Image Processing of Proteinase- and Methylamine-transformed

THE JOURNAL oP BIOLOGICAL CHEMlSTRY 0 1989 by The American Society for Biochemistry and Molecular Biology, Inc.

Vol. 264, No. 20, Issue of July 15, pp. 1204612052,1989 Printed in U. S. A.

Image Processing of Proteinase- and Methylamine-transformed Human ac2-Macroglobulin LOCALIZATION OF THE PROTEINASES*

(Received for publication, December l&1988)

Nicolas BoissetS, Jean-Christophe TaveauS, Franqois Pochone, Annette Tardieull, Martine Barray((, Jean N. Lamy$, and Etienne Delainll** From the $.Laboratoire de Biochimie Fondamentale, 2bis Boulevard Tonne&, F-37042 Tours Cidex, France, the $Znstitut Curie Biologic, 91405 Orsay, France, the Wentre de Gt!nitique Molkdaire, Centre National de la Recherche Scientifiqw, 91190 Gif-sur- Yvette, France, and the I( Laboratoire de Microscopic Cellulaire et Mokculaire, URA 158, Centre National de la Recherche Scientifique, Znstitut Gustave-Roussy, 94805 Villejuif Ckdex, France

Comparative x-ray scattering experiments and electron microscopic observations have been performed on native S-form, and on different F-forms of human plasma cu2-macroglobulin (a2M), obtained by proteinase (chymotrypsin, plasmin, and thrombin) or methylamine treatment. Image processing of electron micrographs of the (w2M molecules transformed by chymotrypsin, plasmin, and methylamine displayed average images which could be compared. The proteinase-com- plexed a2M molecules exhibited the usual H-like structure, but the methylamine-inactivated ones showed a different organization, with almost no stain-excluding material in the central region of the molecule, which therefore presented a central cavity filled with stain. By subtracting average images of cY2M-methylamine from a2M-chymotrypsin or a2M-plasmin, a putative localization of the proteinases inside the (u2M molecule, very close to its center was revealed. The values of the radii of gyration for the S- and F-forms obtained by s!- ray scattering were very different (78 and 67.7 A, respectively). All four scattering curves of the F-forms were comparable in shape and showed maxima and minima different from that of the S-form (w2M. Image processing of electron micrographs and x-ray scattering have provided independent results which indicate that a large cavity exists in the cr2M-methylamine molecule and that the proteinases might be located in a very central position inside the cY2M-proteinase molecules.

Lu2-Macroglobulin (a2M)’ is a tetrameric plasma protein of M, 725,000 that inhibits all classes of endopeptidases by a “trap” mechanism (1). The stoichiometry of the reaction is usually two proteinases for one a2M for small proteinases like chymotrypsin (MC 21,600), or 1:l for larger enzymes like plasmin (Mr 84,000). The proteolytic cleavage of the subunits induces an activation of thiolester sites which trigger an important conformational change. This allows the physical entrapment of the proteinase(s) into the ~u2M molecule, with

* The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “aduertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

** To whom correspondence should be addressed. ’ The abbreviations used are: aZM, a2-macroglobulin; ol2M-MA,

a2M inactivated by methylamine; (uZM-C, (uZM-P, and cu2M-T, aZM-chymotrypsin, cu2M-plasmin, and a2M-thrombin complexes; CORAN, correspondence analysis.

covalent or noncovalent bonding of the proteinase(s) (l-4). The thiolesters can also be hydrolyzed by primary amines (5, 6), resulting in a similar conformational change for human a2M. However, this phenomenon is not general: methylamine treatment does not induce the conformational change of plasma a2M in different species (7-IO), nor that of ovo- macroglobulin (11). The proteinase- and the methylamine- transformed cy2M molecules (fast form) migrate similarly and more rapidly than the native ones (slow form) in acrylamide gels (6). X-ray scattering studies demonstrate the same similarity between proteinase- and methylamine-transformed LY~M (12) and the important differences compared with the native form (13). Spectroscopic and ultrastructural studies of human cu2M have revealed that the proteinase- and methylamine-transformed cv2M molecules are similar but not strictly identical (14, 15). Biophysical studies have provided data on the relative position of two chymotrypsin molecules inside an LYBM (16), and models have proposed that trapped proteinases are localized inside the a2M molecule (8, 17, 18). Immuno- electron microscopy has identified plasmin epitopes at the extreme pole of the a2M-plasmin complexes (19), but no direct visualizations have been described. Image processing of a few cY2M-trypsin complexes has provided preliminary results on the possible localization of two trypsin molecules in LYZM (17). Besides complementary x-ray scattering studies of either native or proteinase- and methylamine-transformed a2M, we present here a more detailed image analysis of (u2M molecules transformed by methylamine, chymotrypsin, or plasmin, which show important differences in the organization of these transformed cu2M molecules. Visualization of the proteinases inside the fast form of human o12M is therefore possible now.

MATERIALS AND METHODS

a2-Macroglobulin-Human plasma LYZM was prepared as described (20). n2M-chymotrypsin (ol2M-C) (I:2 stoichiometry), cr2M-plasmin (a2M-P), and cu2M-thrombin ((uSM-T) (1:l stoichiometry) were prepared and purified as described (21-23, respectively). Methylamine- inactivated cv2M (a2M-MA) was obtained by incubation of native &M in 50 mM NaCl, 0.1 M Tris-HCl, pH 8.0, with 0.2 M methylamine during 3 h.

Small-angle X-ray Scattering Studies-The data presented here were obtained with the new D24 version of the small-angle camera installed at the Synchrotron Radiation Laboratory LURE (Orsay) (24). The x-ray beam was monochromatized andfocused with a bent germanium crystal. The wavelength was 1.608 A and a point collimation geometry was used. The detector was a linear position- sensitive detector with delay line readout. The data were recorded using a CAMAC acquisition system designed at EMBL (25). Absolute scale measurements were obtained by reference to the scattering of a precalibrated carbon black sample. Sample-to-detector distance was

12046

Localization of Proteinases into Transformed a&Macroglobulin 12047

of the order of 1 m (&/channel = 1 x 10e4, 20 = scattering angle, and s = 2 sin B/h). Sample concentrations were varied from 10 to 38 mg/ ml. Exposure time at LURE was 20 min. Samples were contained in quartz capillary tubes, the diameter of which was determined by water absorption measurements. The determination of the radii of gyration (R,) was done as described (26).

Electron Microscopy-Negative staining with uranyl acetate and micrographs in the elastic bright field mode of the Zeiss EM 902 were obtained as described (27). Image processing was performed on these bright field images, although the illustration of the different forms of cu2M was with dark field inelastic images obtained with the same microscope on the O,,s edge of uranium at 96 eV (27). Precautions were taken for identical orientation of the grids in the microscope, as described (28).

Image Processing-Electron micrographs (X 50,000) of the three types of molecules were selected by optical diffract.ion with no astig- matism or drift. For (YAM-MA, ot2M-P, and o2M-C, 350,273, and 191 images of molecules were extracted from three, four, and four negatives, respectively.

Micrographs were digitized with a Perkin-Elmer 1010 G flatbed microdensitometer, using a spot size and a scanning step of 25 pm, corresponding to a pixel size of 0.5 x 0.5 nm.

Digitized images were analyzed using the SPIDER image processing software (29). The hierarchical ascendent classification was performed with the SPAD.N program (30).

cu2M molecules (40 x 40 nm = 80 X 80 pixels) were interactively windowed from the negatives and were submitted to contrast nor- malization. Resealed images were aligned on a reference image as previously reported (31), using the angular and cross-correlation method recommended by Frank et al. (32).

Because of the H-form of the transformed ~u2M molecules, all

images were first aligned on a reference image rendered symmetrical (2-fold symmetry). Then, a first correspondence analysis (CORAN) (31-33) separated two clusters of images according to their up or down orientation (180” rotation range). As previously reported (28), the average map corresponding to the negative values of factor 1 was used as a new reference, according to which all images were realigned with a 360” rotation range. A second CORAN was performed, which allowed the separation of the image set into clusters which were defined by hierarchical ascendent classification (34). The average image of each cluster was submitted to low-pass filtering in the Fourier space up to its resolution limit. The measurement of this resolution was performed on two sub-averages with the differential phase residual criterion (35).

Difference images were computed by subtracting the filtered average image of one cluster from the filtered average image of another cluster.

Computer Graphic Technique-This was performed as previously described (28).

RESULTS

Electron Microscopy of the a2M Molecules--Fig. 1 shows the morphology of three types of a2M used in this study, and for comparison, the native form of cu2M (a). The cu2M transformed by chymotrypsin (LuBM-C) by plasmin (cu2M-P) and by methylamine (cuBM-MA) are displayed in b, c, and d, respectively, with a representative large field and an inset typical molecule. a2M-C and ot2M-P are not easily distin- guishable and show the classical H-like shape; the latter is, however, often more stocky. The ot2M-MA is different from

FIG. 1. Electron micrographs of cu2M molr graphed in inelastic dark field. This mode of imaging gives -images in reverse contrast, where the stain- excluding domains of the molecules appear as dense regions. a), control native ru2M with its typical shape, an (Y letter (black arrows and inset), or in a tetrameric cross-like form (white arrows). b), n2M molecules transformed by chymotrypsin (1:2 stoichiometry). They have the typical H-like structure. c), a2M molecules transformed by plasmin (1:l stoichiometry). They also have a typical H-like structure but look more dumpy than a2M-C. d), tv2M molecules inactivated by methylamine. They are characterized by the absence of stain-excluding domains in their center. Magnifications: ~200,000, insets x 600,000. Scales represent 50 and 10 nm, respectively.

12048 Localization of Proteinases into Transformed a2-Macroglobulin

FIG. 2. Average images of 273 a2M-P molecules (a), 191 a2M-C molecules (b ) , and 350 a2M-MA molecules (c) , obtained after the first CORAN and realignment on the asymmetric reference images. Note that the two proteinase-transformed a2M molecules are rather similar but that the methylamine- transformed a2M presents an important longitudinal cavity with two small stain-excluding domains located between the upper and lower arms of the H. d and e show the difference images obtained by subtracting the a2M-MA average image from that of a2M-P ( d ) or a2M-C ( e ) . Centrally located twin domains are considered to represent the proteinases inside the a2M molecule.

the two proteinase-transformed a2M: there is no strong stain- excluding domain in the central region connecting the two walls, but only two or three small dense spots (inset, Fig. Id).

Image Processing-Fig. 2 shows the average maps of 1x2" P (a), a2M-C ( b ) , and a2M-MA (c) obtained after the first CORAN and realignment on the asymmetric reference images. Fig. 2, d and e show the difference images obtained by subtracting a2M-MA from a2M-P ( d ) and from a2M-C (e). One can clearly see that the difference images reveal the existence of a pair of dense structures in the center of the image. The densities are nearly equal in each paireand are close together. They seem to be better separated (53 A center- to-center for a2M-C which contains two chymotrypsin mol- yules, than for a2M-P which contains only one plasmin (46 A center-to-center). These densities are presumed to correspond to the proteinases inside the a2M molecule.

It is very likely that various orientations of the a2M molecules on the support film and individual variability in their structure explain an absence of homogeneity among the av-

FIG. 3. Correspondence analysis of a2M-MA. The clusters outlined on the CORAN map corresponding to the first two factors have been delimited by hierarchical classification. The insets show the average maps of the main three clusters A, B, and C.

FIG. 4. Corresondence analysis of a2M-P.

ACTOR 2

FIG. 5. Correspondence analysis of a2M-C.

erages of the different types of the a2M images. A clearer visualization of the proteinase(s) location should be obtaina- ble by subtracting more homogeneous averages. Figs. 3-5 show the CORAN maps obtained for aBM-MA, a2M-P, and a2M- C, respectively, with the three subsets of images designed as

Localization of Proteinases into Transformed a2-Macroglobulin 12049

FIG. 6. Two difference images have been selected here. For a2M-P, we show the difference between cluster C of a2M-P and cluster B from aPM-MA. All contour lines have been highlighted (a), and this image in reverse contrast has been superimposed onto the corresponding average image of cluster C of a2M-P ( b ) . In c and d, the same treatment was done from the C clusters of a2M-C and a2M- M A . The contour of the central region delimits the zone in which the proteinases are presumed to be located. The present results confirm those obtained with the first average images (Fig. 2).

A, B and C. The selection of the images from the clusters was determined by hierarchical ascendent classification, then the average image was calculated for each cluster. At present, there is no obvious correlation between the various clusters of each class of molecules, nor any obvious relationship with particular orientations or biochemical characteristics of the a2M molecules.

a2M-C molecules (Fig. 5) have two stain-excluding nuclei in the middle of the walls and a third one located below the horizontal central region connecting these two walls. In fact, there are two stain-excluding domains connected to the two walls, which give rise to the two characteristic domes located in the vertical axis of the H-form. Assigning the lower position to the dome with the maximum density, this orientation was used for the final alignment of the a2M molecules. The central region of the image actually exhibits a zone of less stain- exclusion, and the top-left arm is less stain-excluding than the three others. This gives an asymmetrical image with more or less deep furrows or cavities inside the molecule. The differences in the images from the three clusters after CORAN can be interpreted to be due to some flexibility of the molecules, to slight differences in their orientation on the support film, or to minor conformational changes.

a2M-P molecules (Fig. 4) have roughly the same structural organization, except that two arms (upper right and lower left) present a distortion which makes the molecule dumpy and a little larger than the a2M-C. Stain-filled furrows appear deeper than for a2M-C.

For these two proteinase-transformed a2M, the images show a rather complex organization of the central region which was called the “plateau” in previous work (28). The

I I lo 1 I 0.01 0.02 0.01 0.02

s - 2 sin e /A ( ~ - 1 ) FIG. 7. X-ray scattering intensities recorded with native

and modified a2M. Background was subtracted from the curves which were put on an absolute scale but were otherwise uncorrected. a, native a2M from two different preparations. b-e, a2M transformed by chymotrypsin, plasmin, thrombin, and methylamine, respectively, compared with the native form of the same batch. f, superposition of the last four figures. For each form, the curves were found independ- tn t of concentration except a t very low scattering angles (s < 0.003 A-’). Therefore, the curves presented here are combinations of low concentration curves for the low angle part, and of high concentration curves for high angle part.

present data do not reveal continuous density in this region but rather oblique structures between the two vertical walls, and this is even more conspicuous for a2M-MA.

a2M-MA molecules have the same organization at the level of the two walls, but the main differences from the proteinase(s)-containing molecules are in the central region. Two subsets ( A and C , Fig. 3) possess two cavities located at the right of the vertical axis. The third ( B ) shows a deep furrow in an axial position. As compared with a2M-C and a2M-P, a2M-MA molecules are characterized by the absence of strong stain-excluding domains in the central portion of the molecule. The extremities of the two domes observed in the proteinase-transformed a2M molecules occupy the same posi- tions in a2M-MA, but they look rather like isolated zones of stain-exclusion with less evident connections with the other dense structures of the molecule.

All the possible subtractions between the average images of clusters A, B, and C from the three types of transformed a2M have been done (data not shown). All the difference images between a2M-P or a2M-C and a2M-MA average images reveal the existence of the two central densities already observed in the global average images of the three types. How- ever, slight differences appear between all these subtracted images. Those corresponding to a2M-P were more heterogeneous (from cluster to cluster) than those obtained with a2M- C. This apparent discrepancy suggests that the a2M-P clusters may be composed of more heterogeneous populations of images. In an attempt to find an explanation for this hetero-

12050 Localization of Proteinases into Transformed a2-Macroglobulin

geneity, the variance maps of the three clusters A, B, and C of a2M-P were carefully examined but no particular heterogeneity was found compared with that of a2M-C clusters. Moreover, the filtration of all the images up to the resolution limit computed by the phase-residual method should have removed the meaningless details from all the averages, and consequently, from the difference maps. We thus consider that the cluster-to-cluster heterogeneity of a2M-P is probably related to subtle structural differences or to different orientations on the support film.

When a2M-MA images are subtracted from a2M-protein- asefs) images, the remaining central densities can be interpreted as being the proteinases. In order to have the best view possible of the actual position of the single plasmin and the two chymotrypsin molecules in a2M, we have superimposed the contour lines of the two difference images on the corresponding average image of a2M-P and a2M-C (Fig. 6). If 220 A, the total height of the a2M-C molecule, is taken as a reference, one can measure that the two identical spots in a2M-C are 41-45 A in diameter and with centers distant by 42 A. For a2M-P, the two unequal spots range from 33-37 A, also with 42 A between their centers. All these results con- cerning the position, the shape, and the size of these densities are very similar to those obtained by subtracting of the global average images (Fig. 2). Similar results obtained before and after the separation of the average images into three clusters by CORAN is a good argument in favor of the homogeneity of these average images.

Small-angle X-ray Diffusion Data-Fig. 7 shows the x-ray scattering curves of the different forms of a2M tested, and the comparison between the native and the modified forms. For the native, the experiments were performed with two different batches corresponding to two different preparations (a). The same curve shape was obtained after the central maximum, a slight minimum was observed around s = 0.0065 A" and a second maximum at s = 0.001 A-l. The minimum was, however, slightly shifted toward low angles and the radius of gyration was higher for the first preparation than for the second one. The difference found was reproducible, whatever the protein concentration, which indicates stability of the protein preparations. The curves for the four transformed forms of a2M molecules are quite similar and rather different from thpse of native a2M (b-e). The first minimum at s = 0.0085 A" becomes a change in slope around s = 0.009 4-', and the secondary maximum is shifted toward = 0.012 A-'. The radius of ration is found to be 67.7 f 1.7 A as compared with 78 k 2 R for the native form. Some slight differences may also be observed between the modified forms: after a central peak, nearly identical for all four curves, the secondary maximum was highest for a2M-MA and lowest for a2M-P. The molecular weights 850,000 f 150,000 calculated from the extrapolated intensities at the origin were consistent with the a2M molecular weight, but no significant differences could be observed for the a2M-MA and the a2M-proteinase forms.

X-ray scattering data were found to be similar although not identical to those previously published (12, 13). If the fadius of gyration of the native form is the same, 78 f 2 A, our measured radius of gyration for a2M-MA is slightly smaller, 67.7 f 1.7, as compared with 72 f 2 A. More importantly, the secondary maximum of both native and modified forms seems more pronounced, while the curves for the native andemodified forms definitively cross each other around s = 0.008 A-'. Such a difference may arise from the fact that collimation correc- tions are not needed in this work.

These results show that the shape of all transformed a2M molecules is similar and rather different from that of the

native. Such an important conformational change has only been observed previously in one protein, ATCase (37). In addition, the measured radii of gyration are in good agreement with the open structures seen in electron microscopy. They provide independent arguments for the presence of important cavities inside a2M, since the radius of gyration of a globular protein of the same molecular weight can be calculated to be 48 A (and a radius of 62 A). Finally, the similarity of the R, values for a2M-MA and for the other modified forms provides an independent argument that the proteinases are located in the central region of the a2M molecule.

DISCUSSION

In the present paper, we compare by electron microscopy and x-ray scattering the structural characteristics of native and transformed human a2M. The x-ray data strongly support the homogeneity of the preparations, the structural similarity of the different modified forms, and the large difference between those forms and the native. Electron microscopy, in combination with image processing, demonstrates the differences between the methylamine- and the proteinase-transformed a2M. A recent description of antiplasmin plus colloi- dal gold-labeled IgG binding to a2M-P molecules demon- strated an axial position of the plasmin molecule, with a clear accessibility at one and possibly two extremities of the a2M (19). Owing to the similarity of the conformational change induced in a2M by methylamine or proteinases, these authors considered that a2M-MA could be used as an appropriate control for a2M-proteinases. Beginning with the same hy- pothesis, we subtracted the images of a2M-MA from those of a2M-C or a2M-P, in order to localize the proteinases inside the a2M molecule. This resulted in a clear positioning of the two chymotrypsin and the single plasmin molecule inside the transformed a2M.

In a2M-C, we observed two identical spots which were absent in a2M-MA. TPey were 41-45 A in diameter, and their centers distant by 42 A. The two chymotrypsin molecules (40 A diameter) are thus very close to one another, consistent with distance measurements by energy transfer (16), as well as with the results of Schramm and Schramm (17) who observed in processed images of a2M-trypsin molecules two small protein domains close to each other, and presumed to represent the trypsin molecules in the complex.

In a2M-P we have found a central position for the proteinase from samples in which double binding of plasmin was avoided by purification of a2M-P from the unbound proteinase. This is known to disrupt weak binding of plasmin to a2M, i.e the possible second plasmin molecule. We thus consider it likely that the plasmin could be located near the center of the a2M molecule and that its elongated shape would be perceived in the a2M molecule in an oblique position.

An imperfect subtraction between a2M-P and a2M-MA was observed compared with the clear difference images obtained from a2M-C and a2M-MA. This suggests that two chymotrypsin molecules would fit quite well into the cavity, inside the a2M molecule, whereas a single plasmin molecule would produce some distortion of the a2M structure, and thus cause the difficulties in comparison and subtraction of a2M- P and a2M-MA. a2M-MA may indeed not be a perfect control for a2M-proteinases. Although the conformational change is actually known to be due to thiolester hydrolysis (38), other structural changes directly linked to peptide bond cleavage in the bait region should be taken into account.

A three-dimensional visualization of a2M-MA and a2M- proteinases is shown on Fig. 8 to illustrate a possible inter-

Localization of Proteinases into Transformed a2-Macroglobulin 12051

FIG. 8. Computer graphic schematic drawing illustrating the position of one plasmin and two chymotrypsin molecules inside the a2M molecule, by comparison with the a2M-MA. The three types of molecules are presented successively (a2M-MA a; a2M-P: b; a2M-C: c), with from left to right: the side view, an oblique view, and the top view. The a2M-MA structure is simplified as a two-walled structure with two small spheres; the connection of the spheres with the walls is not specified. The*two chymotrypsin molecules are represented as two spheres of 40 A in diameter. The single plasmin molecule is represented by an ellipsoid profile of about twice the size of a chymotrypsin.

pretation of the results obtained by image processing. This computer-generated model does not claim to show the actual structure of the transformed a2M but only to represent the main results presented here and in previous work (15,27,28).

The a2M-MA molecule is drawn with two vertical walls and two small spheres located between the upper and lower arms. These spheres represent the two small nuclei of stain- exclusion perfectly visible in all the clusters of a2M-MA and may correspond to the NH2-terminal ends of the monomers, linked by the two inter-chain disulfide bridges. At the present time, we fail to understand the connection between these domains and the walls, and this is why a precise indication has not been given. The central portion of the molecule consists of an important cavity delimited by the spheres and the walls.

Two slightly staggered chymotrypsin molecules are located in the central cavity of a2M-C. For a2M-P, a single ellipsoid plasmin molecule is located in an oblique position with respect to the vertical axis of the a2M.

In our computer-drawn three-dimensional model of a2M-C (28), the two domes located in the upper and lower parts of the molecule were shifted away from the vertical axis because of the observed difference of density between the two median stain-excluding domains. This was responsible for a reduction in the symmetry in the molecule (disappearance of the vertical %fold axis). The present data complete and clarify our view of the dome structure, which is presumed to be composed of two parts, one belonging to the a2M molecule (the interarm nuclei of stain-exclusion visible in a2M-MA), and the other to the proteinase (the base of the dome).

The main result of this work demonstrates that the proba- ble position of proteinases is close to the center of the a2M molecule. In addition, the observed absence of stain-excluding material in the center of a2M-MA casts doubt on all previously published models of the F-form a2M (8, 12, 17-19, 27, 28, 39-41). They all describe an H-shaped structure with a central connection between the two walls. Since this central material is apparently mainly comprised of proteinases, new hypotheses on the organization of the a2M-MA molecule are needed so that new models may be built.

Acknowledgments-F. Van Leuven and Lorna Saint-Ange are thanked for critically reading the manuscript.

REFERENCES

1. Barrett, A. J., and Starkey, P. M. (1973) Biochem. J. 133 , 709-

2. Sottrup-Jensen, L., Petersen, T. E., and Magnusson, S. (1980)

3. Howard, J. B. (1981) Proc. Natl. Acud. Sei. U. S. A. 78, 2235-

4. Salvesen, G. S., Sayers, C. A., and Barrett, A. J. (1981) Biochem.

5. Steinbuch, M., Pejaudier, L., Quentin, M., and Martin, V. (1968)

6. Barrett, A. J., Brown, M. A., and Sayers, C. A. (1979) Biochem.

7. Dangott, L. J., and Cunningham, L. W. (1982) Biochem. Biophys.

8. Feldman, S. R., Gonias, S. L., and Pizzo, S. V. (1985) Proc. Natl.

9. Bjork, I., Lindblom, T., and Lindahl, P. (1985) Biochemistry 24,

10. Motoshima, A., Sera, M., Funakoshi, T., Shoji, S., Kubota, Y.,

11. Nishigai, M., Osada, T., and Ikai, A. (1985) Biochim. Biophys.

12. Osterberg, R., and Malmensten, B. (1984) Eur. J. Biochem. 143 ,

13. Branegird, B., Osterberg, R., and Sjoberg, B. (1982) Eur. J.

14. Larsson, L-J., Lindahl, P., Hall&-Sandgren, C., and Bjork, I.

15. Tapon-Bretaudikre, J., Bros, A., Couture-Tosi, E., and Delain, E.

16. Pochon, F., Favaudon, V., Tourbez-Perrin, M., and Bieth, J.

17. Schramm, H. J., and Schramm, W. (1982) Hoppe-Seyler’s 2.

18. Liu, D., Feinman, R. D., and Wang, D. (1987) Biochemistry 2 6 ,

19. Gonias, S. L., Allietta, M. M., Pizzo, S. V., Castellino, F. J., and Tillack, T. W. (1988) J. Biol. Chem. 2 6 3 , 10903-10906

20. Favaudon, V., Tourbez, M., Pochon, F., Mareix, R., and Tourbez, H. (1987) Eur. J. Biol. Chem. 165 , 31-37

21. Pochon, F., Tourbez, M., Favaudon, V., and Delain, E. (1987)

22. Pochon, F. (1987) Biochim Biophys. Acta 915,37-45 23. Pochon, F., Steinbuch, M., Lambin, P., and Kichenin, V. (1983)

24. Depautex, C., Desvignes, C., Feder, P., Lemonnier, M., Bosshard, R., Leboucher, P., Dagneaur, D., Benoit, J. P., and Vachette, P. (1987) Laboratoire pour l’utilisation du Rayonnement Elec- troma: Actively report, August 1985-April 1987, p. 87

25. Bordas, J., Koch, M. H. J., Clout, P. N., Dorrington, E., Boulin, C., and Gabriel, A. (1980) J. Phys. Sei. Instr. E 1 3 , 938-944

26. Tardieu, A., Laporte, D., Licinio, P., Krop, B., and Delaye, M. (1986) J. Mol. Biol. 192 , 711-724

27. Delain, E., Barray, M., Tapon-Bretaudikre, J., Pochon, F., Mar- ynen, P., Cassiman, J.-J., VanDen Berghe, H., and Van Leuven, F. (1988) J. Biol. Chem. 263,2981-2989

28. Boisset, N., Taveau, J.-C., Barray, M., Van Leuven, F., Delain, E., and Lamy, J. (1988) Biol. Cell. 6 4 , 45-55

29. Frank, J., Shimkin, B., and Dowse, H. (1981) Ultramicroscopy 6 ,

724

FEBS Lett. 121,275-279

2239

J. 195,453-461

Biochim. Biophys. Acta 154 , 228-231

J. 181,401-418

Res. Commun. 107,1243-1251

Acad. Sci. U. S. A. 8 2 , 5700-5704

2653-2660

and Ueki, H. (1988) Arch. Biochem. Biophys. 262,517-524

Acta 831,236-241

541-544

Biochem. 122,663-666

(1987) Biochem. J. 243,47-54

(1985) EMBO J. 4,85-89

(1981) J. Biol. Chem. 256,547-550

Physiol. Chem. 363,803-812

5221-5226

FEBS Lett. 217,101-105

FEBS Lett 161,51-54

343-358

12052 Localization of Proteinases into 30. Lebart, L., Morineau, A., and Warwick, K. M. (1986) Multivariate

Descriptive Statistical Analysis. Correspondence Analysis and Related Techniques for Large Matrices, p. 231, John Wiley & Sons, New York

31. Van Heel, M., and Frank, J. (1981) Ultramicroscopy 6, 187-194 32. Frank, J., Goldfarb, W., Eisenberg, D., and Baker, T. S. (1978)

33. Frank, J., Verschoor, A., and Wagenknecht, T. (1985) New Meth- Ultramicroscopy 3, 282-290

odobgies in Studies of Proteins Conjiguration (T. T. Wu, ed) pp. 36-89, Van Nostrand-Reinhold, New York

34. Lebart, L., Morineau, A., and Warwick, K. M. (1982) Multivariate Descriptive Statistical Analysis, J. Wiley & Sons, New York

35. Frank, J., Verschoor, A., and Boublik, M. (1981) Science 214, 1353-1355

Transformed a2-Macroglobulin

36. Castellino, F. J., Brockway, W. J., Thomas, J. K. Liao, H. T.,

37. Hew&, G., Moody, M. F., Tauc, P., Vachette, P., and Jones, P. T.

38. Van Leuven, F., Marynen, P., Cassiman, J.-J., and VanDen

39. BretaudiGre, J-P., Tapon-BretaudiGre, J., and Stoops, J. K. (1988)

40. Ruben, G. C., Harris, E. D., Jr., and Nagase, H. (1988) J. Biol.

41. Ikai, A., Nishigai, M., Osada, T., Arakawa, H., and Kikuchi, M.

and Rawick, A. B. (1971) Biochemistry 12,2787

(1985) J. Mol. Biol. 185,189-199

Berghe, H. (1988) J. Biol. Chern. 263,468-471

Proc. Nutl. Acad. Sei. U. S. A. 85,1437-1441

Chem. 263,2861-2869

(1987) J. Protein Chem. 6,Sl-93

Documents

Image Processing of Proteinase- and Methylamine-transformed