The use of a phospholipase A-less Escherichia coli mutant to establish the action of granulocyte phospholipase A on bacterial phospholipids during killing by a highly purified granulocyte fraction

Phospholipase A₂ present in a highly purified, potently bactericidal, fraction from rabbit granulocytes produces net bacterial phospholipid degradation during killing of a phospholipase A-less strain of Escherichia coli. In the wild-type parent strain phospholipid breakdown is caused not only by the action of phospholipase A₂ but also by phospholipase A₁, indicating activation of the most prominent phospholipase of E. coli. This activation occurs as soon as the bacteria are exposed to the granulocyte fraction. Phospholipid breakdown by both phospholipases A is dose dependent but reaches a plateau after 30–60 min and at higher concentrations of the fraction.Phospholipid degradation is accompanied in both strains by an increase in permeability to actinomycin D that is also dose dependent. Even though net hydrolysis of phospholipids is greater in the parent strain than in the mutant, the increase in permeability is the same in the two strains.The addition of 0.04 M Mg²⁺, after the effects on phospholipids and permeability have become manifest, initiates in both strains the restoration of insensitivity to actinomycin D, the net resynthesis of phospholipids, and the disappearance of monoacylphosphatides and the partial disappearance of free fatty acids that had accumulated. Loss of ability to multiply is not reversed by Mg²⁺ in either strain. Less than 5 μg of granulocyte fraction causes loss of viability of from 90 to 99% of $\text{1 × 10}{}^8$ microorganisms of both strains. However, at lower concentrations the parent strain is considerably more sensitive to the bactericidal effect of the granulocyte fraction than the mutant strain.     

Diverse effects of phospholipids on lipoprotein sorting and ATP hydrolysis by the ABC transporter LolCDE complex

The LolCDE complex of Escherichia coli releases outer membrane-specific lipoproteins from the inner membrane. Lipoproteins with Asp at + 2 remain in the inner membrane since this residue functions as a LolCDE avoidance signal depending on phosphatidylethanolamine. We examined the effects of other phospholipids on lipoprotein sorting in proteoliposomes reconstituted with LolCDE and various synthetic phospholipids. The lipoprotein release and ATP hydrolysis were both low at 2 mM Mg²⁺ but very high at 10 mM Mg²⁺ in proteoliposomes containing cardiolipin alone. However, the Lol avoidance function was abolished at 10 mM Mg²⁺, and the release of lipoproteins with Asp at + 2 was as efficient as that of outer membrane-specific lipoproteins. The addition of phosphatidylethanolamine to cardiolipin stimulated the ATP hydrolysis and increased the Lol avoidance function of Asp at + 2 at 2 mM Mg²⁺. The addition of phosphatidylglycerol to cardiolipin nearly completely inhibited the release of lipoproteins with Asp at + 2 even at 10 mM Mg²⁺, while that of outer membrane-specific lipoproteins was not. Taken together, these results indicate that three major phospholipids of E. coli differently affect lipoprotein sorting and the activity of LolCDE.     

The use of a phospholipase A-less Escherichia coli mutant to establish the action of granulocyte phospholipase A on bacterial phospholipids during killing by a highly purified granulocyte fraction.

Phospholipase A2 present in a highly purified, potently bactericidal, fraction from rabbit graulocytes produces net bacterial phospholipid degradation during killing of a phospholipase A-less strain of Escherichia coli. In the wild-type parent strain phospholipid breakdown is caused not only by the action of phospholipase A2 but also by phospholipase A1, indicating activation of the most prominent phospholipase of E. coli. This activation occurs as soon as the bacteria are exposed to the granulocyte fraction. Phospholipid breakdown by both phospholipases A is dose dependent but reaches a plateau after 30-60 min and at higher concentrations of the fraction. Phospholipid degradation is accompanied in both strains by an increase in permeability to actinomycin D that is also dose dependent. Even though net hydrolysis of phospholipids is greater in the parent strain than in the mutant, the increase in permeability is the same in the two strains. The addition of 0.04 M Mg2+, after the effects on phospholipids and permeability have become manifest, initiates in both strains the restoration of insensitivity to actinomycin D, the net resynthesis of phospholipids, and the disappearance of monoacylphosphatides and the partial disappearance of free fatty acids that had accumulated. Loss of ability to multiply is not reversed by Mg2+ in either strain. Less than 5 micrograms of granulocyte fraction causes loss of viability of from 90 to 99% of 1 X 10(8) microorganisms of both strains. However, at lower concentrations the parent strain is considerably more sensitive to the bactericidal effect of the granulocyte fraction than the mutant strain.     

Inhibition of Mg2+ binding and DNA religation by bacterial topoisomerase I via introduction of an additional positive charge into the active site region

Among bacterial topoisomerase I enzymes, a conserved methionine residue is found at the active site next to the nucleophilic tyrosine. Substitution of this methionine residue with arginine in recombinant Yersinia pestis topoisomerase I (YTOP) was the only substitution at this position found to induce the SOS response in Escherichia coli. Overexpression of the M326R mutant YTOP resulted in ~4 log loss of viability. Biochemical analysis of purified Y. pestis and E. coli mutant topoisomerase I showed that the Met to Arg substitution affected the DNA religation step of the catalytic cycle. The introduction of an additional positive charge into the active site region of the mutant E. coli topoisomerase I activity shifted the pH for optimal activity and decreased the Mg²⁺ binding affinity. This study demonstrated that a substitution outside the TOPRIM motif, which binds Mg²⁺directly, can nonetheless inhibit Mg²⁺ binding and DNA religation by the enzyme, increasing the accumulation of covalent cleavage complex, with bactericidal consequence. Small molecules that can inhibit Mg²⁺ dependent religation by bacterial topoisomerase I specifically could be developed into useful new antibacterial compounds. This approach would be similar to the inhibition of divalent ion dependent strand transfer by HIV integrase in antiviral therapy.     

The dynamic structure of the Escherichia coli cell envelope as probed by 15N nuclear magnetic resonance spectroscopy

Proton decoupled ¹⁵N NMR spectroscopy is shown to be a useful tool for probing the dynamic structure of the bacterial cell envelope. The proton decoupled ¹⁵N NMR spectra of Escherichia coli whole cells, cell envelopes and outer membranes were obtained and displayed resonances originating from protein side-chain groups, phosphatidylethanolamine, and peptidoglycan. Removal of phospholipids from the cell envelope resulted in a decrease in the motional freedom of peptidoglycan and cell envelope proteins. The mobility of the protein Arg side-chain groups is incresed in the absence of peptidoglycan. These data provide insights into the effect of supramolecular organization on the dynamic structure of the E. coli cell envelope.     

Effect of polyamines on synthesis and degradation of guanosine 5′-diphosphate 3′-diphosphate

The effect of polyamines on the in vitro and in vivo synthesis and degradation on guanosine 5′-diphosphate 3′-diphosphate (ppGpp) has been studied in Escherichia coli. The presence of 2 mM spermidine lowered the optimal Mg²⁺ concentration for ppGpp formation from 17 mM to 11 mM. The formation of ppGpp in the presence of 2 mM spermidine and 11 mM Mg²⁺ was about 15% greater than that in the presence of 17 mM Mg²⁺. At a concentration of less than 11 mM Mg²⁺, spermidine was found to stimulate ppGpp formation greatly. Putrescine did not cause any effect. When a polyamine-requiring mutant of E. coli (EWH319) was starved for an amino acid by the addition of valine, spermidine stimulated ppGpp formation. the degradation of ppGpp was not influenced significantly by polyamines.     

Bactericidal effect of Naja nigricollis toxin γ is related to its membrane-damaging activity

The aim of the present study is to investigate the causal relationship between membrane-damaging activity and bactericidal activity of Naja nigricollis toxin γ. Toxin γ showed a similar inhibitory activity on the growth of Staphylococcus aureus (Gram-positive bacteria) and Escherichia coli (Gram-negative bacteria). Antibacterial activity of toxin γ correlated positively with increase in membrane permeability of bacterial cells. Morphological examination showed that toxin γ disrupted the integrity of bacterial membrane. Toxin γ showed similar binding capability with lipopolysaccharide (LPS) and lipoteichoic acid (LTA), and destabilization of LPS layer and inhibition of LTA biosynthesis on cell wall increased bactericidal effect of toxin γ on E. coli and S. aureus, respectively. Although the potency of toxin γ on permeabilzing model membrane of E. coli and S. aureus was similar, the mode of interaction between toxin γ and model membrane of E. coli and S. aureus differed. Membrane-damaging activity of toxin γ was inhibited by either LPS or LTA. Nevertheless, LPS and LTA altered differently membrane-bound conformation of toxin γ. Taken together, our data suggest that bactericidal activity of toxin γ depends on its ability to induce membrane permeability, and that LPS and LTA structurally suppresses bactericidal effect of toxin γ.     

Enhancement of the direct antimicrobial activity of Lysep3 against <Emphasis Type="Italic">Escherichia coli</Emphasis> by inserting cationic peptides into its C terminus

Phage lysins are considered promising antimicrobials against resistant bacterial infections. Some lysins have been reported for the prevention and treatment of Gram-positive bacterial infection. Gram-negative bacterial phage lysins, however, can only destroy the bacterial cell wall from inside because of the obstruction of the bacterial outer membrane that prevents direct hydrolysis of the bacterial wall peptidoglycan from the outside, severely restricting the development of lysins against Gram-negative bacteria. In this study, genetic engineering techniques were used to fuse a 5 cationic amino acid polypeptide (KRKRK), a 10 cationic amino acid polypeptide (KRKRKRKRKR), a 15 cationic amino acid polypeptide (KRKRKRKRKRKRKRK), and a polypeptide including both cationic and hydrophobic amino acids (KRKRKFFVAIIP) to the C-terminus of the Escherichia coli phage lysin Lysep3 to obtain four fusion lysins (5aa, 10aa, 15aa, Mix). The bactericidal effects of those four lysins on E. coli were then compared in vitro. Our results showed that the fusion of hydrophobic and positively charged amino acids, Mix, can kill E. coli effectively; the fusion of positively charged amino acids alone at the C-terminus (5aa, 10aa, 15aa) also showed bactericidal activity against E. coli from the outside, with the bactericidal activity gradually increasing with the positive charge at the C-terminus of the lysin. Collectively, improving the positive charge at the C-terminus of E. coli bacteriophage lysin Lysep3 increases its bactericidal ability from outside E. coli, providing a new practical method for the development of anti-Gram-negative bacterial lysins.     

Analysis of bacterial RNase P RNA and protein interaction by a magnetic biosensor technique

A magnetic sensor technique was applied to analyze the interaction of immobilized bacterial RNase P protein and 3′-biotinylated RNase P RNA bound to streptavidin-coated magnetic beads. Our measurements with three types of beads from different suppliers resulted in K_d values of about 1–2 nM (at 4.5 mM Mg²⁺ and 150 mM NH₄⁺) for Escherichia coli RNase P RNA and protein, consistent with previous analyses using different techniques. We further measured affinity of the E. coli RNase P protein to chimeric RNase P RNA variants, consisting of an E. coli specificity domain and an engineered archaeal catalytic domain. A “bacterial-like” 1-bp insertion and 2-nt deletion in the helix P2/P3 region largely improved affinity, providing independent evidence that these elements are crucial for interaction of the two RNase P subunits. Moreover, our study documents that the properties of the streptavidin-coated magnetic beads decide on success or failure of the technique.     

Purification,kinetic studies,and homology model of Escherichia coli fructose-1,6-bisphosphatase

Previous kinetic characterization of Escherichia coli fructose 1,6-bisphosphatase (FBPase) was performed on enzyme with an estimated purity of only 50%. Contradictory kinetic properties of the partially purified E. coli FBPase have been reported in regard to AMP cooperativity and inactivation by fructose-2,6-bisphosphate. In this investigation, a new purification for E. coli FBPase has been devised yielding enzyme with purity levels as high as 98%. This highly purified E. coli FBPase was characterized and the data compared to that for the pig kidney enzyme. Also, a homology model was created based upon the known three-dimensional structure of the pig kidney enzyme. The k_cat of the E. coli FBPase was 14.6 s⁻¹ as compared to 21 s⁻¹ for the pig kidney enzyme, while the K_m of the E. coli enzyme was approximately 10-fold higher than that of the pig kidney enzyme. The concentration of Mg²⁺ required to bring E. coli FBPase to half maximal activity was estimated to be 0.62 mM Mg²⁺, which is twice that required for the pig kidney enzyme. Unlike the pig kidney enzyme, the Mg²⁺ activation of the E. coli FBPase is not cooperative. AMP inhibition of mammalian FBPases is cooperative with a Hill coefficient of 2; however, the E. coli FBPase displays no cooperativity. Although cooperativity is not observed, the E. coli and pig kidney enzymes show similar AMP affinity. The quaternary structure of the E. coli enzyme is tetrameric, although higher molecular mass aggregates were also observed. The homology model of the E. coli enzyme indicated slight variations in the ligand-binding pockets compared to the pig kidney enzyme. The homology model of the E. coli enzyme also identified significant changes in the interfaces between the subunits, indicating possible changes in the path of communication of the allosteric signal.     

MEMBRANE MODIFICATIONS IN NUTRITIONALLY INDUCED FILAMENTOUS ESCHERICHIA COLI B

Nutritionally induced filamentous cell forms of Escherichia coli B were examined for their morphological and biochemical lesions. The filamentous forms showed no significant alteration in total DNA concentration, RNA synthesis, ability to form β-galactosidase in response to isopropylthiogalactoside, or insensitivity to actinomycin D as compared to the normal cell form. The filamentous cells showed a marked decrease in the ability to incorporate N-acetylglucosamine-UL-¹⁴C into a phenol-soluble glycoprotein fraction relative to the normal cell form or relative to strain E-26 of E. coli grown in the filament-inducing medium. The filaments yielded an envelope-specific phenol-soluble protein fraction markedly reduced in or lacking three proteins as determined by acrylamide gel electrophoresis. Amino acid analysis, and chemical and enzymatic treatments of the envelope-specific phenol-soluble proteins showed striking differences between the fractions obtained from normal and filamentous cells. Electron microscope studies of divalent cation-induced aggregates of the envelope proteins showed different aggregation patterns dependent upon the cell form yielding the protein fraction.     

Interfacial regulation of bacterial sphingomyelinase activity

The objective of this study was to define how the quality of the buffer/membrane interface influences the activity of bacterial sphingomyelinase acting at the interface. The enzyme reaction was carried out in a zero-order trough using a surface barostat. This approach allowed for proper control of the physico-chemical properties of the substrate molecules. Since the molecular area of ceramide is smaller than that of sphingomyelin, the hydrolysis reaction could be followed `on-line' from the monolayer area decrease at constant surface pressure. The hydrolysis reaction could be divided into two separate phases, the first being the lag-phase (time between enzyme addition and commencement of the monolayer area change), and the second phase being the actual hydrolysis reaction (from which a maximal degradation rate could be determined). The activity of sphingomyelinase (Staphylococcus aureus) toward bovine brain sphingomyelin (bb-SM) was markedly enhanced by Mg²⁺ (maximal activation at 5 mM). Mg²⁺ also influenced the lag-phase of the reaction (the lag-time increased markedly when the Mg²⁺ concentration decreased below 1 mM). Saturated sphingomyelins (bb-SM and N-palmitoyl sphingomyelin [N-P-SM]) were more slowly degraded than the mono-unsaturated N-oleoyl sphingomyelin (N-O-SM). Both bb-SM and N-P-SM monolayers underwent a phase-transition at room temperature, whereas the N-O-SM monolayer did not. The phase-transition (liquid-expanded to liquid-condensed) was observed to greatly increase the lag-time of the hydrolysis reaction. The activity of sphingomyelinase was also sensitive to the lateral surface pressure of the monolayer membrane. Maximal degradation rate was achieved at 20 mN/m (with bb-SM, 30°C); above this pressure the lag-time of the reaction increased sharply. The inclusion of 4 mol% of cholesterol into a [³H]sphingomyelin monolayer markedly increased the extent of [³H]sphingomyelin degradation, and shortened the lag-time of the reaction. The inclusion of 10 mol% of zwitterionic or negatively charged phospholipids to the [³H]sphingomyelin monolayer did not affect the sphingomyelinase reaction significantly. In conclusion, this study has demonstrated that the physico-chemical properties of the substrate molecules have a dominating influence on the activity of a bacterial sphingomyelinase acting at the buffer/membrane interface.     

Effect of spermine on peptide-bond formation,catalyzed by ribosomal peptidyltransferase

The effect of spermine on the binding of AcPhe-tRNA to poly(U)-programmed ribosomes (step 1) and on the puromycin reaction (step 2) has been studied in a cell-free system, derived from E. coli.In the absence of ribosomal wash (FWR fraction) and at suboptimal concentration of Mg⁺⁺ (6 mM), spermine stimulated the binding of AcPhe-tRNA at least five fold, while at 10 mM Mg⁺⁺ there was a three fold stimulation. The above stimulatory effect was decreased at 6 mM Mg⁺⁺, or was abolished at 10 mM Mg⁺⁺ by the presence of FWR during the binding. Beside the stimulatory effect, spermine enhanced the stability of initiation complex AcPhe-tRNA-poly(U)-ribosome.In step 2, spermine affected the final degree of puromycin reaction and the activity status of peptidyltransferase. Both stimulatory and inhibitory effects have been observed, depending on the experimental conditions followed during the binding of the donor and during the peptide bond formation.     

Studies of the interactions of ursane-type bioactive terpenes with the model of Escherichia coli inner membrane—Langmuir monolayer approach

Pentacyclic triterpenes (PT), ursolic acid (Urs), and α-amyrin (AMalf) are natural products exhibiting broad spectrum of antibacterial activity. These compounds are membrane-active and can disorder bacterial membranes when incorporated; however, the exact mechanism of their membrane activity is unknown. In our studies, we applied Langmuir monolayer technique supported by Brewster angle microscopy to model the interactions of the selected PT with the lipid matrix of E. coli inner membrane. As the model membrane, we applied mixtures (75/25 mole/.mole %) of the representative Escherichia coli phosphatidylethanolamine (POPE), with the cardiolipin (ECCL) or phosphatidylglycerol (ECPG) extracted from the E. coli inner membrane. On the basis of the recorded isotherms, we performed thermodynamic analysis and calculated free energy of mixing ΔG^exc. It turned out that the phospholipids forming the inner membrane of E. coli are ideally miscible, whereas in binary systems composed of PT and POPE, negative deviations from ideality indicating attractive interactions between the investigated PT and POPE molecules were observed. On the other hand, in ternary systems composed of PT, POPE and one of the E. coli anionic phospholipids large positive changes in ΔG^exc were observed. Thus, both PT exhibit disorganizing effect on the model E. coli membrane. It was also proved that at low terpene proportion, AMalf can be more active than Urs. However, at higher proportion Urs incorporation can lead to the disintegration of cardiolipin-rich domains present in bacterial membrane.     

Acylation of monoacylglycerophosphoethanolamine in the inner and outer membranes of the envelope of an Escherichia coli K12 strain and its phospholipase A-deficient mutant

The site of the Escherichia coli envelope of the conversion of 1-acylglycero-3-phosphoethanolamine to diacylglycerophosphoethanolamine was explored, using two K12 strains with a wild-type phospholipid-degradative apparatus and a K12 mutant lacking detectable phospholipase A₁ and A₂ activity.Experiments with various radioactively labeled substrates show that acylation by crude envelope preparations as well as isolated inner and outer membranes of parent and mutant strains involves neither exogenous fatty acids nor a transacylation reaction with added monoacylglycerophosphoethanolamine. Furthermore, acylation exhibits no absolute requirement for added ATP and coenzyme A.Specific activity of acylating activity is the same in inner membrane preparations of parent and mutant strain and in outer membrane preparations of the mutant deficient in phospholipase A. Although clearly evident, net diacylglycerophosphoethanolamine formation by outer membranes of the parent strain, however, was about 6-fold less. This lower conversion may be attributed to activation during incubation of phospholipases A within the outer membrane, resulting in breakdown of the diacylcompound formed.Reacylation of lysophospholipids formed in the E. coli envelope by the action of endogenous or exogenous phospholipases A provides the organism with the potential of biochemically inexpensive repair and modification of the envelope phospholipids. Moreover, major phospholipids hydrolyzed in the outer membrane of E. coli can be resynthesized in the same location, without need for the transport of the products of hydrolysis to the lipid biosynthetic apparatus associated with the cytoplasmic membrane.     

Comparative Effects of Anesthetics on the Viability and Integrity of Escherichia coli ML30

Cells of Escherichia coli ML30 in a mineral salts medium were exposed to dichlorodifluoromethane (f-12), cyclopropane, halothane, or Ethrane at concentrations of 1.25, 0.2, 0.04, and 0.008× saturation for times up to 1,200 min, and at temperatures in the range of 2 to 37 C. When any of these anesthetics were applied for 300 min at 1.25× saturation, a substantial decrease in number of survivors occurred. Halothane was most bactericidal, cyclopropane and Ethrane were moderately bactericidal, and f-12 was least bactericidal. At saturation values of less than 1.0, none of the four anesthetics had an appreciable effect on viability of E. coli. Greatest increases in cell permeability occurred when anesthetics were used at saturation values of 1.25, and permeability changes generally decreased as the concentrations of the chemicals were reduced. In many instances, anesthetics in the vapor state caused significant increases in cell permeability but little or no loss of viability. This indicated that a close relationship did not exist between loss of viability and increased permeability. All four anesthetics caused E. coli to lose substantial and similar amounts of compounds absorbing at 260 nm. Release of compounds absorbing at 260 nm generally increased as the saturation value of a given chemical was increased. Halothane, Ethrane, and cyclopropane but not f-12 caused lysis of E. coli ML30. Considering all results, E. coli ML30 was damaged more by halothane or cyclopropane than by f-12 or Ethrane. When f-12 was applied at a saturation value of 1.25, the bactericidal effect on E. coli was much greater at 37 or 22 C than at 12 or 2 C.     

Mg2+-induced conformational changes in the btuB riboswitch from E. coli

Mg²⁺ has been shown to modulate the function of riboswitches by facilitating the ligand-riboswitch interactions. The btuB riboswitch from Escherichia coli undergoes a conformational change upon binding to its ligand, coenzyme B₁₂ (adenosyl-cobalamine, AdoCbl), and down-regulates the expression of the B₁₂ transporter protein BtuB in order to control the cellular levels of AdoCbl. Here, we discuss the structural folding attained by the btuB riboswitch from E. coli in response to Mg²⁺ and how it affects the ligand binding competent conformation of the RNA. The btuB riboswitch notably adopts different conformational states depending upon the concentration of Mg²⁺. With the help of in-line probing, we show the existence of at least two specific conformations, one being achieved in the complete absence of Mg²⁺ (or low Mg²⁺ concentration) and the other appearing above ∼0.5 mM Mg²⁺. Distinct regions of the riboswitch exhibit different dissociation constants toward Mg²⁺, indicating a stepwise folding of the btuB RNA. Increasing the Mg²⁺ concentration drives the transition from one conformation toward the other. The conformational state existing above 0.5 mM Mg²⁺ defines the binding competent conformation of the btuB riboswitch which can productively interact with the ligand, coenzyme B₁₂, and switch the RNA conformation. Moreover, raising the Mg²⁺ concentration enhances the ratio of switched RNA in the presence of AdoCbl. The lack of a AdoCbl-induced conformational switch experienced by the btuB riboswitch in the absence of Mg²⁺ indicates a crucial role played by Mg²⁺ for defining an active conformation of the riboswitch.     

Ability of Daphnia Cell-Free Extract to Damage Escherichia coli Cells

A cell-free extract of Daphnia magna was found to lyse Escherichia coli cells as shown by leakage of the enzymes alkaline phosphatase and β-galactosidase from the bacteria. The cell-free extract was separated on Sephadex G-200, and the fractions showing an ability to lyse E. coli cels were isolated. The factor which was responsible for the lysis of the bacterial cells was probably a protein with a molecular weight of several thousands. Mg²⁺ and Ca²⁺ ions augmented the activity of the Daphnia extract on E. coli cells.     

Enzymatic characterization of the enteropathogenic Escherichia coli type III secretion ATPase EscN

Type III secretion is a transport mechanism by which bacteria secrete proteins across their cell envelope. This protein export pathway is used by two different bacterial nanomachines: the flagellum and the injectisome. An indispensable component of these secretion systems is an ATPase similar to the F₁-ATPase β subunit. Here we characterize EscN, an enteropathogenic Escherichia coli type III ATPase. A recombinant version of EscN, which was fully functional in complementation tests, was purified to homogeneity. Our results demonstrate that EscN is a Mg²⁺-dependent ATPase (k_cat 0.35 s⁻¹). We also define optimal conditions for the hydrolysis reaction. EscN displays protein concentration-dependent activity, suggesting that the specific activity changes with the oligomeric state of the protein. The presence of active oligomers was revealed by size exclusion chromatography and native gel electrophoresis.