Dog Erythrocytes Research Articles

(Na,K)-ATPase and Ouabain binding in reticulocytes from dogs with high K and low K erythrocytes and their changes during maturation.

The present study demonstrated that dog reticulocytes had considerable amounts of (Na,K)-ATPase, but lost it rapidly during maturation into erythrocytes. Furthermore, reticulocytes from dogs possessing erythrocytes characterized with high (Na,K)-ATPase activity and high K, low Na concentrations (HK dogs; Maede, Y., Inaba, M., and Taniguchi, N. (1983) Blood 61,493-499) had more ouabain binding sites than cells from normal dogs (LK dogs). Our results were as follows: i) The maximal binding capacities (Bmax) for ouabain binding at equilibrium were approximately 0 and 1,500 binding sites/cell in LK and HK dog erythrocytes, respectively. ii) Reticulocytes from LK dogs possess approximately 5,700 ouabain binding sites/cell. iii) The Bmax value for ouabain in HK reticulocytes was about 10,000 sites/cell, being 2-fold that in LK reticulocytes. iv) Ouabain-sensitive fluxes of 24Na and 42K in each type of reticulocyte were compatible with the number of ouabain binding sites on the cells. v) Ouabain binding capacity, as well as (Na,K)-ATPase activity, in the reticulocytes from LK dogs fell rapidly to nearly zero during the maturation into erythrocytes. vi) Although reticulocytes from HK dogs also showed a similar regression of (Na,K)-ATPase during maturation, they retained a certain number of ouabain binding sites even after maturation, resulting in the high activity of (Na,K)-ATPase in HK erythrocyte membrane.

Purification and characterization of the gonad lectin of Aplysia depilans

Extracts of gonads and fertilized eggs of Aplysia depilans contain a D-galacturonic and D-galactose-binding lectin. This lectin reacts strongly with rabbit and human erythrocytes independent of ABO blood groups, weakly with dog, mouse, rat, and chick erythrocytes and not at all or very weakly with sheep erythrocytes. Purification of the gonad lectin was easily achieved, with a high yield, by heating to 70°C, precipitation with ammonium sulfate and affinity chromatography on Sepharose 4B. The purified lectin was found to be a glucoprotein of molecular mass around 55–60 kDa; it stimulates mitogenesis of human peripheral lymphocytes.

Erythrocyte acid phosphatase: Species specificity in activity modulation by purine analogs

1.1. The previous studies of the interaction of purine analogs and human erythrocyte acid phosphatase isozymes were extended to include erythrocyte acid phosphatase from seven other species.2.2. Consistent responses, similar to the observations with the several genotypically different human isozymes, were observed.3.3. The isozyme from chimpanzee erythrocytes was similar to the human B-type isozyme while the baboon and cow isozymes were at the other extreme in responsiveness and were more divergent from the B-isozyme than was the human A-type isozyme. The ACP from rabbit, dog, sheep and rhesus erythrocytes exhibited intermediate levels of responsiveness but did differ from the human A-type isozyme.4.4. Additional studies indicated some differences between the responsiveness of the partially purified erythrocyte enzyme and the low molecular weight ACP from liver.

NMR relaxation studies of intracellular Na+ in red blood cells.

The state of intracellular Na+ in human and dog erythrocytes was characterized by 23Na-NMR using dysprosium complexes as shift reagents. Intracellular Na+ concentrations were determined using integration of the inner Na+ NMR signals and measurements of the intracellular volume using 59Co-NMR of extracellular Co(CN)3-6. T2 was found to be significantly shorter than T1, indicating some binding to macromolecules. While the longitudinal magnetization decay follows a single exponential, the transverse magnetization could be fitted with a double-exponential function. It was shown that neither the binding to the inner side of the membrane nor binding to hemoglobin contributes to the relaxation enhancement.

Further studies on the red cell glycolipids of various breeds of dogs. A possible assumption about the origin of Japanese dogs.

Genetic polymorphism was observed in the sialic acid species constituting the terminal sugar residues of hematosides from dog erythrocytes. One was N-acetylneuraminic acid and the other phenotype was N-glycolylneuraminic acid, regulated by an autosomal dominant allele (Yasue, S., Handa, S., Miyagawa, S., Inoue, J., Hasegawa, A., & Yamakawa, T. (1978) J. Biochem. 83, 1101-1107). In this study we analyzed blood samples from 1,591 dogs of 36 breeds and demonstrated that the expression of N-glycolylneuraminic acid was limited to several breeds of oriental dogs in spite of its dominant nature. Moreover, the incidence of N-glycolylneuraminic acid was higher in native breeds of northern China, Korea and the southern part of Japan than in other oriental breeds. On the other hand, the Hokkaido-dog is unique in not expressing N-glycolylneuraminic acid. These results suggest that the native breeds in the southern part of Japan came from northern China via the Korean peninsula in contrast with indigenous breeds of the northern part of Japan.

Sodium- and adenosine-triphosphate-dependent calcium movements in membrane vesicles prepared from dog erythrocytes.

Inside-out vesicles from the membranes of dog erythrocytes were obtained by the method of Lew & Seymour (1982) for study of Ca movements. In the absence of ATP, 45Ca accumulation by the vesicles was inhibited by external Na and stimulated by internal Na. The presence of either MgCl2, quinidine sulphate, or LaCl3 in the incubation medium inhibited 45Ca accumulation in the absence of ATP. The release of 45Ca from 45Ca-loaded vesicles was specifically promoted by Na+ in the absence as well as in the presence of ATP. The accumulation of 45Ca by vesicles was stimulated by ATP and the effect of ATP was entirely dependent on the presence of Mg. The Mg- and ATP-dependent 45Ca accumulation was stimulated by the presence of either K or Na in the medium, was hyperbolically activated by increasing the Ca2+ concentration in the medium, was stimulated by calmodulin and inhibited by orthovanadate (10(-4) M) or LaCl3 (10(-3) M). The data demonstrate the presence of two mechanisms for controlling Ca movements in inside-out vesicles from dog erythrocyte membranes, a Na-dependent one similar to the Na-Ca exchange described for squid axons and cardiac muscle and a Ca pump utilizing ATP with characteristics similar to those described for human erythrocytes and squid axons.

Evidence for a specific phosphatidylinositol 4-phosphate phosphatase in human erythrocyte membranes.

Human erythrocyte membranes exhibit a specific phosphatidylinositol 4-phosphate phosphohydrolase (PtdIns4P phosphatase) activity which hydrolyzes PtdIns4P and lysoPtdIns4P but not phosphatidylinositol 4,5-bisphosphate (PtdIns(4,5)P2) or lysoPtdIns(4,5)P2. Phosphatidic acid, lysophosphatidic acid, glycerophosphoinositol 4-phosphate, glycerophosphoinositol 4,5-bisphosphate, inositol mono, bis, and tris phosphates and several other sugar and nucleoside phosphates are not hydrolyzed. The PtdIns4P phosphatase activity is not affected by Ca2+ or Mg2+ ions nor inhibited by EDTA. Maximum in vitro activity requires non-ionic (Triton X-100) detergents. The phosphatase is very stable in isolated membranes at low temperatures but is rapidly inactivated above 35 degrees C. This critical inactivation temperature is lowered to 20-25 degrees C by solubilizing the membranes with non-ionic detergents. Arrhenius plots of the activity show an inflection at these critical temperatures, suggesting a temperature-dependent change in the environment or conformation of the enzyme. Sulfhydryl-reacting reagents are potent inhibitors. Dithioerythritol stimulates only when the membranes are solubilized with non-ionic detergent. The location of cation-independent PtdIns4P phosphatase activity in the membrane and of Mg2+-dependent PtdIns(4,5)P2 phosphatase activity in the cytosol was also observed for monkey, rabbit, rat, and dog erythrocytes. Both activities are located in the cytosol of sheep erythrocytes.

Role of intrinsic arachidonate metabolites in the vascular action of erythrocyte breakdown products.

In helically-cut strips of dog basilar and mesenteric arteries, the isometric tension developed by application of ghost-free hemolysate from dog erythrocytes was recorded. The hemolysate contracted basilar arteries in a concentration-dependent fashion, the response being attenuated by treatment with either aspirin or polyphloretin phosphate, a prostaglandin antagonist. Mesenteric arteries were contracted only slightly by high concentrations of hemolysate. When the mesenteric arteries had partially been contracted with prostaglandin F2 alpha or norepinephrine, the hemolysate induced relaxations, which were abolished by aspirin in approximately half the preparations used. Studies on rat stomach strips exposed to superfusate of dog cerebral arteries showed a release of prostaglandin-like substance by the hemolysate application. It may be concluded that the hemolysate contracts basilar arteries and relaxes mesenteric arteries, mainly through prostaglandins synthesized in and released from the vascular wall. Such a mechanism may be involved in the pathogenesis of cerebral vasospasm following a subarachnoid hemorrhage.

Disposition in rats and dogs of ciglitazone, a new antidiabetic agent.

Oral 14C-ciglitazone was well absorbed by rats to give a maximum plasma level at two hours and an apparent half-life of 4.9 h. In dogs, the plasma level of the compound, after oral administration, reached a plateau at one hour, persisted till ten hours and then declined with a half-life of 23.5 h. In rats, plasma levels of metabolites were higher than those of unchanged ciglitazone, whereas the reverse was noted in dogs. Plasma metabolites were the monohydroxycyclohexyl derivatives (mono-ol) and monoketocyclohexyl derivatives (mono-oxo), together with other components consisting largely of dihydroxycyclohexyl derivatives (di-ol) and unknown polar metabolites. Metabolites found in rats were pharmacologically active trans-4'-ol, 3'-ol, 4'-oxo, cis-4'-ol, 3'-oxo and 2'-ol in the decreasing order listed, and those in dogs were 3'- and/or 4'-ols. Ciglitazone was highly bound to plasma protein of both animals. After oral administration of 14C-ciglitazone to rats, 14C was widely distributed in tissues, with the highest concn. in the gastrointestinal tract, followed by liver, adipose, plasma, adrenal gland, kidney, pancreas, spinal cord, heart and lung, and the lowest in the brain. The concn. of 14C in erythrocytes of rats and dogs was very low, as was the level of 14C in rat fetuses. Elimination of 14C-ciglitazone was complete within 96 h in rats and 144 h in dogs. In both animals, the dosed 14C was excreted largely in faeces as metabolites, with the remainder appearing in urine. Biliary excretion and reabsorption of 14C were obvious in rats. In both rats and dogs, the major metabolites found in faeces were 3'- and/or 4'-ols and other components derived from bile, and those in urine were other components. On repeated oral administration of 14C-ciglitazone to rats for seven days, no accumulation of 14C occurred in plasma and tissues, and 97.5% of the dose was eliminated from the body within 96 h after the last administration.

Exogenous Cu,Zn-superoxide dismutase suppresses the stimulation of neonatal rat hepatocytes' growth by tumor promoters.

A single exposure to a low concentration (10(-10) mol/l) of tumor promoters [such as 12-O-tetradecanoylphorbol-13-acetate (TPA), phenobarbital and nafenopin] or hormones [such as epidermal growth factor (EGF), glucagon and insulin] or drugs [such as imidazole and indomethacin] stimulated the 24-h flow into DNA synthesis and mitosis of primary neonatal rat hepatocytes incubated in high-calcium (1.8 mmol/l) Eagle's FBS(10% v/v)-MEM. However, only tumor promoters acted as enhancers of hepatocytic DNA synthesis when a low-calcium (0.01 mmol/l) FBS-MEM was used. The activity of tumor promoters was totally suppressed by the simultaneous (or nearly such) addition of low doses (from 25.0 to 0.25 micrograms/ml; activity, from 100 to 0.7 U/ml) of exogenous bovine liver and ox and dog erythrocyte superoxide dismutase (SOD), independent of the calcium concentration of the medium. Even at the minimal dose administered, SOD effectively inhibited the stimulatory actions of TPA concentrations up to 10(-6) mol/l. SOD's blocking effect depended upon its enzymatic activity, as it was prevented by a specific inhibitor of SOD, sodium diethyldithiocarbamate (DDC). By contrast, SOD did not inhibit the growth stimulation elicited by hormones and drugs in hepatocytes maintained in high-calcium FBS-MEM. Moreover, several tumor promoters (namely TPA, phenobarbital, nafenopin, saccharin, teleocidin, benzoyl peroxide, BHT, DDT, lindane, clofibrate and melittin) stimulated DNA synthesis even when the hepatocytes were incubated in the serumless HiWoBa2000 medium, whatever its calcium concentration. In this synthetic medium, tumor promoters' stimulatory activity was again completely inhibited by the simultaneous administration of exogenous SOD. Known antioxidants such as retinoids, vitamin E, selenous acid, and 7,8-benzoflavone, when given simultaneously with TPA, also prevented the stimulation of hepatocytic growth. These results disclose the existence of two quite different mechanisms by which the growth of neonatal rat hepatocytes can be stimulated: (i) the physiological-pharmacological extracellular calcium-dependent SOD-insensitive system mediating the effects of EGF, glucagon, insulin, imidazole, and indomethacin; and (ii) the pathological extracellular calcium-independent SOD- and antioxidant-suppressible mechanism operated by agents belonging to the tumor promoters class and involving, as a critical step, the generation of superoxide anions on the surface of the hepatocyte plasmalemma.

Increase of Na+ gradient-dependent L-glutamate and L-aspartate transport in high K+ dog erythrocytes associated with high activity of (Na+, K+)-ATPase.

As reported previously, some dogs possess red cells characterized by low Na+, high K+ concentrations, and high activity of (Na+, K+)-ATPase, although normal dog red cells contain low K+, high Na+, and lack (Na+, K+)-ATPase. Furthermore, these red cells show increased activities of L-glutamate and L-aspartate transport, resulting in high accumulations of such amino acids in their cells. The present study demonstrated: (i) Na+ gradient-dependent L-glutamate and L-aspartate transport in the high K+ and low K+ red cells were dominated by a saturable component obeying Michaelis-Menten kinetics. Although no difference of the Km values was observed between the high K+ and low K+ cells, the Vmax values for both amino acids' transport in the high K+ cells were about three times those of low ones. (ii) L- and D-aspartate, but not D-glutamate, competitively inhibited L-glutamate transport in both types of the cells. (iii) Ouabain decreased the uptake of the amino acids in the high K+ dog red cells, whereas it was not effective on those in the low K+ cells. (iv) The ATP-treated high K+ cells [(K+]i not equal to [K+]o, [Na+]i greater than [Na+]o) showed a marked decrease of both amino acids' uptake rate, which was almost the same as that of the low K+ cells. (v) Valinomycin stimulated the amino acids' transport in both of the high K+ and the ATP-treated low K+ cells [( K+]i greater than [K+]o, [Na+]o), suggesting that the transport system of L-glutamate and L-aspartate in both types of the cells might be electrogenic. These results indicate that the increased transport activity in the high K+ dog red cells was a secondary consequence of the Na+ concentration gradient created by (Na+, K+)-ATPase.

Effects of sulfhydryl and other reagents on the diffusional permeability of dog erythrocytes to small solutes

The effects of p-chloromericuriphenylsulfonic acid (PCMBS), 5,5′-dithiobis (2-nitrobenzoic acid) (DTNB), phloretin and thiourea on the diffusional permeability of dog erythrocytes to tritiated water and to small 14C-labeled lipophilic and hydrophilic solutes were measured at 37°C by means of the linear diffusion technique. Permeability to 3HHO was significantly decreased by PCMBS but was not affected by the other reagents. The permeability to the small hydrophilic solutes acetamide and urea was decreased by phloretin and thiourea but only the permeability to acetamide was reduced to a statistically significant extent by PCMBS. The permeability to the lipophilic solutes methanol, ethanol and antipyrine was not affected by any of these agents. We interpret these results as an indication that the small lipophilic solutes probably move through lipid areas, that the small hydrophilic solutes probably move through protein associated areas in the erythrocyte membrane and that pathways for the small hydrophilic solutes are distinct from those for water. While the pathways for water may be associated with membrane protein they do not appear to be associated specifically with band 3 protein as has been suggested for human erythrocytes. Diffusional water movement through the dog erythrocyte occurs by two distinct pathways.

Arachidonic acid metabolism by erythrocytes.

Rabbit, chicken, rat, and dog erythrocytes (10(9) cells/ml) synthesized immunologically active 12-hydroxyeicosatetraenoic acid (12-HETE) when stimulated by the Ca2+ ionophore, A-23187. The levels of immunologically active hydroxyeicosatetraenoic acid were independent of the number of white blood cells and platelets in the erythrocyte suspensions. Two products were resolved by high performance liquid chromatography; one product was identified as 12-HETE, while a second product appeared to be a dihydroxyeicosatetraenoic acid. Radiolabeled arachidonic acid was incorporated into phospholipids. Phosphatidylcholine and phosphatidylethanolamine were primary sources of the 12-HETE and dihydroxyeicosatetraenoic acid, all of which were released from the cells.

Action and binding of palytoxin, as studied with brain membranes.

Palytoxin in concentrations as low as 10(-11) to 10(-12) M promotes the outflow of the lipophilic [3H]tetraphenylphosphonium ion from particulate brain cortex of guinea-pigs and rats, and from preloaded crude synaptosomes of rats, which indicates depolarization. The outflow is not influenced by tetrodotoxin or the calcium channel blocker nimodipin, or by substitution of choline for Na+ ions. It is increased by Ca2+ and by borate, the latter interacting with the toxin itself. To assess the fixation of palytoxin to biological membranes, a binding step was installed before the depolarization step. Palytoxin binds to membranes from rat brain, liver, kidney, human and dog erythrocytes, and to a lesser degree to liposomes made from rat brain or erythrocyte lipids. Binding is reversible. It is decreased by mild physical pretreatments of crude synaptosomes. Palytoxin binding is increased in the presence of micromolar concentrations of Ca2+ or borate. It is concluded that the potentiation of palytoxin actions by Ca2+ or borate is at least partially due to the promotion of its binding.

Increase of Na-K-ATPase Activity, Glutamate, and Aspartate Uptake in Dog Erythrocytes Associated With Hereditary High Accumulation of GSH, Glutamate, Glutamine, and Aspartate

We have found convincing evidence for the presence of Na-K-ATPase and high potassium (K) and low sodium (Na) concentrations in the erythrocytes of some dogs associated with hereditary high concentrations of erythrocyte glutathione and some amino acids, glutamate, glutamine, and aspartate. The Na-K-ATPase activity of the erythrocyte membranes of the dogs was about 3 times that of human erythrocyte membranes, whereas the enzyme activity was not detected in control dogs with a normal level of blood glutathione. The Michaelis constant of the enzyme for ATP (Km ATP) was 6.6 X 10(-3)M in the dogs' erythrocytes and 5.0 X 10(-4)M in the human erythrocytes. The concentration of K in the erythrocytes in the dogs examined was about 11 times that of the controls, whereas the erythrocyte Na concentration in the dogs was about one-third that of the controls. The concentrations of K and Na in the plasma of the dogs were equal to those of the controls. Furthermore, L-3H-glutamate and L-3H-aspartate uptake by those cells with high activity of Na-K-ATPase greatly increased, while L-3H-glutamine uptake was unchanged. It appeared that Na+ and K+ gradients created by Na-K-ATPase across the cell membrane might stimulate glutamate and aspartate uptake by the cells, thus causing the high accumulation of such amino acids in the cells.

Increase of Na-K-ATPase activity, glutamate, and aspartate uptake in dog erythrocytes associated with hereditary high accumulation of GSH, glutamate, glutamine, and aspartate

We have found convincing evidence for the presence of Na-K-ATPase and high potassium (K) and low sodium (Na) concentrations in the erythrocytes of some dogs associated with hereditary high concentrations of erythrocyte glutathione and some amino acids, glutamate, glutamine, and aspartate. The Na-K-ATPase activity of the erythrocyte membranes of the dogs was about 3 times that of human erythrocyte membranes, whereas the enzyme activity was not detected in control dogs with a normal level of blood glutathione. The Michaelis constant of the enzyme for ATP (Km ATP) was 6.6 X 10(-3)M in the dogs' erythrocytes and 5.0 X 10(-4)M in the human erythrocytes. The concentration of K in the erythrocytes in the dogs examined was about 11 times that of the controls, whereas the erythrocyte Na concentration in the dogs was about one-third that of the controls. The concentrations of K and Na in the plasma of the dogs were equal to those of the controls. Furthermore, L-3H-glutamate and L-3H-aspartate uptake by those cells with high activity of Na-K-ATPase greatly increased, while L-3H-glutamine uptake was unchanged. It appeared that Na+ and K+ gradients created by Na-K-ATPase across the cell membrane might stimulate glutamate and aspartate uptake by the cells, thus causing the high accumulation of such amino acids in the cells.

Amino acid sequence and oligosaccharide attachment sites of the glycosylated domain of dog erythrocyte glycophorin

Glycophorin was prepared from dog erythrocyte membranes by extraction with lithium diiodosalicylate and partition in aqueous phenol. Tryptic and chymotryptic treatments of the glycophorin produced two major glycopeptides labeled T1 and CH1, respectively. The glycopeptides were isolated by gel chromatography followed by ion-exchange chromatography, and subjected to amino acid sequence analysis. Both glycopeptides represented the amino-terminal domain of the major dog glycophorin; T1 of 52 residues and CH1 of 43 residues. The amino-terminal sequence of dog glycophorin does not have significant homology with those of human, horse or porcine glycophorins. This result is in good agreement with our previous proposal that there is no homology in the sequence of the amino-terminal glycosylated domain of glycophorin.

Plasma membrane NADH dehydrogenase and Ca 2+-dependent potassium transport in erythrocytes of several animal species

Ca 2+-dependent K + transport and plasma membrane NADH dehydrogenase activities have been studied in several ‘high-K +’ (human, rabbit and guinea pig) and ‘low-K +’ (dog, cat and sheep) erythrocytes. All the species except sheep showed Ca 2+-dependent K + transport. NADH-ferricyanide reductase was detected in all the species and showed positive correlation with the flavin contents of the membranes. NADH-cytochrome c reductase was very low or absent in dog, sheep and guinea pig membranes. No correlation was found between NADH dehydrogenase and Ca 2+-dependent K + channel activities in the species studied. Nor were any of the above activities correlated with ( Na + + K +)- ATPase activity.

Studies on the anti-inflammatory mechanism of butyl flufenamate (HF-264)

Anti-inflam-matory mechanism of butyl flufenamate (HF-264) was studied by various analysis. HF-264 did not inhibited the heat denaturation of bovine serum albumin, whereas flufenamic acid showed a marked inhibition on it. HF-264 and flufenamic acid prevented the dog erythrocytes from heat induced lysis, and inhibited the production of superoxide anion in macrophages. HF-264 inhibited the synthesis of prostaglandin E1, but its inhibition of HF-264 on it was more weaker than that of flufenamic acid. From the above results, it is suggested that some part of potent anti-inflammatory action of HF-264 is the results of the interaction of HF-264 with biomembrane, as membrane stabilization and inhibition of superoxide anion production.

Hemagglutination by Bordetella bronchiseptica.

A total of 53 isolates of Bordetella bronchiseptica from dogs and pigs were tested for their ability to agglutinate chicken, horse, sheep, dog, pig, and guinea pig erythrocytes. No differences in hemagglutinating activity were attributed to the animal origin of the bordetella isolates. Horse and dog erythrocytes consistently resulted in the strongest hemagglutination reactions, whereas only 4% of the B. bronchiseptica isolates produced weak agglutination of chicken erythrocytes. A total of 85% of the isolates agglutinated horse, sheep, dog, pig, and guinea pig erythrocytes. One canine isolate with hemagglutinating activity, strain 110H, was examined to determine the nature of the hemagglutinin(s) involved. Hemagglutination was always accompanied by hemadsorption, as determined by dark-field or phase-contrast microscopy. Treatment of cells and cell extracts with heat or protease K inhibited the hemagglutination reaction. Sonicated bacterial cells had a greater hemagglutinating ability than did unsonicated live bacteria. The hemagglutination reaction was not inhibited by any of 17 sugars nor by N- acetylglucosamine or ethylene glycol-bis-(beta-aminoethyl ether)-N, N-tetraacetic acid. Hemagglutinins were not detected in sonic extracts nor in several bacterial subunit fractions, including isolated pili. Antigens in some of these preparations were, however, detectable by indirect hemagglutination with anti-B. bronchiseptica serum. Isolated pili could not be detected on the erythrocyte surface by electron microscopy; however, serial sections of erythrocytes agglutinated by the live Bordetella organisms showed that the bacterial outer membrane and the erythrocyte surface were separated by a space of approximately 20 nm. This study provided additional circumstantial evidence that B. bronchiseptica pili or at least heat-labile surface proteins which extend some distance from the bacterial surface are involved in hemagglutination. Multiple hemagglutinins are likely to exist within this species since one isolate lacking pili also agglutinated canine erthyrocytes. The hemagglutinins of B. bronchiseptica need to be isolated and characterized before the hemagglutination reaction can be applied to studies of attachment.