Sialic Acid Content of the Erythrocyte and of an Ascites Tumor Cell of the Mouse

AARON MILLER, JQHN F. SULLIVAN, AND JAY H. KATZ

(P.AJ4ioisotope and Medical Services of the Boston Veterans Adminjairation Hospital, and the Department of Medicine, Boston University @SchoolofMedicine, Boston, Massachusetts)

SUMMARY The sialic acid content of the mouse erythrocyte was compared with that of the Ehrlich ascites carcinoma cell. Neuraminidase treatment of intact tumor cells yielded 36 times more sialic acid than was released from the erythrocyte on a per cell basis. If this were all located on the surface, it may be calculated that the density of sialic acid on the Ehrlich ascites carcinoma cell is about 4 times greater than that on the erythro cyte. Trypsin released a sialic acid-containing fragment from both types of cells, the sialic acid being in the â€œboundâ€•form.The sialic acid released by trypsin treatment of Ehrlich ascites tumor cells was only about one-sixth of that released by neuraminidase. By contrast, approximately equal amounts were liberated from the erythrocytes by these two enzymes. The erythrocyte sialic acid-containing substance was nondialyzable and was not precipitated by trichioroacetic acid. N-acetylneuraminic acid constituted approximately 70 per cent and N-glycolyneuraminic acid the remaining 30 per cent of the sialic acid of the tumor cell as determined by paper chromatography. In the erythrocyte N-acetylneuraminic acid constituted almost all the sialic acid, N-glycolyl neuraminic acid being undetectable in this cell.

The sialic acids are a family of substances ured isoelectric point of about pH 4.0 (18). It formed by the condensation of N- or 0- substitut seemed possible that sialic acid might be respon ed mannosamine with phosphoenol pyruvate sible for the negative charge of the EAT cell. In (17, 923). They are strongly acidic compounds this study the concentration of sialic acid in the whose terminal carboxyl group has a pK of about mouse EAT cell was determined and compared 92.6 (19). Sialic acids have been found on the eryth with that of the mouse erythrocyte. While this rocyte membranes of a number of vertebrate investigation was under way, Wallach and Eylar species (5, 7, 192). Viral or bacterial neuramini reported on the presence of sialic acid in the mouse dases, enzymes which cleave the a-ketosidic link EAT cell (921).The present study corroborates age between a terminal sialic acid group and a their findings and presents further data relating sugar or sugar derivative (9), can liberate free to the subject. sialic acid from such red cell membranes (5, 7). The surface of the mammalian erythrocyte is neg MATERIALS AND METHODS atively charged with a reported isoelectric point Blood was obtained by cardiac puncture from of about pH 92.0 (1, 8). Following enzymatic re 6- to 8-month-old female Swiss mice. Heparin was moval of sialic acid from the human erythrocyte, used as anticoagulant. Blood was pooled from a the isoelectric point of the red cell is greatly in number of animals (up to fifteen), depending on creased, indicating that these groups are primarily the amount of blood needed. The blood was centri responsible for the negative charge on the erythro fuged at 1,400 X g for 920minutes at 4Â°C., after cyte membrane (5, 7). which the plasma and buffy coat were removed. The mouse Ehrlich ascites tumor (EAT) cell The erythrocytes were washed 3 times with cold also has a negatively charged surface with a meas 0.15 M NaCl and then made up to approximately Receivedfor publication October 19, 1962. the original volume. When â€œghostsâ€•were to be 485

prepared, the washed cells were centrifuged again 370 C. in a Dubnoff shaker for 1 hour. Following at 1,700 X g for 45 minutes, and the supernatant incubation, the contents were centrifuged, and the fluid was removed. The ghosts were prepared from supernatants were stored at 4Â°C. until the con 1- or 92-mi. aliquots of the packed red cells by the centration of sialic acid was determined. All exper method of Tishkoff et al. (920). Tumor cells were iments were carried out on triplicate aliquots of obtained 7â€”8days after the intraperitoneal injec cells. tion of 3-month-old female Swiss mice with 0.3 ml. The total sialic acid in mouse red cells was deter of a suspension of a hyperdiploid EAT. The tumor mined on aliquots of ghosts equivalent to 1 or 92 had been previously propagated at 7-day intervals. ml. of packed erythrocytes. Three ml. of 0.1 N Any ascitic fluid grossly contaminated with red H2SO4was added to 1 ml. of ghosts, and the con cells was discarded. The EAT cells were centrifuged tents were then hydrolyzed at 80Â°â€”85Â°C.for 1 at 100 X g at 4Â°C., and the supernatant plas hour. After hydrolysis the ghosts were centrifuged, ma and the few red cells at the top of the EAT cell and the supernatant, together with two 3-mi. 0.1 column were removed and discarded. The cells N H2SO4 washings, were saved for sialic acid meas were then washed S times with 4 volumes of urement. In preliminary studies it was found that Krebs-Ringer bicarbonate, pH 7.4, containing 11 1 hour of hydrolysis was not sufficient for maxi mmoles glucose. At the end of the last wash the mum recovery of red cell sialic acid. Therefore, a cells were returned to their original volume with second 1-hour period of hydrolysis was routinely buffer. performed, the supernatants of both hydrolysis Sialic acid determination.â€”Sialic acid was periods being combined for sialic acid determina measured by the method of Warren (9292),except tion. that 0.5-nd. aliquots of sample were used with a EAT cell homogenates were prepared by freeze commensurate doubling of the volume of periodate thawing whole cells S times. The measurement of added. The reaction product followed Beer-Lam total sialic acid on these homogenates by acid hy bert's law within the limits defined by Warren. drolysis was not feasible, owing to the presence of Crystalline N-acetyl neuraminic acid' (NANA) interfering substances in the sialic acid determina was used as a reference standard with each sialic tions. This problem was also noted by Wallach assay. No correction was made for the N-glycolyl and Eylar (921). The interfering substances could neuraminic acid in EAT cells or for the trace not be eliminated by resin column chromatogra amount of an unknown sialic acid found in eryth phy on Dowex 50W-X8 and Dowex 92-X8 (19) or rocytes. by extraction with a number of organic solvents. In a typical experiment (a) total sialic acid, The use of the optical density readings at 5392and (b) sialic acid released by neuraminidase, and 5692 mjs instead of the 5392 and 549 ms readings as (c) sialic acid released by trypsin were determined. suggested by Warren (@92)and adopted by Wallach All glassware was siliconized. One-half ml. of the (921) did not materially improve the accuracy of packed erythrocytes and 1 ml. of EAT cell suspen the results. When cell homogenates were incubated sion were pipetted into a 50-mi. Erlenmeyer flask. with neuraminidase at 37Â°C. for 1â€”92hours,mate The red cell pipette was rinsed with three 0.5-mi. rials absorbing light at 5392m@ were not released. aliquots of 0.15 M NaCl to insure quantitative de Therefore, only the sialic acid released by neura livery of the packed erythrocytes. Neuraminidase minidase treatment of cell homogenates@could be prepared from Cholera vibrio2 (0.5- to 3.0-ml. ali measured. For the determination, homogenates quots) or trypsin8 three times recrystallized (1â€”8 from I ml. of EAT cells were mixed with 1 ml. of mg.), dissolved in 1 ml. of Krebs-Ringer phos Krebs-bicarbonate buffer pH 7.4 and 1 ml. of the phate, pH 7.6, was added to cell suspensions. Con neuraminidase solution. After incubation for 1 troi flasks without added enzyme were included in hour at 37Â°C., the mixture was centrifuged, and each experiment. The cells were incubated at the free sialic acid was determined on the super natant. In the estimation of total volume of super 1 Obtained from Dr. E. Eylar. Additional material obtained from General Biochemicals Inc. was also found to be satisfac natant, 80 per cent of the packed EAT cell mass tory. was assumed to be due to intracellular water. 2 Behringwerke, West Germany. Enzyme was dissolved in Duplicate 0.1-mi. aliquots of packed red cells 0.15 M NaC1 solution containing 9 m@sCaCl, and 0.05 M Na were diluted with 100 ml. of isotonic saline. The acetate buffer, pH 5.5. One ml. of the enzyme preparation con EAT cells were diluted with S per cent acetic acid. tamed 100 units of enzyme activity, one unit being equal to the Both were counted in a hemacytometer. Hemato amount of enzyme that would liberate 1 @g.ofsialic acid from a glycoprotein substrate during a 15-minute incubation period crits and tumor cell volumes were measured in at 370 C. duplicate. No correction was made for the volume

3 Nutritional Biochemicals and Sigma Chemical Co. of trapped fluid in either of the cell types. The cal

culated mean corpuscular volumes (MCV) were (b) sec-butanol-acetone-acetic acid-water (6:6: @ found to be 54 cu. for the erythrocyte and 1800 3 :5), and (c) n-butanol-n-propanol-0.1 N HC1 Cu. @sforthe EAT cell. According to a modification (1 :92:1). Standards were applied to each paper of the theorem of Pappus (16) and a value of 6.7 z strip. In addition to the crystalline material, NANA for cell diameter (10) , the erythrocyte membrane was prepared from human erythrocytes and area was estimated as 85 sq. @.The surface area of plasma (19). An N-glycolylneuraminic acid stand the EAT cell was calculated as 780 sq. j.@,assuming ard was prepared from bovine plasma (15). After the cell to be spherical. chromatography, the strips were dried and then Chromatopaphic ident@Ca@iOnof sialic acid sub sprayed with an orcinol-trichloroacetic reagent stances.â€”Red cell ghosts from 15 to 920 ml. of (13) to locate the sialic acid spots. packed human or mouse red cells were hydrolyzed with 0.1 N H@SO4at 80Â°C. as previously described. RESULTS The ghosts were then spun down, and the super Mozwe erythrocytes.â€”Acid hydrolysis of the natant was neutralized with 0.1 @iBa(OH)s. After mouse erythrocyte ghosts yielded 0.3924 Â±0.056 removal of the BaSO4 precipitate, the supernatant @moleof sialic acid/mi of packed cells or 0.019 Â± was passed through a cation exchange resin Dowex 0.004 j@mole/109 cells. Neuraminidase treatment of 50 W-X8, 50-100 mesh in the Tris (hydroxymeth intact erythrocytes liberated only 69 per cent ylaminomethane) cycle, so arranged that the con (mean of eight determinations) of the amount of

TABLE 1 SiAuc AcID RELEASEDBYNETThtAMINIDASEANDTRYPSINTREATMENTOFiNTACTMOUSE ERYTaaoCYT@s AND Em@ucn Ascn@s TUMORCELLS

OF CELLI'@,o. IDASE/ [email protected]&ssTa@ssrsNm@uim Taypsix@imo1e/m1*pmole/1O'f@imo1e/m1g@moIe/1O'Erythrocyte

Ascitestumorcell8 70.217(Â±0.019)O.277(Â±0.045)0.01S(Â±O.0O1)0.466(Â±0.082)0.248(Â±0.055)0.043(Â±0.021)0.015(Â±0.004)0.072(Â±0.027)0.88 6.44

* Per ml. = per mL of packed cells. t Per 10@=per 10' cells. Column 4 represents the ratio of neuraminidase to trypsin release of sialic acid calculated for per ml. values only.

tents would drip onto an anion exchange resin, sialic acid obtained from the hydrolyzed ghosts Dowex 92-X8, 50â€”100mesh in the acetate cycle (Table 1). The amount of sialic acid reieased could (19). The sialic acid was eluted from the latter col not be increased by using higher concentrations of umit with 45 ml. of 1 M ammomum acetate, pH the enzyme or increasing the time of incubation. 4.5. The eluted material was lyophilized and re It was of interest that, in contrast to human red dissolved in 92â€”Sml.of water. cells, the mouse erythrocyte did not agglutinate Purification of EAT cell sialic acid as described when suspended in homologous plasma following above yielded impurities in the final eiuate which the neuraminidase treatment. prevented resolution of sialic acid on paper chro Treatment of the erythrocytes with trypsin matograms. However, it was found that prelimi liberated a substance which did not react directly nary dialysis of the homogenates removed much in the Warren procedure but which did following of the interfering substances. Therefore, EAT cell acid-hydrolysis. The fragment split off by trypsin homogenates (equivalent to 10-15 ml. of centri was undialyzable (Visking tubing) but was not fuged cells) were dialyzed for 924hours at 40@60C. precipitated by 10 per cent trichloroacetic acid. against 1800 ml. of water, following which the The amount of sialic acid recovered from this frag homogenate was incubated with S ml. of neuramin ment was slightly greater than that obtained by idase at 87Â°C. for S hours. The supernatant was neuramiuidase treatment (Table 1). then applied to the resin columns as described. As Chromatographic analysis of the free sialic acids cending paper chromatography (Schleicher and obtained from the erythrocytes in three different Schuell paper No. 92043-B) was performed on eryth solvent systems is shown in Table 92.Two distinct rocyte and EAT cell eluates in the following spots were obtained with an orcinol spray. Sub solvents : (a) n-butanoi-acetic acid-water (4:1:5), stance A, which when eluted was found to contain

96 per cent of the total Warren-reaction chromo that had been pre-incubated with the EAT cells gen, had an R@ similar to NANA in all solvents. was found to produce as much sialic acid-contain The other spot (Substance X) moved more rapidly ing substance as did that treated with fresh tryp than both NANA and N-glycolyneuraminic acid sin, indicating that inactivation of the former had in all three systems but was not further identified. not occurred. Ehrlids ascites tumor cells.â€”Measurement of The chromatograms done on the sialic acids re total sialic acid from acid-hydrolyzed EAT cells leased from EAT cells demonstrated that 70 per was not possible owing to technical difficulties (see cent moved with an RF corresponding to NANA â€œMethodsâ€•).The sialic acid released from the cell in the three solvent systems employed (Table 92). homogenates by neuraminidase was equal to The other 30 per cent (Substance B, Table 92),be 0.9287 Â±0.037 @moles/ml of packed cells or haved in the same fashion as the N-glycolylneura 0.485 Â±0.100i/10@ cells. Only slightly less was re minic acid standards. coyered when the intact EAT cells were exposed to the neuraminidase (992per cent of that obtained DISCUSSION from the homogenates, see Table 1). Thus, the The amount of siaiic acid released by neuramin quantity of sialic acid released from the EAT cells idase from the EAT intact cell or homogenate was by neuraminidase was some 86 times greater than far in excess of both the total and neuraminidase

TABLE 2 THE R@'s OF SIALIC ACID SUBSTANCES ISOLATED FROM MOUSE ERYTH ROCYTES AND EmtLICH ASCITES TUMOR CELLs SEPARATED IN TmtEE DIFFERENT SOLVENT SYsTEMs

SYSTSUSec-butanol-acetone Si.&uc @cm SUBSTANCSSOLVENT Butanol-acetic propanolN.acetylneuraminic acetic acid-water. acid-waterHCI-butanol

acid .45-0.48 . 18-0 .20

N-glycolylneuraminic acid 0 .39-0 .41 0 . 16-0 . 17 0.80-0.32

Erythrocyte: SubstanceA 0.47-0.48 0.18-0.19 0.39-0.40 Substance X 0 .52-0 .53 0 .23 0.60

Ascites tumor cell: Substance A 0 .46â€”0.48 0 .18â€”0.20 0.38-0.40 Substance B0 0 .39-0 .410 0 . 16-0 . 170.38-0.40 0.30-0.32

that released from the erythrocytes on a per cell liberated sialic acid of the erythrocyte. The differ basis. ences in total sialic acid may be even greater if the Tryptic digestion of the EAT cells produced a EAT cells contain gangliosides (11) or nucleotide nondialyzabie fragment which, as in the case of the sialic acid compounds (3) , substances whose sialic erythrocytes, did not react directly in the Warren acid may not have been released by the neuramini procedure unless first acid-hydrolyzed. However, dase treatment of cell homogenates. The increased by contrast with the red cells, the trypsin treat siaiic acid found in the EAT cell could merely be ment of EAT cells yielded far less sialic acid than due to the greater membrane area of the much did the neuraminidase (Table 1). larger EAT cell. However, when these values are The possibility that the trypsin in the EAT cell expressed in these terms (using the estimated areas experiments was inhibited or inactivated by some of 780 sq. Jhfor the EAT cell and 85 sq. js for the cellular product was considered. To determine erythrocyte), the siaiic acid released by neuramini whether this were the case, the supernatants were dase from the intact EAT cell was still 4 times removed following the EAT cell digestion. The greater than that liberated from the erythrocyte. supernatant fluid was then incubated with an ali Errors incurred in the counting of the cells, in the quot of red cells, and the sialic acid released com estimation of the membrane area of both cell pared with that obtained with a similar aliquot of types, and differences in the amount of trapped red cells incubated with fresh trypsin. The trypsin extraceliular fluid in the hematocrit and tumor cell

volume measurements probably do not apprecia since it was not precipitated by trichloroacetic bly alter the above values. acid and the acid-soluble fraction could be separat Despite the much greater density of negatively ed by paper chromatography. The trypsinized, charged sialic acid groups on the EAT cell mem nondialyzable fragment of mouse erythrocytes was brane, the isoelectric point of this ceil has been re also not precipitated by trichloroacetic acid, but ported as about pH 4 (18) in contrast to that of this is compatible with the properties of a glyco about pH 92found for the mammalian erythrocyte protein. (1, 8). In this regard, it is pertinent to note that In view of the similar amount of sialic acid lib 38 per cent of the total erythrocyte sialic acid erated by both neuraminidase and trypsin from could not be liberated by neuraminidase. A similar mouse erythrocytes, it seems reasonable to sup situation has been noted for the horse erythrocyte pose that the sialic acid liberated by neuramini (7). It is possible that this siaiic acid, possibly dase is an end group on a large membrane protein, linked to a ganglioside (11), may contribute to the one of whose peptide bonds can be split by trypsin. negative charge of the erythrocyte. However, this In sharp contrast to the erythrocyte, trypsin treat fraction of sialic acid represents a relatively small ment of EAT cells only split off about one-sixth as number of negatively charged groups and could much sialic acid liberated by neuraminidase. not account for the observed findings. It has been These results suggest that most of the sialic acid observed that the anionic polymer, polyxenylphos of the EAT cell is located on proteins or sub phate, is bound to the EAT cell, presumably by stances which do not contain trypsin-sensitive positively charged groups on the membrane (18). sites. The possibility that such sites are present, Since the isoelectric point is determined primarily but are protected from enzyme attack by neigh by the net charge on the cell surface, variability boring molecular groups, cannot be excluded. in the concentration of cationic groups may account for the fact that the reported isoeiectric point in REFERENCES EAT cells is higher than the mammalian erythro 1. BANGRAM,A. D.; PETHICA,B. A.; and Sz@ai@N,G. V. F. cyte. The Charged Groups at the Interface of Some Blood Cells. Another consideration is the possibility, raised Biochem.J.,69:12â€”19,1958. 2. BENNETT, H. S. The Concepts of Membrane Flow and by Waliach and Eylar (921), that not all the sialic Membrane Vesiculation as Mechanisms for Active Trans.. acid liberated by neuraminidase is situated on the port and Ion Pumping. J. Biophys. Biochem. Cytol., 2: surface of the cell. The amount of sialic acid so re 99â€”103,1956. leased in their study was similar to that found by 3. COMB, D. G.; SHIMIZU,F.; and ROSEMAN,S. Isolation of Cytidine-5'-monophospho-N-Acetylneuraminic Acid. J. us and was greatly in excess of that which could be Am. Chem. Soc.,81:5513â€”14,1959. accounted for on the basis of the change in elec 4. CooK, G. M. W.; HEARD, D. H.; and Snss@u, G. V. F. A trophoretic mobility. In view of the proposed dy Sialomucopeptide Liberated by Trypsin from the Human namic continuity of the surface with the ergasto Erythrocyte. Nature, 188: 1011â€”12,1960. plasmic membranes of the cell (92),Eylar and Wai 5. . Sialic Acids and the Electrokinetic Charge of the Human Erythrocyte. Nature, 191:44â€”49,1961. lach suggested that a considerable pool of sialic 6. EYLAR, E. H., and MADOFF,M. A. Isolation ofa Glycopep acid might be present in the interior of the cell at tide from the Red Cell Membrane. Federation Proc., 21: any one time which, although not contributing to 402, 1962. the surface charge, would ultimately be exposed 7. Em.R, E. H. ; MADOFF, M. A. ; BRODY, 0. V. ; and ONCLEY, to the action of the enzyme. J. L. The Contribution of Sialic Acid to the Surface Charge of the Erythrocyte. J. Biol. Chem., 237:1992â€”2000, In contrast to neuraminidase, the sialic acid re 1962. leased by trypsin was not free but was part of a 8. FURCEGOTT, R. F., and PONDER, E. Electrophoretic nondialyzable fragment. Similar observations were Studies on Human Red Blood Cells. J. Gen. Physiol., 24: made by MÃ¤kela and co-workers (14) following 447â€”57,1941. 9. GOTTSCHALK,A. The Influenza Virus Neuraminidase. Na trypsin treatment of human red cells. On the other ture, 181:377â€”78, 1958. hand, Eylar and Madoff found the trypsinized 10. GULLIVER, G. Observations on the Sizes and Shapes of the fragment to be a glycopeptide with a minimum Red Corpuscles of the Blood of Vertebrates with Drawings molecular weight of 4,000â€”5,000 (6). A highly of Them to a Uniform Scale and Extended and Revised asymmetrical molecule of this size might not Tables of Measurements. Proc. Zool. Soc., London, pp. 474â€”95,1875. dialyze through Visking tubing. However, owing 11. Ki@u, E. Neuraminic Acid. In: G. E. W. WOLSTEN to the differences in methods of preparation, it is HOLME and M. O'CONNOR (eds.), Symposium on Chemis also possible that the substance reported by these try and Biology of Mucopolysaccharides, pp. 296â€”802. authors represents a group detached from a larger Boston: Little, Brown & Co., 1958. 12. KLEux, E., and UHLENBRUCK, G. tYber em Neuramin parent substance. Cook et at. (4) also concluded saurehaltiges Mucoproteid am Rindererythrocytenstroma. that the trypsinized fragment was a small peptide, Ztschr. physiol. Chem., 311:227â€”83,1958.

13. KLEVSTRAND,R.,and Noansi@,A. A Spraying Reagent for Cells and the Effect of Polyxenylphosphate. Cancer Bee., Paper Chromatograms Which Is Apparently Specific for 19:913â€”17,1959. Ketoheptoses. Acts Chem. Scandinav., 4:1320, 1950. 19. Svmixnmioui, L. On the Isolation and Characterization 14. MX@xi@,0.; MIrnINEN, T.; and Pxsoi@, R. Release of of N-Acetylsialic Acid. Acta Soc. Med. Upsalien, 61:75- Sialic Acid and Carbohydrates from Human Red Cells by 85, 1956. Trypsin Treatment. Vox Sang., 5:492â€”96,1960. 20. Tis.uxoz'i', G. H.; ROBSCHEITT-ROBBINS,F. S.; ai@i 15. M.ARTENSSON,E.;R@u, A.; and Svisseramou&, L. Sialic Wiswpzn, G. H. Bed Cell Stroma in Dogs. Variations in Acids in Blood Serum. Biochem. Biophys. Acta, 30:124- the Stroma Protein and Upid Fractions Belated to Experi 29,1958. mental Conditions.Blood,8:459-68, 1953. 16. PONDER,E. Red Blood Cell Dimensions: Measurement. 21. Wau@acii,D. F. H., and En@sa, E. H. Sialic Acid in the in: 0. GLAssER (ed.), Medical Physics, 1:1208-14. Chica Cellular Membranes of Ehrlicli Ascites-Careinoma Cells. go: The Year Book Publishers, Inc., 1944. Biochim. Biophys. Ada, 52:594-96, 1961. 17. Roas.at&N, S., Jouiwuii, G. W. ; WA@rsoN, D. ; and RooD, 22. W@tnnzri,L. The Thiobarbituric Acid Assay of Sialic R. Enzymatic Synthesis of Sialic Acid 9-Phosphates. Proc. Acids. J. Biol. c:hem., 234:1971â€”75,1959. Natl. Acad. Sd., 47:958-61, 1961. 23. Wummi, L., and Fziamirzi@, H. The Biosynthesis of 18. STRAUMPJ0RD,J. V., JR., and Huararai@,J. P. Anionic N-AcetylneuraminicAcid. Biochem. Biophys. Bee. Comm., Polymers. IV. Microelectrophoresis of Ascites Tumor 4:232â€”85,1961.

Aaron Miller, John F. Sullivan and Jay H. Katz

Cancer Res 1963;23:485-490.

Updated version Access the most recent version of this article at: http://cancerres.aacrjournals.org/content/23/3/485

E-mail alerts Sign up to receive free email-alerts related to this article or journal.

Reprints and To order reprints of this article or to subscribe to the journal, contact the AACR Publications Subscriptions Department at [email protected].

Permissions To request permission to re-use all or part of this article, use this link http://cancerres.aacrjournals.org/content/23/3/485. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.