130 A. W. NORMAN AND H. F. DELUCA 1964Harrison, H., Harrison, H. & Park, E. A. (1958). Amer. J.

Phy8iol. 192, 432.Joshi, J. G., Dikshit, P. K. & Patwardhan, V. N. (1957).

Indian J. med. Re8. 45, 439.Lees, H. & Kuyper, A. C. (1957). J. biol. Chem. 225, 641.Meyer, J., Bolen, R. J. & Antin, J. (1959). Arch. Biochem.

Biophy8. 81, 340.Neuman, W. F. & Neuman, M. W. (1958). The ChemicalDynamice of Bone Mineral, p. 157. Chicago: Universityof Chicago Press.

Nicolaysen, R. & Nordb0, R. (1943). Acta physiol. scand.5, 212.

Palmer, J. H. (1955). Bull. Conn. agric. Exp. Sta. February,p. 589.

Ramalingaswami, V., Sriramachari, S., Dikshit, P. K.,Tulpule, P. G. & Patwardhan, V. N. (1954). Indian J.med. Sci. 8, 509.

Snyder, F. & Godfrey, P. (1961). J. Lipid Res. 2, 195.Steenbock, H. & Bellin, S. A. (1953). J. biol. Chem. 205,

985.Steenbock, H. & Herting, D. E. (1955). J. Nutr. 57, 449.Umbreit, W. W., Burris, R. H. & Stauffer, J. F. (1957).

Manometric Techniques, p. 149. Minneapolis: BurgessPublishing Co.

Waasjo, E. & Eeg-Larson, N. (1951 a). Acta phy8iol. 8cand.25 (suppl.), 84.

Waasj6, E. & Eeg-Larson, N. (1951 b). Acta phy8iol. scand.25 (suppl.), 89.

Biochem. J. (1964) 91, 130

Studies on Human Casein Preparations from Single Milk Samples

BY F. H. MALPRESS AND F. E. HYTTENDepartment of Biochemi8try, Queen'8 Univer8ity, Belfast, and

Medical Research Council Obstetric Medicine Research Unit, Aberdeen University Mledical School

(Received 14 August 1963)

The reported values for the sialic acid content ofhuman casein vary widely. Malpress (1961, 1962)has given a maximum value of 2x1 % for samplesprepared from milk collected from the fourth to theeighth day after parturition, and a range of 0 30-0-38 % for samples from mature milk. Other valuesgiven for mature milk are 0 76 % (Johansaon &Svennerhohn, 1956), 0-8 and 0-37 % (Alais &Jolles, 1962a, b).

Discrepancies are also present in reports on theextent of rennin action on casein samples: Malpress(1962) found the release in combined form of 30 %of the total sialic acid by this enzyme; Alais &Jolles (1962b) obtained a glycopeptide, containingsialic acid, from some but not all of their caseinpreparations as the result of rennin action.These variations could be the result of the dif-

ferent procedures used for preparing casein.Alternatively they might indicate true individualvariations in caseins from different mothers, orfrom the same mother at different stages of lacta-tion; the complexity of human casein shown by itsheterogeneity on electrophoresis in the presence ofurea (Malpress, 1962; Alais & Jolles, 1962a) sug-gests possibilities of variation in support of thisalternative view.The problem has now been further studied by an

investigation of caseins, prepared by a standardprocedure, from 80 single milk samples taken from47 mothers and covering all stages of lactation.

MATERIALS AND METHODS

Preparation of ca8ein. Milk samples (3-100ml.) werestored at 20 and used for the preparation of casein within1 week of collection. (In a few exceptional cases thesamples were kept frozen for up to 1 month before use.) Themilks were centrifuged at 10 O0Og at 50 for 15 min. toremove fat. The skimmed milks (pH 6.5-7.5) were filteredthrough cotton wool, adjusted to pH 4-7 with HCI and keptat 2° overnight; they were then warmed to 250 for 30 min.(cf. Maeno & Kiyosawa, 1962) and centrifuged at 35 000gat 100 for 30 min. The precipitated caseins were dispersedand recentrifuged twice in 0-01M-acetate adjusted to pH 4*7with N-NaOH. The volume of the skim-milk supernatantswas measured; they were then filtered through cotton wooland retained for the estimation of 'non-casein sialic acid'.The washed casein precipitates were dissolved in water

with minimum amounts of 0-1 N-NaOH, care being taken tokeep the pH less than 10, and the solutions (50-100 ml.)were filtered through a cotton-wool plug and adjusted topH 4-7 with HCI. After storage overnight at 2° the caseinswere collected by centrifuging at 15 0OOg at 100 for 15 min.They were washed twice with ethanol-ether (1:1, v/v) andtwice with ether, and dried in air.

In the present paper 'casein' is defined as the proteinprepared from human milk samples by the foregoingmethod, and 'mature casein' refers to the preparationsderived from milk taken on or after the tenth day oflactation.

Starch-gel electrophore8i. This was carried out at pH 8-6by using the discontinuous tris-citrate-borate buffer systemof Poulik (1957) in the presence of 7M-urea as described byWake & Baldwin (1961). All analyses were made at 20 in a

PROPERTIES OF HUM1AN CASEINPerspex tray (21.5 cm. x 10-5 cm. x 1 cm.) which allowedfour samples to be run simultaneously. Samples, eachrepresenting about 2 mg. of protein, were put into the slotsas 0.3% solutions in 0-076M-tris-citrate buffer, pH 8-6,brought to a glue-like consistency by the addition of smallamounts of starch. The voltage gradient was about 4v/cm.and the analyses were stopped after about 24 hr. when thebrown boundary had travelled 10-15 cm. from the origin.The gels were stained with 0-01% Nigrosin in methanol-acetic acid-water (5:1:4, by vol.).

Estimation of sialic acid. The thiobarbiturate method(Warren, 1959) was used. N-Acetylneuraminic acid (L.Light and Co. Ltd., Colnbrook, Bucks.) was used as astandard and gave a molar extinction coefficient of 70 000at 549 m,u. 'Non-casein sialic acid' values were obtainedafter a twentyfold dilution of the supernatants. Un-hydrolysed-blank estimations were carried out on allsolutions.

Action of rennin. Crystalline rennin (Nutritional Bio-chemicals Corp., Cleveland, Ohio, U.S.A.) was used in afinal concentration of 0-001% to determine the extent ofrennin action on casein preparations. The incubations werecarried out for 2 hr. at 370 in unbuffered solutions of casein(0-1-0.2%) adjusted to pH 5-8; under these conditionsrennin action was usually complete in less than 1 hr., thechange in pH during incubation being negligible. Theenzyme was destroyed by the addition ofan equal volume of24% (w/v) trichloroacetic acid, the solutions were centri-fuged at 35 OOOg at 100 for 30 min. and the clear super-natants were extracted three times with an equal volume ofether. The aqueous layers were evaporated to dryness overP205 in vacuo and the dry residues were dissolved in 0-1 N-H2SO4. Control solutions were incubated without rennin,the enzyme being added after the addition of trichloro-acetic acid.The sialic acid released as glycopeptide soluble in 12%

trichloroacetic acid was measured by the difference in sialicacid values of the test and control solutions, prepared inthis way, after hydrolysis at 800 for 1 hr. It is referred to asthe rennin-sensitive fraction of the casein sialic acid.

RESULTS

Eighty casein preparations were made, from milksamples given by 47 mothers; they representedstages of lactation from the first day to theeleventh month after parturition.

Yields varied from 20 to 500 mg./100 ml. of skimmilk for samples taken on or after the tenth day oflactation (non-colostral milks); these values may becompared with values given by Macy (1949), whoquotes a range of 140-680 mg. of casein/100 ml. ofhuman whole milk, obtained from 166 samples by amethod which involved no reprecipitation of thecasein. In the present work the wide variation inyields showed no correlation with either the stage oflactation at which the sample was taken or with thesample volume. There were differences betweensamples in the readiness with which casein was pre-cipitated at pH 4-7 and 250 from the skimmedmilks, and it is therefore possible that in the presentmethod yields were influenced at this stage. To

Table 1. Relative mobltities of the bands shown byhuman casein in starch-gel electrophoresis

Experimental conditions are given in the text. Thedistance from the origin to the discontinuous boundary was12 cm. The mobilities are related to that of band 1, which istaken as 1-0.

Main Relativebands mobility

0-33- 0-456 0-56- 0-625 0-71

0-764 0-83- 0-863 0-892 0-951 1-00- 1-04- 1-06_ 1-08

1-131-36-1-40

_ 1-43

Subsidiarybands

76'5'

4'

3'

abcd

BoundarysmudgeBoundary

minimize the danger that a low yield might be asso-ciated with uncharacteristic properties in theisolated casein, an arbitrary exclusion of all samplesrepresenting less than 100 mg. of casein/100 ml. ofskimnmilk has been made in the present paper.

Starch-gel electrophoreisi. The characteristic bandpattern given by the human casein samples con-sists of six well-defined bands, all of which movetowards the anode more slowly than the two mainbands given by cow casein under similar conditions(Fig. 1). The bands are here numbered 1-6 inorder, band 1 being that which migrates farthestfrom the point of origin. Table 1 gives the measureof the relative mobilities of the bands (that of band1 being taken as 1-00). In addition values are givenfor a number of other bands which occur both morerarely and, when present, with far less intensity.Of these, bands a, b, c and d have in the presentstudy been mainly, but not exclusively, associatedwith milk samples taken during the first month oflactation; they were present in only a few of thesamples from this period, however, and were notseen in any sample taken earlier than the tenth day.Bands 3', 4', 5' and 6' were rarely detectable withthe casein concentrations normally used (0-3 %);they were sometimes present when this concentra-tion was doubled. Band 7 was seen only once, inassociation with an uncommon pattern (type C).Other features of the stained gels were a weak andconfused smudge, running almost with the boun-dary, given by casein from samples taken early inlactation, and a weak generalized staining extend-ing from the origin to approximately the band 6

9-2

Vol. 91 131

F. H. MALPRESS AND F. E. HYTTENposition which was not always detectable but wasfound in samples at all stages of lactation. This lastobservation gains in significance from the occur-rence ofa similar diffuse and weakly staining area inelectrophoretic patterns of cow casein, where it hasbeen found to represent the rennin-sensitive K-caseinfraction (Wake & Baldwin, 1961). In the presentwork it has been found that the six-band electro-phoretic pattern of human casein is unchanged bythe action of rennin and it is clear that the hetero-geneity ofhuman casein is greater than this patternimplies.A limited study ofcaseins prepared from colostral

milks secreted during the first week of lactationgave a consistent picture of the order of bandappearance, but suggested that the stage of lacta-tion at which any given band appeared might vary.One series is shown in Fig. 2: in this case milksamples were obtained from the third to the sixthday of lactation and the caseins isolated showedbands 1 and 2 increasing in intensity throughout;bands 3 and 4 appeared in trace amounts on thefourth day, and increased in intensity on the follow-ing days; band 5 appeared in trace amounts on thesixth day; band 6 was absent throughout.A second series, started on the first day of lacta-

tion, showed bands 1 and 2 appearing for the firsttime on the third day: two other series showed nobands at all after 4 and 6 days respectively.The full six-band pattern was seen at its earliest

in the present work in a casein sample from milktaken on the tenth day after parturition; in generalthe results indicated that for most mothers whoselactation was graded as 'good' the pattern wasestablished soon after this time, but exceptionswere encountered in which the appearance ofbands, 5 and 6 in particular, was delayed.

Caseins from 32 out of 43 mothers whose milkwas sampled on or after the tenth day of lactationgave a band pattern with intensities conforming totypeA (Fig. 3), in which the intensities decreased inthe order: 2; 4; 1 and 3; 5; 6. During the firstmonth of lactation a modification of this pattemwas sometimes seen in which band 1 was moreintense than band 4; it is thought that this was aconsequence of the serial development of the bandsduring this period, in which band 4 appears laterthan band 1; other minor and temporary dif-ferences might also be expected during this periodfor the same reason, and in assessing the electro-phoretic analyses nine such small variations, occur-ring during the first month of lactation, were classi-fied as 'developmental type A' patterns andincluded in the type A total.

Seven of the mothers gave caseins classified astype B (Fig. 3), in which the intensities decreased inthe order: 2; 1; 4; 3; 5; 6; band 1 had a markedlygreater intensity than band 4, and bands 5 and 6

were either weak or absent; in these cases thevariation from type A persisted beyond 1 monthand could not be regarded as an anomaly of thedevelopmental period. Four of these mothers gavesamples at more than one stage of their lactation,but the type B pattern showed no tendency totranspose to type A as the lactation advanced.Of the remaining four cases two gave electro-

phoretic patterns of type C (Fig. 3), in which theintensities decreased in the order: 4; 2; 3, 5 and 6; 1.In both cases this pattern recurred when furthercaseins were prepared from later milks from thesame mothers. Lastly, two caseins showed a com-plete absence of bands although the lactations wereat 12 and 21 days respectively and subjectivelyassessed as 'fair' and 'good'; later samples werenot available from these subjects.Two mothers (type B and type C) gave samples

simultaneously from both breasts. In both casesthe characteristic pattern (type B or type C) wasfound in one sanple whereas the second showedproperties intermediate between type A and thecharacteristic form; these 'transitional' patternsare referred to in the Discussion section.One mother (type A) gave samples immediately

before and after suckling; the casein band patternswere identical.The six-band pattern was not dependent on the

presence of urea in the starch-gel preparation. Inthe complete absence of urea shrinkage occurredand the bands, although present, and in theircorrect relative intensities, were less well-defined;with 2m-urea the electrophoretic analysis wasidentical with that given in the presence of 7M-urea.

Non-casein 8ialic acid. The sialic acid in com-bined form remaining in the skim-milk super-

EXPLANATION OF PLATE I

Fig. 1. Starch-gel electrophoresis in tris-citrate-boratesystem of: human casein (bands 1-6); cow casein (twobands, the faster travelling almost with the boundary).The experimental conditions are given in the text; 2 mg. ofcasein was applied in each case.

Fig. 2. Starch-gel electrophoresis in tris-citrate-boratesystem of human caseins from milks sampled in earlylactation from the same mother at the following times afterparturition: 3 days; 4 days; 5 days; 6 days. The experi-mental conditions are given in the text; 2 mg. of casein wasapplied in each case. Band 5, which was just present on thesixth day, cannot be seen in the photograph. The prominentbands near the origin (at 3 and 4 days) are not found inthe electrophoresis of mature caseins.

Fig. 3. Starch-gel electrophoresis in tris-citrate-boratesystem of human casein types A, B and C. The experi-mental conditions are given in the text; 2 mg. of casein wasapplied in each case. In the type B example bands b and c(Table 1) can also be seen.

132 1964

The Biochemical Journal, Vol. 91, No. 1

Cow casein

Human casein;

Band ............

tOrigin

Band .....Time afterparturition

6 days

5 days

4 days

3 days

Band

Origin

Type A

Type B

Type C

Band ........

OriginF. H. MALPRESS AND F. E. HYTTEN

6 5 4 32 1

4 32 1

21..i

ps

6 5 4 3 21

Plate 1

Fig. 1

I.

Boundary

.jI

Fig. 2

I.

Boundary

i-

IFig. 3

1Boundary

(Facing p. 132)

*1

Vol. 91

natants after precipitatifrom 237 mg./100 ml. on6-5 mg./l00 ml. after 11:sialic acid associated M

240

6

._q

OQ

6

0

.6

-CE8

c;

CQ0.zcem

I

Time after

PROPERTIES OF HUMAN CASEIN 133on of casein at pH 4-7 fell compared with the non-casein sialic acid, except inthe first day of lactation to advanced lactation when the latter values are verymonths (Fig. 4). Since the low, the distribution of samples in the scattervith casein is very small diagram would be negligibly affected by any

incomplete precipitation of casein at the isoelectricpoint.

Ca8ein 8ialic acid. The sialic acid in the caseinpreparations fell from about 1 0% for samples frommilk taken in the first few days of lactation to about0-5% after 1 month. The decline continued moreslowly to the end of lactation (Fig. 5).A positive correlation existed between the casein

sialic acid and the sialic acid in the casein-freesupernatants (Fig. 6). Since the electrophoreticevidence suggested that mature casein is notpresent until the tenth day of lactation, regressionequations have been calculated separately forperiods before and after this time, in addition to theequation covering all samples. The equation formature samples gave a minimal intercept value of0-34% for the sialic acid content of casein (Table 2).An isolated value of 0-10% for a casein sample

prepared from milk taken immediately after suck-ling is referred to in the Discussion section.

I I I Action of rennin. The amount of the total sialice___________________ acid in casein which was released as glycopeptide

160 240 320 by the action of rennin increased from less thanparturition (days) 10% during the first few days of lactation to values

Fig. 4. Sialic acid, in combined form, in human skim-milksamples from which casein had been removed by precipi-tation at pH 4-7.

1 2r

1-0 I

C)._

C)0

CaC)S

cS

0

c;

r-)

0-81-

0-6

0

0

0,0- 0 1

0

0004p

1.2

1*0

.0n' 0-8

C-

.6, 046c)

0

0-0

.g 046ea

v

0-2 -

I I I

&~~~~

0 0

0 0

0 0

*- . 0

0 40 80 120 210 230Non-casein sialic acid (mg./100 ml. of skim milk)

0 80 160 240 320 Fig. 6. Relationship between casein sialic acid and the non-Time after parturition (days) casein sialic acid of human skim milks. O, Milks sampled

before the tenth day of lactation; 9, milks sampled on orFig. 5. Sialic acid in human casein preparations isolated after the tenth day of lactation. The best-fit straight regres-from milk samples taken at different stages of lactation. sion line is given for the latter (mature) samples only.

1

.11. I .1 I

F. H. MALPRESS AND F. E. HYTTEN 1964

Table 2. Regres8ion equations relating casein 8ialwc acid, and the rennin-sen8itive and rennin-insensitivefractions of casein sialic acid, to the non-casein sialic acid of casein-free skim-milkIsupernatantsIn the equations shown, x is non-casein sialic acid (mg./100 ml.), y the fractions shown (g./100 g. casein), t the

time from parturition (days), and n the number of samples.

yCasein sialic acidCasein sialic acidCasein sialic acidRennin-sensitive fractionRennin-sensitive fractionRennin-insensitive fractionRennin-insensitive fraction

t n Regression equation3 10 36 y =034+00042x< 10 13 y = 058 +00020xAll periods 49 y = 0 37 + 0 0035x3 10 32 y = 015 +00024x< 10 13 y = a*-O0-0007x3 10 32 y = 017+0-0021x< 10 13 y = a*+ 0.0028x

* Calculation of 'a' in these cases leads to unrealistic values

0

*- q

.Q

420

44 20

0 80 160 240 320Time after parturition (days)

Fig. 7. Action of rennin on human casein preparations iso-lated from milk samples taken at different stages of lacta-tion. The mean value ± S.D. of the rennin-sensitive fractionfor 32 samples taken on or after the tenth day of lactationis 48-8±8-8%.

mostly falling between 40 and 60 % when lactationhad been established; the mean value + S.D. for 32samples from milk taken on or after the tenth daywas 48-8 ± 8-8 % (Fig. 7).

In view of the correlation between the sialic acidcontent of casein samples and the sialic acid in thesupernatants from which the casein had beenremoved (Table 2 and Fig. 6), it was decided toanalyse this effect further by considering the rennin-sensitive and rennin-insensitive fractions of thecasein sialic acid separately.

Regression equations for these fractions are givenin Table 2; they show that, for caseins preparedfrom milk sampled on or after the tenth day oflactation, the positive correlation obtained forwhole casein sialic acid is a property shared by boththe rennin-sensitive and rennin-insensitive frac-tions. By contrast, for immature caseins thepositive correlation found with whole casein sialicacid is shown to be the resultant of entirely dif-

04

02

0.

005~ ~~

d~.-

I~ ~~ ~~ ~~~~~ I;I,,I 0

0 40 80 120 210 230

Non-casein sialic acid (mg./100 ml. of skim milk)

Fig. 8. Relationship between the rennin-sensitive fraction ofcasein sialic acid and the non-casein sialic acid of humanskim milks. 0, Milks sampled before the tenth day oflactation; 0, milks sampled on or after the tenth day oflactation. The best-fit straight regression line is given for thelatter (mature) samples only.

ferent relationships in the two fractions: a negativecorrelation with the non-casein sialic acid beingobserved for the rennin-sensitive part, and a positivecorrelation for the rennin-insensitive part (Figs. 8and 9).

DISCUSSION

The results presented above were subject to a

number of uncontrolled factors. Milk samples weretaken from mothers representing all strata of an

urban society; this could imply different standardsboth of maternal nutrition and of the care withwhich lactation was established and maintained.No attempt was made to assess differences of thiskind. Further, the pH of milk samples whenreceived in the laboratory varied from 6-5 to 7 5,but no evidence was obtained to suggest that thisinitial value was influencing the propertiesexamined in any way.

More serious objections to comparison of thecasein preparations arise, however. First, thevariation in casein yields might lead to the sup-

position that some preparations were incomplete

134

r+0-72+057+0-82+0-60-0-38+057+0-59

0

by variations in the relative activities of certainenzymes of the mammary alveolar cell responsible

I 1-0 _ for its synthesis.0

_O ~~~~~~~~~~oThe latter view arises from the electrophoretico 1evidence. The basic pattern of six bands given by

- 0-8 00 casein on starch gel at pH 8-6, which represents01 opart or all of the rennin-insensitive fraction of the

ce 0 0o protein, has been shown to vary with respect to the

20 p~~~~~~~~0-6 o relative intensities of the bands. Three types ofc' 0-6 -

0 0 pattern, A, B and C, have been recorded; they areeasily distinguishable and suitably classify all thecaseins found in the present study, except two in> 04 ° which the band pattern was completely absent.

. ..*s."az Within each type minor variations occurred whichC * . were not thought large enough to warrant further

0-2 * type differentiation. An important aspect of the**. differentiation was that, whenever casein sampleswere prepared from one individual at more than

I1 I IF~l..LWJ one stage of the lactation, the characteristic band0 40 80 120 210 230 pattern for the individual remained the same. This

was ofspecial interestinthosemothersshowingtheless frequently occurring patterns (type B and type

Fig. 9. Relationship between the rennin-insensitive frac- C), one of whom for example gave the type Btion of casein sialic acid and the non-casein sialic acid of pattern with caseins obtained on four occasionshuman skim milks. 0, Milks sampled before the tenth day from the fifteenth to the sixty-third day of herof lactation; 0, milks sampled on or after the tenth day of lactation.lactation. The best-fit straight regression line is given for The existence of different band patterns inthe latter (mature) samples only. mature caseins seems to be best explained on the

supposition that the formation, or possibly theand possibly uncharacteristic. This objection has interconversion, of these casein subfractions is abeen met by the rejection of all samples giving a function of specific enzymes in the alveolar cell, andcasein yield of less than 0.1 % [the mean value for that for different individuals these enzymes havecasein content of human milks quoted by Macy & different relative activities. The existence of minorKelly (1961) is 0x37 %], and by the evidence that, variations from the main types, and also, as awhenever a further small casein preparation was special case, of the 'transitional' types noted in theobtained from a 'casein-free supernatant', the Results section,might then be reasonably expected.properties of the second yield differed very little One particular interpretation may be cited: thefrom the main preparation. Secondly, in most cases near resemblance of the type B pattern to theit was not known at what period after the previous pattern given by certain immature caseins derivedsuckling samples were taken. In one case where from milk taken in the first week of lactation,comparable and extreme data were available from especially the enhanced prominence of band 1 andone mother, the sialic acid content of the casein the weakness or absence of bands 5 and 6, suggestsprepared from milk taken immediately after that this type is the result of a deficiency in at leastsuckling, 0-10 %, was much lower than that found one enzyme normally involved in producing the fullin the casein from milk taken before suckling heterogeneity of the casein molecule, and that thisbegan, 0 35 %. The value, in the context of the results in a permanently immature type casein inpresent work, was uniquely low and suggests that such individuals.the time during which milk rests in the gland may The influence of the rnilk on human caseins isbe an important factor in determining the sialic demonstrated by the direct relationship whichacid content of the isolated protein. Since, how- exists between the sialic acid in the isolated proteinever, only three of the mature caseins examined in and the non-casein sialic acid in the milk. Boththe present work were prepared from milk samples decrease throughout lactation and from the presentwith a volume less than 20ml., it is thought that experiments the sialic acid values for maturethis effect could not have been seriously damaging. samples of casein, obtained by using the derivedThe results suggest that human casein is a protein regression equation (Table 2 and Fig. 6), might be

having variable properties, and that it can be expected to fall to between about 0'70 and 0 35 %.influenced both by the composition of the milk For such caseins, isolated on or after the tenthfrom which it is derived and, more fundamentally, day of lactation, when from the electrophoretic

Vol. 91 PROPERTIES OF HUMAN CASEIN 135

136 F. H. MALPRESS AND F. E. HYTTEN 1964

evidence and from the development of maximumsensitivity to rennin (Fig. 7) there is a reasonablepresumption that the protein is 'mature', the cor-relation with the non-casein sialic acid is a propertyshared in an equal degree by both the rennin-sensitive and the rennin-insensitive fractions of theprotein (Figs. 8 and 9). As a consequence of this theproportion of sialic acid found to be sensitive to theaction of rennin in such specimens is independentof the time of sampling, and in the present studywas found to be about half of the total sialic acidpresent, though the divergence from the meanvalue could be very large in individual cases.In casein prepared from cow's milk, Marier,

Tessier & Rose (1963) have shown that all the sialicacid is present in the K-casein (micelle-stabilizingfraction; Waugh & Hippel, 1956; Waugh, 1958),which has a sialic acid content of about 2-3 %. Thesialic acid content of whole cow casein ranged from0-26 to 0 59% in the work referred to (Marier et al.1963), and this variation is ascribed entirely todifferences in the K-casein content of the wholeprotein, from 11 to 26 %.No fraction, having the high sialic acid content or

showing the specific physicochemical properties ofK-casein, has yet been obtained from maturehuman casein, although Alais & Jolles (1962b) havenoted that the release of a glycopeptide relativelyrich in sialic acid by rennin action is evidence infavour of the existence of a fraction in humancasein similar to K-casein. It is possible thereforethat the effects outlined in the present paper willultimately be found to refer more specifically tovariations in the proportion of one fraction havinga constant high sialic acid content, and formingonly a pa"t of the human casein aggregate, ratherthan to a progressive change in the casein aggregateas a whole. The unchanging character ofthe electro-phoretic band pattern throughout a lactation givessupport to the view that for a given individualmuch of the casein is well defined and invariable.

SU:MMARY

1. Specimens of human casein representing allstages oflactation have been prepared by a standardprocedure from 80 milk samples given by 47mothers.

2. Electrophoresis ofhuman casein on starch gelat pH 8.6 gives a six-band pattern. Three caseintypes (A, B and C) have been characterized, whichdiffer in the relative strengths of the bands withinthis main pattern.

3. The results suggest that for any one individualthe type does not change during the course of alactation.

4. The sialic acid content of casein preparationsfell during lactation from about 1.0 to 0 2 %; itshowed a positive correlation with the amount ofsialic acid present in the skim milk after casein hadbeen removed.

5. About half the casein sialic acid was sensitiveto the action of rennin.

6. It is concluded that the nature and propertiesof human casein may be influenced by the relativeactivity of certain enzymes in the mammaryalveolar cell, and by the composition of the milkfrom which it is derived.

The authors acknowledge the help of Dr H. A. Warnockand his staff at the Belfast Maternity and Child HealthDivision in obtaining milk samples. They also thankProfessor D. C. Harrison for his helpful interest in the work,and Mr J. McGrath for skilled technical assistance.

REFERENCES

Alais, C. & Joll6s, P. (1962a). C.R. Acad. Sci., Paris, 255,2309.

Alais, C. & Jolles, P. (1962b). Nature, Lond., 196, 1098.Johansson, B. & Svennerholm, L. (1956). Acta physiol.

scand. 37, 224.Macy, I. G. (1949). Amer. J. Dis. Child. 78, 589.Macy, I. G. & Kelly, H. J. (1961). In Milk: the Mammary

Gland and its Secretion, vol. 2, p. 265. Ed. by Kon, S. K.& Cowie, A. T. New York and London: Academic PressInc.

Maeno, M. & Kiyosawa, I. (1962). Biochem. J. 83, 271.Malpress, F. H. (1961). Biochem. J. 80, 19P.Malpress, F. H. (1962). Biochem. J. 85, 33P.Marier, J. R., Tessier, H. & Rose, D. (1963). J. Dairy Sci.

46, 373.Poulik, M. D. (1957). Nature, Lond., 180, 1477.Wake, R. G. & Baldwin, R. L. (1961). Biochim. biophys.

Acta, 47, 225.Warren, L. (1959). J. biol. Chem. 234, 1971.Waugh, D. F. (1958). Disc. Faraday Soc. 25, 186.Waugh, D. F. & Hippel, P. H. von (1956). J. Amer. chem.

Soc. 78, 4576.