search
for
 About Bioline  All Journals  Testimonials  Membership  News


Australasian Biotechnology (backfiles)
AusBiotech
ISSN: 1036-7128
Vol. 10, Num. 3, 2000, pp. 24-29
Untitled Document

Australasian Biotechnology, Vol. 10 No. 3, 2000, pp. 24-29

WASTE TREATMENT: THE FILAMENTOUS BACTERIA CAUSING BULKING AND FOAMING IN ACTIVATED SLUDGE PLANTS

Robert J Seviour, Jacques A. Soddell, Elizabeth M. Seviour, Jian Rong Liu and Christine A. McKenzie of Biotechnology Research Centre, La Trobe University, Bendigo, and Christopher P. Saint of the Water Quality Centre, SA Water, Bolivar, South Australia

Code Number: au00033

INTRODUCTION

Although the activated sludge system has operated globally to treat both domestic and industrial wastes for more than 80 years, it is still treated very much as a “black box” by the engineers. It is a biological process, and yet until very recently, little understanding of the microbiology of this huge biotechnological industry has been forthcoming. There are good reasons for this, most of which are methodological. Activated sludge is a highly complex ecosystem and, until quite recently, the methods available for identifying and characterising the bacteria there were inadequate, even in the unlikely event that they could be grown in pure culture. The advent of molecular methods to study natural communities of microbes including those found in activated sludge has revolutionised our ideas on their composition. Even so, many engineers (with some justification it has to be said) would claim that the microbiologists have promised much but in reality have contributed little to our understanding of how these plants work, or how their operational efficiencies might be improved. They could argue that all the microbiologist has achieved is to add to the confusion by showing that the microbial communities are far more diverse than were previously thought likely (Amann et al., 1996, 1998; Snaidr et al., 1997; Seviour & Blackall, 1999), but without suggesting how this information might be beneficial to the operators faced with the more mundane daily tasks of running these plants.

This article will attempt to show how the application of these molecular techniques might help us to deal with the serious operational problems most plants around the world suffer from ie. BULKING and FOAMING. In both, the settling characteristics of the biomass (hopefully organised as aggregates called flocs) in the sedimentation tanks after it leaves the aeration tanks are seriously impaired and several serious operational difficulties can result from this (Jenkins et al., 1993; Seviour & Blackall, 1999). Both problems are caused by a range of filamentous bacteria, which in most cases should probably be considered normal members of the activated sludge community, providing as they do the matrix for floc formation. Eikelboom (1975) in his pioneering work on these filaments recognised about 30 morphotypes (for examples see Fig 1 showing a selection of filaments mentioned in this article), which he “identified” exclusively on the basis of their microscopic appearances like cell and filament dimensions, and staining reactions (eg Gram and Neisser stains) as listed in Table 1, and classified them using these characters (Table 2). Most of these he could not grow and properly characterise to provide them with valid names, and so he referred to them instead as morphotypes ie Type numbers, (eg Type 1863, 021N, 0041 etc), a system of ‘nomenclature” which has persisted in the industry until now (Seviour & Blackall 1999).

THE PROBLEMS WITH THIS IDENTIFICATION SCHEME

This is not the place to discuss the science of bacterial taxonomy, except to say that there are strict rules which have to be followed when bacteria are classified, named and identified, and the scheme of Eikelboom (1975) follows few or none of these. Most attempts to ‘identify’ these filaments use the published manuals of either Eikelboom & van Biujsen (1983) or Jenkins et al., (1993), based on their morphological features (Table 1). Neither will necessarily identify reliably these filaments in activated sludge, since the characters they use are likely to change with different environmental conditions, in ways we do not understand. Furthermore, we have known for some time that because two bacteria happen to look alike under the microscope is no guarantee that they are necessarily closely related to each other (Woese, 1987). Also some filaments can switch their morphology between a unicellular and filamentous form (Seviour & Blackall 1999). Therefore it is unclear whether organisms from different plants or different countries which cannot be distinguished under the microscope are really the same bacteria.

The requirement of being able to unequivocally identify these bacteria in activated sludge plants is not only of academic interest. At the moment it is the only method available to operators who wish to control the bulking and foaming problems these bacteria cause. Several rules of thumb associate certain (but not all) filament types with specific operational parameters (Table 3) and provide a framework for eliminating the filaments present (ie those “identified” microscopically), by adjusting these parameters (Jenkins et al., 1993). This is the most popular control strategy employed by operators in the wastewater industry to cope with bulking and foaming. Clearly such a system will fail if the filaments are wrongly “identified”, and unfortunately it seems that this happens most of the time.

POSSIBLE SOLUTIONS

Many of these problems could be solved if these filamentous bacteria could be grown in pure culture, their taxonomy properly resolved and their physiology and ecology more fully understood. Obtaining these in pure culture is a very difficult task, but if achievable, then progress is possible.16S rRNA sequence data are the current benchmark for sorting out the phylogeny of unknown organisms. Once this sequence is known for a particular bacterium, it is a relatively simple task to design a specific 16s rRNA targeted gene probe based on a unique region of this 16S rRNA sequence (Amann et al., 1995). This probe can be tagged with a fluorescent reporter molecule and the bacterium of interest then identified unequivocally in biomass samples (Amann et al., 1995) using the technique of fluorescent in situ hybridisation (FISH). There is now no need to culture that organism to identify it, and unlike microscopic methods, probes do not rely on filament morphology for this identification.

SO WHAT DO WE KNOW NOW ABOUT THESE BULKING FILAMENTS?

Pure cultures are now available for several of these bulking Eikelboom filaments, and we also have probe sequences for some of these (Seviour & Blackall, 1999). Several important and often unexpected findings have emerged from such studies:

  • Most of the filaments after 16S rRNA sequence analyses are previously undescribed novel bacteria (Blackall et al., 1994, 1996, 2000). These include (Fig 1) “Microthrix parvicella”, “Nostocoida limicola” II and III, and Types 0803, 0411, 0092.
  • Some belong to previously described bacterial genera eg Type 021N (Fig 1) is a Thiothrix sp (Howarth et al., 1999).
  • Several of these genera were not considered to be filamentous in nature eg some Type 1863 (Fig 1) isolates are members of the genus Acinetobacter, a Gram-negative coccobacillus usually in pairs (Seviour et al., 1997)
  • A single morphotype can contain members of several different previously described bacterial groups eg the morphotype Type 1863 (Figure 1) contains bacteria belonging to the genera Acinetobacter and Moraxella as well as the Cytophaga-Flexibacter-Bacteroides subdivision (Seviour et al., 1997). “Nostocoida limicola” I isolates fall into three genera Trichococcus, Lactosphaera and Streptococcus (Liu et al., unpublished). This emphasises the risks associated with relying on cell and filament morphology to identify these bacteria in activated sludge.
  • Members of what were considered historically to be variations of a single morphotype emerge as quite unrelated bacteria eg “Nostocoida limicola” I, II and III are usually collectively described as “Nostocoida limicola”. All three exist as cocci in chains in activated sludge plant and are differentiated from each other almost exclusively on the microscopic dimensions of the individual cells in these chains, with “N. limicola” I being smaller than “N. limicola” II and “N. limicola” III being the largest (Figure 2). 16S rRNA analyses of pure cultures show they are in fact quite unrelated, being members of the low mol% G+C, high mol% G+C and the very unusual Planctomycete subdivisions of the Bacteria respectively (Blackall et al., 2000; Liu et al., unpublished).
  • Surveys using 16S rRNA targeted probes and FISH based methods for “M. parvicella” suggest that this morphotype, seen all over the world and one of the most troublesome bulking (and foaming) bacteria, is probably a single bacterium (Erhardt et al., 1997). Thus the probe designed from a Melbourne isolate grown here lights up “M. parvicella” in plants in several other countries. The situation with some of the other filaments after FISH confirms that a single morphotype may contain several different bacteria eg type 021N (unpublished data).
  • Not all these filaments have been grown in pure culture yet eg Types 0961, 0914, 0041/0675 and so they still await satisfactory identification.
WHAT ABOUT THE FOAMING BACTERIA THEN?

In some regards, the situation with the microbiology of foaming, also a global problem, is a little clearer, although foams are usually less readily controlled than bulking. The formation of a thick stable foam on the surface of both the aeration and sedimentation tanks can be attributed mainly to a few bacterial morphotypes (Soddell & Seviour 1990; Soddell 1999), although some of the bulking filaments mentioned above eg., Nostocoida limicola have been implicated occasionally. These foaming bacteria include;

  • Long unbranched filamentous Gram-positive bacteria which have been identified as Microthrix parvicella (Fig 1), an organism responsible for both bulking and foaming, although little is known about what triggers a bulking or foaming response caused by it.
  • Gram-positive branching filamentous bacteria with acute branching angles, resulting in filaments with branching patterns resembling pine trees (Fig 2a). These pine tree like organisms were originally classified as Nocardia pinensis, but are now included in a new genus of the mycolata, (mycolic acid-producing actinomycetes) as Skermania piniformis. To date, this distinctive morphotype is associated only with the genus Skermania (Soddell & Seviour, 1998a).
  • Gram-positive branching filamentous bacteria with branching angles that are approximately at right angles (Fig 2b). Foams caused by these morphotypes were (and still are) called “Nocardia” foams, because the original isolates of these organisms were identified as being Nocardia spp ie N. amarae (Lechevalier & Lechevalier, 1974).
  • However, later studies have shown that although foaming organisms with this morphotype generally belong to the mycolata, they represent many different genera (eg Gordonia, Rhodococcus, Tsukamurella, Nocardia, Dietzia) (Soddell, 1999).
  • So diversity is the key word when describing these ‘look-alike’ organisms, and it is certain that many more novel members of the mycolata will be isolated from activated sludge foams in the future (eg Goodfellow et al., 1998; Soddell et al., 1998b).
  • This morphotype should be referred to by a broader and more taxonomically correct name than “Nocardia “ (Seviour & Blackall, 1999), and more recently the terms NALO (Nocardia amarae-like organism), and GALO (Gordonia amarae-like organism, a more precise description) have been used. One of the major problems with recognizing this morphotype in foams is that many mycolata undergo a rod-coccus life cycle, and hence do not always appear as branching filaments but instead as Gram-positive cocci, or short rods (Fig 2c) (eg., Soddell & Seviour, 1990; Stratton et al., 1996). Identity of these as members of the mycolata has been either after isolation by micromanipulation and identification by conventional methods and 16S rRNA gene sequencing (Stratton et al., 1996), or through FISH studies with16S rRNA targeted probes (Davenport et al., 1998).

This non-filamentous morphology makes interpretation from microscopic examination of activated sludge foam and mixed liquor more confusing still, and the problem of the occurrence of these unicells has only been addressed in one survey study (Seviour et al., 1994).

A NEED FOR RELIABLE IDENTIFICATION OF THESE FOAMING FILAMENTS TOO

Although M. parvicella and S. piniformis can be identified microscopically with confidence on their morphology alone (at least as far as we know), reliable identification of the other mycolata needs techniques like FISH (de los Reyes, F. et al., 1997; 1998; de los Reyes M. et al., 1998; Oerther et al., 1999; Davenport et al.,1998; 2000) to enable differentiation between the different genera and their morphological variants.

These are rarely if ever used by plant operators, and so the desired level of identification is not achieved. As discussed for the bulking filaments, this information is important, since members of the mycolata differ physiologically in ways that may have considerable relevance to the control of foaming. In particular, we know their growth rates do vary (even among members of the same genus), which has implications in any control strategy based on sludge age manipulation. This is a popular control method which not too surprisingly in the absence of proper identification, only succeeds occasionally. In fact there are no reliable control measures for foaming, almost certainly due, at least in part, to our current lack of understanding of the true level of taxonomic diversity among the foam-forming bacteria, and hence methods for their unequivocal identification in foams.

IDENTIFICATION IS ONLY PART OF THE STORY

Although these FISH techniques are very elegant and powerful for identifying these filaments, they need to be applied carefully to avoid misleading results associated with false negatives from permeability problems and presence of metabolically inactive cells (Head et al., 1998), and from autofluorescence of the activated sludge samples. They also tell us little about what the bacteria are doing in these systems, and whether a single morphotype identified by FISH in many different plants is physiologically constant or can vary its biochemistry depending on the prevailing conditions. Such information will be needed if we are ever to scientifically and rationally control the growth of these bacteria. Recent approaches (Andreason & Nielsen 1997, 1998; Lee et al., 1999) have used microautoradiography (MAR), where activated sludge biomass samples are fed labelled substrates and cells able to assimilate these are then detected by silver grain deposition around them. These cells can then be identified in the same experiments using FISH probing. Studies with some of the important bulking filaments are beginning to provide some answers to these important questions.

A good example is “M. parvicella” (Andreasen & Nielsen, 2000) and its in situ behaviour in activated sludge plants removing the nutrients N and P. Such systems often operate with a long sludge age (mean cell residence time), and the biomass is cycled repeatedly through aerobic and anaerobic regimes (see Blackall et al this issue of Australasian Biotechnology). MAR provides some clues as to why “M. parvicella” does so well in these plants, and suggests possible control strategies for a very difficult organism to eliminate. The following information on its in situ physiology has emerged:

  • there was no assimilation of 33P by “M. parvicella”, consistent with it having no role in biological P removal;
  • unlike several other filaments seen in these systems, “M. parvicella” could assimilate hydrophobic substrates like lipids and oleic acid under aerobic, anoxic and anaerobic conditions, possibly explaining why its growth is not controlled with anaerobic or anoxic selectors;
  • it was able to remain metabolically active for longer periods of time under anaerobic and anoxic conditions than aerobic ones, supporting the view that “M. parvicella” is microaerophilic.

Andreasen & Nielsen (2000) proposed that the ability to assimilate long-chain fatty acids (LCFA) like oleic acid under anaerobic conditions gives it a considerable competitive edge over most other bacteria present. They also suggested that it might explain why “M. parvicella” often prefers a lower bulk liquid temperature where the solubility of lipid substrates would be reduced, providing it with a further selective advantage because of its high uptake capacity for LCFA. Consequently, the control strategy they put forward for this organism based on these data involves minimising the availability of LCFA under all but especially anaerobic conditions. Similar work is urgently required for other bulking and foaming filamentous bacteria and we are currently looking at “Nost. limicola” in the same way. Equally important is whether all strains of “M. parvicella” behave under all operating conditions in the same way. Otherwise, the value of such fascinating and important data will be limited to individual plants.

DO WE UNDERSTAND HOW FOAMS HAPPEN AND WHAT ROLE THESE BACTERIA PLAY IN THEIR FORMATION?

Apart from M. parvicella (described earlier), microautoradiography techniques have not yet been applied to foaming organisms (see above), and so their in situ behaviour is not well defined. However, other evidence would suggest that the most common factor contributing to their growth is the presence in the wastes of fats, oils and greases (Soddell 1999). Although much of this evidence is anecdotal, foam-forming mycolata, which tend to be very hydrophobic, probably because of the mycolic acids in their cell walls, appear to have a strong affinity (both metabolically and physico-chemically) for hydrophobic substrates (Soddell & Seviour 1996; Soddell et al 1998), and may grow even better on oils than on more simple hydrophilic substrates like glucose (Soddell & Seviour 1996). Hence, one approach to controlling foam formation might be to physically remove these hydrophobic chemicals as early as possible in the process. However, such an explanation fails to account for how sudden explosive foaming episodes, common to all plant operators might happen.

Foam is thought to be a flotation event (Soddell & Seviour, 1990). Hydrophobic particles (in this case the bacterial cells) attach to air bubbles (present from the aeration system in the aerobic tank), and detergents produced either by the bacteria or present in the waste, then stabilise the liquid/air films to produce a remarkably persistent foam. Attempts to control it with antifoams or by reducing the hydrophobicity of the cells, attractive in principle, have met with little success in full-scale plants. We are investigating alternative biological control strategies that have the added advantage of specificity of action, but again these depend on a more complete understanding of the microbiology of foam than is presently available.

CONCLUSIONS

We believe that this short article demonstrates the immense value of using molecular biology techniques in attempts to understand wastewater treatment processes, the largest biotechnology industry in the world, better, and how these methods can provide us with information unavailable from the more conventional ones. Furthermore, it is now possible to apply these to understand what favours the growth or otherwise of some of these problems causing filamentous bacteria. In our view these data will provide for the first time the basis for reliable control strategies for them. Of course, there is a long way to go, and funding for this kind of work in Australia is almost impossible to obtain, a situation incidentally in sharp contrast to that seen in Europe. There is still an attitude here that all problems encountered with activated sludge systems must have engineering reasons, and therefore the solutions have to be engineering ones. When the costs involved in containing bulking and foaming episodes are taken account of (and these are considerable), the money spent on trying to find sound scientific solutions is embarrassingly trivial.

ACKOWLEDGEMENTS

We wish to thank our colleagues who have participated in the studies carried out in Bendigo, especially Dr Linda Blackall, University of Queensland.

REFERENCES
  • Amann, R. et al. (1995). Microbiol. Rev. 59, 143-169.
  • Amann, R. et al., (1996). J. Bacteriol. 178, 3496-3500.
  • Andreasen, K. & Nielsen, P.H. (1997). Appl. Environ. Microbiol. 63, 3662-3668.
  • Andreasen, K. & Nielsen, P.H. (2000). Water Res. 34, 1559-1569.
  • Blackall, L.L. et al., (1994). System.Appl. Microbiol. 17, 513-518.
  • Blackall, L.L. et al., (1996). Water Sci. Technol. 34, 137-144.
  • Blackall, L.L. et al., (2000). Int. J. System. Evol. Microbiol. 50, 703-709.
  • Davenport R. J. et al.,(1998).. Anton van Leeuwen. 74, 41-48.
  • Davenport, R.J. et al., (2000). Appl. Environ. Microbiol. 66, 1159-1166.
  • de los Reyes,F.L. et al.,(1997). Appl Environ. Microbiol. 63, 1107-1117
  • de los Reyes F.L. et al.,(1998). Water Sci Technol. 37, (4-5) 485-493.
  • de los Reyes M.F, et al.,(1998). Appl Environ. Microbiol. 64, 2503-2512.
  • Eikelboom, D.H. (1975). Water Res. 9, 365-388.
  • Eikelboom, D.H. & van Buijsen, H.J.J. (1981). TNO Report No. 94a. TNO Research Institute for Environmental Hygiene, Delft.
  • Goodfellow, M. et al (1998). Water Sci. Technol. 37, 511-519.
  • Head, I. M. et al., (1998). Microbial Ecol. 35, 1-21.
  • Howarth, R. et al., (1999). Int. J. System. Bacteriol. 49, 1817-1827.
  • Jenkins, D. (1992). Water Sci. Technol. 25, 215-230.
  • Jenkins, D. et al., (1993). Manual on the Causes and Control of Activated Sludge, Bulking and Foaming, 2nd Edn., Lewis, Boca Raton.
  • Lee, N. et al., (1999). Appl. Environ. Microbiol. 65, 1289-1297.
  • Lechevalier, M.P. & Lechevalier, H.A. (1974). Int. J. System. Bacteriol. 24, 278-288.
  • Oerther D.B.(1999).. FEMS Microbiol. Ecol. 29, 129-136.
  • Seviour, E. M. et al., (1997). J. Appl. Microbiol. 82, 411-421.
  • Seviour, R.J. & Blackall, L.L. (1999). The Microbiology of Activated Sludge. Kluwer Academic, Dordrecht.
  • Snaidr,J., et al., (1997). Appl. Environ. Microbiol. 63, 2884-2896.
  • Soddell, J.A. ( 1999). Foaming IN: The Microbiology of Activated Sludge (Eds R.J. Seviour & L.L.Blackall), pp. 161-202. Kluwer Academic, Dordrecht.
  • Soddell, J.A. & Seviour, R.J. (1990).J. Appl. Bact. 69, 145-176.
  • Soddell, J.A. & Seviour, R.J. (1996). Water Sci. Technol. 34, 113-118.
  • Soddell, J.A. & Seviour, R.J. (1998a). J.Appl. Microbiol. 84, 272-284.
  • Soddell, J.A. & Seviour, R.J. (1998). Water Sci. Technol. 37, 495-502.
  • Stratton, H.M. et al.,(1996). Lett. Appl. Microbiol. 22, 341-346.
  • Stratton, H.M. et al.,(1993). Actinomycetes 4, 40-46.
  • Wanner, J. et al.,(1998). Water Sci. Technol. 37, 271-280.

Copyright 2000 - Australasian Biotechnology


The following images related to this document are available:

Photo images

[au00033f2.jpg] [au00033f1.jpg] [au00033t2.gif] [au00033t3.gif] [au00033t1.jpg]
Home Faq Resources Email Bioline
© Bioline International, 1989 - 2024, Site last up-dated on 01-Sep-2022.
Site created and maintained by the Reference Center on Environmental Information, CRIA, Brazil
System hosted by the Google Cloud Platform, GCP, Brazil