Determining Microbial Biomass and Community
Structure
Traditional isolation and culture techniques are inadequate for
microbial characterization in environmental studies because the techniques
are (1) selective and not quantitative (Vestal and White, 1989; White
et al., 1997), (2) provide little insight into microbial consortium
interactions (White et al., 1997), and (3) may introduce disturbance
artifacts because these techniques involve subsampling and separation
of microorganisms from the environmental matrix (Findlay et al., 1990).
Furthermore, most microorganisms in the environment are viable but
not cultivable (Xu et al., 1982; McCarthy and Murray, 1996). Viable
counts of bacteria in environmental samples determined with classical
methods represent only a small fraction (0.1% to 10%) of the active
microbial community (White et al., 1997).
Catabolic gene and 16S rRNA probes have been used successfully to
assess the presence of specific microorganisms in environmental samples
(e.g., Sayler et al., 1985). Such methods, however, are labor intensive
and experience limitations when measuring community functionality
under stress or competition (Findlay, 1996). Moreover, since the sequence
of the universal primers is based on cultured organisms, the applicability
of this technique for community analysis in environmental samples
remains questionable (Pace, 1996). Such limitations have motivated
the development of chemical characterization techniques to determine
microbial biomass and community structure without prior isolation
and cultivation of microorganisms.
Current approaches used for chemical characterization of microbial
populations in natural environments include two techniques that analyze
the cell membrane phospholipids. These are (1) phospholipid ester-linked
fatty acid (PLFA) analysis by gas chromatography/mass spectrometry
(White et al., 1979), and (2) intact phospholipid profiling (IPP)
using liquid chromatography/electrospray ionization/mass spectrometry
(LC/ESI/MS) analysis of bacterial membrane phospholipids (Fang and
Barcelona, 1998). Both techniques rely on the fact that phospholipids
are found in the membranes of all living cells, but not in storage
lipids, and are rapidly turned over in dead cells. Thus, their quantification
provides an estimation of viable biomass (Balkwill et al., 1988).
Identification of microorganisms by either PLFA or IPP is subject
to potential confounding effects of overlapping phospholipid profiles
and potential changes in phospholipid composition due to differences
in growth conditions (Haack et al., 1993; White et al., 1997). Nevertheless,
both techniques can give valuable insight into microbial community
structure, based on the premise that fatty acids and phospholipids
are major membrane components (Fig. 1) and that there are a great
number of dissimilar fatty acids in bacterial phospholipids and some
bacteria contain unique fatty acids.
Recently, we compared PLFA and IPP techniques in identifying five
Pseudomonas strains (Fang et al., 2000a) and seven different species
of Type I and Type X methanotrophic bacteria (Fang et al., 2000b).
The type I methanotrophs, Methylomonas methanica, Methylomonas rubra,
and Methylomicrobium album BG8 were characterized by PE and PG phospholipids
with predominantly C16:1 fatty acids. The type II methanotrophs, Methylosinus
trichosporium OB3b and CSC1 were characterized by phospholipids of
PG, PME, and PDME with predominantly C18:1 fatty acids. Methylococcus
capsulatus Bath, a representative of type X methanotrophs contained
mostly PE (89% of the total phospholipids). Finally, the intact phospholipid
profiles of a recently isolated acidophilic methanotroph, Methylocella
palustris, showed it had a preponderance of PME phospholipids with
18:1 fatty acids (94% of total). Principal component analysis showed
these methanotrophs could be clearly distinguished based on phospholipid
profiles (Fig. 2). Results from this study suggest that IPP can be
very useful in bacterial chemotaxonomy.
The five reference pseudomonad strains: Pseudomonas putida mt-2, Pseudomonas
putida F1, Burkholderia cepacia G4, Burkholderia pickettii PKO1, and
Pseudomonas mendocina KR1 contained eight major fatty acids in these
pseudomonads, ranging in chain length from C14 to C19. IPP (intact
phospholipid profiling) detected 16 phospholipids in three different
classes: phosphatidylglycerol, phosphatidylethanolamine and phosphatidyl-dimethylethanolamine.
Factor analysis was applied to the PLFA and IPP data to compare the
efficacy of these methods for microbial identification (Fig. 2). The
number of variables with high loadings (>|1.0|) was greater in
IPP data than in PLFA data (2 vs. 1). IPP also provided better separation
of the five pseudomonad strains, whereas using PLFA only separated
F1 from the other four strains, which were closely clustered (Fig.
2). The superior differentiation power of the IPP can be attributed
to the fact that the limited number of fatty acids (eight in this
study), when combined with three different classes of head groups
yields sufficient different phospholipid compounds for microbial identification.
LC/ESI/MS analysis of intact phospholipids simultaneously determines
the class (polar head group) and the structure (individual fatty acids)
of intact phospholipids (Fang and Barcelona, 1998). On the other hand,
the total PLFA analysis detected a mixture of fatty acids (two fatty
acids from each phospholipid), and the information carried in the
original membrane lipid molecules had been lost, suggesting that IPP
is superior to the PLFA technique in microbial differentiation and
identification.
