Critical role of C-terminal residues of the Alzheimer's associated β-amyloid protein in mediating antiviral activity and modulating viral and bacterial interactions with neutrophils

Recent studies have shown that the Alzheimer’s associated β-amyloid protein (βA) can inhibit growth of bacteria, fungi and viruses. We reported that the 42 amino acid βA protein inhibits replication of seasonal and pandemic strains of H3N2 and H1N1 influenza A virus (IAV) in vitro and modulates activation of neutrophils and monocytes exposed IAV. We here show that fragments composed of the N and C terminal domain of βA42, including βA22-42 and the 8 amino acid βA35-42, retain viral neutralizing and viral aggregating activity, whereas fragments lacking the C-terminal amino acids 41 and 42 (e.g. βA1-40, βA1-34, βA1-28, βA22-40 or βA33-40) have markedly diminished activities on these assays. βA22-42 also increased viral uptake, and virus induced respiratory burst responses, by human neutrophils, while peptides lacking residues 41 and 42 did not. Similar results were obtained with regard to bacterial aggregation, or promotion of bacterial uptake by neutrophils. Published structural studies have shown that βA1-42 has a greater propensity to form neurotoxic oligomers than βA1-40 due to a molecular interaction between Met35 and Ala42. Our findings suggest that there is a relationship between neurotoxic and antimicrobial activities of βA1-42. Truncated peptides containing the last 8 C-terminal amino acids of βA1-42 retain antimicrobial and opsonizing activities likely resulting from their ability to induce viral or bacterial aggregation.


Introduction
Accumulation of the Alzheimer's associated β-amyloid protein (βA) is believed to contribute strongly to the pathogenesis of Alzheimer's disease (AD), although the actual physiological function and reason for accumulation of βA in the brain are not known. βA peptides are fragments of the larger β amyloid transmembrane precursor protein (APP. βA1-40 is more abundant than βA1-42, but xA1-42 is the more amyloidogenic and neurotoxic species [1][2][3] neurotoxicity of βA1-42 has been shown to depend on the ability of this peptide to form unstable oligomers (pentamers mainly), whereas the protofibrils or fibrils formed from the peptide are less neurotoxic [4]. There are a variety of studies suggesting links between βA or AD and inflammation or infection [5]. Excess accumulation of βA has been linked to Alzheimer's disease but also to Human Immunodeficiency virus (HIV) related dementia [6], Herpes Simplex Virus (HSV) induced encephalitis [7], and cytomegalovirus infection. Examination of AD brain tissue has shown evidence of fungal infection. These findings suggest that viruses that infect the brain could be triggers for accumulation of βA, perhaps as part of an aberrant or sustained innate immune response.
The structure of βA resembles that of antimicrobial peptides like porcine protegrin and, like protegrin, it can form membrane channels [8]. Importantly, recent studies have demonstrated antibacterial and antifungal activity of βA peptides [9,10]. We showed that βA1-42 inhibited influenza A virus (IAV) through a mechanism that involves viral aggregation. βA1-40 had significantly reduced antiviral activity in these studies as compared to βA1-42. βA has also since shown to have antiviral activity against HSV [11]. Recently Kumar et al showed that transgenic expression of βA1-42 protects against bacterial and fungal infection in mice and worms through a mechanism that involves agglutination of these microbes [12]. They also demonstrated that mice lacking βA due to gene knockout had increased mortality from bacterial meningitis. Most recently, Spitzer et al showed that bacterial and fungal aggregating activities of βA1-42 are retained in peptides βA2-42 or 3-42, but lost in βA1-40, again suggesting that the final two amino acids of βA1-42 may be important to antimicrobial activity [13]. Overall these studies indicate that βA peptides have broad spectrum activity against various potential pathogens (viral, bacterial and fungal), perhaps providing the long sought physiological role for βA.
Antimicrobial peptides can also trigger recruitment and activation of immune cells [14]. βA peptides have been shown to activate glial cells and macrophages [15] in part through binding to toll-like receptor 2 (TLR2) as well as other receptors [16]. We demonstrated that βA1-42 modulates responses of neutrophils and monocytes to IAV [17]. βA1-42 increased neutrophil uptake of IAV and potentiated neutrophil respiratory burst and extracellular trap (NET) formation in response to the virus [17]. Of note, βA1-40 lacked the ability to either increase neutrophil uptake or neutrophil respiratory burst responses to IAV. These findings suggest that βA1-42 can modulate phagocyte responses to pathogens, and that this property also depends on the C-terminal amino acids Ile41 and Ala42.
In this paper we extend on our prior findings to determine the key antiviral domains of βA1-42. Through study of additional peptide fragments of βA1-42 we confirm the importance of Ile41 and Ala42 in mediating antiviral, bacterial aggregating, and neutrophil modulating effects of the protein. In addition we show that significantly shortened peptides containing these amino acids retain or exceed antiviral and/or neutrophil activating activities of βA1-42.

Ethics statement
Blood collection for isolation of neutrophils and monocytes was done with informed consent as approved by the Institutional Review Board of Boston University School of Medicine. The Institutional Review Board specifically approved this study and also approved the consent form for the study. The blood donors were healthy volunteers and they all signed the written consent form prior to each donation.

Virus preparation
Philippines 82/H3N2 (Phil82) strain was kindly provided by Dr. E. Margot Anders (Univ. of Melbourne, Melbourne, Australia) and grown in the chorioallantoic fluid of ten day old chicken eggs and purified on a discontinuous sucrose gradient as previously described [18]. The virus was dialyzed against PBS to remove sucrose, aliquoted and stored at -80˚C until needed. Post thawing the viral stocks contained~5x10 8   . The endotoxin assays were done using the highest concentrations of βA peptides used in the assays in the paper (50μg/ml). The highest concentration found was 2.2 EU/ml in βA22-42 which %1.4 pg/ml of endotoxin. Concentrations considerably higher than this did not have any effect on neutralization, aggregation or neutrophil functional assays in control experiments using purified endotoxin (n!4 experiments for each assays). Table 1 shows the peptides used in this paper.

Fluorescent focus assay of IAV infectivity
MDCK cell monolayers were prepared in 96 well plates and grown to confluence. These layers were then infected with diluted IAV preparations for 45 min. at 37˚C in PBS. MDCK cells were tested for presence of IAV infected cells after 18 hours of virus addition using a monoclonal antibody directed against the influenza A viral nucleoprotein (Millipore, Billerica, Ma) as previously described. IAV was pre-incubated for 30 min. at 37˚C with various concentrations of βA or control buffer, followed by addition of these viral samples to the MDCK cells.

Measurement of viral and bacterial aggregation by βA peptides
Viral aggregation caused by βA was measured by assessing light absorbance at 350nM by suspensions of IAV. This was done using a Perkin Elmer Lambda 35 UV/Vis spectrophotometer as described [19]. Viral aggregation was also assessed by electron microscopy (EM) as previously described [17]. Rhodamine labeled E.coli was incubated with βA preparations and then examined by fluorescent microscopy [20].

Confocal microscopy
For these experiments the IAV was labeled with Alexa Fluor 594. Alexa Fluor 594 carboxylic acid, succinimidyl ester labeling kit was purchased from Molecular Probes and labeling was carried out using manufacturer's recommendations with some modifications. In brief, concentrated virus stock was incubated with the Alexa Fluor in sodium bicarbonate buffer (pH 8.3) for one hour at room temperature. The preparation was then dialyzed overnight against PBS at 4˚C. After this procedure there was no reduction in viral hemagglutination titer. MDCK cells were pre-incubated with the labeled virus for 45 min., followed by washing and fixation using 1% paraformaldehyde. Prior to this the IAV was either pre-incubated with control buffer or βA for 30 minutes at 37˚C in the same manner as in the infectious focus assay. Wheat germ agglutinin (WGA)-Oregon Green 488 (4μg/ml) and DAPI 350 were used to stain the cell membrane and nucleus respectively. Confocal pictures were taken at Zeiss LSM510 (LSEB) on 100x resolution.

Human neutrophil and monocyte preparation
Neutrophils from healthy volunteers were isolated to > 95% purity by using dextran precipitation, followed by Ficoll-Paque gradient separation for the separation of mononuclear cells (layering above the Ficoll-Paque) and neutrophils (below the Ficoll-Paque). The neutrophils were purified further by hypotonic lysis to eliminate any contaminating erythrocytes, as previously described [21]. Cell viability was determined to be >98% by trypan blue staining. The isolated neutrophils were re-suspended at the appropriate concentrations in control buffer (PBS) and used within 2 hours.

Measurement of IAV and bacterial uptake by neutrophils
Fluorescein isothiocyanate (FITC)-labeled IAV (Phil82 strain) was prepared and uptake of virus by neutrophils was measured by flow cytometry as described [22]. In brief, IAV was treated with various doses of βA peptides for 30 min at 37˚C. Then it was incubated with cells for 45 minutes at 37˚C in presence of control buffer. Trypan blue (0.2 mg/ml) was added to these samples to quench extracellular fluorescence. Following washing, the neutrophils were fixed with 1% paraformaldehyde and neutrophil associated fluorescence was measured using flow cytometry. The mean cell fluorescence (>2000 cells counted per sample) was measured. Similar method was used for measuring neutrophil uptake of rhodamine labeled E.coli or S.aureus.

Measurement of neutrophil H 2 O 2 production
H 2 O 2 production was measured by assessing reduction in scopoletin fluorescence as previously described [23]. Measurements were made using a POLARstar OPTIMA fluorescent plate reader (BMG Labtech, Durham NC).

Statistics
Statistical comparisons were made using Student's paired, two-tailed t test or ANOVA with post hoc test (Tukey's). ANOVA was used for multiple comparisons to a single control. Statistics were performed using Excel and Statmost programs.

Neutralization of IAV strains by βA peptides
As shown in Fig 1, several βA peptides significantly inhibited infectivity of a seasonal H3N2 strain of IAV (Phil82), including βA1-42, βA22-42, and βA35-42. βA1-40, βA1-34, βA1-28, βA22-40 and βA33-40, had no or minimal antiviral activity in these experiments. We previously showed that βA1-40 has lower inhibitory activity than βA1-42 for this strain [17]. These results suggest that the C-terminal amino acids Iso41 and Ala42 are critical for antiviral activity. It is notable that the 8 amino acid peptide βA35-42 retained antiviral activity. We tested all of the peptides to determine if they altered cell viability using and LDH assay and no increase in LDH release was noted for any peptide at a 50μg/ml concentration (S1 Fig).

Aggregation of IAV by βA peptides
As shown in Fig 2A and 2B, βA peptides 1-42, 35-42, and 22-42 also caused aggregation of IAV particles, whereas the remaining peptides did not. βA22-42 had the highest activity in this assay. By comparison, βA peptides that lacked the last two amino acids did not cause aggregation of IAV. We also evaluated viral aggregation using electron microscopy (EM) as shown in Fig 2C. Results obtained with surfactant protein D (SP-D) which has well established ability to C-terminal residues of β-amyloid are critical for its antimicrobial peptide activities aggregate IAV and bacteria are presented for comparison in Fig 2B. βA22-42 caused marked viral aggregation, whereas βA22-40 did not. LPS did not cause any viral aggregation (Fig 2C  panel D). We also tested viral aggregation under the conditions of the fluorescent focus assay using confocal microscopy. Aggregates were visible in cases where Alexafluor labeled virus was pre-incubated with βA22-42 but not when pre-incubated with βA1-40 (Fig 3).

Effect of βA peptides on neutrophil H 2 O 2 responses to IAV
H 2 O 2 production by neutrophils was measured based on its ability to quench the fluorescence of scopoletin. Cells stimulated with IAV alone produced more H 2 O 2 than cells in PBS buffer alone. Pre-incubation of IAV with 16μg/ml of βA22-42 caused increased H 2 O 2 formation as compared to IAV alone and slightly increased production compared to IAV pre-incubated with βA22-40 (Fig 5A). βA 35-42 (not shown) and the β-amyloid peptides lacking the last two C-terminal amino acids did not increase H 2 O 2 production (Fig 5B). For comparison βA1-42 was shown to significantly increase the H 2 O 2 response as previously reported (Fig 5B).

Effect of βA peptides on E.coli aggregation and uptake of E.coli by neutrophils
To confirm our findings using a bacteria instead of virus we tested the ability of the peptides to aggregate or increase neutrophil uptake of E.coli. As shown in Fig 6, βA1-42 and βA 22-42 caused aggregation of fluorescent E.coli but βA22-40 did not. We also tested whether βA peptides could increase uptake of bacteria by neutrophils. As shown in Fig 7A, βA1-42 caused significant increase in uptake of fluorescently labeled E.coli. βA1-42 also increased uptake of fluorescently labeled S.aureus (i.e., 32 and 64μg/ml of βA1-42 increased uptake to 171±23 and 165±24% of control; p<0.03 for either one; n = 7). Since the initial experiments shown in Fig  7A appeared to show a continued rise at the highest concentration of peptides used we tested the remaining peptides at higher concentrations as shown in Fig 7B. In these experiments βA22-42 caused a marked increase in uptake of E.coli which greatly exceeded the effect seen with βA22-40. No significant effects were seen with either βA1-34, βA1-28 or βA33-40 (Fig 7A  and 7B). However, some increase in E.coli uptake was seen with 32μg/ml of βA35-42 (Fig 7A).

Discussion
These studies confirm the importance of the N-terminal amino acids Ile41 and Ala42 in the antiviral and antibacterial actions of βA. This was demonstrated by comparing activities of various peptides which contain (or did not contain) these amino acids. These residues are structurally important since they allow βA peptides to form the C-terminal loop between Met35 and Ala42 and this loop has been found to be critical in mediating oligomerization of βA peptides [4,24].  [25,26]. βA1-42 is more neurotoxic than βA1-40 and its concentration is selectively increased in amyloid fibrils in the plaques of Alzheimer's disease patients [1]. The greater neurotoxicity of βA1-42 as compared to βA1-40 is felt to relate to the greater propensity of the full length peptide to form neurotoxic oligomers. Analysis of small fragments of amino acids corresponding to βA 35-40 (MVGGVV), βA37-42 (GGVVIA), and βA35-42 (MVGGVVIA) by Wagoner et al using simulation kinetics and energetics revealed the increased propensity of βA35-42 assembled into unorganized oligomers as well [24]. The findings indicated that this N-terminal domain of protein is involved in self-aggregation. This self-aggregating property also appears to allow for aggregation of pathogens. Overall our findings and recent findings of Kumar and Spitzer [12,13] suggest the hypothesis that the very property that makes βA1-42 neurotoxic also mediate antiviral and antibacterial or antifungal activities.
Our results also indicate that the two C-terminal amino acids of βA1-42 contribute to its ability to modulate pathogen interactions with neutrophils. βA22-42 had particularly potent abilities to increase neutrophil uptake of IAV or E.coli, exceeding that of the full length peptide. In addition, the various N-terminal fragments of βA that did not include the C-terminus lacked the ability to increase viral uptake or neutrophil respiratory burst responses. βA35-42   [4]. Panel A shows a schematic of the structure of βA1-42 including the two intramolecular bonds present in the molecule. Panel B shows how the C-terminal loop formed by the bond between Ala42 and Met35 is involved in self aggregation of the peptide. Panel C shows atomic force microscopic images of oligomers formed by βA1-42.
https://doi.org/10.1371/journal.pone.0194001.g008 also increased slightly neutrophil uptake of IAV and bacteria, while βA33-40 did not. βA1-42 contains another turn formed by a salt bridge between Lys28 and Asp23 forming a β-turn-β conformation in the full length peptide (see Fig 8A). This salt bridge would be retained in βA22-42 but not in βA35-42. This may explain the enhanced antiviral activity of βA22-42 compared to βA35-42. βA1-42 has been reported to bind to some innate immune receptors on phagocytes (e.g. formyl peptide receptor 2, TLR2 and RAGE) [16,27] and future studies could determine if the shortened fragments of βA retain the ability to bind these receptors.