Surf’ace Alterations in Calf’ Lymphocytes Oxidized by Sodium Periodate*

In order to investigate alterations in surface structure in transformed lymphocytes, calf submandibular lymph node cell suspensions were oxidized with NaIO,. Oxidized lymphocytes were morphologically transformed and had higher rates of DNA synthesis by 2 days after treatment. These results were prevented by reduction of the cell suspension with h’aBH,, or by neuraminidase treatment of cells prior to oxidation. The amount of ‘251-labeled Agaricus bisporus lectin bound to cells immediately after oxidation and the affinity constant for binding were increased over 2 -fold, while cells immediately following oxidation and reduction showed decreased receptors with still higher affinity for the lectin compared to untreated cells. The amount of Phoseolus oulgaris lectin bound to oxidized cells was also increased, but affinity was unchanged. Immediately following oxidation and reduction, these receptor sites were unchanged in number and affinity from untreated cells. In contrast, the number and affinity of receptors for concanavalin A were not changed immediately after oxidation or oxidation and reduction. In order to define the extent of compositional changes in surface glycoprotein receptors,

Binding Assays-Lectins were iodinated with '*? by the method of Hunter (8) employing a 5-s exposure to chloramine T. The ability of the lectins to agglutinate cells was unchanged after this procedure. Over 90% of the radioactivity in each lectin could he bound to an excess of cells.
Binding assays were performed as previously described (7). Plastic tubes were incubated for 60 min with 1 ml of 5 mg/ml of bovine serum albumin in 0 N-Acetylneuraminic acid was used as the standard. Galactose and mannose were measured enzymatically following sample hydrolysis in 2 N H,SO, for 4 hours at 100" and deionization on Amberlite MB3, as described by Kornfeld et al. (15). Glucosamine and galactosamine were quantitated on a Spinco automatic amino acid analyzer following hydrolysis in 4 N HCI for 4 hours at 100" i n uacuo. Total hexosamine was determined after similar hydrolysis and adsorption to and elution from Dowex 50, by the method of Reissig et al. (16). Fucose was assayed by the cysteine-H,SO, method (17). 5'-Nucleotidase was measured by the method of Weaver and Boyle (18) as a marker for plasma membrane. @-N-Acetylglucosaminidase was measured as described (19)  rnM Tris/l mM EDTA, pH 7.5, and was homogenized for 10 s in a Waring Blendor.
The homogenate was filtered through cheesecloth and the filtrate was centrifuged at 980 x g for 10 min to remove nuclei and intact cells. The supernatant solution was brought to 0.1 M LiCl by addition of 3 M LiCl. and was then centrifuged at 4000 x g for 15 min. The pellet contained mitochondria and was discarded.
The supernatant solution was centrifuged at 50.000 x g at 4' for 1 hour; the 50,000 x g supernatant solution containing rihosomes was discarded. and the 50.000 x g pellet was dispersed in sucrose/l0 mM Tris. pH 7.5, to a.f'inal concentration of 40', sucrose (w/w) hy homogenization in a 50.ml homogenizer (Kontes Glass Co.). This was distributed into cellulose nitrate tuhes (18 ml/tube) and carefully layered over with 30'~ (w/w) sucrose in 10 mM Tris, pH 7.5 (18.5 ml/tube).
The tubes were centrifuged for 4 hours at 7:1,000 \. g in a Spinco model 30 rotor in a Beckman ultracentrifuge. Four fractions were obtained by aspiration: a cloudy supernatant solution ("upper"). a dense milky band at the interface ("middle"). a clear solution below the interface ("bottom"). and a pellet. These fractions were each centrifuged at 50,000 x g at 4" for 1 hour, and the pellets were washed twice with 10 rnM Tris, pH 7.5. and stored frozen until use.
Fractions thus ohtained were assayed for marker enzymes, DSA. protein, and sialic acid (

Characterization of Calf Lymphocyte
System-Cell suspensions were made f'rom fresh calf submandibular lymph nodes. Untreated cell suspensions contained 875 small, 9.5% medium, and 2% large lymphocytes, no lymphoblasts, and 1.5% After oxidation in NaIO,, the cell suspension consisted of 82% small, 13% medium, and 2% large Iymphocytes, no lymphoblasts, and 3% monocytes. The cells were 99% viable both before and after oxidation as determined by trypan blue exclusion, and 88% viable after NaBH, reduction. After 2 days in short term culture, lymphoblasts accounted for 2.1% + 0.6% (1 S.E.) of cells in the untreated cell suspension, 21.3% + 1.8% of cells in the oxidized cell suspension, and 4.8% l 1.0% of cells in the oxidized and reduced cell suspension.
DNA synthesis was measured by tritiated thymidine incorporation into cultures of calf lymphocytes (Table II). Addition to the cultures of an optimal amount of Phaseolus uulgaris E-PHA, 15 fig/ml, or NaIO, oxidation of the cells resulted in an increased rate of DNA synthesis which was maximal at 48 hours. When NaIO, treatment was followed by NaBH, reduction, the rate of DNA synthesis was equal to that in untreated cells. NaBH, treatment alone resulted in the same rate of DNA synthesis as in untreated cells.
Lectin Binding to Lymphocytes-Untreated, oxidized, and oxidized and reduced calf lymphocyte suspensions immediately after preparation were tested for their ability to bind 'Y-mushroom PHA-B. One experiment is illustrated (Fig. 1). Oxidized cells bound more radioactive lectin than untreated cells. Oxidation plus reduction did not increase the lectin binding to cells. Although increased affinity of the lectin for the cells was more evident in the replicate experiments, apparent affinity of' the lectin for the cells was slightly increased after oxidation, and further increased after oxidation plus reduction in the experiment illustrated (Fig. 1). Scatchard analyses of each of the three curves in this and other experiments with mushroom PHA-B indicated only one class of receptor (data not shown). This is also apparent in the double reciprocal plot (Fig. 1, right). A similar binding study using P. vulgaris E-""I-PHA also suggested increased lectin binding to oxidized cells (Fig. 2). The binding of lectin to cells that had been oxidized and reduced was similar to that with untreated cells. The slight difference in association constant suggested in Fig. 2, right panel, was not a consistent finding. Scatchard analyses demonstrated only one class of P. vulgaris E-PHA receptor for each of the three cell suspensions. Data from four such P. vulgaris E-PHA binding experiments (Table III) indicated a 1.8-fold increase in available binding sites on lymphocytes following oxidation (p < 0.05). Although cells that had been oxidized and reduced had an average of one-third less receptor sites than untreated cells, this was not statistically significant (p > 0.4). Association constant determinations were more variable, and the average 2-fold difference after oxidation and a l.'i-fold decrease after oxidation plus reduction were not statistically different from untreated cells (p > 0.4).

Results
In contrast to the results with mushroom PHA-B and P. vulgaris E-PHA, the binding ot 1251-Con A to lymphocytes was slightly decreased after either oxidation, or oxidation plus reduction ( Fig. 3 and Table III). The differences were small, however, and were not statistically significant (p > 0.4) due to greater variability in quantitative results. Affinity of Con A for the receptor site was similar in the three cell populations (p > 0.4). Scatchard analyses again demonstrated only one class of receptor.  (Lowry) was added 5 ml of 3.7 x lo-' M NaI0,/0.005 M sodium phosphate, pH 7.4/0.15 M NaCl for 15 min at 25" in light, and the membranes were then centrifuged and washed with the same buffer without NaIO,. The membranes were then hydrolyzed in conditions appropriate for the assay of amino acids or sugars as described under "Experimental Procedure." Sugar composition of the untreated and oxidized plasma membranes was similar except for sialic acid (Table IV). The thiobarbituric acid assay revealed only 69% of the original activity in oxidized membranes.
Since mild oxidation of N-acetylneuraminic acid yields a product NeuNAc-7-ald which reacts about half as well in the thiobarbituric acid assay (21), the results suggested incomplete oxidation (about 60%) of sialic acid in lymphocyte plasma membranes to aldehydes. Contents of other sugars were not significantly different in oxidized membranes, and there was no loss of fucose content with oxidation.
Amino acid analysis showed no significant degradation of amino acids by oxidation (Table V). In particular, amino acids sensitive to non-Malapradian oxidation, threonine, serine, proline, cystine, and methionine, were unchanged within experimental error. Sialic Acid Chromatography-Since the only apparent change in oxidized membranes was a decrease in thiobarbituric acid-assayable material, chromatography was performed to more precisely determine quantitative and qualitative aspects of the oxidation. Normal membranes, oxidized membranes, and oxidized membranes reduced by NaBaH, were each hydrolyzed in dilute acid. The hydrolysate after centrifugation was subjected to descending paper chromatography in various solvents (   (Table VII). This enzyme was active, as measured by its ability to remove 0.79 + 0.1 rmol of sialic acid (measured as N-acetylneuraminic acid) from 10" calf lymphocytes (mean * 1 SE. for four experiments) or 0.63 rmol of sialic acid from lOLo human erythrocytes (one experiment) in 90 min at 37".' Sonicated calf lymphocytes hydrolyzed in 0.05 N H,SO, at 80" for 1 hour contained 1.2 pmol of free sialic acid per 10" ceils. Thus, neuraminidase was able to remove 66% of the sialic acid. Lymphocytes which had been treated with neuraminidase prior to NaIO, oxidation exhibited slightly decreased rates of DNA synthesis compared to untreated cells. Neuraminidase treatment did not decrease the stimulation of DNA synthesis by P. uulgaris, indicating that its effect on NaIO,-oxidized cells was not a consequence of general inhibition of DNA synthesis. Thus, surface sialic acids were specifically required for the periodate-induced stimulation of DNA synthesis in calf lymphocytes. Lectin binding to available glycoprotein receptor sites was used as a probe for structural alterations in the surface of oxidized lymphocytes. Induction of heterogeneity in the binding sites, altered affinity of the lectin for the cell surface receptors, or changes in the number of receptors for the lectin each would reflect a structural change. The lectins studied were chosen because of their different glycopeptide specificities. P. uulgaris E-PHA binds to glycopeptides with N-glycosidically linked oligosaccharides containing galactose -N-acetylglucosamine residues, especially if mannose is present in the "core" of the oligosaccharide chain (25). Mushroom PHA-B binds to glycopeptides with O-glycosiditally linked galactose -N-acetylgalactosamine residues (7). Con A binds to glycopeptides containing mannose in an (Y configuration.
Because each of the lectins used is a multivalent molecule, it is likely that the numbers of binding sites calculated from the binding studies underestimates the actual number of receptor sites.
In limited studies of lectin binding at 4", the amount of mushroom PHA-B bound to lymphocytes at 4" was only 37% of that at 25", and the amount of P. uulgaris E-PHA bound at 4" was only 76% of that at 25". Thus, lectin binding at 25", used in all experiments reported under "Results," is probably greater than if similar experiments were performed at O-4", temperatures below the temperature-dependent phase transition proposed to occur in eukaryotic cell membranes (26). The higher temperature was chosen so that the study might also reflect conformational changes in lectin receptors occurring only in more fluid plasma membranes. Binding was not studied after NaBH, reduction of otherwise untreated lymphocytes since NaBH, reduction alone did not alter the rate of DNA synthesis. Although the cell suspensions studied consisted nearly entirely of small and medium lymphocytes, subpopulations of lymphocytes (T cells, B cells, or others) were not evaluated. Scatchard analyses and double reciprocal plots of the binding data both indicated only one class of receptor for each lectin. This suggests that receptors on various subpopulations of cells were identical, although the possibility remains that different lymphocyte subpopulations had different numbers of receptors. It is unlikely that leakage of any intracellular molecules modified the lectin binding to intact cells in these experiments, since all cells were washed in medium prior to the binding assay, since nearly all cells were viable, and since the binding assays were performed immediately after oxidation or oxidation and reduction.
Results of the binding studies documented a significant increase in numbers of P. uulgaris and mushroom PHA-B receptor sites on oxidized lymphocytes. The oxidized receptors appeared to be homogeneous by Scatchard analyses suggesting that the additional receptor sites were glycoproteins with oligosaccharides similar in structure to those of unoxidized lymphocytes. This is compatible with a rearrangement in surface molecules permitting additional lectin to bind to glycoprotein receptors of identical or similar structure, and does not imply production of new receptor sites. New synthesis of additional receptor sites is also unlikely, because of the brief duration of NaIO, exposure prior to performing the binding study.
It is of interest that the apparent affinity of P. oulgaris E-PHA for nonoxidized and oxidized lymphocyte receptors was equal. This is consistent with observations on P. uulgaris E-PHA receptors of human erythrocyte glycopeptides, which bind P. uulgaris E-PHA equally well before and after neuraminidase treatment (7). In contrast, the apparent affinity of mushroom PHA-B for oxidized calf lymphocyte receptors was increased compared to receptors on untreated cells. This conformational change in oxidized lymphocyte receptors may be related to oxidation of membrane sialic acids, since mushroom PHA-B has been shown to.have a much greater affinity for erythrocyte glycopeptides after removal of sialic acid (7). Con A binding to lymphocytes was not significantly changed by oxidation, or oxidation plus reduction, indicating that the changes in mushroom PHA-B and P. uulgaris E-PHA binding were related to specific membrane alterations. Possible explanations for the differences are: (a) Con A binds to a surface structure distinct from P. uulgaris E-PHA and mushroom PHA-B on calf lymphocytes; (b) undetected surface conformational changes are insufficient to permit more Con A molecules to bind; and (c) maximal amounts of Con A are bound even to unoxidized lymphocytes. Further evidence that alterations of receptor sites by oxidation are specific is that NaBH,-reduction of lymphocytes reduced the number of binding sites for both P. uulgaris E-PHA and mushroom PHA-B to levels equal to or below those of untreated cells, but did not change the binding characteristics of Con A. Analysis of the composition of oxidized plasma membranes revealed only a change in sialic acids. This suggests that only mild Malapradian oxidation occurred under the conditions of periodate used in these experiments, and that more severe Malapradian oxidation of hexoses and hexosamines within the oligosaccharide chains of glycoproteins or glycolipids did not occur. However, since analyses were performed on crude membrane preparations containing many different glycoproteins and glycolipids, the techniques used might not detect oxidation of hexose, hexosamine, or amino acid residues accounting for less than 10% of the total. In addition, these experiments did not study whether structural changes occurred in glycolipid fractions as well. However, it is unlikely that oxidation of glycolipids would result in altered binding of the lectins studied, since purified mammalian cell membrane receptors for these lectins are glycoproteins (7,25,27). The lack of major composition changes in oligosaccharide content of the oxidized plasma membranes in this study suggests, therefore, that the altered lectin binding reflects a conformational change in glycoproteins due to oxidation of sialic acids.
The nature of the oxidation of sialic acids was incompletely studied in the experiments reported. Although there is good evidence for complete conversion of N-acetylneuraminic acid to NeuNAc-7-ald by periodate in the calf lymphocytes (as in sialoglycopeptides and erythrocytes (21, 28, 29)), the result of N-glycolylneuraminic acid oxidation is not characterized, since the oxidized products of N-glycolylneuraminic acid cochromatographed with unaltered N-glycolylneuraminic acid. Studies are currently in progress to identify the oxidation products of N-glycolylneuraminic acid and of calf lymphocyte membranes by gas chromatography. This is important since 93% of the incorporation of NaB3H, was into N-glycolylneuraminic acid derivatives and since N-glycolylneuraminic acid accounts for 53% (30) to 63% (31) of bovine fibrinogen or bovine serum glycoproteins.
Neuraminidase treatment of lymphocytes prevented transformation by subsequent periodate oxidation. This suggests that certain sialic acid-containing plasma membrane molecules are necessary for NaIO,-induced transformation of calf lymphocytes. Intracellular oxidation by NaIO,, although it occurs,* is not capable of initiating and sustaining lymphocyte transformation without the presence of surface sialic acid.