Search

Summary Homozygous transferrins of adult cattle are made up of two strong pairs and one weak pair of protein bands on starch gel electrophoresis. Foetal transferrins have only the slower band of each pair with the fastest band of the three being much stronger than in the adult type. Before term the second band of each pair begins to develop and at the same time the fastest pair becomes weaker – attaining the adult type by term or soon after. The ai protein, which is present in early foetal life and almost disappears by 250 days of embryonic development, shows individual variation. Its relationship to fetuin is discussed.

Summary Amylase activity has been estimated in the sera of 213 cattle by an amyloclastic method with potato amylose as the substrate with the addition of CaCl2. The method has been developed from that of Street and Close to overcome the problems of the high colour of bovine serum which would otherwise interfere with the assay. A relationship has been found between the serum amylase genotype detected on starch-gel and the total enzyme activity. The enzymes AmI-B and AmII-B in homozygous combinations have been found to have the highest activities. Age of animal has no significant effect on the total enzyme activity. Activity has been investigated in sera from three separate cattle populations and it has been found to be a highly heritable character. Heritability estimates of 49 % and 39 % by sib analysis and 75 % by the twin method have been obtained.

Summary Two genetically controlled polymorphic systems of amylase enzymes are described in sera from Friesian, Jersey, Guernsey, Hereford, Ayrshire and Charolais cattle. One of these Ami has been described previously and is controlled by two codominant alleles in British cattle, AmIB and AmIC. Evidence presented here suggests that it could be a gamma.amylase. The other enzyme which is described here for the first time has many of the properties of an alpha amylase. It is also polymorphic and controlled by two codominant alleles AmIIA and AmIIB. AmIIA is only present in low frequency. The inheritance of both enzyme systems has been studied in 419 families.

BoLA specificities wl w2 w5 w8 w9 wlO and wll have been re-examined by absorption. All of the clusters except w9 showed evidence of sub-group specificities, but only in some cases (wl, w5, v8 and wll) is it suggested that the definition should be revised. The results show a very high degree of reproducibility between sera from different laboratories with clusters of narrow and wide sera for most specificities. These behave in all the cases reported here as linear sub groups or inclusion groups.

SUMMARY Polymorphism of the bovine DRB, DQA, DQB, DYA, DOB and DIB genes was investigated using restriction fragment length polymorphism (RFLP) analysis, isoelectric focusing (IEF), class II serology and polymerase chain reaction (PCR) based typing techniques. The simultaneous application of multiple typing techniques and the characterization of multiple genes resulted in a greatly enhanced picture of the bovine class II regions. Thirty-eight class IIa (DR-DQ) and 5 class lib (DYA-DOB-DIB) haplotypes were defined. It was found that IEF types were associated with DRB3 polymorphism defined by DRB3 PCR-RFLP and DRB3 microsatellite PCR. Serologically defined polymorphism was associated with distinct molecular/IEF motifs and, therefore, DR and DQ specificities could be tentatively distinguished. Although the DR and DQ genes are tightly linked, neither DR nor DQ typing defined all of the class IIa region polymorphism. Furthermore, even the most powerful DRB3 typing technique, DRB3 PCR-RFLP, failed to detect all expressed DRB3 polymorphism. All detected DRB3 polymorphism could, however, be distinguished with a combination of two molecular techniques: DRB3 PCR-RFLP and DRB3 microsatellite PCR. RFLP typing with transmembrane probes detected significantly less polymorphism than typing with cDNA or exon probes. However, the transmembrane probes were useful because they were locus specific. The presence of only 5 of 12 possible class lib haplotypes was unexpected and indicates that the DYA, DOB and DIB genes are tightly linked.

Altogether 424 Norwegian AI bulls, progeny tested for clinical mastitis, ketosis and fertility (recorded as nonreturn percentage), were typed by Edinburgh and Oslo allo-antisera to bovine lymphocyte antigens (BoLA-A) over a 7-year period. Significant effects of BoLA-A on disease were revealed. A2 was associated with relative resistance to mastitis, a positive influence on fertility, and a possible relative resistance to ketosis, while A13 was associated with relative resistance to ketosis. The previously reported associations of A11 and w16 with relative susceptibility to mastitis were not confirmed in the present material.

Peripheral blood lymphocytes (PBL) from cattle have been separated into T & B cell subpopulations using a panning technique. These T and B cell preparations have then been tested by direct and indirect complement mediated cytotoxicity tests with a series of monoclonal antibodies (Moabs) reacting with HLA class I and class II products. The anti-class I monoclonals were monomorphic or non-reactive and where reactive killed both B and T cells. The anti-class II monoclonals reacted only with B lymphocytes. There was no relationship between their reactivity pattern and any BoLA specificity.

The amylase (Ami) phenotypes have been determined in sera collected from 244 heifers before 6 months and after 2 years of age. Fifty of these animals exhibited a change in phenotype from Ami C to an apparent Ami BC with age. Using a modified electrophoretic technique these apparent heterozygotes were reclassified as Ami C(E). Eight hundred and nineteen cows, which had previously been shown to have a highly significant excess of heterozygotes at the Ami locus, were also retyped by this modified technique. The apparent heterozygote excess was not found as it was now possible to differentiate between Ami C(E) and Ami BC.

An earlier report that there is a high incidence (50 to 60 per cent) of cytotoxic antilymphocyte antibody in the sera of parous ewes is confirmed. This antibody is directed against paternally inherited antigens of the fetus. Antibody could be detected pre partum in four out of 34 primiparous ewes. However, there was no strong association between parity and the incidence of antibody. Antibody was also found in three out of 10 older rams which implies that pregnancy is not the only stimulus for antibody production.

The two stage freezing technique was adapted for use with sheep lymphocytes. Parameters investigated were lymphocyte concentration, composition of thawing medium and DMSO concentration. The main difference between this and earlier published techniques is the use of a high DMSO concentration (17%). The technique adopted was both simple and reliable. It consistently gave lymphocyte viabilities of 95% or more in thawed cell suspensions. The procedure is apparently without effect on lymphocyte antigen expression and thus appears very suitable for use in microlymphocytotoxicity tests.

SUMMARY Lymphocytotoxic sera have been produced in forty-seven cattle following skin grafting or lymphocyte immunization between dam-offspring pairs. From these sera seventy antilymphocyte reagents have been produced. Seven of the sera were operationally monospecific without absorption, thirty-seven others were made operationally monospecific and twenty-six partially purified by absorption with lymphocytes. Thirty-five of these sera were used in France and an additional thirty-five were used in Scotland. The inheritance of the antigenic factors detected by these sera has been studied in 480 cattle families and the evidence is consistent with the hypothesis that at least seventeen of the factors are inherited in a simple Mendelian fashion. Fourteen of these factors show evidence of linkage with at least one other factor and are likely to be controlled by a limited number of closely linked loci. It is suggested that these loci should be identified as the BLA region.

1. 1. Individual components of two of the variants of sheep (Ovis aries) transferrin were isolated. 2. 2. There was one main component with three residues of sialic acid per mole of protein and a faster component with four residues. 3. 3. There was no evidence for pairs of bands with the same number of sialic acid residues, as seen in cattle (Bos taurus). 4. 4. Sedimentation equilibrium ultracentrifugation showed that the bands of a single variant have a molecular weight of 77,000. 5. 5. Finger print studies showed at least 4 peptide differences between the two genetic variants. The amino acid composition was similar to that of cattle transferrin. 6. 6. The results provide evidence that sheep transferrin is a single polypeptide chain.

Homozygous cattle transferrin has been fractionated into six main peaks by DEAE-Sephadex chromatography. These correspond to the six transferrin components seen in starch gel electrophoresis of normal serum. In addition, six more minor components were isolated from DEAE-Sephadex, making 12 in all, and these could be divided into six pairs. Treatment of whole transferrin with neuraminidase yielded only two bands. Treatment of the individual fractionated bands showed that the slower band of each pair had the same mobility as the slow band from treated whole transferrin, while the faster band from each pair corresponded with the fast band of treated whole transferrin. These observations, and the results of sialic acid assays, showed that the difference between the pairs of bands was caused by differing numbers of sialic acid residues (0–5) per molecule of protein, but that sialic acid was not responsible for the difference between the bands within a pair. The mode of genetic control is discussed and probably involves three loci. Other physicochemical properties of cattle transferrins, namely, molecular weight, effect of iron addition, and behavior in isoelectric focusing, were also studied.

Friesian cows were inseminated with semen mixtures containing equal numbers of spermatozoa from a Friesian and a Hereford bull. The five bulls of each breed gave twenty-five possible combinations. The paternity of calves was established by inspection of colour and conformation. Heterospermic indices were calculated to express the relative ability of sires to father offspring after mixed insemination. There were significant differences between the heterospermic indices of bulls, the maximum observed difference being twenty-one-fold. The indices were consistent over two series. The homospermic index was defined as the 16-week non-return rate after normal single first inseminations. The heterospermic index established differences between bulls more efficiently than the homospermic index; one estimate showed that the heterospermic method needed less than 1/170th the number of inseminations required by the homospermic method. The homospermic index was predictable from the heterospermic index, the regression coefficient having a significance level of 0\m=.\05> P > 0\m=.\025.The initial spermatozoan concentration of a bull's semen (before dilution) was highly correlated with the heterospermic index. Measures of semen quality based on the morphology and staining affinity of spermatozoa predicted heterospermic and homospermic indices non-significantly but in the right direction.

A method of micro-titration for cytolytic antibodies is described. Standard drops of diluent are placed on a subdivided siliconed slide and dilutions are made by rotating a wire loop holding 1 drop of serum in each drop sequentially. One drop of a mixture of cells and complement is then added. The slide is incubated in a moist chamber for 90 min at 37° C. after which a small cover-slip is placed on each drop and the amount of cell lysis, judged by phase-contrast microscopy, is graded.An analysis of variance of the method has been performed and the method has been found to be reliable, accurate and unbiased.It allows accurate estimates of cytolytic activity in sera using cell concentrations of 104to 105cells per ml. Eight ten-dilution titrations can be carried out with 1 ml. of such a suspension.

Treatment of detection problems from a Bayesian viewpoint implies the modeling of uncertain parameters as random variables with known a priori distributions. Attempts to compare performance of the certain parameter case with the uncertain parameter case become difficult because, for the former, performance is usually dependent on the value of the parameter in question, whereas for the latter, this dependence is removed by averaging. A method for a meaningful solution to this problem is proposed here. An auxiliary receiver, the Externally Sensed Parameter (ESP) receiver, is defined, and its performance is used as the standard for comparison purposes. The comparison is meaningful because the ESP receiver performance is independent of the value of any particular parameter in question and is an upper bound in the sense of perfect parameter information. The method appears to be particularly applicable to evaluating learning or adaptive acoustic signal processing systems and to determining the desirability of incorporating additional parameter estimation techniques in these systems.

An experiment was conducted to examine the addition of zinc in nutritive rich feed contained lemuru oil to growth and birth weight of Bali cattle. The experiment design used was randomized complete block design. The animal were randomly assigned into four group of treatments, i. e addition of 0, 75, 150 and 225 mg ZnSO4 kg -1 on nutritive rich feed (NRF) containing lemuru oil. The basal diet consist dried grass while NRF consisted of palm sugar, leucaena and glirisidia leaf meal, rice bran fermentation, fish meal, lemuru oil and urea. The crude protein content of NRF was 23%, while total digestible nutrient (TDN) was 70%. Addition ZnSO4 in the diet did not increase significantly zinc status, prostaglandin E2 (PGE2) concentration and ME retention. However, level of ZnSO4 addition increase significantly (P

Seven cases of an abnormality in bovine transferrins are described. In this abnormality the two fastest strong protein bands apparently have the same mobility as the two slow bands. It has been seen in Tfa2, Tfd1, and Tfd2 transferrins. The characteristic which has only been found in the Hereford breed could be explained by a recessive epistatic gene affecting the sialic acid binding to the two faster strong protein bands of the four controlled by each Ifallele.

SUMMARYThe amylase (Am I) genotypes have been determined in sera from 834 cows comprising 562 Friesians, 77 Ayrshires, 106 Guernseys, 62 Jerseys and 27 crossbred cows. All of these cows had completed their third lactation and many were much older. There was a highly significant excess of heterozygotes at the Am I locus. This excess of heterozygotes was highest in high-yielding Friesian herds and there was a significant correlation between percentage heterozygotes and herd yield in the Friesians.

This paper presents evidence to show that previous seemingly conflicting reports of the differential agglutinability of bovine red cells are in fact only different interpretations of the same character. The variation in agglutinability of ox red cells is shown to be consistent in anti-globulin systems using several species of sensitising and anti-globulin antisera. The type of agglutinability is inherited and controlled by genes at one locus Agg. High agglutinability Agg I is recessive to low Agg L. The frequencies of Agg I and Agg L in some British breeds have been studied.

Bovine alloreactive cytotoxic lymphocyte (CTL) lines of known target specificity were infected in vitro with sporozoites of Theileria annulata and T parva and cultured in limiting dilution. The phenotypes of the CTL lines both pre- and post infection were assessed using a panel of monoclonal antibodies specific for defined bovine lymphocyte subpopulations. The effector function of the resultant infected cell lines was determined using a Cr51 release assay and compared to the uninfected control CTL line. The results indicated that T parva sporozoites consistently infected and transformed the CTL lines very efficiently even at the lowest cell doses. In contrast the T annulata sporozoites were largely unable to infect and transform the alloreactive CTL except at the very highest cell and sporozoite doses. A factor which appeared to influence susceptibility to T annulata infection was an increased level of class II expression on the CTL line. None of the cell lines showed cytotoxic effector function after infection with either T annulata or T parva sporozoites.

BoLA w6 is one of 16 specificities agreed upon at the 2nd International BoLA Workshop and many laboratories have produced sera reacting with this specificity. This paper presents evidence for at least four sub-groups in w6. Three Edinburgh sera showing identical reaction patterns in the 1st and 2nd BoLA workshops have been studied by absorption. One serum contained a single antibody population reacting with an epitope common to all w6 positive cells. The other two contained antibodies against the same common epitope, antibodies to epitopes on four sub-groups and an antibody reacting only with one subgroup w6.1. Another Edinburgh serum contained two populations of antibodies. One reacting with the common epitope and one with an epitope on two or possibly three subgroups. Immunisation between w6 positive individuals produced antisera to two subgroups without producing antibodies to common epitopes. At least two additional subgroups are likely to exist. These results indicate the presence of specificities unique to individual subtypes coupled with epitopes common to some and all w6 subgroups.

In this paper family studies are presented which support the hypothesis of polymorphism in the process controlling sialic acid binding to bovine transferrin which modifies its phenotype as seen in starch gel electrophoresis. It has been shown that this polymorphism is controlled by locus Tfs with two alleles Tfs A and Tfs a. Tfs a/a animals have the abnormal phenotype with the two faster bands of the four bands of a normal transferrin allele being virtually absent. Tfs A/a and Tfs A/A are phenotypically normal. Limited evidence is presented which suggests that the Tf and Tfs loci are not linked.

Using monospecific sheep antisera, the mean plasma immunoglobulin concentrations in 82 Ayrshire milking cows were found, by the single radial diffusion method, to be: IgG2 10-91 plus or minus 4-30, IgG1 9-76 plus or minus 2-59 and IgM 2-71 plus or minus 1-22 mg/ml (plus or minus SD). IgA levels in the same group of 82 cows were determined by a simple, sensitive coated tube radioimmunoassay and had a mean value of 0-32 plus or minus 0-16 mg/ml. The frequency distributions of IgG1, IgM and IgA were approximately log normal, whereas that of IgG2 was bimodal with two populations with means of approximately 7 and 14 mg/ml. An 18-fold difference was found between high and low IgG2 levels whereas only three- to four-fold differences were recorded for IgG1 levels in the same population. Eight-fold differences were found for IgM and IgA levels. The effect of age (2-9 years) on sampling could be disregarded for IgG1, IgG, IgM and IgA levels, but cows above 6 years of age had significantly higher IgG2 levels.

Bovine peripheral blood mononuclear cells (PBMC) were labelled with monoclonal antibodies recognizing bovine MHC class II, sIgM, monocyte, T-helper and T-cytotoxic cell phenotypes. They were sorted into positive and negative populations with a fluorescence-activated cell sorter (FACS). The cell populations were infected in vitro with sporozoites of either Theileria annulata or T. parva, and the degree of infection and transformation determined. The results showed that despite the many similarities between these two parasites, they infected different cells of the immune system. T. annulata preferentially infected MHC class II-positive cells but did not infect T cells. Monocytes were infected very efficiently by T. annulata but were uninfectable with T. parva. B cells were infected much more efficiently by T. annulata than T. parva. Cell lines derived from infections with T. annulata were analysed phenotypically. Virtually all reactivity was lost for the anti-sIgM and the anti-monocyte monoclonal antibodies post-infection and no T-cell markers were detected.

Summary 1. The inherited differential agglutinability of cattle erythrocytes is shown to be similarly expressed on ghosts and intact cells. 2. Removal of virtually all sialic acid by prolonged neuraminidase treatment does not alter the agglutinability status of ghosts prepared from either high or low agglutinable cells. Hence the differing sialic acid content of the two cell types is not responsible for the differential agglutinability. 3. The significance of these findings with respect to other well defined agglutination systems and current theories of membrane structure is discussed.

Purified populations of bovine antigen presenting cells (APCs) and T cells have been isolated from peripheral blood and characterised using various monoclonal antibodies (mAbs) for cell surface markers. Bovine APCs were found in an adherent cell fraction and were non-specific esterase positive, phagocytic and expressed bovine major histocompatibility complex (MHC) class II determinants, all of which are typical macrophage characteristics. T cells were rigorously depleted of accessory cell function before being used in an antigen presenting cell assay. The generation of T helper cells in response to the soluble antigen, ovalbumin, was entirely dependent upon a critical number of APCs. Further the proliferative response was inhibited by several mAbs to bovine MHC class II molecules. Thus the interaction between bovine APCs and helper/inducer T lymphocytes (TH/I) appears to be similar to that in other species.

Alloreactive cytotoxic T cells (CTL) were generated in mixed lymphocyte culture against cells bearing subgroups of BoLA w6, as well as in BoLA w4, w10 and w16. Primary cultures were restimulated at weekly intervals with irradiated stimulator cells and tested in a 51Cr-release assay with target cells derived from Theileria annulata-infected cell lines. Generation of CTL was accelerated in animals that had been previously primed in vivo by skin grafting. CTL were generated that were specific for BoLA w6 subgroups and not w6. With w6.1 the specific killing was significant at the 5% level, and with w6.2, 6.3 and 6.4 it was significant at the 0.1% level. Where CTL were potentially generated against two BoLA-A locus allele products at the same time (i.e. with heterozygous stimulator cells), one of which was a w6 subgroup, there was similar CTL activity against both products when w6.4 and w16 or w6.1 and w4 were the combinations involved. In two generations against w10 and either w6.1 or w6.2 there was significantly more killing of w10-bearing targets than those with the w6 subgroup.

This paper reports on a study of1,033 Angoni cattle in the Eastern Province of Zambia. The animals came from two areas, one of which had frequent outbreaks of trypanosomiasis whilst the other was relatively free of clinical trypanosomiasis. Whole body tick collections were made from each animal and blood samples taken together with data on colour, sex, age and skinfold thickness. The blood samples were tested for parasites, red cell antigens and polymorphism at the transferrin (Tf), haemoglobin (Hb), albumin (Alb), amylase I (Am I) and amylase II (Am II) loci.

In an investigation of the expression of major histocompatibility complex (M.H.C.) antigens by cells infected and transformed in vitro by Theileria parasites, Spooner and Brown (1980) reported that the qualitative expression of such antigens remained unaltered by infection and transformation. However, in one instance cells taken from a dizygous twin and transformed by Theileria parva (Muguga) were found to have a different pattern of reactivity with the Edinburgh panel of Bovine Lymphocyte Antigen (BoLA) typing sera from those of the fresh peripheral blood lymphocytes (PBL) from the donor. Conversely PBL from the same animal transformed by T. annulata (Hissar) exhibited the same BoLA phenotype as uninfected PBL. The donor was a chimaera and it was suggested that the T. parva (Muguga) cell line arose from cells of a minor population within the chimaera, the inference being that this minor population was too small to be detectable in the standard BoLA-typing test. The results of BoLA typing of the parents of the cell donor were consistent with this hypothesis.

This paper presents evidence for an IgG1 allotype detected by a sheep antibovine serum. The character which appears to be inherited in a simple Mendelian way has been named G1a1.

The bovine major histocompatibility system (MHS) as defined by bovine lymphocyte antigens (BoLA) detected by microlymphocytotoxicity testing was described by Spooner et al (1978) and Amorena & Stone (1978). Subsequently international collaboration between the laboratories active in this area led to agreement at the 1st International BoLA workshop on eleven specificities (Spooner et al, 1979). Population evidence (Spooner et al, 1979) and family studies (Oliver et al, 1981) indicated that these were controlled by codominant alleles at a single locus. Evidence for the considerable prolongation of graft survival of grafts between haplo-identical unrelated individuals (Spooner & Oliver, unpublished; Amorena & Stone, 1978) supports the assumption that BoLA is the bovine MHS.

Summary. Two hundred and eighty-two alloantisera were submitted by 20 participating laboratories from 13 countries and tested against lymphocytes of 1298 cattle. The cell panel consisted of samples from 38 Bos taurus breeds, 11 Bos taurus crossbreeds, 4 Bos indicus breeds, 6 Bos taurus X Bos indicus, and a variety of other crossbred populations. Using a standardized lymphocytotoxicity test, all 17 previously identified BoLA specificities were confirmed. The workshop produced agreement on 16 new lymphocyte alloantigenic specificities. Three of the new specificities behaved as splits of previously identified BoLA specificities. Four of the new specificities behaved as alleles at the agreed BoLA-A locus. Seven new specificities are tentatively assigned to the BoLA-A locus but require further definition. Two new specificities may represent products of a second closely-linked BoLA locus.

Summary This paper describes the effect of gel buffer pH on the resolution of bovine serum amylase (Amylase I) isozymes in starch gel and the consequences for the understanding of the genetics of this locus. The two main findings are: (1) the existence of a satellite isozyme E to isozyme C which at pH 7.3 has the same mobility as the B isozyme but which at pH 8.0 migrates slower than B, and (2) the finding of three alleles AmI A, AmI B and AmI C in British cattle populations previously reported as having only AmI B and AmI C.

A simple two-stage technique for preserving bovine lymphocytes is described. Lymphocytes from animals chosen at random were used. The experiments indicate that the optimum temperature for freezing and the optimum concentration of dimethylsulphoxide (DMSO) as cryoprotectant were in the range -29 degrees C to -31 degrees C and 17.5% to 20% respectively. These concentrations of DMSO are much greater than those reported in most other studies.

Recovery of calves from tropical theileriosis was accompanied by the disappearance of macroschizonts from lymph nodes and the appearance of cytotoxic cells in the blood and lymph nodes. Acute, fatal disease was associated with incremental parasitosis and parasitaemia and, in general, an absence of detectable cytotoxic cells in the blood or lymph nodes. After recovery from infection, calves were resistant to challenge. Challenge with sporozoites was followed sometimes by an immediate reappearance or by a later peak, or sometimes by twin peaks of cytotoxic cells but macroschizonts were not detected. Histocompatibility (BoLA) typing indicated that calves produced two sequential populations of cytotoxic cells during recovery from primary infection with Theileria annulata. The expression of lysis by the first appeared to be BoLA restricted. In contrast, both the peaks of lysis manifest after challenge appeared to be BoLA restricted. Results suggest that BoLA restricted cells are established in the immunological memory and are probably analogous to cytotoxic T cells, while non-BoLA restricted cytotoxic cells are natural killer like cells. The results suggest a role for cytotoxic cells in recovery from primary infection, in the inhibition of proliferation of macroschizonts which evade mechanisms of acquired resistance and in the lysis of macroschizont infected cells deriving from challenge sporozoites which have evaded serum-mediated inhibition.

Summary Red cell potassium concentrations were determined in 5 breeds of cattle. Most cattle had red cells with less than 50 mmol K/litre and although a few Hereford and Friesian cows had levels approaching 70 mmol K/litre, it was only in the Jersey breed that values above this were found. Attempts were made to correlate potassium levels with serological specificity using cattle S system and sheep M system blood typing reagents. Red cells negative for cattle factors S, H‘, and S were never positive for sheep L; cells positive for H’ were usually positive for M, and those positive for S were always positive for L. No obvious relationship was found between potassium levels and serological type, although some consistent trends were noted in Friesians and Herefords. Limited family data indicated that a major difference in potassium levels of red cells in cattle may be controlled by two codomi-nant alleles.

Three bovine isoantisera to IgC are described. They detect three different specificities. Evidence for the identification of the antigen is presented but so far no evidence for the inheritance of these characteristics has been obtained.

Ox red cell antigens F, V, ε1, J, L, R′, S′ and T′ have been recovered in soluble membrane protein obtained by a butanol extraction method. F, V and ε1 have been studied in more detail and are found in the sialic acid rich fraction of the membrane protein.