Fractionation of the multiple forms of bovine gastric aspartic proteases by chromatofocusing

ANALYTICAL

BIOCHEMISTRY

143, 256-264

(1984)

Fractionation of the Multiple Forms of Bovine Gastric Aspartic Proteases by Chromatofocusing PATRICE MARTIN AND CHRISTIANCORRE Institut

National

de la Recherche Agronomique, 65. rue de Saint-Brieuc.

Laboratoire de Recherches de Technologie 35042 Rennes-Cedex, France

Laitike,

Received April 16, 1984 By extending the chromatofocusing technique to a very acidic pH range (down to pH 2.0) a method which, in a single-step procedure, allows separation of the three main aspartic proteases secreted by the bovine abomasal mucosa i.e., chymosin (EC 3.4.23.4), gastricsin (EC 3.4.23.3), and pepsin A (EC 3.4.23. I), has been developed. Starting materials for separation were crude commercial milk-clotting extracts or abomasal juices. A multistep procedure, using narrower pH gradients, enabled the fractionation of these proteases into their multiple forms. Chymosins A and B, which are known to differ only by a single amino acid substitution (Asp/Sly), were completely resolved. Their elution pHs, 3.75 and 3.80, respectively, though far from their “normal” pls (around 4.7 in isoelectric focusing), demonstrate the resolving power of such a technique. Multiple forms of bovine pepsin A, which differ in their organic phosphate content (O-3 phosphate group(s) per molecule of enzyme) and whose pZs are lower than 2.5, were also separated using 15-20 mM glycine buffer, pH 2.0, as eluent. Although many attempts to get a linear gradient remained unsuccessful within this pH range, resolution appeared quite satisfactory, as judged from analytical isoelectric focusing patterns. In particular, the two subcomponents of bpA,, which presumably have a different site of post-translational phosphorylation, were resolved in this way. o 1984Academic press, IIIC. KEY WORDS: chromatofocusing; chymosin; pepsin A; gastricsin; fractionation; microheterogeneity.

The fourth stomach (or abomasum) of ruminants secretes, as zymogens, three milkclotting aspartic (or acid) proteases called: chymosin (EC 3.4.23.4), pepsin A (EC 3.4.23.1), and gastricsin (EC 3.4.23.3). The secretion of chymosin, which is the main enzyme produced by the abomasum mucosa of the unweaned calf, decreases after weaning (l), and pepsin A becomes the major enzymic component found in the adult bovine abomasum. Developmental changes in the secretion of gastric proteases also occur in most mammals (2). It was suggested that such developmental patterns were related to the presence or absence of the requirement for postnatal uptake of immunoglobulins (3). The developmental stage at which the secretion of gastricsin occurs is not known precisely. Small amounts of gastricsin may be present OQO3-2697184

$3.00

Copyright 0 1984 by Academic F’ms. Inc. All rights of reproduction in any form reserved.

in the newborn pig gastric secretion (2) and it seems that in gastric tissue of human fetus progastricsin appears before pepsinogen (4). In the course of studies in gastric secretion, we have previously purified and characterized bovine gastricsin (5). Its purification, involving three chromatographic steps, was difficult and time consuming. This is a serious drawback for an enzyme whose stability is notably low. Chymosin and gastricsin are separated easily from pepsin A by ion-exchange chromatography on DEAE-cellulose (6), but resolved poorly by such a technique, although they exhibit different p1 values (ranging between 4.5-5.0 and 3.5-4.0, respectively) in isoelectric focusing (5). Chromatofocusing, which displays the high resolution of isoelectric focusing without the drawback of a heating inherent in the application of an 256

CHROMATOFOCUSING

electric current, and which is able to separate proteins according to their p1 values, was expected to separate chymosin from gastric&. Chromatofocusing was successfully applied in the pH range 4.5-2.0, by a single chromatographic procedure, to separate chymosin, gastricsin, and bovine pepsin A present in abomasal juices or crude commercial extracts. Moreover, the microheterogeneity of each protease was also revealed. This is of special interest, as far as bovine pepsin A is concerned, since the only chromatographic procedure described so far, i.e., hydroxyapatite chromatography, fails to resolve bovine pepsin A into homogeneous fractions according to their phosphate contents (7,8). MATERIALS

AND

METHODS

Enzymes. Bovine chymosin, pepsin A, and gastricsin were prepared as described previously (95). A liquid commercial preparation of bovine pepsin (extracted from adult bovine abomasa) was provided by Laboratoire Presure Granday, Beaune, France. Abomasal juice samples, supplied by La Station de Recherches Zootechniques (Institut National de la Recherche Agronomique, Rennes, France) were obtained from calves equipped with an innervated abomasal pouch, cut out in the fundic part (10). Juice, secreted in the pouch, was recovered into a refrigerated flask in 2 M piperazine-HCl buffer, pH 6.0, using a catheter. The juices were filtered, dialyzed against distilled water, and finally freeze dried. Chromatofocusing. Chromatofocusing was performed at 20°C on a 1.5 X 30-cm column, packed with the polybuffer exchanger (PBE)’ 94 (Pharmacia, Uppsala, Sweden) equilibrated in the starting buffer, 25 mM piperazine-HCl (pH 5.5 to 4.4). Samples were applied, either in starting or eluent buffers. Eluents were polybuffer PB 74 (Pharmacia) diluted 1: 10, ’ Abbreviations used: IEF, isoelectric focusing; PBE, polybuffer exchanger; PB, polybuffer; bpA, bovine pepsin A; Hepes, 4-(2-hydroxyethyl)-l-piperazineethanesulfonic acid.

OF

GASTRIC

PROTEASES

257

adjusted to different pH values (4.0 to 2.0) with HCl and supplemented or not with various additives such as propanediol, urea and NaCl. The final step of elution was obtained with 15 or 20 mrvt glycine-HCl buffer, pH 2.0. All buffers were degassed before use. Elution was achieved at a constant flow rate of 25 ml/h and 3- or 5-ml fractions were collected while the absorbance at 280 nm was continuously monitored. A preliminary set of analytical chromatofocusings were first carried out with immunologically pure bovine chymosin, gastricsin, and pepsin A to set up “optimum” experimental conditions before fractionating crude commercial extracts or abomasal juices. Electrophoretic titration curve. A 1% agarose gel slab ( 110 X 110 X 5 mm), containing 2% LKB carrier ampholytes in the pH range 2.5-4.5, was cast with a 2-mm-wide, g-cmlong and 0.25-mm-deep slot in the middle. The pH gradient was preformed by applying, between electrode paper strips impregnated with 0.5 M CH,COOH at the anode and 0.4 M Hepes at the cathode, a 3-W constant wattage for 90 min at 4°C using a Pharmacia ECPS 3000/ 150 power supply. The electrode strips were at right angles to the slot. After removing the paper strips, the pH gradient was measured by means of a surface pH electrode and the gel was turned 90”. New paper strips soaked in the same electrolytes were applied on both sides of the gel perpendicular to the former. The slot was filled with the sample (45 ~1 of 1% w/v bpA) and electrophoresis was carried out, perpendicular to the pH gradient, for 30 min at 500 V. The gel was then fixed, dried, and stained as described previously (11). Enzyme assay. Activity was measured at 30°C using a 0.2% (w/v) K-casein solution in 50 mM sodium citrate, 75 mM NaCl, 0.2% (w/v) NaN3 buffer, pH 5.3 (12). The K*casein was prepared according to Zittle and Custer (13) from milk of a single cow (Francaise Frisonne Pie Noire breed) which was homozygous at the KCnA locus. IdentiJcation of proteases. Fractions cor-

258

MARTIN

AND

responding to each peak were pooled, alternatively saturated or not with ammonium sulfate, to separate protein from polybuffer, dialyzed, and freeze dried. The material thus obtained was analyzed by double-radial immunodiffision according to Ouchterlony ( 14) and (or) by isoelectric focusing (IEF) in ultrathin agarose gels.

CORRE

ment, carried out with PB 74 adjusted to pH 3.0, a linear gradient was obtained between pH 4.8 and 3.0. Chymosin again eluted around pH 3.8, in the middle of the profile, but with some loss in resolution. Therefore subsequent chromatofocusings were performed using narrower pH ranges. Clear-cut separations were obtained when the starting buffer was 25 mM piperazine, 10% (v/v) propanediol, pH 4.40, and the eluting buffer RESULTS was PB 74 (diluted 1: lo), 10% (v/v) propaChromatofocusing runs with pure enzymes. nediol, pH 3.60 (Fig. 1). Such conditions Since the pZ of chymosin was estimated to gave a linear gradient between 3.82 and 3.60 be around pH 4.7 by IEF (5), it was first and chymosin was resolved into chymosins applied on a PBE column equilibrated in 25 A and B as shown by IEF (Fig. 2). Samples mM piperazine-HCl buffer, pH 5.5, and in starting buffer or eluent gave identical eluted with PB 74 which had been adjusted chromatofocusing patterns. However, soluto pH 4.0 with HCl. Under such conditions, bilization of chymosin was easier in PB 74. no elution of chymosin occurred. Elution From our experience with chymosin, it was obtained only between pH 3.5 and 4.0 was expected that gastricsin would elute at by subsequently applying a 15 mM glycinean apparent pH less than its “true” pZ. The HCl buffer, pH 2.0. To get elution of chy- chromatofocusing profile, shown in Fig. 3, mosin closer to the expected pH, i.e., pH was in agreement with such an assumption. 4.7, 50 mM NaCl was added to the eluent. Indeed, elution of gastricsin occurred between By increasing ionic strength, elution of chy- pH 3.413.6 and 3.15, depending upon the mosin began around pH 4.3 to 4.4. However, gradient used; the actual pZ of gastricsin the pH gradient appeared nonlinear and ranges between pH 3.5 and 4.0. The microshowed a pronounced plateau in the early heterogeneity reported previously (5) was stage of the elution. No significant improveconfirmed in these experiments. ment in shifting the elution peak of chymosin With bovine pepsin A elution was achieved nearer to its estimated pZ resulted from low- by applying PB 74 (diluted 1: lO)-HCl, 10% ering the dilution factor of PB 74 to 1:8, nor (v/v) propanediol, pH 2.0, or 15-20 mM by supplementing the eluent with 2 M urea glycine-HCl, 10% (v/v) propanediol buffer, or propanediol (O-20% v/v), which is known pH 2.0, as eluents. Unfortunately, the pH to reduce the dielectric constant of the me- gradient obtained was not linear and elution dium and to modify the pK of ionizable always occurred after the gradient had groups. However, 10% (v/v) propanediol pro- abruptly declined. Attempts made to obtain voked a slight improvement in resolution. a linear gradient were unsuccessful, presumTherefore, it was added systematically in ably because the buffering capacity of the starting and eluting buffers. exchanger PBE 94 is extremely weak below Another approach was intended by ex- pH 3.0. The best results were obtained by tending the pH range in which chromatofoincreasing propanediol concentration to 20% cusing could be performed. In a first experi(v/v) and, as a consequence, by raising the ment PB 74 (1:lO) was adjusted to pH 3.8. pH to 2.5 (Fig. 4). Bovine pepsin A, which Elution of chymosin occurred between pH is known to display microheterogeneity, 3.8 and 3.9. A linear gradient was observed linked to phosphate content (O-3 phosphate between pH 5.0 and 3.8. Moreover, the res- groups per molecule), was resolved into six peaks. Their purity was examined by IEF in olution was improved. In a further experi-

CHROMATOFGCUSING

OF GASTRIC

259

PROTEASES

4

I, 3

ss

SO Fraction

7s Number

FIG. 1. Chromatofocusing of pure chymosin. Piperazine-HCI (25 mM), 10% (v/v) propanediol, pH 4.4, was used as starting buffer. Elution was achieved by applying PB 74 (1:10), 10% (v/v) propanediol, pH 3.5. Activity (continuous line) is l/f, X IO’, t, being the clotting time, expressed in seconds, of a 0.2% w-casein solution.

6

PH

5 e/ -\

4

Chymosin

C

Chymosin

B

Chymosin

A

4 1

23451

2. Analytical isoelectric focusing in agarose gel of pure chymosin and its fractions from chromatofocusing, in the pH range 4-6. The samples were: (1) whole chymosin which contains two main components (A and B) and a third minor component (C); (2) pure chymosin C; (3) pure chymosin B; (4) and (5) peaks 1 and 2 of the chromatofocusing run shown in Fig. 1. Pure chymosin A was not available. The gel was prepared as described under Materials and Methods for electrophoretic titration curve. Focusing was carried out for 3 h at 4 W and 4°C. The electrode solutions were 0.5 M acetic acid (anode) and 0.5 N NaOH (cathode). Samples of 7 pl (0.2-l% w/v) were applied, without a prefocusing run, in the middle of the gel slab. FIG.

260

MARTIN

Fraction

AND CORRE

Numb-r

FIG. 3. Chromatofocusing of pure gastricsin. Starting and eluting buffers were PB 74 (1: lo), pH 3.6, and PB 74 (I: lo), pH 3.0, respectively, both containing 10% (v/v) propanediol. Activity (continuous line) is l/r, X lo’, f, being the clotting time, expressed in seconds, of a 0.2% Kcasein solution.

the pH range 2.5-4.0. Figure 5 shows that materials contained in those peaks were eluted in the order of their isoelectric pHs. Some of the peaks gave several bands, strongly suggesting that there was complex formation between some of the components of bovine pepsin A. However, the fractions corresponding to the four major forms, previously called pAo, pAI, PA*, and pA3 (8) exhibited satisfactory homogeneity. Moreover, pAI, which contains two components, pA I -a and pA I +, presumably differing in the location of their

Frrstion

unique phosphate group, and so far only distinguishable by isoelectric focusing, appeared to be perfectly resolved in that chromatofocusing run (fractions 5 and 6). Chromatofocusing of abomasal juices and commercial crude extracts. A typical elution profile found when the starting material was a sample of calf abomasal juice is seen in Fig. 6. The results showed there was a single peak of chymosin (likely chymosin B), whereas bpA was split into four peaks. Several abomasal juices obtained from unweaned

Numbmr

FIG. 4. Chromatofocusing of pure bovine pepsin A. Starting and eluting buffers were PB 74 (1: lo), pH 3.3, and 20 mM glycine, pH 2.5, respectively, both containing 20% (v/v) propanediol. Activity (continuous line) is l/r, X lo), te being the clotting time, expressed in seconds, of a 0.2% Kcasein solution.

CHROMATOFOCUSING

OF GASTRIC

261

PROTEASES

3

-

PH

c

= -\ -

lli .‘

/ , -

~PAO ~PAI, bpL\lb b-2

‘WA3

2 I

1

967664321

FIG. 5. Analytical isoelectric focusing in agarose gel, in the pH range 2.5-4.0, of bovine pepsin A and its fractions from chromatofocusing. Preparation of the gel was as described under Materials and Methods for electrophoretic titration curve. Focusing conditions were as previously described (11). Samples were (I) whole bpA: (2) dephosphorylated bpA; (3-9) successive peaks emerging from chromatofocusing column in the run shown in Fig. 4.

calves were analyzed by this procedure. Gastricsin was always absent, whereas it was present when the starting material was commercial crude extracts which had been prepared from adult abomasal mucosa (see Fig. 7 and 9). Crude commercial extracts sepa-

FPLCtiDn

rated by a single-step procedure (Fig. 7) provided good separation of the three major aspartic proteases expected to be present. Chymosin was eluted in the flat part of the gradient, thus allowing a rough fractionation of the B and A forms. Gastricsin and bpA

Numbrr

FIG. 6. Typical chromatofocusing profile of calf abomasal juice. The starting buffer was 25 mM piperazine, pH 4.4, and eluting buffers were PB 74 (l:lO), pH 3.25, followed by 15 mrvt glycine, pH 2.0. All buffers contained 10% (v/v) propanediol. The sample was 20 mg freeze-dried abomasal juice from an unweaned calf in 1 ml starting buffer. Activity (continuous line) is I/r, X lo’, t, being the clotting time, expressed in seconds, of a 0.2% k-casein solution. The absorbance at 280 nm is the dotted line.

262

MARTIN

AND CORRE

determined each time but, in experiments performed with pure enzymes, relatively acceptable yields ranging between 70 and 85% were obtained. DISCUSSION

Frrction

Numbrr

FIG. 7. Single-step chromatofocusing run of a crude commercial milk-coagulant extract. The starting and eluting buffers were 25 mM piperazine, pH 4.4, and PB 74 (I: lo), pH 2.0, respectively, both containing 10% (v/ v) propanediol. The sample, which was 50 ml commercial extract dialyzed, freeze-dried, and dissolved in 20 ml starting buffer, contained approximately 10 mg chymosin, 10 mg gastricsin, and 100 mg bpA. Fractions were 3 ml. Activity (continuous line) is l/r, X 103, fe being the clotting time, expressed in seconds, of a 0.2% K-casein solution. The absorbance at 280 nm is the dotted line.

A chromatofocusing procedure was successfully applied to the fractionation of multiple forms of abomasal aspartic proteases. It appears to be a very simple and powerful technique. Indeed, chymosins A and B, which are known to differ only by a substitution Asp/Gly at position 290 in the prochymosin sequence ( 16,17), leading to slightly different pls (see Fig. 2), are perfectly separated by the method we have developed (Fig. 1). Despite many attempts, we were unable to get a linear pH gradient below pH 3.0. The extension of the method down to pH 2.0, using 15-20 I’BM glycine as eluent, allowed the fractionation of bovine pepsins A differing by their post-translational phosphorylation levels. Although resolution is comparable to that observed in analytical IEF, elution in chro-

PH

6

eluted in the steep end of the pH gradient and were resolved satisfactorily judging by the IEF pattern (Fig. 8). A three-step run, derived from the assays carried out with pure enzymes, is shown in Fig. 9. The resolution was greatly improved and chymosin as well as bpA were fractionated into their multiple forms. A third peak of chymosin (the first to be eluted) appeared, presumably corresponding to chymosin C, which is a degradation product of chymosin A (15). Alternatively, this peak could correspond to pseudochymosin, which occurs when activation takes place at pH 2.0 (3). Some fractionation of gastricsin occurred. Moreover, active materials which eluted between chymosin and gastricsin were also resolved. Recoveries of aspartic proteases were not

6

4

3

12346 FIG. 8. Analytical isoelectric focusing in the pH range 2-6 of the fractions obtained from the chromatofocusing run shown in Fig. 7. The samples were: (1) pure chymosin; (2) pure bpA; (3, 4 and 5) peaks 3, 2, and I, respectively.

CHROMATOFOCUSING

OF GASTRIC

E> Ti a

Fr~ctlon

PROTEASES

263

I

Numbsr

Rc. 9. Three-step chromatofocusing run of a crude commercial milk-coagulant extract. The starting buffer was 25 mM piperazine, pH 4.4. Eluting buffers were PB 74 (l:lO), pH 3.6, followed successively by PB 74 (l:lO), pH 3.0, and 15 mM glycine, pH 2.0. All buffers contained 10% (v/v) propanediol. The sample was the same as in Fig. 7; 5-ml fractions were collected. Activity (continuous line) is l/f, X lo’, t, being the clotting time, expressed in seconds, of a 0.2% k-casein solution. The absorbance at 280 nm is the dotted line.

matofocusing does not occur at the expected pZ values. Chymosins and gastricsins were eluted at least 0.5 to 1.0 pH unit below their pZs as estimated by IEF. Various compounds, such as urea, NaCl, and propanediol, were added to PB 74 to influence the elution pH of chymosin. Changing ionic strength was the only modification which accelerated elution of chymosin. Addition of 50 mM NaCl to PB 74 triggered elution, which started at pH 4.5, close to the expected pZ, lying between 4.5 and 5.0. This presumably occurred because of weakening the interactions between the exchanger and chymosin. However, the effect of ionic strength on chromatofocusing behavior of proteins, previously mentioned for tryptophanyl-tRNA ligase (18), also affected linearity of the pH gradient. In attempts made to improve chymosin fractionation, we noticed that when elution occurred either in the first part or at the end of the gradient, elution pHs of chymosin were slightly different, i.e., 3.8 vs 3.9, respec-

tively. This is consistent with the presumed effect of 50 mM NaCI, since ionic strength in the eluent gradually increases as elution progresses because ampholyte concentration is increasing at the same time. This is also consistent with the fact that by bringing the dilution factor of PB 74 up to l/ 13, a later elution was observed, while by lowering it down to l/8 there was no modification compared with the l/10 recommended dilution factor (results not shown). Nevertheless, the abnormally low elution pH of chymosin is likely ascribable to a tendency for this protease to precipitate near its isoelectric pH. Conversely, bpA was eluted above its “normal” pZ. The unphosphorylated form (bp&), whose pZ is close to 2.5 as estimated from the electrophoretic titration curve shown in Fig. 10, starts to elute at pH 2.9 (Fig. 4). Moreover, bpA, and pbAJ, whose pZs are likely lower than 2.0 (Fig. lo), emerge from the column before the eluent reaches pH 2.0, the lower limit of the gradient. This suggests

264

MARTIN

AND CORRE

of multiple forms of aspartic proteases secreted by bovine abomasal mucosa. However, the separation procedure is time consuming. Indeed, 2 days were necessary to perform the experiment shown at Fig. 9. ACKNOWLEDGMENTS

I

1

,

PH

We thank P. Guilloteau and R. Toullec for providing calf abomasal juices, Christiane Hulin for her secretarial assistance, and B. Ribadeau Dumas for critically reading the manumipt. We are grateful to Dr. Head for correcting the English manuscript.

2.2

4.0 3.0 FIG. 10. Electrophotetic titration curve of whole bovine pepsin A. The experimental conditions are given in the text. Multiple forms of bpA differ by their organic phosphate content. The bpAe (the unphosphorylated form) titration curve crosses the slot, where the sample was filled, at pH 2.5 That pH corresponds to its isoelectric point. pIs of bpA, , bpAr, and bpA,, which contain 1, 2, and 3 phosphate group(s) per molecule and whose curves do not cross the slot, are lower than 2.2.

that elution not only depends upon the charge of bpA but also involves competition between ions of the eluent and bpA for the charged groups of the exchanger, as in ampholyte displacement chromatography. This is so unless the buffering capacity of the exchanger at this pH is too low, which would preclude a strong enough binding of bpA to the exchanger. However, addition of propanediol to the eluent should be taken into account. Propanediol lowers the dielectric constant of the medium, and consequently increases apparent ionization pK of charged groups occurring on proteins and carrier ampholytes in the moving phase as well as those of the stationary phase (exchanger). Thus, propanediol influences the apparent pH at which proteins elute. By bringing the propanediol concentration up to 20%, bp& begins to elute at pH 3.25 (Fig. 4) instead of 2.9 as it does in 10% propanediol (result not shown). The chromatofocusing method we have developed and extended to very acidic pHs is easy and reproducible, with great resolving ability. As described it permits fractionation

REFERENCES 1. C¬, P., Toullec, R., Thapon, J.-L., Martin, P., Hoang, M.-T., Mathieu, C.-M., and Ribadeau Dumas, B. (1977) I. Dairy Rex 44, 9-23. 2. Foltmann, B., and Axelsen, N. H. (1980) FEBS Proc. HI,27 l-280. 3. Foltmann, B. (1981) Essays Biochem. 17, 52-84. 4. Hirsh-Marie, H., Loisillier, F., Touboul, J. P., and Burtin, P. (1976) Lab. Invest. 34, 623-632. 5. Martin, P., Trieu-Cuot, P., Collin, J.-C., and Ribadeau Dumas, B. (1982) Eur. J. B&hem. 122, 31-39.

Gamot, P., Thapon, J.-L., Mathieu, C.-M., Maubois, J.-L., and Ribadeau Dumas, B. (1972) I. Dairy Sci. 55, 1641-1650. 7. Lang, H. M., and Kassell, B. (197 I) Biochemistry 10, 2296-230 1. 8. Martin, P., Trieu-Cuot, P., Corre, C., and Ribadeau Dumas, B. (1982) Biochimie 64, 55-64. 9. Martin, P., Raymond, M.-N., Bricas, E., and Ribadeau Dumas, B. (1980) B&him. Biophys. Acta 6.

612,410-420.

10. Guilloteau, P., and LecalvC, J.-L. (1977) Ann. Biol. Anim. Biochim. Biophys. 17, 1047-1060. Il. Martin, P. (1984) Biochimie 66, 371-384. 12. Douillard, R., and Ribadeau Dumas, B. ( 1970) Bull. Sot. Chim. Biol. 52, 1429-1445. 13. Zittle, C. A., and Custer, J. H. (1963) I. Dairy Sci. 46, 1183-l 188. 14. Ouchterlony, 0. (1949) Acta Puthol. Microbial. Stand. 26, 507-515. 15. Foltmann, B. (1964) C. R. Truv. Lab. Carlsberg 34, 319-325. 16. Foltmann, B., Pedersen, V. B., Kauffman, D., and Wybrandt, Cl. (1979) J. Biol. Chem. 254, 84478456. 17. Moir, D., Mao, J., Schumm, J. W., Vovis, G. F., Alford, B. L., and Taunton-Rigby, A. (1982) Gene 19, 127-138. 18. Carter, C. W., and Green, D. C. (1982) Anal. B&hem. 124, 327-332.