The schemes which have been published for their identification are all artificial

or 'special purpose', and consequently all suffer the disadvantages of such

classifications.
Eikelboom (1975) made one of the first, and very influential attempts, by dividing
the 29 different morphological types he recognized into several groups based on
their microscopic properties and staining reactions, and this scheme is still widely
used. These groupings are listed below.
1. Sheath-forming, Gram-negative filaments: Sphaerotilus natans,
Haliscomenobacter
hydrossis, Types 1701, 1702, 0321.
2. Sheath-forming, Gram-positive filaments: Types 0041, 0675, 1851.

3. Sheathless, curled, multicelled filaments: Thiothrix spp., 'Nostocoida limicola',

Cyanobacteria, Type 021N.
4. Slender, coiled filaments: 'Microthrix parviceZZa', Types 0581, 0192.
5. Straight, short, multicelled filaments: Types 0803, 1091, 0092, 0961.
6. Gliding, motile filaments: Beggiatoa spp., Types 0914, 1111, 1501.
7. Others: Nocardia spp., Fungi, Types 1863, 0411.

This classification is, of course, incomplete, and tells us nothing about how these
bacteria are really related to each other. It does not incorporate our more recent
understanding of the taxonomy of some of the Gram-negative filaments (Bradford
et aI., 1996; Seviour et al., 1997) (Table 5.3) or, for example, G. amarae, S.
piniformis and
Rhodococcus spp. (see Chapter 7). The groupings of Eikelboom are based on
heavily
weighted taxonomic characters with no clear justification for their choice. Some of
the selected primary characters may confuse, since many Cyanobacteria for
example
are also motile by gliding and so could be placed equally validly into Group 6
(Waterbury, 1992). Certain of these properties like filament morphology may also
change with alterations in the organisms' environments as with 'M. parvicella' and
Type 1863, and so can not be relied upon (e.g. Foot et al., 1992; Seviour et al.,
1997).
More practical 'classifications' for these bacteria have also been suggested by
Jenkins (1992) and Wanner (1993), who both group filaments based on their
appearance
in plants with particular operating parameters. These schemes were discussed
earlier (Tables 4.9 and 4.10). Again, they are open to criticism. For example, it is
considered unlikely that each of these morphologically similar filament types is a
collection of organisms with identical or very similar physiology and ecology. Both
these schemes mentioned above were designed to assist plant operational staff to
control the problems of bulking and foaming, and not to increase our understanding
of the biology of these filamentous bacteria. They extrapolate from a very small
knowledge base, and will certainly need to be modified as our understanding of the
taxonomy of these filamentous bacteria increases.

5.5 CURRENT STATUS OF FILAMENT IDENTIFICATION PROCEDURES

Most attempts to identify these filaments currently rely on the published
descriptions
given in the two reference identification manuals of Eikelboom and van Buijsen
(1983) and Jenkins et al. (1993b). This identification is based almost exclusively on
morphological characteristics and the response of the filaments to a few staining
techniques, and although these manuals are of great value to, and widely used by,
the industry, they are intellectually unsatisfying taxonomically. The diagnostic
features
currently most widely used to identify these filaments were given earlier in
Table 5.2 and the features each possesses is shown in Table 5.4. It has already been
mentioned that morphological criteria are unreliable as indicators of relatedness for
prokaryotes, a problem exacerbated by the fact that many filaments undergo
reversible
morphological changes to unicellular rods or cocci in response to changes in
culture or plant conditions (e.g. Foot et al., 1992; Wagner et al., 1994a,b; 5eviour
et al.,
1997). Thus, organisms which look the same, even to a trained, experienced eye,
may
not necessarily be so. It is quite possible that filaments from different plants,
particularly
in different countries, may readily fit descriptions given in these manuals, and
yet strains within each morphological type vary considerably in their physiology and
taxonomy. However, probes designed with the 165 rONA sequence of an Australian

5.6 FUTURE PROSPECTS FOR FiLAMENT TAXONOMY

How do we remedy these serious weaknesses in systematics of these filamentous
bacteria in activated sludge? Clearly, more basic microbiology needs to be done,
especially using more pure cultures (Seviour et al., 1994, 1997; Kampfer et al.,
1995;
Kiimpfer, 1997), and methods developed to grow the filaments to high biomass
concentrations so that they can be comprehensively characterized with the
techniques
mentioned earlier. It is also important that this characterization is carried out
on large numbers of isolates of each of these morphological types from a wide
range
of locations and countries, since only then can we assess the degree of variation or
biodiversity existing within each type. In addition, now is the time to take advantage
of the developments in techniques of molecular systematics by analysing the rRNA
gene sequences of these filaments (and other activated sludge organisms) to help
us
resolve their true taxonomic position and provide them with proper names (Blackall
et al., 1996a,b; Bradford et al., 1996). Such data are beginning to appear in the
literature and references to these studies have appeared regularly throughout this
book. This is not the place to provide a detailed background to the principles of
bacterial phylogeny, and the reader is directed towards the excellent reviews of
Woese (1987, 1992) for this. However, some basic information is necessary to
explain
how and why knowing the sequences of particular genes can lead to better systems
for classifying, naming and identifying these filamentous bacteria.

As briefly discussed in Chapter I, all cells need to synthesize proteins to grow, and
the methods they use for this are very similar, regardless of whether the cell is
prokaryotic or eukaryotic. In protein synthesis, the genetic information, stored as
sequences of nucleotides or bases in their DNA, is firstly copied or transcribed into
a sequence of bases in a strand of what is called messenger RNA (mRNA), and this
is then translated into an amino acid sequence of a protein. Details of this process
can
be found in any good cell biology text. The structures inside cells where this process
of translation and protein synthesis occurs are the ribosomes (section 1.5.2), which
are made up of proteins and other RNA molecules called ribosomal RNA (rRNA).
Each ribosome is comprised of several different-sized rRNA molecules which can be
defined by their sedimentation properties in a centrifuge, in terms of their so-called
Svedberg (S) units. In prokaryotic ribosomes, the 5S rRNA, 16S rRNA and 23S rRNA
molecules are found (Alberts et al., 1997). Particular genes in the cells' DNA carry
the
genetic information which codes for the synthesis of these different rRNA molecules.
Because the translation of mRNA into proteins or ribosomes is so important for all
cells, it is probable that the system for translation arose only once in evolution, and
has not changed subsequently. In other words, as it is a conserved process in an
evolutionary sense, then sequences of bases in the 5S, 16S and 23S rRNA would
also
be conserved, as would the genes (the corresponding complementary DNA
sequence)
which encode for them. Woese (1987) convincingly summarizes the argument
that the 16S rONA sequence, which is the gene usually sequenced, can be used
as a 'molecular clock' to indicate how prokaryotes may have evolved. The 16S rRNA
gene (16S rONA) of about 1500 bases of deoxyribonucleotides in length has now
been
sequenced from thousands of bacteria (Maidak et al., 1996), and it is clear that the
molecular diversity of this gene sequence is high among these organisms. It has
also
revolutionized our view of bacterial classification (Woese, 1992; Stackebrandt and
Rainey, 1995). However, some of the conserved regions of it are very similar
between
different bacteria, and within these, short sequences may be recognized that are
unique to a particular bacterium (Wheeler AIm et al., 1996), as indicated in Table
5.5.
These sequences in fact provide a molecular 'fingerprint' for that particular
organism,
and it is these unique oligonucleotide sequences which can be used to identify
individual bacterial cells (Amann et al., 1990; Schleifer et al., 1993; Amann, 1995a-
c;
Stackebrandt and Rainey, 1995). It is also possible to compare the 16S rONA
sequences
between different bacteria, and so an unknown isolate (for example a filament),
once its 16S rONA sequence is known, can then be compared with all other
bacterial16S rONA sequences, and its taxonomic position thus determined. Then a
proper valid biological name can be provided for it (Goodfellow and Stackebrandt,
1991; Triiper and Schleifer 1992; Stackebrandt and Liesack, 1993). The technology
for determining 16S and 23S rONA sequences has improved at a staggering rate
over
the past decade and is now becoming a routine procedure for properly equipped
laboratories (Lane, 1991; Stackebrandt and Liesack, 1993; Stackebrandt and
Rainey,
1995).

Once the sequence of the unique regions of the conserved gene for a particular
bacterium has been determined, a specific probe for its identification can then be
constructed in the laboratory using a special commercially available piece of
apparatus,
an oligonucleotide synthesizer (Schleifer et al., 1993; Amann et al., 1995; Embley
and Stackebrandt, 1996). Probes of about 15-30 nucleotides in length should only
hybridize with complementary strands of target DNA or RNA in samples or pure
cultures of that particular bacterium. For example, if the signature in the 16S
rONA of a particular organism is 'TACGGTAC', the complementary sequence is