Determination of lysozyme activity at low levels with emphasis on the milk enzyme

ANALYTICAL

BIOCHEMISTRY

157,367-374

(1986)

Determination of Lysozyme Activity at Low Levels with Emphasis on the Milk Enzyme H. A. MCKENZIE Protein

Chemistry Australian

Group, John Curtin National University,

AND F. H. WHITE,

JR.’

School of Medical Research, Institute of Advanced G.P.O. Box 334, Canberra, A.C.T. 2601, Australia

Studies,

Received February 25, 1986 A method is described for the determination of lysozyme (muramidase) activity, whereby sensitivity is maximized by incubation of the reaction mixture (sample, buffer, and substrate (Micrococcus luteus)) over an extended period. This approach is made feasible by exploiting our observation that the lytic reaction follows simple kinetic order during this time (e.g., 700 min for bovine lysozyme and 960 min for the eggwhite enzyme at low concentrations). After this period, the reaction rates diminish, indicating biphasic behavior, and eventually become negligible. The kinetic order may vary with both the type of lysozyme and the buffer system used. The limit of detection for bovine milk lysozyme is 100 &ml reaction mixture, equivalent to 6 r&ml milk, for a 50-~1 sample (with reference to hen eggwhite lysozyme). With these limits, the method has proven valuable in our comparative studies, particularly for low levels of activity in bovine milk, but also in secretions and tissue extracts from various other eutherian, metatherian, and prototherian mammals. The method may also be applied to investigation of structure and function in modified forms of the enzyme. 0 1986 Academic F’ress, Inc. KEY WORDS: lysozyme; kinetic order; Micrococcus luteus; hen eggwhite; bovine milk; picogram.

Lysozymes (muramidases) are widely distributed throughout nature. Despite species differences, lysozymes of bird eggs and mammalian secretions fit into two classes, c and g. Within a given class they exhibit sequence homology. Their characteristic enzymic activity involves cleavage of the glycosidic bond in the alternating P-linked (l-4) copolymer of Nacetyl-D-glucosamine and N-acetyl-D-muramic acid. It is believed that this muramidase activity is the basis of lytic activity against bacteria where such polysaccharide units are present in the cell wall. Most methods for the determination of lysozyme activity, in fact, involve measurement of the initial turbidity clearing rate in bacterial cell suspension containing the lysozyme sample (e.g., Shugar (1)). Although such methods are rapid (5-10 min), their limit of detection

is usually about 0.1 pg lysozyme/ml of reaction mixture. For our comparative studies on lysozymes in the secretions and tissues of eutherian, metatherian, and prototherian mammals, we have found need for a more sensitive method. The concentration of lysozyme in milk and some other secretions of bovine origin is so low that its occurrence from such sources has been questioned (2,3). Chandan et al. (4), nevertheless, have isolated a bacteriolytic enzyme from bovine milk. They gained sensitivity by adding large volumes of sample relative to substrate (e.g., 1 ml to 1.5 ml of cell suspension) and allowing the reaction to proceed for 20 min (5). However, with the requirement of such large samples, their method does not permit determination of activity directly on milk because of its opacity and other potential complications. Thus, they had recourse to the use of “whey” samples obtained by acid pre-

’ To whom correspondence should be addressed. 367

0003-2697186 $3.00 Copyright 0 1986 by Academic Pms, Inc. All rights of reproduction in any form reserved.

368

MC

KENZIE

cipitation of casein and its removal. This procedure does not necessarily give a true picture of the lysozyme level in the milk itself. Furthermore, in our studies of milk from prototherian mammals, economy of sample is of the utmost importance. Selsted and Martinez (6) gained sensitivity for their turbidimetric method by prolonging the incubation period to 18 h. Their procedure has not proven satisfactory for our purposes, although we also use extended incubation periods (albeit somewhat shorter than theirs). However, we stress the kinetic behavior of cell lysis. It is shown that, while the reaction is biphasic, simple kinetics are obeyed over a sufficiently long period that this property can be exploited in the determination. Moreover, the kinetic order must be taken into consideration for quantitative treatment of the results. The method so derived has proven to be especially suitable for milk samples from a variety of mammals, and also applicable to other secretions and tissue extracts. For a preliminary report see (7). MATERIALS

AND

METHODS

Materials. Killed cells of Micrococcus luteus (Micrococcus lysodeikticus) were obtained from the Sigma Chemical Company, St. Louis, MO. (Lot Nos. 45C-0167 and 62F-0250), as was domestic hen eggwhite lysozyme (Lot No. 69C-8055). “Desicote,” for siliconization of glassware, was obtained from Beckman Instrument Corporation, Fullerton, California. Polycarbonate tubes ( 16 X 102 mm, Nalgene, Rochester, N.Y.) for containing the reaction mixtures were fitted with Teflon caps of our own manufacture. Micropore filtration. Buffers and samples were filtered through a Millex-GS disposable filter unit (pore diameter: 0.22 pm, Millipore Corp., Bedford, Mass.). Absorbance measurements. A Zeiss Model PM-II spectrophotometer, equipped with a double monochromator MM 12 (Carl Zeiss, Oberkochen, F.R.G.), or a Cary Model 14M Recording spectrophotometer (Varian Instruments, Palo Alto, Calif.) was used.

AND

WHITE

Special precautions. The procedures outlined below were designed primarily to maintain sterility and to prevent occasional spontaneous (nonenzymatic) clearing of the cell suspension on prolonged incubation (see Results, Initial experiments). (i) Pipetting by mouth was avoided. The desired volumes of sample or standard lysozyme solution were added to the reaction tubes with a “Transferpettor” micropipette (Rudolf Brand and Co., Wertheim, Main, F.R.G.). The cell suspension was added with an “Optifix Universal Dispenser” (Walter Graf and Co., Wertheim, Main, F.R.G.). (ii) Rigorous trace analysis procedures were used for the cleaning and sterilization of all glassware and polycarbonate tubes (details available from authors). (iii) All glassware used for dilution of enzyme solutions was first siliconized to prevent loss of enzyme by adsorption on glass surfaces. (iv) All buffer solutions were sterilized by filtration with Millex filters, and washed previously with 10 ml buffer to remove any residual soluble material remaining from manufacture of the filter. (v) Absorbance at 450 nm (A& at zero time and for any given time during the reaction was determined only once for a given reaction tube, which was then discarded. Thus, the loss of sterility that would have resulted from multiple readings on the same mixture was avoided. Preparation of buffer solutions and substrate. For studies of pH and buffer effects, a series of buffers were prepared with KH2P04K2HP04, imidazole-HCl-NaCl, or Tris-HClNaCl, to give the required ionic strengths (I) and pH values. For determination of lysozyme activity in bovine milk, 0.1 M imidazole-O.0 16 M HCl-0.084 M NaCl (I = 0.1, pH 7.5, at 25”C), in the presence of 0.003 M NazEDTA, was the preferred buffer. The substrate was prepared by mixing a sample of the killed cell preparation (7.5 mg) with buffer (50 ml), sterilized as above. All equipment (weighing beaker, stainless steel spatula, and glass stoppered bottle for containing the suspension) was sterilized before use. The cells were suspended by gentle swirl-

LYSOZYME

ACTIVITY

ing, and the mixture was incubated 2 h at 37 “C prior to use. Preparation of stock eggrvhite lysozyme solutions. A stock solution of domestic hen eggwhite lysozyme (1 mg ml-‘) was prepared in buffer of the same kind chosen for the reaction, passed through a Millex filter, and stored at 2°C. The lysozyme concentration was found by measurement of the absorption at 280 nm, taking Ai g1 = 27.3 (8). Milk samples and milk lysozyme @actionation. Milk samples obtained locally were cooled in ice immediately on completion of milking, and lysozyme determination commenced within 1 h. Samples from more distant locations were frozen immediately after milking and transported to Canberra by air. These precautions were necessary, since loss of lysozyme activity may occur over several days in the milk of some species, e.g., bovine milk, even at 2°C. However, the enzyme was relatively stable, once fractionation (7) of the milk was commenced, or on dilution concomitant with activity determination, indicating destabilization factors in the milk, rather than inherent instability of the enzyme as the cause. Milk samples of all species were skimmed by centrifugation at 2000g for 20 min at 2°C in a Sorvall RC-5 Centrifuge, with either a GSA or SS-34 rotor. Bovine milk lysozyme was purified 500- to 7000-fold by the method of McKenzie et al. (for a brief account, see (7) and to be published in detail). Equine and echidna milk lysozymes were purified as described respectively by Bell et al. (9) and Griffiths et al. (10). Stomach mucosal samples. Stomach mucosal extracts were prepared in dilute ammonium acetate buffer as described by Dobson et al. (11). Lysozyme activity determination: Absorbance corrections. In view of the low lysozyme activities as well as the extended incubation periods involved, it is critical to monitor any change in absorbance of the M. luteus suspension with time in the absence of enzyme. This change was usually small (O-0.0 15 over a 16-

369

DETERMINATION

h period), but allowance was made for it by including an appropriate control tube, which contained an equivalent volume of buffer in place of sample. The “corrected” absorbance (A&,) at any time t in the presence of enzyme is then given by A&, = A&, - AA&,

PI

where A& is the uncorrected absorbance, and AAijo = A& - A2&,

PI

where the latter two terms are respectively the absorbances of control at zero time and time t. The change in absorbance throughout the reaction is then given by AA 450

= A&o - A&o,

131

where AiS is the absorbance at 450 nm in the presence of enzyme at zero time. The value of A$,0 is equal to A2,00when addition of sample to the cell suspension causes no increase in initial absorbance. However, there is usually a small increase when milk (or other opaque samples) are added (see Results). Thus, Ait cannot be assumed invariably to be equal to AiSo, and an experimental determination of each is required. In all determinations the reaction mixture contained 3.0 ml cell suspension in the required buffer, 0.01-0.05 ~1 sample or “reference” enzyme solution, and sufficient buffer solution (0.04-0.00 ~1, respectively) to maintain the total volume of the reaction mixture constant. The control contained no enzyme solution, but only the added buffer. The stock solution of hen eggwhite lysozyme was diluted with required buffer (usually 104- to 105-fold) before addition to the cell suspension. Kinetic studies. In kinetic studies of a given enzyme at a particular concentration all reaction tubes were identical, the variable being time. Each reaction tube was paired with a control tube. Individual experiments usually involved six pairs of tubes, each pair being removed from the constant temperature bath at the desired time for measurement of absorbances at 450 nm, from which AA450 was

370

MC

KENZIE

found with Eq. [3]. The resulting values were then treated kinetically as in Results. RESULTS

Initial experiments. In some of our early experiments, based essentially on the procedure of Selsted and Martinez (6), a spontaneous and variable decrease in absorbance of the cell suspension sometimes occurred within 2 h after the start of incubation. The effect was one of “clearing” and not aggregation. It was greater at 37°C than at 20°C and occurred irrespective of the presence or absence of lysozyme. This problem was resolved by taking the special precautions, especially with respect to sterility, as given under Materials and Methods. They did not generally include micropore filtration of milk samples, since equivalent results were obtained with and without filtration, indicating that the milk sample was not normally a source of contamination contributing to the clearing effect. Eficts of pH and bu&r. Phosphate buffers have been widely used in determination of lysozyme activity. In development of the present procedures, such buffers were first used, since they had proven satisfactory when higher lysozyme concentrations were being studied. However, with the prolonged incubation method, there was occasional difficulty, especially at lysozyme levels below 1 ng ml-’ in the reaction mixture, where loss of sensitivity was observed. Absorbance values for controls were relatively high, particularly when NazEDTA was included in the reaction mixtures for milk samples (see below), and reproducibility between different cell lots was unsatisfactory. These problems were overcome when an imidazole-HCl buffer system was adopted in place of the phosphate buffer. Essentially the same improvement was achieved in Tris-HCl buffer. The pH value for optimum activity of domestic hen eggwhite lysozyme was found to be between 6.8 and 7.4 (where I = 0.1 - 0.13 mol liter-‘) in agreement with other workers. The optimum value for bovine lysozyme

AND

WHITE

found in the present work was 7.5, slightly less than the value of 7.9, given by Parry et al. (5). A pH value of 7.5 was chosen for investigation of kinetic order and routine determination of activities for hen eggwhite and bovine lysozymes. E&t of temperature. Incubations were conducted at 20 and 37”C, and values of AAds after incubation for 17 h were 0.095 and 0.302, respectively. Thus, a temperature of 37°C offered greater sensitivity and was chosen for routine determinations. Efict of mixing. Various means were examined for continuous mixing of the reaction mixture during incubation. They did not, however, have any apparent effect on either the rate or extent of reaction. Thus, mixing was eliminated, except for manual mixing immediately prior to determination of absorbance. Kinetics of cell lysis. Under the conditions described here, it was demonstrated that &e was linearly related to M. luteus concentration. If a is the concentration of cells in the reaction mixture at zero time, and A&, is the corresponding absorbance at 450 nm, and if x is the concentration of cells lysed at time t, then the concentration of cells remaining at t is a-x, with absorbance of A&. The term x is proportional to A&0 (defined by Eq. [3]). If the kinetics of the reaction are zero, first, or second order, then plots of x, ln(a(a-x)-l), or a(a-x)-l vs twill be linear, with slopes equal to kb, k\ , or k;a, and with intercepts of zero, zero, or unity, respectively. The values kb, k’,, and k)2are the zero, first, and second order velocity constants, respectively (e.g., see (12)). In terms of the present experimental parameters, the dependent variables are proportional to A-4450, M&d&d, andA%o/&o, rewctively. The slopes of these parameters vs t are b, k’, and k2AiSo,where kc, and k2 are empirical zero and second order constants. (There is no need to introduce an empirical fnst order constant, owing to the form of the first order equation.) The corresponding half times (tI,Z) are Ao450/2ko, 0.693/k\, and 1/A&ok2. Kinetic plots are shown in Fig. 1 for

LYSOZYME

-

ACTIVITY

DETERMINATION

371

A

Minutes

FIG. 1. Kinetics of esgwhite (A) and bovine milk (B) lysozymes, 0, AA 450;n , ln(u/(a-x)); A, a/@-x). The latter two terms are equivalent to ln(Ai,/A450) and A&/AdH) respectively, as described in the text. The bovine enzyme had been purified approximately 7000-fold. (That for (C) had been purified approximately 9% fold.) The buffer was imidazole-HCl-NaCl-EDTA, pH 7.5 (see Methods). For (C), the kinetics of bovine lysozyme were studied in the presence of skim milk, the latter having been found enzymatically inactive. The enzyme preparation contained 17 pg lysozyme ml-‘, expressed as hen eggwhite equivalent, estimated from Fig. 2. The preparation was diluted 20-fold with milk, and the aliquot size was 50 ~1. Other experimental details were as given for kinetic experiments in Methods.

eggwhite (A) and bovine milk (B, C) lysozymes. The second order plot is linear over the time periods shown for each sample; thus the reactions shown are second order over this period. Typically, after the maximal times for which these reactions follow simple kinetic order, the reaction rates diminish, indicating biphasic behavior, with negligible rates within 1 h. The tl,2 of the reaction for eggwhite lysozyme in Fig. lA, where the enzyme concentration in the reaction mixture is 2.4 ng ml-‘, is 1282 min. Thus, the maximum time of 960 mm corresponds to 0.75 tli2. When the concentration of enzyme is increased to 3 1.5 ng ml-‘, this time becomes 244 min, or 1.8 t,,2, the maximal value we have found for this enzyme. Bovine lysozyme at a concentration 3.8 ng ml-’ followed second order for 558 min, or 0.6 t1,2, as shown in Fig. 1B. From other experiments (not shown) the longest time yet found for this enzyme has been 700 min. When the concentration was increased to 19 ng ml-‘, the duration of second order became 2 18 min, or 1.3 t1,2 (maximal). These results demonstrate the necessity of

determining the length of time for each new lysozyme to be studied during which a specific order is observed for the concentration range being considered. It is clear that the ratio of the time during which simple kinetic order is observed to the half time is greater at higher concentrations of lysozyme. Limit of detection.It can be seen from Fig. 2 that determination of eggwhite lysozyme is possible to levels at least as low as 0.2 ng ml-’ in the reaction mixture. The limit of detection has been found from the standard deviation (SD) of 18 determinations, involving incubation of tubes containing reaction mixture in the absence of enzyme for 16 h at 37°C. The SD was kO.006. If we adopt the recognized convention for limit of detection ( 13) as 3. SD, then AL&,, = 0.018 for a minimally significant lysozyme content. Taking A!& = 0.800, and t = 700 min, this A&, value corresponds to a k2of 4 X 10e5 min-r. (A conventional second order velocity constant, k;. has units of (concentration)-’ t-‘, e.g., (g/ liter))’ min-I. Since the present empirical constant, k2, incorporates an arbitrary constant relating absorbance of M. luteus to concentration, it has the unit of t-l.) The above

312

MC KENZIE

AND WHITE

EDTA, usually gives an absorbance increase of about 0.05, although the range of variation is from 0 to 0.1. Allowance is made for this change with the A!& value (see Methods). With the sample aliquot again assumed to be 50 ~1, the limit of detection for lysozyme in milk is 6 ng ml-‘, given the experimental conditions as above under Limit of detection, with the minimal lysozyme concentration in the reaction mixture of 100 pg ml-‘. The validity of the method for determinaIO 15 20 5 tion of lysozyme in milk samples was checked Nanogram ml-’ in two ways: FlG. 2. Change in the second order reaction rate constant First, a kinetic study was made on a mixture (k2) with concentration of hen eggwhite lysozyme in the of partially purified bovine milk lysozyme with reaction mixture. Buffer as in Fig. 1. Other experimental details given in text. Inset shows the plot on a larger scale a skim milk sample that was devoid of lysofor the region 0- 1 ng ml-‘. zyme activity. The kinetics were second order over the period shown in Fig. 1C. In an additional experiment (not shown) in which the value for k2 is equivalent to a lysozyme con- same lysozyme preparation was diluted into centration of 0.1 ng ml-’ in the reaction mixwater instead of skim milk, identical results were obtained. The kinetics are, of course, in ture (Fig. 2). Determination of lysozymeactivity in bovine accordance with those of Fig. 1B. milk. For the study of lysozyme in milk samSecond, the validity of the procedure was ples, there was an additional problem accruing checked by the method of standard additions from the opacity of the casein micelle system, (16). When either aliquots of eggwhite lysoeven with the small sample volumes used in zyme or of partially purified bovine lysozyme the present work. Thus, a transient rise in Ads0 (7000-fold) are added to bovine milk, which was found when the milk sample was mixed is then subjected to lysozyme determination, with cell suspension. This rise diminished over the resulting activity agrees with the sum of a period of 30 min to a constant level, usually the added activity and that of the naturally occurring lysozyme in the milk. somewhat higher than the control absorbance. Thus, there is no indication that milk conThis problem, however, was overcome with the inclusion of EDTA, in the presence of stituents interfere with the expression of cell which the increase in absorbance was less pro- lytic activity. nounced, with a steady value reached immeWe have used our method in a survey of milk samples from 20 cows, at different stages diately. of lactation, with lysozyme concentrations These phenomena result from the micelle system of milk. As has been shown elsewhere from zero to 1120 ng ml-‘. Although the method as described is well suited to this pur(14,15) the constituents are not in rapid equipose, the sensitivity could be increased even librium, and when constraints (e.g., dilution) are applied, the system is slow to reach a final further by adding larger sample volumes, in state. The integrity of the system depends on which case it would be necessary to increase the EDTA concentration proportionally. the presence of calcium. Removal of calcium by EDTA circumvents the above difficulty by Other lysozymes.Lysozymes m equine milk immediately dispersing the system and does and bovine stomach mucosal extract exhibited second order kinetics in either imidazole or not interfere with the reaction. A milk aliquot of 50 ~1, in the presence of Tris buffer, as did that from hen eggwhite, in

LYSOZYME

ACTIVITY

contrast to first order when phosphate buffer was used. Isolated lysozymes from human and echidna milk, and eggwhite of the black swan (with both c and g types) showed no such differences, giving second order regardless of buffer. Lysozyme in platypus milk and stomach mucosal tissue of the tammar wallaby both exhibited zero order in phosphate and imidazole buffers. The reaction order for all of the above lysozymes endured from 1 to 1.8 t1,2. Thus, discrete kinetic behavior may prove to be the rule among various lysozymes, although dependent upon reaction conditions. Second order predominates in imidazole or Tris buffer. DISCUSSION

Development of the method described here arose from the necessity of determining lysozyme in trace concentrations, as it appears in a variety of mammalian tissues and secretions. A procedure involving prolonged incubation of the reaction mixture appeared to offer the greatest hope of achieving the desired sensitivity. Since a variety of lysozymes have been studied in our work, however, it was necessary to give careful attention to possible differences in kinetic behavior. The critical observation that makes our approach feasible is that simple kinetic order obtains, for low concentrations of lysozyme over prolonged periods, and this observation is exploited in the quantitative treatment of the results. The present work appears to be the first in which kinetics of cell lysis have been followed over such an extended period. Other studies have been conducted at comparatively high concentrations of lysozyme, for which half times of a few minutes occur. The kinetics in these studies have been considered as zero ( 1,17), first (18,19), and second order (20-22). The results reported here are consistent with those of Howard and Glazer (2 1) and of Prasad and Litwack (22) since both of these groups found second order, albeit at relatively high lysozyme concentrations, and most of the lysozymes tested in our system have exhibited second order kinetics.

DETERMINATION

373

The former group (21) suggested that two points of attack on the cell wall may be responsible for the biphasic nature of the reaction observed by them and by Smith et al. (20). Prasad and Litwack (22), on the other hand, speculated that two points of attack could produce the second order kinetics observed by them. Such explanations are only speculative, however, and much remains to be learned about so complex a substrate as a cell wall. We concur with Parry et al. (5) that application of cell lysis by lysozyme to the study of lysozyme kinetics “should be viewed with caution.” Accordingly, any kinetic treatment of this reaction must be regarded as empirical, and it is necessary to perform an initial kinetic study for each new lysozyme to be examined, so as to establish not only its kinetic order, but also the length of time through which this order is maintained. Then it becomes practical to employ kinetic treatment of the results. Moreover, it is clear that the kinetic behavior must be taken into account if the results of lysozyme determination by prolonged incubation are to have quantitative significance. An estimate of the lysozyme concentration is then possible by relating the empirical rate constant to a calibration curve, such as that in Fig. 2. It is stressed, however, that the result is an approximation, since it is based on the assumption of equivalence between specific activities of sample and hen eggwhite lysozyme. Determination of the true concentration is only possible after the enzyme has been highly purified. The lowest lysozyme concentration that can be detected is shown here, based on the standard deviation from a series of blank determinations, to be 100 pg ml-‘. The lowest level previously reported seems to be that of Selsted and Martinez (6) where prolonged incubation permits detection of 5 pg ml-‘. They were, however, assuming A&, > 0.002 to be significant. It is our opinion, based on statistical grounds, that an absorbance change below 0.0 18 should not be considered as indicating significant lysozyme activity.

374

MC KENZIE

The lowest value for lysozyme concentration previously determined in milk appears to be that given by Parry et al. (5) as 100 ng ml-‘. Our limit of detection in the reaction mixture (see above) corresponds to 6 ng ml-’ in milk, and this value represents a substantial improvement in sensitivity. It is to be especially noted that, although this level was adequate for our studies on milk, even greater sensitivity could be gained by increasing the sample aliquot above 50 ~1, for 3 ml of substrate, with a proportionally higher EDTA concentration. The present methodology may be of use for measurement of lysozyme activity from many sources, where the enzyme is suspected of being present in minute concentrations. Another potential use of our method lies in detecting and studying possible traces of lysozyme activity remaining after structural modification (e.g., reduction of disulfide bonds CWW. ACKNOWLEDGMENTS Grateful acknowledgment is due R. J. Pearce, C.S.I.R.O. Division of Food Research, Highett, Victoria, for his help and advice, and to M. Brown, Milking Research Centre, Gilbert Chandler Institute, V.C.A.H., Werribee, Victoria, for milk samples and their collection. We thank K. Ma&in, St. Vincent’s School of Medical Research, Fitzroy, Victoria, for samples of lysozymes c and g from eggwhite of the black swan. Suzanne Kelly is especially thanked for skilled technical assistance One of us (F.H.W.) is grateful to the Australian National University for the award of a Visiting Fellowship.

REFERENCES 1. Shugar, D. (1952) Bioc~im. B&?&x. Acta 8, 302309. 2. Pahud, J. -J., and Widmer, F. (1982) Biochem. J. 201, 661-664.

AND WHITE 3. Banyard, M. R. C., and McKenzie, H. A. ( 1982) Mol. Cell. Biochem. 41, 115- 124. 4. Chandan, R. C., Parry, R. M., and Shabani, K. M. (1965) Biochim. Biophys. Acta 110,389-398. 5. Parry, R. M., Chandan, R. C., and Shahani, K. M. (1965) Proc. Sot. Exp. Biol. Med. 119,384-386. 6. Selsted, M. E., and Martinez, R. J. (1980) Anal. Biochem. 109,67-IO. 7. McKenzie, H. A., Pearce, R. J., and White, F. H., Jr. (1985) Proc. Aust. Biochem. Sot. 17,32. 8. Glazer, A. M. (1959) Aust. J. Chem. 12,304-307. 9. Bell, K., McKenzie, H. A., Muller, V., Rogers, C., and Shaw, D. C. ( 198 1) Camp. Biochem. Physiol. 68B, 225-236.

10. Griffiths, M. E., McKenzie, H. A., Shaw, D. C., and Tcaban, C. G. (1985) Proc. Aust. Biochem. Sot. 17,25.

11. Dobson, D. E., Prager, E. M., and Wilson, A. C. ( 1984) J. Biol. Chem. 259, 11607-l 1616. 12. Roberts, D. V. (1977) Enzyme Kinetics, Cambridge University Press, New York. 13. Commission on Spectrochemical and other Optical Procedures for Analysis (I.U.B.) (1976) Pure Appl. Chem. 45,99-103. 14. McKenzie, H. A. (1967) Adv. Protein Chem. 22, 55234. 15. Waugh, D. F. (1971) in Milk Proteins (McKenzie, H. A., ed.), Vol. 2, pp. 3-85, Academic Press, New York. 16. Morrison, G. H. (1975) Pure Appl. Chem. 41, 395403. 17. Smolelis, A. N., and Hartsell, S. E. ( 1949) J. Bucteriol. 58,131-136. 18. Dickman, S. R., and Proctor, C. M. (1952) Arch. Biochem. Biophys. 40, 364-312. 19. Kerby, G. P., and Eadie, G. S. ( 1953) Proc. Sot. Exp. Biol.Med.83, 111-113. 20. Smith, E. L., Kimmel, J. R., Brown, D. M., and Thompson, E. 0. P. (1955) J. Biol. Chem. 215, 76-89. 2 1. Howard, J. B., and Glazer, A. N. (1969) J. Biol. Chem. 244, 1399-1409. 22. Prasad, A. L. N., and Litwack, G. (1963) Anal. Biochem. 6,328-334. 23. White, F. H., Jr. (1982) Biochemistry 21,967-977. 24. White, F. H., Jr. (1984) Int. J. Pept. Protein Res. 24, 453-461.