Molecular Dynamics in Protein—Single Stranded DNA Complexes. Two Distinct Nucleoside Mobilities in Poly(deoxythymidylic acid)—poly-L-lysine Complexes

Abstract

Stoichiometric amounts of poly-L-lysine were added to site-specifically spin labeled single stranded nucleic acids and the resulting complexes analyzed by electron spin resonance spectroscopy (ESR). The nucleic acids were spin labeled to different extents and with labels of varying tether length. The ESR data are used to determine nucleoside dynamics and some structural features in these complexes. It is concluded that two distinct base mobilities exist in the complexes; one set is characterized by a mean correlation time τR = 2 ns, and the other one by a τR ≤ 50 ns. A model is proposed which suggests that a poly-L-lys single stranded nucleic acid complex consists of low mobility segments flanked by more mobile bases. An interesting feature of the proposed model is its applicability to explain ESR data of single strand binding protein-spin labeled nucleic acid complexes, which can also be interpreted in terms of two distinct nucleoside mobility states. It is hypothesized that this phenomenon could be of biological significance for the release of protein ligands from a protein-nucleic acid complex.     

Molecular dynamics in protein-single stranded DNA complexes. Two distinct nucleoside mobilities, in poly(deoxythymidylic acid)-poly-L-lysine complexes

Stoichiometric amounts of poly-L-lysine were added to site-specifically spin labeled single stranded nucleic acids and the resulting complexes analyzed by electron spin resonance spectroscopy (ESR). The nucleic acids were spin labeled to different extents and with labels of varying tether length. The ESR data are used to determine nucleoside dynamics and some structural features in these complexes. It is concluded that two distinct base mobilities exist in the complexes; one set is characterized by a mean correlation time tau -R = 2 ns, and the other one by a tau -R greater than or equal to 50 ns. A model is proposed which suggests that a poly-L-lys single stranded nucleic acid complex consists of low mobility segments flanked by more mobile bases. An interesting feature of the proposed model is its applicability to explain ESR data of single strand binding protein-spin labeled nucleic acid complexes, which can also be interpreted in terms of two distinct nucleoside mobility states. It is hypothesized that this phenomenon could be of biological significance for the release of protein ligands from a protein-nucleic acid complex.     

ESR study of 5'-nucleotidase from bull seminal plasma

5'-Nucleotidase of bull seminal plasma has been spin labeled with the sulfhydryl reagent 3-maleimidoproxyl. ESR analysis reveals the presence of two classes of labeled sites. The first is characterized by a long spin label rotational correlation time, from which a protein diameter of about 70 A can be estimated, under the assumption of a spherical shape. The second class is characterized by a shorter correlation time of the covalently bound spin labels and binding of the substrate sodium thymidine 5'-monophosphate to 5'-nucleotidase results in a reduction of their mobility. Low-temperature ESR analysis shows that no paramagnetic ion is bound to the native protein.     

Electron paramagnetic resonance study of the interactions between steroid hormones and binding proteins]

Interaction of a spin labeled corticosteroid (desoxycorticosterone nitroxyde: DOC -NO) with three purified proteins (albumin, transcortin, progesterone binding protein: PBG) was studied by electron spin resonance (ESR) spectroscopy. DOC-NO was competitive with natural corticosteroids and therefore bound at the same site to specific binding proteins. ESR spectra in the presence of each of the proteins showed an immobilized (bound) form of the spin labeled steroid and allowed the calculation of the corresponding association constant (Ka) at equilibrium. The three binding proteins could be characterized by the ESR parameters of the DOC-NO bound form. The thermodynamic parameters (deltaH, deltaS) of the steroid-protein interactions were calculated from the ESR data obtained within a wide temperature range (3--40 degrees C). The ESR spectra width (2T) was used to evaluate the polarity of the spin label environment within the steroid binding site: a hydrophobic character was observed for transcortin whereas PBG exhibited a more hydrophilic steroid binding sits. The rotational correlation time of the three protein DOC-NO complexes at equilibrium were calculated from ESR data; the results were correlated with the protein molecular size and suggested a non spherical shape for the binding macromolecule in solution. Spin labelling of biologically active steroids thus provides a novel approach for the study of the interaction of these hormones with their binding protein. Providing a suitable spin label, the ESR parameters may allow the characterization of several types of binding sites of different biological significance for the same hormone, in biological fluids as well as in target tissues.     

An ESR study of the influence of some physico-chemical factors on the conformation of a postsynaptic acetylcholinesterase.

1. In a previous ESR study of a membrane acetylcholinesterase (EC 3.1.1.7) we found, contrary to observations by other authors, spectra indicating that the active serine might be located in a pocket of the enzyme surface. In order to inquire into this possibility, ESR spectra were studied under the influence of different physico-chemical factors known to cause an unfolding of proteins. 2. The active serine of the postsynaptic membrane acetylcholinesterase of Torpedo marmorata electric organ was spin labeled using 1-oxyl-2, 2, 6, 6-tetramethyl-4-piperidinyletoxyphosphonofluoridate. 3. The effect of the chosen physico-chemical factors was an increase in the rotational freedom of spin labels; this result corroborates the suggestion that the active center of our acetylcholinesterase preparation is located in a pocket.     

Evidence that the two free sulfhydryl groups of plasma fibronectin are in different local environments. Saturation-recovery electron spin resonance study.

Human plasma fibronectin is a dimer consisting of two subunits; each contains two cryptic thiol groups that were selectively labeled with an 15N,2H-maleimide spin label. Previous studies using conventional X-band electron spin resonance (ESR) methods showed that the spectrum of the labeled protein displays a single strongly immobilized component with an effective rotational correlation time of approximately 17 ns, suggesting that the physical environments of the two labeled sites per chain are indistinguishable. Here we have used saturation-recovery ESR to measure directly electron spin-lattice relaxation time (T1) of the labeled protein in solution at 27 degrees C. Interestingly, the time evolution of the signal was found to be biphasic, which was deconvoluted into two T1 values of 1.37 and 4.53 microseconds. Thus, the two spin-labeled sulfhydryl sites of plasma fibronectin (Fn), being similar in rates of rotational diffusion, differ by a factor of 3.2 in T1. Parallel experiments using various fibronectin fragments showed that the 1.37-microseconds component is associated with the label attached onto the thiol located in between the DNA-binding and the cell-binding domains, and the 4.53-microseconds component is associated with the label attached onto the thiol located within the carboxyl-terminal fibrin-binding domain. The data suggest that the saturation-recovery ESR is a useful method for differentiating multiple spin-labeled sites on macromolecules in which the labels undergo similar rates of rotational motion.     

Spin label study of slow molecular rotations of globular proteins using microwave saturation effects in electron paramagnetic resonance

Viscosity, temperature and ionic strength dependences of ESR microwave saturation parameters of spin labelled human oxyhemoglobin (Hb) and bovine serum albumin (BSA) have been studied. The piperidine and pyrrolidine nitroxyl derivatives of maleimide were used as covalent SH reagents for Hb and BSA and the same two derivatives of gamma-benzocarboline and spin labelled stearic acid were used as noncovalent spin probes for BSA. The effects of label binding tightness on ESR spectral parameters were considered. The rotational correlation times were determined using viscosity dependences of the separation of the outer hyperfine extrema and Stokes extrapolations at high viscosities. The ESR microwave saturation parameters of the spin labels were shown to depend just weakly on temperature (at constant eta/t) over the range 0-25 degrees and on g, A values but to be sensitive to protein rotational correlation times up to 10(-4) sec and also to the rotational anisotropy and to the relative motion of the spin label.     

Structural analysis of nucleic acids by labeled oligonucleotides

In this report, the characterization of labeled oligonucleotides was discussed from the view points of base sequence analysis and structural analysis of nucleic acids in solution. Oligonucleotides site specifically spin labeled with TEMPO and fluorescent labeled with fluorescein were prepared and used for those analyses. The changes of ESR lines and rotational correlation time (tau) of the spin labeled oligonucleotide (S-probe) were dependent on the base sequence of S-probe, diastereoisomers, and the manner of hybridization. These results suggest that the conformation of the hybrid largely affected the local mobility of TEMPO and that tau value of S-probe reflected the local structure of the hybrid. When S-probe which was complementary to a single strand region of 5S RNA, was mixed with 5S RNA, tau value largely changed, indicating that the S-probe could form hybrid with 5S RNA in solution. Similar results were also obtained in the fluorescence depolarization analysis using fluorescent labeled oligonucleotide (F-probe). These results suggest that S-probe and F-probe are capable for the recognition of the secondary structure of 5S RNA in solution and useful for the analysis of the secondary structure of other nucleic acids in solution.     

Alternations in lipid membrane fluidity and the physical state of cell-surface sialic acid in zinc-deficient rat erythrocyte ghosts

Erythrocyte ghosts, prepared from the blood of rats fed zinc-deficient diets, were evaluated for membrane fluidity and surface sialic acid properties using spin-labeled probes and electron spin resonance (ESR) spectroscopy. These physical parameters of the erythrocyte ghosts from the zinc-deficient group were compared to those for erythrocyte ghosts obtained from ad libitum and pair fed controls consuming zinc-adequate diets. As the animals became progressively zinc deficient, the erythrocyte ghost membranes became more fluid than those from the control groups. In addition, the apparent rotational correlation time of Tempamine spin probes on surface sialic acid residues was smaller for the zinc deficient group, indicative of an increased rotational mobility of the spin label. These results suggest that zinc deficiency can have pronounced effects on the physical state of membrane bilayer lipids and cell surface carbohydrates and supports the view that many of the pathological signs of zinc deficiency are due to a general membrane defect.     

A spin-label study of membrane proteins and internal microviscosity of erythrocytes in hereditary spherocytosis

Electron spin resonance (ESR) spectra of erythrocyte membranes of patients with hereditary spherocytosis (HS) and of healthy controls labeled with a maleimide spin label did not differ significantly both before and after prolonged incubation at 37 degrees C. It suggests that the different behavior of spin-labeled HS erythrocyte membranes upon incubation at a higher temperature reported previously is due indeed to structural abnormalities of HS red cell membranes and not to alterations in their proteolytic activity. Measurements of the rotational correlation time of Tempamine spin probe demonstrated a significant elevation of internal microviscosity of erythrocytes in HS, more pronounced in non-splenectomized patients.     

Lipid protein interactions in mitochondria. VIII. Effect of general anesthetics on the mobility of spin labels in lipid vesticles and mitochondrial membranes

We have studied the effect of general anesthetics on the mobility of two stearic acid spin labels (5-doxyl stearic acid and 16-doxyl stearic acid) in bovine heart mitochondria and in phospholipid vesicles made from either mitochondrial lipids or commercial soybean phospholipids. The general anesthetics used include nonpolar compounds (alcohols, halothane, pentrane, diethyl ether, chloroform) and the amphipathic compound, ketamine. All anesthetics tested increase the mobility of the spin labels in phospholipid vesicles to a limited extent up to a concentration where the ESR spectra become those of free spin labels. On the other hand, anesthetics have a pronounced effect on mitochondrial membranes at concentrations as low as those known to produce general anesthesia; the effect is lower near the bilayer surface (5-doxyl stearic acid) and very strong in the bilayer core (16-doxyl stearic acid). The effects of anesthetics are mimicked by the detergent, Triton X-100. We suggest that the discrepancy between the action of anesthetics in mobilizing the spin labels in lipid vesicles and in membranes results from labilization of lipid protein interactions.     

Spin labeling of calcium-dependent phospholipid-binding proteins

Bovine lung annexins p32 and p34 were spin labeled with an iodoacetamidoproxyl spin label, a reagent that reportedly couples with protein methionine residues. Labeling conditions and stoichiometry were studied with the radiolabeled analogue [1-14C]iodoacetamide. As judged by this method, carboxamidomethylation of both p32 and p34 occurred up to a 0.7 mol ratio after 60 h of reaction at 37 degrees C and at pH 4. The two proteins retained Ca2(+)-dependent phospholipid-binding ability both in radiolabeled and in spin-labeled forms. Electron resonance spectra of spin-labeled p32 and p34 showed the features of a partially immobilized spin probe, with rotational correlation time values of 1.15 and 1.25 ns, respectively, which definitely indicate successful spin labeling. Quantitation of ESR spectra by computer double integration indicated 70% spin labeling of both proteins, as anticipated by radiolabeling. The use of spin-labeled p32 and p34 in the study of Ca2(+)-dependent interaction of annexins with biomembranes is proposed.     

A multifrequency electron spin resonance study of T4 lysozyme dynamics.

Electron spin resonance (ESR) spectroscopy at 250 GHz and 9 GHz is utilized to study the dynamics and local structural ordering of a nitroxide-labeled enzyme, T4 lysozyme (EC 3.2.1.17), in aqueous solution from 10 degrees C to 35 degrees C. Two separate derivatives, labeled at sites 44 and 69, were analyzed. The 250-GHz ESR spectra are well described by a microscopic ordering with macroscopic disordering (MOMD) model, which includes the influence of the tether connecting the probe to the protein. In the faster "time scale" of the 250-GHz ESR experiment, the overall rotational diffusion rate of the enzyme is too slow to significantly affect the spectrum, whereas for the 9-GHz ESR spectra, the overall rotational diffusion must be accounted for in the analysis. This is accomplished by using a slowly relaxing local structure model (SRLS) for the dynamics, wherein the tether motion and the overall motion are both included. In this way a simultaneous fit is successfully obtained for both the 250-GHz and 9-GHz ESR spectra. Two distinct motional/ordering modes of the probe are found for both lysozyme derivatives, indicating that the tether exists in two distinct conformations on the ESR time scale. The probe diffuses more rapidly about an axis perpendicular to its tether, which may result from fluctuations of the peptide backbone at the point of attachment of the spin probe.     

ESR studies of the erythrocyte membrane skeletal protein network: influence of the state of aggregation of spectrin on the physical state of membrane proteins, bilayer lipids, and cell surface carbohydrates

The stability of the human erythrocyte membrane skeletal network is reported to be dependent on the state of aggregation of spectrin and decreased or increased by polyphosphate anions or the polyamine, spermine, respectively. We have employed polyacrylamide gel electrophoresis and electron spin resonance (ESR) utilizing spin labels specific for membrane proteins, bilayer lipids, or cell-surface sialic acid in order to gain insight into these observations and into the reliability of the ESR spectra of the protein-specific spin label used to correctly report the interactions of the skeletal protein network. The major findings are: (1) We confirm previous reports that the preferred state of spectrin aggregation in the skeletal network is tetrameric and that spectrin can be reversibly transformed to dimeric spectrin and back to tetrameric spectrin on the membrane. (2) The ESR spectra of the protein specific maleimide spin label employed accurately reflect the state of aggregation of spectrin. (3) As dimeric spectrin is increased on the membrane or when 2,3-bisphosphoglycerate was added to spin-labeled membranes, increased segmental motion of protein spin label binding sites reflecting decreased protein-protein interactions in the skeletal network is observed (P less than 0.002 and P less than 0.005, respectively). (4) Conversely, as protein-protein interactions between skeletal proteins or between skeletal proteins and the bilayer are increased by spermine (reflected in the total inability to extract spectrin from the membrane in contrast to control membranes), highly decreased segmental motion of the protein specific spin label binding site is observed (P less than 0.005). (5) The dimeric-tetrameric state of spectrin aggregation on the membrane does not have influence on the order or motion of bilayer lipids nor on the rotational rate of spin-labeled, cell-surface sialic acid, a result also observed when protein-protein interactions were decreased by 2,3-bisphosphoglycerate. In contrast, increased protein-protein interactions by addition of spermine produced a small, but significant, increase in order and decrease in motion of bilayer lipids near the membrane surface as well as a nearly 40% decrease in the apparent rotational correlation time of spin labeled, cell surface sialic acid (P less than 0.002). These latter observations are discussed with reference to possible associations of phospholipids and the major, transmembrane sialoglycoprotein with the skeletal protein network.     

Rotational dynamics of protein and boundary lipid in sarcoplasmic reticulum membrane

We have used spin labels and electron paramagnetic resonance (EPR) to study the correlation between the rotational dynamics of protein and lipid in sarcoplasmic reticulum (SR) membranes. A short-chain maleimide spin label was used to monitor the submillisecond rotational mobility of the Ca-ATPase enzyme (using saturation transfer EPR); a free fatty acid spin label was used to monitor the submicrosecond rotational mobility of the bulk lipid hydrocarbon chains (using conventional EPR); and a fatty acid spin label derivative (long-chain maleimide) attached to the enzyme was used to monitor the mobility of hydrocarbon chains adjacent to the protein (i.e., boundary lipid). In the native SR membranes, the protein was highly mobile (effective correlation time 50 microseconds). The spectra of the hydrocarbon probes both contained at least two components. For the unattached probe, the major component indicated nearly as much mobility as in the absence of protein (effective rotational correlation time 3 ns), while a minor component, corresponding to 25-30% of the total signal, indicated strong immobilization (effective correlation time greater than or equal to 10 ns). For the attached hydrocarbon probe, the major component (approximately 70% of the total) was strongly immobilized, and the mobile component was less mobile than that of the unattached probe. When the lipid-to-protein ratio was reduced 55% by treatment with deoxycholate, protein mobility decreased considerably, suggesting protein aggregation. A concomitant increase was observed in the fraction of immobilized spin labels for both the free and attached hydrocarbon probes. The observed hydrocarbon immobilization probably arises in part from immobilization at the protein-lipid boundary, but protein-protein interactions that trap hydrocarbon chains may also contribute. When protein aggregation was induced by glutaraldehyde crosslinking, submillisecond protein mobility was eliminated, but there was no effect on either hydrocarbon probe. Thus protein aggregation does not necessarily cause hydrocarbon chain immobilization.     

Interaction of delta-sleep-inducing peptide with cell membranes in vitro]

The effect of delta-sleep-inducing peptide (DSIP) on erythrocytic membranes of human donor blood was studied by the spin label and spin probe methods. The spin-labeled derivative of DSIP containing the N-terminal residue of 1-oxyl-2,2,5,5-tetramethylpyrroline-3-carboxylic acid was synthesized. An analysis of the ESR spectra of the spin-labeled DSIP derivative recorded after its incubation with a human erythrocyte suspension at 37 degrees C revealed a decrease in the rotational correlation time (tau c) and molecular order parameter (S) in comparison with the control solutions of the peptide in phosphate buffer (pH 7.4). The application of paramagnetic probes, 5-, 12-, and 16-doxylstearic acids and 3-doxylandrostanol, demonstrated that the introduction of DSIP in an erythrocytic suspension significantly increased the mobility of the hydrophobic area of the membrane bilayer both at a depth of 20-22 A and in the subsurface area (4-6 A). The dependence of these effects on the DSIP concentration was shown to have the form of a curve with well-defined extremes. The maximal disordering of membrane lipids was observed at peptide concentrations of 10(-9) and 10(-6) M. These results suggested that DSIP significantly affected the structure of plasmatic membranes in vitro by changing the physical state of their lipid components.     

Analysis of existing models of spin label movement during spin labeling of biological macromolecules

The spin-labeled bovine serum albumin and IgG were studied in search of an experimental approach for comparison of different models of rotational mobility of spin label. These models are: the model of isotropic motion of spin label together with the macromolecule (IM); the model of highly anisotropic motion of spin label (HAM); and the model of slow isotropic motion of label around the binding site (SIML). The experimental spectra were measured on a common X-band ESR spectrometer and on the unique 140 GHZ (lambda = 2 mm) ESR spectrometer under the same conditions. Theoretical spectra were computer-calculated according to Freed's theory. We have found, that the results of temperature-viscosity experiments in X-band are contradictory to the model of IM both for the BSA and IgG species. The models of HAM and SIML for the BSA give identical X-band spectra. The bovine serum albumin spectra in the 2 mm region strongly contradict to the assumptions of the HAM model. Also, the SIML model fails to describe the experimental spectra in terms of isotropic motion of the spin label around the binding site. X-band spectra of IgG can not be explained by the SIML model, while the same spectra in the 2 mm region can not be explained by the HAM model.     

Rotational mobility of an erythrocyte membrane integral protein band 3 in dimyristoylphosphatidylcholine reconstituted vesicles and effect of binding of cytoskeletal peripheral proteins

Band 3 protein was isolated from human erythrocyte membranes, purified, and reconstituted into a well-defined phospholipid bilayer matrix (dimyristoylphosphatidylcholine). The preparation yielded uniform single-bilayered vesicles of the diameter 40--80 nm. The rotational motion of band 3 was studied by saturation transfer electron spin resonance (ESR) spectroscopy of covalently attached maleimide spin-labels. The rotational mobility changed in response to the host lipid phase transition. The rotational correlation time was in a range from 73 (37 degrees C) to 94 microseconds (26 degrees C) in the fluid phase and from 240 (15 degrees C) to 420 microseconds (5 degrees C) in the solid phase. The motion was analyzed based on the anisotropic rotation of band 3 in the reconstituted vesicles. To obtain information on the rotational diffusion constant around the axis parallel to the membrane normal, we made an attempt to measure the angle between the spin-label magnetic axis and the membrane normal. The result gave 3.9 x 10(4) s-1 at 37 degrees C as a rough estimate for the diffusion constant. This is compatible to anisotropic rotation of a cylinder of radius 3.3 nm in a two-dimensional matrix with inner viscosity 2 P and inner thickness 4 nm. The cytoskeletal peripheral proteins caused a definite increase in the rotational correlation time (from 73 to 180 microseconds at 37 degrees C, for example). The restriction of the rotational mobility was shown to be due to the ankyrin-linked interaction between band 3 and spectrin-actin-band 4.1 proteins in the reconstituted membranes.     

Electron spin resonance study of human alpha 1-acid glycoprotein interaction with a spin labelled steroid.

The interaction of human alpha 1-acid glycoprotein (AAG) with a corticosteroid was studied using nitroxide labeled deoxycorticosterone and electron spin resonance (ESR) spectroscopy. The ESR spectra of the spin labeled steroid in the presence of AAG could be used to characterize the ligand-protein interaction at equilibrium without the need of a separation between bound and free species. An association constant Ka of 6.10(5) M-1 at 20 degrees C and a binding capacity of one site per mole protein were found. ESR spectra recorded at equilibrium at various temperatures allowed the calculation of enthalpy and entropy variations for the steroid-protein interaction; these thermodynamic parameters exhibited a rapid change above 45 degrees C which may be related to a protein conformational modification above this temperature, as detected by circular dichroism study. The ESR spectra width could be used to define a polar character for the spin label environment in the steroid binding site of AAG and to calculate an apparent rotational correlation time of 2.8 x 10(-8) sec for the steroid-protein complex in aqueous solution at 20 degrees C. It can be concluded that spin labeling and ESR methodology is of value in the study of steroid-protein interactions of biological significance above all because it can provide direct physico-chemical information concerning the local environment of the ligand in its binding site at equilibrium.     

ESR studies of the erythrocyte membrane skeletal protein network: Influence of the state of aggregation of spectrin on the physical state of membrane proteins, bilayer lipids, and cell surface carbohydrates

The stability of the human erythrocyte membrane skeletal network is reported to be dependent on the state of aggregation of spectrin and decreased or increased by polyphosphate anions or the polyamine, spermine, respectively. We have employed polyacrylamide gel electrophoresis and electron spin resonance (ESR) utilizing spin labels specific for membrane proteins, bilayer lipids, or cell-surface sialic acid in order to gain insight into these observations and into the reliability of the ESR spectra of the protein-specific spin label used to correctly report the interactions of the skeletal protein network. The major findings are: (1) We confirm previous reports that the preferred state of spectrin aggregation in the skeletal network is tetrameric and that spectrin can be reversibly transformed to dimeric spectrin and back to tetrameric spectrin on the membrane. (2) The ESR spectra of the protein specific maleimide spin label employed accurately reflect the state of aggregation of spectrin. (3) As dimeric spectrin is increased on the membrane or when 2,3-bis-phosphoglycerate was added to spin-labeled membranes, increased segmental motion of protein spin label binding sites reflecting decreased protein-protein interactions in the skeletal network is observed (P < 0.002 and P < 0.005, respectively). (4) Conversely, as protein-protein interactions between skeletal proteins or between skeletal proteins and the bilayer are increased by spermine (reflected in the total inability to extract spectrin from the membrane in contrast to control membranes), highly decreased segmental motion of the protein specific spin label binding sites is observed (P < 0.005). (5) The dimeric-tetrameric state of spectrin aggregation on the membrane does not have influence on the order or motion of bilayer lipids nor on the rotational rate of spin-labeled, cell-surface sialic acid, a result also observed when protein-protein interactions were decreased by 2,3-bisphosphoglycerate. In contrast, increased protein-protein interactions by addition of spermine produced a small, but significant, increase in order and decrease in motion of bilayer lipids near the membrane surface as well as a nearly 40% decrease in the apparent rotational correlation time of spin labeled, cell surface sialic acid (P < 0.002). These latter observations are discussed with reference to possible associations of phospholipids and the major, transmembrane sialoglycoprotein with the skeletal protein network.