Amyloid-Like Adhesins Produced by Floc-Forming and Filamentous Bacteria in Activated Sludge

Abstract

Amyloid proteins (fimbriae or other microbial surface-associated structures) are expressed by many types of bacteria, not yet identified, in biofilms from various habitats, where they likely are of key importance to biofilm formation and biofilm properties. As these amyloids are potentially of great importance to the floc properties in activated sludge wastewater treatment plants (WWTP), the abundance of amyloid adhesins in activated sludge flocs from different WWTP and the identity of bacteria producing these were investigated. Amyloid adhesins were quantified using a combination of conformationally specific antibodies targeting amyloid fibrils, propidium iodide to target all fixed bacterial cells, confocal laser scanning microscopy, and digital image analysis. The biovolume fraction containing amyloid adhesins ranged from 10 to 40% in activated sludge from 10 different WWTP. The identity of bacteria producing amyloid adhesins was determined using fluorescence in situ hybridization with oligonucleotide probes in combination with antibodies or thioflavin T staining. Among the microcolony-forming bacteria, amyloids were primarily detected among Alpha- and Betaproteobacteria and Actinobacteria. A more detailed analysis revealed that many denitrifiers (from Thauera, Azoarcus, Zoogloea, and Aquaspirillum-related organisms) and Actinobacteria-related polyphosphate-accumulating organisms most likely produced amyloid adhesins, whereas nitrifiers did not. Many filamentous bacteria also expressed amyloid adhesins, including several Alphaproteobacteria (e.g., Meganema perideroedes), some Betaproteobacteria (e.g., Aquaspirillum-related filaments), Gammaproteobacteria (Thiothrix), Bacteroidetes, Chloroflexi (e.g., Eikelboom type 1851), and some foam-forming Actinobacteria (e.g., Gordonia amarae). The results show that amyloid adhesins were an abundant component of activated sludge extracellular polymeric substances and seem to have unexpected, divers functions.

Among the most important factors for activated sludge wastewater treatment plants (WWTPs) to operate well is the solid-liquid separation, which includes flocculation, clarification, settling, and dewatering (44). The outcome of these steps is to a large extent determined by the number of filamentous bacteria that may cause bulking (poor settling) if present in large numbers and the characteristics of the activated sludge flocs (37, 44, 69). Particularly for the flocs, the amount, composition, and properties of the extracellular polymeric substances (EPS) are assumed to be very important for flocculation and dewatering (40, 68). The EPS fraction consists mainly of a mixture of proteins, polysaccharides, lipids, nucleic acids, and humic substances (44). Proteins are the major component, but the detailed composition has not yet been adequately described.

Recently, we found that proteinaceous amyloid adhesins are an abundant component of the EPS fraction in biofilms from different habitats among these activated sludge flocs (32). Bacterial amyloids are usually thin fibrils, insoluble, and extremely resistant to various denaturants (8). This is primarily due to the chemical structure of the amyloid proteins, which are folded as beta sheets and stacked perpendicularly to the long fibril axis (60). They share many similarities to amyloid fibrils, which are known to cause neurodegenerative diseases such as Parkinson's and Alzheimer's disease in humans (6). Although many different bacteria are presumably able to produce amyloid adhesins (32), detailed descriptions of the size, amount, and properties of bacterial amyloids are restricted to pure culture studies of mainly Escherichia coli, Salmonella species, and the gram-positive Streptomyces coelicolor (48). Amyloid adhesins (named curli) produced by E. coli are 4 to 12 nm wide and 0.1 to 10 μm long (6, 15).

The bacterial amyloids in activated sludge and other biofilms are expressed by a broad range of phylogenetically distant species in the phyla Proteobacteria, Bacteroidetes, Chloroflexi, Firmicutes, and Actinobacteria (32). Generally, the function of amyloid fibrils is assumed to be related to enhanced adhesion to surfaces (49) and biofilm formation, but they may also increase resistance to chlorine (55) and resistance to chemical and enzymatic digestion (45). The function in activated sludge flocs is still unknown but may be related to the aggregation of microorganisms internally in microcolonies, whereas it is more uncertain what the function is in the filamentous bacteria, which were also shown to produce amyloids (32). Activated sludge flocs are often described as having a strongly and loosely bound fraction of cells and EPS (23, 57), so amyloids might be a good candidate for contributing to the high stability of the strongly bound fraction. It is, however, still unknown which species produce amyloids and what roles they have among the various functional groups in activated sludge and biofilm systems, such as nitrifiers, denitrifiers, polyphosphate-accumulating organisms (PAO), glycogen-accumulating organisms (GAO), and filamentous bacteria.

Detection of sessile bacteria producing amyloids can be performed by staining with thioflavin T (ThT) or by labeling with antibodies targeting a generic conformational epitope on amyloidal proteins (32). Antibodies have been found to be very specific for labeling of amyloid adhesins (32, 47), whereas ThT suffers from some nonspecificity, since it can also bind to cellulose and DNA (19, 51). As most bacteria in environmental biofilms are still uncultured, a combination of ThT or antibodies with fluorescence in situ hybridization (FISH) and oligonucleotide probes can be used to identify bacteria producing amyloid adhesins (32). This approach is particularly well suited for activated sludge systems, as most abundant bacteria can now be identified by culture-independent methods to species or genus level and thus are detectable by available oligonucleotide probes (28).

The aim of this study was to investigate the extent of amyloid adhesins in a range of activated sludge treatment plants and to identify the phylogenetic affiliation of microcolony-forming and filamentous bacteria producing amyloids by culture-independent methods.

MATERIALS AND METHODS

Sample collection and preparation.

Samples of activated sludge were collected from the aeration tanks of 43 different WWTPs treating both municipal and industrial wastewater. The samples were chosen from a large collection of activated sludge samples used in various other studies for FISH analysis. For each oligonucleotide probe tested together with antibody, two samples with high abundance of the target organisms (typically, 2 to 10% of the biomass) were chosen from the sample collection. Samples used with oligonucleotide probes targeting gram-negative bacteria were fixed with 4% paraformaldehyde for 3 h at 4°C, followed by washing in sterile-filtered (0.22-μm-pore-size polycarbonate filter) tap water. After the final washing step, the samples were resuspended in phosphate-buffered solution (PBS)-ethanol (diluted 1:1) and stored at −20°C. Samples used to target gram-positive bacteria were fixed with 50% ethanol and stored at −20°C. We observed the same degree of antibody labeling of a fresh sample and a 2-year-old fixed sample. For the quantitative analysis of the fraction of microorganisms producing amyloid adhesin in sludge, 10 out of the 43 WWTP samples were chosen. These samples were taken from five plants with and five plants without biological phosphorus removal. All sludge samples tested were gently homogenized to a level where the probe-defined microcolonies and filamentous bacteria were exposed as free floating in the bulk liquid or exposed on the surface of the activated sludge flocs. This ensured that problems with the access of antibodies to the EPS-embedded microcolonies were minimized.

Pure cultures.

Pure cultures of Haliscomenobacter hydrossis (DSM 1100) and Sphaerotilus natans (DSM 6575) were acquired from Deutsche Sammlung von Mikroorganismen (DSMZ), and Zoogloea ramigera Itzigzohn (ATCC 19544) was acquired from the American Type Culture Collection (ATCC). The media employed were DSMZ 134, DSMZ 51, and ATCC 1858 for H. hydrossis, S. natans, and Z. ramigera, respectively. Further details about growth conditions and medium description of the cultures are available from the DSMZ and ATCC. Paraformaldehyde-fixed samples of Chloroflexi morphotype 1851 strain EU25 (30) and M. perideroedes strain Gr1 (62) were used to support staining/labeling results from activated sludge samples.

Detection of amyloid adhesins in activated sludge samples.

Detection of amyloid adhesins was performed using conformationally specific antibodies (32). The protocol was the same on fresh and fixed samples, apart from the addition of 0.05% sodium azide to the blocking buffer when fresh samples were used. Briefly, two conformationally specific antibodies, WO1 and WO2 (47), were used. The sludge samples were homogenized using a glass tissue grinder (Thomas Scientific), centrifuged (9,500 × g for 5 min), and resuspended to an optical density at 650 nm of approximately 1 in PBS containing 1% (wt/vol) gelatin as a blocking buffer. The samples were preincubated at 37°C for 1 h, after which antibodies WO1 or WO2 and Tween 20 were added to a final concentration of 10 nM and 0.05% (wt/vol), respectively. The samples were incubated for 2 h at 37°C. The primary antibody was removed by centrifugation (9,500 × g for 5 min), and the pellets were washed once in 80 μl of PBS containing gelatin (1%, wt/vol) and 0.1% (vol/vol) Triton X-100. Incubation with a 1:256 dilution of fluorescein isothiocyanate-conjugated secondary immunoglobulin M antibody (Sigma), 0.025% Tween 20, and PBS containing gelatin (1%, wt/vol) was carried out for 1 h at 37°C. Finally, the samples were washed three times as described above but without gelatin in the washing buffer. After the final washing step, the samples were resuspended in 80 μl of PBS and stored at 4°C until microscopy. Antibody staining of fresh samples was supplemented with ThT staining (32).

FISH.

FISH was performed using 16S rRNA-targeted oligonucleotide probes (Tables 1 and 2) hybridized in suspended samples (14). Paraformaldehyde-fixed sludge samples were gently homogenized using a glass tissue grinder (Thomas Scientific). A total of 40 μl of homogenized sludge was pelleted by centrifugation for 5 min at 9,500 × g, washed, and resuspended in 80 μl of hybridization buffer. The sample was vortexed and pelleted again, followed by resuspension in 80 μl of hybridization buffer with the oligonucleotide probes used (final concentration, 50 ng of probe ml⁻¹). Oligonucleotide probes were labeled with 5(6)-carboxyfluorescein-N-hydroxy-succinimide ester (FLUOS) or with the sulfoindocyanine dyes (Cy3 and Cy5) (Biomers, Germany).

TABLE 1.

Identity of microcolony-forming bacteria in activated sludge samples examined with combination of oligonucleotide probes, antibody, and ThT

Functional group	Overall phylogenetic group	Phylogenetic affiliation	Gene probe(s)	Reference(s)	Antibody binding (WWTP 1/2)a	ThT binding (WWTP 1/2/3)b
All phylogenetic groupsc	Alphaproteobacteria	Alphaproteobacteria	ALF 968	42	+	ND
	Betaproteobacteria	Betaproteobacteria	BET42a	36	+	+
	Gammaproteobacteria	Gammaproteobacteria	GAM 42a	36	−	−
	Deltaproteobacteria	Deltaproteobacteria	SRB385+SRB385Db	50	−	ND
	Firmicutes	Firmicutes	LGC354mix	38	−	−
	Actinobacteria	Actinobacteria	HGC69a	52	+	+
Denitrifiers	Betaproteobacteria	Aquaspirillum-related bacteria	Aqs997	64	−/−	+/+/+
	Betaproteobacteria	Zoogloea ramigera	ZRA	54	−/−	+/ND/ND
	Betaproteobacteria	Thauera spp.	Thau646	31	−/−	−/+/ ND
	Betaproteobacteria	Most Azoarcus	Azo644	17	−/−	+/−/−
PAO	Betaproteobacteria	Rhodocyclus-related PAO	PAOmix: PAO462+PAO651+ PAO846	10	−/−	−/−
	Actinobacteria	Actinobacterial PAO	Actino-221+c1Actino-221+c2Actino-221, Actino-658+c1Actino-658+c2Actino-658	26	−/−	−/+/+
GAO	Alphaproteobacteria	Defluviicoccus spp.	DF2Mix: DF988+DF1020+H966 +H1038	39	−/−
	Gammaproteobacteria	Competibacter-related GAO	GAOmix: GAOQ989+GB_G2	9, 27	−/−	−/−
Nitrifiers	Betaproteobacteria	Betaproteobacterial ammonia-oxidizing bacteria	Nso190, Nso1225	41	−/−	−/−
	Betaproteobacteria	Nitrosomonas oligotropha lineage	Cluster 6a	1	−/−	−/−
	Nitrospira	Nitrospira	Ntspa662	11	−/−	−/−

^a−, no binding of antibody to probe-defined group; +, binding of antibody to probe-defined group determined in a sample from one WWTP.

^bND, not determined.

^cTested in only one sample.

TABLE 2.

Identity of filamentous bacteria in activated sludge samples examined with the combination of oligonucleotide probes and antibody

Overall phylogenetic group	Phylogenetic affiliation	Gene probe	Reference(s)	Antibody bindingb
				WWTP1	WWTP2
Class Alphaproteobacteria	“Candidatus Meganema perideroedes”	Meg983+Meg1028	62	+	+
	“Candidatus Monilibacter batavus”	MC2-649	58	−	−
	“Candidatus Sphaeronema italicum”	Nost993+Helper1010	29	+c	+c
	“Candidatus Alysiosphaera europaea”	Noli-644	58	+	+c
	“Candidatus Alysiomicrobium bavaricum”	PPx1002	29	−c	−
Class Betaproteobacteria	Aquaspirillum-related bacteria	Aqs997	64	+	+
	Sphaerotilus natans	SNA23a	65	−	+c,d
Class Gammaproteobacteria	Thiothrix nivea	TNI	65	−	+a
Phylum Bacteroidetes	Cytophaga-Flavobacterium	CFB319a+b	35	+	−
	Members of family Flavobacteriaceae	CFB 563	66	+	−
	Environmental isolates; most members of the class Bacteriodetes	CFB719	66	+	−
	Members of the family Saprospiraceae	Sap-309	56	−	−
	Haliscomenobacter hydrossis	HHY	65	−	−
Phylum Chloroflexi	Phylum Chloroflexi	CFXmix: CFX1223+GNSB941	4, 16	+	+
	Chloroflexi subdivision 3	CFX109	4	+	+
	Chloroflexi subdivision 1a-1b	CFX784	4	+	ND
	Eikelboom morphotype 1851	Chl1851	2	+	+
Candidate phylum TM7	Most of Candidate division TM7	TM7-905	18	−c	−c
Phylum Actinobacteria	Microthrix sp.	MPA-all-1410	34	−c	−c
	Skermania piniformis	Spin1490	13	−	−a
	Gordonia amarae	Gor596	12	−	−a

^aFor some filamentous bacteria, the combination of antibody and FISH did not work, but some of these filamentous bacteria could be targeted and identified by FISH alone. Labeling the same sample with antibody alone enabled identification of these filamentous bacteria by comparison with the morphology of filamentous bacteria in the FISH-probed sample.

^b−, no binding of antibody to probe-defined group; +, binding of antibody to probe-defined group; ND, not determined because of poor binding of oligonucleotide probe and no possibility of morphological identification.

^cIdentified only on the basis on morphology.

^dOnly empty sheaths were positive. Filaments containing FISH-positive bacteria were negative.

Combination of antibody and propidium iodide.

Samples labeled with primary and secondary antibodies were stained with propidium iodide by mixing equal volumes of sludge sample with a stock solution of propidium iodide (60 μM) (Molecular Probes, The Netherlands), followed by a 15-min incubation in the dark. Staining with propidium iodide was performed only on prefixed samples to ensure penetration of propidium iodide.

For quantification of amyloid adhesins in multispecies biofilms, the combination of propidium iodide and antibody proved to be superior to the combination of DAPI (4′,6′-diamidino-2-phenylindole) and antibody (32) in two ways. First, it avoided a labor-intensive relocation step, and, second, it ensured a higher quality of the images by using the same focal layer on the confocal laser scanning microscope (CLSM). Each sample was tested in replicates, and 10 images were acquired on each replicate.

Combination of antibody and FISH.

The activated sludge samples hybridized with oligonucleotide probes were fixed once again in 4% paraformaldehyde overnight. After this step the samples were washed twice in PBS, followed by labeling with primary and secondary antibodies as described above. This combination of FISH and antibody labeling is a slight modification of previous studies (32). The protocol used in this study leads to increased fluorescence signals from oligonucleotide probes and less interference from blocking agents.

Microscopy.

A CLSM (LSM 510 META; Carl Zeiss) equipped with an Ar ion laser (458 and 488 nm) and two HeNe lasers (543 and 633 nm) was used to record digital images for quantification of the antibody-positive fraction of the activated sludge samples. Images recorded as documentation of positive antibody labeling of different probe-defined groups were recorded using either this CLSM or a Zeiss Axioscop 2 epifluorescence microscope. The settings for the CLSM were calibrated on a negative control sample with only the secondary antibody.

RESULTS

Abundance of bacteria producing amyloid adhesins.

Amyloid adhesins in activated sludge originating from full-scale WWTPs were quantified by antibody and propidium iodide staining. As the process configuration of the treatment plants affects the microbial community composition and therefore also potentially the abundance of amyloids, five treatment plants with N removal and five plants with both N removal and enhanced biological phosphorus removal were selected. The amyloid-positive fraction in the different plants ranged from 10% to 40% of the total biovolume with consistent and similar results, using the two independent primary antibodies WO1 and WO2 (Fig. 1). In general, plants with biological phosphorus removal showed a higher level of binding of both antibodies (20 to 40% of the total biovolume). In all treatment plants, it was observed that certain microcolonies in the flocs produced amyloids, but filamentous bacteria were the dominant amyloid producers. Amyloids were primarily detected directly on the surface of the microorganisms, often as a thin layer or “cloud” outside the bacterial cells stained by propidium iodide (Fig. 2a and b). Some of the amyloid adhesins were also detected in the EPS fractions of the sludge, where the amyloids could not be directly linked to microorganisms. This appeared both as amorphous aggregates and some elongated tubes, probably partly degraded filamentous bacteria or their sheaths.

FIG. 1.

Fraction of microorganisms positive with antibodies (WO1 and WO2) in activated sludge from 10 different WWTPs. Values were determined by dual staining with antibody and propidium iodide. The error bars indicate the standard error. BioP, biological phosphorus.

FIG. 2.

Detection of amyloid adhesins in activated sludge by antibody or ThT staining in combination with propidium iodide (a and b) or oligonucleotide probes (c to f). Double staining with antibody WO1 (green) and propidium iodide (red) on paraformaldehyde-fixed sludge. (c) Simultaneous labeling of Aquaspirillum-related organisms with ThT (green) and oligonucleotide probes (Aqs997, red; EUBmix, blue). (d) ThT staining (green) of Zoogloea microcolony with finger-like morphology. (e) Simultaneous labeling of Actinobacteria-related PAO with ThT (green) and oligonucleotide probes (Actino658, red; EUBmix, blue). (f) Simultaneous labeling of Thauera with ThT (green) and oligonucleotide probes (Thau646, red; EUBmix, blue). Scale bar, 20 μm.

Identity of bacteria producing amyloid adhesins in microcolonies.

The identity of amyloid-producing microorganisms was first investigated using the broad oligonucleotide probes in combination with WO1, revealing some microcolony-forming bacteria among the Alpha- and Betaproteobacteria and Actinobacteria. Subsequently, a selection of probes targeting microcolony-forming nitrifiers, denitrifiers, PAO, and GAO were investigated for amyloid adhesins. However, all these specific probe-defined groups tested negative for amyloid adhesins by antibody labeling in all samples investigated (Table 1). The combination of antibody and oligonucleotide probes worked without problems in most cases. Only for probe Nsv443 targeting Nitrosospira spp. and probe NmV targeting Nitrosococcus mobilis was the FISH signal lost when it was used in combination with antibodies (results not shown).

In order to verify these results, the less specific dye ThT was also combined with FISH and applied to the most abundant probe-defined groups. Initially, each probe-defined group was analyzed in two WWTPs where they were abundant. If the results were different in these plants, one to two more plants were investigated to verify the results. ThT-positive bacteria were identified among the denitrifiers and the actinobacterial PAO, whereas all ammonium and nitrite-oxidizing nitrifiers were found to be ThT negative (Table 1). Among the denitrifiers, a ThT-positive subpopulation of Aquaspirillum-related bacteria was detected in all WWTPs (Fig. 2c). This population of coccoid cells formed characteristic spherical microcolonies. Thauera and Azoarcus tested both ThT positive and negative in the WWTPs investigated, indicating that these oligonucleotide probes target different subpopulations in different WWTPs or that these organisms express amyloid adhesins only in some cases (Fig. 2f shows an example of a ThT-positive Thauera microcolony). Zoogloea microcolonies were present in only some of the WWTPs and in low abundance, which complicated in situ analysis of this probe-defined group. Therefore, a pure culture of Z. ramigera was tested with ThT staining showing that Zoogloea was highly positive close to the cell membrane but negative in the large space between the cells in the characteristic finger-like microcolonies. The same ThT signal distribution was found on Zoogloea microcolonies when they were tested in situ (Fig. 2d). When the pure culture was tested with antibodies, no binding was found, indicating that the amyloids were masked by other exopolymers or that ThT stained nonamyloidal material. The ThT-positive actinobacterial PAO were identified with probe Actino-658 targeting bacteria with rod morphology (Fig. 2e). Tetrad-forming PAO targeted with probe Actino-221 were not positive with ThT in any of the WWTPs investigated.

Identity of filamentous bacteria producing amyloid adhesions.

Many filamentous bacteria producing amyloid adhesins were observed belonging to several phylogenetic groups, such as Alpha- and Betaproteobacteria, Bacteroidetes, Chloroflexi, and Actinobacteria (mycolata) (Table 2). The FISH signal was highly variable between each filament, which made quantification of the fraction of positive filaments within each probe-defined group impossible. Therefore, the result in Table 2 for each sample is given as positive or negative, indicating whether bacteria clearly positive with the oligonucleotide probe were also positive with antibody WO1.

Within the Alphaproteobacteria, antibody-positive filaments were identified among M. perideroedes, “Candidatus Sphaeronema italicum,” and “Candidatus Alysiosphaera europaea.” The amyloids of M. perideroedes were almost exclusively found on the tips of the filaments and never on the sheath (Fig. 3a). Interestingly, in some cases the antibody also bound to the septum dividing the two outermost cells at the tip but rarely between cells in the middle of the filament. This indicates either that the production of amyloid adhesins primarily took place by the cells at the end of the filaments or that the relatively large primary immunoglobulin M antibody was unable to penetrate through the sheath. To test these hypotheses, ThT staining was applied since the ThT molecule is much smaller than the antibody. A fresh sludge sample with M. perideroedes was stained and gave a very bright signal on all cells (Fig. 3d). This shows that M. perideroedes most likely contained amyloids on the surface of all cells underneath a sheath that prevented penetration and binding of antibodies. Among filaments of “Ca. Sphaeronema italicum” and “Ca. Alysiosphaera europaea,” approximately half were amyloid positive and half were negative in the same samples, suggesting that the oligonucleotide probe targeted different subspecies or that only some of the bacteria expressed amyloids at the time of fixation. Antibody was detected on the entire surface of the filament of “Ca. Sphaeronema italicum,” whereas “Ca. Alysiosphaera europaea” showed antibody binding only at the tips. It was not possible to get fresh samples of these species for ThT staining.

FIG. 3.

Detection of amyloid adhesins on filamentous bacteria in activated sludge by antibody and ThT staining. (a) Simultaneous labeling of M. perideroedes filaments with antibody (green) and oligonucleotide probes (red). (b) Bright-field micrograph of Thiothrix with sulphur inclusions. (c) Differential interference contrast microscopy of G. amarae. (d) Labeling of M. perideroedes in sludge with ThT. (e) ThT staining (green) of the same view shown in panel b. (f) ThT staining (green) of the same view shown in panel c. (g to i) Simultaneous labeling of filaments with antibody (green) and oligonucleotide probes (red): phylum Bacteroidetes (g), phylum Chloroflexi (h), and phylum Chloroflexi subdivision 3 (i). Scale bar, 20 μm.

Among the Betaproteobacteria, filamentous Aquaspirillum-related bacteria detected by probe Aqs997 were positive in both samples. Some of the filaments had attached cells that hybridized with probe Sap309 targeting the family Saprospiraceae, but these did not show any presence of amyloids. Probe-defined S. natans was difficult to investigate as the sludge samples mainly contained many empty sheaths, presumably from this organism, and these were amyloid positive (Table 2). However, this could not be verified as no amyloids could be detected in a pure culture of S. natans either with antibodies or with ThT.

The gammaproteobacterial Thiothrix showed a positive signal with antibody in one sample and was negative in the other sample. In the sample containing antibody-positive Thiothrix, the positive signal was detected only at the tips of the filaments forming rosettes. With ThT staining a positive signal was found on all the cells underneath the sheath. In some places, a single cell under the sheath was ThT negative although it was covered by a sheath, indicating that the amyloids were bound to the cells and not to the sheath (Fig. 3b and e).

In the phylum Bacteroidetes, most of the probes tested showed amyloid-positive filamentous bacteria. The probes for the family Flavobacteriaceae and the class Bacteroidetes revealed filaments with amyloids on the filament sheaths (Fig. 3g). Only the family Saprospiraceae, which includes the species H. hydrossis, did not express amyloid adhesins. A pure culture of H. hydrossis was also tested with both antibody and ThT without revealing the presence of any amyloids.

Most filamentous bacteria belonging to the phylum Chloroflexi expressed amyloids. This was observed using both the broad and the more specific oligonucleotide probes. Chloroflexi filaments belonging to Eikelboom type 1851 (positive with the specific probe Chl1851) had amyloids. A pure culture strain of this type 1851 (strain EU25) also tested amyloid positive. Filaments from the broader subdivision 3 (probe CFX109), which were negative with probe Chl1851, were also amyloid positive. Many filamentous bacteria belonging to this subdivision expressed amyloids along the entire filament, despite the fact that, in some cases, the filaments showed regions without any cells or signals from the oligonucleotide probe. This suggested that a sheath with amyloids persisted after death of the cells (Fig. 3i). The combination of probe CFX784 and antibodies was unsuccessful, so a direct link between the production of amyloid adhesins and Chloroflexi subdivision 1a-1b was impossible. However, as some amyloid-positive filamentous bacteria were targeted by the overall probe for Chloroflexi but had a morphology different from type 1851, it is very likely that representatives of Chloroflexi subdivision 1a-1b also produced amyloids. Most of the Chloroflexi filaments had attached epiflora that hybridized with probe Sap309 targeting the family Saprospiraceae, but these did not show any amyloid presence. Some filamentous Chloroflexi without attached growth, only detected with the broad Chloroflexi probe, were primarily amyloid positive at the septa (Fig. 3h).

Among filamentous gram-positive bacteria, the production of amyloid adhesins was investigated by branched mycolata and the unbranched “Candidatus Microthrix parvicella” in the phylum Actinobacteria. “Ca. Microthrix parvicella” did not produce any amyloids. The production of amyloid adhesins by mycolata was tested on probe-defined Gordonia amarae and Skermania piniformis. Both groups stained negative with antibodies, so a fresh sample from a plant with foam problems containing both S. piniformis and G. amarae was further tested by ThT staining. Contrary to the antibody staining, ThT staining showed that all branched organisms in the sample (thus both S. piniformis and G. amarae) produce amyloids but that these were masked by other surface components (Fig. 3c and f). The gram-positive filamentous bacteria belonging to the candidate phylum TM7 (primarily Eikelboom type 0041) did not express any amyloids.

DISCUSSION

The structure of the microbial communities in activated sludge plants with nutrient removal is now relatively well described (28), making it possible to target most of the abundant populations with FISH, using genus or species-specific oligonucleotide probes. This allows a detailed study of the presence of amyloids in the individual populations under the actual in situ conditions. Recently, it was shown that amyloid adhesins are abundant in natural biofilms, including activated sludge, by combining FISH with monoclonal antibodies specific to a generic epitope on amyloid fibrils, a method that proved to be highly specific and reliable (32, 47). In this study, however, it was found in some cases that the amyloids were apparently masked by other EPS components quenching the antibody binding. ThT staining proved to be a useful tool for the analysis of amyloid adhesin layers quenched for detection with antibodies. This means that amyloid adhesins can be analyzed in at least two functional forms in different populations: surface-exposed amyloids (can bind antibody) as described for E. coli and other Enterobacteriaceae (73) and amyloids in a sublayer integrated into other EPS components or perhaps into the cell wall (no antibody binding but ThT positive). The latter type has not been described before, and the two types of amyloids may have different functions for the bacteria, as will be discussed below. A drawback of application of ThT is its low specificity compared to the antibodies, which means that the bacteria staining positive only with ThT can be designated only as potential producers of amyloid protein. However, for the Gordonia species we have been able to purify “hidden” amyloids underneath an EPS layer, indicating that most of the ThT-positive signals are truly amyloids (P. L. Jensen, P. Larsen, P. H. Nielsen, and D. Otzen, unpublished results).

Distribution of amyloid adhesins in activated sludge.

In the 10 different WWTPs investigated, the amyloidal proteins constituted 10 to 40% of the cell area, based on positive staining with propidium iodide, highlighting that amyloids are an important EPS component in activated sludge. The higher fraction of amyloid adhesins in treatment plants with biological phosphorus removal than in plants with only N removal seemed primarily to be associated with a relatively larger fraction of amyloid-producing filamentous bacteria. In many plants of this type, filaments belonging to Chloroflexi and Aquaspirillum are very common (4, 30, 63). Microcolonies and non-cell-bound amyloids were also present in flocs from all plants, but were, as detected by antibodies, not as abundant. However, this fraction must be regarded as a minimum fraction due to the shielding provided by other EPS components against antibody binding.

Identity of amyloid-producing populations in activated sludge.

The most abundant probe-defined populations in nutrient-removing activated sludge plants were investigated for the presence of amyloids in at least two different WWTPs, where previous investigations had shown that the probe-defined populations were abundant. This approach ensured that their potential expression of amyloids was investigated in plants where they were actively growing, and the results can thus be taken as being representative of that particular probe-defined population. Surprisingly, none of the specific populations tested growing in microcolonies produced surface-exposed amyloids that could be labeled with antibody. Only some species hybridizing with the broad probes for Alpha- and Betaproteobacteria and Actinobacteria expressed surface-exposed amyloids, but they could not be identified further by specific probes. When most of the same microcolony-forming probe-defined populations were investigated with ThT for sublayers of amyloid adhesins, great diversity in the bacteria potentially producing amyloid adhesions was revealed. These were detected primarily among the denitrifiers: Azoarcus, Thauera, Zoogloea, and a subpopulation of Aquaspirillum-related bacteria. Gram-positive Actinobacteria also tested ThT positive. A characteristic microcolony-forming bacterium in activated sludge staining with ThT has previously been shown not be targeted by any of the broader probes. Using the specific probes, these microcolonies were identified as a subgroup of Aquaspirillum-related bacteria, which fits well with the fact that they did not hybridize with probe BET42a (64). In typical WWTPs with N and P removal, the microcolony-forming groups, which were ThT positive in this study but not antibody positive, comprise 30 to 50% of the total bacterial biovolume (28, 61). This indicates that sublayers of amyloidal material are an important EPS component for these microcolony-forming species, which generally belong to the strongest floc formers in activated sludge (3, 22, 24).

The probe-defined filamentous bacteria investigated are all commonly encountered in WWTPs, where they form an integrated part of the flocs. However, if they are present in large numbers, they grow into the bulk water and prevent flocculation and settling (cause bulking), making them a nuisance (14, 21). Most of these probe-defined filamentous bacteria produced amyloid adhesins. Interestingly, in most phyla where filamentous organisms with amyloids were found, examples of closely related species with and without amyloids were observed. For example, the alphaproteobacterial M. perideroedes, “Ca. Sphaeronema italicum,” and “Ca. Alysiosphaera europea” produced amyloids, whereas two other types (“Candidatus Monilibacter batavus” and “Candidatus Alysiomicrobium bavaricum”) did not. The results from the probe-defined filamentous organisms were generally supported by pure culture studies (M. perideroedes, Eikelboom type 1851, and H. hydrossis), and only S. natans gave a different result. This strongly indicates that the production of amyloids is widely distributed among filamentous bacteria in several phyla but that not all species expressed amyloids, either because they did not have the capability or because growth conditions triggering the expression of amyloid adhesins were absent. For E. coli and Salmonella species, amyloid production is known to depend on a range of environmental factors such as temperature, osmolarity, nutrient limitation (nitrogen, phosphate, and iron), and oxygen (15, 46, 53).

Possible function of amyloid adhesions in activated sludge.

The term adhesin is an overall term covering bacterial surface structures implicated in specific adhesion of bacteria to different surfaces, and different adhesins are specific for adhesion to target molecules in a lock-and-key fashion like enzymes and immunoglobulins (25). Many pure-culture studies performed on bacteria related to the family Enterobacteriaceae have also shown that the amyloid-like fibrils expressed under different environmental conditions are involved in bacterial adhesion 49); hence, the term amyloid adhesins was adopted (32). However, the results obtained in this study suggest that the function of amyloids in activated sludge is more diverse than hitherto expected.

Microcolony-forming bacteria expressing surface-exposed amyloid adhesins were not present to the same extent as groups producing amyloid sublayers (detected by ThT), indicating that adhesion within microcolonies is promoted by interaction between different surface molecules. This is in accordance with pure-culture studies performed on Salmonella enterica and E. coli, where amyloid fibers, cellulose, colanic acid, and other capsular polysaccharides are integrated to form a highly hydrophobic extracellular matrix under environmental stress (59, 67, 72). It also ties in with the suggestion that the amyloid is essential for biofilm formation and biofilm structure (49). Production of a variety of chemically different surface structures in parallel by ThT-positive bacteria in activated sludge will most likely increase their adhesion strength by a suite of different adhesion mechanisms, such as hydrophobic and electrostatic interactions and physical entanglement, leading to strong microcolonies. We have observed this, in fact, in recent studies (24, 33). Several filamentous bacteria with ThT-positive sublayers were also detected. M. perideroedes, G. amarae, and S. piniformis are known to be highly hydrophobic (20, 29, 43), which might be a consequence of the presence of amyloid-like material together with other substances such as mycolic acid for G. amarae and S. piniformis, in the same way that amyloidal chaplins are responsible for the surface hydrophobicity of filamentous streptomyces hyphae (7). For Thiothrix, antibody-positive filament tips were detected in a sample containing Thiothrix rosettes, which could be the fimbriae previously described on the tips of Thiothrix rosettes (70). However, the ThT staining also indicated the presence of amyloids underneath the sheath, and the function of these remains uncertain, as Thiothrix is usually hydrophilic (P. H. Nielsen, unpublished results).

The function of surface-exposed amyloids on the filamentous Chloroflexi and Aquaspirillum may be related to the presence of a very strong sheath or skeleton. These filamentous bacteria were often observed to have dead cells in the middle of the filament; but the sheath material kept the filament intact, and it was still antibody positive in these segments, confirming its relatively high mechanical and/or biological stability. They may thus be very important backbones in the floc and are also the most common filamentous bacteria in many N- and P-removing treatment plants (30). Interestingly, these two particular groups are also frequently observed with attached growth of other bacteria. This epiflora is primarily affiliated to the family Saprospiraceae (28), and the bacteria are highly specialized in protein hydrolysis (71). Whether they degrade the amyloids or use these for anchorage to the sheaths remains unknown.

In summary, this study clearly showed that amyloid adhesins constitute a relatively large fraction of EPS in activated sludge flocs and are expressed by a diverse range of probe-defined groups, growing in microcolonies and as filaments. The function of the amyloidal material seems surprisingly divers, including a role in internal adhesion within the microcolonies, strengthening the sheaths of filamentous bacteria important for the backbone of the flocs, and increasing surface hydrophobicity. Most likely, future studies will reveal additional functions.