ApoE, a 34 kDa heterodimeric glycoprotein, is a well-known lipid-binding protein that plays a main role in the metabolism and transport of lipids (1An apolipoprotein preferentially enriched in cholesteryl ester-rich very low density lipoproteins.). ApoE is primarily produced by the liver and macrophages in the peripheral tissues and by astrocytes in the central nervous system, where apoE plays a critical role in cholesterol homeostasis (2Martinez-Martinez A.B. Torres-Perez E. Devanney N. Del Moral R. Johnson L.A. Arbones-Mainar J.M. Beyond the CNS: the many peripheral roles of APOE.). In humans, apoE occurs in three isoforms: E2, E3, or E4 (3Weisgraber K.H. Innerarity T.L. Mahley R.W. Abnormal lipoprotein receptor-binding activity of the human E apoprotein due to cysteine-arginine interchange at a single site.). Although the isoforms differ from each other by just one or two amino acids at positions 112 and 158, their structures and functions vary widely (4Apolipoprotein E: structure-function relationships., 5Tudorache I.F. Trusca V.G. Gafencu A.V. Apolipoprotein E - a multifunctional protein with implications in various pathologies as a result of its structural features.). E3 is the most common variant (78%) in the human population and has been well characterized in terms of structure and function (6Davignon J. Gregg R.E. Sing C.F. Apolipoprotein E polymorphism and atherosclerosis., 7Davignon J. Cohn J.S. Mabile L. Bernier L. Apolipoprotein E and atherosclerosis: insight from animal and human studies., 8Mahley R.W. Weisgraber K.H. Huang Y. Apolipoprotein E: structure determines function, from atherosclerosis to Alzheimer's disease to AIDS.). The E4 isoform (14%) has been linked to multiple diseases such as atherosclerosis, Alzheimer’s disease (AD), and multiple sclerosis (6Davignon J. Gregg R.E. Sing C.F. Apolipoprotein E polymorphism and atherosclerosis., 9Wishart H.A. Saykin A.J. Rabin L.A. Santulli R.B. Flashman L.A. Guerin S.J. Mamourian A.C. Belloni D.R. Rhodes C.H. McAllister T.W. Increased brain activation during working memory in cognitively intact adults with the APOE epsilon4 allele.). The least common isoform is E2 (8%), which is connected to hyperlipidemia (6Davignon J. Gregg R.E. Sing C.F. Apolipoprotein E polymorphism and atherosclerosis., 10Mahley R.W. Huang Y. Rall Jr., S.C. Pathogenesis of type III hyperlipoproteinemia (dysbetalipoproteinemia). Questions, quandaries, and paradoxes.).More recently, various peptides from the receptor-binding region of apoE (amino acid residues 130–150), which is essential for its biological function in lipid metabolism, have been shown to exert both antiviral and antibacterial activity (14Dobson C.B. Sales S.D. Hoggard P. Wozniak M.A. Crutcher K.A. The receptor-binding region of human apolipoprotein E has direct anti-infective activity., 15Forbes S. McBain A.J. Felton-Smith S. Jowitt T.A. Birchenough H.L. Dobson C.B. Comparative surface antimicrobial properties of synthetic biocides and novel human apolipoprotein E derived antimicrobial peptides., 16Pane K. Sgambati V. Zanfardino A. Smaldone G. Cafaro V. Angrisano T. Pedone E. Di Gaetano S. Capasso D. Haney E.F. Izzo V. Varcamonti M. Notomista E. Hancock R.E. Di Donato A. et al.A new cryptic cationic antimicrobial peptide from human apolipoprotein E with antibacterial activity and immunomodulatory effects on human cells., 17Zanfardino A. Bosso A. Gallo G. Pistorio V. Di Napoli M. Gaglione R. Dell'Olmo E. Varcamonti M. Notomista E. Arciello A. Pizzo E. Human apolipoprotein E as a reservoir of cryptic bioactive peptides: the case of ApoE 133-167.). This region is responsible for the binding of apoE to the LDL receptor family, and also, within this sequence, residues 142–147 (the heparin-binding domain) mediate attachment of apoE to cell surfaces (18Clay M.A. Anantharamaiah G.M. Mistry M.J. Balasubramaniam A. Harmony J.A. Localization of a domain in apolipoprotein E with both cytostatic and cytotoxic activity.). In particular, some apoE-derived peptides were shown to be effective against multiresistant bacteria and reduced lipopolysaccharide (LPS)-induced storm of cytokines in THP-1 cells (16Pane K. Sgambati V. Zanfardino A. Smaldone G. Cafaro V. Angrisano T. Pedone E. Di Gaetano S. Capasso D. Haney E.F. Izzo V. Varcamonti M. Notomista E. Hancock R.E. Di Donato A. et al.A new cryptic cationic antimicrobial peptide from human apolipoprotein E with antibacterial activity and immunomodulatory effects on human cells., 17Zanfardino A. Bosso A. Gallo G. Pistorio V. Di Napoli M. Gaglione R. Dell'Olmo E. Varcamonti M. Notomista E. Arciello A. Pizzo E. Human apolipoprotein E as a reservoir of cryptic bioactive peptides: the case of ApoE 133-167.). Therefore, in addition to their anti-infective activity, these peptides also show immunomodulatory activity (19Wang C.Q. Yang C.S. Yang Y. Pan F. He L.Y. Wang A.M. An apolipoprotein E mimetic peptide with activities against multidrug-resistant bacteria and immunomodulatory effects.). Two peptides from the same region were shown to prevent neurodegeneration and restore cognition in an AD model in Drosophila melanogaster (20Sarantseva S. Timoshenko S. Bolshakova O. Karaseva E. Rodin D. Schwarzman A.L. Vitek M.P. Apolipoprotein E-mimetics inhibit neurodegeneration and restore cognitive functions in a transgenic Drosophila model of Alzheimer's disease.), thus mimicking the protective effects of the full-length protein.

In this work, we hypothesized that the full-length version of apoE may have an impact on bacterial survival. Here, we demonstrated for the first time the antibacterial activity of full-length apoE in vitro and in vivo. Using advanced imaging techniques, we showed that bacteria aggregate in the presence of apoE. We also detected the formation of high-molecular-weight complexes of apoE after binding to bacterial endotoxins. This interaction provides a molecular explanation for the neutralizing effect of apoE on endotoxins.

Materials and methods Bacterial strains

Escherichia coli (25922) and Staphylococcus aureus (29213) were purchased from the American Type Culture Collection. S. aureus SAP229 was kindly provided by Dr Roger Plaut (Division of Bacterial, Parasitic, and Allergenic Products, FDA, Bethesda, MD). Pseudomonas aeruginosa (PA01) was kindly provided by Dr B. Iglewski (University of Rochester), and P. aeruginosa XEN41 was purchased from PerkinElmer (Akron, OH).

 Endotoxins

LPS from E. coli (serotype 0111:B4, cat# L3024), LPS from P. aeruginosa 10 (cat# L8643), and lipoteichoic acid (LTA) from S. aureus (cat# tlrl-pslta) were purchased from Sigma-Aldrich. Lipid A from E. coli (serotype R515, cat# ALX-581-200-L002) was purchased form AH Diagnostics.

 Proteins and peptides

Human plasma apoE (cat# IHUAPOE) and human plasma apoA1 (cat# IRHPL0059) were purchased from Innovative Research. The thrombin-derived peptide GKY25 (GKYGFYTHVFRLKKWIQKVIDQFGE) (97% purity, acetate salt) was synthetized by AmbioPharm (Madrid, Spain).

 Animals

SKH-1 hairless and BALB/c male mice (Charles River Laboratories), 8–12 weeks old, were used for in vivo experiments. The animals were housed under standard conditions of light and temperature and had free access to standard laboratory chow and water.

 Viable count assayPotential antibacterial activity of apoE on E. coli, P. aeruginosa, and S. aureus was explored by incubating one colony overnight in 5 ml of Todd-Hewitt (TH) medium. The next morning, the bacterial culture was refreshed and grown to the mid-logarithmic phase (absorbance at 620 nm of 0.4). The bacteria were then centrifuged, washed, and diluted 1:1,000 in 10 mM Tris buffer at pH 7.4 to obtain an approximate concentration of bacteria amounting to 2 × 106 cfu/ml. Next, 50 μl of bacterial suspension was incubated with 1 and 5 μM of apoE, 5 μM of GKY25 (used as a positive control), or buffer control (10 mM Tris buffer at pH 7.4) for 2 h at 37°C. After 2 h, serial dilutions of the samples were plated on TH agar plates, incubated overnight at 37°C, and followed by colony counting the next day based on the following equation:

cfu/ml=coloniesxdilutionfactorvolume(perspot)onplate

Antimicrobial activity of apoE was also assessed by the viable count assay (VCA) after preincubation for 30 min with 200 and 500 μg/ml heparin (Sigma-Aldrich). Four independent experiments were performed for each bacterial strain (21Petrlova J. Petruk G. Huber R.G. McBurnie E.W. van der Plas M.J.A. Bond P.J. Puthia M. Schmidtchen A. Thrombin-derived C-terminal fragments aggregate and scavenge bacteria and their proinflammatory products., 22Puthia M. Butrym M. Petrlova J. Stromdahl A.C. Andersson M.A. Kjellstrom S. Schmidtchen A. A dual-action peptide-containing hydrogel targets wound infection and inflammation.). Radial diffusion assayWe used E. coli, P. aeruginosa, and S. aureus for the radial diffusion assay (RDA). The bacteria were grown to the mid-log phase in 10 ml of TH medium, spun down, washed, and suspended in 10 ml of 10 mM Tris buffer, pH 7.4, as for VCA. This step was followed by the addition of bacteria (4 × 106 cfu) to 15 ml of under-lay agarose gel, consisting of 0.03% TH media, 1% (w/v) low-electroendosmosis-type agarose (Sigma-Aldrich), and 0.02% (v/v) Tween 20 (Sigma-Aldrich). The underlay was poured into a 144 mm diameter Petri dish. After solidification, 4 mm diameter wells were punched in the underlay, which were subsequently loaded with 6 μl of the buffer (negative control), 5 μM GKY25 (positive control), and apoE or apoAI in 10 mM Tris buffer at pH 7.4. The plates were thereafter incubated for 3 h at 37°C. Molten overlay gel (15 ml, 6% TH, and 1% low-electroendosmosis-type agarose in water) was added to the plate. We measured the antimicrobial activity of the peptides by measuring the radius of the clearing zone surrounding the wells after 18–24 h of incubation at 37°C (21Petrlova J. Petruk G. Huber R.G. McBurnie E.W. van der Plas M.J.A. Bond P.J. Puthia M. Schmidtchen A. Thrombin-derived C-terminal fragments aggregate and scavenge bacteria and their proinflammatory products., 22Puthia M. Butrym M. Petrlova J. Stromdahl A.C. Andersson M.A. Kjellstrom S. Schmidtchen A. A dual-action peptide-containing hydrogel targets wound infection and inflammation.). Mouse model of subcutaneous infectionOvernight cultures of bioluminescent P. aeruginosa Xen41 or S. aureus 229 were refreshed and grown to the mid-logarithmic phase in TH media. Bacteria were washed (5.6 × 1,000 rpm, 15 min) and incubated with apoE (5 μM) for 2 h or injected directly without preincubation. A total of 200 μl of the mixture (1 × 106 cfu/mouse) was injected subcutaneously into the dorsum of male SKH-1 hairless mice or shaved dorsum of BALB/c mice (8–12 weeks), which were anesthetized before injections, using a mixture of 2% isoflurane and oxygen. In vivo bacterial infection was then imaged by measuring bioluminescence in anesthetized mice in Vivo Imaging System (IVIS Spectrum, PerkinElmer Life Sciences), and the data obtained were analyzed using Living Image 4.0 Software (PerkinElmer). Five or six mice per treatment group were used. The animal model was previously described by Petrlova et al. and Puthia et al. (21Petrlova J. Petruk G. Huber R.G. McBurnie E.W. van der Plas M.J.A. Bond P.J. Puthia M. Schmidtchen A. Thrombin-derived C-terminal fragments aggregate and scavenge bacteria and their proinflammatory products., 22Puthia M. Butrym M. Petrlova J. Stromdahl A.C. Andersson M.A. Kjellstrom S. Schmidtchen A. A dual-action peptide-containing hydrogel targets wound infection and inflammation.). Blue native PAGE and Western blotTwenty-one microliters of ApoE (5 μM) was mixed with either 10 mM Tris as control or endotoxins (200 μg/ml final concentration). Samples were incubated for 30 min at 37°C before mixing with the loading buffer (4× loading buffer native gel, cat#BN2003, Life Technologies), and subsequent 28 μl was loaded onto 4–16% Bis-Tris Native Gels (cat#BN1002BOX, Life Technologies). Control experiments with 5 μM apoA1 and apoE were performed with or without 100 μg/ml LPS from E. coli and loaded onto gel after a 30-min incubation at 37°C. Samples were run in parallel with a marker (NativeMark Unstained Protein Standard, cat#LC0725, Life Technologies) at 150 V for 100 min. Gels were run in duplicates for each experiment: one for gel analysis after destaining from Coomassie and subsequent staining with GelCode Blue Safe Protein (cat# 1860983, Thermo Scientific), while the other was transferred to a 0.2 μm polyvinylidene fluoride membranes (Trans-Blot Turbo Transfer Pack, cat #1704156, Bio-Rad) via a Trans Turbo Blot system (Bio-Rad). Thereafter, the membrane was destained with 70% ethanol and blocked with 5% milk in 1× PBS-Tween (PBS-T) for 30 min at room temperature. The membrane was incubated with mouse mAb anti-human apoE (cat#ab1906, Abcam), at a concentration of 1 μg/ml diluted in 1% fat-free milk in 1× PBS-T, overnight at 4°C. ApoE and its high-molecular-weight complexes were then detected using a secondary rabbit anti-mouse polyclonal antibody that was conjugated to HRP (cat#P0260, Dako) (diluted 1:1,000 in 1× PBS-T complemented with 5% milk) after incubation for 60 min at room temperature. PBS-T was used to wash the membrane after each step (3 × 10 min), and the last wash after the secondary antibody was performed five times. The bands were revealed by incubating the membrane in the developing substrate (Super Signal West Pico PLUS Chemiluminescent Substrate, cat#34580, Thermo Scientific). Signal was acquired by a ChemiDoc (Bio-Rad) system. All the experiments were performed at least three times (21Petrlova J. Petruk G. Huber R.G. McBurnie E.W. van der Plas M.J.A. Bond P.J. Puthia M. Schmidtchen A. Thrombin-derived C-terminal fragments aggregate and scavenge bacteria and their proinflammatory products.). CD

CD measurements were performed using a Jasco J-810 spectropolarimeter equipped with a Jasco CDF-426S Peltier that was set to 25°C. Protein was diluted to 5 μM in 10 mM Tris at pH 7.4 and incubated with or without 200 μg/ml LPS from E. coli or P. aeruginosa, Lipid A from E. coli, LTA from S. aureus, or heparin (200 and 500 μg/ml). Measurements were performed after a 1–5 min or a 30 min incubation at room temperature in a 0.1 cm quartz cuvette. Scans were measured over the far-UV wavelength interval 200–260 nm, with a scan speed of 20 nm/min. An average of five scans for each sample was collected. The baseline (10 mM Tris buffer alone or with different ligands) was subtracted from the spectrum of each sample for normalization.

The α-helical content was calculated according to previously published equations (23Morrow J.A. Segall M.L. Lund-Katz S. Phillips M.C. Knapp M. Rupp B. Weisgraber K.H. Differences in stability among the human apolipoprotein E isoforms determined by the amino-terminal domain.) and reported below:

%α−helicalcontent=([θ]222+3000)(−39000+3000)

Where −3,000 and −39,000 have been established previously as constants based on the helicity of poly-L-lysine as described by (24Greenfield N. Fasman G.D. Computed circular dichroism spectra for the evaluation of protein conformation.):

[Θ]Molarellipticity=(θ222)×(MRW)(10×c×l)

Where θ222 = observed ellipticity at 222 nm in millidegrees, c = concentration in g/l, l = path length of the cuvette (cm), and MRW = mean residual weight, that is, molecular weight of the protein (Da)/(number of amino acids). All the experiments were performed at least three times (21Petrlova J. Petruk G. Huber R.G. McBurnie E.W. van der Plas M.J.A. Bond P.J. Puthia M. Schmidtchen A. Thrombin-derived C-terminal fragments aggregate and scavenge bacteria and their proinflammatory products., 25Petrlova J. Hansen F.C. van der Plas M.J.A. Huber R.G. Morgelin M. Malmsten M. Bond P.J. Schmidtchen A. Aggregation of thrombin-derived C-terminal fragments as a previously undisclosed host defense mechanism.). Transmission electron microscopyE. coli and P. aeruginosa were visualized using transmission electron microscopy (TEM) (Jeol JEM 1230; Jeol, Japan) in combination with negative staining after incubation with apoE (5 μM) or buffer as described in the VCA method. Images of endotoxins (200 μg/ml) in the presence or absence of apoE (5 μM) were taken after incubation for 120 min at 37°C. For the mounted samples, 10 view fields were examined on the grid (pitch 62 μm) from three independent sample preparations. Samples were adsorbed onto carbon-coated grids (Copper mesh, 400) for 60 s and stained with 7 μl of 2% uranyl acetate for 30 s. The grids were rendered hydrophilic via glow discharge at low air pressure (21Petrlova J. Petruk G. Huber R.G. McBurnie E.W. van der Plas M.J.A. Bond P.J. Puthia M. Schmidtchen A. Thrombin-derived C-terminal fragments aggregate and scavenge bacteria and their proinflammatory products., 25Petrlova J. Hansen F.C. van der Plas M.J.A. Huber R.G. Morgelin M. Malmsten M. Bond P.J. Schmidtchen A. Aggregation of thrombin-derived C-terminal fragments as a previously undisclosed host defense mechanism.). The size of aggregates was analyzed as the mean of gray value/μm ± SD by ImageJ 1.52k, after all the images were converted to 8 bit and the threshold was manually adjusted. Fluorescence microscopic analysis of live/dead bacteriaE. coli and P. aeruginosa viability in the aggregates was assessed by using LIVE/DEAD® BacLightTM Bacterial Viability Kit (Invitrogen, Molecular Probes, Carlsbad, CA). Bacterial suspensions were prepared as described above for the VCA. Strains were treated with 5 μM apoE, 5 μM GKY25, or 10 mM Tris at pH 7.4. After a 1 h incubation at 37°C, samples were mixed 1:1 with the dye mixture, followed by incubation for 15 min in the dark at room temperature. The dye mixture was prepared according to the manufacturer's protocol, that is, 1.5 μl of component A (SYTO 9 green-fluorescent nucleic acid stain) and 1.5 μl of component B (red-fluorescent nucleic acid stain propidium iodide) were dissolved in 1 ml of 10 mM Tris at pH 7.4. Five microliters of stained bacterial suspension was trapped between a slide and an 18 mm square coverslip. Ten view fields (1 × 1 mm) were examined from three independent sample preparations using a Zeiss AxioScope A.1 fluorescence microscope (objectives: Zeiss EC Plan-Neofluar 100×/1.3 oil and 40×; camera: Zeiss AxioCam MRm; acquisition software: Zeiss Zen 2.6 [blue edition]) (21Petrlova J. Petruk G. Huber R.G. McBurnie E.W. van der Plas M.J.A. Bond P.J. Puthia M. Schmidtchen A. Thrombin-derived C-terminal fragments aggregate and scavenge bacteria and their proinflammatory products.). In silico prediction of antimicrobial peptidesApoE was scanned for antimicrobial peptides (AMPs) using two in silico predictors: AmpGram (26Burdukiewicz M. Sidorczuk K. Rafacz D. Pietluch F. Chilimoniuk J. Rodiger S. Gagat P. Proteomic screening for prediction and design of antimicrobial peptides with AmpGram.) and a method for detection of “cryptic” AMPs, utilizing amino acid properties and peptide length developed by Pane et al. (27Pane K. Durante L. Crescenzi O. Cafaro V. Pizzo E. Varcamonti M. Zanfardino A. Izzo V. Di Donato A. Notomista E. Antimicrobial potency of cationic antimicrobial peptides can be predicted from their amino acid composition: application to the detection of “cryptic” antimicrobial peptides.). AmpGram is an AMP predictor, utilizing n-grams and a random forest classifier to predict AMPs with high accuracy while allowing for high-throughput proteomic screening. The AmpGram package (https://cran.r-project.org/web/packages/AmpGram/index.html) was used in R 4.0.2 (in the RStudio 1.3.1073 IDE). The model was applied to the FASTA sequence of apoE fetched from UniProt (P02649), and the results were visualized using the AmpGram package.Thereafter, the sequence was inserted into Microsoft Excel containing the algorithm presented by Pane et al. (REF) using the values m = 0.700 and n = 0.800. The helical property of the peptide with the highest antimicrobial score was visualized using NetWheels (28Mól A.R. Castro M.S. Fontes W. NetWheels: a web application to create high quality peptide helical wheel and net projections.). ApoE and the predicted AMPs were visualized in the NMR structure of human apoE (PDB accession 2L7B, https://www.rcsb.org/structure/2L7B) as a 3D model using PyMOL 2.3.4 (29

PyMOL, The PyMOL Molecular Graphics System, Version 2.0, Schrödinger, LLC; New York, NY.

). Fluorescence spectroscopyThe emission fluorescence spectra of tryptophan in apoE were measured between 300 and 450 nm, after excitation at 280 nm. Intrinsic fluorescence of 5 μM apoE (10 mM Tris at pH 7.4) was measured in a 3×3 mm quartz cuvette using a Jasco J-810 spectropolarimeter equipped with an FMO-427S fluorescence module, a scan speed of 200 nm/min, and a 2-nm slit width. The interactions between apoE and bacterial ligands were performed by measuring intrinsic fluorescence of protein at 25°C, immediately after addition of increasing concentrations of ligands (0–100 μg/ml). Then, the dissociation constant (Kd) was calculated from the spectral shift of maximum absorbance (λmax) as a function of the concentration of the bacterial ligands. The signal obtained for the protein alone was subtracted from all spectra. The results are expressed as an average of three independent experiments ± SEM (30Petruk G. Petrlova J. Samsudin F. Giudice R.D. Bond P.J. Schmidtchen A. Concentration- and pH-dependent oligomerization of the thrombin-derived C-terminal peptide TCP-25.). Statistical analysis

The graphs of the VCA, Kd (from fluorescence spectroscopy), and α-helical content (from CD measurements) are presented as the mean ± SD (VCA and α-helical content) or SEM (Kd) from at least three independent experiments. We assessed differences in these assays using the Brown-Forsythe ANOVA test for CD and one-way ANOVA with Dunnett’s multiple comparison tests for the VCA and Kd. All data were analyzed using GraphPad Prism (GraphPad Software, Inc.). In addition, P-values less than 0.05 were considered to be statistically significant (∗P < 0.05, ∗∗P < 0.01, ∗∗∗P < 0.001, and ∗∗∗∗P < 0.0001).

 Ethics statement

All animal experiments were performed according to the Swedish Animal Welfare Act SFS 1988:534 and were approved by the Animal Ethics Committee of Malmö/Lund, Sweden (permit numbers M88-91/14, M5934-19, M8871-19). Animals were kept under standard conditions of light and temperature and water ad libitum.

DiscussionResearch in the field of lipoproteins has the potential to discover new host defense molecules and the mechanisms by which they participate in the immune response to infection (12Daniels T.F. Killinger K.M. Michal J.J. Wright Jr., R.W. Jiang Z. Lipoproteins, cholesterol homeostasis and cardiac health., 13Plasma lipoproteins are important components of the immune system., 23Morrow J.A. Segall M.L. Lund-Katz S. Phillips M.C. Knapp M. Rupp B. Weisgraber K.H. Differences in stability among the human apolipoprotein E isoforms determined by the amino-terminal domain., 32van den Elzen P. Garg S. Leon L. Brigl M. Leadbetter E.A. Gumperz J.E. Dascher C.C. Cheng T.Y. Sacks F.M. Illarionov P.A. Besra G.S. Kent S.C. Moody D.B. Brenner M.B. Apolipoprotein-mediated pathways of lipid antigen presentation.). Several in vitro studies have shown that prototypic peptides with the sequence corresponding to the apoE receptor-binding region have strong antimicrobial activity (15