Immunologic comparison of pancreatic ribonucleases

Immunochemistry, 1976, Vol. 13, pp. 6534158. Pergamon Press. Printed in Great Britain

IMMUNOLOGIC COMPARISON OF PANCREATIC RIBONUCLEASES GJALT W. WELLING, GERDA GROEN, JAAP J. BEINTEMA, M A R C H I E N U S E M M E N S and F. PETER SCHRODEW Biochemisch Laboratorium, Rijksuniversiteit, Zernikelaan, de Paddepoel, Groningen; and 1Streeklaboratorium voor de Volksgezondheid, van Ketwich Verschuurlaan 92, Groningen, The Netherlands (First received 9 February 1976; in revised form 25 March 1976)

Abslract--Ouchterlony double immunodiffusion and micro-complement fixation were used in crossreactivity studies with 9 pancreatic ribonucleases differing 3-28~o in amino acid sequence and rabbit antisera to cow, gnu, reindeer and whale ribonuclease. Generally a correlation was observed between the extent of cross-reactivity and amino acid sequence resemblance. The antiserum against whale ribonuclease however, reacted to a larger degree with several antigens than expected from the amino acid sequence difference. Dromedary ribonuclease, which differs at 26~o of the positions, showed the highest cross-reaction. By comparing the antigenic reactivities and the differences in sequence, an attempt was made to localize antigenically relevant regions.

Amino acid sequence studies of homologous proteins may yield valuable information on their molecular evolution and on the relationship between protein structure and function. In the case of pancreatic ribonucleases, sequences are manifold (Welling et al., 1975a) and the evolutionary history of these enzymes was investigated (Welling et al., 1975b). A correlation between the degree of cross-reactivity as measured by various immunologic tests and amino acid sequence resemblance has been observed for animal lysozymes (Prager & Wilson, 1971a, b), cytochromes c (Margoliash et al., 1970) and azurins (Champion et al., 1975). A search for a similar relationship between pancreatic ribonucleases from different species seems of interest. Moreover, the antigenic differences can be related with the known tertiary structure of bovine ribonuclease A (Kartha et al., 1967; Carlisle et al., 1974) and ribonuclease S (Wyckoff et al., 1970), ultimately resulting in knowledge about the location and conformation of antigenic determinants. The interaction of native and oxidized bovine ribonuclease with their antibodies was extensively studied by Brown et al. (1959). Native bovine ribonuclease did not react with antibodies to the oxidized enzyme, indicating that in the native protein only conformational dependent antigenic determinants are present. Changes in the structure by removal of the N-terminal amino acid (Brown, 1960), by using homologous proteins (Brown et al., 1960) or by rupturing the peptide bond between residue 20 and 21 (Singer & Richards, 1959), have been shown to result in less antigenic reactivity with antiserum directed against the native molecule. In this study an immunologic comparison is made for the following 9 pancreatic ribonucleases: from cow (Smyth et al., 1963), goat (Welling et al., 1974a) giraffe (Gaastra et al., 1974), reindeer (Leijenaar-van den Berg & Beintema, 1975), pig (Jackson & Hirs, 1970; Wierenga et al., 1973), dromedary (Welling et al., 1975a), horse (Scheffer & Beintema, 1974), gnu (Groen et al., 1975), and pike whale or lesser rorqual (Emmens et al., 1976). Four 653

different antisera against cow, gnu, reindeer and whale ribonuclease respectively, were utilised. Ouchterlony double immunodiffusion and quantitative micro-complement fixation, which has been shown to be very sensitive to small differences in primary structure (Prager & Wilson, 1971a, b) have been used to measure the antigen-antibody reaction. MATERIALS A N D M E T H O D S

Materials

Bovine ribonuclease was a product from Miles (Maidenhead, U.K.). Bovine ribonuclease A (carbohydrate free) was purchased from Schwarz/Mann (Orangeburg, NY, U.S.A.). Agarose was from BDH (Poole, Dorset, U.K.) and Freund's complete adjuvant was purchased from Difco (Detroit, MI, U.S.A.). All other reagents were of analytical grade. Ribonucleases were isolated by affinity chromatography as described by Wierenga et al. (1973). Antisera

Antisera to cow, gnu, reindeer and whale ribonucleases respectively were produced in rabbits as described by Westendorp Boerma et al. (1974) for hemoglobins with some modifications. One rabbit was used for each immunogen. At days 1and 10, 0.5 ml of an 0.5% ribonuclease solution in 0.9% NaC1 was mixed with an equal volume of Freund's adjuvant and subcutaneously injected in each thigh of a rabbit. At day 22, 0.2 ml of an 0.25% ribonuclease solution was injected subcutaneously, followed after 30min by 0.2 ml intravenously. Increasing amounts (0.5-2.5ml) of the same solution were injected intravenously at days 24, 26, 29, 31, 33, 35 and 38. Bleedings were taken at day 45. After 3 weeks rest, four booster injections (0.2-2 ml) were given at intervals of 2 days. Rabbits were bled one week after the last injection. Since only a limited amount of reindeer ribonuclease was available, this antigen was injected once a week during the whole immunisation period. The concentration of the bovine ribonuclease solution used, was twice that of the other ribonucleases. In total, 80rag of bovine ribonuclease, 40 mg of gnu and whale ribonuclease and 12.5 mg of reindeer ribonuclease were used for immunisation. Most of the work described here was performed with second bleeding sera (after three months of immunisation).

GJALT W. WELLING et al.

654

Immunologic methods Double immunodiffusions were performed in 2% agarose gel with antigens at a concentration of 0.5 mg/ml in 0.9% NaCI solution. Antigen and antiserum wells were 4mm and 6mm dia, respectively. The centres of antiserum and antigen wells were 9 mm apart. About 12.5/A of undiluted antiserum was used. The antiserum directed to reindeer ribonuclease was 4-fold concentrated with an Amicon S 125 serum concentrator (Amicon Corp., Lexington, MA). Micro-complement fixation was carried out in disposable styrene U-plates containing 96 wells as described by Casey (1965). 4 C'H~0-units complement were used. When comparing very related antigens, the antiserum was diluted by steps of 20%. The titer difference between two proteins was a measure of antigenic difference. The titer in the heterologous system is sometimes equal to but generally lower than in the homologous system. The ratio of the antiserum dilutions used, is defined as the titer difference. Antigen concentrations were determined by amino acid analyses with a Technicon TSM-1 amino acid analyser. Samples were hydrolysed in 0.4 ml 6M HCI at about 110C in evacuated sealed glass tubes for 18-22 hr. Because of differences in mol. wts, quantities of the various ribonucleases were expressed in molar amounts. RESULTS

Antiserum production Cow, gnu, reindeer, and whale ribonucleases were rather immunogenic in rabbits giving antiserum titers of about 4000, 66,000, 1000 and 1000, respectively as measured by the micro-complement fixation method.

Immunod!ffusion Despite the relative insensitivity of immunodiffusion to small differences in structure as reported by Arnheim & Wilson (1967) spurs were frequently observed (see Fig. 1). Matrices showing the results of the comparison of the 9 ribonucleases using the 4 antisera, are shown in Figs. 2a-d. With these data, a relative order of antigenic reactivity was drawn up, as shown in Table 1. Immunodiffusions were also carried out with 2 first course antisera directed to bovine ribonuclease and 1 first course antiserum directed to bovine ribonuclease A (carbohydrate free). Only minor differences were observed compared with the results given in the matrix (Fig. 2a).

sitive to small differences in primary structure. Employing antisera directed against cow, gnu and reindeer ribonuclease, a rather good correlation with the differences in sequence is obtained. From the results with antiserum to whale ribonuclease it is obvious that a method which uses only one antiserum does not hold to predict sequence differences between homologous proteins. The ribonucleases, varying 18-26% in amino acid sequence from whale ribonuclease, show titer differences ranging from 4-128 (Table 2). Reciprocal measurements were done for four species (cow, gnu, reindeer and whale). Expressing our data in 'immunologic distances' as defined by Prager & Wilson (1971a), made possible a comparison between our results and those of others (Champion et al., 1975). In Fig. 3, it is shown that our data fit fairly well to the line drawn by Champion et al., (1975); one point (-~), representing the reaction of dromedary ribonuclease with antiserum to whale ribonuclease, is rather off the straight line; this will be discussed below (see Section 4 below).

Interpretation of antigenic differences In Fig. 4 part of the amino acid sequences of the 9 ribonucleases are shown. Residues common to all sequences have been omitted. By comparing the differences in sequence and the titer difference (listed in Table 2), we attempted to localise antigenically relevant regions. It was assumed that most of the amino acid substitutions fixed during evolution had only a small or no effect on the main chain conformation of a series of homologous proteins (Champion et al., 1975). (1) Reaction with antiserum to bovine ribonuclease. No difference in antigenic reactivity was found using bovine and goat ribonuclease, suggesting that Thr 3,

I!

Micro-complement fixation In Table 2, the titer differences (see methods) and the amino acid differences are given. F r o m these data, a relative order of antigenic reactivity could be deduced. The results are compared with the relative order obtained from the differences in amino acid sequence (see Table 1). In Table 3, the amino acid sequence differences between the ribonucleases from four deer species, reindeer, moose, fallow deer (Leijenaar-van den Berg & Beintema, 1975) and red deer (Zwiers et al., 1973; Welling et al., 1975b) are given with the titer differences observed by using antiserum directed against reindeer ribonuclease.

/

/

[] DISCUSSION

Sequence-antigenicity correlation The results from immunodiffusion and micro-complement fixation experiments coincide rather well (see Table 1). Evidently, complement fixation is more sen-

Fig. 1. Immunodiffusion of antiserum to bovine ribonuclease (central well) against different ribonucleases; from reindeer (1), pig (2 and 4), gnu (3), giraffe (5) and dromedary (6). Dromedary ribonuclease gave a faint precipitation line which is hardly visible on the photograph.

Immunologic Comparison of Pancreatic Ribonucleases (a)

• Cow Goat Gnu Giraffe Reindeer Pig Dromedary Whale Horse

(b)

Gnu Giraffe Cow Goat Reindeer Pig Dromedary Whale Horse

(c)

Reindeer Goat Gnu Cow Giraffe Dromedary Pig Whale Horse

(d)

Whale Dromedary Goat Cow Giraffe Gnu Reindeer Pig Horse

655

i i i p p p n n Cow

i i p p p n •n Goat

i i p p n n Gnu

i p p n n Gir.

p p n n Reind.

p n n Pig

n n Drom.

n Whale

i i i p p p p p Gnu

i p p p p p p Gir.

i p p p p p Cow

p p p p p Goat

p p p p Reind.

p p p Pig

P P Drom.

Whale

i i i p p p p p Reind.

i i i p p p p Goat

i i i i i p Gnu

i i i p i Cow

i i p p Gir.

i i i Drom.

i i Pig

i Whale

p p p p p p p p Whale

p p p p p p p Drom

i p i i i p Goat

i i i i p Cow

i i i p Gir.

i i p Gnu

i P Rein&

P Pig

Fig. 2. Matrices with immunodiffusion results using antiserum to bovine ribonuclease (a), gnu ribonuclease (b), reindeer ribonuclease (c) and whale ribonuclease (d), in which the relationships between the ribonucleases are shown. (i, lines of identity; p, partial identity, i.e. spur formation by a ribonuclease of the horizontal line over a ribonuclease of the vertical row; n, no identity.)

Ala 19, Lys 37 and Asn 103 are not part of an antigenic determinant. Reaction with gnu ribonuclease, differing more from bovine ribonuclease at only two positions (residues 50 and 99), results in a titer difference of 1.8. This suggests that Ser 50 and/or Thr 99, are parts of antigenic determinants. Using the titer difference of giraffe and reindeer ribonuclease (2.6 and 3.2 respectively), it is possible to predict that these differences should result from substitutions at positions 13, 15, 17, 20, 31, 34, 59, 70, 76, 78, 80, 89, 96, 98 and 120. The amino acid side chains in the 5 antigenic determinants of myoglobin (Atassi, 1975) generally are accessible to solvent molecules (Lee & Richards, 1971). If a side chain of an amino acid in an antigenic region is not accessible, then at least the immediate surroundings is accessible. In ribonuclease, the side chains of residues 13 and 120 together with their surroundings are largely inaccessible to solvent molecules (Lee & Richards, 1971), indicating that these residues are unimportant for antigenicity. (2) Reaction with antiserum to gnu ribonuclease. A small titer difference (1.2) was found for goat ribonuclease, indicating that the amino acids which it differs from gnu ribonuclease (residues 50, 99 and 103) are not in an antigenic region. Bovine ribonuclease shows a titer difference of 2.5, which should be the result of amino acids which it differs from gnu ribonuclease IMM. 13,8

12

in addition to those it already differs from goat ribonuclease. This suggests that residues 3, 19 and/or 37 are part of an antigenic region. The same procedure can be applied to giraffe ribonuclease (titer difference 2.6), indicating that residues 13, 20, 31, 78, 89, 98 and 120 are in an antigenic reactive region. Residues 13 and 120 are less probable to be in such a region (see Section 1). Reindeer ribonuclease, which differs only 6% in amino acid sequence from gnu ribonuclease, shows the relatively high titer difference of 4.0, which should be caused by one or more of the residues 15, 17, 34, 59, 70, 76, 80 and 96. (3) Reaction with antiserum to reindeer ribonuclease. The reaction of goat and gnu ribonuclease with antiserum to reindeer ribonuclease, shows a relatively high difference (titer difference 1.6 and 3.2, respectively) with regard to the small differences in primary structure. This suggests that residues 50, 99 or 103, in which they differ from each other, may be part of an antigenic determinant. Moose ribonuclease differs like fallow deer and red deer ribonuclease at 6 positions from reindeer ribonuclease. However, it shows a titer difference of 2.5 upon reaction with antiserum to reindeer ribonuclease. This should be the result of amino acids at positions different from the other three ribonucleases. From Table 3 it can be deduced that this concerns residues 20 and 91, sug-

656

GJALT W. WELLING et al. Table 1. Relative order of antigenic reactivity of different ribonucleases determined with immunodiffusion and micro-complement fixation using antiserum directed against bovine ribonuclease (A), gnu ribonuclease (B), reindeer ribonuclease (C) and whale ribonuclease (D). The relative order of resemblance in amino acid sequence is also shown

Imm. dif.

A cow gnu whale goat > giraffe > reindeer > pig > dromedary > horse

CF

cow goat > gnu > giraffe > reindeer > dromedary > pig > whale > horse

Sequence

cow > goat > gnu > giraffe > reindeer > pig > dromedary > whale > horse B

Imm. dif.

gnu whale cow > giraffe > reindeer > pig > dromedary > horse goat

CF Sequence

gnu > goat > cow > giraffe > reindeer > pig > dromedary > whale > horse gnu > goat > cow > giraffe > reindeer > pig > dromedary > whale > horse C

Imm. dif.

gnu dromedary whale reindeer > goat > cow > giraffe > pig > horse

CF

cow reindeer > goat > gnu > giraffe > dromedary > pig > whale > horse

Sequence

cow horse reindeer > goat > gnu > giraffe > pig > whale > dromedary D COW

Imm. dif.

CF Sequence

whale > dromedary > goat > reindeer > giraffe > horse pig gnu whale > dromedary > goat > cow > gnu > reindeer > giraffe > pig > horse cow whale > pig > goat > gnu > horse > giraffe reindeer dromedary

With immunodiffusion the relative order was determined as follows: with the data from the second row of Fig. 2,a, a first classification was made: cow, goat, gnu, giraffe > reindeer, pig, dromedary > whale, horse. With the data from the fourth and fifth row, a further classification could be made: cow, goat > gnu, giraffe > reindeer > pig, dromedary > whale, horse. The spur formed by pig over dromedary ribonuclease supports the final classification listed in Table 1.

gesting that one of these residues or b o t h are part of an antigenically i m p o r t a n t region. Residue 91 is most drastically changed in moose ribonuclease: from positively charged lysine in the other deer ribonucleases to negatively charged aspartic acid. (4) Reaction with antiserum to whale ribonuclease. As mentioned before, dromedary ribonuclease differs 26~o in amino acid sequence from whale ribonuclease and shows a relatively low titer difference after reaction with antiserum to whale ribonuclease (see Table 2). Also the immunodiffusion experiments reveal a remarkable similarity between dromedary a n d whale ribonuclease (see Fig. 2d). These results could be explained by the presence of identical antigenic structures in whale and dromedary ribonuclease. We have c o m p a r e d the amino acid sequences (Fig. 4) to find short stretches of sequence which dromedary and whale ribonuclease b o t h have in c o m m o n but different from all other ribonucleases: (a) the region containing a m i n o acid residue Asn 22. However, this region is believed to have a different conformation in b o t h enzymes. Subtilisin Carlsberg cleaves drome-

dary ribonuclease in an external loop of the polypeptide chain between residues 19 and 20 (Welling et al., 1974b, 1975b) whereas whale ribonuclease is not cleaved; (b) the region containing amino acid residue Glu 52. Glutamic acid residue 52 is accessible to solvent molecules as has been calculated by Lee & Richards (1971) a n d is surrounded by a m i n o acid residues (Ser, Leu, Asp, Val) which have been found earlier in antigenic reactive regions (Atassi, 1975). Experiments are in progress to determine which part of the surroundings of residue 52 also belongs to a n antigenic reactive region. It is interesting to note that in cow, reindeer a n d whale ribonuclease (see Sections 1, 3 a n d 4) apparently the same region of the molecule is part of a n antigenic reactive region (residues 50 a n d 52 are rather close to each other, as c a n : b e seen in a 3-dimensional model of bovine ribonuclease S), whereas in gnu ribonuclease, serine 50 is substituted by a proline residue which p r o b a b l y is the reason why this part of gnu ribonuclease is not in an antigenic region (see Section 2).

Immunologic Comparison of Pancreatic Ribonucleases Table 2. Titer differences, optimal antigen concentrations in p-mol/ml, and % difference in amino acid sequence

Ribonuclease

Difference in amino acid sequence (%)

Titer difference

Optimal Ag conc. p-mol/ml

With antiserum to bovine ribonuclease Goat 3 1.0 Gnu 5 1.8 Giraffe 9 2.6 Reindeer 11 3.2 Dromedary 23 8 Pig 21 8 Whale 24 8 Horse 28 32

6 14 9 32 8 27 116 187

With antiserum to gnu ribonuclease Goat 3 1.2 Cow 5 2.5 Giraffe 8 2.6 Reindeer 9 4.0 Pig 17 16 Dromedary 22 32 Whale 24 128 Horse 25 256

3 2-7 2 8 14 8 58 94

With antiserum to reindeer ribonuclease Goat 8 1.6 Gnu 9 3.2 Cow 11 4.0 Giraffe 11 4.0 Dromedary 26 8 Pig 20 8 Whale 24 16 Horse 26 64

3 7 2 2 4 7 58 94

657

Influence of carbohydrate chain P a r t of the surface structure comprising antigenic determinants could be rendered inaccessible to antibodies by carbohydrate chains attached to ribonucleases. In 7 out of the 9 ribonucleases, carbohydrate is attached to residues at positions 21, 34, 62 or 76. Some of them, giraffe a n d pig ribonuclease, are completely glycosidated. Nevertheless this has no m a r k e d influence on the relative order of reactivity as compared with their relative differences in amino acid sequences, see Table I(A-C). For example, the completely glycosidated giraffe ribonuclease reacts with antiserum directed to bovine ribonuclease in a manner expected from its relative difference in amino acid sequence. This behaviour has also been verified by using antisera to bovine ribonuclease (containing a

ooe°~

15C -0

',3 .o I OC - -

o

ow ._E /,o,

[] o

Q,I

o 50--

~

With antiserum to whale ribonuclease Dromedary 26 4.0 Goat 22 6.3 Cow 24 6.4 Gnu 24 8.0 Reindeer 24 10.0 Giraffe 26 12.5 Pig 18 16 Horse 25 128

°

I--

2 3 4 8 8 2 7 47

-Z~n

I I0

Optimal antigen concentrations for the homologous combinations were 4 p-mol/ml, for bovine ribonuclease, 7 p-mol/ml for gnu ribonuclease, 4 p-mol/ml for reindeer ribonuclease and 15 p-mol/ml for whale ribonuclease. The relative order was deduced from the titer difference and after that, a further classification was made by using the optimal antigen concentration. The more antigen used to obtain a certain complement fixation, the less was the antigenic reactivity.

I

k

20 Sequence difference,

30 %

Fig. 3. Dependence of immunologic distance on % sequence difference among pancreatic ribonucleases, bacterial azurins, bird lysozymes and ~ subunits of bacterial tryptophan synthetases. The expression immunologic distance (Prager & Wilson, 1971a) was used to compare our results with those of others. The immunologic distance between different ribonucleases (O, II) is defined as 100 × log titer difference (Table 2). The data for bacterial azurins, bird lysozymes and ~ subunits of bacterial tryptophan synthetases (O, I ) are from Champion et al. (1975). Solid points represent the averages of reciprocal tests.

Table 3. Amino acid sequence differences between reindeer, red deer, moose and fallow deer ribonuclease with titer differences obtained after reaction with antiserum to reindeer ribonuclease

Reindeer Red deer Moose Fallow deer

17

20

34

35

87

89

91

99

Titer difference

Pro Thr Ala Met

Ala Ala lie Ala

Asp Lys Asn* Lys

Leu Met Leu Met

Thr Ser Ser Ser

Ser Asn Asn Asn

Lys Lys Asp Lys

Thr Ala Thr Ala

1.3 2.5 1.6

* Carbohydrate attached to this residue.

658

GJALT W. WELLING et al. I

134 KTA K SA KSA KSA KSA KS P

2

3

6 9 3 5678901238 1 A E M S STSAAS SSQ K A E M S S T S S A S S SQ K A EM S STSSAS SSQ K A E 1 S S T S S VS S S Q T A EMPSPSSAS SSQQ KQM PDS SS SNSS L S

45

24 5 78 SN L KD SN L QD SN L QD SN L QD S D L QD RNMQG

9 R R R R R R

6

7

70257912349 VSAQVSKNVAQ V PAQV S KNVAQ VSAQVSKNVAQ V SAQVSKNVAQ VSAQVFKNVAQ VSAQVS INVNQ

8

9

0346780689689 TYQYSTSEGSAKT TYQY S T S EG S AKA TYQYSTSEGSAKT T Y Q Y SA S E G N A Q T SYQNSAHEGSVKT TYQNSTHQGSAKA

10 0 12348 T QANKV T QAKKV T QAEKV T QAEKV T QAEKV S QEQKV

STAEEMSYSSSSSNSQKREMNG--ISEQVSKSVTQTHQSTSHEGSAKASNL RSPMQM SGNSPGNNPQM RKMQG KSPM E M SGSTS SNPTQ K RNMQG

R VS E K V S K N V L R TYENSTHQGSAKT VPAQI LKNI TQ SYQS S SHL SGAQT

KQKl

S QKEKV S QKERV

12 0 F F F Y F F

256 A A A A A A

Cow Gnu Goat Giraffe Reindeer Pig

FA

Dromedary

F N F AQT

Whale Horse

Fig. 4. Amino acids at variant positions in the nine ribonucleases studied. One-letter code is used: A, alanine; D, aspartic acid; E, glutamic acid; F, phenylalanine; G, glycine; H, histidine; I, isoleucine; K, lysine; L, leucine; M, methionine; N, asparagine; P, proline; Q, glutamine; R, arginine; S, serine; T, threonine; V, valine; Y, tyrosine. The numbering of bovine ribonuclease is used.

glycosidated and a non-glycosidated component) produced in different rabbits and by using antiserum to bovine ribonuclease A (carbohydrate free). Only minor differences in the relative order of reactivity were observed. For instance with antiserum to bovine ribonuclease A: cow dromedary goat > gnu > giraffe > reindeer > pig > horse whale and with the second course antiserum directed to bovine ribonuclease: cow gnu goat > giraffe > reindeer > pig > dromedary >

whale horse "

So we conclude that the presence of a carbohydrate chain in pancreatic ribonucleases is not of major importance for the antigenicity of these proteins. In conclusion, ribonucleases from different animal species with different amino acid sequences have been used to determine which part of their molecular structure is important for their antigenicity. Differences in sequences from 3-28~o affect the antigenic reactivity, which could be used to elucidate the evolutionary relationship between pancreatic ribonucleases. A serious drawback of this approach when used with small proteins like ribonucleases with only a few antigenic regions, is that the extent to which particular amino acid replacements will affect the antigenic reactivity is not known in advance.

Acknowledgements--We thank Dr. F. Westendorp Boerma and Mrs. J. SchriSder-Nijboer for their help in preparing antiserum to bovine ribonuclease and Mr. J. S. Bouwer for his assistance in preparing the other antisera. We thank Drs. H. G. Seijen, A. C. Wilson, E. M. Prager and J. Drenth for helpful comments. Part of this work has been carried out under auspices of the Netherlands Foundation for Chemical Research (S.O.N.) and with financial aid from the Netherlands Organisation for the Advancement of Pure Research (Z.W.O.).

REFERENCES

Amheim N. & Wilson A. C. (1967) J. biol. Chem. 242, 3951-3956. Atassi M. Z. (1975) Immunochemistry 12, 423-438.

Brown R. K., Delaney R., Levine L. & Van Vunakis H. (1959) J. biol. Chem. 234, 2043-2049. Brown R. K. (1960) Fedn Proc. 19, 201. Brown R. K., Tacey B. C. & Anfinsen C. B. (1960) Biochim. biophys Acta 39, 528-530. Carlisle C. H., Palmer R. A., Mazumdar S. K., Gorinsky B. A. & Yeates D. G. R. (1974) J. molec. Biol. 85, 1-18. Casey H. L. (1965) Public Health Monograph No. 74, United States Government Printing Office, Washington, DC. Champion A. B., Soderberg K. L., Wilson A. C. & Ambler R. P. (1975) J. molec. Evol. 5, 291-305. Emmens M., Welling G. W. & Beintema J. J. (1976) Biochem. J. 157, 317-323. Gaastra W., Groen G., Welling G. W. & Beintema J. J. (1974) FEBS Lett. 41, 227-233. Groen G., Welling G. W. & Beintema J. J. (1975) FEBS Lett. 60, 300-304. Jackson R. L. & Hirs C. H. W. (1970) J. biol. Chem. 245, 637-653. Kartha G., Bello J. & Harker D. (1967) Nature, Loud. 213, 862-865. Lee B. & Richards F. M. (1971) d. molec. Biol. 55, 379-400. Leijenaar-van den Berg G. & Beintema J. J. (1975) FEBS Lett. $6, 101-107. Margoliash E., Nisonoff A. & Reichlin M. (1970) d. biol. Chem. 245, 931-939. Prager E. M. & Wilson A. C. (1971a) J. biol. Chem. 246, 5978-5989. Prager E. M. & Wilson A. C. (1971b) J. biol. Chem. 246, 7010-7017. Scheffer A. J. & Beintema J. J. (1974) Eur. J. Biochem. 46, 221-233. Singer S. J. & Richards F. M. (1959) J. biol. Chem. 234, 2911-2914. Smyth D. G., Stein W. H. & Moore S. (1963) d. biol. Chem. 238, 227-234. Welling G. W., Scheffer A. J. & Beintema J. J. (1974a) FEBS Lett. 41, 58-61. Welling G. W., Groen G., Gabel D., Gaastra W. & Beintema J. J. (1974b) FEBS Lett. 40, 134-138. Welling G. W., Groen G. & Beintema J. J. (1975a) Biochem. J. 147, 505-511. Welling G. W., Leijenaar-van den Berg G., van Dijk B., van den Berg A., Groen G., Gaastra W., Emmens M. & Beintema J. J. (1975b) Biosystems 6, 239-245. Westendorp Boerma F., Nijboer J., Vella F., Wong S. C. & Huisman T. H. J. (1974) Clin. Chim. Acta 55, 4~55. Wierenga R. K., Huizinga J. D., Gaastra W., Welling G. W. & Beintema J. J. (1973) FEBS Lett. 31, 181-185. Wyckoff H. W., Tsernoglou D., Hanson A. W., Knox J. R., Lee B. & Richards F. M. (1970) J. biol. Chem. 245, 305-328. Zwiers H., Scheffer A. J. & Beintema J. J. (1973) Eur. J. Biochem. 36, 569-574.