Effects of Secondary Metabolite Extract from Phomopsis occulta on β-Amyloid Aggregation

Inhibition of β-amyloid (Aβ) aggregation is an attractive therapeutic and preventive strategy for the discovery of disease-modifying agents in Alzheimer's disease (AD). Phomopsis occulta is a new, salt-tolerant fungus isolated from mangrove Pongamia pinnata (L.) Pierre. We report here the inhibitory effects of secondary metabolites from Ph. occulta on the aggregation of Aβ42. It was found that mycelia extracts (MEs) from Ph. occulta cultured with 0, 2, and 3 M NaCl exhibited inhibitory activity in an E. coli model of Aβ aggregation. A water-soluble fraction, ME0-W-F1, composed of mainly small peptides, was able to reduce aggregation of an Aβ42-EGFP fusion protein and an early onset familial mutation Aβ42E22G-mCherry fusion protein in transfected HEK293 cells. ME0-W-F1 also antagonized the cytotoxicity of Aβ42 in the neural cell line SH-SY5Y in dose-dependent manner. Moreover, SDS-PAGE and FT-IR analysis confirmed an inhibitory effect of ME0-W-F1 on the aggregation of Aβ42 in vitro. ME0-W-F1 blocked the conformational transition of Aβ42 from α-helix/random coil to β-sheet, and thereby inhibited formation of Aβ42 tetramers and high molecular weight oligomers. ME0-W-F1 and other water-soluble secondary metabolites from Ph. occulta therefore represent new candidate natural products against aggregation of Aβ42, and illustrate the potential of salt tolerant fungi from mangrove as resources for the treatment of AD and other diseases.


Introduction
Alzheimer's disease (AD) is a devastating condition leading to progressive cognitive decline, functional impairment and loss of independence, and is the major cause of dementia in the elderly worldwide [1]. Its prevalence will continue to increase as life expectancy increases. AD therefore represents a major and rising public health concern. However, as none of the medicines currently in use are able to cure this neurodegenerative disorder [2], understanding its etiology and developing new protective medicines have become the primary research goals in AD research.
Many clinicopathological studies have demonstrated that the deposition of beta-amyloid (Ab) peptides, fragments of the amyloid precursor protein (APP), in brain parenchyma and cerebral blood vessels is one of the hallmarks of AD [3,4]. Although the molecular mechanism of its involvement in the development and progression of AD is not clear, a critical role for Ab is universally acknowledged [5]. Ab fibrils were once thought to be the main molecular culprit in AD, but recent studies show a more decisive correlation between the levels of soluble, non-fibrillar Ab oligomers and the extent of synaptic loss and cognitive impairment [6][7][8]. Compared with Ab fibrils and plaques, Ab oligomers are more potent as neurotoxins that cause disruption of neuronal synaptic plasticity [9,10]. The relationships between Ab peptides, oligomerisation, cellular dysfunction and AD suggest that inhibition of Ab oligomerisation might lead to novel therapeutics for the treatment of AD [11].
Marine microorganisms are a source of potentially useful natural extracts for the treatment of multifaceted diseases such as AD [21,22], and we focus here on microbes associated with mangroves, which are salt-tolerant, woody trees that grow in coastal habitats. Recently, we isolated and identified a new salttolerant endophytic fungus, Phomopsis occulta SN3-2 (CCTCC No. 2011044), from mangrove Pongamia pinnata (L.) Pierre, and have assessed water-soluble secondary metabolites from Ph. occulta for inhibitory effects on the aggregation of Ab42 in mammalian cells and in vitro. Here we show that a bioactive fraction, ME0-W-F1, from Ph. occulta mycelia extract can reduce formation of high molecular weight (HMW) Ab42 oligomer and tetramer in vitro by inhibiting the formation of b-sheet secondary structure. Moreover, ME0-W-F1 is able to reduce the neurotoxic effect of Ab42 in SH-SY5Y cells. Culture of Phomopsis occulta and preparation of its secondary metabolite extracts Axenic cultures of Ph. occulta were maintained on potato dextrose agar. The cultures were transferred to liquid medium LB for 5-7 days, and then incubated in LB medium containing 0, 1, 2 or 3 M NaCl at 28uC without shaking for 40 days. These cultures were separated by filtration into mycelia and filtrates. The filtrates were concentrated to 2 L below 45uC in the dark, and extracted five times by shaking with an equal volume of ethyl acetate (EtOAc). After drying using anhydrous Na 2 SO 4 , collection and evaporation of EtOAc at 50uC in vacuo using a rotary evaporator (RV06-ML 1-B, IKA, Germany) yielded the fermentation broth extracts BE0, BE1, BE2 and BE3 (corresponding to cultures at 0, 1, 2 or 3 M NaCl, respectively). The mycelia were dried under vacuum and extracted three times using 2 L methanol for 72 h. Combination and evaporation of methanol yielded the mycelia extracts ME0,ME1, ME2 and ME3 (corresponding to cultures at 0, 1, 2 or 3 M NaCl, respectively). Escherichia coli cell model E. coli cell models of Ab aggregation have been developed by others previously [23][24][25]. Briefly, E. coli cultures capable of producing a secretable form of Ab42 fused to b-lactamase were grown overnight in LB supplemented with chloramphenicol (Cam) and then diluted 1:100 and grown for another 3 h at 37uC. These exponential phase cultures were diluted 1:50 in 96-well plates containing LB supplemented with 12.5 mg/mL Cam, 1 mM isopropyl-b-D-thiogalactopyranoside, 50 mg/mL ampicillin (Amp) and, as required, 200 mg/mL test samples, and EGCG was used as positive control (100 mg/ml). The plates were incubated at 37uC for 20 h without shaking. The OD 600 was read and relative growth rate (%) calculated according to the following formula.

Reagents
Here, D E.coli+Sample+Amp and D E.coli+sample represent the changes in OD 600 in E. coli cell and sample interaction systems after 20 h in the presence of Amp or not, respectively. D Sample+Amp and D Sample represent the changes in OD 600 in sample systems after 20 h in the presence of Amp or not, respectively. E. coli cells are normally killed by Amp because they are unable to export blactamase linked to aggregated Ab42 peptide. If Ab42 aggregation is inhibited, b-lactamase can be exported and degrade Amp, allowing cell growth.

Purification of active fractions and identification by TLC
For the active Ph. occulta secondary metabolite extract, extraction and column chromatography were used for further purification. The active fraction ME0 was distributed between nbutyl alcohol and water phases. The water soluble components, ME0-W, were separated by column chromatography filled with Diaion-20 resin. Methanol/water was used as mobile phase, and five fractions (ME0-W-F1 to F5; 0, 5, 10, 30 and 50% methanol/ water respectively (v/v)) were collected when the gradient elution was finished. Components soluble in n-butyl alcohol were not separated because of the absence of bioactivity. The inhibitory effect of these fractions on Ab42 aggregation was assessed using the E. coli model described above.
The bioactive fraction, ME0-W-F1 (10 ml), was applied to cellulose precoated (20620 cm) thin layer chromatography (TLC) plates (Merck, Germany). TLC plates were developed in a chloroform:methanol:water system (1:3:1 v/v), then air dried and visualized with iodine. Dried TLC plates were sprayed with ninhydrin reagent and heated at 80uC for 6 min. Peptide complexes became visible as intensely pink and purple-coloured bands and spots [26].

Flp-In T-REx 293 anti-aggregation assay
The Flp-In T-REx 293 (Invitrogen) cell line, a derivative of HEK293 cells containing a stably integrated FRT site and a TetR repressor, was maintained in DMEM media (Sigma D6171) supplemented with 10% fetal bovine serum (FBS), 5 mM Lglutamine, 5 mg/ml blasticidin. T-REx 293 cells were grown at 37uC under a 5% CO 2 atmosphere. The anti-aggregation screen was performed essentially as described [27]. Briefly 20,000 cells per well were seeded into a 24-well plate and allowed to attach for 48 h. Transient transfections were performed using GeneJammer (Agilent Technologies) as per manufacturer's instructions with either 0.75 mg of pcDNA3-Ab42-EGFP [27] or pATNRW20. The latter construct expresses an early onset familial form of Ab42 (Ab42E22G) fused with the fluorescent protein mCherry (pcDNA3.3-Ab42E22G-mCherry, N. Williamson et al., in preparation). ME0-W-F1 was added three hours post-transfection at the indicated concentrations, and gene expression was allowed to proceed for a further 48 h. An equivalent volume of dimethyl sulfoxide (DMSO) was used as a negative control and 10 mM epigallocatechin gallate (EGCG), a compound known to inhibit amyloid formation, was used as a positive control.
Quantification of Ab42 aggregates was performed as described previously [27]. Approximately 200 GFP (Ab42) or mCherry (Ab42E22G) positive cells were counted for each treatment and cells were scored as positive if they contained one or more aggregates. Images were acquired on an Olympus IX81 inverted wide field microscope and all experiments were performed in triplicate and odds ratio analysis of aggregation data was performed using the statistical package GraphPad Instat 3. The nature of Ab42 aggregates was also demonstrated by confocal microscopy, performed as described [27].
The inhibitory effect of ME0-W-F1 on Ab42 fibril formation was monitored by sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) under reducing conditions on 15% Tricine gels (Invitrogen) followed by Coomassie blue staining. In each experiment, Ab42 solution was incubated with ME0-W-F1 at 37uC and 8 ml samples were removed at various time points, then pooled and analyzed by SDS-PAGE. Gel band intensities were quantified using Quantitative One software (Bio-Rad).

Fourier transform infrared (FTIR) spectroscopy
Measurements and evaluation were as described [25]. Spectra were collected on a NICOLET-6700 (Thermo Nicolet, USA) spectrometer at room temperature using a CaF 2 cell with a 50 mm Teflon spacer. Ab42 stock solution (10 mg/mL in DMSO) was prepared according to section 2.7. ME0-W-F1 was prepared in DMSO at a concentration of 1 mg/mL. Mixtures were prepared by addition of Ab42 and ME0-W-F1 stock solutions to unbuffered D 2 O in a mass ratio of 1:1 and measurements taken at various time points. IR spectra were collected at 2 cm 21 resolution. Electrode readings were uncorrected for deuterium effects. CO 2 was removed and the air moisture inside the chamber was reduced by flushing the chamber with nitrogen gas. During each experiment, spectra were scanned 32 times over the range 4000-400 cm 21 . In some cases, the residual overlapping band was eliminated by subtraction from the final spectrum. The OMNIC software package (Thermo Nicolet, USA) was used for analysis of FT-IR spectra. Second derivative spectra were generated by using a 9-data point (9 cm 21 ) function included in the OMNIC software package.

Data analysis
The data were expressed as mean 6 SD, or mean of means 6 SE, and were evaluated by two-way analysis of variance (ANOVA) followed by a post hoc test, or t-test. P,0.05 was considered to be significant.  prepared separately from each culture and labeled according to salt concentration (i.e. BE0, BE1, ME0, ME1 etc.), then purified as described in Materials and Methods. The strategy is outlined in Figure 1A. Peptides were the main components of MEs, as shown by TLC and stains such as iodide ( Figure 1B) or ninhydrin ( Figure 1C).

Preparation of
The effect of Ph. occulta secondary metabolites on the aggregation of Ab42 was evaluated using an E. coli cell model. The fusion protein, ssTorA-Ab42-Bla, was expressed in E. coli. In the presence of samples with Ab42 aggregation inbititory effect, ssTorA-Ab42-Bla can be transported into the extracellular space and degrade Amp. Thus, E. coli growth is proportional to the inhibitory effect of samples on Ab42 aggregation [25]. In most cases, growth rates were higher in the presence of MEs than in the presence of BEs. This indicated an inhibitory effect of MEs on the aggregation of Ab42 in E. coli. Relative growth rates of E. coli cells with ME0, ME2 and ME3 were 57%, 98% and 48%, respectively, showing that all were at least as effective as the positive control, EGCG, which gave a relative growth rate of 42% (Figure 2A). The E. coli growth rate in the presence of ME1 was the same as that of the negative control (no additive), suggesting it had no effect on Ab42 aggregation.

Effect of ME0-W-F1 on Ab42-induced cytotoxicity in SH-SY5Y cells
An MTT assay in the neuronal cell line SH-SY5Y was employed to explore the cytoprotective activity of ME0-W-F1. We showed that ME0-W-F1 did not affect the viability of SH-SY5Y cells, even at concentrations up to 200 mg/mL (Figure 3). In contrast, exposure to freshly prepared Ab42 for 48 h was cytotoxic, producing a sharp decrease in SH-SY5Y viability, down to about 62% of control values. When ME0-W-F1 was added, however, the toxic effect of Ab42 was significantly reduced in a dose-dependent manner, with cell viability of 77%, 84% and 89% at 10 mg/mL, 100 mg/mL and 200 mg/mL, respectively ( Figure 3). Thus, ME0-W-F1 can reduce the cytotoxicity of Ab42 significantly in vitro, in a similar fashion to the positive control, EGCG.

ME0-W-F1 reduces aggregation of fluorescently tagged Ab42 in HEK293 cells
We have previously used Ab42 aggregation in human cells as a screening tool to identify small molecules with anti-aggregation activity [27]. The effect of ME0-W-F1 was therefore tested in HEK293 cells transiently transfected with genes encoding either Ab42-EGFP or Ab42E22G-mCherry; the latter is a mutant form of Ab42 associated with early onset AD. Both fluorescently tagged forms of Ab42 aggregated in the human cell line (Figure 4; Figure  S1). When ME0-W-F1 was added to cultures 3 h after transfection, however, the number of cells containing aggregates was reduced in an apparently dose-dependent manner. Therefore, ME0-W-F1 contains active components that can suppress aggregation of fluorescently tagged forms of Ab42 in human cells, similar to the findings in bacteria.
Inhibitory effect of ME0-W-F1 on Ab42 aggregation SDS-PAGE was used to investigate the effect of ME0-W-F1 on Ab42 aggregation in vitro. When Ab42 was incubated at 37uC for 7 d in the absence of ME0-W-F1, four main forms of Ab42 were visible on gels, i.e. monomer, dimer, tetramer and high weight molecular (HMW) oligomers, with the latter comprising about 50% of the material ( Figure 5). However, the proportion of Ab42 forming HMW oligomers was reduced to 32% and 7% after a 7 d incubation under the same conditions in the presence of low and high concentrations of ME0-W-F1, respectively ( Figure 5; Figure  S2 & S3). There was a corresponding, dose-dependent increase in the amount of Ab42 monomer when ME0-W-F1 was present, suggesting that the water soluble fraction has an inhibitory effect on Ab42 oligomerisation.
The formation of Ab42 aggregates is characterised by a shift in conformation of the protein secondary structure from a-helix to bsheet [30]. FT-IR spectroscopy allows this structural transition to be observed in the amide I band, 1600-1700 cm 21 , in which bands at ,1670 cm 21 and ,1627 cm 21 are characteristic of ahelix and b-sheet, respectively [31,32]. For Ab42 alone, there is a progressive shift from a-helix to b-sheets over a 4 d period ( Figure 6A). However, this transition is markedly reduced in the presence of ME0-W-F1 at 1 mg/mL ( Figure 6B), with the proportion of b-sheet reducing from about 44% to 63% ( Figure 6C). These data demonstrate that ME0-W-F1 can disrupt the transformation of a-helix to b-sheet associated with inhibition of the oligomerisation and aggregation of Ab42.

Discussion
Aggregation of Ab into plaques is a hallmark pathogenic feature of dementia and therefore is a primary target for amelioration of the disease [33]. Numerous chemical ligands have been developed as Ab aggregation inhibitors in recent years including EGCG [34], curcumin [35], scyllo-inositol [36] and LPFFD [37], but very few have progressed to clinical trials. In light of this disappointing situation, it is appropriate to search for alternative Ab aggregation inhibitors among natural products. Some herb and fungal extracts have remarkable anti-AD activities in vivo and in vitro due to inhibition of Ab aggregation [12][13][14][15][16][17][18][19][20]. Such studies justify further research on natural products, which could identify candidate lead compounds for AD treatment.
Evidence is accumulating that fungi are more likely to produce novel chemicals when they live in extreme environments [38]. Mangrove endophytic mycelium as a source of new microorganisms with potential pharmaceutical value has been intensively researched in recent years [39,40]. In this paper, the inhibitory The bioactivity of secondary metabolites from Ph. occulta is affected by the salt conditions during its growth. Thus, ME1, extracted from fungi grown at 1 M NaCl, has no clear growthpromoting, and hence anti-aggregation, effect in an E. coli model of Ab42 aggregation. This might be because Ph. occulta grows naturally in sea water, which contains about 0.75 M NaCl, and is less stressed at 1 M NaCl than lower or higher salt concentrations. In contrast, MEs from Ph. occulta grown at 0 M, 2 M and 3 M NaCl exhibited strong growth-promoting effects in the E. coli model, suggesting that certain secondary metabolites produced under such salt stress conditions have anti-aggregation activity. The water-soluble peptides in the selected bioactive fraction, ME0-W-F1, are candidates for such secondary metabolites. Because growth of Ph. occulta is very slow under high salt (2 M and 3 M NaCl) conditions, it was not possible to obtain sufficient quantities of material to test fractions such as ME2, which had strong bioactivity (Figure 2). Therefore, our analysis was limited to ME0 and related fractions. The BEs had no inhibitory effects on the aggregation of Ab42.
ME0-W-F1 is active in human cells, as well as the E. coli model, reducing the cytotoxicity of Ab42 in the SH-SY5Y cell line. Ab oligomerisation and fibril formation are toxic to neurons, and these processes mediate Ab toxicity mainly through interaction with other factors, e.g. Tau, in AD [5]. This suggests that ME0-W-F1 antagonises the oligomerisation and aggregation of Ab42. An effect of ME0-W-F1 on intracellular Ab42 aggregation was demonstrated in a HEK293 cell line, in which the water-soluble fraction reduced aggregation of both Ab42 expressed as a fusion protein with EGFP and also an early onset form, Ab42E22G, fused to mCherry; the fluorescent fusion partners allowed visualisation of aggregates within cells.
During the aggregation of Ab in vivo, it is suggested that native Ab peptides undergo conformational changes to form misfolded intermediates and various aggregated structures rich in b-sheet [41]. The transformation from a-helix to b-sheet is thought to be the rate-limiting step in the formation of soluble Ab intermediates and oligomers, which are the most toxic Ab species and are typically unstable, undergoing further aggregation to form higherorder oligomers and fibrillar deposits [7,42]. In vitro, ME0-W-F1 inhibits this structural transition from a-helix to b-sheet, as shown by both SDS-PAGE and FT-IR spectroscopy. Thus, SDS-PAGE demonstrated that the formation of tetramers and HMW oligomers of Ab42 was disrupted in the presence of ME0-W-F1 in a dose-and time-dependent manner. Similarly, FT-IR spectroscopy showed that the shift from a-helix to b-sheet as Ab42 aggregated was markedly reduced when ME0-W-F1 was present. These results suggest that ME0-W-F1 inhibits the oligomerisation and aggregation of Ab42 through blocking the transformation of secondary structure and preventing subunit assembly.
Since ME0-W-F1 prevented or reduced the aggregation of intracellular Ab42 fusion proteins in both bacteria and human cells, and since in vitro studies suggest an interaction with Ab42 species, it seems likely that the active components of the watersoluble fraction must gain access to the intracellular space. The most likely mechanism of action is that these components interfere with Ab42 aggregation within cells, as occurs in vitro, but we cannot rule out, for example, a stimulatory effect on molecular chaperone surveillance systems. The nature of the active components of ME0-W-F1 and the molecular mechanism of their action are currently under investigation.
In summary, water-soluble secondary metabolites from Ph. occulta exhibited inhibitory effects on the oligomerisation and aggregation of Ab42 in cells and in vitro. Therefore, ME0-W-F1 and Ph. occulta are novel natural materials worthy of further investigation as potential therapeutic agents for AD.