Entamoeba histolytica - Infection and Immunity

4 downloads 4 Views 2MB Size Report
Proceedings of the. 80th Ciba Symposium on Adhesion of Microorganisms and Pathogenicity, London. 15. Niedermeier, W., T. Kirkland, R. T. Acton, and J. C..

INFECTION AND IMMUNITY, Apr. 1982, p. 0019-9567/82/040396-11$02.00/0

Vol. 36, No. 1

396-406

Attachment and Ingestion of Bacteria by Trophozoites of Entamoeba histolytica R. BRACHA, D. KOBILER, AND D. MIRELMAN* Department of Biophysics and Unit for Molecular Biology of Parasitic Diseases, Weizmann Institute of Science, Rehovoth, Israel

Received 30 July 1981/Accepted 3 December 1981

Entamoeba histolytica trophozoites were found to be very selective in their interactions with bacteria. Two principal mechanisms were shown to be responsible for these interactions. Certain bacteria, such as a number of Escherichia coli and Serratia marcescens strains which are known to contain mannose-binding components on their cell surface, bound to mannose receptors on the amoeba membrane. This attachment was markedly inhibited by ot-methylmannoside (0.5%), especially when the incubations were done at low temperature (5°C). Other bacterial species, such as Shigella flexneri and Staphylococcus aureus, which do not possess a mannose-binding capacity, attached to the amoebae, but only with the aid of concanavalin A or after opsonization of the bacteria with immune serum. In both types of attachment, between 40 and 100 bacteria bound per amoeba, and considerable ingestion of bacteria into amoeba vacuoles was observed at 37°C. The attachment of opsonized bacteria to the amoebae does not appear to be mediated by Fc receptors since Fab' dimers obtained after pepsin digestion of immunoglobulin were capable of mediating adherence. Furthermore, preincubation of the amoebae with aggregated human immunoglobulin G or with heat-inactivated immune serum and EDTA did not inhibit the attachment of opsonized bacteria. The attachment of opsonized bacteria was markedly inhibited by N-acetylglucosamine-containing glycoconjugates, such as peptidoglycan and chitin oligosaccharides, as well as by N-acetylgalactosamine. These results indicate that amoebae can attach and ingest bacteria either by using their membrane-associated carbohydrate-binding protein or by having their mannosecontaining cell surface components serve as receptors. It is generally recognized that pathogenicity of Entamoeba histolytica is related in part to association of the amoebae with suitable bacterial species. Early observations by Westphal (24), Phillips and Gorstein (17), Wittner and Rosenbaum (25), and others have demonstrated that various bacterial species, when allowed to associate with amoebae, augment their virulence, as measured by their pathogenicity upon inoculation into animals. Although these studies have tended to confirm this notion, the precise nature of the amoeba-bacteria relation has remained speculative. Practically nothing is known about the molecular mechanism of interaction between amoebae and bacteria or about the contribution made by the bacteria to the enhancement of amoebic virulence. The development of axenic in vitro methods for E. histolytica trophozoites (8) has made it possible to obtain amoebic inocula free from all other microorganisms with which E. histolytica is usually associated in the intestine. It has thus provided the basis by which studies of various aspects of the amoeba-bacterium-host interrela-

tionship could be undertaken. We were interested, therefore, in determining the molecular basis of the association between bacteria and amoebae. A convenient method was developed for the separation of amoebae from non-associated bacteria, and the attachment and ingestion of various bacterial strains by trophozoites of an axenized strain of E. histolytica were investigated. MATERIALS AND METHODS

Trophozoites of E. histolytica strain NIH:200 were axenically grown in TYI-S-33 medium according to the methods of Diamond et al. (8). Trophozoites were grown for 40 to 72 h (exponentially growing culture) and harvested by chilling in an ice water bath for 10 min to release those attached to the culture tubes. The trophozoites were washed twice in saline by low-speed centrifugation (600 x g, 5 min) and resuspended in saline to a final concentration of 106 amoebae per ml. Counting of the amoebae was done under a microscope with a hemacytometer. Growth of bacterial cells. Escherichia coli strain 7343 and Serratia marcescens were grown overnight at 37°C in a medium containing yeast extract (1%), 396

VOL. 36, 1982

UPTAKE OF BACTERIA BY E. HISTOLYTICA

peptone (0.5%; Difco Laboratories), NaCl (0.5%), and [14C]glucose (1 p.Ci/ml; specific activity, 329 mCi/ mmol; Radiochemical Centre, Amersham, England). Shigella flexneri cells were grown in Luria broth medium with the addition of Ca2+ (5 mM) and ["4C]glucose as described above. Bacteria were harvested by centrifugation at 9,000 x g for 10 min, washed three times with saline, and resuspended in saline to a concentration of 5 x 109 to 10 x 109 bacteria per ml. Staphylococcus aureus strain 52A2 (7) was grown in Difco antibiotic medium no. 3 with [14C]glucose as described above. The specific radioactivity of the bacteria obtained varied between 3 and 10 cpm per 104 bacteria. Preparation of specific antisera. Antisera were prepared against S. flexneri and S. aureus 52A2. Rabbits were inoculated in the footpads with 1010 glutaraldehyde-fixed bacteria. After 2 weeks they were bled, and the sera were separated and checked for their agglutination titers. A positive specific agglutination at a 1:100 dilution was considered satisfactory. Cleavage and removal of the Fc fragment of the immunoglobulins prepared against S. flexneri were done by pepsin digestion and staphylococcal protein A precipitation as described (16). Pretreatments of bacteria and amoebae. Precoating of bacteria or amoebae with concanavalin A (ConA) (Miles Yeda, Rehovoth, Israel) or with immune and nonimmune sera was done by preincubating the bacteria or amoebae (15 min at room temperature) at the optimal concentration which gave agglutination. After the reaction, the bacteria were washed two times with saline by centrifugation (9,000 x g, 5 min) and resuspended in saline to their original concentration. Washing of the amoebae was done by sedimentation at 600 x g for 10 min. Fixation of bacteria or amoebae was done with 0.25% glutaraldehyde in saline for 15 min at room temperature followed by two washings with 0.1 M glycine (to inactivate free aldehyde groups) and two washings with saline. Attachment of bacteria to amoebae. Bacteria (5 x 10' to 10 x 108) were incubated with trophozoites (106) in a total volume of 1 ml of saline solution for the desired time and at the desired temperature in glass tubes (13 by 100 mm). The incubation mixture was rotated at 150 rpm to avoid settling of the amoebae. The reaction was terminated by filling the tubes with saline (5 ml) and centrifuging the suspension at low speed (600 x g, 5 min). Under these conditions, the sedimented pellet contained mostly amoebae and the associated bacteria. To differentiate between bacteria that attached to the amoebae and those that did not, a discontinuous density gradient centrifugation with Percoll (Pharmacia Fine Chemicals, Inc., Sweden) was used. The Percoll stock solution (100%) was adjusted to an osmolality of 230 mosmol/kg of water by the addition of 1 part of lOx concentrated physiological salts solution to 9 parts of Percoll. Further dilutions of the Percoll stock solution were done with saline. The gradient was built on top of the sedimented pellet of amoebae and bacteria by layering 1 ml of 100% Percoll followed by 1 ml each of 80 and 60% Percoll, 2 ml of 50% Percoll, and 1 ml of 30% Percoll. After the discontinuous gradient was formed, centrifugation was carried out at 2,000 x g for 15 min

397

at room temperature. Under these conditions, bacteria layered between 80 and 60% Percoll, whereas amoebae layered between 50 and 30% Percoll. Only attached bacteria were observed in the amoeba layer. The amoeba layer was collected with a Pasteur pipette, suspended in a large volume of saline (30 ml), and sedimented again at 600 x g for 10 min to recover the amoebae and the labeled bacteria associated to them. The final pellet was resuspended in 0.5 ml of saline, and the labeled bacteria were counted in a Triton X-100 toluene-based scintillation fluid. Electron microscopy. Association of bacteria to amoebae was visualized by scanning of the amoeba surface with a scanning electron microscope. Samples were prepared after incubating the amoebae with bacteria for 15 min and separating the amoebae by Percoll gradient centrifugation as described above. Glutaraldehyde in saline was slowly added to a final concentration of 2%, and fixation took place during a 1-h incubation time. The samples were postfixed with OS04 (1%), dehydrated with ascending concentrations of alcohol, and dried by the critical point drying method with Freon. The preparations were coated with gold and observed with a Jeol 35 microscope. Ingestion of bacteria by the amoebae was observed in thin sections of the amoebae. Amoebae which had interacted with bacteria were embedded in Epon, and thin sections were prepared as described previously

(19).

Inhibitors. a-Methylmannoside (aMM) was from Sigma Chemical Co., St. Louis, Mo. Crab shell chitin (Sigma) was purified by refluxing in 1 M HCl for 2 h to extract protein contaminants. Partial acid hydrolysis of chitin and preparation of the N-acetylglucosamine oligosaccharides were done according to Rupley (18). Insoluble peptidoglycan was prepared from a lipoproteinless mutant of E. coli as previously described (12). Aggregated human immunoglobulin G (IgG) was a gift from Z. Bar-Shavit, Weizmann Institute, Rehovoth, Israel. N-Acetyl-D-galactosamine was from Pfanstiehl. Cytochalasin B was obtained from Sigma Chemical Co. RESULTS

Previous studies have shown that certain strains of gram-negative organisms possess a mannose-specific lectin-like substance (13) which mediates attachment to mannose residues on phagocytic cell surfaces (4). Since trophozoites of E. histolytica have been shown to contain receptors for ConA (21), it was logical to assume that adherence of bacteria to amoebae would be similar to that observed for macrophages (3). The attachment of the mannose-binding organism E. coli 7343 to E. histolytica trophozoites was found to be time and concentration dependent (Fig. 1 and 2). The average number of bacteria that attached per trophozoite at a high multiplicity of bacteria per amoeba was between 40 and 100. Standard experiments contained, therefore, a ratio of 1,000 bacteria per amoeba. When observed by scanning electron microscopy (Fig. 3), the attachment of bacteria to the amoebae did not appear to be uniform. Some trophozoites had large numbers of bacteria at-

398 0

-o

BRACHA, KOBILER, AND MIRELMAN

I

I

I

I

r

-

0

E a D

0 I

0 x

E C. 0 0

.._ 0 D

0 -Q

0

30 "4 5 60 min FIG. 1 Time dependence of attachment of "Clabeled E coli 7343 cells (103 bacteria per amoeba) to trophozoiites of E. histolytica strain NIH:200. Incubations were done in suspension at 37°C as described in the text. The specific activity of the radiolabeled bacteria vwas 1 cpm per 4 x 103 bacteria.

5 10 15

tached to them, whereas others had only a few bacteria. Similar differences of attachment were observed with a fluorescent microscope when fluorescein-labeled bacteria were coupled with amoebae (not shown). The attachment of E. coli 7343 was found to be markedly inhibited by the presence of aMM (Fig. 3 and 4). The inhibition was slightly more efficient when interactions with the bacteria were done at 5 rather than at 37°C. Bacteria attached at 37°C very rapidly and became irreversibly bound. Removal of the associated bacteria from the amoebae after 15 min of interaction by aMM was more efficient when the interaction was done at 5°C (Fig. 5), probably due to the lack of ingestion of the bacteria by the amoebae at the low temperature. Precoating of the mannose receptors of the amoebae by ConA also markedly blocked the attachment of E. coli cells (Fig. 4). Glutaraldehyde-fixed E. coli cells retained their mannose-binding activity, and they readily attached to glutaraldehyde-fixed amoebae (Table 1). This attachment was markedly blocked by aMM, and most bacteria which had been attached (-90%) could be removed from the amoebae by the inhibitor. Results very similar to those described above for E. coli 7343 have been obtained with S. marcescens (13), another mannose-binding organism (data not shown). The rapid irreversibility of the bacterial attachment to the amoebae indicated that after the

INFECT. IMMUN.

initial interaction the amoebae began ingesting the bacteria. This prevented their removal upon the addition of the aMM inhibitor. When no ingestion occurred (low temperature or glutaraldehyde-fixed amoebae), removal by aMM was more efficient. Evidence to support these results was obtained from electron microscopy observations of thin sections of amoebae. Amoebae that had come in contact with E. coli cells at 37°C contained a considerable number of phagocytized bacteria in their vacuoles (Fig. 6). On the other hand, when the interaction was done at 5°C, practically no bacteria could be seen inside the amoebae, although some could be seen adhered to the surface. The mannose-containing receptors present on the surface of the amoebae appear to regenerate after ingestion of the bacteria. Interaction of the amoebae with unlabeled E. coli cells for a period of 15 min at 37°C, followed by sedimentation of the amoebae and subsequent interaction with labeled bacteria, showed that considerable numbers of labeled bacteria attached to the amoebae during the second incubation period (Fig. 7). In experiments in which the incubations were done at 5°C, very few receptors remained free, and only small numbers of bacteria attached during the second incubation period. Controls in which

109 2xl0w Bacteria added /amoeba (106) FIG. 2. Attachment of various amounts of 14Clabeled E. coli strain 7343 to trophozoites (106) of E. histolytica strain NIH:200. Nonadherent bacteria were separated by a Percoll gradient centrifugation as described in the text. The final concentration of aMM was 1% (filled-in symbols). The reaction was carried out in suspension for 30 min at 5 or 37°C as described in the text (see also Fig. 1).

VOL. 36, 1982

UPTAKE OF BACTERIA BY E. HISTOLYTICA

399

E 0

2 L

cs

0

L)

- Q5 5 25 50 100

mg/mi

a-Methyl Mannoside

(mM) L

FIG. 4. Effect of increasing concentrations of aMM and precoating of trophozoites with ConA on the adherence of E. coli to amoebae. The reaction was

camed out for 15 min at 37°C. the attachment was. done by interacting labeled bacteria followed by unlabeled bacteria showed that at 37°C there was almost no displacement or loss of labeled bacteria by the unlabeled ones, whereas at 5°C, there was a slight decrease in the number of labeled bacteria that remained attached. A further indication that the amoebae internalized bacteria was obtained by subjecting the trophozoites with the adhered bacteria to mild detergent (Triton X-100, 0.2%). As was earlier found with the Dictyostelium amoeba (20), phagocytized bacteria became very sensitive to Triton X-100 (0.2%), whereas bacteria that remained attached to the surface (incubation done at 5°C) were not sensitive to the detergent. As

FIG. 3. Scanning electron microscope micrographs showing the adherence of E. coli 7343 to surfaces of trophozoites of E. histolytica strain NIH:200. The fine attachment of bacteria to filopodia can be clearly seen in the bottom photograph.

400

BRACHA, KOBILER, AND MIRELMAN

INFECT. IMMUN.

E

0

0

50

370

E

a

---

--

_-

--

I-2

,,'t

10

20

t

10

20

INCUBATION TIME (min)

FIG. 5. Removal of E. coli cells attached to trophozoites by the addition of 1% aMM 7.5 min after the initial interaction between them. Reactions were done at 5°C (left) and 37°C (right). Incubations in the presence of aMM are indicated by (---).

shown in Table 2, interaction of bacteria at low

internalized and surface-attached bacteria, such as detecting the surface-attached bacteria by coating them with fluorescent antibodies, did not give clear-cut or reproducible results, apparthe labeled bacteria became solubilized by the ently due to the fact that amoebae bind and detergent. Only intact amoebae were capable of ingest immunoglobulins and other serum composensitizing the bacteria. Control incubations of nents (6; see also below). bacteria with lysates of amoebae did not render Trophozoites of E. histolytica were found to them sensitive to the detergent. shun bacteria that did not have a mannoseOther techniques for distinguishing between binding capacity. Thus, various strains of Shitemperature did not cause them to become sensitive to solubilization by the detergent, whereas at 37°C, over 50% of the radioactive content of

TABLE 1. Attachment of bacteria to amoebae under various conditionsa Pretreatment of bacteria

Bacteria E. coli 7343

S. flexneri

None None None None None None None None

Incubation Pretreatment of amoebae temp (C) aobetm 0)

None None

Glutaraldehyde fixed Glutaraldehyde fixed ConA coated ConA coated Antibody coated Cytochalasin B

37 5 37 5 37 5 37 37

Bacteria attached control) ((% offcnrl

100 70 103 130 26 28 104 95

100 37 None Antibody coated 37 3 None None 490 37 ConA coated None 37 190 None ConA coated 4 5 None Antibody coated 19 37 Cytochalasin B Antibody coated 70 37 None Heat-inactivated antibody coated 70 37 Antibody coated + 0.01 M EDTA None 97 37 None Fab' dimer coated 54 37 Antibody coated Antibody coated a Incubations between 14C-labeled bacteria and trophozoites were done as described in the text. A ratio of 103 bacteria/amoeba was routinely used. ConA was added at a concentration of 1 mg/ml. Antibody coating was done with rabbit immune serum at a concentration of 100 ,ug/ml. This concentration gave optimal agglutination and adherence. Preparation of Fab' dimers was as described previously (16). Cytochalasin B was added at 5 ,ug/ml before the addition of the bacteria.

UPTAKE OF BACTERIA BY E. HISTOLYTICA

VOL. 36, 1982

401

TABLE 2. Effect of detergents on solubilization of bacteria after incubation with trophozoitesa Incubatio Incubation condition

(°C) Temp CQ

Without amoebae Lysate of amoebae Glutaraldehyde-fixed amoebae Intact amoebae Intact amoebae Without amoebae Lysate of amoebae

37

Time (min) 15

37 5 37 37

15 15 L5 30

Antibody-coated Shigella cells

Intact amoebae

37

30

50

ConA-coated Shigella cells

Intact amoebae

37

30

8

Bacteria Bacteria

E. coli 7343

S. flexneri

Detergent solubilization (% of total cpm) 7 9 11 24 42 3 10

Shigella cells 37 30 ConA-coated amoebae 5 a Incubations of "4C-labeled bacteria as well as pretreatments of amoebae were done as described in the text. After separation of the amoebae and associated bacteria on the Percoll gradient, they were suspended for 15 min at room temperature with 0.2% Triton X-100. The suspension was sedimented for 10 min at 10,000 x g, and the sedimented pellet and the clear soluble supernatant were counted for radioactivity. A control containing an amoeba lysate obtained by the addition of 0.2% Triton X-100 before the addition of bacteria was incubated under the same conditions.

gella (sonnei, flexneri, and dysenteriae), Bacteroides fragilis, Staphylococcus aureus, and Micrococcus luteus did not attach to the amoebae. S. flexneri cells were found to agglutinate with ConA. Precoating of these bacteria with ConA converted them into mannose-binding bacteria that readily attached to the amoebae, and this adherence was very sensitive to 1% aMM (>90% inhibition). Maximal attachment of ConA-precoated bacteria to the amoebae was achieved with concentrations of 1 mg of ConA and 5 x 108 bacteria per ml. Interestingly, their binding at 5°C was almost twice that at 37°C. Addition of aMM 30 min after the initial interactionr readily released most of the bacteria that attached at 5°C, whereas only 69% were removed at 37°C. In these experiments, the inhibitor (aMM) displaced both the ConA that bound to the amoebae as well as the ConA that attached to the bacteria. Examination of thin sections of amoebae that interacted with ConAcoated Shigella cells revealed that practically no bacteria were internalized by the amoebae, and all of the bacteria remained attached outside the cell. Moreover, detergent solubilization of the complex of amoebae and ConA-coated bacteria revealed that very few of the bacteria became solubilized by the detergent (Table 2). Amoebae that were precoated with ConA were also capable of binding uncoated S. flexneri cells. The number of bacteria attached was almost twice that of E. coli or ConA-precoated Shigella cells that bound to the same number of amoebae (Fig. 8). Addition of aMM 30 min after the initial interaction removed most of the adhered bacteria (>90%) at both 5 and 37°C.

As mentioned above, neither Shigella (that were not coated with ConA) nor S. aureus cells attached to the amoebae. On the other hand, Shigella or Staphylococcus cells that were preincubated with their respective rabbit immune sera adhered quite readily to the amoebae. Incubation of Shigella cells with human sera obtained from three shigellosis patients gave similar results. Incubation with heterologous or nonimmune rabbit or human serum did not lead to attachment (Fig. 9 and 10). Attachment of opsonized Shigella or Staphylococcus cells was dependent on the amount of the specific immune serum added (Fig. 9 and 10) and was very sensitive to temperature. In contrast to what was observed with the mannose-binding bacteria, practically no attachment was observed at 5°C (Table 3). Heat inactivation (57°C, 30 min) of the immune serum before coating of the Shigella cells or addition of EDTA (0.01 M) to the incubation mixtures had almost no effect on the adherence and ingestion of the bacteria (Table

1). Electron microscopy of thin sections showed that amoebae internalized numerous opsonized bacteria (Fig. 6b). Moreover, a considerable proportion of the attached bacteria were found to be detergent sensitive, indicating that they had been internalized by the amoebae (Table 2). The nature of the interaction between opsonized bacteria and the amoebae was also investigated. The attachment of the opsonized bacteria does not appear to be mediated by Fc receptors on the amoebae. No inhibition of attachment was observed when Shigella cells were precoated only with the Fab' dimers of anti-Shigella immunoglobulins after cleavage of the Fc frag-

402

BRACHA, KOBILER, AND MIRELMAN

INFECT. IMMUN.

V, .' i.

Ix

b

Y.

.[ .:

FIG. 6. Electron micrographs of thin sections of amoebae that interacted for 30 min at 37°C with (a) E. coli 7343 or (b) antibody-coated S. flexneri. Thin sections were prepared after Epon embedding as described in the text. Bacteria were almost exclusively found inside vacuoles. ment by pepsin digestion (16) and removal of the Fc fragment by staphylococcal protein A precip-

itation. Moreover, attachment of opsonized Shigella cells was only slightly blocked by aggregated human IgG, the well-known Fc receptor blocker (5). This same preparation of aggregated human IgG very efficiently inhibited (>90%) the ingestion of E. coli by mouse peritoneal macrophages (3, 4). Numerous conditions and materials were tested for their capability to interfere with the attachment of opsonized bacteria to the amoe-

bae. Precoating of the amoebae with their specific rabbit antisera did not affect the attachment of E. coli 7343 to the amoebae (Table 1). However, it markedly inhibited the adherence of opsonized Shigella cells (Table 3). No competition was detected by binding unlabeled E. coli 7343 to the amoebae before the attachment of opsonized Shigella cells. Chitin oligosaccharides and a purified preparation of peptidoglycan, which were previously shown by us to be potent inhibitors of the amoebic lectin activity (9), prevented the attachment of opsonized Shigella or Staphy-

UPTAKE OF BACTERIA BY E. HISTOLYTICA

VOL. 36, 1982

A. Agglutination

0

x

of Shigella

a

B.

403

Addition

to amoeba

a

.0

w 0

E 0

0 s0

0

1C

E 0

-

0

6

10 0

4

0)

0 .

10

5

25

15 25 35 L__J ~~~INCUBATION9 TlIME (min) FIG. 7. Regeneration of binding sites for "4C-labeled E. coli after incubation with unlabeled E. coli. All incubations were done at a ratio of 2,000 bacteria/ amoeba, and after 15 min, amoebae were sedimented and removed from the incubation mixture and subjected to a second incubation period of 15 min (arrow) The attachment of radiolabeled bacteria after each step was checked separately. The incubation period with "4C-labeled E. coli is indicated by ( ~) and that with unlabeled E. coli is indicated by (---). 15

35

5

m0. 0

_

2

iiIl

z E~~~W Qn Im mnS 88

Immune Serum Added Q*1I/Ml) FIG. 9. Effects of pretreatments of S. flexneri cells with increasing amounts of rabbit immune serum on the attachment of bacteria to trophozoites. Nonimmune serum (NIS) or precoating of the amoebae with immune serum did not lead to any attachment of Shigella cells. The agglutination of the cells by the immune serum is noted.

lococcus cells to the amoebae (Table 3 and Fig.

10). In addition, galactose, N-acetylgalactosamine, and lactose also had considerable inhibitory effects (Table 3). Addition of cytochalasin B A. ConA+ Shigella Observed Agglutination

_ +

B. ConA- Amoeba

+4++4+++

+

+++

4+

0

8 20

~In o

I

16

-

~~5040-

E 12

C)

0

8 L-

0 .2

(5 ,ug/ml) to the amoebae 5 min before interaction with the bacteria did not affect the binding of E. coli 7343 but markedly inhibited the attachment of opsonized Shigella cells (Table 1).

I02D-O

t

4

_~~~~1

U

0

m

I0

Concanavalin A added (pg/ml) FIG. 8. Effects of pretreatments of S. flexneri cells (left side) or trophozoites (right side) with various concentrations of ConA on the attachment of the bacteria to amoebae. The observed agglutination of cells caused by the different concentrations of the lectin is noted.

DISCUSSION Although it was known that E. histolytica ingests bacteria in axenic cultures (17, 24, 25), the molecular nature of the amoeba-bacterium associations had not been investigated. The major finding of the present work was the identification of the two possible mechanisms for attachment of bacteria to E. histolytica trophozoites, which precedes the phagocyotic process.

Previous studies had shown that certain strains of gram-negative bacteria possess a mannose-specific, lectin-like substance on their cell surfaces which mediates attachment to mannose residues on phagocytic cell surfaces (3, 4). Since trophozoites of E. histolytica had been shown to contain receptors for ConA (21), it was not surprising to find that adherence of such bacteria to the amoebae was mediated by mannose receptors on the amoeba surface. This adherence, when carried out at 37°C, was followed by considerable phagocytosis of the bacteria. After

404

INFECT. IMMUN.

BRACHA, KOBILER, AND MIRELMAN

ing of the Shigella cells with the multivalent, mannose-specific lectin ConA converted them into mannose-binding particles that avidly bound to the amoebae. A similar result was 0 1-~ obtained upon precoating of the amoeba receptors with ConA. The binding of the Shigella cells 0 under both of these conditions, however, did not Agglutination of Staphylococci xz lead to any significant phagocytosis, and no + E _ + ++ + bacteria could be observed by electron microscopy inside the amoebae in any of the thin sections. The lack of phagocytosis is probably 6 due to the extensive cross-linkage of the amoeba membrane that occurs at the rather high concentration of ConA (0.5 to 1.0 mg/ml) (1) that is 4-4 required for these attachments (Fig. 8). Phagocytic cells are also known to attach and co 2 ingest opsonized bacteria by virtue of recognizr,H~~~~~~~4/ ing the Fc region of the immunoglobulin molecule. The attachment between the receptor and the Fc region of the immunoglobulin has been 0.5 100 20 50 lO0 200 suggested to be mediated by a lectin-like activity NIS Immune serum Peptidoglycon Added(LI/mtl) (mg/ml) of the Fc receptor which recognized the carbohydrate side chain of the glycoprotein (2, 10, 20). FIG. 10. Effects of pretreatments of S. aureus Classic inhibitors of the interaction between strain 52A2 with increasing amounts of rabbit immune activated macrophages and opsonized bacteria serum on the attachment of Staphylococcus cells to the trophozoites. Nonimmune serum (NIS) did not are preparations of aggregated IgG in which the lead to significant attachment. Addition of peptidogly- carbohydrates seem to be better exposed and recognized by the receptor (5). Coating of the can to coated Staphylococcus cells (100 ,ug/ml) caused inhibition of attachment. The agglutination of the Shigella or Staphylococcus cells with their spebacteria by the immune serum is noted. cific antisera made possible the attachment and phagocytosis of the bacteria by the amoebae. This process was dependent on the concentraingestion of bacteria at 37°C, the mannose recep- tion of the immune serum and did not occur at tor capability of the amoebae seems to be par- low temperatures. tially restored (Fig. 7). Whether the mechanism The attachment of opsonized bacteria to the of this process is the result of unmasking of amoebae differed, however, from that of peritoneal macrophages by the fact that Fab' dimers mannose receptors that existed in the membrane obtained after pepsin digestion and staphylococor involves synthesis of new ones is not yet clear. Both adherence and subsequent phagocytosis of E. coli cells, however, were markedly inhibited by aoMM or by blocking of the amoeba TABLE 3. Inhibitory effects of various compounds mannose residues with ConA. In addition, no on the attachment of opsonized S. flexneri to E. interference with attachment of bacteria was histolyticaa observed when amoebae were precoated with Attachment rabbit antisera prepared against intact trophozo(% of Concn (mg/ml) Compound added ites. control) have like other cells, phagocytic Amoebae, 100 been shown to phagocytize latex beads (0.1-,um None 26 1 Insoluble peptidoglycan diameter) having hydrophobic surfaces (E. Calef N-acetylglucosamine 60 5 and C. Gitler, personal communication). Latex N-acetylgalactosamine 25 1 beads do not carry functional groups that can be Galactose 30 10 imagined to interact specifically with a cell sur- Lactose 15 20 face component. Thus, mere physical forces Mannose 98 20 90 20 Glucose seem to promote adhesion between nonspecific 73 0.2 receptors on the amoebae and the latex parti- Aggregated IgG 116 E. coli 7343 (103 bacteria/ cles. amoeba) A wide variety of bacterial strains, however, such as S. flexneri and S. aureus, which do not Incubations with 14C-labeled, antibody-coated possess the mannose-specific cell surface lectin Shigella cells and trophozoites were done as described (13) did not interact with the amoebae. Precoat- in the text. w

z

o Z

a

VOL. 36, 1982

UPTAKE OF BACTERIA BY E. HISTOLYTICA

cal protein A precipitation of immunoglobulin preparations (16) were still capable of mediating adherence of Shigella cells to the parasite. Furthermore, the aggregated human IgG preparation, which was very efficient in blocking the adherence and phagocytosis by macrophages, had very little effect on the binding of opsonized bacteria to the amoebae. Inhibition of attachment of opsonized Shigella cells was observed, however, upon precoating of the amoebae with immune antisera. The attachment of opsonized bacteria was also not mediated by binding of complement, as heat-inactivated immune sera or additions of EDTA, conditions known to prevent complement binding (23), had almost no effect. The attachment of opsonized bacteria, however, apparently requires microfilament and pseudopod formation, since, under conditions of low temperature (5°C) or in the presence of cytochalasin B, no interaction occurred. Since E. histolytica trophozoites had been previously shown to have a cell membraneassociated lectin with a specificity for N-acetylglucosamine (9), we suspected that it might be involved in binding to the immunoglobulin by recognizing the carbohydrate-containing residues which are present both in the Fc and Fab regions (light chains) (15). Chitin oligosaccharides and peptidoglycan, all potent inhibitors of erythrophagocytosis as well as of adherence of intact trophozoites to tissue-cultured mammalian cells (14), were found to markedly inhibit the attachment of opsonized Shigella or Staphylococcus cells to the amoebae (Table 3 and Fig. 10). The reason for the observed inhibition with galactose and N-acetylgalactosamine is not yet clear. These carbohydrates did not inhibit the adherence of intact trophozoites to tissue-cultured cells or the attachment of erythrocytes to the amoebae (14). N-Acetylgalactosamine, however, was found to inhibit the adherence of Chinese hamster ovary cells to the trophozoites of E. histolytica (J. I. Ravdin and R. L. Guerrant, J. Clin. Invest., in press). In summary, our studies indicate that E. histolytica, which inhabits the colon region of the human intestine, has a variety of options by which to attach and ingest local microbial flora. The mannose-containing cell surface components on the amoebae can serve as attachment sites for mannose-binding bacteria, and trophozoites can adhere and phagocytize opsonized bacteria by binding specific carbohydrate moieties present on side chains of the immunoglobulin that coats the bacteria (15). The binding and recognition mechanisms of the bacteria and amoebae function independently and do not seem to interfere with or affect one another. Investigations of the molecular contribution

405

which the ingested bacteria provides to the amoebae to augment their virulence are in progress. ACKNOWLEDGMENT This investigation was supported by a grant from the Rockefeller Foundation. LITERATURE CITED 1. Aley, S. B., W. A. Scott, and Z. A. Cohn. 1980. Plasma membrane of Entamoeba histolytica. J. Exp. Med. 152:391-404. 1. Anderson, C. L., and H. M. Grey. 1977. Solubilization and partial characterization of cell membrane Fc receptors. J. Immunol. 118:810-825. 3. Bar-Shavit, Z., R. Goldman, I. Ofek, N. Sharon, and D. Mirelman. 1980. Mannose-binding activity of Escherichia coli: a determinant of attachment and ingestion of the bacteria by macrophages. Infect. Immun. 29:417-424. 4. Bar-Shavit, Z., I. Ofek, R. Goldman, D. Mirelman, and N. Sharon. 1977. Mannose residues on phagocytes as receptors for the attachment of Escherichia coli and Salmonella typhi. Biochem. Biophys. Res. Commun. 78:455-460. 5. Basten, A., J. F. A. P. Miler, N. L. Warner, R. Abraham, E. Chia, and J. Gamble. 1975. A subpopulation of T-cells bearing Fc receptors. J. Immunol. 115:1159-1165. 6. Calderon, J., M. A. de Lourdes Munoz, and H. M. Acosta. 1980. Surface redistribution and release of antibodyinduced caps in Entamoeba. J. Exp. Med. 151:184-193. 7. Chatterjee, A. N. 1969. Use of bacteriophage-resistant mutants to study the nature of the bacteriophage receptor site of Staphylococcus aureus. J. Bacteriol. 98:519-527. 8. Diamond, L. S., D. R. Harlow, and C. C. Cunnick. 1978. A new medium for the axenic cultivation of Entamoeba histolytica and other Entamoeba. Trans. R. Soc. Trop. Med. Hyg. 72:431-432. 9. Kobiler, D., and D. Mirelman. 1980. A lectin activity in Entamoeba histolytica trophozoites. Infect. Immun. 29:221-225. 10. Koide, N., M. Nose, and T. Muramatsu. 1977. Recognition of IgG by Fc receptor complement: effects of glycosidase digestion. Biochem. Biophys. Res. Commun. 75:838-844. 11. Kurkinen, M., J. Wartiovaara, and A. Vaheri. 1978. Cytochalasin B releases a major surface associated glycoprotein, fibronectin, from cultured fibroblasts. Exp. Cell Res. 111:127-137. 12. Mett, H., R. Bracha, and D. Mirelman. 1980. Soluble nascent peptidoglycan in growing Escherichia coli cells. J. Biol. Chem. 255:9884-9890. 13. Mirelman, D., G. Altmann, and Y. Eshdat. 1980. Screening of bacterial isolates for mannose-specific lectin activity by agglutination of yeasts. J. Clin. Microbiol. 11:328331. 14. Mirelman, D., and D. Kobiler. 1981. Adhesion properties of Entamoeba histolytica, p. 17-30. Proceedings of the 80th Ciba Symposium on Adhesion of Microorganisms and Pathogenicity, London. 15. Niedermeier, W., T. Kirkland, R. T. Acton, and J. C. Bennett. 1971. The carbohydrate composition of immunoglobulin G. Biochim. Biophys. Acta 237:442-449. 16. Nisonoff, A., F. C. Wissler, and L. N. Lipman. 1960. Properties of the major component of a peptic digest of rabbit antibody. Science 132:1770-1771. 17. Phillips, B. P., and F. Gorstein. 1966. Effects of different species of bacteria on the pathology of enteric amebiasis in monocontaminated guinea pigs. Am. J. Trop. Med. Hyg. 15:863-868. 18. Rupley, J. A. 1964. The hydrolysis of chitin by concentrated hydrochloric acid and the preparation of low molecular weight substrates for lysozyme. Biochim. Biophys. Acta 83:245-255.

406

BRACHA, KOBILER, AND MIRELMAN

19. Spurr, A. R. 1969. A low viscosity epoxy resin embedding medium for electron microscopy. J. Ultrastruct. Res. 26:31-43. 20. Thornburg, R. W., J. F. Day, J. W. Baynes, and S. R. Thorpe. 1980. Carbohydrate-mediated clearance of immune complexes from the circulation. A role for galactose residues in the hepatic uptake of IgG-antigen complexes. J. Biol. Chem. 255:6820-6825. 21. Trissl, D., A. Martinez-Palomo, C. Arguello, M. de la Torre, and R. de la Hoz. 1977. Surface properties related to Concanavalin A induced agglutination. A comparative study of several Entamoeba strains. J. Exp. Med. 145:652-665.

INFECT. IMMUN. 22. Vogel, G., L. Thilo, H. Schwarz, and R. Steinhart. 1980. Mechanism of phagocytosis in Dictyostelium discoideum phagocytosis is mediated by different recognition sites and disclosed by mutants with altered phagocytic properties. J. Cell Biol. 86:456-465. 23. Waltraut, H. L., and V. Nussenzweig. 1968. Receptors for complement on leukocytes. J. Exp. Med. 128:991-1007. 24. Westphal, A. 1937. Betrachtungen und experimentelle Untersuchungen Zur Virulenz der Entamoeba histolytica beim Menschen. Arch. Schiffs Trophenhyg. 41:262-279. 25. Wittner, M., and R. M. Rosenbaum. 1970. Role of bacteria in modifying virulence of Entamoeba histolytica. Am. J. Trop. Med. Hyg. 19:755-761.