Proc. NatL Acad. Sci. USA Vol. 79, pp. 2912-2916, May 1982 Cell Biology

Receptor-mediated endocytosis of diphtheria toxin by cells in culture (clustering/fluorescence/video intensification microscopy/clathrin-coated pits) JAMES H. KEEN*, FREDERICK R. MAXFIELDt, M. CAROLYN HARDEGREE*, AND WILLIAM H. HABIGt *Fels Research Institute, Temple University School of Medicine, Philadelphia, Pennsylvania 19140; tDepartment of Pharmacology, New York University Medical Center, New York, New York 10016; and Wood and Drug Administration, Bureau of Biologics, Bethesda, Maryland 20205 Communicated by Sidney Weinhouse, January 4, 1982

ABSTRACT The binding and uptake of fluorescently labeled oftoxin with cells at physiological temperature resulted in much diphtheria toxin by cells in culture has been examined byusing epi- ofthe toxin becoming inaccessible to external agents-suggesting fluorescence video intensification microscopy. Rhodamine-la- entry into the cell. Finally, extensive degradation of toxin and beled diphtheria toxin retained significant toxicity on bioassay and inhibition of protein synthesis were observed (7). in cell culture and was tested for uptake by human WI-38 and The exact mechanisms by which toxin reaches the cell cy- mouse 3T3 fibroblasts grown in culture. When added to cells at toplasm and the factors responsible for the phenomenon of re- 37rC, toxin was observed to become concentrated and internalized sistance remain unknown. We have prepared a rhodamine-la- in discrete vesicles in both cell lines. The appearance of fluores- beled derivative cent clusters could be prevented by addition of excess unlabeled of diphtheria toxin (R-DT) that retained diphtheria toxin to the medium or by addition of ATP (which has 25-50% ofthe biological activity ofcontrol preparations. Using been shown to block toxin binding to cells), indicating that the rho- this derivative and video intensification microscopy (8), we have damine-labeled toxin was binding to diphtheria toxin-specific cell found that toxin undergoes receptor-mediated endocytosis§ (9): surface binding sites. When the simultaneous uptake of rhoda- it is recognized by specific cell surface binding sites and is in- mine-labeled diphtheria toxin and fluorescein-labeled a2-mac- ternalized in endocytic vesicles. Furthermore, this process oc- roglobulin was monitored, the two proteins appeared in the same curs in both diphtheria toxin-sensitive and toxin-resistant cell clusters indicating that the toxin undergoes receptor-mediated lines. The implications of this uptake process for the toxicity of endocytosis. Despite the difference in susceptibility to diphtheria the protein are discussed. toxin ofcells derived from sensitive (human) and resistant (mouse) tissues, the behavior of the rhodamine-labeled derivative in both MATERIALS AND METHODS cell lines was indistinguishable in terms of toxin required for for- Swiss mouse 3T3 clone A and human WI-38 cells were grown mation of clusters or inhibition by unlabeled toxin or by ATP. as described (10). The WI-38 cells were used between passages These results demonstrate that diphtheria toxin-specific cell sur- 16 and 34 and a single lot of calf serum (GIBCO no. 12K3101) face binding sites occur on both insensitive and sensitive cells and was used throughout these suggest that toxin is processed similarly by both cell types during studies (11). its initial cell surface binding and internalization by this pathway. DT was obtained from John Robinson ofVanderbilt Univer- The possible involvement of this uptake system in the mechanism sity. The preparation contained 20,000 minimum skin-reactive of action of diphtheria toxin in cells is discussed. doses (MRD) per Aug of protein (12) and was homogeneous on NaDodSO4 gel electrophoresis. R-DTwas prepared by reaction Extensive studies on the structure and mechanism of action of ofDT with tetramethylrhodamine isothiocyanate; DT at 10 mg/ diphtheria toxin (DT) have resulted in a detailed understanding ml in 0.05 M sodium borate (pH 9.2) was dialyzed for 24 hr at ofthe biochemical pathogenesis ofthe toxin (for review, see refs. 4°C against 100 ml of0.05 M sodium borate (pH 9.2) containing 1 and 2). DT is synthesized by Corynebacterium diphtheriae 2.0 mg of tetramethylrhodamine isothiocyanate (Cappel Lab- lysogenic for the ,3tox+ phage and is secreted as a single poly- oratories, Cochranville, PA, lot no. 8612). The labeled protein peptide chain of Mr 62,000. The toxin can be nicked by pro- was then dialyzed extensively against calcium- and magnesium- teases to yield A and B chains (Mr 21,000 and 41,000, respec- free Dulbecco's phosphate-buffered saline (Pi/NaCl) and then tively) that remain disulfide linked. The A chain catalyzes dialyzed overnight against Pi/NaCl supplemented with 1 M transfer ofthe ADP-ribose moiety ofNAD to cytoplasmic elon- NaCl. Finally, the toxin was again dialyzed against Pi/NaCl. gation factor 2 (EF-2), resulting in the arrest ofcellular protein Fluorescein-labeled a2-macroglobulin (F-a2M) was prepared synthesis. The B chain is catalytically inactive but appears to as described (13). be responsible for cell surface binding. Although EF-2 from Bioactivity of the toxins was monitored in two ways. (i) Skin eukaryotic sources is highly conserved and isolated toxin A chain reactivity was assayed by (0.1 ml) intradermal injections of di- will ADP-ribosylate EF-2 proteins from all eukaryotic sources lutions of the toxin in Pi/NaCl containing 0.2% gelatin into in cell-free extracts (3), rodents are relatively resistant to the guinea pigs (12). The area of erythema was estimated as the actions of DT (1). This resistance is continued in cell lines de- product oftwo diameters measured at right angles to each other. rived from mouse or rat tissues. (ii) In vitro protein synthesis assays were performed by using Following earlier studies of the binding of DT to sensitive confluent monolayers ofcells in 35-mm plastic dishes. The cells cells (4) and to membranes derived from both sensitive and in- were washed and dilutions of DT or R-DT were added in 1.0 sensitive cells (5), the existence ofspecific cell surface receptors on intact cells (6) has recently been documented. Incubation Abbreviations: DT, diphtheria toxin; R-DT, rhodamine-labeled diph- theria toxin; a2M, a2-macroglobulin; F-a2M, fluorescein-labeled a2M; Pi/NaCl, phosphate-buffered saline; EF-2, elongation factor. The publication costs ofthis article were defrayed in part by page charge § The phrase "receptor-mediated endocytosis" is used to describe the payment. This article must therefore be hereby marked "advertise- process by which ligands are bound to specific cell surface binding ment" in accordance with 18 U. S. C. §1734 solely to indicate this fact. sites, clustered and internalized into primary endocytic vesicles. 2912 Downloaded by guest on September 24, 2021 Cell Biology: Keen et aL Proc. Natd Acad. Sci. USA 79 (1982) 2913 ml of complete medium. After 4 hr at 37C, 0.5 /Ci of L- ['4C(U)]leucine (339 mCi/mmol; 1 Ci = 3.7 X 1010 becquerels; N Amersham) was added and incubation was continued for 3 hr. E The cells were then washed three times with P1/NaCl (con- E taining calcium and magnesium) at 00C and solubilized by ad- dition of1.0 ml of0.1 M NaOH. The solution was acidified with 200 1.0 ml of 30% trichloroacetic acid; 75 Ag of bovine serum al- w bumin (Sigma) was added and the protein precipitate was w washed with 1.0 ml of 10% trichloroacetic acid. Pellets were solubilized with 0.5 ml of 1 M NaOH and 0.5 ml of H20; 10 ml U- of ACS (aqueous counting scintillant, Amersham) was added 0 100 and radioactivity in the samples was counted in an Intertech- nique scintillation counter. Control incubations with cyclohex- wuUR imide (5 ,ug/ml, Sigma) were routinely performed. For assay ofR-DT interaction with cultured cells, cells were plated at 1.7 x 104 cells per cm2 in 35-mm plastic dishes. The following day the dishes were washed three times with serum- -3 -4 -5 -6 -7 free medium (3TC) and R-DT was added in 1.0 ml of serum- DILUTION (log) free medium containing 0.5% bovine serum albumin (370C). After incubation for 15 or 30 min at 370C, the dishes were washed three times with serum-free medium, then washed 100 three times with Pi/NaCl, and finally were fixed by immersion U) in Pi/NaCl containing 3.7% formaldehyde. After several min- e) LLJ 80 utes, the dishes were rinsed in Pi/NaCl and mounted for _ microscopy. Observation of the cells by phase-contrast and epi-fluores- 0z C 60 cence microscopy was performed as described (8) by using a Dage MTI model 65 silicon intensifier camera. For comparison, z0 40 all images within each figure were obtained at constant manual LL -,0 settings of the camera gain, kV, and recorder video level controls. 20 Protein was assayed by the method ofLowry et aL (14) or that of Bradford (15). Bovine serum albumin or gamma globulin 0 (Bio-Rad), respectively, was used for standards. All other chem- icals were reagent grade or better. 10-4 10-3 10-2 10-1 100 101 102 TOXIN (Mg/ml) RESULTS FIG. 1. Bioactivity of control DT and R-DT. (A) Guinea pig skin Rhodamine-labeled derivatives of DT have been prepared to reactivity assay of control DT (o) and R-DT prepared with 2.0 (i) or permit visualization of toxin interaction with cells in culture. 4.0(e) mgof tetramethyirhodamine isothiocyanate (see Materials and Skin in Methods). All stock toxin concentrations were 3.0 mg/ml. (B) Effect of reactivity assays guinea pigs revealed that increasing control DT (open symbols) and R-DT (closed symbols) on protein syn- derivatization resulted in a progressive loss ofbiotoxicity (Fig. thesis in human WI-38 (o, *) or mouse 3T3 (n, *) cells. 1A). DT treated with 2.0 mg of tetramethylrhodamine isothio- cyanate per 100 ml was found to retain 25% of the toxicity of the control; this preparation was used in all subsequent exper- Although this punctate pattern offluorescence could be rec- iments. In short-term assays of inhibition of protein synthesis ognized by using these concentrations, higher levels of R-DT by sensitive (human WI-38) cells in culture, this R-DT prepa- (0.32 AM) increased the signal compared to the background ration retained ==50% of the activity of control preparations autofluorescence and gave clearer patterns (Fig. 3). At these (Fig. 1B). In agreement with earlier studies (1, 16, 17), mouse toxin concentrations, rhodamine-specific fluorescence is quite (3T3) cells were resistant (Fig. 1B). apparent (Fig. 3 A-C). Toxin Uptake by Human Cells. To study the binding and uptake of toxin by human fibroblasts in culture, R-DT (65 nM) was incubated with WI-38 cells for 30 min. When rhodamine A excitation-emission of the washed and fixed cells was moni- tored, a punctate pattern offluorescence in the perinuclear re- gion and extending throughout the cell was observed (Fig. 2A). Prominent autofluorescence in the WI-38 cells accounted for some of this signal as shown by examination of the same field by fluorescein excitation-emission (Fig. 2B). The autofluores- cence component was generally diffuse or in large spots, was observed on untreated cells (not shown), and was characterized by a similar distribution and intensity with either the rhodamine or fluorescein filter set. Therefore, only that fluorescence on FIG. 2. Uptake of R-DT by human WI-38 cells. WI-38 cells were toxin-treated cells that was observed with the rhodamine but incubated in serum-free medium with R-DT (65nM) for 30 min at37TC. not with the fluorescein filter set was judged to reflect the pres- The cells were then observed by rhodamine (A) or fluorescein (B) epi- ence of R-DT. fluorescence. (x360.) Downloaded by guest on September 24, 2021 2914 Cell Biology: Keen et al. Proc. NatL Acad. Sci. USA 79 (1982)

FIG. 3. Effect of unlabeled DT and ATP on the uptake of R-DT by WI-38 cells. WI-38 cells were incubated in serum-free medium with R-DT (0.32 ,uM) for 30 min at 37TC alone (AWC), in the presence of unlabeled toxin (39 pM) (D-F), or in the presence of ATP (5 mM) (G-I). In each ex- periment, the same field was observed by rhodamine (A, D, and G) or fluorescein (B, E, and H) epi-fluorescence or by phase-contrast optics (C, F, andI). (x370.) The specificity of this pattern was investigated in two addi- ever, when inhibition of protein synthesis is measured in cell- tional ways. Incubation of WI-38 cells with R-DT in the pres- free extracts, all eukaryotic preparations are uniformly suscep- ence of a 120-fold excess of unlabeled toxin blocked the for- tible to the action ofthe toxin (3). Therefore, it was ofparticular mation of all rhodamine-specific fluorescent clusters (Fig. 3 interest to ask whether R-DT would interact specifically with D-F), indicating that binding of the rhodamine-labeled toxin resistant rodent cells. Incubation of R-DT with Swiss mouse is saturable and is competitively inhibited by the native toxin. 3T3 fibroblasts-under conditions similar to those employed Nucleotides-in particular ATP and other triphosphates-have with human WI-38 fibroblasts-resulted in the formation of been shown to block DT binding to cell surface receptors and fluorescent clusters indistinguishable in appearance from those to prevent the expression of biotoxicity (6, 18). When ATP (5 observed with the human cells (Fig. 4A and B). The appearance mM) was included with R-DT in the incubation medium of WI- of these clusters was blocked by simultaneous incubation with 38 cells, formation of the fluorescent pattern described above excess unlabeled toxin (Fig. 4 C and D) and, as with the human was blocked (Fig. 3 G-I). Substantial effect was also observed cells, ATP (5 mM) also prevented the formation of the R-DT at lower ATP concentrations (0.5-1.0 mM), although CTP (up clusters on the mouse fibroblasts (Fig. 4 E and F). Thus, these to 3 mM) had much less effect (not shown), in agreement with resistant cells also possess specific toxin binding sites. earlier findings (18). Nucleotides were entirely without effect Co-Localization of R-DT and F-a2M. To specifically deter- on the clustering of F-a2M (not shown). These results and the mine whether the fluorescent clusters ofR-DTobserved on 3T3 competition by unlabeled toxin indicate that the R-DT is bind- cells were likely to have been formed in clathrin-coated pits, ing to specific cell surface sites and that these sites have prop- we studied the simultaneous uptake ofR-DT with F-a2M. The erties consistent with those observed for toxin receptors in other distribution ofR-DT in (and on) a cell after a 30-min incubation systems (6).¶ is shown in Fig. 5A and the localization of F-a2M in the same Toxin Uptake by Mouse Cells. Rats and mice (and cultured field is shown in Fig. 5B. Superimposition ofpositive transpar- cells derived from rodent tissues) are much less sensitive to the encies ofA and B reveals that most of the spots are coincident actions of DT than their nonrodent counterparts (16, 17). How- in both fields (Fig. 5C), demonstrating that F-a2M and R-DT are present in the same structures. Coincidence of the fluo- I Because the R-DT has not been physically isolated, we cannot rule rescence due to random overlap is rare as shown in Fig. 5D; here out the formal possibility that it is being processed differently from one of the transparencies has been shifted (about 1 mm in the the unlabeled material in some way that we do not detect. scale ofthe figure) and very few fluorescent spots are observed. Downloaded by guest on September 24, 2021 Cell Biology: Keen et aL Proc. Natd Acad. Sci. USA 79 (1982) 2915

FIG. 5. Simultaneous uptake of R-DT and F-a2M by Swiss mouse 3T3 cells. Cells were incubated with R-DT (0.32 /AM) and F-a2M (50 ,gg/ml) for 30 min at 370C. The same field was then observed by rho- damine (A) and fluorescein (B) epi-fluorescence. Alignment ofpositive transparencies ofA andB resulted in C, whereas slight misalignment (about 1 mm) produced D. (x340.)

cells or on isolated membranes (5, 18). ATP entirely blocked the formation offluorescent clusters of R-DT on both sensitive and resistant cells whereas CTP was ineffective, in agreement with previously published studies on "MI-labeled DT binding to cells (18). The coincidence ofR-DT and F-a2M fluorescence -GaAs at both 15 and 30 min indicates that both proteins are present in the same endocytic vesicle at these times and suggests that DT-like a2M (10, 19)-is internalized after clustering in clath- FIG. 4. Uptake of R-DT by Swiss mouse 3T3 cells. Cells were in- rin-coated pit regions of the plasma membrane. Recently, ul- cubated in serum-free medium for30 min at 370C with R-DT (0.48 AM) trastructural studies with another protein toxin, Pseudomonas alone (A and B) or in the presence of unlabeled toxin (39 pM) (C and and in coated D) or ATP (5 mM) (E andF); The same field was then observed by rho- exotoxin A, revealed that it was bound clustered damine epi-fluorescence (A, C, and E) or by phase-contrast optics (B, pits on mouse LM cells (22). D, and F). No- image was observed by fluorescein epi-fluorescence. The toxin concentrations (o10-d M) used in this study are (x360.) substantially greater than LD50values (- 10-1o M) (6); this raises ,a question concerning the relevance of our observations to the R-DT and F-a2M fluorescence was also coincident when a 15- toxic process in cells. In toxicity studies, a small number of min incubation was performed (data not shown). Electron mi- molecules [possibly only one (23)] enters the cytoplasm and acts croscopic studies with 3T3 cells have shown that a2-macroglob- catalytically for hours. Higher toxin concentrations were re- ulin (a2M) is internalized through clathrin-coated pits (19) and quired to study early events in the toxic process-the cell sur- within the first 30 min (13, 20) is localized predominantly in face binding and internalization that occur during the first 30 primary endocytic vesicles (receptosomes). Thus, these results min. However, it should be noted that the toxin concentrations suggest that R-DT is also being processed in this manner by used approximate toxin-receptor dissociation constants (10-9- mouse cells. 10-' M) that have been measured with intact cells (6, 24) or isolated membranes (5). DISCUSSION There are several arguments in support ofthe hypothesis that Although the broad outline of toxin-cell interactions is clear, the process ofreceptor-mediated endocytosis that we have ob- the details of the process by which active toxin traverses the served is a physiologically relevant pathway for toxin entry. The membrane and reaches the cell cytoplasm remain unclear. Fur- binding of R-DT is specific and saturable, and the appearance thermore, the nature of the differences responsible for sensi- of R-DT in endocytic vesicles can be inhibited by nucleotides tivity and resistance is unknown-despite the equal suscepti- that block DT toxicity. The rate at which ligands such as a2M bility of EF-2 from all eukaryotic sources to toxin in vitro. The are cleared from the cell surface by receptor-mediated endo- results reported here demonstrate that DT undergoes inter- cytosis (9) is similar to the rate at which "I-labeled DT is in- nalization by receptor mediated endocytosis as do serum pro- ternalized (7). Toxin-specific antibody and concanavalin A have teins, peptide hormones, and viruses (9, 21). In addition, we been shown to inhibit the internalization of "MI-labeled DT: demonstrate that this uptake pathway operates in both toxin- both also block the cytotoxic effect (7). sensitive and toxin-resistant cells. Rapid degradation of internalized toxin, characteristic of ly- The formation offluorescent R-DT clusters on cells could be sosomal involvement, has also been documented (7). Chloro- blocked by an excess ofunlabeled toxin, indicating the involve- quine (25) and other lysosomotropic agents (26-28) that prevent ment of specific and saturable binding sites. Nucleotides have this lysosomal degradation of toxin also block its toxic effect. been shown both to block the expression oftoxicity in sensitive Finally, reduction of the extracellular pH to a level approxi- cells and to inhibit DT binding to specific high-affinity sites on mating that in the lysosome results in rapid intoxication of sen- Downloaded by guest on September 24, 2021 ,0QQU16- Cell Biology: Keen et d Proc. Natl. Acad. Sci. USA 79 (1982) sitive cells (27, 29) and in association of toxin with synthetic 5. Chang, T.-M. & Neville, D. M., Jr. (1978) J. Biol. Chem. 253, membranes (30). These observations are consistentwith passage 6866-6871. across the limiting membrane and entry of toxin into the cy- 6. Middlebrook, J. L., Dorland, R. B. & Leppla, S. H. (1978) J. Biol. Chem. 253, 7325-7330. toplasm only after delivery to an acidic environment. In fact, 7. Dorland, R. B., Middlebrook, J. L. & Leppla, S. H. (1979) J. Tycko and Maxfield (31) have recently shown that primary en- BioL. Chem. 254, 11337-11342. docytic vesicles containing a2M, in which we also find R-DT, 8. Willingham, M. C. & Pastan, I. (1978) Cell 13, 501-507. have a pH of5.0 ± 0.2 (±SEM). Thus, the receptor-mediated 9. Goldstein, J. L., Anderson, R. G. W. & Brown, M. S. (1979) Na- endocytosis of R-DT that we have observed, whereby ligands ture (London) 279, 679-685. are internalized into primary endocytic vesicles and delivered 10. Maxfield, F. R., Schlessinger, J., Shechter, Y., Pastan, I. & Wil- lingham, M. C. (1978) Cell 14, 805-810. to lysosomes (9, 13), has many of the properties of the toxic 11. Middlebrook, J. L. & Dorland, R. B. (1977) Can.J. Microbiol 23, pathway. Nonetheless, we cannot exclude the possibility that 175-182. additional entry pathways-which we were unable to de- 12. Barile, M. F., Kolb, R. W. & Pittman, M. (1971) Infect. Immun. tect-share these properties and contribute to or are entirely 4, 295-306. responsible for toxicity. 13. Willingham, M. C.; Maxfield, F. R. & Pastan, I. (1980) J. His- We have observed receptor-mediated endocytosis of DT in tochem. Cytochem. 28, 818-823. 14. Lowry, 0. H., Rosebrough, N. J., Farr, A. L. & Randall, R. J. resistant mouse fibroblasts as well as in sensitive cells. These (1951)1. Biol. Chem. 193, 265-275. experiments confirm the earlier observations of Chang and 15. Bradford, M. M. (1976) AnaL Biochem. 72, 248-254. Neville (5) that DT-specific binding sites exist on resistant cells. 16. Gabliks, J. & Solotorovsky, M. (1962)J. Immunol 88, 505-512. Sandvig and Olsnes (29) noted that acidification of the extra- 17. Bonventre, P. F. & Imhoff, J. G. (1968) J. Exp. Med. 126, cellular medium did not result in rapid intoxication ofresistant 1079-1086. mouse cells. This observation and the results reported here sug- 18. Middlebrook, J, L. & Dorland, R. B. (1979) Can.J. Microbiol. 25, 285-290. gest that post-internalization factors may be responsible for the 19. Willingham, M. C., Maxfield, F. R. & Pastan, I. (1979) J. Cell resistance manifested by rodent cells-e.g., retention of toxin Biol 82, 614-625. in the lysosome despite the acidic environment (29)' or the ab- 20. Willingham, M. C. & Pastan, I. (1980) Cell 21, 67-77. sence of a gene product responsible for translocation into the 21. Helenius, A., Kartenbeck, J., Simons, K. & Fries, E. (1980) J. cytoplasm (32). Alternatively, the endocytosis pathway may be Cell Biol 84, 404-420. unrelated to the toxic process. The present data do not allow 22. Fitzgerald, D., Morris, R. E. & Saelinger, C. B. (1980) Cell 21, 867-873. us to definitively distinguish among these possibilities and fur- 23. Yamaizumi, M., Mekada, E., Uchida, T. & Okada, Y. (1978) Cell ther studies are clearly indicated. 15, 245-250. 24. Uchida, T., Pappenheimer, A. M., Jr. & Harper, A. (1973) J. We are grateful to Charles 0. Boyd and Gwen Koths for assistance Biol Chem. 248, 3845-3850. in preparing the figures and to Sherry Battaglia for typing the manu- 25. Leppla, S. H., Dorland, R. B. & Middlebrook, J. L. (1980) J. script. This study was supported by National Institutes of Health re- Biol Chem. 255, 2247-2250. search Grants GM 28526 (J.H.K.) and CA 12227 (to the Fels Research 26. Kim, K. & Groman, N. B. (1965) J. Bacteriot 90, 1552-1556. Institute). 27. Draper, R. & Simon, M. I. (1980) J. Cell Biol 87, 849-854. 28. Mekada, E., Uchida, T. & Okada, Y. (1981)J. Biol Chem. 256, 1225-1228. 1. Collier, R. J. (1975) Bacteriol Rev. 39, 54-85. 29. Sandvig, K. & Olsnes, S. (1980) J. Cell Biol 87, 828-832. 2. Pappenheimer, A. M., Jr. (1977) Annu. Rev. Biochem. 46, 69-94. 30. Donovan, J. J., Simon, M. I., Draper, R. K. & Montal, M. (1981) 3. Johnson, W., Kuchler, R. J. & Solotorovsky, M. (1968) Bacteri- Proc. Natl Acad. Sci. USA 78, 172-176. ology 96, 1089-1098. 31. Tycko, B. & Maxfield, F. R. (1982) Cell, in press. 4. Boquet, P. & Pappenheimer, A. M., Jr. (1976) J. Biol Chem. 32. Creagan, R. P., Chen, S. & Ruddle, F. H. (1975) Proc. Natl 251, 5770-5778. Acad Sci. USA 72, 2237-2241. Downloaded by guest on September 24, 2021