OUP user menu

Identity and ecophysiology of filamentous bacteria in activated sludge

Per Halkjær Nielsen, Caroline Kragelund, Robert J. Seviour, Jeppe Lund Nielsen
DOI: http://dx.doi.org/10.1111/j.1574-6976.2009.00186.x 969-998 First published online: 1 November 2009


Excessive growth of filamentous bacteria in activated sludge wastewater treatment plants (WWTPs) can cause serious operational problems. With some filaments there may be the problem of bulking, where inadequate flocculation and settling of the biomass in the secondary clarifier results in a carryover of solids with the final treated liquid effluent. Their proliferation often encourages the development of stable foams on the surface of the reactors, and these foams may impact negatively on plant performance and operation. The availability of culture-independent molecular methods now allows us to identify many of the more common filamentous organisms encountered in WWTPs, which are phylogenetically diverse, affiliating to seven separate bacterial phyla. Furthermore, the extensive data published in the past decade on their in situ behaviour from the application of these culture-independent methods have not been summarized or reviewed critically. Hence, here, we attempt to discuss what we now know about their identity, ecophysiology and ecological niches and its practical value in better managing activated sludge processes. Some of this knowledge is already being applied to control and manage full-scale WWTPs better, and the hope is that this review will contribute towards further developments in this field of environmental microbiology.

  • bulking
  • foaming
  • activated sludge
  • FISH
  • ecophysiology
  • filamentous bacteria


Excessive growth of filamentous bacteria in activated sludge wastewater treatment plants (WWTPs) has caused serious operational problems for many years. An overgrowth of some filaments, an event called bulking, can prevent adequate flocculation and settling of the biomass in the secondary clarifier, leading to a carryover of solids with the final treated liquid effluent. Most conventional treatment plants with carbon removal suffer sporadically from severe bulking problems, as do the more recently introduced biological nutrient removal (BNR) processes designed to remove nitrogen and phosphorus. Also, foaming problems caused by several filamentous bacteria can be a severe problem in all types of plants. Since the early 1970s, numerous studies have attempted to develop suitable methods to control these filamentous organisms based on a better understanding of their identity, physiology and ecology. More than 30 different filament morphotypes have been described in activated sludge systems treating primarily municipal wastewater (Eikelboom, 1975, 2000) and many more are encountered in industrial treatment plant communities (Eikelboom & Geurkink, 2002; Eikelboom, 2006). However, it is clear that only a few filaments can be identified reliably based on their morphological features; hence, much of the early published information on activated sludge microbiology is of little more than historical interest and should be viewed as such. This is largely because of the problems associated with the culture-dependent methodologies used in these early studies to reveal the true level of population biodiversity present.

With the development of culture-independent molecular methods, our knowledge of the identification of filaments and other populations in activated sludge has increased dramatically. Applying a range of molecular methods, especially FISH with rRNA-directed oligonucleotides, has been particularly productive. However, as most of these organisms are uncultured, our understanding of their physiology and ecology is still unsatisfactory. With the availability of tools such as microautoradiography in combination with FISH (Lee, 1999), we are now learning much about their ecophysiology, information that might provide clues as to how they and the problems they cause may be controlled.

The extensive work published over the past 10 years on elucidating the in situ behaviour of these filamentous bacteria with culture-independent methods has not been summarized or reviewed, and so here we attempt to discuss critically what we now know about the identity, ecophysiology and ecological niches of these bacteria in activated sludge and its practical implications for better management of these processes. Some of this knowledge is already being used to better control and manage full-scale WWTPs, and the hope is that this review will contribute towards further developments in this field of environmental microbiology.

Identifying filamentous bacteria

A high number of different filamentous morphotypes are present in activated sludge systems. These have been identified conventionally by their morphological features using the manuals by Eikelboom (2000) and Jenkins (2004). However, we now know that only a few species can be identified reliably in this way. These include Candidatus‘Microthrix parvicella’, Thiothrix spp. and perhaps a few Mycolata (formerly referred to as Nocardia or nocardioforms) (Soddell & Seviour, 1990). Instead, molecular methods such as FISH should always be used if possible, for their identification. The available 16S rRNA gene sequences reveal that these filaments belong to at least seven different bacterial phyla (Fig. 1a–c). Some are commonly present in WWTPs whereas others are found only rarely. Those filamentous bacteria whose phylogeny is known are shown in Table 1, together with the rRNA-targeted probes used for their identification, and all the other potential filament morphotypes that hybridize with each of the probes [sequences of the different probes can be found in probeBase (Loy, 2003, 2007) or in Kragelund (2009)]. The most commonly encountered filamentous bacteria causing bulking problems belong to the Alphaproteobacteria (‘Nostocoida’-like), the Gammaproteobacteria (Thiothrix and type 021N), the Actinobacteria (Candidatus‘Microthrix’, Mycolata) and the Chloroflexi (types 1851, 0041 and 0092).

Figure 1

(a) Maximum likelihood phylogenetic tree of the 16S rRNA gene sequences. (a) Important filamentous Proteobacteria and related sequences. (b) Important filamentous Chloroflexi, Planctomycetales and Bacteroidetes, and related sequences. (c) Important filamentous Gram-positive bacteria and related sequences. All sequences were at least 1000-bp long except AF244377 (541 bp), which was added later using the quick add function in arb. The scale bar corresponds to 0.1 substitutions per nucleotide position.

View this table:
Table 1

Filamentous bacteria organized according to phylogenetic affiliation, oligonucleotide probe and morphotype

PhylumSpecies, genus, class or phylumOligonucleotide probeMorphotypeReferences
ProteobacteriaClass AlphaproteobacteriaALF968Neef (1997)
Meganema perideroedesMeg983+Meg1028Nostocoida limicola II/021NThomsen (2006a)
Meganema perideroedesEU12-645+EU26-653Nostocoida limicola II/021NLevantesi (2004)
Candidatus‘Alysiomicrobium bavaricum’PPx3-1428Nostocoida limicola IILevantesi (2004)
Candidatus‘Alysiomicrobium bavaricum’PPx1002Nostocoida limicola IIKragelund (2006)
Candidatus‘Monilibacter batavus’MC2-649Nostocoida limicola IILevantesi (2004)
Candidatus‘Alysiosphaera europaea’Noli-644Nostocoida limicola IILevantesi (2004)
Candidatus‘Sphaeronema italicum’Sita-649+compsita-649Nostocoida limicola IILevantesi (2004)
Candidatus ‘Sphaeronema italicum’Nost 993+helper1010Nostocoida limicola IIKragelund (2006)
Candidatus‘Combothrix italica’Combo1031Nostocoida limicola IILevantesi (2004)
Class BetaproteobacteriaBET42a (use with GAM42a)Manz (1992)
Curvibacter-related (formerly Aquaspirillum)Curvi997+Comp1curvi997+Comp2curvi997Type1701, type 0041/0675Thomsen (2004)
LeptothrixdiscophoraLDILeptothrixWagner (1994a)
SphaerotilusnatansSNASphaerotilusWagner (1994a)
Class GammaproteobacteriaGAM42a (use with BET42a)Manz (1992)
Thiothrix eikelboomii, T. nivea, T. unzii, T. fructosivorans, T. defluvii, Eikelboom type 021N group I, II, IIIG123T+G123T-CThiothrix and type 021N groupKanagawa (2000)
Eikelboom type 021N group I (T. disciformis)G1BType 021NKanagawa (2000)
Eikelboom type 021N group II (T. eikelboomii)G2MType 021NKanagawa (2000)
Eikelboom type 021N group III (T. flexilis)G3MType 021NKanagawa (2000)
ThiothrixniveaTNIThiothrixWagner (1994a)
Eikelboom type 021N strain II-2621NType 021NWagner (1994a)
AcinetobacterACA23aType 1863Wagner (1994b)
LeucothrixmucorLMULeucothrixWagner (1994a)
BacteroidetesNo phylum probe (sum of probes below)
Most FlavobacteriaCF319aH. hydrossis-likeManz (1996)
Some FlavobacteriaCF319bH. hydrossis-likeManz (1996)
Most members of the class BacteroidetesCFB719H. hydrossis-likeWeller (2000)
Unclassified FlexibacteraceaeCFB655Thin filamentous bacteriaKindaichi (2004)
Unclassified FlexibacteraceaeCFB730Thin filamentous bacteriaKindaichi (2004)
Candidatus Magnospira bakiiBak655Curled filamentSnaidr (1999)
Most FlavobacteriaCFB563H. hydrossis-likeWeller (2000)
Most members of the genus Tannerella and the genus Prevotella of the class BacteroidetesCFB286H. hydrossis-likeWeller (2000)
Family SaprospiraceaeSAP309H. hydrossis-likeSchauer & Hahn (2005)
Haliscomenobacter hydrossis and Iso10BHHY-654H. hydrossis+H.hydrossis-likeKragelund (2008)
Haliscomenobacter hydrossisHHYH. hydrossisWagner (1994a, b)
Clone T5HHY-T5H. hydrossis-likeKragelund (2008)
ChloroflexiPhylum ChloroflexiCFX1223 and GNSB941Group specific Some type 1851 and type 0041/0675, H. hydrossis-likeGich (2001) and Bjornsson (2002)
Subdivision 1 of ChloroflexiCFX784Some type 1851 and type 0041/0675, H. hydrossis-likeBjornsson (2002)
Subdivision 3 of ChloroflexiCFX109Some type 1851 and type 0041/0675, H. hydrossis-likeBjornsson (2002)
Eikelboom type 1851Chl1851Type 1851Beer (2002)
Isolate EU25EU25-1238Type 1851Kragelund (2007a)
AnaerolineaGNSB633Very small thin filamentsSekiguchi (1999)
Unclassified CaldilineaceaGNS667Resemble type 1851 without attached growthKindaichi (2004)
Eikelboom type 0092 (variant 1)CFX197Thicker (0.80 μm) type 0092 trichomeSpeirs (2009)
Eikelboom type 0092 (variant 2)CFX223Thinner (0.67 μm) type 0092 trichomeSpeirs (2009)
ActinobacteriaActinobacteriaHGC69aRoller (1994)
NostocoidalimicolaNLII65Nostocoida limicola IIBradford (1997)
Candidatus‘Nostocoida limicola’ (Tetrasphaera jenkinsii and T. veronensis)NLIMII175Nostocoida limicola IILiu & Seviour (2001)
Candidatus‘Microthrix parvicella’MPAmix (MPA60, MPA223, MPA645)Microthrix parvicellaErhart (1997)
Candidatus ’Microthrixparvicella’ and Candidatus‘M. calida’Mpa-all-1410Microthrix parvicellaLevantesi (2006)
Candidatus‘M. calida’Mpa-T1-1260Thin Microthrix parvicellaLevantesi (2006)
Mycobacterium subdivision (mycolic acid-containing actinomycetes)Myc657MycolataDavenport (2000)
Gordonia spp.Gor596Mycolatade los Reyes (1997)
GordoniaamaraeG.am205Mycolatade los Reyes (1998)
SkermaniapiniformisSpin1449MycolataEales (2006)
FirmicutesFirmicutesLGC354A-CMeier (1999)
Streptococcus spp.Str/StreptStreptococcusTrebesius (2000)
Trichococcus Nostocoida limicola INlimI91Nostocoida limicola ILiu & Seviour (2001)
PlantomycetesPlanctomycetalesPla46Neef (1998)
IsosphaeraNlimIII301Nostocoida limicola IIILiu & Seviour (2001)
TM7Phylum specificTM7905Type 0041/0675Hugenholtz (2001)
Subdivision 1TM7305Type 0041/0675Hugenholtz (2001)
  • * Microthrix sp. is not targeted by this probe.

The data in Table 1 clearly show that each filamentous morphotype often includes more than one species. This is exemplified by the morphotype Nostocoida limicola, which affiliates to at least four different bacterial phyla: Proteobacteria (Snaidr, 2002; van der Waarde, 2002; Levantesi, 2004; Thomsen, 2006a), Firmicutes (Liu, 2000), Actinobacteria (Liu, 2001) and Chloroflexi (Schade, 2002). Also, the thin needle-like filaments morphologically identified as Haliscomenobacter hydrossis in the phylum Bacteroidetes in fact may be short and thin Chloroflexi filaments that can be detected using subdivision probes (Bjornsson, 2002). Furthermore, many H. hydrossis-like species do not hybridize with the probe designed to target this species (HHY), but only with broad group-specific probes, indicating the presence of a large yet unidentifiable population of thin needle-like filaments in activated sludge (Kindaichi, 2004; Schauer & Hahn, 2005; Kragelund, 2008). The filament morphotypes with attached epiflora (types 0041/0675, 1701 and 1851) are also difficult or impossible to identify without FISH analysis (Thomsen, 2002, 2006b; Kragelund, 2007a; Xia, 2008). For these reasons, it is highly recommended to apply FISH or other molecular methods in future plant surveys to supplement the more traditional examination methods. Some morphotypes (e.g. types 0914 and 0803) are not presently identifiable by FISH because of a lack of rRNA sequence data required for targeted probe design.

Investigating important filament ecophysiological traits

A range of physiological characteristics determine the ecology and competitive behaviour of filamentous and other bacteria in activated sludge. Their substrate uptake capability is important, whereby some populations can consume a wide range of substrates (general consumers) while others are restricted to only a few (specialist consumers). Their uptake rates and substrate affinities (Km) are also critical parameters in the competition for substrates. A key issue for many bacteria is their respiratoric capabilities, i.e. whether they are obligate aerobes, able to use nitrite and/or nitrate as electron acceptors (e-acceptor) (e.g. denitrification) or whether they are fermentative will affect their abilities to compete in different reactor zones of plants. An ability to store organic substrates intracellularly as carbon and energy reserves can be especially advantageous under the feast–famine regimes encountered in most activated sludge systems 5w?>(Martins, 2004). Such storage compounds include polyhydroxyalkanoates (PHA) or glycogen, or in some filaments, polyphosphate (polyP) and elemental sulphur (S°) granules are synthesized as energy reserves (Eikelboom, 2000; Nielsen, 2000). Some bacterial populations rely mainly on low-molecular-weight soluble substrates while others utilize macromolecules, which require excretion of appropriate exoenzymes for their hydrolysis (Eikelboom, 2000; Nielsen, 2002). The surface properties of cells are also important for adsorption of some classes of substrates (e.g. lipids to hydrophobic surfaces), and may also promote cell adhesion to floc material. Also, their sensitivity to starvation, oxygen levels or toxic chemicals will all help to determine their competitive potential.

Only a few axenic cultures of the relevant filamentous organisms from activated sludge are available to study these parameters so important for their survival and competitive abilities. However, as is evident from the discussions below, it is often very difficult to extrapolate physiological information obtained from pure cultures of filaments to their in situ physiology in WWTPs. There are several reasons for this. For example, different closely related isolated strains of a single species often grow on different substrates. Thus, the identity of the strain actually growing in the plant is unknown and is likely to be a closely related, but uncultured one. Therefore, studies with isolates can only provide general information about the actual physiology of individual filamentous species and in situ studies are needed to support these findings. A good example is C. ‘Microthrix parvicella’, which may grow on acetate in pure culture, but has never been shown to take up acetate in situ (Andreasen & Nielsen, 2000), as discussed further below.

Another reason is that the selective pressures are very different in pure cultures and in treatment plants, and so physiological traits important in situ may be unimportant and not induced/expressed in a pure culture, and vice versa. In treatment plants, most organisms grow under dynamic conditions, which are often sporadic feast and famine periods (Martins, 2004; Tandoi, 2006), and very different from those experienced by pure cultures. For example, an ability to take up substrates for storage under conditions where the bacteria cannot grow may be an important fitness parameter in mixed, dynamic systems, but is difficult to study in pure culture. Such a situation is seen with some of the filamentous Alphaproteobacteria (Kragelund, 2005, 2006). They assimilate large amounts of substrate under different e-acceptor conditions and store these as polyhydroxyalkanoates. It is unclear whether they can actually grow with nitrate or nitrite as e-acceptors, but an ability to take up substrates under these conditions and to grow when oxygen is present may be important in competing for substrates (Kragelund, 2005, 2006).

In the absence of cultured representatives for many of these filaments, and the concern that their phenotypes in an axenic culture may not reflect those in situ, their physiology is best determined by culture-independent methods, typically based on resolution at a single cell level and readily visualized by microscopic analyses. Several novel methods have been developed over the past 10 years that allow key traits of their ecophysiology to be elucidated (Table 2).

View this table:
Table 2

Overview of methods used for ecophysiological characterization on single cell level on filamentous bacteria in activated sludge

IdentityFISHQuantitative FISHAmann (1990); Daims (2006)
Choice of substrates and e-acceptor conditionsMARMAR-FISH HetCO2-MAR-FISH Q-MAR-FISHLee (1999); Hesselsoe (2005); Nielsen (2003b)
Exoenzyme expressionELFELF-FISHNielsen (2002); Van Ommen Kloecke & Geesey (1999)
BODIPY-labelled fluorescenceBODIPY-label-FISHXia (2007)
Surface properties and componentsMAC Presence of amyloidsMAC-FISH Antibodies-FISHNielsen (2001); Zita & Hermansson (1997a, b); Larsen (2007)
Storage compounds (PHA)Sudan Black Nile BluePHA quantificationDawes (1991) Kragelund (2005)
Storage compounds
Elemental S°ThiosulphateNielsen (2000)
PolyPNeisser, DAPIEikelboom (2000); Crocetti (2000)
  • ELF, enzyme-labelled fluorescence; MAC microsphere adhesion to cells; MAR, microautoradiography; PHA, poly-β-hydroxyalkanoate; polyP, polyphosphate.

Identification of filamentous bacteria in these ecophysiological studies is based on FISH analyses. However, it means that their identification is no better than the specificity of the oligonucleotide probe applied. This feature varies within the currently published probes, but is being continuously addressed as more is known of the phylogenetic diversity existing within each filament morphotype. Where this information is unavailable, as is often still the case, the assumption is usually made that each FISH probe targets a single filament ‘species’. What is known about the specificity of most probes can be found in probeBase (Loy, 2003, 2007) and in the recently published book on FISH detection and quantification of microorganisms in activated sludge (Nielsen et al., 2009), which includes a chapter dedicated to filamentous bacteria and their detection (Kragelund, 2009). FISH can be combined with a range of other methods to detect important ecophysiological traits on a single cell level. Microautoradiography is a powerful technique to study directly the uptake of radiolabelled substrates under defined incubation conditions in complex environments (Andreasen & Nielsen, 1997). Combining microautoradiography and FISH enables the link to be made between physiological capability and identity (Lee, 1999; Ouverney & Fuhrman, 1999; Nielsen & Nielsen, 2005). Variations of the microautoradiography (MAR) protocol such as quantitative MAR and HetCO2-MAR allow for quantification of substrate uptake and more detailed physiological investigations (Nielsen, 2003b; Hesselsoe, 2005). The redox dye 5-cyano-2,3-ditolyl tetrazolium chloride, which fluoresces after reduction, can be used to show respiratoric activity in individual cells (Nielsen, 2003a), while Sudan Black or Nile blue staining detects whether intracellular polyhydroxyalkanoates (or other lipidic storage compounds) are synthesized following incubation with individual substrates under defined conditions, and subsequently its levels are quantified using image analysis software (Kragelund, 2005, 2006). Levels of polyhydroxyalkanoates stored in different filamentous bacteria have been determined in situ based on the author's subjective experiences (low, medium and high), and these are indicated in the tables when appropriate. Intracellular polyphosphate inclusions are readily detected by Neisser or 4′-6-diamidino-2-phenylindole staining (Crocetti, 2000).

Surface properties of cells can be visualized by microsphere adhesion to cell (MAC) analyses (Zita & Hermansson, 1997a, b). This method uses small hydrophobic microspheres (0.01 μm) that adhere to other hydrophobic surfaces in activated sludge populations including some filamentous bacteria, especially those involved in foaming (Nielsen, 2001; Kragelund, 2007c). The presence of specific components on their cell surface, such as proteinaceous amyloidal fibrils, can be detected with appropriate antibodies (Larsen, 2007, 2008), and exoenzymes can be detected by applying enzyme-labelled-fluorescence (ELF) methods (Van Ommen Kloecke & Geesey, 1999; Nielsen, 2002). ELF substrates are commercially available for in situ detection of a limited range of six enzymes only (esterase, lipase, chitinase, galactosidase, glucuronidase and phosphatase), although BODIPY-labelled proteins and starch-based assays have now been developed to elucidate the populations responsible for their enzymatic degradation in situ (Xia, 2007, 2008).

Ecophysiology of filamentous bacteria



Filamentous members of the class Alphaproteobacteria are commonly seen in activated sludge biomass. Currently, six filamentous phylotypes are recognized, all with an appearance resembling the morphotypes N. limicola or type 021N (Levantesi, 2004). They were initially described by Snaidr (2002). Their phylogeny has since been resolved in more detail and they were named by Levantesi (2004). As shown in Fig. 1a, three clusters exist: one containing Candidatus‘Monilibacter batavus’ and Candidatus‘Sphaeronema italicum’, one consisting of Candidatus‘Alysiosphaera europaea’, Candidatus‘Alysiomicrobium bavaricum’, and Candidatus‘Combothrix italica’, and one with a single cultured member, Meganema perideroedes (Kragelund, 2006). Each can be detected and identified in situ by applying the FISH probes listed in Table 1, but other, so far unidentified, filamentous Alphaproteobacteria are often encountered (Levantesi 2004; C. Kragelund & P.H. Nielsen, unpublished data; T. Nittami et al., unpublished data).

FISH-based surveys demonstrated that these filamentous Alphaproteobacteria were abundant, especially in biomass samples from industrial treatment plants, and were often involved in bulking and possibly foaming episodes (Eikelboom & Geurkink, 2002; Snaidr, 2002; van der Waarde, 2002). They were present in about 65% of the samples investigated (86 plants) and seen in excessive amounts in 20–25% of these (van der Waarde, 2002; Levantesi, 2004). Earlier surveys based entirely on morphological identification may have misidentified these filamentous Alphaproteobacteria as N. limicola or type 021N morphotypes.


Pure cultures exist only for M. perideroedes (Levantesi, 2004; Thomsen, 2006a). Like pure cultures of other filaments (Blackall, 2000), some strains failed to grow on minimal media supplemented with different carbon and nitrogen sources, making any detailed characterization impossible (Table 3). Other M. perideroedes isolates were culturable on MSV medium supplemented with a vitamin solution and a range of carbon sources (Levantesi, 2004). Physiological characterizations were performed on three isolates and revealed utilization of short-chain fatty acids (SCFAs), Tween 80, sugars, starch, alcohols and amino acids. Polyphosphates and polyhydroxyalkanoates were stored intracellularly. None of the isolates were able to denitrify or grow anaerobically on glucose and none could oxidize reduced sulphur compounds (Thomsen, 2006a, b).

View this table:
Table 3

Comparison between pure culture physiology and in situ physiology for filamentous Alphaproteobacteria in activated sludge

ProbePure cultureIn situ
Pure culture EU6, EU15 and EU26Candidatus‘Monilibacter batavus’Candidatus‘Sphaeronema italicum’Candidatus‘Alysiomicrobium bavaricum‘Candidatus‘Alysiosphaera europaea’M. perideroedes
MC2-649Nost993+ helper1010Sita-649+ compsita-649PPx1002PPx3-1428Noli-644Meg983+ Meg1028
Use of e-acceptorsO2O2, (NO3)α, (NO2O2O2, (NO3)α, (NO2O2, (NO3), (NO2)O2O2, (NO3)α, (NO2O2, (NO3)α, (NO2
Storage compounds (PHA)HighHighHighHighHighHighHighHigh
Surface propertiesNDHydrophobicHydrophobicHydrophobicHydrophobicHydrophobicHydrophobicHydrophobic
Short-chain fatty acidsAce, Prop, PyrAce, Prop, PyrAce, PyrAce, But, Prop, PyrAce, But, Prop, PyrAce, PyrAce, But, PyrAce, But, Prop
Long-chain fatty acidsTween 80OleicOleicOleic
SugarsGlu, frucGlu, Gal, ManGlu, Gal, ManGlu, Gal, ManGlu, Gal, Man
Di- or polysaccharidesStarchNDNDNDNDNDNDND
AlcoholsEth, propa, butaEth
Amino acidsAsp, SerLeuLeuLeuGly, Leu
Bicarbonate with thiosulphate
  • * Levantesi (2004).

  • Kragelund (2006).

  • Kragelund (2005).

  • § § Only observed once.

  • ¶ Not all isolates.

  • α, only reduced number of substrates are taken up.

  • Ace, acetate; Asp, aspartate; But, butyrate; Buta, butanol; Cas, casamino acids; Chit, chitinase; Est, esterase; Eth, ethanol; Fru, fructose; Fum, fumerate; Gal, galactose; Gala, galactosidase; Gel, gelatine; Glu, glucose; Glucu, glucuronidase; Gly, glycine; Glyce, glycerol; Lip, lipase; Mal, malate; Man, mannose; Mann, mannitol; Mel, melezitose; Oleic, oleic acid; Pal, palmitate; Phos, phosphatase; Prop, propionate; Propa, propanol; Prot, proteases; Pyr, pyruvate; Ser, serine; Suc, succinate; Sucr, sucrose; Tre, trehalose; Triol, trioleic acid; ND, not determined; −, no substrate uptake; +, active substrate uptake.

Other studies have focused on the ecophysiology of these filamentous Alphaproteobacteria (Kragelund, 2005, 2006). All except C. ‘Combothrix italica’ were investigated in samples from industrial WWTP in four European countries. These filamentous Alphaproteobacteria seem to be limited to utilizing soluble substrates only. They have a high affinity for acetate (low Km value) and comparatively high substrate uptake rates (Nielsen, 2003a, b). They could be divided into two groups, each with very similar ecophysiologies, as shown in Table 3. These groupings corresponded well to the phylogenetically defined clusters. Thus, group 1 (C. ‘Monilibacter batavus’ and C. ‘Sphaeronema italicum’) assimilated several SCFAs. Group 2 members (C. ‘Alysiosphaera europaea’, C. ‘Alysiomicrobium bavaricum’ and M. perideroedes) were nutritionally more versatile, and, in addition to SCFAs, took up a range of sugars, ethanol and amino acids. They were all aerobic, although most assimilated substrates with nitrate or nitrite as an e-acceptor. Whether they could denitrify is unclear. All filaments had a large storage capacity in the form of polyhydroxyalkanoates, and several substrates were converted into polyhydroxyalkanoates under different e-acceptor incubation conditions. Excretion of exoenzymes by these Alphaproteobacteria was not commonly observed, although some weak activity was detected (Kragelund, 2005). They all had relatively hydrophobic surfaces, but the functional importance of this feature is unknown. However, each possessed amyloidal surface components that may influence their cell surface hydrophobicity (Larsen, 2008).


For many years, the betaproteobacterial Sphaerotilus (Sphaerotilus natans) was believed to be the most problematic filamentous bacterium in activated sludge systems (Van Veen, 1973; Eikelboom, 1975). Today, we know that it was frequently misidentified and, together with Leptothrix sp., is now known to occur only occasionally there (Van Veen, 1973; Eikelboom, 1975; Garrity, 2002; van der Waarde, 2002). FISH probes targeting S. natans and Leptothrix discophora have been designed based on 16S rRNA gene sequence data from selected cultured strains (Wagner, 1994a). Sequence analysis of an isolated morphotype 1701 (RC2, ATCC49750) indicates that this is closely related to S. natans and also contains the site for the FISH probe targeting S. natans (Howarth, 1998). However, to our knowledge, such probe-defined type 1701 filaments are not common in full-scale plants.

Several other betaproteobacterial filamentous bacteria have been reported as abundant in activated sludge systems. Morphotype 0803 is an example, but full sequence information for it is not available (Bradford, 1996), and so it is uncertain whether they actually belong to the Betaproteobacteria. Aquaspirillum-related filamentous bacteria may be common in full-scale plants, although rarely in high numbers (Thomsen, 2006b). Interestingly, these probe-defined bacteria also form microcolonies in biomass samples (Thomsen, 2004). The FISH probe targeting these (Aqs997) is relatively broad and covers many bacteria related to Aquaspirillum delicatum [reclassified as Curvibacter delicates comb. nov. (Ding & Yokota, 2004)], Pseudomonas lanceolata (reclassified to Curvibacter lanceolatus comb. nov.) and other members of the genus Curvibacter. Thus, the precise identity of these probe-defined bacteria is uncertain. The same probe detected two filamentous morphotypes, some of which had attached epiphytic bacteria belonging to Saprospiraceae in the Bacteroidetes (Xia, 2007, 2008). The type A morphotype is believed to belong to the genus Curvibacter, whereas the identity of type B remains unclear. It seems likely that these Curvibacter-related filaments may include some of the other filament morphotypes with epiphytic growth (types 0041/0675, 1701 and 1851). They probably play an important role in floc formation because they are integrated generally within the floc matrix.

The abundance of the S. natans morphotype has been monitored frequently in full-scale conventional plants, and in some countries, early surveys suggest that it occurs commonly in bulking samples, [e.g. the fourth most frequently observed filamentous organisms in Germany (Wagner, 1982)]. Likewise, morphotype 1701 was the second most abundant filament in an early microscopy-based US survey (Richard, 1982; Strom & Jenkins, 1984), but as discussed above, the reliability of filament identification in such studies remains in doubt and no FISH analyses have confirmed these data.


Several pure cultures of S. natans and L. discophora exist and some information on their physiology is available (Van Veen, 1978; Richard, 1985a; Williams & Unz, 1985b; Mulder & Deinema, 1992; Kämpfer, 1995). In general, both S. natans and L. discophora are obligate aerobes, able to consume a wide variety of organic acids, sugars, alcohols and amino acids (Table 4). More complex substrates such as cellobiose, starch and inulin were not utilized. No similar investigations have been performed on the single isolate of morphotype 0803 due to its slow growth rate in axenic culture (Williams & Unz, 1985b). Several studies have been conducted on isolates of the morphotype 1701, but as no sequence data are available to confirm their identity, these physiological data are not included in this review. No pure cultures exist for the Curvibacter-related filaments.

View this table:
Table 4

Comparison between pure culture physiology and in situ physiology for filamentous Betaproteobacteria in activated sludge

ProbePure cultureIn situ
S. natansL. discophoraCurvibacter-related bacteria
Use of e-acceptorsO2O2O2, (NO3), (NO2)
Storage compounds (PHA)PresentPresentPresent
Surface propertiesNDNDHydrophilic
Short-chain fatty acidsAceAce, But, PropBut
Long-chain fatty acids
SugarsGlu, Fru, Man, Gal+othersFru, Gal, GluGal, Man
Di- and polysaccharidesStarchND
AlcoholsEth+sugar alcoholsButa, Eth
Amino acids++/−
Amino acid mixtureNDND+
Bicarbonate with thiosulphateNDNDND
  • * Richard (1985a).

  • Van Veen (1978), Mulder & Deinema (1992).

  • Thomsen et al. (2006b), Xia (2007).

  • ¶ Not all isolates.

  • ( ), reduced uptake.

  • For abbreviations, see Table 3.

Ecophysiological studies have so far been conducted only with the Curvibacter-related filaments, as shown in Table 4 (Thomsen, 2006b). Investigations on samples from two municipal WWTPs showed that acetate and glucose were not taken up under aerobic conditions, while mannose and a few other substrates including amino acids were. Most of these substrates were still assimilated with nitrate or nitrite as the e-acceptor, but again whether these bacteria can denitrify is unclear. Regardless of the e-acceptor conditions used, in many cases, substrates were converted into intracellular polyhydroxyalkanoates (Thomsen, 2006a, b). No or very little anaerobic substrate assimilation was recorded. It was not possible to distinguish between the type A and B morphotypes, but type A was believed to resemble the microcolony-forming Curvibacter, only taking up an amino acid mixture, which suggests that these filaments are involved in protein degradation, a proposal confirmed by production of exoenzymatic proteases (Xia, 2007). No other exoenzymatic activity has been detected with these, which is not to say that it does not exist, bearing in mind the limited range of ELF substrates available for studies of this kind. Their surface properties were consistent with relatively hydrophilic filaments, and with epiphytic bacteria present (Thomsen, 2006b).


Thiothrix, Beggiatoa and Leucothrix all affiliate within the family Thiotrichaceae in the class Gammaproteobacteria (Cole, 2005). In the past, Thiothrix spp. have been associated mainly with sulphide-containing water and marine habitats, but are frequently implicated in bulking incidents, particularly in industrial treatment systems (Eikelboom, 2000; Eikelboom & Geurkink, 2002; van der Waarde, 2002). Gliding Beggiatoa are often located in the sediment (Nelson, 1992), but are only occasionally seen in activated sludge and always in low numbers.

Currently, seven species of Thiothrix are recognized and several studies have focused on resolving their detailed phylogeny (Howarth, 1999; Kanagawa, 2000; Aruga, 2002). Thiothrix nivea (Teske, 1996) and Thiothrix ramosa (Polz, 1996) have been identified. The morphotype 021N has been included as a separate species Thiothrix eikelboomii and two additional species are proposed as further representatives of the Eikelboom 021N morphotype isolates: Thiothrix disciformis and Thiothrix flexilis (Kanagawa, 2000). The original two FISH probes designed to target Thiothrix and type 021N have now been supplemented with others designed to target the entire ‘type 021N and Thiothrix group’, as well as the three subdivisions represented by T. disciformis (group I), T. eikelboomii (group II) and T. flexilis (group III) (Aruga, 2002). A re-evaluation of the probes TNI and 21N showed that probe 21N targeted group II members, and several mismatches to the target sequences for the TNI probe were found (Kanagawa, 2000).

A FISH probe was also designed to target Leucothrix spp. based on 16S rRNA gene sequence analysis of a pure culture (Wagner, 1994a), but published data suggest that they are not seen very often in activated sludge, and do not appear to play a role in either bulking or foaming. No probes are available for Beggiatoa species in activated sludge, but these exist for marine species (Teske, 1996; Mussmann, 2003).

Some filamentous members of the genera Acinetobacter and Moraxella are occasionally present in activated sludge systems. Under the microscope, these fit the description for morphotype 1863 (Rossetti, 1997; Seviour, 1997), an example of a polyphyletic filament morphotype as some strains also affiliate within the Cytophaga group (phylum Bacteroidetes). Those affiliated within the Gammaproteobacteria were targeted by the probe ACA23a designed against Acinetobacter (Wagner, 1994b).


Several studies have investigated the physiology of Thiothrix sp. and type 021N morphotype isolates in pure cultures (Richards, 1985b; Williams & Unz, 1985a, b; Tandoi, 1994; Aruga, 2002; Rossetti, 2003). In general, Thiothrix isolates can use organic acids and gelatine and require reduced sulphur compounds for growth (Table 5) (Williams & Unz, 1985a, b). Also, both autotrophic and mixotrophic growth by Thiothrix has been observed (Williams & Unz, 1985b; Tandoi, 1994; Rossetti, 2003) as well as an ability to reduce nitrate to nitrite (Aruga, 2002). Aruga (2002) characterized 15 isolates of Eikelboom morphotype 021N in terms of their substrate uptake capabilities. Three physiological groups emerged, each corresponding to a phylogenetically defined group described by Kanagawa, (2000). All strains grew on a wide range of different sugars (both mono- and polysaccharides) and SCFAs (Table 5). An ability to reduce nitrate to nitrite was also recorded for some isolates in group II and all in group III. Moreover, group III isolates showed only a limited ability to oxidize reduced S, whereas large sulphur deposits were seen in filaments of members of groups I and II. The physiology of type 1863 affiliating to the Acinetobacter has been investigated in two studies. Strain RT2 was investigated using the API 20E system, which showed the isolate to be a strict aerobe, utilizing acetate, but no carbohydrates (Rossetti, 1997). The other study confirmed these observations and showed the utilization of several amino acids and different Tween surfactants (Seviour, 1997).

View this table:
Table 5

Comparison between pure culture physiology and in situ physiology for filamentous Gammaproteobacteria in activated sludge

Pure cultureIn situ
ThiothrixType 021NThiothrixType 021N
Use of e-acceptorsO2, NO3O2, (NO3)O2, NO3 (anaerobic)O2
Storage compoundsPHA, S°PHA, PolyPPHA (low), S°PHA, S°
Surface propertiesNDNDNDHydrophilic
Short-chain fatty acidsAce, Fum Lac, Prop, Pyr, SucAce, But, Fum Lac, Mal, Prop, Pyr, SucAceAce, But, Pyr
Long-chain fatty acidsGlyceNDOleic
SugarsFru, Gal, Glu, ManGal, Glu, Man
Di- and polysaccharidesMel, Suc, TreNDND
Amino acidsNDAla, AspGly, Leu
Amino acid mixtureNDNDNDND
Bicarbonate with thiosulphate++
MiscellaneousGel Auto and mixotrophicGelAuto and mixotrophic
  • * Williams & Unz (1989); Tandoi (1994); Aruga (2002); Rossetti (2003).

  • Richards (1985b); Williams & Unz (1989); Aruga (2002).

  • Nielsen (1998); Nielsen (2000); Nielsen (2003a, b).

  • § Andreasen & Nielsen (1997); Nielsen (1998); C. Kragelund & P.H. Nielsen (unpublished data).

  • ¶ Not all isolates or filaments.

  • ( ), reduced uptake.

  • For abbreviations, see Table 3.

Others have analysed aspects of the ecophysiology of Thiothrix spp. in activated sludge samples, as shown in Table 5 (Nielsen, 1998, 2000, 2003b). Of six substrates tested, only acetate and ethanol uptake was observed, while glucose was never assimilated (Nielsen, 1998), in accordance with data from the later pure culture studies. The heterotrophic, mixotrophic and chemolithotrophic activity in an industrial treatment plant sample of one Thiothrix species closely related to T. ramosa has been reported in more detail (Nielsen, 1998, 2000). Both acetate and/or bicarbonate uptake were stimulated by the presence of thiosulphate, with the highest activity observed under aerobic conditions. However, some activity was also detected with nitrate as an e-acceptor and under strictly anaerobic conditions when cells possessed intracellular sulphur granules (Nielsen, 2000). The filaments had a high affinity for acetate (low Km value), but a relatively low substrate uptake rate (Nielsen, 2003b). They were hydrophilic by MAC assays, formed polyhydroxyalkanoates, but only as small granules, and did not appear to excrete exoenzymes (Kragelund, 2005).

Filaments with the 021N morphotype and targeted by probe 021N (the group II members as defined by Kanagawa (2000), corresponding to T. eikelboomii) have also been studied in this way. Only a limited number of substrates were tested and some variation in uptake capability by the probe-defined morphotype 021N (T. eikelboomii) was seen with these filaments in samples from four different treatment plants. Acetate and glucose were taken up, and in half of the plants some assimilation of oleic acid, ethanol, glycine and leucine was observed under the experimental conditions used. Only uptake under aerobic conditions was tested. Assimilation of several SCFAs, sugars and amino acids has been reported for the FISH probe-defined 021N morphotype, but only under aerobic conditions (C. Kragelund & P.H. Nielsen, unpublished data). No data exist for any other members of the morphotype 021N groups.

A comparison of the physiological features of pure culture of Thiothrix species with those obtained in situ shows some agreement. Thus, under both conditions, active uptake of acetate alone and in combination with thiosulphate and/or bicarbonate could be demonstrated (Nielsen, 2000). Some are also capable of autotrophic growth. Uptake of organic acids was also seen in pure culture experiments, and moreover, sugars were never used (Williams & Unz, 1989; Aruga, 2002). An ability to reduce nitrate to nitrite occurred in both, while substrates were only taken up under strictly anaerobic conditions in situ (Nielsen, 2000). However, only a few substrates were tested with the in situ experiments. What is clear is that the pure culture and in situ physiological data for morphotype 021N (T. eikelboomii) show that they are nutritionally very versatile and flexible filaments, assimilating a wide range of SCFAs, sugars, amino acids and alcohol under aerobic conditions.


Filamentous members of the Bacteroidetes phylum have been isolated from many environments including salt marshes (Lydell, 2004), marine environments (Kämpfer, 1995; Eilers, 2001) and activated sludge (Williams & Unz, 1985b; Kämpfer, 1995). The phylum Bacteroidetes [formerly the Cytophaga–Flavobacter–Bacteroides (CFB) phylum] contains three classes, Bacteroidetes, Flavobacteria and Sphingobacteria, with the latter consisting of the genera Cytophaga and Flexibacter (Garrity, 2002).

Only members of two Bacteroidetes genera have been isolated from activated sludge: H. hydrossis (Van Veen, 1973) and two isolates affiliated within the Cytophaga subgroup with a morphotype of type 1863 (Seviour, 1997). Morphotypes 0411 and 0092 have also been placed in this phylum based on partial 16S rRNA gene sequence analysis obtained on micromanipulated filaments subsequently grown in a pure culture (Bradford, 1996). However, the phylogenetic affiliation of the type 0092 morphotype has now been confirmed as belonging to the Chloroflexi, and FISH probes have been designed for its in situ identification (Speirs, 2009). Furthermore, C.‘Magnospira bakii’, a motile filament with a corkscrew-shaped appearance, belonging to the family Flexibacteraceae, is also occasionally present in low amounts in activated sludge (Snaidr, 1999).

Several FISH probes have been designed to target members of the Bacteroidetes (Wagner, 1994a; Manz, 1996; Snaidr, 1999; Schauer & Hahn, 2005). The species-specific probe (HHY) targeting H. hydrossis was based on the type strain (Wagner, 1994a) and a broader probe (HHY-654) was designed against both the H. hydrossis type strain and a sequence from a micromanipulated filament (Kragelund, 2008). Two additional probes have also been designed based on cloned 16S rRNA gene sequences from extracted DNA from a biofilm sample (CFB655 and CFB730) (Kindaichi, 2004). The abundances in activated sludge of these H. hydrossis-like bacteria have been investigated only in a few Danish WWTPs, and no FISH-defined populations were detected here (C. Kragelund & P.H. Nielsen, unpublished data). The probes do not target sequences within the Saprospiraceae family to which H. hydrossis belongs (blast search using the Ribosomal Database Projects website, http://rdp.cme.msu.edu/). The identity of the most abundant filamentous Bacteroidetes, including H. hydrossis, in activated sludge can be demonstrated adequately with probes CFB719, SAP309 (family Saprospiracee) and HHY-654 (Kragelund, 2008).

Filamentous Bacteroidetes are very common in activated sludge plants, but rarely cause bulking problems, even though they may have a slight effect on biomass settling properties (van der Waarde, 2002; Kragelund, 2008). The needle-shaped H. hydrossis and H. hydrossis-related filaments are the most commonly seen Bacteroidetes in industrial and municipal treatment plants (Kragelund, 2008).

Interestingly, epiphytic growth of bacteria belonging to Saprospiraceae members of the Bacteroidetes is often observed on some filamentous Chloroflexi, filamentous members of the phylum TM7, and several other filaments in activated sludge (Xia, 2007). The most abundant of these epiphytic bacteria were named Candidatus‘Epiflobacter spp.’, and evidence suggests they are actively involved in protein hydrolysis and amino acid consumption (Xia, 2008).


Only one study (Kragelund, 2008) has focused on the ecophysiology of filamentous members of Bacteroidetes in activated sludge (Table 6). Probe-defined H. hydrossis and other Bacteroidetes appear to be highly specialized bacteria involved in polysaccharide degradation, as suggested by their ability to uptake glucose and N-acetylglucosamine, a monomer of the glycan of bacterial cell wall peptidoglycan. They are aerobic, and no anoxic or anaerobic substrate assimilation activity has been observed. Many surface-associated exoenzymes are synthesized by these hydrophilic filaments, including chitinase, glucuronidase, esterase and phosphatase, supporting the proposition that they are involved in metabolizing polysaccharides and possibly bacterial cell debris. Most filaments are hidden within flocs, although they occasionally extend into the bulk liquid. A similar ecophysiology was also reported for filamentous Bacteroidetes in a biofilm originating from activated sludge (Kindaichi, 2004; Okabe, 2005), where probe-defined filaments showed a substantial uptake of N-acetylglucosamine and mixtures of amino acids.

View this table:
Table 6

Comparison between pure culture physiology and in situ physiology for filamentous Bacteroidetes in activated sludge

ProbePure cultureIn situ
H. hydrossis H. hydrossis CFB environment IsolatesAutotrophic biofilm
HHYCFB 719CFB655, CFB730, CF319a+b
Use of e-acceptorsO2O2O2Only O2 tested
Storage compounds (PHA)PresentLowLowND
Surface propertiesNDNDHydrophilicND
ExoenzymesNDChit, Glucu, Est, PhosChit, GlucuND
Short chain fatty acidsProp(Prop)
Long chain fatty acidsND
SugarsFru, GluGluND
Di- and polysaccharidesStarchNDNDND
Amino acids
Amino acid mixtureNDND+
Bicarbonate with thiosulphateNDND
  • * Kämpfer (1995).

  • Kragelund (2008).

  • Kindaichi (2004); Okabe (2005).

  • § Observed in one study/WWTP only.

  • ( ), reduced uptake.

  • For abbreviations, see Table 3.

Many pure culture studies have been conducted with H. hydrossis, as illustrated in Table 6 (Van Veen, 1973; Krul, 1977; Ziegler, 1990; Mulder & Deinema, 1992; Kämpfer, 1995). These data generally show that glucose, N-acetylglucosamine, d-glucosamine and a few other compounds are utilized. Also, some data on type 1863 isolates affiliating to the Chryseobacterium have been investigated and showed utilization of a range of polysaccharides and some amino acids (Seviour, 1997). However, no in situ data exist, and so it is not possible to compare these. Both the pure culture and the in situ data suggest that H. hydrossis is a specialized aerobic bacterium occurring in both municipal and industrial WWTPs. Uptake of N-acetylglucosamine and excretion of chitinase suggest an active involvement in the degradation of bacterial cell wall components such as lipopolysaccharides and peptidoglycan. Thus, they seem to be autochthonous members of the activated sludge community, and it might prove a difficult task to remove those substrates sustaining their growth and hence limit their abundance there.


Members of the phylum Chloroflexi, previously known as the green nonsulphur bacteria (Garrity & Holt, 2001), were once associated primarily with extreme habitats, including microbial mats in hot springs (Boomer, 2002; Nubel, 2002) and hypersaline environments (Nubel, 2001), where they behave as (filamentous) anoxygenic phototrophs (Hanada & Pierson, 2002; Hanada, 2002; Nubel, 2002). However, nonphototrophic filamentous Chloroflexi are common in activated sludge plants, especially those removing phosphate by enhanced biological phosphorous removal (EBPR) (Beer, 2002; Bjornsson, 2002; Speirs, 2009) and in some industrial WWTPs with a simple configuration (Kragelund, 2007a), and they have been associated occasionally with bulking episodes (Schade, 2002; Jenkins, 2004). The phylum Chloroflexi currently contains two classes: the Chloroflexi and the anaerobic Anaerolineae.

Several filamentous Chloroflexi have been isolated successfully by micromanipulation from activated sludge (Kohno, 2002; Kragelund, 2007a). For example, 16S rRNA gene sequence analyses of many sharing the Eikelboom type 1851 morphotype belong to the Chloroflexi (Beer, 2002), and five type 1851 isolates were named Kouleothrix aurantiaca (Kohno, 2002). Filamentous members of Herpetosiphon affiliating within Chloroflexi have also been identified in activated sludge (Senghas & Lingens, 1985; Reichenbach, 1992).

FISH probes are available that target members of this phylum and two of the four subdivisions (Bjornsson, 2002) as well as more specific probes for the morphotype 1851, probe Chl1851 (Beer, 2002) and probe EU25-1238 (Kragelund, 2007a). Probes targeting two morphological variants of type 0092 differing consistently in their trichome diameters are also available (Speirs, 2009). Another probe AHW183 was designed to target an N. limicola morphotype belonging to the Chloroflexi (Schade, 2002), but no accession number is available for the sequence. Surveys also suggest that this type is rarely seen in plants and it will not be discussed further here. An additional probe GNS667 was designed from cloned sequences recovered from DNA extracted from a biofilm sample (Kindaichi, 2004). The abundance of filamentous Chloroflexi targeted by this probe in activated sludge is presumably low (C. Kragelund, unpublished data). Evidence suggests that Herpetosiphon seems to play little or no role in bulking or foaming incidents, and hence no further data on it are included here.

Particularly important in any discussion on FISH detection of Chloroflexi in activated sludge is that some (filaments and single cells) lack the target sites for all the EUBmix probes used to detect all (or most) Bacteria (Kragelund, 2007a; Morgan-Sagastume, 2008; Speirs, 2009). Hence, these filaments can easily be overlooked in FISH-based studies. However, this property is a useful diagnostic characteristic for screening 16S rRNA gene clones prepared from activated sludge, and helped to show that some type 0092 members are Chloroflexi (Speirs, 2009).

Filamentous Chloroflexi are commonly seen in full-scale municipal and industrial WWTPs, but as they are usually hidden inside the flocs, they rarely cause bulking (Kragelund, 2007a). They frequently harbour Saprospiraceae epiphytic bacteria (C. ‘Epiflobacter spp.’) (Xia, 2007, 2008), a diagnostic characteristic helpful in identifying them microscopically. However, in higher numbers, they form distinctive bundles, which, as interfloc bridges, encourage bulking, and then epiphytic bacteria are generally not so abundant (Beer, 2002). Chloroflexi filaments seem to be robust organisms and only degrade slowly, a feature that might be ascribed to the presence of amyloid proteins on their cell surfaces (Larsen, 2008). These proteins may also function as attachment sites for the epiphytic bacteria.


Ecophysiological studies (Kindaichi, 2004; Okabe, 2005; Kragelund, 2007a; Miura, 2007; Miura & Okabe, 2008) reveal that the Chloroflexi constitute a nutritionally specialized group of filamentous bacteria consuming primarily carbohydrates, such as glucose and N-acetylglucosamine, and amino acids (Table 7). Most of their probe-defined populations also use butyrate and some pyruvate as well, but never acetate. From the available data, they only appear generally to take up substrates under aerobic conditions. The filaments usually contain polyhydroxyalkanoates granules, but in small amounts, and many exoenzymes including chitinase, glucuronidase, galactosidase (Kragelund, 2007a), as well as proteases (Xia, 2007) are excreted, suggesting that these grow primarily on complex polysaccharides and proteins. Surfaces of Chloroflexi filaments from MAC assays appear to be more hydrophilic than those of many other filaments in activated sludge. Biofilm studies on filamentous Chloroflexi detected by the probe GNS667, which targets sequences within the Anaerolineae, support these observations (Kindaichi, 2004; Okabe, 2005). Filamentous Chloroflexi originating from a membrane bioreactor assimilated several sugars, mixtures of amino acid, protein hydrolysate and bacterial detritus under aerobic conditions (Miura, 2007; Miura & Okabe, 2008). However, some uptake of glucose and mixed amino acids was still observed under anoxic conditions with nitrate and under strict anaerobic conditions (Miura, 2007).

View this table:
Table 7

Comparison between pure culture physiology and in situ physiology for filamentous Chloroflexi in activated sludge

ProbePure cultureIn situ
K. aurantiaca EU25 & Ver9Iso2K. aurantiaca likePhylumSubdivision 3Type 1851Autotrophic biofilmMembrane bioreactor
EU25-1238Pure cultureCFX1223CFX109Chl1851+EU25-1238GNS667, CFX mixCFX mix
Use of e-acceptorsO2, fermentationO2, fermentationO2, (NO3)O2O2O2Only O2 testedO2,(NO3), (anaerobic)
Storage compounds (PHA)NDNDLowLowLowLowNDND
Surface propertiesNDNDHydrophilicHydrophilicNDHydrophilicNDND
ExoenzymesNDNDEst, GalaChit, Glucu ProtND, ProtEst, (Glucu), (Gala)NDND
Short-chain fatty acidsPyrAce, Pyr(Ace, But, Pyr)(But)(Pyr), (But)
Long-chain fatty acidsND
SugarsGlu, LacFruc, Glu, LacGlu, ManGlu, (Gal), (Man)GluGluNDArabinose, Fucose, Gal, Glu, Lac, Man
Di- or polysaccharidesNDNDNDNDND
Amino acidscas
Amino acid mixturecasNDND+ (mix)+ (mix)
Bicarbonate with thiosulphateNDND
MiscellaneousYeast extractProtein hydrolysate, bacterial detritus
  • * Kohno (2002).

  • Kragelund (2007a).

  • Kindaichi (2004); Okabe (2005).

  • § Miura (2007); Miura & Okabe (2008).

  • ¶ Observed in one study only.

  • ++No preincubation with unlabelled substrate was performed; ( ), reduced uptake.

  • For abbreviations, see Table 3.

There is some disagreement between pure culture data and ecophysiological data for type 1851-related strains (Table 7). A range of substrates were utilized by the five isolates obtained by Kohno (2002) including glucose, lactose and pyruvate. In contrast to the in situ data, two of these strains reduced nitrate to nitrite and grew anaerobically on glucose and fructose. Furthermore, a detailed study with isolates EU25 and Ver9Iso2 (Kragelund, 2007a) belonging to the same species, K. aurantiaca, showed that they were also capable of anaerobic growth on R2A agar, and a number of sugars including glucose, fructose, lactose and galactose were also utilized under aerobic conditions. In addition, several SCFAs acetate, propionate and pyruvate supported growth, which contrasts with the in situ data (Kragelund, 2007a).

Thus, the filamentous Chloroflexi appear to be specialized bacteria, capable of degrading complex macromolecules such as polysaccharides and proteins. Uptake of soluble substrates has only been observed under aerobic conditions, indicating that the fermentative metabolism observed in pure culture studies probably does not play an important role in these organisms in activated sludge. The uptake of N-acetylglucosamine, combined with detection of chitinase activity, reveals an ability to use N-acetylglucosamine units found in lipopolysaccharides, peptidoglycan and chitin (Kindaichi, 2004; Okabe, 2005; Kragelund, 2007a; Miura, 2007), and certain to be present in activated sludge. The Cytophagales also contain many chitinolytic members (Reichenbach, 1999). Galactosidase and glucuronidase activity in the Chloroflexi filaments is consistent with their ability to assimilate sugars and polysaccharides. Surface-associated esterase activity has also been detected in association with some Chloroflexi by ELF (C. Kragelund & P.H. Nielsen, unpublished data).

Candidate phylum TM7

Members of the Candidate division TM7 are widely distributed in nature, and have been detected in soil, groundwater, seawater, mouse faeces, the human oral cavity and activated sludge (Hugenholtz, 1998). Based on environmental 16S rRNA gene sequences, three subdivisions within this phylum have been defined, and primers and FISH probes are available that target members of the subdivision 1 and almost the entire phylum (Hugenholtz, 2001). Some filamentous bacteria respond to the subdivision FISH probe when applied to activated sludge (Hugenholtz, 2001). These were ‘identified’ as the Eikelboom morphotype 0041/0675 based on their morphological attributes, as they often possess epiphytic bacteria on their surface (Xia, 2008). However, bacteria with the type 0041/0675 morphotype also belong to the Chloroflexi and the Curvibacter-related bacteria within the Betaproteobacteria (Thomsen, 2006b). Thus, it is extremely difficult to identify these adequately solely from their morphology, and Thomsen (2002) have shown that only 15% of the 0041/0675 morphotype population hybridized with the TM7 subdivision probe.


The ecophysiology of TM7 probe-defined populations of morphotype 0041/0675 was investigated in two Danish treatment plants (Table 8) (Thomsen, 2002). These populations could take up several monosaccharides and leucine under aerobic conditions. In addition, some of the FISH-probed filaments consumed galactose, while only a few assimilated glycine. No uptake of acetate was observed. Under anoxic conditions with nitrate as an e-acceptor, glucose and galactose uptake levels were comparable to those achieved under aerobic conditions (as assessed by the density of silver grains), and glucose was also utilized under strictly anaerobic conditions. However, it is unclear whether these are fermentative organisms. Unfortunately, no pure culture members of the TM7 division exist (Hugenholtz, 2001). Nine strains of morphotype 0041 have been reportedly isolated on the basis of their morphology, but no information about their exact identity and nutritional requirements has been forthcoming (Williams & Unz, 1985b).

View this table:
Table 8

Comparison between in situ studies of TM7 defined type 0041 and nontargeted type 0041

ProbeIn situ
Type 0041Type 0041
Use of e-acceptorsO2, (NO3), (NO2), (anaerobic)O2
Storage compounds (PHA)PresentND
Surface propertiesNDND
Short-chain fatty acids
Long-chain fatty acids(Oleic)
SugarsGal, Glu, ManGlu
Di- or polysaccharidesNDND
Amino acids(Gly), Leu(Leu)
Bicarbonate with thiosulphateNDND
  • * Thomsen (2002).

  • Xia (2007).

  • Andreasen & Nielsen (1997).

  • ( ), reduced uptake.

  • For abbreviations, see Table 3.


Filamentous bacteria with a streptococcus-like morphology are occasionally observed in activated sludge plants (Eikelboom & Geurkink, 2002), leading to the assumption that these belong to the genus Streptococcus in the class Bacilli and the family Streptococcaceae. This is supported by the isolation of eight strains obtained from different plants belonging to two different genera within the phylum Firmicutes, some closely related to Trichococcus flocculiformis (family Carnobacteriaceae) and others to Streptococcus (Liu, 2000, 2002). All isolates had the morphology described for N. limicola type I. Probes have been developed for some of the isolates (Liu & Seviour, 2001). The T. flocculiformis isolates have been characterized in pure culture, and growth on glucose produced lactate, acetate (oxically) and also ethanol and formate anoxically (Liu, 2002). In general, some sugars were used (sucrose, mannose and cellobiose) and some strains could also reduce nitrate. No in situ study has been carried out.

A recent study on identifying the active fermentative bacterial populations in activated sludge revealed that probe-defined filamentous bacteria hybridizing with probe Strept (denoted Str in probeBase) designed for streptococci and closely related genera (Trebesius, 2000) were commonly seen in plants with phosphate removal, constituting 1–4% of the biovolume (Kong, 2008). These are facultative anaerobes, utilizing glucose under both aerobic and anaerobic conditions (but not mannose or galactose), and unable to consume acetate (Y. Kong & P.H. Nielsen, unpublished data). They probably ferment substrates in the anaerobic reactors, thus being important in providing readily metabolizable substrates for the denitrifying and phosphate-removing communities in these processes (Kong, 2008).


Two groups of filamentous members of the Actinobacteria are known to cause severe operational problems in activated sludge. These are the C. ‘Microthrix parvicella’ morphotype and several mycolic acid-producing species of the Mycolata. Comprehensive reviews have been published on both (Rossetti, 2005; Seviour, 2008); hence, they will not be discussed in such detail here. The third group of filamentous Actinobacteria implicated in foaming and bulking episodes in plants globally containing the N. limicola II morphotype (Blackall, 2000) are now classified as several species in the genus Tetrasphaera (McKenzie, 2006).

Candidatus‘Microthrix parvicella’

From current evidence, Candidatus‘Microthrix parvicella’ seems to occur exclusively in activated sludge plants, where it has been associated with both bulking and foaming incidents (Van Veen, 1973; Blackall, 1994; Eikelboom, 2000; Jenkins, 2004). This morphotype was named Microthrix parvicella (Van Veen, 1973), and several isolates have been obtained in pure culture (Van Veen, 1973; Eikelboom, 1975; Slijkhuis, 1983; Blackall, 1994; Tandoi, 1998). Phylogenetic analyses of some of these show that they affiliate to a deep branching member of the class Actinobacteria within the phylum of Actinobacteria, and the name C. ‘Microthrix parvicella’ was proposed (Blackall, 1996). Four FISH probes were designed to target this filament (Erhart, 1997). It does not hybridize with the actinobacterial phylum probe (HGC69a), having three mismatches with it (Erhart, 1997). Pretreatment before FISH significantly enhances cell wall permeabilization, and thus hybridization signal strength (Beimfohr, 1993; Erhart, 1997; Carr, 2005; Kragelund, 2007c).

More recently, another species, Candidatus‘Microthrix calida’, was isolated from industrial activated sludge samples (Levantesi, 2006). Its phylogenetic affiliation also falls within the unclassified Actinobacteria, although three of the strains affiliate within the genus Acidimicrobium. This filament appears as a thinner version of C. ‘Microthrix parvicella’, with 95.7–96.7% shared sequence similarity. FISH probes are available that target C. ‘Microthrix calida’ and all known Microthrix species (Levantesi, 2006).

Candidatus‘Microthrix parvicella’ seems to play an important role in bulking incidences worldwide, particularly in municipal and, to some extent, in industrial plants. Current evidence based on very few studies would suggest that C. ‘Microthrix calida’ is not common and probably plays no role in bulking incidents.


Several in situ investigations have been performed with the C. ‘Microthrix parvicella’ morphotype or FISH probe-defined populations, as can be seen in Table 9 (Andreasen & Nielsen, 1998, 2000; Nielsen, 2002, 2005; Hesselsoe, 2005). Only the assimilation of long-chain fatty acids (LCFAs) such as oleic acid, trioleic acid and palmitate has ever been observed by C. ‘Microthrix parvicella’in situ. Substrates are taken up under both aerobic and anaerobic conditions and stored as lipid reserves, but the filaments grow only under conditions with nitrate or oxygen as an e-acceptor. They show surface-associated esterase and lipase activity, and as they are highly hydrophobic, they are well adapted to hydrolysis, uptake and growth on lipids and greases (Nielsen, 2002). They appear to be sensitive to high oxygen tensions, suggesting a microaerophilic preference (Rossetti, 2005). No comparable in situ data are available for C. ‘Microthrix calida’.

View this table:
Table 9

Comparison between pure culture physiology and in situ physiology for filamentous Actinobacteria in activated sludge

ProbePure cultureIn situ
Candidatus ‘Microthrix parvicella’Candidatus ‘Microthrix parvicella’Candidatus ‘Microthrix calida’Candidatus ‘Microthrix parvicella’
Use of e-acceptorsO2O2,NO3O2, NO3O2, NO3, NO2β, anaerobic β
Storage compounds (lipidic compound)PresentPresentPresentLipid, PolyP
Surface propertiesNDNDNDHydrophobic
ExoenzymesNDNDEst, Gala, LipEst, Lip
Short-chain fatty acidsAce, Pyr
Long-chain fatty acidsTween 40, 60, 80OleateOleic, Pal, Triol
Di- or polysaccharidesNDND
Amino acidsCasCas
Bicarbonate with thiosulphateNDNDND
MiscellaneousR2AYeast extractMSVND
  • * Slijkhuis & Deinema (1982); Slijkhuis (1984).

  • Tandoi (1998); Rossetti (2005).

  • Levantesi (2006).

  • § Andreasen & Nielsen (1997); Andreasen & Nielsen (2000); Nielsen (2002).

  • ∥ Only positive growth on MSV agar without added carbon source

  • β, only substrate uptake but no growth.

  • For abbreviations, see Table 3.

In general, information obtained from the few pure culture experiments with existing isolates (Slijkhuis & Deinema, 1982; Slijkhuis, 1984) is comparable to the in situ data, but some inconsistencies also exist. Thus, Tandoi (1998) showed that C. ‘Microthrix parvicella’ grew well in an axenic culture on SCFAs, a feature that has never been demonstrated in situ (Table 9).


Members of the mycolic acid-containing Mycolata have been identified in many different environments, including soil, fresh water and marine habitats (Goodfellow, 1998), and contain pathogenic members that have been isolated from animals and humans (Roth, 2003), as well as from activated sludge foams (Eikelboom, 2000; Jenkins, 2004). The phylogeny of the Mycolata has been evolving as new members are described, and is reviewed elsewhere (Goodfellow & Maldonado, 2006). The Mycolata fall within the suborder Corynebacterineae within the class Actinobacteria and encompass members of several genera, all with a characteristic branching morphology. In activated sludge, species from the two genera Gordonia and Skermania are most frequently encountered, where they are often associated with severe foaming episodes. Their morphotypes are commonly described, respectively, as the right-angled branching Gordonia amarae-like organisms (Soddell, 1999) and the acute-angled pine tree-like organisms (Blackall & Marshall, 1989; Chun, 1997). However, FISH analyses suggest that neither always consistently shows this characteristic morphology (Stainsby, 2002; Kragelund, 2007c).

Several FISH probes target members of the Mycolata common in activated sludge (de los Reyes, 1997, 1998; Davenport, 2000; Eales, 2005). Detection by FISH is difficult because of the highly impermeable cell wall/cell envelope from the presence of mycolic acids. Pretreatment of cells is essential, and although several protocols have been published (Schuppler, 1998; Davenport, 2000; Carr, 2005; Kragelund, 2007c), the treatment described by Kragelund (2007c) generally works the best.

Many of the filamentous bacteria commonly observed in bulking and foaming samples from the United States (Richard, 1982; Strom & Jenkins, 1984), Italy (Rossetti, 1994) and Australia (Seviour, 1994) were members of the Mycolata. Gordonia spp. and Skermania piniformis appear to be among the most common, but some branched filaments in foams do not fluoresce with the FISH probes currently available to target these. Instead, they respond only to the broad Myc657-probe targeting most Mycolata (van der Waarde, 2002; Kragelund, 2007c), confirming what many have long believed is a large as yet undescribed diversity among these foaming Mycolata.


Studies on their ecophysiology have been carried out but restricted to just a few foaming Mycolata; see Table 10 (Eales, 2005, 2006; Carr, 2006; Kragelund, 2007c). Many appear to be metabolically inactive in situ, particularly in foam, and fail to fluoresce with the EUB mix probes. Often, the fluorescent signal is concentrated at the tip of the growing filaments. Nevertheless, the available evidence suggests that the foaming Mycolata are nutritionally more diverse than once thought in being able to assimilate a range of both hydrophilic and hydrophobic substrates. They also store large amounts of polyhydroxyalkanoates and polyphosphate, are generally hydrophobic by the MAC assay and excrete a battery of exoenzymes. Several clear differences emerge when the in situ data are compared with those from pure culture studies. For example, substrates such as oleic acid, on which G. amarae grows well in a pure culture, are not always assimilated in situ (Carr, 2006; Kragelund, 2007c). There is also strong evidence for in situ substrate uptake with nitrate as an e-acceptor, which questions whether G. amarae is obligately aerobic, as the pure culture data suggest (Chun, 1997; Stainsby, 2002). Similar conflicting data are reported for S. piniformis (Eales, 2006), and so the ecophysiology of Mycolata appears to be very complex. Clearly, more studies are needed to further elucidate their diversity and ecology.

View this table:
Table 10

Comparison between pure culture physiology and in situ physiology for filamentous Mycolata in activated sludge

ProbePure cultureIn situ
S. piniformis G. amarae S. piniformis S. piniformis G. amarae G. amarae S. piniformis Mycolata, only morphological identification
Use of e-acceptorsO2O2O2, (NO3)α, (NO2)α, anaerobicO2, NO3O2, (NO3)α, (NO2)α, (anaerobic)O2, (NO3), (NO2)O2, (NO3O2
Storage compoundsNDNDPHAPHANDPHA, PolyPPHA, PolyPND
Surface propertiesNDNDHydrophobicHydrophobicNDHydrophobicHydrophobicND
ExoenzymesEst, Est, Lip (API-ZYM)ND(Chit) Est, Glucu, Phos,(Chit), Est, Glucu, PhosNDLip, Est, PhosLip, Est, PhosND
Short-chain fatty acidsAceAce, But, Prop, Pyr, Suc, MalAceAce, (Prop)Ace, (Prop)Ace, Prop
Long-chain fatty acids/lipidsGlyce trioleate, Tween 20, 40, 60Glyce Tween 20, 40, 60OleicGlyce, Oleic, Pal, TriolGlyce, PalOleic, Pal, Triol
SugarsGlu, ManFru, Glu, Man, SucGluGlu
Di- or polysaccharidesSomeTreNDNDNDND
AlcoholsSugars alcoholsEthEth
Amino acidsNDGlyGlyGlyLeu, Mix
Bicarbonate with thiosulphateNDNDNDNDNDNDND
MiscellaneousYeast extractBenzoic acid, cholesterol
  • * Soddell & Seviour (1998); Blackall (1991).

  • Soddell & Seviour (1998).

  • Eales (2005).

  • § Eales (2006).

  • Carr (2006).

  • Kragelund (2007c).

  • α, reduced number of substrates are taken up.

  • ( ), reduced activity.

  • For abbreviations, see Table 3.


The actinobacterial member of the polyphyletic N. limicola morphotype (Seviour, 2002; Snaidr, 2002) has been shown to be a member of the genus Tetrasphaera in the family Intrasporangiaceae (McKenzie, 2006). Several isolates exist (Blackall, 2000). One FISH probe is available for the actinobacterial morhotype N. limicola II (Liu & Seviour, 2001), although several separate species are targeted by it (Seviour, 2006, 2008). This filament is often present, but in low quantities in activated sludge systems, and suggestions have been made that it may, on rare occasions, participate in foaming incidents (Wanner, 1994).

Ecophysiological studies on samples from plants in Australia and Japan suggest that it has a broad substrate assimilation profile, utilizing both hydrophilic and hydrophobic substrates, including LCFAs (see Table 11; Seviour, 2006; E.M. Seviour & A. Chua, unpublished data), in some cases aerobically, anoxically and anaerobically. However, in the latter case, whether these are used for growth or storage is not clear. Acetate is not assimilated, unlike the situation with pure cultures (Blackall, 2000). The formation of polyhydroxyalkanoates does not seem to be a universal trait among the actinobacterial N. limicola II (Liu, 2001) as opposed to the alphaproteobacterial N. limicola II morphotype (Kragelund, 2006). Thus, the nature of any storage material synthesized remains unclear, although some evidence has suggested a considerable polyphosphate storage capacity (Blackall, 2000; Liu, 2001).

View this table:
Table 11

Comparison between pure culture physiology and in situ physiology of Candidatus‘Nostocoida limicola’ in activated sludge

ProbePure cultureIn situ
N. limicola IIN. limicola IIN. limicola II
Use of e-acceptorsO2, fermentationO2, NO3O2, NO3, NO2, anaerobic
Storage compoundsNDLipophilic inclusionsPolyP
Surface propertiesNDNDND
Short-chain fatty acidsAceAce, Prop, Pyr
Long-chain fatty acids/lipidsOleicGlyce, Tween 80Glyce, (Pal)
SugarsGluGlu, Fru, Lac, Man, Suc
Di- or polysaccharidesNDNDND
Amino acidsNDND(Gly)
Bicarbonate with thiosulphateNDNDND
MiscellaneousPeptoneBenzoic acid
  • * Seviour (2006).

  • Blackall (2000).

  • ( ), reduced activity.

  • § § Only uptake under anaerobic conditions.

  • For abbreviations, see Table 3.


Five isolates with a morphology resembling that of N. limicola type III were obtained from an Australian WWTP. All isolates affiliated to the genus Isosphaera (family Planctomycetaceae), filamentous bacteria originally isolated from hotsprings (Liu, 2001). FISH probes were designed to target these isolates and enabled a differentiation between otherwise unrecognized bacteria with an indistinguishable morphology, but different phylogenetic affiliations (Liu & Seviour, 2001). No pure culture investigation of the isolates has been performed and no ecophysiological data exist. The abundance of the FISH probe-defined groups in activated sludge is not known.


From published surveys, several other morphotypes described in the identification manuals of Eikelboom (2000) and Jenkins (2004) are found frequently on a global basis, but their true identity is not yet resolved. An example is the morphotype 0914. The presence of sulphur granules inside cells suggests a mixotrophic behaviour, where oxidation of reduced sulphur compounds as energy sources is carried out simultaneously using organic substrates as carbon sources. This filament is not commonly associated with bulking or foaming (Eikelboom, 2000; Jenkins, 2004). Other morphotypes 0211, 0581, 0961 and 1852 appear to be less common, and in the absence of any pure cultures for these, no reliable information about their identity or physiology is available.

Ecophysiological groups and WWTPs

Numerous studies have screened for the presence of predominant and secondary filamentous morphotypes in differently configured WWTPs (e.g. Eikelboom, 2000; Jenkins, 2004). Based on the outcomes, it became clear that certain morphotypes were related usually to specific types of plants differing in their design and operation and influent wastewater characteristics. One good example is the occurrence of Thiothrix primarily in treatment plants with only carbon removal and with sulphide in the influent (Strom & Jenkins, 1984). The results from most of these survey studies were integrated by Wanner & Grau (1989) and Wanner (1994) and they proposed that the filaments could be divided into four groups: oxic zone growers S (consisting of morphotypes S. natans, type 1701, H. hydrossis and perhaps type 0041); oxic zone growers V (type 021N and Thiothrix); all zone growers A (Microthrix, type 0092 and N. limicola); and finally the group of foam-formers F [Microthrix, N. limicola, nocardioforms (now known as Mycolata) and possibly type 0041]. Unfortunately, the validity of any of these four groups has rarely been confirmed by FISH analyses. Hence, they can only indicate at best tentative relationships between the identity of filaments, their ecophysiology as well as their preference for treatment plants with certain operational features.

The information now available linking filament identity and in situ ecophysiology as critically evaluated in this review allows us to understand much better the ecological niches these bacteria may occupy in different WWTPs. Interestingly, what emerges from this work is that there seems to be a clear relationship for most groups between their phylogenetic affiliation and ecophysiology. For example, most filamentous Alphaproteobacteria share many common physiological traits of importance, which encourage their presence in certain WWTPs (Table 3). The same holds for filaments in the Chloroflexi and Bacteroidetes and other groups. This is helpful as it allows us to construct a new grouping of all suitably described filaments based on their in situ physiology, and this is presented in Table 12. However, despite the large apparent overall physiological similarities within one phylogenetic unit (e.g. the Alphaproteobacteria), differences will always be present between members of the different taxa (species, genera or families) and these may determine which strains are found in certain plants.

View this table:
Table 12

Overview of ecophysiology exhibited by the different physiological groups found in WWTPs

Group of filamentsVersatile, soluble substrate-dependent filamentsVersatile foam formersFermenting filamentous bacteria
Species and probesAlphaproteobacteria (probes for group 1 or 2)Thiothrix/type 021N (TNI and 21N)Mycolata (Skermania and Gordonia) (Gor596, G.am205 and Spin1449)Streptococci (Str/Strept)
Organic and inorganic substrates
Other SCFA+++ND
Other sugars++ND
Amino acids+++
N-acetyl -glucosamineNDNDNDND
Bicarbonate with thiosulphate+ND
e-acceptor conditions
None (anaerobic)(+)(+)+
ExoenzymesNone/few(Chit), Est, Glu, Phos, LipND
Specialized filamentous bacteria involved in degradation of complex matter
Group of filamentsPolysaccharide degradationLipid degradationProtein degradation
Species and probesBacteroidetes (CFB719, HHY, CFB655, CFB 730)Chloroflexi (EU25-1238, Chl1851, CFX1223, CFXmix)Candidatus ‘Microthrix parvicella’Curvibacter (Curvi997+ competitorsTM7 (TM7305)Chloroflexi (EU25-1238, Chl1851, CFX1223, CFXmix)
Organic and inorganic substrates
Other SCFA++++
Other sugars++++
Amino acids+++++
Bicarbonate with thiosulphateNDND
e-acceptor conditions
None (anaerobic)(+)(+)(+)
ExoenzymesChit, Est, Glu, PhosChit, Est, Gala, Glu, ProtLip, EstProtProtChit, Est, Gala, Glu, Prot
PHAVery lowSomeHigh, lipidsSomeSomeSome
  • * Not Skermania piniformis.

  • † Some species, not all.

  • ‡ Not all studies observe this.

  • LCFA, long-chain fatty acids; SCFA, short-chain fatty acids; (), reduced activity; α, reduced number of substrates are taken up; β, substrate uptake but no growth; ND, not determined; −, no substrate uptake; +, active substrate uptake.

The data show that one common feature of these filamentous organisms is that oxygen is their primary e-acceptor. Only Trichococcus/streptococci can also grow under anaerobic conditions. Most probably, none of the known filamentous bacteria denitrify. Some reduce nitrate to nitrite (C. ‘Microthrix parvicella’ and a few others) and either grow or obtain energy to assimilate substrates for polyhydroxyalkanoates storage. In addition, very few are active in the presence of nitrite (but not in its absence), but whether these can carry out denitrification is unknown. One interesting observation, however, is that for a given species, for example as with some members of the Alphaproteobacteria, the range of organic substrates assimilated under aerobic conditions is markedly reduced under anoxic conditions, and suggests that the ecological niches for such populations may be very complex.

A few probe-defined filamentous bacteria have not been allocated to any of the groups below as we lack the necessary ecophysiological information and most of them are of low abundance in WWTP communities as evaluated by the available FISH probes. These include the Acinetobacter (type 1863), Tetrasphaera-related N. limicola and Isosphaera-related N. limicola.

Versatile, soluble substrate-dependent filaments

Members of this group are nutritionally versatile, preferring soluble substrates for growth and therefore do not express exoenzymes. They are consumers of many substrates including acetate, have a medium to very high capability to store polyhydroxyalkanoates and are aerobic, although activity with nitrate as an e-acceptor may also occur. The group contains the filamentous Alphaproteobacteria and gammaproteobacterial Thiothrix/021N group. They can also use many different substrate groups under aerobic conditions, for example SCFAs, LCFAs, and several carbohydrates (glucose and others), amino acids and ethanol (Table 12). Thus, the presence of soluble substrates is essential for their persistence and excessive proliferation, and thus they cause serious bulking problems in industrial WTPPs. Type 021N and Thiothrix are rare in plants with denitrification, whereas the Alphaproteobacteria can be found there, consistent with their potential use of nitrate/nitrite as an e-acceptor. A remarkable ability to form polyhydroxyalkanoates from many substrates with different e-acceptors is a feature of the filamentous Alphaproteobacteria. They are generally hydrophobic, perhaps from the presence of amyloidic surface structures, and are occasionally responsible for foam formation. Some members of the Thiothrix/021N group also appear to depend on the presence of sulphides, as some have mixotrophic behaviour. This group contains both the N. limicola and Thiothrix/type021N morphotypes.

Specialized filamentous bacteria involved in the degradation of complex matter

The members of this group are highly specialized in relying on hydrolysis and consumption of complex macromolecules such as proteins, polysaccharides or lipids (Table 12). They express few or many different exoenzymes for their degradation, have low or no polyhydroxyalkanoates storage capability, large resistance to conditions of starvation and are usually only aerobic. They never consume acetate in situ.

Protein degradation

Some filaments belonging to Chloroflexi, Curvibacter-related populations and the TM7 group express proteases and consume amino acids. Some (probably not Chloroflexi members) can take up substrates under conditions with nitrate or nitrite as e-acceptors, and with some Curvibacter-related filaments, under strictly anaerobic conditions. They generally form small polyhydroxyalkanoates granules (as seen with the Curvibacter and the Chloroflexi, but TM7 has never been examined for its formation). Often, but not always, many protein-degrading epiphytic C. ‘Epiflobacter spp.’ are attached to their trichomes. The filamentous bacteria in this group are always hydrophilic and appear to play a structural role inside flocs where they are rarely abundant, and so can be easily overlooked if FISH is not applied. If they become more numerous (filament index, FI<2), interfloc bridging can occur and biomass settling properties are affected. This group contains types 1701, 1851 and 0041 morphotypes.

Polysaccharide degradation

Most filamentous members of Bacteroidetes and Chloroflexi express exoenzymes including glucuronidase, galactosidase and chitinase, esterase and sometimes lipases. Different monosaccharides are consumed, for example glucose and in some Bacteroidetes N-acetylglucosamine and propionate, but again never acetate. Activity only occurs under aerobic conditions. Some of these filaments may form polyhydroxyalkanoates (only small granules for Chloroflexi and absent in Bacteroidetes). They are hydrophilic and in the case of the Chloroflexi and Bacteroidetes (primarily H. hydrossis) filaments are located within or protrude from hydrophilic sludge flocs. Few other bacteria in activated sludge can degrade N-acetylglucosamine, a component of lipopolysaccharides and peptidoglycan constituting the bacterial cell wall. This group contains Haliscomenobacter and the type 1701, 1851 and 0041 morphotypes.

Lipid degradation

Candidatus‘Microthrix parvicella’ is a specialized lipid degrader able to express lipase and esterase activity in situ and assimilate LCFA under all e-acceptor conditions. Under anaerobic conditions, substrate uptake is used exclusively for the synthesis of lipidic storage inclusions and not for growth. Filaments have a hydrophobic surface that may aid in substrate accessibility. Not too surprisingly, C. ‘Microthrix parvicella’ is found primarily in municipal BNR plants that operate with anaerobic: aerobic biomass recycling, where it is often responsible for bulking and foaming incidents.

Versatile foam formers

These bacteria include the Gram-positive Mycolata (Gordonia, Skermania and others) and form stable foam as a consequence of having very hydrophobic surfaces. Currently, it is not possible to resolve totally the phylogenetic diversity among the Mycolata in activated sludge and foam using FISH. The reasons for this include an insufficient permeabilization of cells and lack of species-specific FISH probes.

From what is known, most are highly versatile in their substrate uptake profile, assimilating SCFA, monosaccharides, amino acids and LCFA under aerobic conditions. They often express exoenzymes consistent with an ability to degrade lipids and polysaccharides. These Mycolata can store large amounts of polyhydroxyalkanoates. Uptake of substrates with nitrate and nitrite as e-acceptors has been observed, but whether they can denitrify is unclear. Expression of lipase and esterase and sometimes glucuronidase and phosphatase has been observed, particularly in the foam itself, perhaps because the more readily assimilable substrates are not found there. These bacteria grow primarily in the mixed liquor and not in the foam (C. Kragelund & P.H. Nielsen, unpublished data). They are rarely seen in bulking incidents, but are frequently held responsible for episodes of foaming, particularly in conventional plants without nitorogen removal. Candidatus‘Microthrix parvicella’ is not placed in this group as foam is not always formed by it and its physiology is very different.

Fermenting filamentous bacteria

Among the taxa containing filamentous bacteria able to grow under anaerobic conditions in activated sludge, only the streptococci (Firmicutes) and other closely related species have been reported. They are facultative anaerobes, unable to assimilate acetate, but consume glucose (but not mannose or galactose) under both aerobic and anaerobic conditions. They probably carry out fermentation under anaerobic conditions, but little other in situ physiological information is available. The Tetrasphaera-related N. limicola may also ferment as other members of this genus are facultative fermentative bacteria.

Presence of filamentous species in differently configured WWTPs

Several operational characteristics of treatment plants are used to group them in relation to potential bulking problems, as these largely determine their ecosystem characteristics. These parameters include levels of plant organic loading, the presence of anoxic/anaerobic tanks (processes) and the type of wastewater. Loading is usually expressed as high- or low-loaded plants (or high/low food/microorganisms ratio or F/M ratio), which are also referred to as short/long sludge age (biomass retention time) plants, respectively. The presence of anoxic or anaerobic conditions selects for denitrifying and anaerobic populations, and such configured processes are often low loaded (long sludge age). Furthermore, the type of wastewater, industrial (primarily soluble) or municipal (primarily particulate), is considered influential. In the following discussion, we will refine these descriptions and attempt to indicate as to which physiological filament groups typically thrive there and why, explanations based in most cases on the authors' personal experiences, because little reliable data have been published for many of the filamentous bacteria.

1. Conventional WWTPs without nitrogen removal (carbon removal±nitrification, but no denitrification) treating soluble industrial wastewater. These plants are usually highly loaded and operate with low sludge age. This treatment plant type mostly contains bacteria from the ‘versatile, soluble substrate-dependent filament’ group, i.e. the Alphaproteobacteria and Thiothrix/type 021N, especially if oxygen depletion occurs, sulphide enters the plant or there is nutrient (nitrogen, phosphorus or micronutrient) limitation. We have occasionally observed unusual partly unidentified filaments in such plants treating particular organic substrates, and in those running at high temperatures or at high salinity. They are often atypical Alphaproteobacteria or in some cases Chloroflexi without epiflora. Undoubtedly, many unidentified filaments can be selected for in such plants, as was suggested by Eikelboom (2006).

2. Conventional WWTPs without nitrogen removal (carbon removal±nitrification but no denitrification) treating primarily municipal wastewater. These plants are usually medium/low loaded with medium/long sludge age. Most substrates are only available after hydrolysis of complex macromolecules from particulate and colloid substrates. Such plants contain mainly bacteria from the ‘specialized filamentous bacteria involved in degradation of complex matter’ group, such as members of Bacteroidetes, Chloroflexi, Curvibacter and TM7 filaments. Candidatus‘Microthrix parvicella’ from this group are rarely seen as they thrive best in BNR plants with anaerobic/anoxic conditions. These conventional plants also often contain Mycolata from the group of ‘versatile foam formers’, and occasionally, if the fraction of soluble substrate is large and oxygen depletion occurs or sulphide enters the plant, also Thiothrix/type 021N.

3. BNR treatment plants (nitrogen removal±biological phosphorous removal) treating soluble industrial wastewater. These plants are usually medium loaded with a medium sludge age. Little is known about the filament populations in these plants, but members of Alphaproteobacteria and Chloroflexi (without attached epiphytes) may occur.

4. BNR municipal treatment plants (nitrogen removal±biological phosphorus removal). These plants are usually low loaded with long sludge ages. Most substrates are available after hydrolysis of complex macromolecules in the form of particulates and colloid forms. This type of plant contains bacteria from ‘specialized filamentous bacteria involved in degradation of complex matter’ group, and primarily Bacteroidetes, Chloroflexi, Curvibacter, TM7 and C. ‘Microthrix parvicella’. Such plants rarely contain many Mycolata.

Ecophysiology and control of bulking in WWTPs

Filamentous bulking and foaming can be controlled by nonspecific measures or by addressing the causes of individual filament growth and then developing targeted control strategies against each. As most filamentous organisms in activated sludge can be placed into four groups with shared ecophysiologies, the reasons why different filamentous bacteria may grow in different types of treatment plants become clearer and, importantly, provide clues as to how they might be controlled.

The group of physiologically versatile bacteria growing on soluble substrates mainly cause problems in conventional plants without nitrogen removal, treating industrial wastewater with a low fraction of particles. The Alphaproteobacteria have an extremely high substrate uptake and polyhydroxyalkanoate-storage capacity and can cause severe bulking problems. Thiothrix/type 021N can be controlled by including an nitrogen removal stage and/or incorporating designed tanks (selectors), to encourage the floc-formers to assimilate most of the soluble substrate (Wanner, 2000; Martins, 2004). This is best achieved in a selector using nitrate or nitrite as the e-acceptor. Under such conditions, it is important to ensure that diffusion limitation of the floc-formers is not a problem (Martins, 2004). Removal of any sulphide in influent wastewater is important. A similar approach can be used to some extent to control Alphaproteobacteria, but is not always successful. We know of several cases where a denitrification step is of no influence in their control, probably because they store large amounts of substrate as polyhydroxyalkanoates under denitrifying conditions.

All the specialized filamentous bacteria degrading complex organic matter are more abundant in low-loaded conventional or BNR plants with a large fraction of particulate substrates in the influent. As few floc-formers seem able to degrade complex macromolecules (Nielsen, 2002; Xia, 2007), these hydrolysing filamentous organisms (Bacteroidetes, Chloroflexi, Curvibacter, TM7 and C. ‘Microthrix parvicella’) should compete very effectively with the floc-formers. When present in industrial treatment plants with a large fraction of soluble substrates, it is most likely that they will survive by growing instead on complex substrates produced by other bacteria, as described with an autotrophic biofilm (see Table 7), and will rarely occur in high numbers. In general, these filamentous bacteria (except C. ‘Microthrix parvicella’) rarely cause severe bulking problems, but at high abundances may contribute towards an open floc structure with impaired settling properties. It is difficult to control these filaments by changes in plant operation or design. Candidatus‘Microthrix parvicella’ might be controlled potentially by removing its preferred lipid substrates from the wastewater, but more practically by adding chemicals (polyaluminium chloride, Roels, 2002; Nielsen, 2005) that probably inhibit the activities of their surface-associated lipases (Nielsen, 2005). As they are known to be microaerophilic, operating at higher oxygen concentrations may also control C. ‘Microthrix parvicella’ (Rossetti, 2005). The Bacteroidetes filaments never cause serious operational problems, whereas Chloroflexi species are relatively common, and in some cases, bulking problems can arise (Kragelund, 2007a). Chloroflexi are capable of protein and polysaccharide hydrolysis and degradation; hence, it is difficult to remove their preferred substrates from the influent. Consequently, no efficient control measures are known for these or for the TM7 and Curvibacter-related filaments. Increasing sludge age in order to increase biomass levels relative to those of incoming particulate substrates was suggested to reduce their growth (Wanner, 1994), but how successful it might be is unclear. The fermentative Gram-positive streptococci have only been detected in EBPR plants with anaerobic tanks. We have no observations or records to believe that these or Trichococcus cause bulking (we have seen it as a bulking filament in lab-scale reactors occasionally).

The actinobacterial foam-forming Mycolata (Gordonia, Skermania, etc.) are very versatile in their substrate assimilation profiles and can also express exoenzymes for lipid and polysaccharide degradation (Eales, 2005, 2006; Carr, 2006; Kragelund, 2007c). It was believed for a long time that they prefer to grow on lipids, but recent data suggest that some grow on soluble substrates including acetate and glucose. It is also important to repeat that they grow only in the mixed liquor, and not in the foam. Selectors (anoxic or anaerobic) may control some of them, while removal of lipids may control others, although available published data suggest that because of their considerable phylogenetic diversity, a single control strategy may not be feasible. Polyaluminium chloride, which selectively eliminates C. ‘Microthrix parvicella’, does not work with the Mycolata (C. Kragelund & P.H. Nielsen, unpublished data), but other chemicals under development and trial appear to be promising (Kragelund, 2007b). Alternatively, foam removal as soon as it appears has been successful in preventing further episodes in some plants, but not in others (Jenkins, 2004). Published reports on the influence of changes in sludge age or removal of lipids on foaming are contradictory, most probably because no reliable identification of the causative organisms was performed, and in each study, it seems likely that different Mycolata may have been responsible.

Research needs

As this review has shown, over the past 10–15 years, we have achieved great improvements in understanding the identity, ecophysiology and ecology of many of the dominant filamentous bacteria in activated sludge communities. However, some important information is still missing before a comprehensive understanding of the entire activated sludge ecosystem and the functions of the filamentous organisms in it can be attained. The phylogenetic diversity of some groups is still poorly resolved, and only broadly targeted oligonucleotide FISH probes exist for many. Although large phylogenetic groups such as the Alphaproteobacteria contain many species with a relatively similar physiology and ecology, the same level of shared properties is absent from many others. For example, we need a much better description of the diversity among the Chloroflexi and better probes to detect more of the abundant populations, because distinctively quite different physiologies are exhibited by its members in spite of some overall similarities. The same situation exists with the Mycolata.

Furthermore, the identity and ecophysiology of the still unidentified filament morphotypes is needed. This is particularly true for type 0803 and a few others. Some type 0092 belong to the Chloroflexi (Speirs, 2009), but this morphotype also seems to include several phylogenetically unrelated populations.

The ultimate goal of most of the research presented here is to be able to control these populations in activated sludge systems at a level where bulking and foaming problems rarely occur. Therefore, we need more full-scale plant trials with reliable and, if possible, quantitative analyses of the important populations (including the filaments). The ecological factors responsible for controlling individual population densities of the activated sludge community and why different filaments can reach excessive abundances may then become clearer. Besides monitoring changes in process design or operation, development of novel targeted control methods should be examined, including phage biotherapy and addition of effective ‘green’ chemicals.


This study was supported by Aalborg University. We thank S. McIlroy for making the phylogenetic trees.


  • Editor: Bernardo González


View Abstract