Infection at the Subcellular Level I. Localization of Intravenously Injected Brucellae in the Vacuolar Apparatus of Cells of Guinea Pig Liver

The distribution of Brucella melitensis in various tissues and in subcellular fractions obtained from liver was investigated to evaluate the initial phases of brucellosis in the guinea pig. Fifty minutes after intravenous infection, brucellae were found principally in the blood and liver, with a substantial number recovered from spleen. Fractionation of liver established that most bacteria were found in the mitochondrial plus lysosomal (M + L) fraction; a significant number, however, sedimented in the nuclear (N) fraction. With time, there was a progressive shift of bacteria from the M + L to the N fraction, accompanied by a similar shift in acid phosphatase activities. Isopycnic centrifugation of mixtures of M + L fractions and brucellae permitted complete separation of acid phosphatase-bearing particles from bacteria. Similar experiments with fractions from infected animals showed that viable bacteria were found in both the acid phosphatase and free brucellae regions of the gradient. At 10 min postinfection, 52% of the recovered organisms were in the acid phosphatase region; at 30 min, 65%; at 60 min, 85%; and at 315 min, 79%. Detergent plus sonic treatment of an M + L fraction from the liver of an animal killed 50 min after infection caused most of the bacteria in the acid phosphatase region to shift to the region where free bacteria were found. These data suggested that brucellae sequestered in the liver were located primarily in the vacuolar apparatus of the cells which phagocytized them.     

Effect of hypoxia on protein tyrosine phosphatase activity and expression of protein tyrosine phosphatases PTP-1B, PTP-SH1 and PTP-SH2 in the cerebral cortex of guinea pig fetus

The present study tested the hypothesis that the hypoxia in utero results in decreased protein tyrosine phosphatase (PTP) activity in cytosolic and membrane fractions and increased expression of PTPs (PTP-1B, PTP-SH1 and PTP-SH2) in the cytosol and the membrane fraction of the cerebral cortex of guinea pig fetus. In addition, we hypothesize that the increased expression is mediated by nitric oxide (NO). To test this hypothesis, PTP activity in cytosol and cell membrane, and expression in the cytosol and membrane fraction were measured in the cerebral cortex of normoxic, hypoxic and L-nitro-L-arginine methyl ester (L-NAME), an inhibitor of nitric oxide synthase (NOS), pretreated hypoxic (L-NAME+Hx) guinea pig fetuses. PTP activity in the cytosolic and membrane fractions was significantly lower in the Hx group as compared to the Nx group. The density of cytosolic PTP-1B, cytosolic PTP-SH1 and PTP-SH2 was increased in the Hx group and this increase was prevented in the L-NAME+Hx group. The data show that pretreatment with L-NAME, an inhibitor of NOS, prevents the hypoxia-induced increased expression of PTP-1B, PTP-SH1 and PTP-SH2 in the membrane and cytosolic fractions of the cerebral cortex of the guinea pig fetus. We conclude that the decrease in PTP activity during hypoxia is not due to protein modification of PTP and due to alteration in PTP expression.     

Pharmacokinetics, subcellular distribution, and covalent binding of [3H]bleomycin in hamsters after intratracheal administration

Pharmacokinetics, subcellular distribution (SCD), and covalent binding of a single dose of 1 microCi of [S-methyl-3H]bleomycin ([3H]-BLM]) in combination with one unit of unlabeled bleomycin were studied in hamsters following intratracheal (IT) injection. The radioactivity decreased from the lung biexponentially with time. The apparent half-time of absorption for the alpha-phase was 1.1 and 17.9 hr for the beta-phase. The plasma disappearance curve of [3H]BLM fits to a two-compartmental model with the apparent half-life removal for the alpha-phase being 1.6 hr and for the beta-phase 116.9 hr. The radioactivity was detected in all studied tissues. The radioactivity from spleen, testicle, liver, fat, RBC, brain, adrenal, and kidney manifested only the alpha-phase of the disappearance curve, while the beta-phase was complicated by redistribution processes. Of the eight tissues, the spleen had the shortest (2.0 hr) and kidney the longest (12.1 hr), and the remaining tissues had half-lives which ranged from 4 to 10 hr. The SCD study revealed that 85 to 95% of the total radioactivity in the lung and liver homogenate was associated with the soluble fraction (SF) at 30 min after treatment, thereafter, the radioactivity from both tissues gradually decreased to 60% of the total at 24 hr. The SF of the lung homogenate had the highest specific radioactivity (SRA) of any of the fractions during the period between 0.5 and 6 hr. The SRA, however, decreased biexponentially and attained a value similar to that of the mitochondrial and microsomal fractions at 12 and 24 hr after treatment. In the case of liver, the SF had the highest, the nuclear the lowest, and mitochondrial and microsomal fractions the same level of SRA at 30 min. Thereafter, the SRA of all fractions were increased with time. A significant amount of radioactivity from [3H]BLM was covalently bound to lung, liver, and plasma proteins. The SF of the lung contained an increasing amount of radioactivity covalently bound after 1.5 hr of [3H]BLM injection and nearly all radioactivity measured in the plasma was covalently bound. It was concluded from the findings of this study that the presence of a major portion of [3H]BLM in the SF of the lung and its covalent linkage with the proteins of this fraction might initiate the complex sequence of events at the metabolic level necessary for the pneumotoxicity.     

Sensitization and recall of anti-Brucella immunity in Guinea pig macrophages by attenuated and virulent Brucella

Peritoneal macrophages from guinea pigs vaccinated with strain Rev I of Brucella melitensis were only moderately activated thereby to limit, in an in vitro system, the intracellular growth of Rev I bacilli. Nevertheless, the appropriate memory cells had been primed, as demonstrated by the observation that reinfection of animals with virulent B. melitensis followed by intraperitoneal inoculation of mineral oil called forth macrophages in immunized guinea pigs which inhibited strongly the intracellular growth of brucellae. These macrophages slowed the growth of brucellae in the absence of immune serum. The intensity of the recall response was related to the challenge route and to the virulence of the challenge strain. After equal doses of attenuated or virulent brucellae, resistance was highest in macrophages recalled by the virulent strain. An important basis for the attenuation of the Rev I strain may lie in its initially low degree of macrophage activation during primary infection, although still retaining the capacity to prime stem cells. This property is associated with a protein found in fraction I, because 600 μg/ml in Freund's adjuvant primed guinea pigs so that challenge by strain 6015 evoked activated macrophages. This was seen microscopically as a reduced spread of infection in and amongst the macrophage population. Immune serum further reduced this spread and limited the number of viable intracellular brucellae.     

Kinetics of killing Listeria monocytogenes by macrophages: rapid killing accompanying phagocytosis

The kinetics of bactericidal activity of activated macrophages can be precisely described by a mathematical model in which phagocytosis, killing, digestion, and release of degraded bacterial material are considered to occur continuously. To gain a better understanding of these events, I have determined the period of time between first contact of bacteria with macrophages and the onset of killing. Activated rat peritoneal macrophages were incubated for various times up to 15 min with Listeria monocytogenes previously labeled with 3H-thymidine and the unassociated bacteria removed by two centrifugations through a density interface. Both cell-associated radioactivity and cell-associated viable bacteria, determined as colony forming units after sonication of the cell pellet, increased with time of incubation. However, the specific viability of these bacteria, expressed as the ratio of number of viable bacteria per unit radioactivity declined with time, as an approximate inverse exponential, after a lag period of 2.9 +/- 0.8 min. Evidence is given that other possible causes for this decline in specific viability, other than death of the bacteria, such as preferential ingestion of dead Listeria, clumping of bacteria, variations in autolytic activity, or release of Listericidins are unlikely. I conclude therefore that activated macrophages kill Listeria approximately 3 min after the cell and the bacterium first make contact.     

Distribution of 3H-labeled staphylococcal alpha-toxin and a toxin fragment in mice.

Staphylococcal alpha-toxin and a toxin fragment were labeled with N-succinimidyl[2,3-3H]propionate. The labeled compounds retained greater than 95% biological activity. The distribution of labeled staphylococcal alpha-toxin and alpha-toxin fragment after intravenous administration to BALB/c mice was studied with whole-body and microautoradiography. The animals were divided into three groups that received (i) labeled alpha-toxin only, labeled alpha-toxin after prior injection of unlabeled fragment, or labeled fragment only. After 5 min, the distribution patterns were similar in groups 1 and 2, with the highest amounts of radioactivity found in the blood vessels, liver, spleen, lungs, and kidneys, whereas the labeled fragment alone showed no initial accumulation in the lungs. The kidneys continued to show a high concentration of radioactivity, whereas the levels at 60 min had decreased in the other organs. The toxin showed continued stable binding to the proximal tubuli, whereas the toxin fragment seemed to dissociate and was found only in small amounts in the glomeruli. No radioactivity was found in the central nervous system.     

Purification, identification and elimination of a natural human serum antibody with cytotoxic effect on guinea pig thymocytes

Normal human sera contain one or several factors cytotoxic for normal guinea pig thymocytes, and when serum is precipitated with ammonium sulphate (60% saturated) and the precipitate dissolved and dialyzed, the activity is preserved. Gel chromatography with Sephadex G-150 and Sepharose CL-6B indicated a molecular weight of approximately 900,000 daltons. The active fractions contained a high amount of IgM according to single radial immunodiffusion and two-dimensional gel electrophoresis. Quantitation of the IgM band in one-dimensional gel electrophoresis preparations by gel scanning indicated that IgM accounted for 65% of the eluted proteins in active fractions. Purified human IgM from myeloma patients eluted as the active factor during gel chromatography. Elimination of IgM from serum by affinity chromatography eliminated the cytotoxic activity. The serum could also be inactivated by heating. The mixing of IgM-depleted serum with either polyclonal IgM or heat-inactivated serum restored the activity. Thus, the cytotoxic activity is due to IgM antibodies plus a heat-labile component (presumably complement). The presence of the cytotoxic activity in autologous (guinea pig) serum was recently demonstrated. The possible functional role of these antibodies in the elimination process of a large number of cortical thymocytes is suggested.     

Release of Mycoplasma pneumoniae substances after phagocytosis by guinea pig alveolar macrophages.

Antibody-opsonized Mycoplasma pneumoniae cells with various radioactive markers were sedimented onto monolayers of guinea pig alveolar macrophages (AM). After 2 h of incubation, about 50% of the activity of [3H]palmitate-labeled mycoplasmas was associated with AM. Nonspecific attachment of the opsonized mycoplasmas to AM-free plastic surface areas was negligible. The occurrence of phagocytosis was proven by electron microscopy and monitoring of AM surface-bound antigen by 125I-labeled F(ab)2 fragments. The activity of [3H]palmitic acid-labeled mycoplasmas was only slowly released into the supernatant. About 55% of the activity remained AM-associated up to 70 h after phagocytosis. After phagocytosis of [3H]thymidine-labeled cells, about 70% of the radioactivity found non-precipitable by trichloracetic acid. 3H-amino acid-labeled protein was released to 50% within 8 h. Supernatants and AM were tested for M. pneumoniae antigen with enzyme-linked immunosorbent assay. Considerable amounts of antigenically active material could be found in the supernatant within 8 h. This antigen was totally inactivated by heat (80 degrees C). Trypsin treatment (1 mg/ml, 10 min) reduced the antigenicity by 80%. The results suggest a selective release of microbial material after phagocytosis.     

Metabolism of primaquine by liver homogenate fractions. Evidence for monoamine oxidase and cytochrome P450 involvement in the oxidative deamination of primaquine to carboxyprimaquine.

The role of monoamine oxidase (MAO) and cytochrome P450 (P450) in the oxidative deamination of primaquine by rat liver fractions was studied. Rat liver fractions including liver homogenate, mitochondria, microsomes and 100,000 g supematant fractions were prepared from a pool of rat livers and characterised using benzylamine as a probe for MAO activity and N,N-dimethylbenzamide as a probe for P450 N-dealkylation activity. Incubation of all fractions with primaquine yielded carboxyprimaquine as the only metabolite detectable by HPLC. The mitochondrial fraction, which contained MAO activity but not P450 activity, presented the highest Vmax/K(M) value for the formation of carboxyprimaquine (8.5 x 10(-6) dm3mg(-1)h(-1). A substantially lower Vmax/K(M) value (1.3 x 10(-6) dm3mg(-1)h(-1)) was obtained in the microsomal fraction, which contained P450 but not MAO activity. The liver homogenate fraction presented a similar value (1.8 x 10(-6) dm3mg(-1)h(-1), though it contained both enzyme systems. Incubations of all the fractions that presented MAO activity, in presence of the MAO inhibitor pargiline, resulted in a marked inhibition of primaquine oxidation. P450 inhibitor SKF 525-A effectively inhibited primaquine metabolism in the microsomal fraction but inhibition in the liver homogenate was less effective. The results are consistent with an important role for MAO in primaquine biotransformation, though clearly metabolism by P450 has a contribution role.     

Subcellular distribution of styrene oxide in rat liver

The subcellular distribution of (3H )-styrene-7,8-oxide was studied in the rat liver. The compound was added to liver homogenate to give a final concentration of 2 X 10(-5); 2 X 10(-4) and 2 X 10(-3) M. Subcellular fractions were obtained by differential centrifugation. Most of styrene oxide (59-88%) was associated with the cytosolic fraction. Less than 15 percent of the compound was retrieved in each of the nuclear, mitochondrial and microsomal fractions. A considerable percentage of radioactivity was found unextractable with the organic solvents, suggesting that styrene oxide reacted with the endogenous compounds. The intracellular distribution of this epoxide was also studied in the perfused rat liver. Comparable results with those previously described were obtained. The binding of styrene oxide to the cytosolic protein was investigated by equilibrium dialysis and ultrafiltration. Only a small percentage of the compound was bound to protein.     

Residual latent activity of acid phosphatase: Autophagy-related variations and effects of cycloheximide

The patterns of activation of two lysosomal hydrolases, acid phosphatase and N-acetyl β-d-glucosaminidase, by hypotonic shock have been measured in heavy and light mitochondrial, and in microsomal fractions from rat liver. The osmotic sensitivity of the particles, as inferred from enzyme activation, was highest in the heavy mitochondrial and lowest in the microsomal fraction. An aliquot of activity (referred to as residual latent activity, RLA) remained latent even after incubation in distilled water for 30 min at 0°C. The proportion of RLA has been shown to vary in relation to changes in the dynamic state of the vacuolar apparatus. A significant decrease of RLA occurred in rats fasted for 24 hr as well as after glucagon administration, two situations known to be correlated with increased autophagy. A significant increase of RLA has been found in regenerating liver, where autophagy is reduced. The effect of glucagon was prevented by previous injection of cycloheximide, which has been reported to inhibit the stimulation of autophagy by the hormone. In addition we have found that cycloheximide alone causes a very conspicuous increase of RLA, already noticeable 30 min after the treatment.     

A new assay system for guinea pig interferon biological activity.

We have developed an assay system for guinea pig interferon (IFN) based on reduction of viral cytopathic effect (CPE) in various cell lines. CPE inhibition was detected optimally in the guinea pig fibroblast cell line 104C1 infected with encephalomyocarditis virus (EMCV). The amount of biologically active guinea pig IFN was quantified by estimating viable cell numbers colorimetrically by means of a tetrazolium compound, 2-(4-iodophenyl)-3-(4-nitrophenyl)-5-(2,4-disulfophenyl)-2H-tetrazolium monosodium salt (WST-1) and 1-methoxy-5-methylphenazinium methylsulfate (PMS). WST-1 color developed until stopped by the addition of sulfuric acid. This had no effect on the colorimetric assay, and the color was stable for at least 24 h. The acid also inactivated the EMCV and, thus, eliminated the viral hazard. Inhibition of CPE activity was highly correlated with the concentration of culture supernatants from BCG-vaccinated guinea pig splenocytes stimulated in vitro with tuberculin or an immunostimulatory oligoDNA. This assay detected guinea pig IFN and human IFN-alpha, but not IFN-gamma from human, mouse, rat, pig, or dog. This assay system has proved useful for the titration of guinea pig IFN, being easy to perform, free from viral hazard, relatively species specific, highly reproducible, and inexpensive.     

Guinea pig Hageman factor as a vascular permeability enhancement factor.

Hageman factor was purified from guinea pig plasma by successive column chromatography. The guinea pig Hageman factor appeared homogeneous as a single-chain protein on polyacrylamide gels in the presence of sodium dodecyl sulfate (SDS) and beta-mercaptoethanol. The apparent molecular weight was 76,000 daltons by SDS--polyacrylamide gel electrophoresis and 105,000 daltons by gel filtration with a Sephadex G-150 column. Amino acid composition of the guinea pig Hageman factor was similar to that reported for human, bovine, and rabbit Hageman factors. The purified guinea pig Hageman factor, as well as guinea pig plasma, showed strong clotting time correction activity in Hageman-factor--deficient human plasma. The activity could be blocked by the IgG fraction of antiserums against guinea pig Hageman factor raised in rabbits or a goat. The concentration of Hageman factor in guinea pig plasma was determined to be 120 microgram/ml by quantitative radial immunodiffusion assay. The 28,000-dalton active form of Hageman factor (beta-HFa) was prepared from guinea pig Hageman factor by treatment with plasma kallikrein. beta-HFa caused an increase in vascular permeability when injected into guinea pig skin at concentrations as low as 3 x 10(-10) M (0.8 ng). Native, or zymogen Hageman factor did not cause an increase in permeability at concentrations of up to 2 x 10(-7) M. The increased permeability induced by beta-HFa was short lasting, with about a 50% decrease in activity apparent within 6 minutes after intradermal injection. The permeability enhancement activity of beta-HFa was inhibited by pretreatment of beta-HFa with diisopropylfluorophosphate. It may be concluded that active Hageman factor in the interstitial space of guinea pigs acts as a vascular permeability factor of far greater potency than bradykinin.     

Demonstration of a Complement-Dependent Migration Inhibitory Activity in Normal Guinea Pig Serum

A cell migration inhibitory effect was evidentiated in normal guinea pig serum as compared with heat inactivated serum, Granuloeytes when used as target showed a greater sensitivity to this effect than lymphomonocytes or macrophages. The migration inhibitory activity of GPS is abrogated or decreased by using complement destroying agents such as: heating at 56°C for as little as 5 min, absorption on immune complexes in presence of calcium, on zymosan, on Sephadex G-50 or by adding EDTA or heparin to culture medium. The GPS dialysation fractions while exhibiting neither complement haemolytic effect nor migration inhibitory activity when tested alone, restored these functions by recombination. Absorption of GPS on homologous blood cells abrogated the migration inhibitory effect but retained the complement haemolytic function. When GPS absorbed on homologous blood cells was mixed 1:5 with heat-inactivated serum (5min at 56°G), the migration inhibitory activity was regained, suggesting that the complement factors from the first sample were necessary for manifestation for the migration inhibitory activity from the heat-inactivated serum.     

Characterization of Ovarian RNA

Ovaries were obtained from 24–25 day old Sprague-Dawley rats. Groups of 20 ovaries were collected corresponding to a wet weight of 100–120 mg, and incubated for 20, 30, 60 and 240 min at 37° C in Krebs bicarbonate buffer containing 5.5 mM glucose and uridine-³H. Following incubation, ovarian RNA was extracted with a modified phenol extraction method described in detail. Fractionation was made by composite agarose-polyacrylamide gel electrophoresis which gave good separation and high resolution of the various RN.4 fractions. Ovarian RNA separated in four well-defined fractions classified as 28S, 18S, 5s and 4s. This classification was based upon comparisons with simultaneously run liver RNA. In addition to the four main fractions. a weak band just before 28S and two smaller bands between 28S and 18S were consistently observed. Labelling of ovarian RNA with uridine-³H showed a polydisperse pattern after the shorter incubation periods with most of the radioactivity concentrated to RNA fractions of molecular weights higher than 28S, but also clear labelling of the 5S—4S region. After 60 min, radioactivity peaks were also associated with the 28S and 18s fractions and with fractions between those, while after 240 min incubation, the radioactivity was concentrated to the four main fractions. The incorporation of radioactivity was found to be markedly decreased in presence of actinomycin D.     

Increased lipid peroxidation in the liver and kidney and their mitochondrial fractions following buthionine sulfoximine induced glutathione depletion in guinea pigs

Malondialdehyde (MDA) and diene conjugates (DC) and vitamin C levels and the activities of glutathione peroxidase (GSH-Px) and superoxide dismutase (SOD) were determined in the liver and kidney and their mitochondrial fractions of guinea pigs 48 h after the injection of L-buthionine-(S,R)-sulfoximine (BSO), a glutathione (GSH) depleting agent. In BSO-induced GSH depletion, lipid peroxidation and SOD activities were found to be increased but GSH-Px activities did not change in the liver and kidney and their mitochondrial fractions. In addition, vitamin C levels remained unchanged in the liver and kidney homogenates. These results indicate that GSH depletion may influence oxidative stress in the liver and kidney and their mitochondrial fractions of guinea pigs.     

Demonstration of epidermis-specific heteroantigens in thymic epithelial cells.

Heterologous anti-epidermal cell serum was obtained by immunizing rabbits with enzymatically dispersed, viable epidermal cells of guinea pigs, followed by absorption of the antiserum with red blood cells (sheep and guinea pig), lymphoid cells and liver powder (guinea pig). Immunofluorescence demonstrated that the antiserum reacted specifically with stratified squamous epithelial cells and thymus epithelial cells including Hassall's bodies of the guinea pig, monkey and man. It is suggested that the epithelial cells of the skin and thymus have common heteroantigens.     

Biosynthesis of the subcomponents C1q, C1r and C1s of the first component of complement (C1) by guinea pig hepatocyte primary cultures

Thus far, the synthesis of C1q by liver cells has not been demonstrated. To investigate this possibility, viable hepatocytes were isolated from the liver of guinea pigs and primary cultures were established. The cells (10(6) cells/ml) were cultured under serum-free conditions for 8 days and the culture medium was changed every 24 h. The few contaminating Kupffer cells were lysed by preincubating the cell cultures with a monoclonal (22C4-8) antibody directed against a nonpolymorphic Ia determinant and preabsorbed rabbit serum. The hemolytic activity of C1 and its subcomponents C1q and C1r/C1s was tested in the supernatants. Guinea pig hepatocyte primary cultures synthesize and secrete up to 3 X 10(3) effective C1q molecules/cell/24 h and 34 X 10(3) effective C1r/C1s molecules/cell/24 h. The synthesis of C1q and C1r/C1s could be reversibly inhibited by cycloheximide (50 micrograms/ml). Furthermore, to demonstrate de novo synthesis of the C1q subcomponent, endogeneous labeling with 3H-proline (or 14C-proline) was performed. The immunoprecipitated C1q from cellular lysates and culture medium was analyzed by sodium dodecyl sulfate-polyacrylamide gel electrophoresis and fluorography. Compared to biosynthetically labeled guinea pig C1q from peritoneal macrophages, three corresponding bands (30, 28 and 24 kDa, respectively) were detectable in the fluorograph. The data show that guinea pig hepatocytes are able to synthesize C1 subcomponents, whereby the synthesis of C1q and C1r/C1s occurs independently.     

Susceptibility of pathogenic and nonpathogenic Naegleria spp. to complement-mediated lysis.

The susceptibility of four species of Naegleria amoebae to complement-mediated lysis was determined. The amoebicidal activity of normal human serum (NHS) and normal guinea pig serum (NGPS) for Naegleria amoebae was measured by an in vitro cytotoxicity assay. Release of radioactivity from amoebae labeled with [3H]uridine and visual observation with a compound microscope were used as indices of lysis. Highly pathogenic mouse-passaged N. fowleri was less susceptible to the lytic effects of NHS and NGPS than the weakly pathogenic, axenically grown N. fowleri or N. australiensis and the nonpathogenic amoebae N. gruberi and N. lovaniensis. However, both pathogenic and nonpathogenic Naegleria spp. depleted complement as assessed by total hemolytic activity. Treatment of serum with EDTA, heat (56 degrees C, 30 min), cobra venom factor, or antibody to C3 or C9 complement components decreased the amoebicidal activity of NHS. The presence of specific agglutinating antibody to N. fowleri enhanced the amoebicidal activity of NGPS for N. fowleri.     

Natural hemolytic activity of snake serum. I. Natural antibody and complement.

Normal snake sera from five species were highly active in lysing sheep erythrocytes (SRBC). No significant hemolytic activity, however, was found in the sera of tortoises, lizards, or bullfrogs. Natural hemolytic activity of snake sera was dependent upon natural antibody and complement. The antibody was heat-labile and inactivated by treatment at 55° for 15 min. A component (C_S), probably the first component, of snake complement may be snake-specific; thus, sera from guinea pigs, bullfrogs, tortoises, chickens, and rabbits were ineffective in lysing SRBC sensitized with snake anti-SRBC natural antibody. Components (C_G) of snake complement other than C_S could be replaced by those of guinea pig complement. The activity of the snake complement system was inactivated by treatment of sera at 55° for 15 min. The titer of complement exceeded that of natural antibody, indicating that the limiting factor in the natural hemolytic activity of snake sera depended upon the level of natural antibody.