Innate Immunity in : The Role of Multiple, Endogenous Serum Lectins in the Recognition of Foreign Invaders in the , discoidalis This information is current as of September 25, 2021. Raymond Wilson, Changwei Chen and Norman A. Ratcliffe J Immunol 1999; 162:1590-1596; ; http://www.jimmunol.org/content/162/3/1590 Downloaded from

References This article cites 27 articles, 7 of which you can access for free at: http://www.jimmunol.org/content/162/3/1590.full#ref-list-1 http://www.jimmunol.org/ Why The JI? Submit online.

• Rapid Reviews! 30 days* from submission to initial decision

• No Triage! Every submission reviewed by practicing scientists

• Fast Publication! 4 weeks from acceptance to publication

*average by guest on September 25, 2021

Subscription Information about subscribing to The Journal of Immunology is online at: http://jimmunol.org/subscription Permissions Submit copyright permission requests at: http://www.aai.org/About/Publications/JI/copyright.html Email Alerts Receive free email-alerts when new articles cite this article. Sign up at: http://jimmunol.org/alerts

The Journal of Immunology is published twice each month by The American Association of Immunologists, Inc., 1451 Rockville Pike, Suite 650, Rockville, MD 20852 Copyright © 1999 by The American Association of Immunologists All rights reserved. Print ISSN: 0022-1767 Online ISSN: 1550-6606. Innate Immunity in Insects: The Role of Multiple, Endogenous Serum Lectins in the Recognition of Foreign Invaders in the Cockroach, Blaberus discoidalis1

Raymond Wilson,2 Changwei Chen, and Norman A. Ratcliffe3

Unlike vertebrates, insects do not have an Ab-based nonself recognition system, and must rely totally on innate immunity to defend themselves from microbial invaders. The most likely candidates for recognizing foreign material in insects are the lectins, which have already been shown to be important in mammalian innate immunity. The hemolymph of the cockroach, , contains multiple lectins, designated BDL1, BDL2, BDL3, and GSL (␤-1,3-glucan-specific lectin), two of which, namely BDL1 and GSL, have close similarities to acute phase reactants. These endogenous molecules, as well as Con A, wheat germ agglutinin, and

Helix pomatia agglutinin, have been shown to induce an enhanced phagocytic response by B. discoidalis plasmatocytes. This effect Downloaded from is related to the carbohydrates presented on the surface of the microorganism and to the sugar specificities of the lectins. Thus, the mannose-specific lectins, BDL1 and Con A, both increase the phagocytosis of baker’s yeast and Escherichia coli, whereas the N-acetyl-D-glucosamine/N-acetyl-D-galactosamine-specific lectins, BDL2, wheat germ agglutinin, and H. pomatia agglutinin, induce the phagocytosis of Bacillus cereus and E. coli. GSL, specific for ␤-1,3-glucan, and the N-acetyl-D-galactosamine-specific BDL3, only enhance the phagocytosis of yeast and B. cereus, respectively. Phenylthiourea, an inhibitor of the prophenoloxidase system,

caused either total, partial, or no inhibition of the lectin-induced increase in phagocytosis, indicating that this immune enhance- http://www.jimmunol.org/ ment results, in some cases, from at least two closely linked mechanisms. These results show that the endogenous lectins in the cockroach hemolymph are capable of acting as nonself recognition molecules for a wide range of microorganisms, and thus obviate the necessity of Abs in these . The Journal of Immunology, 1999, 162: 1590–1596.

n vertebrates, the role of lectins as mediators of nonself rec- carbohydrate-binding specificities, and thus able to recognize a ognition in the innate immune response has been well doc- wide variety of invading organisms. To date, multiple lectins have I umented (1), with several mammalian lectins having an op- been purified from only three species of insects, namely, the silk- sonic function and involved in the clearance of microbial agents worm, Bombyx mori (15, 16); the American cockroach, Peripla- (e.g., 2–5). The best studied of these proteins are the mannose- neta americana (17–21); and the West Indian leaf cockroach, Bla- by guest on September 25, 2021 binding lectins, MBLs4 (6), which are an essential component of berus discoidalis (22, 23). Two of the lectins from P. americana, the vertebrate innate immune system, since MBL-deficient indi- i.e., the LPS-binding protein (11) and the Periplaneta lectin (14), viduals are prone to recurrent infections during infancy (7). MBLs have been reported to bind to, and increase the in vivo clearance of not only enhance the phagocytosis of virulent bacteria (2), but also Escherichia coli from the hemolymph. In B. discoidalis, at least activate the complement system through the classical pathway (8). four lectins occur, namely, BDL1, BDL2, BDL3, and GSL, of In invertebrates, including insects, due to the lack of Ab-based which three, BDL1, BDL2 (22), and GSL (23), have been purified. immunity, lectins probably play a major role in nonself recogni- Of these lectins, BDL1 has been shown to have properties similar tion, with several reports of endogenous serum lectins having op- to MBLs in terms of specificity, structure, and activation of com- sonic activity for invading pathogens (e.g., 9–14). If, however, plement (our unpublished data), and GSL has been shown to be lectins act in an analogous manner to Abs, then it is essential that similar to invertebrate C-reactive proteins.5 These endogenous lec- multiple lectins are present in the circulation, each with different tins are also capable of enhancing the activation of the proph- enoloxidase system (24), a melanization cascade that is an impor- tant component of the immune defense system involved in Biomedical and Physiological Research Group, University of Wales Swansea, Sin- gleton Park, Swansea, United Kingdom numerous defense reactions, including encapsulation, nodulation, and phagocytosis (25). Received for publication May 21, 1998. Accepted for publication October 9, 1998. The present study examines in detail, for the first time in an The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance invertebrate, the role of multiple, endogenous lectins in immune with 18 U.S.C. Section 1734 solely to indicate this fact. recognition, in particular, the effects of the four endogenous serum 1 This work was funded by a Biotechnological and Biological Sciences Research lectins from B. discoidalis, and four exogenous lectins of compa- Council studentship to R.W. rable carbohydrate specificity, on the in vitro phagocytosis of rep- 2 Current address: Department of Zoology, University of Aberdeen, Tillydrone Av- resentatives from three major groups of microorganisms, namely enue, Old Aberdeen, Aberdeen, U.K., AB9 2TN. yeast and Gram-positive and Gram-negative bacteria. Each of 3 Address correspondence and reprint requests to Dr. Norman A. Ratcliffe, Biomed- these microorganisms has very different surface carbohydrate ical and Physiological Research Group, School of Biological Sciences, University of Wales Swansea, Singleton Park, Swansea, U.K. SA2 8PP. 4 Abbreviations used in this paper: MBL, mannose-binding lectin; BDL, Blaberus 5 discoidalis lectin; GalNAc, N-acetyl-D-galactosamine; GIM, Grace’s Insect Medium; C. Chen, A. F. Rowley, R. P. Newton, and N. A. Ratcliffe. 1999. Identification, GlcNAc, N-acetyl-D-glucosamine; GSL, ␤-1,3-glucan-specific lectin; HPA, Helix po- purification and properties of a B-1,3-glucan-specific lectin from the serum of the matia agglutinin; proPO, prophenoloxidase; PTU, phenylthiourea; TPA, Tetragonolo- cockroach, Blaberus discoidalis which is implicated in immune defense reactions. bus purpureas agglutinin; WGA, wheat germ agglutinin. Submitted for publication.

Copyright © 1999 by The American Association of Immunologists 0022-1767/99/$02.00 The Journal of Immunology 1591 properties (26, 27), and we have shown that the activation of 25°C, respectively. The resultant cultures were then processed in the same phagocytosis by each lectin is related both to the carbohydrates way as the yeast, except that they were centrifuged at 2000 ϫ g. exposed on the surface of the phagocytosed particle and to the Preparation of the B. discoidalis lectins sugar specificity of the lectin. The endogenous lectins BDL1, BDL2, and BDL3 were purified from the serum of adult by a combination of ammonium sulfate pre- Materials and Methods cipitation, DEAE-cellulose cation-exchange chromatography, and affinity Insects chromatography, as described in Chen et al. (22). GSL was purified by a West Indian leaf cockroaches (B. discoidalis) were kept in fiber-glass con- combination of gel filtration and blue-Sepharose chromatography, accord- tainers at 28°C, with a constant 12-h light/dark cycle. They were fed on ing to the method of Chen (23). The purified lectins were dialyzed against dried cat food, banana skins, and apple cores, and given water ad libitum. GIM for use in the phagocytosis assays. The BDL1 used for the dose- Adult males and females were used for all of the experiments. dependency curves was purified by the following immunoaffinity method, which yields much greater volumes of almost pure lectin than the method Bleeding of insects and preparation of monolayers of Chen et al. (22), with only one major contaminating protein. Mouse anti-BDL1 mAb 6B4 was purified from 500 ml of hybridoma Cockroaches were cooled at Ϫ20°C for approximately 20 min and surface culture supernatant using a protein G-Sepharose 4B column (Sigma) (29). sterilized with 70% ethanol. Using a sterile 26G needle, 0.4 ml of antico- Fractions containing pure Ab were pooled (total of 12 mg Ab), concen- agulant (0.01 M EDTA (free acid), 0.1 M glucose, 0.062 M NaCl, 0.03 M trated down to 2 ml, and washed with carbonate buffer (pH 8) (0.1 M trisodium citrate, 0.026 M citric acid (pH 4.6), 370 mOsm) was injected sodium carbonate in 0.5 M NaCl), using a Centriprep-30 concentrator. The between the second and third abdominal sclerites. Insects were massaged concentrated Ab was then coupled to 0.5 g (dry weight) CNBr-activated gently to circulate anticoagulant before bleeding by piercing the arthrodial Sepharose (Sigma), according to the manufacturer’s instructions, packed membrane of a posterior limb with a 21G needle, and 0.1 ml of hemolymph into a polypropylene minicolumn (Pharmacia, Piscataway, NJ), and equil- Downloaded from collected in a syringe containing 0.4 ml of anticoagulant. The contents of ibrated with TBS (pH 7.4). the syringe were mixed and ejected into a sterile 7-ml Bijou bottle con- A 2-ml aliquot of a 30–50% ammonium sulfate fraction of B. discoi- taining 2 ml of Grace’s Insect Medium (GIM; Sigma, Poole, Dorset, U.K.). dalis serum was applied to the column and allowed to incubate for1hat In cases in which incubation in PTU (phenylthiourea) was required to 4°C. The throughflow was collected and reapplied to the column. This was inhibit the prophenoloxidase (proPO) system, the contents of the syringe repeated an additional two times. The column was then washed thoroughly, were ejected into 2 ml of GIM containing a few crystals of PTU. Bleeding to remove all nonspecifically bound protein, using a stepped salt gradient of the insects was performed in a 4°C constant temperature room to prevent starting with TBS (pH 7.4) alone, and increasing to 2 M NaCl in 0.5 M hemolymph coagulation. increments, then returning to TBS (pH 7.4) alone. http://www.jimmunol.org/ For hemocyte monolayer preparation, sterile 5-mm diameter glass cov- Bound protein was eluted with 10 ml of 0.1 M glycine (pH 2.5), and erslips were placed into the wells of a sterile, 96-well, flat-bottom micro- fractions were collected in 0.5-ml aliquots, in Eppendorf tubes containing titer plate (Nunc, Roskilde, Denmark), and 50 ␮l of diluted hemolymph 100 ␮l of 1 M Tris/HCl (pH 8). Protein-containing fractions (determined was pipetted into each well. The plate was then incubated at 30°C for 20 by absorbance at 280 nm) were tested for purity on SDS-PAGE, and the min to allow the cells to attach to the glass coverslip. To remove any purest fractions were pooled, washed with GIM, and concentrated using a unattached cells, the monolayers were washed three times with 100 ␮lof Centriprep-30 concentrator. GIM. The activities of purified BDL1, BDL2, and BDL3 were determined by Lectin labeling of hemocytes their ability to agglutinate rabbit erythrocytes, as described by Chen et al. (22), or, for GSL, the agglutination of yeast (23), and their concentrations

A number of FITC-labeled lectins were used to probe the surface carbo- were adjusted to 17,000 U/ml for BDL1, 5000 U/ml for BDL2, 10,000 by guest on September 25, 2021 hydrates of the B. discoidalis hemocytes, namely Con A, WGA, HPA, and U/ml for BDL3, and 760 U/ml for GSL, in which 1 U equals an aggluti- TPA (Sigma). All lectins were dissolved in GIM at a concentration of 100 nation titer of 1/mg of purified lectin. ␮g/ml, aliquoted, and stored at Ϫ20°C. Monolayers were overlaid with 50 ␮l of FITC-labeled lectin and incu- Opsonization of yeast and bacteria bated at 29°C for 15 min in the dark. Unbound lectin was gently washed ␮ ␮ FITC-labeled bacteria and yeast suspensions (50 l used per monolayer) off three times with 100 l of GIM. The cells were fixed with GIM-buff- were pelleted at 10,000 ϫ g for 20 s. The supernatants were removed, and ered 4% formaldehyde for at least 30 min at 29°C. The fixative was washed 50 ␮l of lectin were added with gentle agitation to resuspend the pellets, off with GIM, and the 5-mm coverslips, containing the monolayers, were and incubated at 29°C for1hinthedark. The cells were pelleted by removed from the microtiter plate wells, placed on microscope slides, then centrifugation, washed twice in 50 ␮l of GIM to remove any unbound mounted using Kaiser’s glycerol gelatin (50% glycerin, 8% gelatin, 0.01% lectin, and finally resuspended in GIM (50 ␮l). Inhibited lectin controls merthiolate). At least 250 cells per monolayer were then examined under were prepared, as above, except the lectins were preincubated in their re- UV light using a Zeiss photomicroscope II, and equipped with a barrier and spective inhibitory sugars for1hat29°C before addition to the bacteria or excitation filter set number 4877009 for FITC epifluorescence. yeast. The inhibitory sugars used were: 0.2 M mannose for BDL1 and Con Lectins preincubated for 1 h with their respective inhibitory sugars were A; 0.2 M GlcNAc for BDL2; 1 mg/ml GalNAc for BDL3 and HPA; 1% used as negative controls. The inhibitory sugars used for each lectin were laminarin for GSL; and 0.2 M fucose for TPA. BSA (100 ␮g/ml) was used as follows (the lectin that each sugar inhibits is included in the parenthe- (ϩ) as a proteinaceous nonlectin control. ses): D- -mannose (0.2 M; Con A); N,NЈ,NЉ-triacetylchitotriose, a trimer of N-acetyl-D-glucosamine (4.5 mM; WGA); N-acetyl-D-galactosamine Phagocytosis experiments (1.5 mM; HPA); and ␣-D-(ϩ)-fucose (0.2 M; TPA). To investigate whether Con A was binding to the same ligands as The method described below is a modification of the technique described BDL1, monolayers were incubated with 50 ␮l of either BDL1 or GIM for by Rohloff et al. (28). To each hemocyte monolayer, 50 ␮l of opsonized 30 min at 29°C. The cells were then washed five times with 50 ␮l of GIM, bacteria or yeast (1 ϫ 107 cells/ml), or 50 ␮l bacteria or yeast treated with before incubating the cells in 50 ␮l of FITC-labeled Con A for 15 min at BSA or GIM alone, were added, and incubated at 29°C. After 1 h, the 29°C. The monolayers were washed five times with GIM and fixed and monolayers were washed five times with 100 ␮l GIM in each well, to mounted as above. eliminate any noninternalized particles. Monolayers were then overlaid with 50 ␮l of 0.2% trypan blue, and incubated in the dark for 5 min at Culture and labeling of yeast and bacteria 29°C. All traces of trypan blue were washed away with GIM, and the cells were fixed as above for the lectin-labeling experiments. At least 250 cells Baker’s yeast (Saccharomyces cerivisiae; Westmill Foods, Maidenhead, per monolayer were then examined under UV light for phagocytosed bac- Berks, U.K.) was grown up in YM broth (Difco, E. Molesley, Surrey, teria or yeast using a Zeiss photomicroscope II, as above. U.K.), overnight at 25°C, in a shaking water bath. The cultured cells were heat killed in a boiling water bath for 1 h, then washed three times in pH 2ϩ Treatment of monolayers with laminarin 7.4 TBS/Ca (20 mM Trizma base, 77 mM NaCl, 10 mM CaCl2)by centrifugation at 650 ϫ g for 10 min. The washed cells were then labeled To demonstrate the effect of the activation of the proPO system on phago- with FITC by the method of Rohloff et al. (28) and resuspended at a con- cytosis, monolayers were preincubated with laminarin, a known activator centration of 1 ϫ 107 cells/ml in GIM. of proPO. Since laminarin, a ␤-1,3-glucan, is also an inhibitor of GSL E. coli K12 and Bacillus cereus PC2599 were treated in a similar man- receptor recognition, the proPO system was activated before the addition of ner to the yeast cells, except they were grown in nutrient broth at 37°C and the test particles. Adult cockroaches were bled, as described previously, 1592 NONSELF RECOGNITION IN INSECTS

Table I. Staining of B. discoidalis hemocytes with FITC-labeled exogenous lectins a

FITC-Labeled Labeling of Labeling of Lectin Inhibitory Sugar Plasmatocytes Granular Cells

Con A D-(ϩ)-mannose (0.2 M) ϩϩϩ ϩϩ WGA N,N,NЈ-triacetylchitotriose ϩϩ Ϫ (a trimer of GlcNAc; 1 mg/ml) HPA GalNAc (1 mg/ml) ϩϩ Ϫ TPA ␣-D-(ϩ)-fucose (0.2 M) ϪϪ

a Monolayers were incubated with FITC-labeled lectin solutions for 10 min in the dark at 29°C. At least eight monolayers were examined for each treatment. ϩϩϩ, intense staining; ϩϩ, significantly brighter staining than controls; Ϫ, no significant staining compared with controls.

except the contents of the syringe were ejected into a Bijou bottle contain- ing 2 ml of a GIM-buffered solution of 1% (w/v) laminarin. Monolayers were prepared immediately from the resulting cell suspension, and the

phagocytosis assay was performed as described above. Downloaded from Results Comparison of the purification methods of BDL1 The original method of purifying BDL1 by Chen et al. (22) yields approximately 100 ␮g of pure protein from 10 ml of serum. In FIGURE 1. Effect of laminarin and the endogenous lectins from B. dis- comparison, the immunoaffinity method, described in this work, http://www.jimmunol.org/ produces 100 ␮g of protein with a similar sp. act. to that purified coidalis on phagocytosis. FITC-labeled yeast or bacteria was opsonized with lectins, or treated with BSA or GIM alone, for1hat29°C in the dark. by Chen et al. (22), but from only 1 ml of serum. The purified Test particles (50 ␮l) were overlaid on monolayers and incubated for 1 h BDL1 is, however, not totally pure, as it is always copurified with at 29°C. The fluorescence of the external particles was quenched by incu- a protein that has a mass of approximately 40 kDa under both bation with trypan blue for 5 min at 29°C. At least 250 plasmatocytes, from reducing and nonreducing conditions in SDS-PAGE. Since the im- five to seven randomly chosen fields of view, were scored for internalized munoaffinity-purified lectin and the pure lectin, when used at com- particles. For sugar-inhibited controls, lectins were incubated with their parable sp. act., induced phagocytosis by the same amount, it was respective sugar ligands for1hat4°Cbefore opsonization of the yeast or concluded that the contaminating protein was not contributing to bacteria. Laminarin-treated monolayers were prepared from hemolymph the observed effects, and therefore the immunoaffinity-purified lec- bled into a 1% solution of laminarin in GIM. The data are expressed as by guest on September 25, 2021 tin was used in a limited number of experiments. mean percentage of plasmatocytes containing phagocytosed yeast or bac- -p Ͻ 0.005 compared with inhibited controls (Mann ,ء .teria Ϯ SD; n ϭ 6 Lectin labeling of hemocytes Whitney test). Inhibitors used were: 0.2 M D-(ϩ)-mannose for BDL1; 1 mg/ml GlcNAc for BDL2; 1 mg/ml GalNAc for BDL3; and 1% laminarin Con A, HPA, and WGA stained 100% of plasmatocytes from adult for GSL. Laminarin-induced activation of the proPO system was inhibited B. discoidalis, with Con A giving the most intense staining and by the addition of a few crystals of PTU to the hemolymph suspension WGA the weakest (Table I). Only Con A was observed to also before preparation of the monolayers. bind to the granular cells. TPA did not bind to any of the cell types. Lectin binding was completely inhibited by incubating the lectins with their inhibitory sugars before their addition to the monolayers. There was no significant difference in the pattern of labeling be- a maximum level of phagocytosis at a concentration of 13,000 tween fixed and nonfixed hemocytes. U/ml. Incubation of the yeast with GSL also induced a significant The monolayers incubated with BDL1 before the addition of increase in phagocytosis from 6.22 Ϯ 1.48% to 11.86 Ϯ 1.02%, Con A showed no observable difference in the binding of Con A which was inhibited by preincubation of the lectin with laminarin to the hemocytes. (Fig. 1a). In contrast, pretreatment of yeast by either BDL2 or BDL3 had no effect on phagocytosis (Fig. 1a). Phagocytosis of yeast Of the exogenous lectins tested (Fig. 3a), incubation of yeast Nonopsonized yeast, incubated in GIM alone or in 100 ␮g/ml with Con A caused a dose-dependent (Fig. 4a), threefold increase BSA, was phagocytosed by 6.56 Ϯ 0.59% and 5.87 Ϯ 0.56% of in phagocytosis from 6.56 Ϯ 0.59% to 19.13 Ϯ 0.97%, which was plasmatocytes, respectively (Fig. 1a). Granular cells did not ex- inhibitable by 0.2 M mannose. WGA also induced a significant hibit any observable phagocytosis. This basal, nonspecific phago- enhancement in phagocytosis from 6.56 Ϯ 0.59% to 11.86 Ϯ cytosis was increased to over 11% (Fig. 1a) when the monolayers 1.02%, which was approximately the same level as that observed were incubated with laminarin, a known activator of the proPO with laminarin-incubated monolayers (Fig. 1a). HPA and TPA had cascade, before the addition of yeast. This increase in phagocytosis no effect on the phagocytosis of yeast. The effects of both the due to laminarin was completely inhibitable by PTU (Fig. 1a), an endogenous and exogenous lectins on the phagocytosis of yeast are inhibitor of phenoloxidase (30). Regarding phagocytosis of yeast summarized in Table II. in the presence of the endogenous lectins BDL1, BDL2, BDL3, and GSL, a fivefold increase in internalization, from 6.56 Ϯ 0.56% Phagocytosis of E. coli to 31.54 Ϯ 2.03%, was observed when yeast was opsonized with The basal level of phagocytosis of E. coli in GIM (5.36 Ϯ 0.78%) BDL1 (17,000 U/ml; Fig. 1a). This effect was completely inhibited and BSA (6.65 Ϯ 1.86%), and the activation due to laminarin by mannose (Fig. 1a), and was dose dependent (Fig. 2a), reaching (10.10 Ϯ 2.25%; Fig. 1b) were similar to those observed with The Journal of Immunology 1593 Downloaded from http://www.jimmunol.org/

FIGURE 2. Dose-dependency curves for the induction of phagocytosis of yeast and E. coli by BDL1 (a) and B. cereus by BDL2 (b). The phago- cytosis assay was performed as described in the legend to Fig. 1, using various concentrations of lectin to opsonize the yeast or bacteria. Results are expressed as the mean percentage of phagocytosis Ϯ SEM; n ϭ 6. FIGURE 3. Effect of the exogenous lectins on phagocytosis. Phagocy- tosis assays were conducted as described in the legend to Fig. 1. Data are expressed as mean percentage of plasmatocytes containing phagocytosed p Ͻ 0.005. Inhibitors used are shown in ,ء .yeast. The endogenous lectins BDL1, BDL2, and BDL3 all in- yeast or bacteria Ϯ SD; n ϭ 6 duced significant increases in the phagocytosis of E. coli from Table I. by guest on September 25, 2021 4.80 Ϯ 0.96% to 26.05 Ϯ 2.39%, 15.02 Ϯ 2.96%, and 35.04 Ϯ 9.01%, respectively. All of the increases were inhibitable by their respective sugar ligands, namely mannose (BDL1), GlcNAc 3.97%), and the level of phagocytosis due to the activation of the (BDL2), and GalNAc (BDL3). The enhancement of phagocytosis proPO cascade with laminarin (9.60 Ϯ 1.57%) were similar to of E. coli by BDL1 was dose dependent (Fig. 2a), reaching a those observed with yeast and E. coli (Fig. 1c). In this case, pre- maximum level of phagocytosis at approximately 9000 U/ml. incubation of bacteria with BDL1, BDL3, and GSL caused no BDL2 and BDL3 were not tested for dose dependency due to in- activation of phagocytosis, whereas BDL2 induced a very high adequate supplies of purified lectin. In contrast to the other endog- level at 49.25 Ϯ 6.91%, which was almost completely inhibitable enous lectins, preincubation of E. coli with GSL (Fig. 1b) resulted by 1 mg/ml GlcNAc (Fig. 1c). The dose dependency of BDL2 was in no increase in the internalization of E. coli. When BDL1 (4500 sigmoidal in character (Fig. 2b), with a minimum activation con- U/ml) and BDL2 (5000 U/ml) were incubated together, the ob- centration of approximately 900 U/ml and a phagocytic maximum served level of phagocytosis (14.91 Ϯ 5.88%) was not signifi- at 5000 U/ml. cantly different from that seen with each lectin individually at com- WGA and HPA both gave similar highly significant enhance- parable concentrations (17.09 Ϯ 3.73% and 15.02 Ϯ 2.96%, ments of phagocytosis (28.09 Ϯ 7.36% and 28.42 Ϯ 10.08%, re- respectively). spectively (Fig. 3c)), which were significantly inhibited by their With the exogenous lectins (Fig. 3b), those with specificity for specific sugar ligands, while Con A and TPA had no effect on the mannose, GlcNAc, or GalNAc, i.e., Con A, WGA, and HPA, all internalization of B. cereus. Both the enhancements due to WGA augmented the uptake of E. coli to a similar extent from 4.80 Ϯ and HPA were dose dependent and exhibited their maximum aug- 0.96% for the GIM control to 29.66 Ϯ 3.52%, 31.49 Ϯ 6.79%, and mentation of phagocytosis at approximately 80 ␮g/ml, with HPA 26.82 Ϯ 6.81%, respectively. The effects produced by the exoge- also showing a minimum activation concentration of 40 ␮g/ml nous lectins were all dose dependent and reached a maximum at (Fig. 4c). The effects of both endogenous and exogenous lectins on approximately 80 ␮g/ml (Fig. 4b). In the case of WGA, there also the phagocytosis of B. cereus are summarized in Table II. appeared to be a minimum activation concentration (20 ␮g/ml) The presence of laminarin in monolayers treated with lectin- below which no significant increase occurred (Fig. 4b). The effects opsonized test particles had no effect on the lectin-induced en- of both the endogenous and exogenous lectins on the phagocytosis hancement of phagocytosis for any of the microorganisms tested. of E. coli are summarized in Table II. Effect of PTU on the lectin-mediated enhancement of Phagocytosis of B. cereus phagocytosis The basal level of phagocytosis of B. cereus in GIM alone (6.74 Ϯ PTU is a potent inhibitor of the proPO system, a melanization 1.72%), the level of phagocytosis in the presence of BSA (8.10 Ϯ cascade that is involved in numerous immunologic reactions in 1594 NONSELF RECOGNITION IN INSECTS Downloaded from

FIGURE 5. Effect of the inhibition of the proPO system on the lectin- enhanced phagocytic rates. Cockroaches were bled into PTU-containing GIM or into GIM alone, and monolayers were prepared from the resultant http://www.jimmunol.org/ hemocyte suspensions. The phagocytosis assay was otherwise performed as described in the legend to Fig. 1. Y, yeast; B, B. cereus;E,E. coli. Results are expressed as the mean percentage of phagocytosis Ϯ SEM; n ϭ 6.

FIGURE 4. Dose-dependency curves for the induction of phagocytosis different from the basal level of phagocytosis in GIM alone. In by the exogenous lectins. a, Effect of Con A on the phagocytosis of yeast. contrast, phagocytosis was completely unaffected by PTU when by guest on September 25, 2021 b, Effect of Con A, WGA, and HPA on the phagocytosis of E. coli. c, Effect BDL1, or BDL2, was incubated with E. coli (Fig. 5a). A different of WGA and HPA on the phagocytosis of B. cereus. The phagocytosis effect again was seen when BDL2 was incubated with B. cereus, assays were performed as described in the legend to Fig. 1, using various with the presence of PTU causing a 40% reduction in phagocyto- concentrations of lectin to opsonize the yeast or bacteria. Results are ex- sis, from 48.11 Ϯ 6.99% to 28.93 Ϯ 3.54% (Fig. 5a). This PTU- pressed as the mean percentage of phagocytosis Ϯ SEM; n ϭ 6. inhibited phagocytosis was still significantly ( p Ͻ 0.0001) higher than the nonstimulated level due to GIM alone. invertebrates, including nonself recognition processes (31). In the With the exogenous lectins, PTU had no significant effect on the present study, since it has been shown previously that lectins and phagocytosis of WGA-incubated yeast or Con A-incubated E. coli the proPO system interact (24), PTU was used to determine (Fig. 5b). Only partial inhibition of the lectin-enhanced phagocy- whether proPO was involved in the lectin-mediated enhancement tosis was observed with all the other lectin/test particle combina- of phagocytosis. tions that induced an increase in phagocytosis. All of these par- The presence of PTU produced a total inhibition of the BDL1 tially inhibited levels of phagocytosis were still significantly and GSL-induced phagocytosis of yeast (Fig. 5a), to levels (7.71 Ϯ higher ( p Ͻ 0.005) than the respective basal levels of 2.53% and 5.47 Ϯ 1.32%, respectively) that were not significantly phagocytosis.

Table II. Summary of the effects of endogenous and exogenous lectins on the phagocytic rate by the B. discoidalis plasmatocytes.

Induction of phagocytosis ofa

Lectin Specificity Yeast E. coli B. cereus

BDL1 Man Ͼ GalNAc Ͼ Glc, Gal ϩϩϩ ϩϩ Ϫ BDL2 GalNAc Ͼ Gal Ͼ GlcNAc Ϫ ϩ ϩϩϩ BDL3 GalNAc Ͼ Gal Ͼ Fuc Ϫ ϩϩϩ Ϫ GSL ␤-1,3-glucan ϩϪ Ϫ Con A Man Ͼ Glc ϩϩϩϪ WGA Chitotriose ӷ GlcNAc ϩ ϩϩϩ ϩϩ HPA GalNAc Ͼ GlcNAc Ϫϩϩϩϩ TPA Fuc ϪϪ Ϫ

a Symbols are as follows: ϩϩϩ, enhanced phagocytic rate to over 30%; ϩϩ, enhanced phagocytic rate to between 20% and 30%; ϩ, enhanced phagocytic rate significantly but to a level below 20%; and Ϫ, no enhancement of the phagocytic rate. The Journal of Immunology 1595

Discussion in sufficient quantities to induce a large phagocytic response by B. discoidalis plasmatocytes. Since insects lack the Ab-based, nonself recognition system that is E. coli is a Gram-negative bacterium, and, therefore, the major so important in the vertebrate immune response, they have to rely constituent of its outer cell membrane is LPS (26). LPS is basically entirely upon innate immune defense mechanisms. Like mammals, composed of a membrane-bound lipid A moiety bound to a poly- in which the MBLs activate the complement system and enhance saccharide chain of variable length and composition dependent on phagocytosis, lectins have also been shown to play an important the strain of bacterium. For these experiments, the laboratory strain role as mediators of innate immunity in insects. For example, the K12 was used, which has a relatively simple carbohydrate struc- multiple lectins in B. discoidalis hemolymph are capable of en- ture (26, 34) since it lacks the highly variable O-antigenic portion hancing nonself recognition and ingestion of foreign invaders by of the polysaccharide chain. Of the lectins tested, only GSL and activating the proPO system (24), which, like mammalian com- TPA failed to significantly increase the phagocytosis of E. coli, plement, consists of a complex cascade of immune proteins. which, on examination of the structure of the K12 LPS, is fairly Previously, reports of opsonic activities of insect lectins have predictable since neither ␤-1,3-glucan nor fucose is present. There demonstrated an enhanced phagocytic rate against Gram-negative are, however, numerous binding sites for BDL1 and Con A, since bacteria (11, 14), fungal blastospores (9, 13), or erythrocytes (10). both have activity toward glucose (Table II), while BDL2 and In the cockroach, B. discoidalis, four lectins, each exhibiting dif- BDL3 can bind to the galactose residues, and BDL2, WGA, and ferent carbohydrate specificities, have been isolated (22, 23). Two HPA can interact with the terminal GlcNAc residue of E. coli K12 of these lectins, BDL1 and GSL, have been shown to share func- LPS (Table II). However, the structure of LPS alone does not 5 tional and structural similarities to known acute phase reactants. explain the high levels of phagocytosis obtained with either WGA Downloaded from The present study demonstrates for the first time in an insect that or HPA, since the terminal GlcNAc residue is not always present multiple endogenous serum lectins are capable of potentiating the (26). The levels of phagocytosis observed with WGA and HPA phagocytosis of a variety of very different microorganisms. The are, therefore, probably due to binding of the enterobacterial com- observed effects are inhibited significantly by the sugar ligands of mon Ag that is found on the cell surface of all members of the the lectins and are, therefore, dependent on the carbohydrate rec- Enterobactericidae. This is a 7-kDa molecule composed of repeat- ognition domain of the lectins. Thus, the lectins are probably act- ing units of GlcNAc and N-acetyl-D-mannosaminuronic acid cross- http://www.jimmunol.org/ ing as bridging molecules, by binding to the external polysaccha- linked to palmitic acid residues. BDL2 is probably also capable of rides of the bacteria and yeast and then to receptors on the surface binding to the GlcNAc residues in this molecule. of the plasmatocytes. This is substantiated by the use of the ex- Finally, B. cereus, like all Gram-positive bacteria, has a thick ogenous lectins, of similar sugar specificities to the B. discoidalis peptidoglycan layer covering the entire surface of the bacterium endogenous lectins, which produce analogous opsonic effects to (26). Peptidoglycan is comprised of repeating units of GlcNAc and those observed with the endogenous lectins, and which were also N-acetylmuramic acid linked by short polypeptide chains, and shown to be capable of binding to the surface of the plasmatocytes. from this it is obvious why only those lectins with a specificity The binding of the endogenous lectins to the hemocyte receptors is toward GlcNAc (BDL2, WGA, and HPA; Table II) were able to probably independent of the carbohydrate recognition domain, at induce a significant increase in the phagocytosis of B. cereus. by guest on September 25, 2021 least in the case of BDL1, since no observable reduction in the It has previously been shown that both endogenous B. discoi- intensity of binding to the plasmatocytes by Con A was observed dalis lectins and exogenous lectins are capable of activating the in the presence of BDL1. proPO system (24), a melanization cascade important in many in- To examine the role of the B. discoidalis lectins in the immune vertebrate immune reactions. Also, phagocytosis of bacteria, by system of B. discoidalis, it is essential to first have a knowledge of insect hemocytes, is enhanced by the activation of the proPO sys- the polysaccharides exposed on the surface of each microorganism tem by laminarin, a ␤-1,3-glucan, as demonstrated both in the tested. For this reason, well-studied bacteria and yeast were used. present study and in previous work (35). Therefore, to determine First, baker’s yeast, S. cerivisiae, is known to have an abundant whether the lectin-induced enhancement of phagocytosis was me- supply of polymannans on its cell surface (32), which would pro- diated through the activation of the proPO system, or was operat- vide numerous binding sites for the mannose-binding lectins, ing through an independent mechanism, the proPO system was BDL1 and Con A. Both of these lectins induced a high level of specifically inhibited by PTU. phagocytosis of yeast. Another group of polysaccharides, abundant Three different effects were observed with the various combi- in yeast cell walls, are the ␤-1,3-glucans (27). GSL, as its name nations of lectins and microorganisms when the proPO system was suggests, binds exclusively to ␤-1,3-glucans (23), and induced inhibited with PTU. First, as was the case with BDL1- and GSL- phagocytosis to a level comparable with laminarin-activated cells. induced phagocytosis, the presence of PTU completely inhibited This level is surprisingly small, considering the high agglutinating the lectin-induced increase in phagocytosis of yeast. Since both of activity of GSL toward yeast, and may suggest that the physiologic these lectins have been demonstrated to cause an increase in the role of GSL is more orientated toward the activation of the proPO activation of the proPO system (24), it is probable that BDL1 and system, or some other unknown function, than directly toward GSL are inducing the phagocytosis of yeast via the activation of nonself recognition. Chitin, a polysaccharide composed of repeat- the proPO system. Second, the opposite effect was observed, with ing units of GlcNAc and N-acetylmuramic acid, is also present, in PTU having no effect on the enhanced phagocytosis of E. coli in small amounts, on the yeast cell surface (33). The presence of the the presence of BDL1, BDL2, or Con A, or the WGA-induced GlcNAc residues should provide binding opportunities for BDL2 phagocytosis of yeast, indicating that a proPO-independent acti- and WGA. HPA should also bind to the GlcNAc residues, since, vation pathway is involved. It may be that the lectin-mediated although its highest specificity is for GalNAc, it also has an affinity binding of E. coli to the surface of the hemocytes causes direct for GlcNAc (Table II). However, neither BDL2 nor HPA had any activation of the phagocytic machinery, possibly through a sec- effect on phagocytosis of yeast, and WGA, which has a very high ondary messenger cascade, such as the binding of small GTPases, specificity for chitin, only induced an increase comparable with which are known to be involved in mammalian phagocytic mech- that obtained by activation of the proPO cascade by laminarin. anisms (36). Finally, an intermediate effect was observed with These results seem to suggest that the chitin in yeast is not present BDL2-opsonized B. cereus, and with the majority of the treatments 1596 NONSELF RECOGNITION IN INSECTS opsonized with the exogenous lectins. In all of these cases, the 11. Jomori, T., and S. Natori. 1992. Function of the lipopolysaccharide-binding pro- level of enhancement of phagocytosis induced by the lectins was tein from the hemolymph of the American cockroach, Periplaneta americana. J. Biol. Chem. 266:13318. reduced by 40–50%, but remained at a level significantly higher 12. McKenzie, A. N. J., and T. M. Preston. 1992. Biological characteristics of the than the basal level, indicating that both the proPO-dependent and Calliphora vomitaria agglutinin. Dev. Comp. Immunol. 16:85. proPO-independent mechanisms are involved in producing the lec- 13. Wheeler, M. B., G. S. Stuart, and K. D. Hapner. 1993. Agglutinin mediated opsonization of fungal blastospores in Melanoplus differentialis (Insecta). J. In- tin-induced level of phagocytosis. sect Physiol. 39:477. The results presented in this work demonstrate that the multiple 14. Kawasaki, K., T. Kubo, and S. Natori. 1993. A novel role of Periplaneta lectin as an opsonin to recognize 2-keto-3-deoxy octonate residues of bacterial lipo- serum lectins found in B. discoidalis, and also exogenous lectins, polysaccharide. Comp. Biochem. Physiol. 106B:675. are capable of acting as nonself recognition molecules for a variety 15. Amanai, K., S. Sakurai, and T. Ohtaki. 1990. Purification and characterization of of potential pathogens, with each lectin potentiating a response haemagglutinin in the hemolymph of the silkworm, Bombyx mori. Comp. Bio- chem. Physiol. 97B:471. with a different set of microbes (summarized in Table II). Since 16. Yoshida, H., K. Kuninori, and M. Ashida. 1996. Purification of a peptidoglycan insects lack the Ab-based recognition system found in the verte- recognition protein from hemolymph of the silkworm, Bombyx mori. J. Biol. brates, the ability of a panel of lectins to potentiate the uptake of Chem. 271:13854. 17. Kubo, T., and S. Natori. 1987. Purification and some properties of a lectin from a variety of microorganisms is especially important in terms of the hemolymph of Periplaneta americana (American cockroach). Eur. J. Bio- invertebrate nonself recognition. This study represents the first ev- chem. 16:875. idence in an invertebrate that multiple, endogenous serum lectins, 18. Kubo, T., K. Kawasaki, and S. Natori. 1990. Sucrose-binding lectin in regener- ating cockroach (Periplaneta americana) legs: its purification from adult hemo- of different carbohydrate specificities, are capable of acting as op- lymph. Insect Biochem. 20:585. sonins toward a range of very different microorganisms. Such mul- 19. Jomori, T., T. Kubo, and S. Natori. 1990. Purification and characterization of tiple lectins in cockroaches and related insects not only obviate the lipopolysaccharide-binding protein from hemolymph of American cockroach Downloaded from Periplaneta americana. Eur. J. Biochem. 190:201. need for an Ab-based immune system, but also the presence of 20. Kubo, T., K. Kawasaki, and S. Natori. 1993. Transient appearance and localiza- antibacterial peptides, which, although present in many other in- tion of a 26-kDa lectin, a novel member of the Periplaneta lectin family, in sect species, are apparently absent in cockroaches (unpublished regenerating cockroach legs. Dev. Biol. 156:381. 21. Kawasaki, K., T. Kubo, and S. Natori. 1996. Presence of the Periplaneta lectin- observation). Once an invading microorganism has been recog- related protein family in the American cockroach Periplaneta americana. Insect nized and phagocytosed by the numerous circulating hemocytes in Biochem. Mol. Biol. 26:355.

B. discoidalis, it can be dealt with by intracellular killing mecha- 22. Chen, C., N. A. Ratcliffe, and A. F. Rowley. 1993. Detection, isolation and http://www.jimmunol.org/ characterization of multiple lectins from the hemolymph of the cockroach, Bla- nisms such as those based on superoxide generation (37). berus discoidalis. Biochem. J. 294:181. In conclusion, this present study emphasizes the importance of 23. Chen, C. 1993. Detection, isolation and characterization of multiple lectins from innate immunity based on multiple lectins and activation of a com- the hemolymph of the cockroach, Blaberus discoidalis. Doctoral dissertation, University of Wales, Swansea, U.K. plex, complement-like proPO cascade that is sufficient in cock- 24. Chen, C., H. J. Durrant, R. P. Newton, and N. A. Ratcliffe. 1995. A study of novel roaches to obviate the need for circulating antibacterial peptides. lectins and their involvement in the activation of the prophenoloxidase system in The efficacy of the system is clearly illustrated by the huge success Blaberus discoidalis. Biochem. J. 310:23. 25. Ratcliffe, N. A. 1991. The prophenoloxidase system and its role in of these insects that apparently thrive in extremely inhospitable, immunity. In Phylogenesis of Immune Functions. G. W. Warr and N. Cohen, eds. microbial-rich environments. CRC Press, Boca Raton, pp. 46–71. 26. Hammond, S. M., P. A. Lambert, and A. N. Rycroft. 1984. The Bacterial Cell by guest on September 25, 2021 Surface. Croom Helm, London & Sydney. References 27. Gooday, G. W., and A. P. J. Trinci. 1980. Wall structure and biosynthesis in fungi. In The Eukaryotic Microbial Cell. G. W. Gooday, D. Lloyd, and 1. Epstein, J., Q. Eichbaum, S. Sheriff, and R. A. B. Ezekowitz. 1996. The collectins A. P. J. Trinci, eds. Cambridge University Press, Cambridge, p. 207. in innate immunity. Curr. Opin. Immunol. 8:29. 28. Rohloff, L.-H., A. Wiesner, and P. Go¨tz. 1994. A fluorescence assay demonstrat- 2. Kuhlman, M., K. Joiner, and R. A. B. Ezekowitz. 1989. The human mannose- ing stimulation of phagocytosis by hemolymph molecules of Galleria mellonella. binding protein functions as an opsonin. J. Exp. Med. 169:1733. J. Insect Physiol. 40:1045. 3. Cooper, D., C. M. Butcher, M. C. Berndt, and M. A. Vadas. 1994. P-selectin 29. Harlow, E., and D. Lane. 1988. Antibodies: A Laboratory Manual. Cold Spring interacts with a ␤ -integrin to enhance phagocytosis. J. Immunol. 153:3199. 2 Harbor Laboratory Press, Plainview, NY. 4. Tino, M. J., and J. R. Wright. 1996. Surfactant protein A stimulates phagocytosis 30. Mayer, A. M., and E. Harel. 1979. Polyphenol oxidases in plants. Phytochemistry of specific pulmonary pathogens by alveolar macrophages. Am. J. Physiol. 270: 18:193. L677. 5. Matsushita, M., Y. Endo, S. Taira, Y. Sato, T. Fujita, N. Ichikawa, M. Nakata, 31. Ratcliffe, N. A., C. M. Leonard, and A. F. Rowley. 1984. Prophenoloxidase and T. Mizuochi. 1996. A novel human serum lectin with collagen- and fibrin- activation, non-self recognition and cell cooperation in insect immunity. Science ogen-like domains that functions as an opsonin. J. Biol. Chem. 271:2448. 226:557. 6. Turner, M. W. 1996. Mannose-binding lectin: the pluripotent molecule of the 32. Stewart, T. S., and C. E. Ballou. 1968. A comparison of yeast mannans and innate immune system. Immunol. Today 17:532. phosphomannans by acetolysis. Biochemistry 7:1855. 7. Super, M., S. D. Gillies, S. Foley, K. Sastry, J.-E. Schweinle, V. J. Silverman, and 33. Aronson, J. M. 1965. The cell wall. In The Fungi, Vol. 1. Academic Press, New R. A. B. Ezekowitz. 1992. Distinct and overlapping functions of allelic forms of York, pp. 49–76. human mannose binding protein. Nat. Genet. 2:50. 34. Lugtenberg, B., and L. van Alphen. 1983. Molecular architecture and functioning 8. Matsushita, M., and T. Fujita. 1992. Activation of the classical complement path- of the outer membrane of Escherichia coli and other Gram-negative bacteria. way by mannose-binding protein in association with a novel C1s-like serine Biochim. Biophys. Acta 737:51. protease. J. Exp. Med. 176:1497. 35. Leonard, C., N. A. Ratcliffe, and A. F. Rowley. 1985. The role of prophenoloxi- 9. Pendland, J. C., M. A. Heath, and D. G. Boucias. 1988. Function of a galactose- dase activation in non-self recognition and phagocytosis by insect blood cells. binding lectin from Spodoptera exigua larval hemolymph: opsonization of blas- J. Insect Physiol. 31:789. tospores from entomogenous hyphomycetes. J. Insect Physiol. 34:533. 36. Greenberg, S. 1995. Signal transduction of phagocytosis. Trends Cell Biol. 5:93. 10. Richards, E. H., and N. A. Ratcliffe. 1990. Direct binding and lectin-mediated 37. Whitten, M. M. A., and N. A. Ratcliffe. 1998. In vitro superoxide activity in the binding of erythrocytes to hemocytes of the insect, Extatosoma tiaratum. Dev. haemolymph of the West Indian leaf cockroach, Blaberus discoidalis. J. Insect Comp. Immunol. 14:269. Physiol. In press.