Damaging effects of Clostridium perfringens delta toxin on blood platelets and their relevance to ganglioside GM2

The lytic effect of Clostridium perfringens delta toxin was investigated on goat, human, rabbit, and guinea pig platelets. In contrast to erythrocytes from the latter three species, which are insensitive to the toxin, the platelets were equally lysed by the same amount of toxin. These results suggest the presence of GM2 or GM2-like ganglioside(s) as a specific recognition site of the toxin on platelet plasmic membrane as previously established for sensitive erythrocytes. Plasmic membrane damage of human platelets was evidenced by the release of entrapped alpha-[14C]aminoisobutyric acid used as a cytoplasmic marker. The specific binding of hemolytically active 125I-delta toxin by human and rabbit platelets was practically identical, dose dependent, and inhibitable by GM2. Labeled toxin was also bound by various subcellular organelles separated from rabbit platelets except the 5-hydroxytryptamine (5-HT)-containing dense bodies, suggesting the absence or inaccessibility of GM2 on the surface of the latter organelles. This result correlates with the low amounts of 5-[3H]HT liberated after platelet challenge with delta toxin whereas this mediator was massively liberated upon lysis by the sulfhydryl-activated toxin alveolysin. The levels of M and P forms of phenol sulfotransferase (PST), involved in 5-HT catabolism, were determined in human platelet lysates after challenge with delta toxin, alveolysin, and other disruptive treatments. The low PST-M activities detected after lysis by delta toxin suggest that this isoenzyme is very likely associated to dense bodies in contrast to PST-P which is cytoplasmic. Platelet lysis by the toxin allows easy separation of these organelles.     

Effects of monoclonal antibodies on α-staphylotoxin action against erythrocytes and model phospholipid membranes

It was found that one of twenty tested monoclonal antibodies (MABs) existed which drastically enhanced ability of Staphylococcus aureus α-tosin (ST) to both lysis of human erythrocytes and increase of planar phospholipid bilayer conductance more than 10 and 1000 times respectively. Other 19 MABs possessed only neutralized effect. The activation could only be observed if the activating MAB (AMAB) interacted with ST in solution but not in membrane. The one molecule of AMAB was able to activate approximately 2–4 molecules of ST. It was assumed that this activation was a result of the AMAB-induced transition of ST from a hydrophilic to an amphiphilic form. The activation could not be observed when the activity of AMAB/ST mixtures was tested on highly sensitive rabbit erythrocytes. All the tested MABs (including AMAB) were able to inhibit the ST-induced lysis of rabbit erythrocytes. The activating effects of AMAB on ST action in BLM and in human erythrocytes as well as their inhibiting influence on the ability of toxin to cause a lysis of rabbit erythrocytes indicate the presence of an ST-specific receptor on the membrane of rabbit erythrocytes.     

Binding of Clostridium perfringens 125I-labeled delta-toxin to erythrocytes

Hemolytically active, 125I-labeled delta-toxin from Clostridium perfringens was used to study the binding of this cytolysin to sheep, goat, human, rabbit, horse, mouse, and guinea pig erythrocytes. The extent of toxin binding was correlated with the known hemolytic specificity of the toxin. Detailed studies of the binding were carried out on sheep erythrocytes which showed the highest sensitivity to lysis by delta-toxin. Simultaneous determination of toxin binding and release of intracellular 86Rb+ and hemoglobin suggested that toxin binding and membrane damage were separate sequential events. Toxin binding was rapid (2-5 min) and temperature-dependent. The extent of binding was temperature-independent. Binding was saturable, specific, relatively tight (Ka = 4.4 X 10(8) M-1) and largely irreversible. A single type of binding site (7,000/sheep erythrocyte) was found. Cell-bound toxin was extractable by chaotropic ions. Preincubation of the toxin with N-acetylgalactosaminyl-[N-acetylneuraminyl]-galactosylglucosylceramide (GM2 ganglioside) inhibited both binding and hemolysis. Toxin binding was affected by pretreatment of sheep erythrocytes with pronase but not with trypsin or chymotrypsin. Cell treatment with neuraminidase prevented toxin binding by 30%. Preincubation of the toxin with specific immune sera blocked its binding on target cells. It is suggested that GM2 ganglioside, a more complex membrane component containing this glycolipid or a structurally related molecule is the binding site for delta-toxin on the surface of sensitive erythrocytes.     

Adenylate cyclase toxin from Bordetella pertussis. The relationship between induction of cAMP and hemolysis.

Bordetella pertussis produces a calmodulin-activated adenylate cyclase (AC) that exists in several forms. Only one form of AC, of apparent 200 kDa, is a toxin that penetrates eukaryotic cells and generates uncontrolled levels of intracellular cAMP. Recombination studies in transposon Tn5-insertion mutants of B. pertussis and amino acid sequence homology with alpha-hemolysin of Escherichia coli suggested that AC toxin may also have a hemolytic activity. Here, we demonstrate that only the toxic form of B. pertussis AC possesses hemolytic activity. Immunoblotting of membranes from sheep erythrocytes throughout the process of cell lysis detects the presence and accumulation of only the 200-kDa form of B. pertussis AC. cAMP generation induced by AC toxin in sheep erythrocytes is immediate whereas appearance of hemolysis is delayed by about 1 h and requires a higher level of AC toxin activity. Addition of exogenous calmodulin to sheep erythrocyte incubation medium potentiates the hemolytic activity of AC toxin but blocks cAMP generation. Extracellular Ca2+ at mM concentrations is absolutely required for cAMP generation but not for hemolysis. However, binding of AC toxin to sheep erythrocytes in the absence of exogenous Ca2+ followed by reincubation of cells in a toxin-free buffer containing Ca2+ leads to an immediate rise in intracellular cAMP. Human erythrocytes bind AC toxin and generate cAMP but are resistant to lysis. These results show that binding of AC toxin to erythrocytes can cause both cAMP generation and hemolysis or only one of these depending on conditions applied and cell type used.     

Binding and aggregation of the 25-kilodalton toxin of Bacillus thuringiensis subsp. israelensis to cell membranes and alteration by monoclonal antibodies and amino acid modifiers

The 25-kilodalton toxin of Bacillus thuringiensis subsp. israelensis binds irreversibly to Aedes albopictus cells, Choristoneura fumiferana cells, and erythrocytes. The binding to cells increased with both toxin concentration and time and when the cells were first preincubated with unlabeled toxin. Binding data indicated a two- to threefold increase in the rate of binding after the amount of the membrane-bound toxin reached approximately 3.5 fmol/3 x 10(5) A. albopictus cells or 3.3. fmol/2 x 10(5) C. fumiferana cells. When this level of bound toxin was reached, the toxins also began forming aggregates at the cell membrane. The toxin aggregates were extracted with 10% Triton X-100 and separated from the monomers with a 5 to 20% sucrose density gradient. The toxin aggregates isolated from A. albopictus and C. fumiferana cell membranes were ca. 400 kilodaltons, while those isolated from human erythrocytes were significantly smaller. The proportion of the toxin found in aggregate form increased rapidly with the amount of toxin bound; however, the molecular size of the aggregates remained constant. Eleven monoclonal antibodies raised against the native form of the toxin blocked 80 to 97% of the toxin binding to cells. The epitope of one of these monoclonal antibodies was mapped to a domain which included the cysteine, suggesting the importance of the domain around this amino acid to binding. Toxin binding and cell lysis were also inhibited by treating the toxin with HgCl2, further indicating the importance of the C-terminal hydrophobic cysteine-containing domain in cytolytic activity of the 25-kilodalton protein.     

Binding and aggregation of the 25-kilodalton toxin of Bacillus thuringiensis subsp. israelensis to cell membranes and alteration by monoclonal antibodies and amino acid modifiers.

The 25-kilodalton toxin of Bacillus thuringiensis subsp. israelensis binds irreversibly to Aedes albopictus cells, Choristoneura fumiferana cells, and erythrocytes. The binding to cells increased with both toxin concentration and time and when the cells were first preincubated with unlabeled toxin. Binding data indicated a two- to threefold increase in the rate of binding after the amount of the membrane-bound toxin reached approximately 3.5 fmol/3 x 10(5) A. albopictus cells or 3.3. fmol/2 x 10(5) C. fumiferana cells. When this level of bound toxin was reached, the toxins also began forming aggregates at the cell membrane. The toxin aggregates were extracted with 10% Triton X-100 and separated from the monomers with a 5 to 20% sucrose density gradient. The toxin aggregates isolated from A. albopictus and C. fumiferana cell membranes were ca. 400 kilodaltons, while those isolated from human erythrocytes were significantly smaller. The proportion of the toxin found in aggregate form increased rapidly with the amount of toxin bound; however, the molecular size of the aggregates remained constant. Eleven monoclonal antibodies raised against the native form of the toxin blocked 80 to 97% of the toxin binding to cells. The epitope of one of these monoclonal antibodies was mapped to a domain which included the cysteine, suggesting the importance of the domain around this amino acid to binding. Toxin binding and cell lysis were also inhibited by treating the toxin with HgCl2, further indicating the importance of the C-terminal hydrophobic cysteine-containing domain in cytolytic activity of the 25-kilodalton protein.     

Biological activity of sea anemone proteins: II. Cytolysis and cell line toxicity

Potent cytolytic activity was exhibited by proteins extracted from three sea anemones viz. Heteractis magnifica, Stichodactyla haddoni and Paracodylactis sinensis by affecting the red blood corpuscles (RBC) and the mouse fibroblast cell line (L929) and leukemia cell line (P388). Crude toxin of all the three anemone species induced spontaneous hemolysis of chicken, goat and human erythrocytes. The crude toxin of H. magnifica (0.98 mg/ml) elicited hemolysis at levels of 4096, 512 and 4096 HU (hemolytic unit) in chicken, goat and human erythrocytes respectively. Subsequently, the crude toxin of S. haddoni (0.82 mg/ml) exhibited a hemolytic activity of 256, 128 and 512 HU and that of P. sinensis (0.60 mg/ml) had a hemolytic activity of 128, 4096 and 512 HU. Most of the partially purified proteins of these anemones also exhibited the activity against the three different erythrocytes. The viability of L929 and P388 was adversely affected on adding the crude toxins. The symptoms of toxicity shown by the cells were rounding, lysis and detachment from the substratum. These effects were the least in S. haddoni, as compared to those the crude toxins of the other two species. Inhibition of growth of L929 exhibited by the toxin of the three species ranged between 61.08 and 93.38%. Similarly, inhibition of the growth of P388 ranged between 51.32 and 86.16%. The present investigation reveal the cytotoxic nature of anemone toxins.     

Relationship between haemolytic and sphingomyelinase activities in a partially purified beta-like toxin from Staphylococcus schleiferi

beta-Toxins of staphylococcal species possess dual activity in that they can both lyse erythrocytes (by 'hot-cold' lysis) and catalyse hydrolysis of membrane-associated sphingomyelin. However, the precise relationship between these two activities has not been extensively studied. We have partially purified a beta-like toxin from culture supernatants of Staphylococcus schleiferi N860375 which exhibits both 'hot-cold' lysis of erythrocytes and neutral sphingomyelinase activities. This toxin has a strong preference for sheep erythrocytes, the membranes of which are rich in sphingomyelin. Kinetic analysis suggests that haemolysis and sphingomyelinase activities are very closely associated obeying identical Michaelis-Menten kinetics. However, pre-treatment with antibodies to Staphylococcus aureus beta-toxin, Ca(2+), dithiothreitol and phenylmethylsulfonyl fluoride appear to inhibit sphingomyelinase activity significantly more strongly than haemolysis while Mg(2+) activates sphingomyelinase activity more strongly than haemolysis. We attribute these effects to differences in binding properties in the two assays. Micropurification by both sphingosylphosphocholine-agarose affinity chromatography and preparative electrophoresis revealed that the 34-kDa toxin associates non-covalently with individual proteins.     

Interaction of Staphylococcal α-Toxin with Artificial and Natural Membranes

Comparison of hemolytic activity and chromate-releasing activity of partially purified preparations of staphylococcal alpha-toxin indicated the presence of a lytic factor other than alpha-toxin. This lytic release factor (RF) was isolated from the preparations and was shown to be active against both lipid spherules and erythrocytes. Heat-purified alpha-toxin (HP alpha-toxin) disrupted spherules, with the formation of fragments which always showed the presence of ring structures similar in dimensions (ca. 90 A) to pure alpha 12S-toxin. The interaction of HP alpha-toxin with spherules was accompanied by loss of hemolytic activity and adsorption of toxic protein. The alpha 12S-toxin, although only weakly hemolytic, was shown to be lytic for spherules. An alpha 12S-free toxin rapidly disrupted spherules, with formation of fragments with attached rings similar in dimensions to the alpha 12S molecule. Lipid monolayer experiments showed that HP alpha-toxin could penetrate lipid monolayers by virtue of a hydrophobic interaction. Effects of HP alpha-toxin on rabbit and human erythrocyte ghosts were similar to its effects on spherules, in that rings appeared on membrane fragments. Toxin-lysed rabbit erythrocytes showed similar rings on the resulting membrane fragments. However, rings were not seen on toxin-treated rabbit erythrocytes in the prelytic lag phase; this result and the fact that human erythrocytes are largely insensitive to alpha-toxin were interpreted as evidence against a lytic mechanism involving ring formation as the primary event. Rings were interpreted as toxin polymers similar to alpha 12S molecules, formed from specifically orientated active toxin molecules at the surface of lipid structures. Possible mechanisms for toxin lysis of spherules and erythrocytes are discussed.     

Streptavidin-induced lysis of homologous biotinylated erythrocytes. Evidence against the key role of the avidin charge in complement activation via the alternative pathway

It is shown that non-covalent attachment of streptavidin, as well as of avidin, to biotinylated human erythrocytes induces homologous hemolysis by complement. Rabbit antiserum against human C3 is found to inhibit the lysis specifically as compared with non-immune rabbit serum. Efficiency of lysis inhibition is greater for avidin- and streptavidin-induced lysis of biotinylated human erythrocytes than for antibody-sensitized sheep erythrocytes. In contrast to positively charged avidin (pI 11), streptavidin is a neutral protein. Hence, hemolysis of streptavidin-carrying erythrocytes is inconsistent with the suggestion on the crucial role of avidin charge in lysis. Membrane alterations (cross-linking and clusterization of biotinylated components) induced by avidin (streptavidin) seem to be a more plausible explanation for the lysis.     

Membrane damage by Cerebratulus lacteus cytolysin A-III. Effects of monovalent and divalent cations on A-III hemolytic activity

The effects of monovalent and divalent cations on the hemolytic activity of Cerebratulus lacteus toxin A-III were studied. The activity of cytolysin A-III is remarkably increased in isotonic, low ionic strength buffer, the HC50 (the toxin concentration yielding 50% lysis of a 1% suspension of erythrocytes after 45 min at 37 degrees C) being shifted from 2 micrograms per ml in Tris or phosphate-buffered saline to 20-30 ng per ml in sucrose or mannitol buffered with Hepes, corresponding to a 50-100-fold increase in potency. On the contrary, hemolytic activity decreases progressively as the monovalent cation concentration in the medium increases for Na+, K+, or choline salts. The divalent cations Ca2+ and Zn2+ likewise inhibit the cytolysin A-III activity, but more strongly than do the monovalent cations specified above. Zn2+ at a concentration of 0.3 mM totally abolishes both toxin A-III-dependent hemolysis of human erythrocytes and toxin-induced leakage from liposomes. The observation of similar effects in both natural membranes and artificial bilayers suggests an effect of Zn2+ on the toxin A-III-induced membrane lesion, especially since Zn2+ does not alter binding of the cytolysin. The dose-response curve for toxin A-III exhibits positive cooperativity, with a Hill coefficient of 2 to 3. However, analysis of toxin molecular weight by analytical ultracentrifugation reveals no tendency to aggregate at protein concentrations up to 2 mg per ml. These data are consistent with a post-binding aggregational step which may be affected by the ionic strength of the medium.     

A reexamination of the role of clusterin as a complement regulator.

Clusterin is a highly conserved glycoprotein which has been proposed to protect host cells against complement-mediated cytolysis. We tested the hypothesis that clusterin is a complement regulator using erythrocytes and cells which had been stably transfected with a membrane-anchored form of clusterin as targets for complement-mediated cytolysis. Clusterin gave dose-dependent protection of antibody-coated sheep erythrocytes against complement-mediated lysis by diluted normal human serum. There was a linear relationship between the concentration of clusterin giving 50% protection and the concentration of serum; extrapolation of this to the case of undiluted human serum showed that a clusterin concentration at least two orders of magnitude greater than its physiological plasma concentration would be needed to confer protection against complement-mediated cytolysis under physiological conditions. Physiological concentrations of clusterin did not protect rabbit erythrocytes against alternative complement pathway-mediated lysis using dilute human serum. Exogenous clusterin had no effect on lysis of human erythrocytes triggered by the addition of inulin to autologous human serum. Induction of cell-surface clusterin expression by L929 (murine fibroblast) cells which had been stably transfected with cDNA for human clusterin linked to DNA coding for the 44 C-terminal amino acid residues of CD55 did not protect the cells against complement-mediated lysis by either normal or clusterin-depleted human serum. These data suggest that clusterin may not be a physiologically relevant regulator of complement activation.     

Clostridium perfringens epsilon toxin binds to erythrocyte MAL receptors and triggers phosphatidylserine exposure

Epsilon toxin (ETX) is a 33‐kDa pore‐forming toxin produced by type B and D strains of Clostridium perfringens. We previously found that ETX caused haemolysis of human red blood cells, but not of erythrocytes from other species. The cellular and molecular mechanisms of ETX‐mediated haemolysis are not well understood. Here, we investigated the effects of ETX on erythrocyte volume and the role of the putative myelin and lymphocyte (MAL) receptors in ETX‐mediated haemolysis. We observed that ETX initially decreased erythrocyte size, followed by a gradual increase in volume until lysis. Moreover, ETX triggered phosphatidylserine (PS) exposure and enhanced ceramide abundance in erythrocytes. Cell shrinkage, PS exposure and enhanced ceramide abundance were preceded by increases in intracellular Ca²⁺ concentration. Interestingly, lentivirus‐mediated RNA interference studies in the human erythroleukaemia cell line (HEL) cells confirmed that MAL contributes to ETX‐induced cytotoxicity. Additionally, ETX was shown to bind to MAL in vitro. The results of this study recommend that ETX‐mediated haemolysis is associated with MAL receptor activation in human erythrocytes. These data imply that interventions affecting local MAL‐mediated autocrine and paracrine signalling may prevent ETX‐mediated erythrocyte damage.     

Interaction of thermostable direct hemolysin of Vibrio parahaemolyticus with human erythrocytes.

The interaction between thermostable direct hemolysin produced by Vibrio parahaemolyticus WP-1 and human erythrocytes was studied. The lysis of human erythrocytes by the hemolysin was dependent of temperature and no hemolysis occurred at low temperature (0-4 C), but the hemolysin was adsorbed on human erythrocytes even at low temperature. No hemolysis was observed when antihemolysin antiserum was mixed with the hemolysin and human erythrocytes at zero time. On the other hand, lysis of the cells by hemolysin was not completely inhibited when the antiserum was added during the lag time and the inhibitory effect decreased with delay in the time of addition of antiserum. The inhibitory effect of the antiserum decreased with increase in the incubation temperature, increase in the concentration of divalent cations, and decrease in pH. These results suggest that lysis of human erythrocytes by the hemolysin is at least a two-step process consisting of adsorption of the hemolysin to human erythrocytes and the step(s) following adsorption.     

Effects of temperature and bovine serum albumin on lysis of erythrocytes induced by dilauroylglycerophosphocholine and didecanoylglycerophosphocholine.

The effects of the incubation temperature and bovine serum albumin on hemolysis induced by short-chain phosphatidylcholine were examined. The rate of hemolysis of human, monkey, rabbit, and rat erythrocytes by dilauroylglycerophosphocholine showed biphasic temperature-dependence: hemolysis was rapid at 5-10 degrees C and above 40 degrees C, but slow at around 25 degrees C. In contrast, the rate of lysis of cow, calf, sheep, pig, cat, and dog erythrocytes did not show biphasic temperature-dependence, but increased progressively with increase in the incubation temperature. Bovine serum albumin increased the hemolysis of human erythrocytes induced by dilauroylglycerophosphocholine or didecanoylglycerophosphocholine: it shortened the lag time of lysis and reduced the amount of phosphatidylcholine required for lysis. A shift-down of the incubation temperature from 40 to below 10 degrees C also shortened the lag time of lysis of human erythrocytes induced by dilauroylglycerophosphocholine and reduced the amount of phosphatidylcholine required for lysis.     

Comparison of photodynamic effect with respect to human and rabbit erythrocytes

Parameters of photoinduced lysis are studied in human and rabbit erythrocytes (photosensitizer-Radachlorine, the light source Shatl, λ = 633 nm). The higher sensitivity to irradiation is revealed in rabbit erythrocytes. Treatment of erythrocytes with trypsin showed the surface proteins in human cells to produce a protective action. Trypspnization of rabbit erythrocytes produced the opposite effect-the rate of photohemolysis increased. Results of the study indicate the differences in sensitivity to the photoinduced lysis of erythrocytes of different species and participation of erythrocyte proteins in the effect of photohemolysis.     

Effect of modification of tyrosine residues in cytotoxin-1 from Indian cobra venom. (Naja naja naja)

The role of tyrosine residues in the biological activity of cytotoxin-1 was evaluated using N-bromo succinimide. N-bromo succinimide effected the oxidation of tyrosine residues in cytotoxin-1 with an increase in absorption at 260 nm. N-chloro succinimide was ineffective in the oxidation of tyrosine residues in the toxin. Oxidation of a single tyrosine residue (at 3.50 equivalents of N-bromo succinimide/mole of the toxin) resulted in complete loss of lethal activity of the toxin. The lytic activity of the toxin (lysis of erythrocytes) remained uneffected even after three of the four tyrosine residues in the toxin were oxidised.     

Binding of a radiolabeled sea anemone cytolysin to erythrocyte membranes

Stichodactyla helianthus cytolysin III, a 17 kDa basic polypeptide isolated from a Caribbean sea anemone, is one of the most potent hemolysins yet found in a living organism. This toxin has been reported to form new ion channels in artificial lipid bilayer membranes. The ability of this toxin to attack cell membranes is greatly enhanced by the presence of sphingomyelin. In order to investigate the mechanism by which the cytolysin causes cell lysis, we have prepared a highly active [3H]cytolysin derivative by reductive methylation with sodium cyanoborohydride and [3H]formaldehyde. A dimethylated toxin derivative was used to investigate the basis for the differential lytic activity of this polypeptide upon erythrocytes from six mammalian species. Using both direct [3H]toxin binding and indirect (Thron method) binding techniques, we found that the interspecies differences are due to variable membrane susceptibilities toward the bound toxin, rather than to differences in membrane affinity for the toxin. Similarly, we showed the enhanced lytic activity of the toxin for rat erythrocytes at elevated pH to be caused by enhanced activity of the bound toxin.     

Assembly of Staphylococcus aureus leukocidin into a pore-forming ring-shaped oligomer on human polymorphonuclear leukocytes and rabbit erythrocytes.

Staphylococcal leukocidin consists of two separate proteins, LukS and LukF, which cooperatively lyse human and rabbit polymorphonuclear leukocytes and rabbit erythrocytes. Here we studied the pore-forming properties of leukocidin and the molecular architecture of the leukocidin pore. (1) Leukocidin caused an efflux of potassium ions from rabbit erythrocytes and swelling of the cells before hemolysis. However, ultimate lysis of the toxin-treated swollen erythrocytes did not occur when polyethylene glycols with hydrodynamic diameters of > or = 2.1 nm were present in the extracellular space. (2) Electron microscopy showed the presence of a ring-shaped structure with outer and inner diameters of 9 and 3 nm, respectively, on leukocidin-treated human polymorphonuclear leukocytes and rabbit erythrocytes. (3) Ring-shaped structures of the same dimensions were isolated from the target cells, and they contained LukS and LukF in a molar ratio of 1:1. (4) A single ring-shaped toxin complex had a molecular size of 205 kDa. These results indicated that LukS and LukF assemble into a ring-shaped oligomer of approximately 200 kDa on the target cells, forming a membrane pore with a functional diameter of approximately 2 nm.     

CD16 on human gamma delta T lymphocytes: expression, function, and specificity for mouse IgG isotypes.

We examined the expression, the signal transduction capacity and mouse IgG-isotype specificity of CD16 on human gamma delta T cells. CD16 is expressed by the majority of gamma delta T cells in peripheral blood and by part of the gamma delta T cell clones. The amount of CD16 expressed on gamma delta T cell clones varied considerably with passaging of the cells, but was always significantly less than on freshly isolated gamma delta T cells. Like CD16 on CD3- CD16+ natural killer (NK) cells, CD16 on gamma delta T cells can act as an activation site triggering cytotoxic activity. CD16+ gamma delta T cell clones exerted antibody-dependent cellular cytotoxicity (ADCC) which could be blocked by anti-CD16 mAb. ADCC activity of gamma delta T cell clones was also inhibited by anti-CD3 mAb, suggesting a functional linkage between the CD16 and CD3 activation pathways. MAb directed against CD16 induced lysis of Fc gamma R+ target cells by CD16+ gamma delta T cell clones. The mouse IgG-isotype specificity of CD16 on gamma delta T cells was analyzed using isotype switch variants of a murine anti-glycophorin A mAb in EA rosette assays, and was found to be identical to that of CD16 on CD3- CD16+ NK cells, i.e., highest affinity for mIgG2a, intermediate affinity for mIgG2b, and undetectable binding of mIgG1-sensitized erythrocytes. CD16 was partly modulated from the cell surface of both gamma delta T cells and NK cells after rosette formation with mIgG2a-sensitized erythrocytes, indicating that the rosette formation was indeed mediated via the CD16 molecule.