVDG

364 616 139

193 190 189 187 191 177

FIG. 4. Sequence homology between B. mori GNBP and B. circulans ,3-1,3 glucanase Al, bacterial ,3-1,3-1,4 glucanases, and Tachypleus tridentatus coagulation factor Ga. The identical amino acid residues between B. mori GNBP and B. circulans X3-1,3 glucanase Al are indicated by dots. Conserved sequences are indicated by shaded areas. The alignment is optimized by introducing gaps using the BESTFIT program. Numbers on the left and right refer respectively to the residue numbers of the first and last amino acid residues in each line. Bci, B. circulans ,3-1,3 glucanase; Fsu, F. succinogenes /3-1,3-1,4 glucanase; Cth, Clostridium thermocellum ,3-1,3-1,4 glucanase; Bsu, Bacillus subtilis ,B-1,3-1,4 glucanase; Bam, Bacillus amyloliquefaciens (3-1,3-1,4 glucanase; Bma, Bacillus macerans ,3-1,3-1,4 glucanase; Bli, B. licheniformis (3-1,3-1,4 glucanase; and FGa, (1,3)-(3-D-glucansensitive coagulation factor G a-chain.

435

H

W

so

195

V

R

Bli FGC

S

P

Bsu Bma

175

A

1426

1906 1966 2026 2086 2146 2206

GNICP

55

K

F

***-*

75

gcgccattcgatgaccacttctacataacgctaggcgtggcggcaggaggcataacggag

A

*

35

cgcagagacggcacgtcgtggggcgacagttttcatacatattccgttcaatggacacct R R D G T S W G D S F H T Y S V 0 W T P gatt tcatagccctgtctgtggacggcgaggagtgggcgcgagtggaggcgccgcgggac D F I A L S V D G E E W A R V E A P R D gc cctgccggctgtctgcgcgcacgccccgcggcacctgctgcaggccggctcgcagatg A L P A V C A H A P R H L L Q A G S Q M

**

243 503

ttggagccgttcttaaagaagtatggaagtatgaattatgcgtccggcgtagtgaagata L

v

7891

SGRI.TSIGFAFTYGTVEIRAKLPQGDWLYPEILLEPFLKKYGSSONYASGVV 293 SGKINTKDKLSLKYGRVDFRAKLPTGDGVWPALWMLPKDSVYGTWA .ASGEI 553

GNBP Bci

15

ggaacagtcgaaatcagagcg'aattaccgcaaggggactggctgtatccggaaattcta G

826

a

gaggatccgggtaccatgggtggacgcgtgttgtgtctgattttatttataaaaatatcg M

106


gagttttgcattattattacctttcttgagcgagatgaggactactgacgtgacctggag tgatattttggagcaagcctaagacagggtcaagtggagacagctctaaggcccctatgt accggcggggtgtcttagggctgcaataacaacattacctttcttaaccaaacagcgata ttatctgtaatattatagttagaaagcctgcatcttatactcttattgacacacgtacct ggatgattttaacgatgaagaaataacatcgtgtaaaaaatttagctggaaaagattttg aggaatcaagaaataccctatgaaccagcacaggtacgtgtcaccactctggctaattct gccacgttgcggttggtttgaagtttgagacaacctttgcacgatacaattgagactaag acctcatggctcggagtgagtggagtcattcaagttgcgactggctccagtaaccactta taaccaggtagatcgtgagcctatcagctcatttagagcaacaaaaaacaaaataatgta ctcttgaaaaaatcattttaataaaagaccagagtgaaaaaaaaaaaaaaa

FIG. 3. The nucleic acid and deduced amino acid sequences of cDNA-encoding GNBP. Underlined amino acid residues were identified by Edman degradation. Boxed amino acid residues were not clearly identified by the amino acid sequencing analysis. Amino acid residues in boldface type were used for the synthesis of degenerated primers for MOPAC analysis. The circled asparagine (N) residues indicate potential N-glycosylation sites. Possible polyadenylylation signals (AATAAA) are indicated in shaded areas. Putative RNA instability signals (ATTTA) are indicated by boldface type.

(AATAAA), which could allow for mRNAs of different sizes. The 3' untranslated region also contains six ATTTA putative mRNA instability sequences (28), which are commonly found in insect antibacterial peptide mRNAs. However, GNBP mRNA stability was not examined experimentally. The calculated molecular weight of the deduced protein was 52,225 daltons of which about 2,000 daltons corresponds to the putative signal peptide. This is in good agreement with the molecular weight of the in vitro translation product using the

cloned cDNA (data not shown). Two potential N-glycosylation sites were identified at amino acid positions 119 and 182. When analyzed using a hydrophobicity projection, the first 18 amino acids gave a typical signal peptide profile (data not shown). Even though GNBP was isolated from hemolymph as a soluble protein, the hydrophobic C-terminal portion of GNBP suggested that a membrane bound form of this molecule may exist. In fact, the C terminus exhibited the amino acid sequence requirements necessary for a glycosyl-phosphatidylinositol anchor attachment (29); however, this was not proven experimentally. Comparison of the Amino Acid Sequence of GNBP with That of Other Proteins. Searching various databases for sequence similarity to the GNBP open reading frame revealed significant sequence homology with bacterial glucanases. The

highest sequence homology was ascribed to the cDNA of 13-1,3 glucanase Al (EC 3.2.1.39) of Bacillus circulans WL-12 (24), 34% identity and 56% similarity over a stretch of 124 amino acids (Fig. 4). In this region we found the highly conserved 19 amino acids that are commonly shared by B. circulans ,B-1,3 glucanase and numerous bacterial ,B-1,3-1,4 glucanases (EC 3.2.1.73) (30, 31) (Fig. 4). In this specific region, GNBP shows 53% identity and 63% similarity with B. circulans 13-1,3 glu-

2 3 4 5 6 9.5-

7.5 4.424-

v

1.4-

0.24-

1 23 4 5 6

B 9.5_

7.5

2.4-_ 1.4-

0.24-

FIG. 5. Induction of a 1.6-kb GNBP mRNA in axenicB. mori larvae following a procuticular abrasion. (A) Total RNA (50 ,tg) was extracted from larval fat body (lanes 1, 3, and 5) and epidermal cells (lanes 2, 4, and 6) at 0 hr (lanes 1 and 2) and 6 hr (lanes 3 and 4) following a procuticular abrasion in the presence of E. cloacae and 6 hr after abrasion (lanes 5 and 6) following a sterile procuticular abrasion in the absence of bacteria. The blot was hybridized with a random-primed product of the PstI/EcoRI digested ABP5001 (1.2 kb) fragment. An RNA ladder (Bethesda Research Laboratories) was used for the size marker (in kb). (B) The same blot was dehybridized and rehybridized with constitutively expressed a-tubulin cDNA probe (open arrowhead) and a B. mori cecropin B2-inducible immune gene cDNA probe (solid arrowhead).

Immunology: Lee et al.

7892


canase and 53% identity and 90% similarity with Fibrobacter succinogenes (3-1,3-1,4 glucanase (32). It is suspected that this region might be pertinent to the recognition of the j3-(1 ->3) linkage of polysaccharides (24). Interestingly, this region is also highly conserved in the a-chain of the T. tridentatus intracellular serine protease zymogen heterodimer called (1,3)-f3-Dglucan-sensitive coagulation factor G (33) (Fig. 4). We found 37% identity and 68% similarity in the region between GNBP and Tachypleus factor G a-chain. In Tachypleus, the a-chain of (1,3)-(3-D-glucan-sensitive coagulation factor G, which also contains three xylanase A repeats, is thought to be involved in (1,3)-p3-D-glucan binding (33); however, no xylanase motifs were detected in GNBP. The database search also revealed 21% homology and 46% similarity with the overall amino acid sequence of the human LPS receptor, CD14 (34). Induction of GNBP mRNA in Response to Bacterial Challenge. Given the elevated level of GNBP in the hemolymph plasma capable of binding Gram(-) bacteria following a bacterial challenge by injection or cuticular abrasion, we examined the induction of GNBP mRNA following a procuticular abrasion using axenic B. mori larvae. The 32P randomprimed product of the ABP5001 PstI/EcoRI-digested fragment was used as a probe for Northern blot analysis. Only a 1.6-kb GNBP mRNA transcript was constitutively expressed at low levels in the fat body and in epidermal cells of naive larvae (Fig. SA). Six hours after procuticular abrasion in the presence of bacteria the same mRNA was strongly induced in fat body and to a lesser degree in cuticular epidermal cells underlying the abraded zone. Even sterile abrasion in the strict absence of bacteria was capable of slightly inducing mRNAs in both tissues at 6 hr after abrasion. As a positive control, the blot was dehybridized and rehybridized with a constitutively expressed gene (a-tubulin) and an inducible antibacterial peptide gene (cecropin B2) (Fig. SB). These results strongly suggest that GNBP is an acute phase protein that may play a role in the immune response of the silkworm. To examine the existence of a putative homologue of GNBP in the immune competent 1(2)mbn Drosophila cell line (12, 35), we performed northern blotting of total RNA from LPS-induced and noninduced 1(2)mbn cells using our GNBP probe. A 2.4-kb transcript of comparable intensity was detected in both control and LPSinduced cells (data not shown). Lack of 8-1,3 Glucanase Activity. The degree of homology with B. circulans (-1,3 glucanase led us to investigate a putative glucanase activity of GNBP in silkworm hemolymph. No

A

B 1 2 3

c

123 94

-*-43 ...4 -43 67

67

....

1

H~-9

I~-6 WN_4 I

-30

-67 -43

FIG. 6. Immune detection of GNBP and its putative Drosophila homologue using goat anti-rabbit CD14 antiserum. (A) Bombyx hemolymph (2 ,ul): Ponceau staining (lane 1), immunoblotting of lane 1 with goat anti-rabbit CD14 antiserum (lane 2), immunoblotting of lane 1 with pre-immune goat serum (lane 3). (B) Supernatant of Drosophila 1(2)mbn cells (350 tlI of supernatant was precipitated by acetone and dissolved in SDS/PAGE sample buffer before loading): Ponceau staining (lane 1), immunoblotting of lane 1 with goat anti-rabbit CD14 antiserum (lane 2), immunoblotting of lane 1 with pre-immune goat serum (lane 3). (C) Western blot analysis of partially purified GNBP extracted from the surface of E. cloacae following the in vitro binding assay of immune hemolymph plasma with goat anti-rabbit CD14 goat

antiserum.

glucanase activity could be detected when assayed against laminarin and parchyman substrates as described in the Materials and Methods (data not shown). Detection of GNBP Using a Goat Anti-Rabbit CD14 Polyclonal Antiserum. A goat anti-rabbit CD14 polyclonal serum specifically crossreacted with GNBP from naive hemolymph (Fig. 6A) and with a 50-kDa protein, possibly a GNBP homologue, from the supernatant of 1(2)mbn cells (Fig. 6B), as well as partially purified GNBP from the binding assay (Fig. 6C). Antibodies from pre-immune serum did not detectably crossreact with any proteins in these experiments (Fig. 6 A and B). Although the crossreactivity of GNBP with the anti-CD14 serum suggests that GNBP might be an invertebrate CD14 homologue, data on the function of GNBP are not yet available to confirm the possibility.

DISCUSSION Immune pattern recognition and signal transduction of the immune response are subjects of great interest in contemporary biology (36). Insects, like vertebrates, have been shown to use the Rel family of transcription factors to activate the transcription of immune effector molecules following injury and microbial invasion (37-41). However, in insects the upstream signal transduction cascade(s) leading to the transactivation of the transcription factors remains unknown. At present, only three inducible Gram(-) bacterial binding proteins have been discovered in insect hemolymph: a Periplaneta americana LPS binding protein (42, 43); hemolin, a bacterial surface-binding protein belonging to the immunoglobulin superfamily (13, 44); and GNBP as described in this paper. None of these binding proteins, however, have been proven unambiguously to assume the role of a receptor. These insect binding proteins appear to be functionally similar by having affinity to the Gram(-) bacterial cell wall and by their inducibility during injury or infection. Nevertheless, no primary structure similarity was found between them. On the contrary, GNBP showed significant homology to certain putative polysaccharide binding motifs of B. circulans (3-1,3 glucanase and bacterial (3-1,3-1,4 glucanases (24, 30-32), which have also been recently found in T. tridentatus (1,3)-,BD-glucan-sensitive coagulation factor G a-chain (33). This homology suggests a structural and perhaps a functional relationship between these molecules. Furthermore, another invertebrate defense protein, 1,3-(3-D-glucan-binding protein from the freshwater crayfish Pacifastacus leniusculus was found to contain a short sequence motif with similarity to the active site of bacterial (3-1,3-1,4 glucanases (45). Interestingly, although highest homology was in or near the putative catalytic sites of the bacterial glucanases, neither the T. tridentatus (1,3)-(3-D-glucan-sensitive coagulation factor G a-chain, the P. leniusculus 1,3-(3-D-glucan-binding protein, nor GNBP exhibited glucanase activity. One could speculate that these invertebrate proteins, containing the bacterial glucanase motif, constitute a novel family of microbial polysaccharide pattern recognition immune proteins. This family could possibly be extended to other insects given that we detected a putative GNBP homologue transcript by Northern blot analysis in Drosophila 1(2)mbn cells. It is intriguing to note that an analogous situation was recently found in the chitinase (EC 3.2.1.14) protein family. Human cartilage inflammatory glycoprotein, HC gp-39 (46); an unpublished mouse macrophage secretory glycoprotein, YM-1 (PIR accession no. S27879); and very recently a secretory glycoprotein, DS47 (47) from Drosophila immunocompetent Schneider line-2 cells were found to contain conserved motifs in the catalytic region of microbial chitinases, but again did not manifest any enzymatic activity. It was suggested that DS47 is involved in the binding of specific carbohydrate moieties on the cell surface and may function in tissue remod-

Immunology: Lee et al. eling and/or protection against microbes (47). Studies with recombinant GNBP and GNBP-transfected insect and mammalian immunocompetent cells will help elucidate the precise biological function(s) of GNBP. The predicted protein sequence of silkworm GNBP also showed 21% identity and 46% similarity to the full-length mammalian CD14, a 55-kDa glycoprotein shown to be a receptor for LPS (48). This protein has since been implicated in endotoxin recognition and subsequent immune signal transduction (11). CD14 exists both as a soluble and membrane bound form via a phosphatidylinositol glycan anchor (49). In addition to some amino acid similarity, we also observed some biological similarities with GNBP, e.g., hydrophobic C terminal, soluble forms in serum, capacity to directly bind to Gram(-) bacteria, etc., which prompted us to examine antiCD14 polyclonal antibodies for crossreactivity to GNBP. Interestingly, a goat anti-rabbit CD14 antiserum crossreacted specifically with GNBP in silkworm hemolymph and with a 50-kDa protein from the supernatant of Drosophila 1(2)mbn cells. We cannot explain the biological significance of this result, but certainly some epitopes are shared by CD14, GNBP, and the 50-kDa Drosophila protein. Further studies are necessary to examine the nature of this crossreactivity and the apparent functional similitude between these proteins. We thank P. Couble and J.-C. Prudhomme (Centre de Genetique Moleculaire et Cellulaire, University of Lyon I, France) for their encouragement and valuable discussions. We are grateful to Prof. I. Morishima (Tottori University, Japan) for providing us with his silkworm fat body cDNA library and to Prof. E. Gateff (Johannes Gutenberg University Mainz, Germany) for providing us with her 1(2)mbn cell line. We also acknowledge Dr. J. D'Alayer (Laboratoire de Micros6quencage, Institut Pasteur) for amino acid sequencing analysis. Special thanks are extended to S. Perrot and M. Francois for their excellent technical assistance. This work was supported by the Institut Pasteur, The Pasteur-Weizmann Joint Research Program, and by fellowships to W.-J.L. from the Korean Pohang Steel Company Scholarship Society, Direction des Applications de la Recherche de l'Institut Pasteur, and La Fondation pour la Recherche Medicale. 1. Boman, H. G. (1991) Cell 65, 205-207. 2. Wigglesworth, V. B. (1972) The Principles of Insect Physiology (Chapman & Hall, London). 3. Ashida, M. & Brey, P. T. (1995) Proc. Natl. Acad. Sci. USA 92, 10698-10702. 4. Brey, P. T., Lee, W.-J., Yamakawa, M., Koizumi, Y., Perrot, S., Francois, M. & Ashida, M. (1993) Proc. Natl. Acad. Sci. USA 90, 6275-6279. 5. Boman, H. G., Boman, A. & Pigon, A. (1981) Insect Biochem. 11, 33-42. 6. Dimarcq, J.-L., Zachary, D., Hoffmann, J. A., Hoffmann, D. & Reichhart, J.-M. (1990) EMBO J. 9, 2507-2515. 7. Hultmark, D. (1993) Trends Genet. 9, 178-183. 8. Cociancich, S., Bulet, P., Hetru, C. & Hoffmann, J. A. (1994) Parasitol. Today 10, 133-139. 9. Boman, H. G. (1995) Annu. Rev. Immunol. 13, 61-92. 10. Schumann, R. R., Leong, S. R., Flaggs, G. W., Gray, P. W., Wright, S. D., Mathison, J. C., Tobias, P. S. & Ulevitch, R. J.

(1990) Science 249, 1429-1431. 11. Ulevitch, R. J. & Tobias, P. S. (1994) Curr. Opin. Immunol. 6, 125-130. 12. Gateff, E. (1978) Science 200, 1448-1459. 13. Sun, S.-C., Lindstrom, I., Boman, H. G., Faye, I. & Schmidt, 0.

(1990) Science 250, 1729-1732.


7893

14. Laemmli, U. K. (1970) Nature (London) 227, 680-685. 15. Sambrook, J., Fritsch, E. F. & Maniatis, T. (1989) Molecular Cloning: A Laboratory Manual (Cold Spring Harbor Lab. Press, Plainview, NY), 2nd Ed. 16. Auffray, C. & Rougeon, F. (1980) Eur. J. Biochem. 107, 303-314. 17. Lee, C. C., Wu, X., Gibbs, R. A., Cook, R. G., Muzny, D. M. & Caskey, C. T. (1988) Science 239, 1288-1291. 18. Lee, C. C. & Caskey, C. T. (1992) Methods Enzymol. 216, 69-72. 19. Taniai, K., Kato, Y., Hirochika, H. & Yamakawa, M. (1992) Biochem. Biophys. Acta 1132, 203-206. 20. Kalfayan, L. & Winsink, P. C. (1981) Cell 24, 97-106. 21. Argos, P. (1990) Methods Enzymol. 182, 751-776. 22. Marck, C. (1988) Nucleic Acids Res. 16, 1829-1836. 23. Ashwell, G. (1957) Methods Enzymol. 3, 73-105. 24. Yahata, N., Watanabe, T., Nakamura, Y., Yamamoto, Y., Kamimiya, S. & Tanaka, H. (1990) Gene 86, 113-117. 25. Pugin, J., Heumann, D., Tomasz, A., Kravchenko, V. V., Akamatsu, Y., Nishijima, M., Glauser, M. P., Tobias, P. S. & Ulevitch, R. J. (1994) Immunity 1, 509-516. 26. Towbin, H., Staehelin, T. & Gordon, J. (1979) Proc. Natl. Acad. Sci. USA 76, 4350-4354. 27. Lee, W.-J. & Brey, P. T. (1995) Gene 161, 199-203. 28. Shaw, G. & Kamen, R. (1986) Cell 46, 659-667. 29. Coyne, K. E., Crisci, A. & Lublin, D. M. (1993)J. Biol. Chem. 268, 6689-6693. 30. Hofemeister, J., Kurtz, A., Borris, R. & Knowles, J. (1986) Gene 49, 177-187. 31. Schimming, S., Schwarz, W. H. & Staudenbauer, W. L. (1992) Eur. J. Biochem. 204, 13-19. 32. Teather, R. M. & Erfle, J. D. (1990) J. Bacteriol. 172, 3837-3841. 33. Seki, N., Muta, T., Oda, T., Iwaki, D., Kuma, K.-I., Miyata, T. & Iwanaga, S. (1994) J. Biol. Chem. 269, 1370-1374. 34. Haziot, A., Chen, S., Ferrero, E., Low, M. G., Silber, R. & Goyert, S. M. (1988) J. Immunol. 141, 547-552. 35. Samakovlis, C., Asling, B., Boman, H. G., Gateff, E. & Hultmark, D. (1992) Biochem. Biophys. Res. Commun. 188, 1169-1175. 36. Janeway, C. A. (1994) in Phylogenetic Perspectives in Immunity: The Insect Host Defense, eds. Hoffmann, J. A., Janeway, C. & Natori, S. (R. J. Landis, Houston), pp. 115-122. 37. Sun, S.-C. & Faye, I. (1992) Comp. Biochem. Physiol. 103, 225-233. 38. Kobayashi, A., Matsui, M., Kubo, T. & Natori, S. (1993) Mol. Cell. Biol. 13, 4049-4056. 39. Ip, Y. T., Reach, M., Engstrom, Y., Kadalayil, L., Cai, H., Gonzalez-Crespo, S., Tatei, K. & Levine, M. (1993) Cell 75, 753-763. 40. Reichhart, J. M., Georgel, P., Meister, M., Lemaitre, B., Kappler, C. & Hoffmann, J. A. (1993) C. R. Acad. Sci. III 316, 1218-1224. 41. Hultmark, D. (1994) Nature (London) 367 116-117. 42. Jomori, T., Kubo, T. & Natori, S. (1990) Eur. J. Biochem. 190, 201-206. 43. Jomori, T. & Natori, S. (1991) J. Biol. Chem. 266, 13318-13323. 44. Ladenforff, N. E. & Kanost, M. R. (1990) Arch. Insect Biochem. Physiol. 15, 33-41. 45. Cerenius, L., Liang, Z., Duvic, B., Keyser, P., Hellmann, U., Tapio Palva, E., Iwanaga, S. & Soderhall, K. (1994)1. Biol. Chem. 269, 29462-29467. 46. Hakala, B. E., White, C. & Ricklies, A. D. (1993) J. Biol. Chem. 268, 25803-25810. 47. Kirkpatrick, R. B., Matico, R. E., McNulty, D. E., Strickler, J. E. & Rosenberg, M. (1995) Gene 153, 147-154. 48. Wright, S. D., Ramos, R. A., Tobias, P. S., Ulevitch, R. J. & Mathison, J. C. (1990) Science 249, 1431-1433. 49. Bazil, V., Horejsi, V., Baudys, M., Kristofova, H., Strominger, J. L., Kostka, W. & Hilgert, I. (1986) Eur. J. Immunol. 16, 1583-1589.