Surface phagocytosis of Staphylococcus epidermidis and Escherichia coli by human neutrophils: serum requirements for opsonization and chemiluminescence

Abstract We examined the serum requirements for surface phagocytosis of Staphylococcus epidermidis and Eschericia coli and for the subsequent chemiluminescent response of human neutrophils. Substantial surface phagocytosis of S. epidermidis occured in the absence of opsonins, although the presence of 10% pooled or heat-inactivated serum significantly increased phagocytosis. There was no significant difference between these opsonins, indicating that surface phagocytosis of S. epidermidis did not require complement, Unopsonized E. coli were not as readily phagocytized as S. epidermidis (33% versus 57%). In contrast to S. epidermidis optimal phagocytosis of E. coli required complement as 10% heat inactivated donor serum (HHS) was significantly less effective as an opsonin than 10% pooled healthy donor serum (PHS). The time kinetics for phagocytosis of each organism were similar, with most of the phagocytosis iluminescent response of neutrophils produced discrepant results. Maximal chemiluminescence was observed when neutrophils were stimulated with bacteria opsonized in PHS. The response to HHS-opsonized bacteria was less, and chemiluminescence to unopsonized bacteria was only marginally higher than the control, even though there was relatively good phagocytosis. These results define the opsonic requirements for surface phagocytosis of S. epidermidis and E. coli and indicate that although complement may not be required for phagocytosis, it is necessary for generation of a maximal oxidative burst, and thus may be essential for efficient intracellular killing.     

Interaction of Candida albicans with neutrophils: effect of phenotypic changes in yeast cell-surface composition

The susceptibility of four strains of Candida albicans to phagocytosis and intracellular killing by rabbit peritoneal neutrophils was investigated. Two of the strains, isolated from active infections, were known to synthesize a surface layer of mannoprotein fibrils in response to growth on 500 mm-galactose; the other strains, from asymptomatic carriers, lacked this capability. The presence of serum opsonins greatly enhanced phagocytosis of all four strains and, following opsonization, phagocytosis of an infective strain was equally rapid after growth on either 500 mm-galactose or 50 mm-glucose. In the absence of opsonins, galactose-grown infective strains were phagocytosed faster than either glucose-grown infective strains or galactose-grown carrier strains. These differences in phagocytic uptake were paralleled by differences in neutrophil chemiluminescence response. Intracellular killing of galactose-grown infective strains was only half that of glucose-grown infective strains or galactose-grown carrier strains after incubation for 60 min. Pretreatment of neutrophils with extracellular polymeric material, which contains the surface fibrils, completely inhibited intracellular killing. These results indicate that production of the fibrillar layer promotes yeast virulence by increasing resistance to intracellular killing, although it may enhance phagocytosis in locations where opsonic activity is poor.     

Metabolic and immunological effects of endotoxin in the plaice, Pleuronectes platessa L.

In an effort to identify factors contributing to the resistance of fish to endotoxin toxicity, the metabolic effects of an intraperitoneal injection of a Boivin preparation of E. coli lipopoly-saccharide (LPS) were studied in plaice over 4 days. Significant changes were found in serum concentrations of glucose, cortisol and non-esterified fatty acids but not in total or free cholesterol. The effect of injected LPS on the phagocytic capacity of kidney neutrophils was examined because of the possible use of LPS to promote non-specific immunity. There was no enhancement of the chemiluminescent response to the phagocytosis of unopsonized bacteria by neutrophils from plaice 24 h after LPS treatment.     

Human neutrophil response to recombinant neisserial Opa proteins

Interactions of human neutrophils with recombinant Escherichia coli expressing gonococcal outer membrane Opa proteins were examined using chemiluminescent and biological assays. Seven opa loci from Neisseria gonorrhoeae MS11 4.8 were expressed as beta-lactamase-Opa fusion proteins that contained all but the mature N-terminal amino acid of the full-length Opa protein fused to three N-terminal amino acids derived from the mature beta-lactamase. The Opa fusion proteins were exported and assembled in the outer membrane of E. coli in a manner similar to that of Opa in N. gonorrhoeae, as evaluated by antibody binding and in situ proteolytic cleavage. All fusion proteins exhibited the characteristic heat-modifiable migration in SDS-polyacrylamide gel electrophoresis that typifies Opa proteins of neisseriae. Opa fusion proteins conferred on E. coli the ability to stimulate a chemiluminescent response from human neutrophils in the absence of antibody or complement. The nature of the response in terms of chemiluminescence, phagocytosis, and killing was in all cases analogous to that seen using N. gonorrhoeae expressing the equivalent Opa protein. Neither E. coli nor gonococci expressing OpaA elicited a response from neutrophils. Use of E. coli expressing Opa fusions should be useful in defining their biological activities and pathogenic roles.     

Nonspecific resistance in germ-free and E. coli monocontaminated miniature piglets]

Phagocytic activity of leukocytes, as well as the complement, properdin, and lysozyme levels in the blood serum of miniature piglets, germfree and monocontaminated with E. coli 055 and E. coli 083, were studied. E. coli 055 phagocytosis was decreased in the presence of autologous serum and complement and increased under the effect of specific opsonins (antibodies to E. coli 055). Complement, properdin, and lysozyme levels were decreased in the germfree, in comparison with conventional animals. In the E. coli contaminated piglets properdin and complement production was stimulated most, and lysozyme formation--less. No antibodies to E. coli 055 were revealed in monocontaminated piglets. The highest lysozyme levels were found in the ex-germfree animals, this indicating the participation of factors other than E. coli contamination in lysozyme stimulation. It is concluded that microbial contamination played an important role in the development of cellular and humoral factors of the organism resistance.     

The microbial factor in the chemiluminescence of the peripheral blood neutrophils of patients with suppurative surgical infection

The luminol-dependent chemiluminescence of neutrophils in the peripheral blood of 30 healthy adults and 39 patients with the local and generalized forms of purulent infection was studied. Nonstimulated chemiluminescence and the index of chemiluminescence stimulation in the presence of opsonized Staphylococcus aureus added in vitro were determined. The former characteristic was found to be directly and the latter one, inversely related to the concentration of S. aureus, Escherichia coli and Candida albicans, but not E. epidermidis, Pseudomonas aeruginosa or Citrobacter, in the primary focus. At the microbial concentration exceeding 10(4) cells/g of tissue, the former characteristic was essentially higher than the level of chemiluminescence in healthy persons. With the improvement of the general state of the patients and in the absence of microorganisms in the wound as the result of complex treatment this characteristic decreased to values comparable with the reaction of neutrophils in healthy persons.     

Lucigenin chemiluminescence as a probe for measuring reactive oxygen species production in Escherichia coli

Addition of oxygen to whole cells of Escherichia coli suspended in the presence of the chemiluminescent probe bis-N-methylacridinium nitrate (lucigenin) resulted in a light emission increase of 200% of control. Addition of air to cells showed a chemiluminescent response far less than the response to oxygen. The redox cycling agents paraquat and menadione, which are known to increase intracellular production of O2- and H2O2, were also found to cause a measurable increase in lucigenin chemiluminescence in E. coli cells when added at concentrations of 1 and 0.1 mM, respectively. The oxygen-induced chemiluminescent response was not suppressed by extracellularly added superoxide dismutase or catalase. Further, the lucigenin-dependent chemiluminescent response of aerobically grown E. coli to oxygen was significantly greater than that of cells grown anaerobically. Heat-killed cells showed no increase in chemiluminescence on the addition of either oxygen, paraquat, or menadione. These results show that lucigenin may be used as a chemiluminescent probe to demonstrate continuous intracellular production of reactive oxygen metabolites in E. coli.     

Mannose-binding lectin (MBL) facilitates opsonophagocytosis of yeasts but not of bacteria despite MBL binding

Mannose-binding lectin (MBL) is a serum protein of the innate immune system. After binding to a microorganism, MBL in complex with MBL-associated serine proteases activates the complement system, resulting in cleavage of complement factor C3. Cleaved C3 on the surface of the microorganism mediates opsonization for clearance, but the impact of MBL on subsequent phagocytosis has not been widely studied. We investigated the role of MBL in complement activation and phagocytosis of various bacteria and yeast species by flow cytometry. We measured both the C3 deposition during serum opsonization of fluorescent-labeled microorganisms as well as subsequent uptake of these microorganisms by human neutrophils. In MBL-deficient sera, a consistently decreased C3 deposition on both zymosan and Candida albicans was found and a reduced phagocytosis by neutrophils that was restored by exogenous MBL. This indicates that the lectin pathway of complement activation is important for the opsonophagocytosis of yeasts. In contrast, the C1q-dependent classical pathway dominated in the opsonization and phagocytosis of Staphylococcus aureus, Streptococcus pneumoniae, and Escherichia coli, whereas no effect of MBL was found. Both the lectin and the classical pathway of complement activation were highly amplified by the alternative route for opsonophagocytosis by neutrophils of yeast as well as microbial species. In summary, our data demonstrate that yeast species are preferentially opsonized and subsequently phagocytosed via activation of the lectin pathway of complement, whereas the uptake of bacterial strains was found to be largely MBL independent.     

Mechanisms of the mediated neutrophil reactivity to Escherichia coli peptidoglycan

The opsonic properties of normal serum with respect to E. coli peptidoglycan was studied under the actual conditions of the oxygen-dependent metabolism of neutrophils. In the course of the differentiated study of the influence of antibodies, the classical and the alternative cascades of complement the serum was heated, treated with ethylenediaminetetraacetate and ethylene glycol tetraacetate, exhausted in the cold. In serial experiments the opsonic activity of purified fibronectin was studied. The indirect reactions were shown to be the leading mechanisms of the neutrophil-stimulated activity of E. coli peptidoglycan. IgG was found to be in the center of the opsonic cooperation and thus to determine the quantitative manifestation of the total phenomenon. Complement proved to be of lesser importance; depending on the conditions of the experiment, the activation of complement occurred by the alternative way (after the removal of antibodies) or the classical way (whole serum). The actual contribution of IgG-independent and complement-independent opsonins was insignificant. Fibronectin in physiological concentrations showed no opsonic activity.     

Staphylococcus aureus metalloprotease aureolysin cleaves complement C3 to mediate immune evasion

Complement is one of the first host defense barriers against bacteria. Activated complement attracts neutrophils to the site of infection and opsonizes bacteria to facilitate phagocytosis. The human pathogen Staphylococcus aureus has successfully developed ways to evade the complement system, for example by secretion of specific complement inhibitors. However, the influence of S. aureus proteases on the host complement system is still poorly understood. In this study, we identify the metalloprotease aureolysin as a potent complement inhibitor. Aureolysin effectively inhibits phagocytosis and killing of bacteria by neutrophils. Furthermore, we show that aureolysin inhibits the deposition of C3b on bacterial surfaces and the release of the chemoattractant C5a. Cleavage analyses show that aureolysin cleaves the central complement protein C3. Strikingly, there was a clear difference between the cleavages of C3 in serum versus purified conditions. Aureolysin cleaves purified C3 specifically in the α-chain, close to the C3 convertase cleavage site, yielding active C3a and C3b. However, in serum we observe that the aureolysin-generated C3b is further degraded by host factors. We pinpointed these factors to be factor H and factor I. Using an aureolysin mutant in S. aureus USA300, we show that aureolysin is essential and sufficient for C3 cleavage by bacterial supernatant. In short, aureolysin acts in synergy with host regulators to inactivate C3 thereby effectively dampening the host immune response.     

The respiratory burst response to Histoplasma capsulatum by human neutrophils. Evidence for intracellular trapping of superoxide anion

Human neutrophils (PMN) have received little attention as to the role they play in host defense against Histoplasma capsulatum (Hc). We have characterized the binding and phagocytosis of Hc yeasts by human PMN and quantified the PMN respiratory burst in response to this organism. mAb specific for CD11a, CD11b, and CD11c all partially blocked the attachment of unopsonized yeasts to PMN; a mAb to CD18 inhibited attachment by greater than 90%. Thus, human PMN recognize and bind Hc yeasts via CD18 adhesion receptors as has been found for human cultured macrophages and alveolar macrophages. Unopsonized yeasts were phagocytosed by PMN, but phagocytosis was increased markedly by heat-labile and heat-stable serum opsonins. These opsonins promoted enhanced phagocytosis of yeasts by increasing the attachment of Hc yeasts to the PMN membrane. Phagocytosis of viable or heat-killed Hc yeasts by PMN did not induce the secretion of superoxide anion (O2-) as quantified by the reduction of cytochrome c. O2- was not detected when yeasts were opsonized in normal serum or immune serum, or at a ratio of yeasts to PMN of up to a 100:1. However, phagocytosis of opsonized yeasts by PMN did not prevent them from subsequently releasing O2- after further incubation with opsonized zymosan or PMA. Opsonized Hc yeasts clearly stimulated the PMN respiratory burst as quantified by intracellular reduction of nitroblue tetrazolium, reduction of cytochrome c in the presence of cytochalasin D, oxygen consumption, luminol-enhanced and nonenhanced chemiluminescence, and H2O2 production. These data suggest that phagocytosis of Hc yeasts by PMN is associated with intracellular entrapment of O2- that is not detectable by reduction of extracellular cytochrome c.     

Measurement of phagocyte chemiluminescence using a microtitre plate luminometer

The use of chemiluminescence techniques to study the interaction between bacteria and phagocytes has been useful for examining the extent to which serum factors, such as opsonins, are important in internalization of the organisms and the response of the cell to phagocytosed bacteria. However, such methods have been limited by the number of experiments which can be performed at one time using most commercial luminometers. However, the recent introduction of the Amerlite microtitre plate luminometer allows the measurement of chemiluminescence responses in 96-well microtitre plates. Using this instrument, lucigenin-enhanced chemiluminescence can be detected from as few as 5000 cells (polymorphonuclear leukocytes or monocytes) per well with a 1:10 ratio of cells to zymosan particles opsonized with 10% serum. The opsonic capacity of up to 100 sera can be measured in triplicate wells in a single experiment using four microtitre plates and polymorphonuclear leukocytes prepared from less than 40 ml freshly obtained venous blood. We are currently using this technique to investigate the effect of serum opsonins on the interaction between normal human polymorphonuclear leukocytes and monocytes with mycobacteria of three species (Mycobacterium leprae, M. tuberculosis, and M. aviumintracellulare). Other possible applications of this method are discussed.     

Binding of C-reactive protein to the pneumococcal capsule or cell wall results in differential localization of C3 and stimulation of phagocytosis

C-reactive protein (CRP) is a serum protein that shows rapid increases of as much as 1000-fold in concentration in response to infection, traumatic injury, or inflammation. CRP reacts with the phosphocholine moiety of pneumococcal cell wall C-polysaccharide, and this reaction can lead to complement activation in vitro and protection against pneumococcal infection in vivo. We have previously studied the chemiluminescence response of human neutrophils to Streptococcus pneumoniae as a measure of in vitro opsonophagocytosis by CRP and complement. CRP in the presence of complement was an effective opsonin for S. pneumoniae serotype 27 (Pn27), but not for serotypes 3 or 6. Because Pn27 differs from most serotypes of S. pneumoniae in containing phosphocholine in its capsular polysaccharide, we have determined the sites of CRP and C3 fixation to Pn27 and S. pneumoniae serotype 4 (Pn4), and related these to the ability of CRP and complement to opsonize these serotypes in vitro. By using a chemiluminescence (CL) assay to measure opsonophagocytosis, CRP was shown to enhance the response of human neutrophils and monocytes to Pn27 in the presence of normal human serum. The CL response of neutrophils and monocytes to Pn4 was not affected by the addition of CRP to serum. The addition of anti-capsular antibody to Pn4 and Pn27 enhanced the CL responses of both neutrophils and monocytes to both bacteria. The localization of bound CRP and C3 on Pn4 and Pn27 was determined by immunoelectron microscopy. CRP bound to Pn4 only in the cell wall region and C3 was located in this area whether or not CRP was present. Anti-capsular antibody deposited C3 in the capsule of Pn4. In contrast, Pn27 bound CRP throughout the capsule and cell wall areas. C3 was deposited in the cell wall region of Pn27 by serum alone and in the cell wall region and capsule when CRP or anti-capsular antibody was present. Because C3 fixation to the capsule was consistently associated with enhanced responses by phagocytic cells, it appears that the site of CRP binding and subsequent complement activation may be critical in the opsonophagocytosis of S. pneumoniae. These findings extend the correlation between capsular C3 and opsonization to a nonimmune system. By using CRP and different pneumococcal serotypes we have shown that the same molecules that are effective in the stimulation of phagocytic cells when bound to the capsule are not effective when bound to the cell wall.     

Phagocytosis by bovine polymorphs of glass-adherent Escherichia coli in the presence and absence of serum

The influence of serum on the phagocytosis and killing of Escherichia coli adherent to glass (‘surface phagocytosis’) was investigated by using 2 strains known to require opsonisation when phagocytosis takes place in suspension. Although phagocytosis of the unencapsulated strain occurred in the absence of serum, serum was necessary for the uptake of the encapsulated strain, and improved the uptake of both strains. Killing of ingested bacteria appeared to be independent of serum. Antibodies were not an absolute requirement for the promoting effect of serum, and the necessity for complement (Cp) appeared to depend on the presence of antibodies and on the strain tested. The results suggest that serum components other than opsonins (antibodies and Cp) contributed to the serum-enhanced uptake of bacteria.     

Characterization of Staphylococcus aureus cardiolipin synthases 1 and 2 and their contribution to accumulation of cardiolipin in stationary phase and within phagocytes

In many bacteria, including Staphylococcus aureus, progression from the logarithmic to the stationary phase is accompanied by conversion of most of bacterial membrane phosphatidylglycerol (PG) to cardiolipin (CL). Phagocytosis of S. aureus by human neutrophils also induces the conversion of most bacterial PG to CL. The genome of all sequenced strains of S. aureus contains two open reading frames (ORFs) predicting proteins encoded with ～30% identity to the principal CL synthase (cls) of Escherichia coli. To test whether these ORFs (cls1 and cls2) encode cardiolipin synthases and contribute to CL accumulation in S. aureus, we expressed these proteins in a cls strain of E. coli and created isogenic single and double mutants in S. aureus. The expression of either Cls1 or Cls2 in CL-deficient E. coli resulted in CL accumulation in the stationary phase. S. aureus with deletion of both cls1 and cls2 showed no detectable CL accumulation in the stationary phase or after phagocytosis by neutrophils. CL accumulation in the stationary phase was due almost solely to Cls2, whereas both Cls1 and Cls2 contributed to CL accumulation following phagocytosis by neutrophils. Differences in the relative contributions of Cls1 and Cls2 to CL accumulation under different triggering conditions suggest differences in the role and regulation of these two enzymes.     

Surface properties of Treponema pallidum in relation to phagocytosis by human polymorphonuclear leucocytes in vitro

Surface charge and hydrophobicity of Treponema pallidum have been investigated in relation to phagocytosis by human polymorphonuclear leucocytes (PMNs) in vitro. The treponemal surface was relatively hydrophobic and negatively charged but despite these properties, phagocytosis, as assessed by luminol-enhanced chemiluminescence, was minimal in the absence of serum. Preopsonization of bacteria with serum reduced surface hydrophobicity but promoted phagocytosis, suggesting that specific immune mechanisms may be more important in controlling phagocytosis of T. pallidum in vitro than non-specific surface properties. T. pallidum evoked a much weaker chemiluminescence response from PMNs than the non-pathogenic treponeme Treponema phagedenis biotype Reiterii even though similar numbers of bacteria were phagocytosed, suggesting differences in the reactivity of the surface components of the two organisms toward PMNs. The reactivity of T. pallidum towards PMNs could be increased by removal of the bacterial outer membrane by Triton X-100 treatment. These observations reinforce the suggestion that the outer surface of T. pallidum is inherently inert.     

Production of granulocyte colony-stimulating factor in the nonspecific acute phase response enhances host resistance to bacterial infection

Mice mounting an acute phase response, induced by sterile inflammation after a single s.c. injection of casein 24 h beforehand, were remarkably protected against lethal infection with Gram-positive or Gram-negative bacteria. This was associated with enhanced early clearance of bacteremia, greater phagocytosis and oxidative burst responses by neutrophils, and enhanced recruitment of neutrophils into tissues compared with control, nonacute phase mice. Casein-induced inflammation was also associated with increased concentrations of G-CSF in serum, and administration of neutralizing Ab to this cytokine completely abrogated protection against Escherichia coli infection after casein pretreatment. Injection of recombinant murine G-CSF between 3 and 24 h before infection conferred the same protection as casein injection. In contrast, the casein-induced acute phase response affected neither serum values of TNF-alpha, IL-1 beta, or IL-6 after E. coli infection nor susceptibility to LPS toxicity. Furthermore, protection against infection was unaffected in IL-1R knockout mice, which have deficient acute phase plasma protein responses, or after nonspecific inhibition of acute phase protein synthesis by D-galactosamine or specific depletion of complement C3 by cobra venom factor. Increased production of G-CSF in the acute phase response is thus a key physiological component of host defense, and pretreatment with G-CSF to prevent bacterial infection in at-risk patients now merits further study, especially in view of increasing bacterial resistance to antibiotics.     

Role of myeloperoxidase in the killing of Staphylococcus aureus by human neutrophils: studies with the myeloperoxidase inhibitor salicylhydroxamic acid

We have used salicylhydroxamic acid (SHAM) to inhibit intraphagosomal myeloperoxidase activity in order to evaluate the role of this enzyme in the killing of Staphylococcus aureus by human neutrophils. 50 microM-SHAM reduced the luminol-dependent chemiluminescence response stimulated during phagocytosis of unopsonized latex beads and opsonized S. aureus by over 80% and 60%, respectively. When opsonized S. aureus were incubated with neutrophils, 45% were killed within 15 min incubation and 60% by 1 h. However, in neutrophil suspensions incubated with 50 microM-SHAM, only 13% were killed by 15 min whilst 71% still remained viable after 1 h. This inhibitor had no effect upon the number of bacteria phagocytosed or upon degranulation. In a cell-free system, 2.5 microM-H2O2 alone killed 55% of the bacteria, whereas in the presence of myeloperoxidase (i.e. 10 mU myeloperoxidase and 2.5 microM-H2O2) virtually all of the bacteria were killed: the addition of 50 microM-SHAM abolished this myeloperoxidase-enhanced killing but did not affect the H2O2-dependent killing. We therefore conclude that in normal neutrophils whilst H2O2 is required for killing of this pathogen, both myeloperoxidase-dependent and -independent pathways exist.     

Curvularia haloperoxidase: antimicrobial activity and potential application as a surface disinfectant

A presumed antimicrobial enzyme system, the Curvularia haloperoxidase system, was examined with the aim of evaluating its potential as a sanitizing agent. In the presence of hydrogen peroxide, Curvularia haloperoxidase facilitates the oxidation of halides, such as chloride, bromide, and iodide, to antimicrobial compounds. The Curvularia haloperoxidase system caused several-log-unit reductions in counts of bacteria (Pseudomonas spp., Escherichia coli, Serratia marcescens, Aeromonas salmonicida, Shewanella putrefaciens, Staphylococcus epidermidis, and Listeria monocytogenes), yeasts (Candida sp. and Rhodotorula sp.), and filamentous fungi (Aspergillus niger, Aspergillus tubigensis, Aspergillus versicolor, Fusarium oxysporum, Penicillium chrysogenum, and Penicillium paxilli) cultured in suspension. Also, bacteria adhering to the surfaces of contact lenses were killed. The numbers of S. marcescens and S. epidermidis cells adhering to contact lenses were reduced from 4.0 and 4.9 log CFU to 1.2 and 2.7 log CFU, respectively, after treatment with the Curvularia haloperoxidase system. The killing effect of the Curvularia haloperoxidase system was rapid, and 10(6) CFU of E. coli cells/ml were eliminated within 10 min of treatment. Furthermore, the antimicrobial effect was short lived, causing no antibacterial effect against E. coli 10 min after the system was mixed. Bovine serum albumin (1%) and alginate (1%) inhibited the antimicrobial activity of the Curvularia haloperoxidase system, whereas glucose and Tween 20 did not affect its activity. In conclusion, the Curvularia haloperoxidase system is an effective sanitizing system and has the potential for a vast range of applications, for instance, for disinfection of contact lenses or medical devices.     

Studies on the phagocytic activity of the reticuloendothelial system (RES) to rough and smoothEscherichia coli in young conventional and germfree guinea pigs

The functional (phagocytic) capacity of the reticuloendothelial system (RES) of young conventional and germfree guinea pigs was studied using thein vivo blood clearance test of living bacteria (rough and smoothEscherichia coli). It was found that as previously shown in newborn germfree piglets, the smooth strain was taken up from the blood stream of germfree guinea pigs very slowly whereas roughEscherichia coli was phagocytosed effectively. The inability of the RES of germfree guinea pigs to phagocytose the smooth strain is not due to a functional incapability of phagocytic cells, but it reflects rather the lack of serum opsonins to this strain. This was demonstrated in experiments in which smooth bacteria, sensitized prior to injection into the blood circulation with specific antiserum, were phagocytosed as effectively as the rough strain. It is assumed that effective phagocytosis of rough strain is due to the presence of non-specific opsonins (e.g. components of the complement system). In young conventional guinea pigs both strains,i.e. smooth and rough, were taken up from the blood stream very effectively thus indicating that sufficient levels of serum opsonins for both strains were present. This fact could be correlated with the finding that in sera of conventional guinea pigs haemagglutinating antibodies to both strains ofEscherichia coli could be detected, whereas in sera of young germfree guinea pigs, no antibodies to usedEscherichia coli strain were found. The importance of serum opsonins for effective phagocytosis of bacteria by RE cellsin vivo is discussed.