Biomedical subjects
L D Powell
Publications and source records attributed to L D Powell.
Characterization of sialyloligosaccharide binding by recombinant soluble and native cell-associated CD22. Evidence for a minimal structural recognition motif and the potential importance of multisite binding.
CD22, a B cell-specific receptor of the immunoglobulin superfamily, has been demonstrated to bind to oligosaccharides containing alpha 2-6-linked sialic acid (Sia) residues. Previously, we demonstrated that the minimal structure recognized by this lectin is the trisaccharide Sia alpha 2-6Gal beta 1-4GlcNAc, as found on N-linked, O-linked, or glycolipid structures (Powell, L., and Varki, A. (1994) J. Biol. Chem. 269, 10628-10636). Here we utilize a soluble immunoglobulin fusion construct (CD22Rg) to determine directly by equilibrium dialysis the stoichiometry (2:1) and dissociation constant (32 microM) for Neu5Ac alpha 2-6Gal beta 1-4Glc binding. Inhibition assays performed with over 30 different natural and synthetic sialylated and/or sulfated compounds are utilized to define in greater detail specific structural features involved in oligosaccharide-protein binding. Specifically, the critical features required for binding include the exocyclic hydroxylated side chain of the Sia residue and the alpha 2-6 linkage position to the underlying Gal unit. Surprisingly, alterations of the 2-, 3-, and 4-positions of the latter residue have limited effect on the binding. The nature of the residue to which the Gal is attached may affect binding. Bi(alpha 2-6)-sialylated biantennary oligosaccharides are capable of simultaneously interacting with both lectin sites present on the dimeric CD22Rg fusion construct, giving a marked improvement in binding over monosialylated compounds. Furthermore, data are presented indicating that full-length native CD22, expressed on the surface of Chinese hamster ovary cells, is structurally and functionally a multimeric protein, demonstrating a higher apparent affinity for multiply sialylated compounds over monosialylated compounds. These observations provide a mechanism for strong CD22-dependent cell adhesion despite the relatively low Kd for protein-sugar binding.
CD22-mediated cell adhesion to cytokine-activated human endothelial cells. Positive and negative regulation by alpha 2-6-sialylation of cellular glycoproteins.
We previously showed that cultured human umbilical vein endothelial cells (HEC) exposed to the inflammatory cytokines tumor necrosis factor-alpha or interleukin-1 display increased activity of beta-galactoside alpha 2,6-sialyltransferase. This is associated with enhanced expression of ligands for the B cell receptor CD22 beta, which recognizes alpha 2-6-linked sialic acids (Hanasaki, K., Varki, A., Stamenkovic, I., and Bevilacqua, M. P. (1994) J. Biol. Chem. 269, 10637-10643). Here we report that increased expression of CD22 ligands is a feature of dermal microvascular endothelial cells as well, and is also observed in response to the cytokine interleukin-4. Tumor necrosis factor-alpha stimulation of HEC causes no change in the profile of endothelial glycoproteins recognized by CD22, but doubles the proportion of total cellular N-linked oligosaccharides capable of binding tightly to CD22. This modest change is sufficient to cause a marked increase in alpha 2-6-linked sialic acid-dependent binding of Chinese hamster ovary (CHO) cells expressing recombinant human CD22. In contrast, B lymphoma cell lines expressing higher levels of cell surface CD22 do not show such sialic acid-dependent binding to activated HEC. Since B lymphoma cells themselves also express high levels of alpha 2-6-linked sialic acids, their CD22 molecules might be rendered nonfunctional by endogenous ligands. In support of this, the lectin function of CD22 can be directly detected on transfected CHO cells, but not on B lymphoma cells. Furthermore, coexpression of beta-galactoside alpha 2,6-sialyltransferase with CD22 in the CHO cells abrogates sialic acid-dependent binding to cytokine-activated HEC. However, such co-transfected cells can bind to B lymphoma cells in a manner apparently less dependent upon alpha 2-6-linked sialic acid, suggesting CD22-mediated interactions that may not be directly dependent on its lectin function. Thus, CD22-mediated interactions between B cells and activated vascular endothelium may be positively regulated by induction of alpha 2-6-linked sialic acid-bearing endothelial cell ligands, but negatively regulated by such ligands on the B cells expressing CD22. Since expression of both CD22 and beta-galactoside alpha 2,6-sialyltransferase are regulated during B cell ontogeny, these findings could be of importance in B cell function and/or trafficking.
Binding of human plasma sialoglycoproteins by the B cell-specific lectin CD22. Selective recognition of immunoglobulin M and haptoglobin.
CD22 is a cell-surface receptor of resting mature B cells that recognizes sialic acid (Sia) in the natural structure Sia alpha 2-6Gal beta 1-4GlcNAc (Powell, L. D., Jain, R. K., Matta, K. L., Sabesan, S., and Varki, A. (1995) J. Biol. Chem. 270, 7523-7532). Human umbilical vein endothelial cells (HEC) treated with inflammatory cytokines such as tumor necrosis factor-alpha (TNF-alpha) display increases in cell-surface CD22 ligands, caused by increased expression of the enzyme beta-galactoside alpha 2,6-sialyltransferase (Hanasaki, K., Varki, A., Stamenkovic, I., and Bevilacqua, M. P. (1994) J. Biol. Chem. 269, 10637-10643; Hanasaki, K., Varki, A., and Powell, L. D. (1995) J. Biol Chem. 270, 7533-7542). Thus, CD22 could direct potential interactions between mature B cells and endothelial cells during inflammatory states. However, this would have to occur in the presence of blood plasma, which contains many sialoglycoproteins known to carry alpha 2-6-linked sialic acids. We show here that human plasma can indeed inhibit Sia-dependent binding of a recombinant soluble chimeric form of human CD22 (CD22Rg) to TNF-alpha activated HEC. Affinity adsorption of individual human plasma samples with immobilized CD22Rg showed that, of the numerous alpha 2-6-sialic acid containing glycoproteins in plasma, only three polypeptides with apparent molecular mass (under reducing conditions) of 74, 44, and 25 kDa bound, and were specifically eluted with alpha 2-6-sialyllactose. NH2-terminal amino acid sequencing of these high affinity CD22 ligands revealed that they are subunits of immunoglobulin M (IgM) and haptoglobin. Purified human IgM from pooled human plasma can be quantitatively bound by CD22Rg, and binding is blocked by alpha 2-6-sialyllactose, but not by alpha 2-3-sialyllactose. Pretreatment by sialidase or by mild periodate oxidation of sialic acid side chains abolishes these interactions. IgM at physiological concentrations also inhibits CD22Rg binding to TNF-alpha-activated HEC in a manner dependent not only upon its sialylation but also requiring its intact multimeric structure. These data show that CD22 is capable of highly selective recognition of certain multimeric plasma sialoglycoproteins that carry alpha 2-6-linked sialic acids. Notably, the two proteins that are selectively recognized are known to be involved in immune and inflammatory responses. Haptoglobin synthesis by the liver is markedly increased during the "acute phase response" to systemic inflammation, while IgM is the major product resulting from activation of resting CD22-positive B cells.
The oligosaccharide binding specificities of CD22 beta, a sialic acid-specific lectin of B cells.
CD22 beta is a B cell surface glycoprotein involved in cell adhesion and activation. We previously reported that a recombinant soluble form termed CD22 beta Rg is capable of binding alpha 2-6 sialylated complex N-linked oligosaccharides purified from lymphocyte glycoprotein ligands (Powell, L. D., Sgroi, D., Sjoberg, E. R., Stamenkovic, I., and Varki, A. (1993) J. Biol. Chem. 268, 7019-7027). Here, we utilize a number of naturally and enzymatically sialylated oligosaccharides and sialoglycoproteins to further define its lectin specificity and demonstrate that the minimal structure recognized is Neu5Ac alpha 2-6Gal beta 1-4Glc(NAc). Reduction of the glucose residue of Neu5-Ac alpha 2-6Gal beta 1-4Glc diminishes the interaction, while truncation of the sialic acid side chain by mild periodate oxidation abolishes it. Branched oligosaccharides with two alpha 2-6-sialyl residues bind better, regardless of whether they were derived from N- or O-linked oligosaccharides or from gangliosides. alpha 2-3-Sialyl residues have no effect on binding, whereas increasing the number of alpha 2-6-sialyl residues on multiantennary oligosaccharides progressively improves binding. No specific feature of the core region affects binding, although the spacing of the alpha 2-6-sialyl residues on tetraantennary chains appears to have a significant effect. Of several model sialoglycoproteins examined, fetuin and transferrin had an apparent affinity no greater than that observed with free sialylated N-linked oligosaccharides. Some subfractions of these proteins displayed unexpectedly weak binding, suggesting that the protein backbone can exert a negative effect. In contrast, a subfraction of alpha 1-acid glycoprotein was identified as having a substantially higher apparent affinity than free oligosaccharides derived from it, indicating that multiple glycosylation sites may increase the apparent binding affinity. Thus, CD22 beta Rg contains a lectin activity specific for the minimal motif Neu5Ac alpha 2-6Gal beta 1-4Glc(NAc), and branched, multisialylated oligosaccharides are better ligands, regardless of the core sequences. Intact sialoglycoproteins can also interact, although with a variable affinity not directly predictable from the precise structure of their sialylated oligosaccharides chains. These data may help to explain why certain T and B cell surface sialoglycoproteins with the Neu5Ac alpha 2-6Gal beta 1-4Glc(NAc) motif are superior ligands, capable of mediating CD22 beta-mediated adhesion and activation events.
Natural ligands of the B cell adhesion molecule CD22 beta can be masked by 9-O-acetylation of sialic acids.
CD22 beta is a B cell-restricted phosphoprotein expressed on the surface of mature resting B cells. It mediates interactions with other cells partly or exclusively via recognition of alpha 2-6-linked sialic acids on glycoconjugates. The sialylated N-linked oligosaccharides recognized best by CD22 beta are common to many glycoproteins, suggesting that additional regulatory mechanisms may exist. Since the exocyclic side chain of sialic acid is required for recognition, we explored the effects of a naturally occurring modification of the side chain, 9-O-acetylation. Semisynthetic N-linked oligosaccharides terminating with 9-O-acetylated, alpha 2-6-linked sialic acids showed markedly reduced binding to CD22 beta relative to their non-O-acetylated counterparts. Murine lymphoid cells were probed for natural CD22 beta ligands that might be O-acetylated using recombinant soluble forms of CD22 beta (CD22 beta Rg) and influenza C esterase (CHE-Fc, which specifically removes 9-O-acetyl esters from sialic acids). By flow cytometry analysis, CD22 beta Rg binding to splenic B cells and a subset of T cells was increased by pretreatment with CHE-Fc, indicating that some potential CD22 beta ligands are naturally "masked" by 9-O-acetylation. Unmasking of these CD22 beta ligands by removal of 9-O-acetyl esters from intact splenocytes substantially increases their CD22 beta-dependent adhesion in an in vitro adhesion assay. Probing of murine lymphoid tissue sections by CD22 beta Rg and CHE-Fc treatment demonstrates regionally restricted and differentially expressed patterns of distribution between masked and unmasked ligands. For example, lymph node-associated follicular B cells express high levels of CD22 beta ligands, none of which are masked by 9-O-acetylation. In contrast, the ligands on lymph node-associated dendritic cells are almost completely masked by 9-O-acetylation, suggesting that masking may regulate interactions between CD22 beta-positive B cells and dendritic cells. In the thymus, only medullary cells express CD22 beta ligands, and a significant portion of these are masked by 9-O-acetylation, particularly at the cortical-medullary junction. Thus, 9-O-acetylation of sialic acids on immune cells is in a position to negatively regulate CD22 beta adhesion events in a manner depending on both cell type and tissue localization.
Carbohydrate-deficient glycoprotein syndrome: not an N-linked oligosaccharide processing defect, but an abnormality in lipid-linked oligosaccharide biosynthesis?
The carbohydrate-deficient glycoprotein syndrome (CDGS) is a developmental disease associated with an abnormally high isoelectric point of serum transferrin. Carbohydrate analyses of this glycoprotein initially suggested a defect in N-linked oligosaccharide processing, although more recent studies indicate a defect in the attachment of these sugar chains to the protein. We studied both serum glycoproteins and fibroblast-derived [2-3H]mannose-labeled oligosaccharides from CDGS patients and normal controls. While there was a decrease in the glycosylation of serum glycoproteins of affected individuals, differences were not seen in either monosaccharide composition or oligosaccharide structures. The lectin-binding profiles of glycopeptides from [2-3H]-mannose-labeled fibroblasts were likewise indistinguishable. However, the incorporation of [2-3H]mannose into both glycoproteins and the dolichol-linked oligosaccharide precursor was significantly reduced. Thus, at least in some patients, CDGS is not due to a defect in processing of N-linked oligosaccharides, but rather to defective synthesis and transfer of nascent dolichol-linked oligosaccharide precursors. This abnormality could result in both a failure to glycosylate some sites on some proteins, as well as secondary abnormalities in overall glycoprotein processing and/or function.
Natural ligands of the B cell adhesion molecule CD22 beta carry N-linked oligosaccharides with alpha-2,6-linked sialic acids that are required for recognition.
CD22 beta is a glycoprotein found on the surface of B cells during restricted stages of development. It is believed to play a role in cell-cell interactions and B cell activation. The accompanying paper (Sgroi, D., Varki, A., Braesch-Andersen, S., and Stamenkovic, I. (1993) J. Biol. Chem. 268, 7011-7018) shows that CD22 beta recognizes multiple glycoproteins on the surfaces of T and B cells and that sialylation of these ligands is essential for binding. To identify the structure(s) of the sialylated oligosaccharide(s) recognized by CD22 beta, [3H]glucosamine-labeled glycoproteins were purified from Daudi cells by adsorption onto a CD22 beta recombinant immunoglobulin (CD22 beta Rg) chimera attached to protein A-Sepharose (PAS), and the N-linked oligosaccharides were released by peptide N-glycosidase F. These released oligosaccharides failed to bind to CD22 beta Rg-PAS under the conditions used initially to adsorb the glycoproteins, but their elution from a column of CD22 beta Rg-PAS was significantly retarded. Populations of oligosaccharides with different affinities could be identified by their order of elution. Specific sialidases were used to determine the content of alpha-2,3- and alpha-2,6-linked sialic acid in these different populations and their contribution to binding. Multiantennary oligosaccharides with one alpha-2,6-linked residue bound marginally, and those with two or more bound more tightly. alpha-2,3-Linked sialic acid residues were without effect. Binding did not require divalent cations and was abrogated by mild periodate oxidation of the outer side chain of sialic acid. No marked differences in size or fucose content were found between the populations of high and low affinity oligosaccharides. However, the low affinity population could be partially converted into higher affinity by treatment with beta-galactoside alpha-2,6 sialyltransferase and CMP-sialic acid. Thus, CD22 beta is a mammalian lectin that can recognize specific N-linked oligosaccharide structures containing alpha-2,6-linked sialic acids.
Site specific glycosylation patterns of H-2K: effects of allelic polymorphism and mitogenic stimulation.
The site-specific glycosylation patterns of two H-2K alleles, k and b, were determined on splenic T cells metabolically labeled with [3H]mannose. Cells from B10, B10.A, (B10 X B10.A)F1, and C3H mice were examined, along with the effect of short- (8 hr) and long-term (36 hr) mitogenic stimulation. For both glycosylation sites (Asn86 and Asn176) of both antigens, 80% of the structures consisted of mono- and bisialylated biantennary N-linked complex oligosaccharides, with the remaining consisting of smaller (probably high mannose) structures. Asn176 of both H-2Kk and H-2Kb contained the same ratio (2.8 to 1) of bi- to monosialylated chains. However, Asn86 of H-2Kb contained a higher ratio (5 to 1), while Asn86 of H-2Kk a lower ratio (1.5 to 1). This difference was seen on antigens isolated from cells of the parental strains as well as from the F1 cross. The glycosylation of H-2Kk did not vary between B10.A and C3H mice. Mitogenic stimulation increased markedly both total [3H]mannose incorporation and the spectrum of N-linked oligosaccharides labeled. For H-2Kk, it had no effect on sialylation, but resulted in a slight under galactosylation of the monosialylated structures at both sites. A comparison of the patterns seen here, determined on nontransformed T cells, with those previously determined on H-2Kk from a B lymphoma line, revealed marked differences in sialylation and branching patterns at both sites. These data indicate that glycosylation differences may be found between highly homologous (91%) alleles of an H-2 gene, even when co-dominantly expressed by F1 cells; however, the patterns do change with mitogenic stimulation, and between normal and transformed cells.
Cell surface sialic acid influences tumor cell recognition in the mixed lymphocyte reaction.
The Ia+ B cell lymphoma, AKTB-1b, fails to stimulate thymic lymphocytes in a one-way mixed lymphocyte reaction unless pretreated with sialidase or inhibitors of N-linked oligosaccharide processing. A comparison of different sialidases and sialyltransferases suggests that the removal of only a subset of total surface sialic acid, rather than net desialylation of the cell surface, is required. Three sialidases were compared, including Vibrio cholerae (VC) and Clostridium perfringens (CP), which will cleave alpha 2-3, alpha 2-6, and alpha 2-8, sialic acid linkages, and Newcastle Disease virus (NDV), which will remove only alpha 2-3 and alpha 2-8 linked sialic acid. When treated with equivalent units of sialidase, CP-, VC-, and NDV-treated cells were 24-fold, sixfold, and threefold better stimulators than untreated cells. In contrast, VC released 1.3-fold and 2.5-fold more sialic acid per cell than did CP or NDV, respectively. Furthermore, VC was superior in reducing the levels of binding of the sialic acid-specific lectin, Limulus polyphemus agglutinin, in exposing Gal beta 1-3GalNAc and Gal beta 1-4GlcNAc residues, and in desialylating gangliosides. Two-dimensional gel analysis indicated that VC and CP were both equal and superior to NDV in the desialylation of iodinatable cell-surface proteins, including H-2Kk, I-A beta k, and a highly sialylated 65,000 dalton protein of unknown identity. Maximal resialylation of CP-treated cells with exogenously added CMP-NANA and either the alpha 2-3(Gal beta 1-3GalNAc) or alpha 2-6(Gal beta 1-4GlcNAc) sialyltransferase did not reduce the stimulatory capacity of these cells. However, resialylation of VC-treated cells with just CMP-NANA alone resulted in 49% reversal of their stimulatory capacity, and no additional reversal could be achieved with either of the sialyltransferases. Although the alpha 2-6(Gal beta 1-4GlcNAc) sialyltransferase was capable of adding back approximately 10% of the sialic acid removed, the endogenous activity added back approximately 0.1% of the total sialic acid removed. SDS-PAGE gels of the sialylated cells indicated that the exogenously added sialyltransferase labeled many different proteins, whereas the endogenous activity labeled far fewer proteins, predominantly in 46,000 and 25,000 m.w. range. Both the desialylation and resialylation data suggest that the sialidase-dependent stimulation is due to the desialylation of specific membrane structures. Together with previous studies, these data suggest that the sialic acids involved are probably alpha 2-6 linked to N-linked glycosyl moieties.
Quantitation of picomole levels of N-acetyl- and N-glycolylneuraminic acids by a HPLC-adaptation of the thiobarbituric acid assay.
A simple HPLC adaptation of the periodate-TBA assay for free N-acetyl- and N-glycolylneuraminic acids greatly extends the sensitivity and increases the specificity of this standard colorimetric assay. The method, employing a C18 reverse-phase column eluted isocratically with a phosphoric acid-MeOH buffer, is linear over a range of 2 pmol to 20 nmol. Analyses can be performed directly on cell lysates and digests without prior purification of released sialic acids from contaminating salts and biological materials. Interference from 2-deoxysugars is completely eliminated as the chromophore from these compounds is completely resolved from that derived from sialic acids. The application of the technique to quantify cell-surface and total cellular TBA-reactive sialic acids on the surfaces of a variety of tumor cells is described. Additionally, the extent of desialylation of erythrocytes necessary to expose the T antigen is determined.
Influence of asparagine-linked oligosaccharides on tumor cell recognition in the mixed lymphocyte reaction.
The ability of an Ia+ B cell lymphoma, AKTB-1b, to stimulate thymocytes in the allogeneic mixed lymphocyte reaction is dependent on its prior treatment with either swainsonine or deoxynojirimycin, two inhibitors of the processing of asparagine-linked oligosaccharides. In the absence of drug treatment, the tumor cells fail to stimulate thymocytes, whereas pretreatment of the tumor cells with either drug results in a five- to 10-fold increase in their ability to induce thymocyte proliferation. Drug-treated AKTB-1b stimulates thymocytes at levels comparable to those obtained with allogeneic splenocytes. In contrast, the untreated lymphoma does stimulate splenic lymphocytes, and pretreatment with either inhibitor only marginally increases the response. Genetic studies demonstrate that the thymocyte response is still H-2 locus restricted and can be blocked by monoclonal antibodies against two tumor cell major histocompatibility antigens, H-2K and I-A. Drug treatment does not change cell surface I-A expression, and H-2K levels are apparently decreased one-third by deoxynojirimycin but are not affected by swainsonine. To verify that the drug protocol used was capable of altering the glycoconjugates of membrane-associated proteins, the endo-beta-N-acetylglucosaminidase H (endo H)-sensitivity of immunopurified H-2K and I-A was analyzed by SDS-PAGE. These studies demonstrated that swainsonine treatment does result in cell surface expression of glycoconjugates with altered oligosaccharide moieties. Likewise, deoxynojirimycin treatment results in the cell surface expression of an I-A alpha polypeptide with altered oligosaccharide chains while only marginally affecting H-2K and not affecting the I-A beta chain. An intracellular form of the I-A beta chain sensitive to endo H digestion in the presence of deoxynojirimycin is not detectable at the cell surface. Neuraminidase-digested AKTB-1b are also capable of stimulating allogeneic thymocytes. These studies demonstrate that changes in the glycosylation state of the tumor cell can markedly influence its recognition by allogeneic lymphocytes, and further, that different T cell populations differ in their response to such changes.
Reconstitution and regulation of actin gel-sol transformation with purified filamin and villin.
Gel-sol transformation of actin filaments, a process essential for cell motility, can be reconstituted in vitro and regulated in a predictable fashion by the combined action of villin and filamin. Measurements made in a low shear falling ball viscometer show that mixtures of actin, villin, and filamin exist either as a gel (yield point greater than or equal to 140 dynes/cm2) or as a low viscosity liquid depending on the relative ration of villin:actin. Filamin induces gelation of F-actin by forming stable cross-links between actin filaments. Villin inhibits filamin-induced F-actin gelation, but the effect can be overcome by increasing the amount of filamin. Sedimentation assays show that villin does not inhibit gelation of actin by preventing filamin from binding to F-actin. Results from viscosity measurements and filament length determinations show that villin increases actin filament number by reducing the average filament length without altering the total amount of polymer. Because the gel point of a fixed amount of polymer is sharply dependent on the ratio of cross-links to number of polymers, the solation effect of villin might be explained by its effect on filament number. Based on the network theory of gel formation, calculations of the amount of additional cross-linker required to overcome the effect of a known increase in the number of actin filaments agree reasonably well with experimental findings. These results document the existence of cellular proteins which could regulate gel-sol transformation in vivo by their effect on actin polymer length and, therefore, on actin filament number.
Structural changes in (Na+ + K+)-ATPase accompanying detergent inactivation.
Structural changes in the purified (Na+ + K+)-ATPase accompanying detergent inactivation were investigated by monitoring changes in light scattering, intrinsic protein fluorescence, and tryptophan to beta-parinaric acid fluorescence resonance energy transfer. Two phases of inactivation were observed using the non-ionic detergents, digitonin, Lubrol WX and Triton X-100. The rapid phase involves detergent monomer insertion but little change in protein structure or little displacement of closely associated lipids as judged by intrinsic protein fluorescence and fluorescence resonance energy transfer. Lubrol WX and Triton X-100 also caused membrane fragmentation during the rapid phase. The slower phase of inactivation results in a completely inactive enzyme in a particle of 400 000 daltons with 20 mol/mol of associated phospholipid. Fluorescence changes during the course of the slow phase indicate some dissociation of protein-associated lipids and an accompanying protein conformational change. It is concluded that non-parallel inhibition of (Na+ + K+)-ATPase and p-nitrophenylphosphate activity by digitonin (which occurs during the rapid phase of inactivation) is unlikey to require a change in the oligomeric state of the enzyme. It is also concluded that at least 20 mol/mol of tightly associated lipid are necessary for either (Na+ + K+)-ATPase or p-nitrophenylphosphatase activity and that the rate-limiting step in the slow inactivation phase involves dissociation of an essential lipid.
Regulation of actin polymerization by villin, a 95,000 dalton cytoskeletal component of intestinal brush borders.
A 95,000 dalton actin-binding polypeptide, villin, has been purified to 98% homogeneity from brush border cytoskeletons of chicken intestinal epithelial cells. In vitro, this protein exerts control over the polymerization of actin. In the presence of villin, the lag phase preceding detectable actin polymerization is shortened and the steady state equilibrium viscosity is reduced in proportion to the amount of villin present. A molar ratio of villin:actin of 1:40 results in a 70% reduction of the Ostwald viscosity. Significant effects can be detected at a ratio of 1:600. These ratios are physiologically relevant because the ratio of villin:actin in brush borders is 1:13 and in isolated microvilli is 1:9-12. Reduction of viscosity is mirrored by an increase in the amount of protein which fails to sediment at 150,000 X g for 60 min. An assay of the nonsedimentable protein for actin monomers by the inhibition of DNAase I showed that the concentration of monomer was not significantly altered by the presence of villin. Electron microscopic examination of negatively stained, nonsedimentable actin demonstrated that the presence of villin during actin polymerization results in the production of short oligomers which cannot anneal with each other to form long filaments. Villin is also effective in reducing the viscosity of F-actin when it is added to a fuly polymerized actin sample. In view of these striking properties, villin is likely to be an important in vivo regulator of cytoskeletal structure and, by implication, of cell shape and motility.