Stereological assessment of extracellular polymeric substances, exo-enzymes, andspecific bacterial strains in bioaggregates using fluorescence experiments
Sunil S. Adav a, Justin Chun-Te Lin a, Zhen Yang b, Chris G. Whiteley c, Duu-Jong Lee a,⁎,Xiao-Feng Peng b, Zhen-Peng Zhang d
a Department of Chemical Engineering, National Taiwan University, Taipei 10617, Taiwanb Department of Thermal Engineering, Tsinghua University, Beijing, 100084, Chinac Department of Biochemistry, Microbiology, Biotechnology, Rhodes University, Grahamstown, South Africad Beijing Enterprises Water Group Limited, BLK 25, No. 3 Minzhuang Road, Beijing 100195, China
Abbreviations: BODIPY, bovine serum albumin conjugfluid dynamics; CLSM, confocal laser scanning microscopdeoxyribonucleic acid; EPS, extracellular polymeric suautoradiography; NTSC, National Television System ComSTAR, substrate tracking autoradiography; TAMRA, tetra⁎ Corresponding author. Tel.: +886 2 23625632; fax:
This review addresses the introduction of fluorescent molecular tags into exo-enzymes and extra polymericsubstances of bioaggregates and the use of confocal laser scanningmicroscopy (CLSM) tomap their role, purposeand quantitative description of the biological processes they undertake. Multiple color staining coupled withCLSM and fluorescent in situ hybridisation (FISH) and flow cytometry have identified the individual polymericsubstances, whether they are proteins, lipids, polysaccharides, nucleic acids or antibodies, as well as themicroorganisms in the bioaggregate. Procedures are presented for simultaneous multicolor staining with sevendifferent fluorochromes — SYTOX Blue for nucleic acids; Nile red for lipids; Calcofluor white [CW] for β-polysaccharides; concanavalin A [Con A] for α-poly-saccharides; fluorescein-isothiocyanate [FITC] for proteins;SYTO 63 for live microbial cells and Calcium Green for monitoring calcium levels in the microbial cells. For thedistribution of certain microbial strains, metabolic enzymes and extrapolymeric substances to be quantitativelydescribed the generated colored images are converted into digital forms under specific predefined criteria.Procedures and computer software programs (Amira;MATLAB) are presented in order to quantitatively establishgrid patterns from the CLSM images. The image is digitized using a threshholding algorithm followed by areconstruction of the image as a volumetric grid for finite element simulation. The original color image is firstconverted to a grey followedby resizing, detectionandmodificationof bilevel imagesandfinally a total reversal ofthe image colors. The grid file is then used by specific computer software (Gambit, Fluent) for further numericalstudies incorporating chemical reactions, transport processes and computational fluid dynamics including intra-bioaggregate fluid flow, and heat and mass transfer within the bioaggregate matrix.
ates of 4,4-difluoro-5,7-dimethyl-4-bora-3a,4a-diazas-ine; COD, chemical oxygen demand; Con A, concanavalin A;bstance; FISH, fluorescence in situ hybridation; FITC, flmittee; PHOs, protein-hydrolysing organisms; SCOD, solumethyl-6-carboxyrhodamine; TEM, transmission electro+886 2 23623040.
Bioaggregates such as activated sludge flocs, aerobic and anaerobicbiofilms and marine snow are made up of numerous microorganisms,immobilized in EPS and/or matrices constituting polymers of proteins,polysaccharides, humic acids, and lipids (Nielsen et al., 1992). Over 99%of the bacteria are present in biofilms (Dalton and March, 1998) orgranules that protects the incorporated bacteria from antibiotics(Goldberg, 2002), disinfectants (Peng et al., 2002), or threat by envi-ronmental shock (Chen et al., 1998). The intercellular communicationwithin biofilms and granules regulates gene expression that enablestemporal adaptation to phenotypic variation (Kjelleberg and Molin,2002) and enhances survival rate in a nutrient deficient condition (Kochet al., 2001). The enzymes secreted by constituent cells acceleratebiological reactions for functions of maintenance and reproduction. Toquantitatively describe the biological processes that are involvedwithinthe bioaggregates acquires spatial distribution of functional bacterialstrains, anticipated enzymes, and substrates and metabolic productsinside the bioaggregates. For instance, Kloeke and Geesey (1999)determined the locations of phosphatase as well as phosphatase-excreting microorganisms in activated sludge using precipitatedfluorescent crystals from the reaction between the enzyme and theELF® 97 palmitate, using the FISH technique. With this combination ofanalytical tools, these authors used phosphatase activity to quantita-tively describe the biological processes that occurred in the system.
Images of very high resolution could be obtained using SEMor TEM technique (Gerhardt et al., 1994; Erlandsen et al., 2004).However, these technologies were not able to differentiate the func-tional constituents of the bioaggregates such as protein, lipids, specificbacterial strains and enzymes. The basic concept of confocal micros-copy was originally developed by Marvin Minsky in the mid-1950s(patented in 1961) (Amos and White, 2003). The use of CLSM hasgained popularity since it can (1) control depth of field of scanning;(2) eliminate background information away from the focal plane;(3) collect serial optical sections from thick specimens; (4) eliminateout-of-focus light; (5) provide quality scanned images; and (6) inparticular, provide fluorescent scanning of substances bound withdifferent fluorochromes that would be excited using light of differ-ent wavelengths. The multiple color staining technique and CLSM
together visualize the distribution of components of EPS in biologicalaggregates (Bockelmann et al., 2002; Strathmann et al., 2002;Lawrence et al., 2004; Boessmann et al., 2004; Staudt et al., 2004;Neu et al., 2004; Lawrence et al., 2005; McSwain et al., 2005; Wanget al., 2005, Chen et al., 2007a,b, Adav et al., 2007a,c,d; Adav and Lee,2008). FISH coupled with CLSM were able to locate specific bacterialstrain(s) in bioaggregates that were under investigation (Lemaireet al., 2008). Furthermore the immunohistochemical staining tech-nique and CLSM were able to locate the specific enzymes in a sludgefloc or a biological granule (Whiteley and Lee, 2006; Lee et al., 2009a).
To have a comprehensive three-dimensional (3D) view of the“functional map” of the bioaggregate under investigation was of greatacademic and practical interest. To achieve this goal, however, anintegrated protocol with the above-mentioned techniques should beavailable to simultaneously allocate EPS, functional cells and enzymes onthe same target. FITC labeled lectins have been used to stain theglycoconjugate fraction of biofilm (Lawrence et al., 1998; Michael andSmith, 1995; Neu 2000; Cerca et al., 2005) or of flocs and granules (DeBeer et al., 1996;Wang et al., 2005). Schmid et al. (2003) applied FITC tostain the protein fraction in their sludge floc. Two fluorochromes wereused to probe the EPS in single biofilms specimen (Strathmann et al.,2002; Boessmann et al., 2003, 2004; Staudt et al., 2004). Furthermore,three fluorochromes were applied to locate the EPS components inbiofilms (NeuandLawrence, 1997;Neuet al., 2001, 2004; Lawrenceet al.,2003, 2004, 2005); aerobic granules (McSwain et al., 2005). Bockelmannet al. (2002) combined the FISH technique with a lectin fluorochromein their double labeled experiments on river snow to provide fantastic,3D images for visual observations. To quantitatively describe the spatialcorrelations among those probed substances, such as the distribution ofdistance of certain strain andof specific enzyme, the colored images haveto be converted to digital formswith criteria set by the acquired analysis.The 3Dgridmodel for the studied aggregate could thereby be establishedwith locations of functional components realized for analyzing its steady-state behavior and transient response under changes in environmentalconditions (Chu and Lee, 2004; Chu et al., 2005).
Chen et al. (2007b) proposed a six-fold labeling scheme for stainingtotal cells, dead cells, proteins, lipids and α- and β-polysaccharides inbioaggregates. For instance, a merged image for a phenol-fed aerobicgranule, scanned at 680 μmfrom the granule surface, is shown in Fig. 1.
Fig. 1. Merged image of phenol-fed granule. Bar=200 μm. Green: protein (FITC); lightblue: α-D-glucopyranose polysaccharides (Con A); red: nucleic acids (SYTO 63); blue:β-D-glucopyranose (calcofluor white); yellow: lipid (Nile red). (For interpretation ofthe references to colour in this figure legend, the reader is referred to the web version ofthis article.)
257S.S. Adav et al. / Biotechnology Advances 28 (2010) 255–280
In adopting the scheme developed by Chen et al. (2007b), Chenet al. (2007a) compared the EPS distributions in acetate-fed andphenol-fed granules. Adav et al. (2007a) identified the detailedstructural changes that occurred after aeration shearing of aerobicgranules. Chen et al. (2006) examined the distributions of proteins,polysaccharides and DNA in the fouling layer on membranes. Bycombining multicolor fluorescent staining and the microelectrodemeasurement technique, Chiu et al. (2007a,b) demonstrated howdissolved oxygen was consumed in active aerobic granules. Extracel-lular enzymes have a crucial role in biological processes (Teuber andBrodisch, 1977). Measurement of extracellular enzymes was analternative method for assessing the microbial biomass and activity(Nybroe et al., 1992). The matrix of EPS in sludge flocs coulddeteriorate using exo-enzyme such as amylase, glucosidase andprotease (Ayol et al., 2008). The addition of the glycosidic enzymesto an anaerobic digester improved dewatering properties of sludge(Recktenwald et al., 2008). Activities of catechol 1,2-dioxygenase andcatechol 2,3-dioxygenase in phenol degrading granules (Jiang et al.,2004a, b) and proteolytic enzyme activities in stored granules (Adavet al., 2009) were determined. Immunohistochemical staining is theprocess of localizing proteins in cells of a tissue section exploiting theprinciple of antibodies binding specifically to antigens in biologicaltissues, and has been applied together withmultiple fluorochromes tolocalize α-amylase enzyme and EPS in aerobic granules (Lee et al.,2009a). In this study, the polyclonal antibodies raised in rabbit for theimmunogen α-amylase were employed for the detection anddistribution profile of α-amylase. These authors also noted that thesecreted amylase was located close to the living cells in the aerobicgranule specimen, demonstrating the polysaccharide hydrolysisactivity near the functional strains in the granule core. Xia et al.(2007) detected surface-associated exo-proteases activity using caseinor BODIPY andmicrobial diversity by FISH in activated sludge. Xia et al.(2008) applied fluorescence in situ enzyme staining with BODIPYfluorescein-labeled starch for labeling the starch hydrolyzing organ-ism expressing α-amylase in activated sludge.
This currentpaper reviews theuseof a staining/hybridisationprotocolusing multiple color staining, for proteins, lipids, α-, β-polysaccharidesand total DNA, FISH and the immunohistochemical staining forsimultaneous location of components of interests in a bioaggregatesuch as sludge floc or aerobic granule. Also, this review provides
procedures and detailed computer programs to quantitatively establishthe 3D grid models from the CLSM images. The grid models could beprocessed further to extract stereological information for analysis.
2. Probing spatial distributions of fluorescent componentsin bioaggregates
The spatial distribution of proteins, polysaccharides and livingcells have been profiled in biofilm, sludge floc and sludge granules.This section lists some commonly used fluorochromes for thesestudies.
2.1. Fluorochromes for CLSM imaging
Successful fluorescence labeling and detection relies on thespecificity and sensitivity rather than surrounding light absorbance,transmittance and reflection of the applied fluorochromes. The fluo-rochromes carrying reactive groups for covalent attachment to targetbiomolecules that had minimal spectral sensitivity to environmentalchanges are widely used. For example, isothiocyanate, succinimidyl andpentafluorophenyl esters incorporated in these dyes allowed for theircoupling with amino group on the target macromolecules. Fluorescentnanocrystals and quantum dots exhibited interesting properties forbiological labeling due to their broad absorption spectra with molarabsorptivities of more than six million at 450 nm. Their relatively largesize and high mass, however, prohibited their use in applicationsrequiring high diffusional mobility. Lanthanide chelates were anothergroup of fluorescent reagents that had special spectral properties withfluorescence lifetimes in the region of microseconds and could bedistinguished from typical nanosecond auto-fluorescence background(Seveus et al., 1994; Vereb et al., 1998). Temporal separation of probefluorescence from background signal could give a very high signal tonoise ratio and highly specific detection even though the brightness ofthe labeling reagent was only modest. In membrane potentialdetermination, freely diffusing anionic oxonol dyes have been pairedwith dyes anchored at the cell surface permitting fast ratiometricdetection of membrane potential changes. (Gonzalez and Tsien, 1995,1997; Gonzalez and Maher, 2002). The numbers of pH indicators werereviewed by Tsien (1989) and more recently by Yip and Kurtz (2002).
2.1.1. Fluorochromes for EPS stainingThe EPS as well as the embedded proteins and DNA, often provided
and maintained the structural integrity for the three-dimensionalbiofilm matrix (Blenkinsopp and Costerton, 1991; Nivens et al., 2001;Whitchurch et al., 2002; Jackson et al., 2004; Friedman and Kolter,2004; Sarkisova et al., 2005; Allesen-Holm et al., 2006; Adav et al.,2007d, f, 2008c). The distribution of extracellular polysaccharides wasstudied using fluorochromes that bind with specific oligosaccharidesubunits. The wide ranges of the fluorochromes (more than 100, website reference: http://probes.invitrogen.com) are available for eachcomponents of the EPS and their selection depends upon the researchneed, sample pH, and excitation/emission properties. Fluorescein andboratedipyrromethene (BODIPY) dye series; amine reactive AlexaFluor series were used to stain proteins (Nielsen and Jahn, 1999; Neuet al., 2001; Tay et al., 2002; Boessmann et al., 2003).
Fluorescently labeled lectins such as Con A stain glycoconjugateswere used to stain the glycoprotein conjugate fraction of biofilm(Lawrence et al., 1998; Michael and Smith 1995; Neu, 2000; Cercaet al., 2005), flocs and granules (de Beer et al., 1996; Wang et al.,2005). Schmid et al. (2003) applied FITC to stain the protein fractionin their sludge floc. The cyanine dyes (Mujumdar et al., 1993), BODIPYcomplexes (Wories et al., 1985; Kang et al., 1988) and the Alexa Fluordyes (Panchuk-Voloshina et al., 1999) complement the use of tra-ditional fluorescein and rhodamine dyes on staining experiments.The Nile red localized lipids such as phospholipids, triglycerides andneutral lipid droplets within cells (Lawrence et al., 2003, Chen et al.,
2007a,b, Adav et al., 2007a,b,c; Adav and Lee, 2008). The cell-impermeable and cell-permeable cyanine dyes (SYTO series) differ-entiated the live and dead cells in bioaggregates (Strathmann et al.,2002; Bockelmann et al., 2002; Lawrence et al., 2004; Boessmannet al., 2004; Staudt et al., 2004; Neu et al., 2004; Lawrence et al., 2005;McSwain et al., 2005; Wang et al., 2005, Chen et al., 2007a,b, Adavet al., 2007a,b,c; Adav and Lee, 2008).
2.1.2. Fluorochrome for FISHThe FISH technique applies fluorescently labeled nucleic acids that
are specific for RNA molecules for characterizing bacterial distribu-tions and, in some cases, expression activity in situ in biomass(DeLong et al., 1989; Amann, 1995; Moller et al., 1996; Xu et al., 2000;Jang et al., 2005). The FISH to ribosomal rRNA has proved effective forspecific fluorescent labeling of microorganisms (Amann et al., 1990),andwas also used in flow cytometry (Wallner et al., 1993, 1995; Veseyet al., 1999). One drawback for the FISH technique was the lowersensitivity than PCR-based techniques since the former needed highercell concentrations for detection (Hogardt et al., 2000; Moreno et al.,2003; Poppert et al., 2005).
By using different colored fluorescently labeled dyes, distributionof different species of bacteria on single specimen samples can beprobed. The genetic probe used in FISH could be fluorescently labeledwith Cy5, Cy3, Biotin, TAMRA and FITC. The dense clusters of nitrifyingbacteria, Nitrosospira and Nitrospira spp. were detected in aggregatesobtained from nitrifying reactors (Schramm et al., 1996, 1998) whileOkabe et al. (1999) profiled their distribution in biofilms. Sohierand Lonvaud-Funel (1998) adopted the FISH technique to detectand identify lactic acid bacteria in wine. Kolloffel et al. (1999) appliedFISH to probe brevibacteria on the surface of gruyere cheese. Theenumeration of probiotic bifidobacteria in the fermented food(Kaufmann et al., 1997; Lahtinen et al., 2005) and Carnobacteria andLactobacillus brevis from sea food (Connil et al., 1998) were conductedusing FISH. The FISH-CLSM technique elucidated the abundance andspatial distribution of phenol degrading strain PG-01 in the aerobicgranules (Jiang et al., 2004a, b), the sulfate-reducing bacteria insulfidogenic bioreactors (Dar et al., 2007), and the microbial diversityand community structure of aerobic and anaerobic methane oxidizersat the Haakon Mosby Mud Volcano, Barents Sea (Losekann et al.,2007).
Recent developments in FISH technology with new probe designshave increased cell detection efficiency and also overcome some of thedifficulties of probe penetration through the EPS matrix. Lehtola et al.(2006) developed the FISH technique that could incorporatepeptide nucleic acid (PNA) probes and hence the technique wascalled PNA-FISH and was used to detect Mycobacterium spp. inpotable-water biofilms. Another technique, the CARD-FISH, wasrefined to provide high permeability without serious cell loss(Pernthaler et al., 2002; Fazi et al., 2005; Ferrari et al., 2006). A lowcopy number of mRNA from an intact cell was amplified by in situreverse-transcription polymerase chain reaction and the amplifiedproduct was further detected using a fluorochrome (Magnuson et al.,2004). Furthermore, the MAR-FISH technique was developed toprovide information on in situ phylogenetic identification andmetabolic capabilities within complex microbial community (Leeet al., 1999a,b). The MAR-FISH technique has been recognized underdifferent names like STAR-FISH (Ouverney and Fuhrman, 1999), andMAR-FISH (Cottrell and Kirchman, 2000).
2.1.3. Fluorochrome for antibodiesThe immunohistochemical staining can localize proteins in
biological samples, exploiting the principle of antibodies bindingspecifically to antigens. The fluorescence labeled primary or second-ary antibodies are employed to detect the proteins and/or enzymedistribution profile in the biological samples.
Antibodies labeledwithfluorochromes are useful both in direct andindirect detection methods and also have applications in flowcytometry. In immunofluorescence staining techniques, in order toprevent the occurrence of false positives, the secondary antibodiesshould specifically recognize only one of theprimary antibodieswhich,in turn, should be specific in their recognition of the substance orenzymes against which they were raised. The fluorochrome attachedto a primary or secondary antibody absorbs energy when exposed to aparticular wavelength of light causing excitation of electrons to ahigher energy level and subsequently emit excess energy as light. Thefluorescent antibody conjugates exhibit high specific fluorescencewith minimal nonspecific staining. Four fluorochromes, rhodamine,fluorescein, Texas red and phycoerythrin are used for immunofluo-rescence assays as they differ in spectral properties and selectiondepends upon particular applications (Mao, 2008). The monoclonalantibodies to two different targetable antigens were conjugated toeach of four commercially available cyanine fluorochromes Cy3, Cy5,Cy5.5, and Cy7 and employed to locate tumors (Ballou et al., 1998).The performance of antibody–fluorochrome conjugate in an immu-nofluorescence technique largely depends upon the purity andconcentration of the antibody. Antibodies labeled with Cy5 and Cy3have a brightness comparable to or brighter than fluorescein-labeledantibodies and have little tendency to precipitate from solution(Wessendorf and Brelje, 1992). The quantum dots are flexible inspectral properties and could also be used in a simultaneousimmunofluorescence detection of multiple targets (Wu et al., 2003).
Activities of different exoenzymes such as proteases, galactosi-dases, glucosidases, lipases, chitinases and phosphatases have beenreported in activated sludge (Teuber and Brodisch, 1977; Boczar et al.,1992; Nybroe et al., 1992). The protein conjugated probes allow in situlocalization of enzymatic activities in tissues (Mook et al., 2003) andorganisms (Farber et al., 2001). Numerous surface associated exo-enzymes were probed in microbial aggregates (Confer and Logan,1998; Goel et al., 1998; Kloeke and Geesey, 1999; Nielsen et al., 2002).Once quenched BODIPY dye-labeled casein and BSA were hydrolyzed,fluorescent precipitate enable labeling of PHOs and explore theirdistribution in complex aggregates including biofilm and granularsludge by CLSM. However, some hydrolysates were released into thebulk medium (Yoshioka et al., 2003), so bacteria utilizing theseoligopeptides were also labeled.
For evaluating their post-translational modifications, Murray et al.(2004) raised monoclonal antibodies that were capable of immuno-capturing five enzyme complexes [Complex I (NADH dehydrogenase),complex II (succinate dehydrogenase), complex III (cytochrome creductase), complex IV (cytochrome c oxidase), and complex V (F1F0ATP synthase)] from bovine heart mitochondria that were involved inoxidative phosphorylation.
2.2. Selection and applications of fluorochromes
While selecting the fluorochromes for the biological samples,researchers should consider the following criteria:
(1) The fluorochrome must be delivered to the site of interest andremain there for long enough to acquire the experimental data.
(2) The fluorochrome and fluorochrome-probe conjugate shouldnot induce significant perturbations of the specimen samplestructure under study.
(3) Molecular association, transport, and metabolic fate of thefluorochromes and probes are important during selection andits application.
(4) The fluorochrome-fluorochrome interactions and environmen-tal conditions of temperature and sample pH are significantlyimportant so that the obtained data will not be biased.
(5) Spectral properties and cross-reference or overlap of spectralproperties need to consider before selecting fluorochromes.
259S.S. Adav et al. / Biotechnology Advances 28 (2010) 255–280
2.2.1. Distribution profile of EPS by fluorescence and CLSM techniqueThe EPS, a complex bio-polymer secreted by the microbial com-
munity as a response to stress condition, consists of polysaccharides(Costerton et al., 1981), proteins, nucleic acids (Frolund et al., 1996;Nielsen et al., 1996) and lipids (Neu, 1996; Takeda et al., 1998). Anin situ composition and spatial distribution of EPS in bioaggregateshas been conducted (Neu et al., 2002; Strathmann et al., 2003; Staudtet al., 2004) and the accumulation of the secreted EPS correlatedwith the occurrence of biological adhesion and microbial co-aggregation of constituent cells (Costerton et al., 1981; Cammarotaand Sant'Anna, 1998; Adav and Lee, 2008; Adav et al., 2008b).Heydorn et al. (2000a,b, 2002) used the CLSM images for quantifyingkey biofilm parameters, though, the said quantification was basedon fluorescence intensities or signals in CLSM image that would besubject to errors by CLSM settings including pinhole size, detectorgain, laser intensities, amplifier offset/gain and other protocol param-eters (Sekar et al., 2002). Cross-research comparisons of CLSM dataare generally not meaningfully conducted and so it is important thata quantitative analysis of a CLSM image needs consistency of allexperimental parameters in testing.
Studies have used different pre-treatment methods and applieddifferent fluorochromes for EPS in bioaggregates. For instance, theEPS contents of activated sludge floc (Neu et al., 2002; Staudt et al.,2004), nitrifying biofilm (Boessmann et al., 2004), and granular sludge(McSwain et al., 2005; Chen et al., 2007a,b) were stained withoutany pre-treatment or fixation. Lawrence et al. (2007) first fixedtheir biofilm with formaldehyde solution and then used FISH toinvestigate the composition and spatial distribution of EPS and themicrobial community. A similar fixation protocol was employed tostain Pseudomonas aeruginosa biofilm directly on polycarbonate-membrane filters (Strathmann et al., 2002). Kim et al. (2006) fixedthe biofilm on membranes with 4% Paraformaldehyde solution anddehydrated it at high ethanol concentrations. Effects of fixationmethods on the structure and form of biofilm have not been deter-mined. Nosyk et al. (2008) compared the effects of different pre-treatment and fixation methods while staining EPS components ofbiofilm and activated sludge; and concluded that pre-treatment withparaformaldehyde and formaldehyde are appropriate with the leastimpact on their structure.. The dehydration using ethanol leads tosample shrinking. Other fixative and/or pretreatments procedureswith formaldehyde or para-formaldehyde stop metabolic function inorder to preserve the microstructure and are not recommended forimmunofluorescence staining. The sequential processes of specimenfixation, dehydration and embedding often affect the proteinantigenicity adversely, which may result in some cases of failure(Brandtzaeg, 1982; Puchtler and Meloan, 1985; Larsson, 1988, 1993;Griffiths, 1993). Aldehyde fixation has been recognized as one of themajor causes for immunohistochemical failures in the past (Berodet al., 1981; Brandtzaeg, 1982). The coagulating fixative techniquesare suitable for immunofluorescence staining as they are easy toapply, reproducible and preserve the antigen recognition sites forimmunolabeling. Profound shrinkage of the specimen, however, is amajor problem in applying coagulative fixatives. The ideal fixativeshould penetrate the specimen samples quickly, act rapidly, preservethe cellular structure before the cell can react to produce structuralartifacts and it must be determined by experimental requirements.
2.2.2. Protein distributionFITC, a fluorochromes used for localizing protein, has the excitation
and emission wavelengths of 488 nm/520 nm and is a derivative offluorescein used in wide range applications including flow cytometry.FITC is the original fluorescein molecule functionalized with anisothiocyanate reactive group, replacing a hydrogen atom on thebottom ring of the structure. This derivative is reactive towardsnucleophiles including amine and sulfhydryl groups on proteins.
Proteins have been reported to be located at the center of the aerobicsludge floc, peptone and glucose fed aerobic granules (McSwain et al.,2005) and anaerobicfloc and granules (DeBeer et al., 1996). Conversely,proteins were noted to be distributed throughout the entire phenol fedandacetate-fed aerobic granules (Chen et al., 2007a,b;Adavet al., 2007a,2008b; Adav and Lee, 2008). Even though the FITC staining demon-strated high fluorescent intensities of FITC at the center of the granulesthe fact that a negative SYTO 63 stain suggested that the center wascomposed of dead cells, which had leaked proteins and other amine-containing compounds into the granule center (McSwain et al., 2005;Chen et al., 2007a,b; Adav et al., 2007a, 2008b; Adav and Lee, 2008). Aword of caution regarding the use of SYTO 63 should be mentionedbecause of the possible incorporated mass transfer limits throughcompact bioaggregate layers.
A new era of fluorescent probe technology and cell biology wasopened by the discovery of the green fluorescent protein (GFP) fromjellyfish and the development of mutant spectral variants. In nativegreen fluorescent protein, the fluorescent moiety was a tripeptidederivative of serine, tyrosine, and glycine that requires molecularoxygen for activation, but no additional cofactors or enzymes (Zimmer,2002). This discovery of fluorescent protein, initiated multicolorinvestigation of protein distribution, intermolecular interactions andtrafficking in living cell cultures (Miyawaki et al., 2003; Zhang et al.,2002; Lippincott-Schwartz and Patterson, 2003; Tsien, 2005).
2.2.3. Lipid distributionNile red is an uncharged heterocyclic molecule and thus is quite
soluble in organic solvents and lipids, but relatively insoluble in water.Nile red (also known as Nile blue oxazone) is a lipophilic stain havingan excitation wavelength of around 450–500 nm and an emissionwavelength of greater than 528 nm and acting on intracellular lipiddroplets. Nile red acts as a hydrophobic probe and its fluorescencemaxima vary depending on the relative hydrophobicity of the sur-rounding environment. For example, when it is dissolved in hy-drocarbon solvents such as heptane or in neutral lipids such astriacylglycerol or cholesteryl ester droplets, it fluoresces yellow-gold, but when dissolved in more polar solvents such as ethanol,phosphatidylcholine, the dye fluoresces red. Fowler and Greenspan(1985) used this novel property of Nile red to develop a sensitivefluorescent histochemical stain for tissue lipids. Nile redwas applied tothe environmental samples such as microbial aggregates to study thedistribution of lipids (Lawrence et al., 1998, Chen et al., 2007a,b; Adavand Lee, 2008). The lipids were found distributed at the outer edge ofthe phenol and acetate-fed granules (Chen et al., 2007a,b; Adav et al.,2007a, 2008b; Adav and Lee, 2008).
2.2.4. Polysaccharide distributionLectins are versatile probes for detecting glycoconjugates in
histochemical and flow cytometric applications and for localizingglycoproteins in gels. In neutral and alkaline solutions, concanavalin Aexists as a tetramerwhile in acidic solutions (pH below 5.0), it exists asa dimer. The fluorescence of FITC-Con A was quenched by forming anFITC-Con A–glycogen conjugate and de-quenched upon addition ofsugars to the conjugate solution due to disaggregation of the conjugate(Sato and Anzai, 2006). Chen et al. (2007a,b) used Con A–rhodamineconjugate to stain α-mannopyranosyl and α-glucopyranosyl residuesin bioaggregates.
The fluorescence brightener, CW, detects cell wall materialsin fungi algae and plants. The CW has high affinity for cellulose andchitin but obviously also interacts with polysaccharides and iscommonly used in clinical studies and environmental biology. Forinstance, orosomucoid (α1-acid glycoprotein), a small acute-phaseglycoprotein is negatively charged at physiological pH, consistingchain of 181 amino acids (40% carbohydrate by weight), and up to 16sialic acid residues (10–14% by weight) was detected using fluoro-chrome calcofluor white (Kute and Westphal, 1976). Calcofluor
interacts with sialic acids, the terminal glycan residues, and thusinteracts with the carbohydrates at the surface of the protein. Inanother study, Pneumocystis carini in respiratory specimens of ratswas detected by CW staining technology. The CW has also beenapplied to probe polysaccharide distribution in aerobic granules(McSwain et al., 2005; Chen et al., 2007a,b; Adav et al., 2007a, 2008b,Adav and Lee, 2008) and sludge flocs (Chen et al., 2007a,b).The polysaccharides excreted by cells assist cell-to-cell attachmentby bridging cells (Ross, 1984) and/or reversing the surface charge ofdispersed cells to coagulate them together (Shen et al., 1993; Schmidtand Ahring, 1994). The polysaccharides were shown to be heteroge-neously distributed in biofilms (Neu et al., 2001; Lawrence et al.,2007; Laue et al., 2006; Strathmann et al., 2002). Laue et al. (2006)characterized the distribution of three polysaccharides in biofilmsthat were produced by the plant pathogen Pseudomonas syringae andfound that the polysaccharide, levan, localized in the biofilm interiorand the polysaccharide that binds the Naja naja lectin was distributedin a fibrous structure in regions that were not occupied by levan.
2.2.5. Live and dead cells distributionThe SYTOdyes are cell-permeant nucleic acid stains that differ from
each other in one or more characteristics, including cell permeability,fluorescence enhancement upon bindingwith nucleic acids, excitationand emission spectra, DNA/RNA selectivity and binding affinity. Theeukaryotic cells incubatedwith SYTOdyes generally show cytoplasmicor mitochondrial staining as well as nuclear staining. Mitochondrialstaining predominates in yeast and animal cells stainedwith SYTO 59–64 stains. In addition, SYTOdyeswill stainmost live and permeabilizedbacteria and arewidely used in environmental biological samples suchas sludge granules. The SYTO dyes are suitable for staining of livingsamples without the need for prior cell fixation (Neu et al., 2002).These dyes can be excited at 633 nm and thus can be used with CLSM-systems that do not have expensive UV laser, as needed by othercommonly used nuclear stains such as DAPI (McSwain et al., 2005;Chen et al., 2007a,b; Adav et al., 2007a, 2008b; Adav and Lee 2008;Zhang et al., 2008).
The SYTOX Blue stain has high-affinity for nucleic acid and caneasily penetrate the cell with compromised plasma membranes butwill not cross uncompromised cellmembranes. After a brief incubationwith SYTOX Blue stain, the nucleic acids of dead cells fluoresces brightblue when excited at 405 nm. This stain has been widely used inflow cytometry and DNA damage study in clinical and environmentalbiology such as sludge floc, both aerobic and anaerobic granule andbiofilms.
2.3. Multicolor fluorescence experiments
Multicolor fluorescence experiments have received increasinginterest in recentyears for exploringhowEPS and/or cells are distributedin bioaggregates. The principal obstacle for applying multiple fluo-rochromes on the same specimen is the overlapping of the spectralproperties (excitation and emissionwavelengths) of the fluorochromes.Highly selected fluorochromes with minimum spectral cross over wereadopted (Chen et al., 2007a,b). To date, amaximumof sixfluorochromesona single biological specimenhadbeen appliedonaerobic granules andmembrane (Chen et al., 2006) and on anaerobic granules (Zhang et al.,2008). The “recipe” for the multicolor fluorescence experiments isnow discussed and a new test of simultaneous staining with sevenfluorochromes is proposed.
2.3.1. Scheme designThe general principles for designing a multicolor staining scheme
for granules have been presented (Chen et al. 2007b). That is, one ofthe following conditionsmust bemet: (1) if spectra of all fluorophoresdo not overlap, the fluorophores are excited in a one-by-one mannerwhen adequate light sources are available; and, (2) if parts of the
emission spectra of all fluorophores do not overlap, then the emittedlight can be collected in a one-by-one manner using limited collectionwavelength width.
Table 1 shows the cross check between two channels. Owing tooverlapping excitation and emission wavelengths the Con A, NileRed, and TRITC cannot be applied to the same sample nor can CWand DAPI for the same reason. Moreover, the application of Nile Redinterferes with application of many other stains. In parallel tests theaerobic granule section reveals auto-fluorescence at the sameemission regime for Nile Red. Table 1 can be used to select appropriatestains for the proposed scheme. For instance, based on cross-checkdata in Table 1, the distributions of total DNA and lipid in the samespecimen can be determined if Nile Red and DAPI, rather than SYTO63, were selected.
Additional limits for immunohistochemical staining are thatprimary and secondary antibodies must not be obtained from thesame host, and secondary body must be against the IgG of the host inwhich the primary antibody has been raised.
Table 2 shows the excitation and emission wavelengths for thepresently used dyes and the associated targets for EPS and specificenzymes.
2.3.2. Demonstrative seven fluorochrome testLee et al. (1999a,b) proposed using Calcium Green in addition six
other fluorochromes used by Chen et al. (2007a, b) (Fig. 2).
1. The sample was suspended in the sterile distilled water. The waterwas removed by decantation and samples washed two–three timeswith sterile distilled water and then suspended in 1× PBS buffer.[200 μl].
2. SYTO 63 (20 μM, in 1× PBS buffer, pH 7.2) was first added to thesample and placed on a rotary shaker (rpm: 100) for 30 min.The excess dye solution was removed by washing with 1× PBSbuffer.
3. The sodium bicarbonate buffer (0.1 M, pH 9.0) was added, tomaintain the amine group in non-protonated form, followed by aFITC solution (1 mg/ml in DMSO), and the sample was incubatedon a rotary shaker for 30 min. After staining the excess stain wasremoved by PBS buffer.
4. The sample was incubated with Con A–tetramethylrhodamineconjugate solution (0.25 mg/ml in 1× PBS buffer, pH 7.2) for afurther 60 min. After staining the excess stain was removed by PBSbuffer.
5. The sample was successively stained by CW (30 mg/ml, in 1× PBSbuffer, pH 7.2) for 30 min. After staining the excess stain wasremoved by PBS buffer.
6. The Nile red solution 100 μl (1 mg/ml), in acetone was diluted to10 ml with PBS buffer, pH 7.2) was added and samples wereincubated for 25 min. Excess stain was washed with PBS buffer.
7. The Calcium Green solution 100 μl (0.12 mg/ml in DMSO) wasadded and samples were incubated for 30 min. PBS buffer washingremoved the excess stain.
8. Before observation, the sample was stained with SYTOX Blue(2.5 μM in PBS buffer. pH 7.2) for 5 min without further washing.
The phase contrast image in Fig. 2a demonstrated that thegranule was a non-spherical object with internal cracks. The stainingresult in Fig. 2, however, conveys much more information onthe EPS distributions in the granule. It revealed that the proteinsand β-D-glucopyranose polysaccharides were distributed over theentire granule interior, probably supporting the architecture of thegranule; the live cells accumulated at the outer layers of granules,probably taking the advantage for easy assessment of incomingnutrient; while the lipid shown may be bound with cell membranes.Furthermore the calciumwas distributed mainly in the outer layers ofthe granule, in contrast to the literature results (Ren et al., 2008). It
Table 1Cross-talk check between stains at prescribed ex/em wavelengths (modified from Lee et al. (2009b) with permission).
aExcitation/emission wavelengths (nm/nm) at: 561/570-590 for Con A, UV/410-480 for CW, 488/500-550 for FITC, 514/625-700 for Nile red, 633/650-700 for SYTO 63, 458/460-480 for SYTO blue, 561/597-607 for Alex 568, 633/650-670 for Alex 633, 514/644-658 for Mega 520, UV/400-480 for DAPI, 561/590-610 for TRITC. bCW: calcofluor white fromSigma. cFAM: Fluorescein phosphoramidite from Sigma; calcium green from Invitrogen. dSYTO blue from Molecular Probes. eAlexa Fluor 568 goat anti-rabbit IgG Cy3-conjugatedA11031 antibody from Abcam. fAlexa Fluor 633 goat anti-sheep IgG Cy5-conjugated A21071 antibody from Abcam. gMeg520 Anti-rabbit IgG from Sigma.
261S.S. Adav et al. / Biotechnology Advances 28 (2010) 255–280
has been demonstrated herein that information can be extracted fromthe scanned CLSM slices with multiple fluorescence experiments.
2.3.3. EPS-enzyme assayThis section demonstrates a staining result with three fluoro-
chromes on the surface of the aerobic granule specimen (Fig. 3). Thestaining protocol is outlined here.
1. The collected granules were kept fully hydrated during staining.The granules were embedded for cryosectioning and froze at−20 °C, after which 10 μm sections were cut on a cryomicrotomeand mounted onto the microscope slides for observation.
2. During staining, DAPI (10 mg ml−1) (100 μl) was first dripped ontothe sample and was placed on a shaker table for 30 min.
3. The β-amylase ab617 from rabbit and galactosidase antibodyab9361 from chicken at 1:20 dilution were dripped onto thesections, and were incubated for 2 h at 37 °C [100 μl].
4. After the residual primary antibody solution was removed, thesections were incubated with 2 mg ml−1 Alexa Fluor® 568 goatanti-rabbit IgG Cy3-conjugated A11031 antibody [50 μl] and2 mg ml−1 Alexa Fluor® 488 goat anti-chicken IgG A11039 anti-body at 1:50 dilution for 40 min at 37 °C. [50 μl].
5. Finally, the stained section was washed twice in 1% v/v Tween 20PBS solution.
Table 2Excitation and emission wavelengths for dyes and associated targets.
Dye Ex(nm)
Em(nm)
Targets
Calcofluor white 400 410–480 Beta polysaccharideSYTOX Blue 458 460–480 Dead cellsFITC/FAM/Calcium Green 488 500–550 Protein/DNA/Ca2+
Nile red 514 625–700 LipidCon A(tetramethylrhodamineconjugate )
6. After immunohistochemical staining, 0.1 M sodium bicarbonatebuffer (pH 9.0, 100 μl) was added; then, the FITC solution (10 g l−1)[100 μl] was added to the sample for 1 h at room temperature.
Clearly the filamentous microorganism on the surface of phenol-fed granule (identified by DAPI) was surrounded by α-polysaccharidelayer (identified by Con A), and has certain activity of α-amylase(Alexa 568 goat anti-rabbit IgG Cy3-conjugated A1031). Detailedstudy on the correlation of the secreted enzyme on the surroundingpolysaccharide layer can be done with this type of staining scheme.
2.3.4. EPS-FISH testThe FISH technique detects and localizes the presence of specific
bacteria in granules, biofilms and specific DNA sequences on chromo-somes in clinical biology. This technique is based on the hybridizationof labeled DNA probes to taxon-specific regions of the bacterial ribo-somes and can be detected by fluorescence microscopy or flowcytometry (Amann et al., 1990; Wallner et al., 1993). Using FISHtechnique, Chen et al. (2009) located the aerobic ammonia-oxidizingbacteria (AOB) and aerobic heterotrophic bacteria in the aerobic outerpart of the biofilm; and proved the presence of active ANAMMOX(anaerobic ammonium oxidation) and denitrifying bacteria in theanaerobic inner part of the biofilm. Multiple staining of biofilms withnucleic acid specific fluorescent compounds such as acridine orange orDAPI have been reported (Neu et al., 2002). Okabe et al. (1999)stained the Gram-negative filamentous bacteria in a nitrifying biofilmusing oligonucleotide probes with different labels. Triple hybridiza-tions using FITC, Cy3, and Cy5 labeled oligonucleotide probes forsimultaneous in situ visualization of the genetic diversity of activatedsludge organisms or in a trickling filter biofilm have been deter-mined (Amann et al., 1996; Schmid et al., 2000). The biofilm washybridized respectively with probes ANA103-Cy3, EUK516-FITC, andSOB174-Cy5 to identify Actinomyces naeslundii, Candida albicans, andStreptococcus sobrinus, and was stained with Calcofluor for probingthe (polysaccharides of) EPS (Thurnheer et al., 2004). Combined EPSstaining and FISH technique helped interpret biofilm formationmechanisms in situ. The unlabelled species in the biofilm, however,yielded errors for estimating void volume and pore structure in thebiofilm.
This section demonstrates a staining result with CW for β-polysaccharides and the probe 5′-GCA CTT AAG CCG ACA CCT-3′ for
Fig. 3. The CLSM images of filaments seen on the surface of phenol-fed granule. Bar=2 μm. (a) Alexa 568 goat anti-rabbit IgG Cy3-conjugated A1031 (α-amylase), (b) Con A(α-polysaccharides), (c) DAPI (total cells), (d) merged image of a–c. (For interpretation of the references to colour in this figure legend, the reader is referred to the webversion of this article.)
Fig. 2. The images of acetate-fed granule. (a) Phase contrast image; (b) green (FITC) proteins; (c) yellow (Nile red) lipids; (d) cyan blue (ConA) α-D-glucopyranose polysaccharides;(e) red (SYTO 63) nucleic acids; (f) pink (SYTOX Blue) dead cells; (g) blue (calcofluor white) β-D-glucopyranose polysaccharides; (h) purple blue (Calcium Green) calcium.Bar=300 μm. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
263S.S. Adav et al. / Biotechnology Advances 28 (2010) 255–280
Bacteroides sp. in a stored granule. The staining protocol is outlinedhere.
1. The collected granules were kept fully hydrated during stainingand FISH.
2. The staining was conducted by adding CW (fluorescent brightener28, Sigma, USA) solution (300 mg l−1, 100 μl) to the sampled granulefor 30 min.
3. The stained granule was washed twice with phosphate-bufferedsaline (pH 7.2) to remove excess stain.
4. The granule was then hybridized for probing the anaerobic Bac-teroides strain using hybridization buffer (0.9 M NaCl, 20 mM Tris–HCl, pH 7.4, 0.01% sodium dodecyl sulfate, 50% formamide)containing 5 ng μl−1 of probe (5′-GCA CTT AAG CCG ACA CCT-3′)labeled at the 5′-end with Fluorescein phosphoramidite (FAM dye)for 2 h at 48 °C.
5. The stained and hybridized samples were analyzed by CLSM (LeicaTCS SP5, Confocal Spectral Microscope Imaging System GmbH,Germany).
The staining/hybridization result revealed that the strict Bacter-oides anaerobe was present in the core regime of stored granules(Fig. 4). These strains generally perform anaerobic digestion reactionto hydrolyze and degrade organic compounds in an anaerobicenvironment. This type of fluorescence experiments can provideinformation on the embedded biological reactions involved inbioaggregates.
2.3.5. Effects of sample pre-treatments methodsNosyk et al. (2008) determined that the washing, immobilization
and fixation steps did not deteriorate the sludge floc structure. Theseauthors mentioned, however, that the fixation chemicals in FISHwould strongly affect the floc shape. To determine the possible effectsof pretreatment methods on the distributions for EPS inside anaerobic granule, acetate-fed aerobic granule sampleswere testedusingdifferent pre-treatment methods.
The granule sample (6.0 g wet weight) was washed with de-ionized water, and then divided into three parts.
The group (i) sample was suspended in a 1× PBS buffer (pH 7.2)and stained as described below.
Fig. 4. FISH-CLSM image at the core regime for the Bacteroides sp. (5′-GCA CTT AAG CCGACA CCT-3′) in a stored granule (Bacteroides sp.: green; β-polysaccharides: blue). (Forinterpretation of the references to colour in this figure legend, the reader is referred tothe web version of this article.)
The group (ii) sample was further divided into two parts andseparately fixed by the addition of 3.7% formaldehyde solution, withpart 1 being incubated for 2 h at room temperature and part 2, 2 hat 4 °C. The fixation solution was removed by decantation andthe samples were washed with de-ionized water and resuspended in1× PBS buffer (pH 7.2).
The group (iii) sample was divided into two parts and fixed withparaformaldehyde-phosphate buffer solution (4%, 2 h) with part 1at room temperature and part 2 at 4 °C. The fixation solution wasremoved, samples were washed with de-ionized waster and resus-pended in 1× PBS buffer (pH 7.2).
All samples from the three groups were stained for EPS usingprocedures modified from Chen et al. (2007a, b) (Fig. 5).
The comparison made here indicated that there was no need forfixation in case the following steps for fluorescent observation couldbe done within a short period of time (b 1 h). A simple deionisedwater washing stepwas sufficient to secure the following staining andobservation steps. Fixation with group (ii) sample would not affectthe dye penetration, but group (iii) hindered the penetration of dyeCW and SYTO63. In case one needs to store the stained sample fora long time before observation, fixation by 4% paraformaldehyde–phosphate buffer solution was not recommended.
2.3.6. Staining and mass transfer limitBioaggregates, such as aerobic granules, have a dense surface
layer and compact interior core that can generate significant masstransfer resistance to oxygen and nutrient intake. The surface layers ofthe granules comprise active cells, lipids and polysaccharides onthe outer edge of the granules at 200–300 μm thickness (Chen et al.,2007a,b). Oxygen diffusion limit probed by microelectrode measure-ment (Chiu et al., 2006, 2007a,b) and granule interior permeabilityby size exclusion chromatography (Adav et al., 2008a) have beenreported. Successful staining requires efficient fluorochrome pen-etration through the compact layer of cells, lipids and polysaccharides.Modeling of the substrate diffusion in biofilms and granules werestudied (Wanner and Gujer, 1985; Picioreanu et al., 1998; Li andLiu, 2005). Pores and channels presented in microbial aggregates(Massol-Deya et al., 1995), thereby may reduce the mass transferlimit for fluorochrome staining. Tay et al. (2002) reported apenetration depth of 800–900 μm from the granule surface of amodel dye using FISH. Meanwhile, Jang et al. (2003) reported a pen-etration depth of 700 μm for dissolved oxygen considering both masstransfer limit and chemical reaction consumption in aerobic granularsystem.
Tsai et al. (2008) monitored in situ penetration dynamics offluorochromes Con A–tetramethylrhodamine conjugate, Nile red, CWand SYTO 63 in phenol-fed aerobic granules. These authors deter-mined minimal mass transfer limits for the former three dyes, but alimited mass transfer rate for the SYTO 63 into the granule. All testsusing SYTO 63 as a nucleic acid dye thereby needed re-examination.Lee et al. (2009b) tested the penetration of antibody into aerobicgranules with a compact interior. A total of 100 μl of tested antibody(0.1 mg ml−1) was suddenly injected into the section-containingtunnel, and 18 CLSM images of the section were collected at a 156 sinterval at 30 µm thick from the granule top surface. These authorsnoted that the antibodies they used experienced no serious masstransfer limit.
It was difficult to predict whether mass transfer limit was presentin specific staining/hybridization systems. Transient tests as demon-strated in the following example were suggested before conductingcomprehensive study. Fig. 6 shows the multicolor staining test withsudden injection of 100 μl Alexa 430 goat anti-sheep IgG-Cy5-conjugated A21071 antibody. It was noted that this antibody couldpenetrate into the granule at b600 s. One can set a staining time of15 min with this antibody for releasing mass transfer limit.
Fig. 6. Time course of intensities of CLSM images taken at 90 s interval (top left to down right) with sudden injection of Alexa Fluor® 430 goat anti-sheep IgG Cy5-conjugated A21071antibody.
Fig. 5. The images of acetate-fed granule fixed with formaldehyde. (a) green (FITC) proteins; (b) yellow (Nile red) lipids; (c) cyan blue (ConA) α-D-glucopyranose polysaccharides;(d) pink (SYTOX Blue) dead cells; (e) blue (CW) β-D-glucopyranose polysaccharides; (f) red (SYTO 63) nucleic acids. Bar=200 μm. (For interpretation of the references to colour inthis figure legend, the reader is referred to the web version of this article.)
265S.S. Adav et al. / Biotechnology Advances 28 (2010) 255–280
3. Assessing stereological information from acquired images
Quantitative description of the spatial correlations among theprobed substances using fluorescence needs conversion of scannedcolor images to digital formswith criterion set by the acquired analysis.
Detailed modeling and analysis can be conducted by imageprocessing procedures on the acquired CLSM images.We demonstrateherein the procedures using a classical hydrodynamic question: howa porous floc moves steadily through an infinitely large Newtonianfluid (Yang et al., 2007a,b,c). The presented procedures aregenerally applicable to a series of images from a 3D object regardlessof its scanning methods. We also demonstrate an example how touse correlative analysis for extracting stereological data from thescanned images of a 3D object (Section 3.6).
3.1. Sample images
We utilized the activated sludge collected at the Min-ShengMunicipalWastewater Treatment Plant in Taipei as the testing sample.It was gravitationally settled to a solid content of 14,700 mg l−1, andwas the tested sample. The pH value of the sludge was approximately6.4. The COD of the sludge was 24,400 mg l−1, determined using adirect reading spectrometer (DR/2000, Hach Co.). The COD of thefiltrate of the sludge sample after it was filtered through a 0.45-μmmembrane was called SCOD, which was 410 mg l−1 for the originalsludge.
The aggregate samples were first chemically fixed and thenembedded in agarose to perform FISH. This study used DNA probesEub 338 (labeled by rhodamine) to detect most eubacteria in bio-solids. Its nucleotide sequence is, 5′-GCTGCCTCCCGTAGGAGT-3′. Afterthe probe was added, hybridization was performed at 50 °C for 1 h.The stained samples were washed three times at the same condi-tion using hybridization buffer solution (40% formamide, 0.9 M NaCl,0.01 M sodium dodecyl sulfate, and 20 mM Tris–HCl) to removeexcess probes.
The interior structure of the stained samples was observed directlyusing a confocal laser scanning microscope (CLSM; Leica TCS SP2,Germany) or the sliced (2 μm thickness) images using a phase-contrast microscope. The CLSM was equipped with image processingsoftware and an argon laser source was used to stimulate fluores-cence. The sludge floc was imaged using 20×, 40×, or 60× objectives,depending on the required resolution. The microscope scanned thesamples at a fixed depth and the obtained image was digitized.Samples at two depths (47 μm and 94 μm) from the sample surfaceare shown in Fig. 7.
3.2. Acquisition of two-dimensional images
The obtained images may be unable to demonstrate the real cross-sectional structure of the samples owing to interference by lightreflection and scattering. The light intensity of each obtained imagedecreases with scanning depth; as a result, the average intensity of theCLSM images varied over the depth. The CLSM images are processedusing the following five steps to remove the inherent intensity bias forthe collected images.
3.2.1. Converting original color images to grey imagesEach pixel of the CLSM image is composed of red, green, and blue
(RGB) color information. The void and solid phases of the object canbe identified by bileveling the image to black (value=0) and white(value=1). The NTSC color category was introduced to use aluminance-chrominance encoding system, whereas luminance takesthe place of the original monochrome signal and chrominance carriesthe composite color information. Complete separation between theluminance and the chrominance information was realized by usingthe NTSC. The components of the NTSC color are Y (the luminance
component), I (the cyan-orange component), and Q (the green-purple component). The luminance component (Y) can be directlyconverted to a grey-scale image and the convert algorithm betweenRGB and NTSC can be expressed by
where R, G and B are the intensities of the red, green and blue in thecolor images. Fig. 7c and d show the grey images corresponding toFig. 7a and b converted based on Eq. (1).
A MATLAB® program, Appendix A, is applied to convert the RGBimages to grey images. The pixels in grey images have values between0 (black) and 255 (white) and the light portions correspond to thesolid matrix and the heavy portions correspond to the void space.
3.2.2. Resizing the grey imagesThe grey images obtained in Section 3.2.1. contain complete
luminance information. The spatial intervals between the neighboringCLSM images in the series are generally greater than the spatial size ofthe neighboring pixels on the 2-D image. For instance, the sampleimages discussed herein have a distance between neighboring pixels of0.221 μm, which is much smaller than that between two neighboringpixels from the CLSM images (1.118 μm). These very different spatialdistances will yield distorted object elements in a reconstruction step(described later). The pixels on each image are hence resized usingthe following relation:
∑SR
pi⋅dSi = pRSR ð2Þ
where SR is the area occupied by a resized pixel, dSi is the areaoccupied by a pixel in SR, and pi is the pixel value (0≤pi≤255).
A MATLAB® program, Appendix B, is applied to resize pixels. Fig. 8shows the comparison between images before and after resizing.
3.2.3. Bileveling of the grey imagesAssume that all analyzed images are 8-bit grey-scale images
(HGREY). Several variables are first defined for the sake of followinganalysis,
pðiÞ = HGREYðiÞ
∑255
0HGREYðiÞ
ð3Þ
PðiÞ = ∑i
0p; P0ðitÞ = ∑
it
0p; P1ðitÞ = ∑
255
it + 1p = 1−P0ðitÞ ð4a–cÞ
G0ðitÞ = ∑it
0p2; G1ðitÞ = ∑
255
it + 1p2 ð5a–bÞ
μT = ∑255
0ðipÞ; μ0ðitÞ =
∑it
0ðipÞP0
; μ1ðitÞ =∑255
it
ðipÞ
1−P0ð6a–cÞ
σ2T = ∑
255
0½ði−μT Þ2p�; σ2
0 ðitÞ =∑it
0½ði−μ0Þ2p�
P0; σ2
1 ðitÞ =∑255
it + 1½ði−μ1Þ2p�
1−P0ð7a–cÞ
HT = −∑255
0ðp lnpÞ; HtðitÞ = −∑
it
0ðp lnpÞ ð8a–bÞ
Fig. 7. Color images (above) and the converted grey images (bottom): scanned at thickness of a and c (61 μm) and b and d (94 μm).
The volume-average pore size dp[4,3] and surface-average poresize dp[3,2] were determined from the following definitions,
dp½4;3� =∑d4p;ini
∑d3p;ini=
∑dp;iVi
∑Við9Þ
dp½3;2� =∑d3p;ini
∑d2p;ini=
∑Vi
∑d−1p;i Vi
ð10Þ
where ni is the number fraction of dp,i. The same definition can beapplied on the floc size to determine df[4,3]. The specific surface areaσwas then estimated by dividing the sum of the pore perimeters withthe area of analyzed region,
σ =∑pp;iL2equi
ð11Þ
where Lequi is the characteristic length.Selecting appropriate grey value, i, to minimize within-class vari-
ance, σL2, or to maximize between-class variance, σJ
2, yields theoptimum i value. Fig. 9 demonstrates the variation of σL
2 with respectto it in a given between-class variance structure. A MATLAB® program,Appendix C, is applied to bilevel the image. Fig. 10 demonstratesthe believeled image of Fig. 3 using program in Appendix C.
3.2.4. Modifying the bileveled imagesSpecific part of the object of interest can be cut off from the original
image for further processing. For instance, the aforementioned CLSMimage series can be cut and shaped into a spherical object. The first
step is to determine the center and radius of the sampling space basedon the scanning distances between the collected two-dimensional,bileveled images, hence determining the circles on each slides basedon their spatial positions in the floc (Fig. 11). Then the pixels locatingoutside the circle are set as 255 (white, void).
A MATLAB® program, Appendix D, is applied for the mentionedimage cutting function. For instance, the position of circle center ofimage #31 is at (50.0, 50.5), with corresponding circle radius of 30.5pixels.
3.2.5. Reversing image colorsThe colors of void and solid phases of the bileveled images are
reversed for three-dimensional reconstruction of two-dimensionalslices (Fig. 12). A MATLAB® program, Appendix E, is applied for colorreversion.
After all the processing steps, all images, each with 101×101×81pixels should be saved as “Image_000.tif” in a series of TIF files in thesame folder.
3.3. Three-dimensional reconstruction of two-dimensional images
Numerous software programs, including the commercial software,Amira™ 3.1 (TGS Inc., USA), can build up three-dimensional objectfrom a series of bileveled two-dimensional images from the files savedin Section 3.2.5.
After starting the Amira™ 3.1 program, choose File\Load todownload the bileveled image files. In the present example, a totalnumber of 81 images are loaded. The spatial size of a pixel on eachimage has been adjusted to be the intervals between the neighboringimages and set the Voxel size to 1 (default), This indicates the lengths
Fig. 10. Image of Fig. 8b: (a) before bileveling and (b) after bileveling.Fig. 8. Grey images at the depth of 81 μm: (a) before the action of ‘resizing’ (pixel: 512by 512); (b) after the action of ‘resizing’ (pixel: 512 by 512).
267S.S. Adav et al. / Biotechnology Advances 28 (2010) 255–280
of a pixel in the x-, y- and z-directions are all unity in length. Theimages will be sequentially arranged as in Fig. 13.
The image file is first uploaded to the Amira™ 3.1 program. Thethreshold value to differentiate the void and solid phases can be anyvalue between 0 and 255 since the images have been bileveled. Aftersetting the threshold value, the steps listed Appendix F are performedto obtain Fig. 14a and b.
Fig. 9. The between-class variance (y-axis) versus ‘threshold value’ (x-axis) from theimage of Fig. 8b.
The grids are then checked for their validity in subsequent analysissteps (AppendixG). The three-dimensional object viewcanbeproduced(Fig. 15) and saved as “Image_000.grid.” The grid file can be used forfurther numerical study on incorporated chemical reactions and trans-port processes in the studied matrix. In the next section, the producedgrid file (Fig. 15) is used in CFD applications.
Fig. 11. Determining the circles on each slides based on their spatial positions in thefloc.
Fig. 14. Surface generation composing a series of 2D bileveled images following steps inAppendix F.Fig. 12. One typical image: (a) before color-conversion and (b) after color-conversion.
Yang et al. (2008) confirmed the image acquisition processes byexamining images from two deterministic multifractals.
3.4. From three-dimensional grid file to CFD mesh file
Numerous CFD software packages are available for fluid flow andheat and mass transfer simulations. In this section the saved grid file
Fig. 13. Assembly of series of 2-D bileveled images.
of Fig. 15 is placed into a uniform flow field at constant translationvelocity U (Fig. 16).
The produced grid file from Fig. 15 is converted by GAMBIT (FluentInc., USA) to a mesh file for the CFD package, Fluent. One key issue isto modify the meshes in the object and add surrounding meshes.Sample procedures are listed (Appendix H) with the steps 1–42
Fig. 15. The three-dimensional object view of grid file produced in Appendix G.
Fig. 16. Computational domain and boundary conditions: D/df=80 and L/df=80.
269S.S. Adav et al. / Biotechnology Advances 28 (2010) 255–280
producing the structural model (Fig. 17), the steps 44–60 producingthe structural model (Fig. 18) and the steps 61–65 giving the meshmodel (Fig. 19) which is then ready to be loaded into CFD packageFluent for simulation.
Fig. 17. The mesh file at st
3.5. Numerical simulation case
We simulated the flow fields surrounding and inside a modelfloc subjected to an incoming unbound Newtonian fluid. The fluid of
density ρ and viscosity μ is flowing at a uniform velocity of u→∞ frominfinity towards the fixed floc (Fig. 16). The mesh file in Fig. 19 wasapplied to obtain fluid flow field inside and surrounding the modelfloc by solving the following equation:
ð⇀u ′ ⋅∇Þ⇀u ′ +P0ρV2 ∇P′ =
1Re
∇2⇀u ′ ð12Þ
where u⇀′= u⇀/u∞ and P′=P/P0. The corresponding boundary condi-tions are as follows:
⇀u = →1 at tube’s surfaces ð13aÞ
Fig. 19. The completed mesh file for nume
∂→u∂r = 0 at r = 0 ðoutside flocÞ ð13bÞ
→u = →0 at rugged solid surface ðinside flocÞ ð13cÞ
The calculation was performed at a maximum relative error of0.01%. To describe the intra-floc flow field for a floc moving in asedimentation tank, Re=0.1 in this demonstration simulation.
Sample procedures for executing CFD package Fluent™ ver. 6.0(ANSYS, Inc., Santa Clara, CA, USA) are listed in Appendix I. The steps1–27 produce the simulation results in Fig. 20.
rical simulation output from GAMBIT.
Fig. 20. The streamlines of intrafloc flow in studied floc subjected to the incoming flow in the (a) x-direction; (b) y-direction; and (c) z-direction, colored by static pressure(Pa, gauge).
271S.S. Adav et al. / Biotechnology Advances 28 (2010) 255–280
The advectiveflowpatternwas displayed by the intrafloc path linesof the spherical, constructed model floc, with incoming flow along thex-, y- and z-directions (Fig. 20), colored by the corresponding staticpressure. The streamlines of the structuredmodel are in all casesmuchmore tortuous than those of the homogeneous or radially-varyingmodels, yielded by the complicated, inter-connected floc pores withrugged surfaces.
Fig. 21. Angular distributions of flow directional velocity at specific radii of the internal
The angular distributions of flow directional velocity at specificradii of the internal planes of the studied floc are shown in Fig. 21. Thehorizontal lines denote the homogeneous velocity approximation. It isclear that the internal flow is not described by a homogeneous flowmodel as described by most literature works.
The said simulation can be extended to cases with higher Re, suchas in the case for an aeration tank. The fluid stream can flow fast, but
planes of studied floc. Horizontal lines: homogeneous flow approximation models.
the relative velocity between the fluid and the moving floc would bemuch lower. Yang et al. (2007c) noted a minimal effect of Re on thedrag forces exerted on a moving floc with Re up to 100 based on flocrelative velocity. These authors claimed that the pore tortuosity ratherthan Re has a profound effect on the drag force exerted on a movingfloc. Comprehensive simulation considering turbulent flow field in anaeration tank should be conducted in future study.
3.6. Correlating stereological information from scanned object
The constructed 3D grid file based on procedures in Sections 3.1–3.4 can be used for analyzing stereological correlation betweenfluorescent substances in the 3D scanned object. For instance, Ma et al.(2007) probed the spatial distributions of extracellular polymericsubstances and cells in the original wastewater flocs and the yieldedfragments following weak ultrasonication. With the six fluorochromeexperiments proposed by Chen et al. (2007a, b), the stereologicalinformation of the three-dimensional architectures of originalwastewater floc and the ultrasonicated fragments was presentedFig. 22.
The local architecture of the sampled flocs was quantified using astereological method with the pixels of each fluorescent image forproteins,α- and β-polysaccharides, lipids, or cells first bileveled usingOtsu's method, then the pair cross-correlation functions, g(r), foreach pair evaluated. With g(r)N1, the two substances are spatially“attracting”; with g(r)b1, they are spatially “repelling.” The g(r)functions for the original flocs revealed a mono-dispersed character-istic and peaked at around 2–10 μm in distance, indicating that theprobed substances were largely associated with cell membranes andthe peak was correspondent to the distance between neighboringcells. Following ultrasonication, numerous minor peaks emerged onthe g(r) function over a wide range of distance, indicating disturbanceof local architecture. The main peak of g(r), however, remainedunchanged, revealing that the gross structure of floc matrix had notdeteriorated. Following weak ultrasonication, the gross architectureof the floc matrix remained intact, but the disturbances of local
Fig. 22. Combined fluorescent images of sludge floc and tMa et al., 2007.
architecture were probed by the reconstructed pair cross-correlationfunctions. Enhanced methane production from weakly ultrasonicatedsludge is thereby attributed to the reduced interior mass transferresistances associated with the ultrasonicated disturbed structure.
4. Perspectives
In order to investigate any transformation of any biological systemit will be an absolute necessity to exploit and explore the depths of thesystem at a molecular level. Even the complete identification ofparticular genes, proteins and biomacromolecules has specific limita-tions unless their role and function can be characterized. A rationalapproach would be to systematically map the functional propertiesand structural features of the different molecules that play a fun-damental role in the biological system. These features would need tobe integrated, combined, compared and tested to monitor that anydesired effect was accomplished. Unfortunately despite extensiveprogress in biotechnology this requires multiple sets of structural andbiochemical information on every probable biomolecule involved.Furthermore there is no guarantee that the results and parametersobtained from such an in vitro study will reflect the exact biologicalreaction that occurs in vivo. Even directed molecular evolution wheremultiple environmental factors develop or select biomolecules tomeetthe multitude of challenges posed by biological transformations iswrought with uncertainties. The design of a totally new biomoleculethat would be specific for a particular biological transformation isan enormous challenge for biotechnologists. In view of the complexarray of biological conditions the ‘new’ molecule would have to beengineered and/or designed to overcome harsh unfriendly environ-ments such as acidity, temperature changes, solvents and chemicals.Over the next 20 years the biomolecule–bioaggregate model will needto be exploited at a molecular level from its rational design to itsspecific delivery to the target system. The multidisciplinary field ofnanobiotechnology is in its infancywhen it comes to biomolecules andbioprocesses. Nanoparticles will revolutionise current thinking inbiotechnology since their small size and large surface area to mass
he yielded fragment following weak ultrasonication.
A
273S.S. Adav et al. / Biotechnology Advances 28 (2010) 255–280
ratio offers superior advantages over present paradigms. Theseparticles can act as scaffolds to enable the attachment of specificbiomolecules.
Rather than challenge the design and synthesis of these moleculesor their delivery vectors it would make sense to consider theirmechanism of action and develop a bioreactive roadmap instead. Dueto the localisation of biomolecules in every biological transformationwhether they be lipids, proteins, nucleic acids or polysaccharides thereis much to gain with the enhancement and use of specific detectablemarkers. The fluorescence of multicolored stains has shown to beextremely versatile in this regard. Of course as this technologybecomes more and more sophisticated so too is the need forinstrumentation to be equally developed. Confocal laser scanningmicroscopy (CLSM), originally developed in the mid 1950s, stillremains a powerful tool when it comes to investigating multicolorstaining techniques. An enhancement of this technology to take itbeyond its limitation of only using light is currently being investigated.In tandem with the search for better instrument resolution is theprogression of dual technologies such as CLSM-FISH (fluorescence insitu hybridisation), CLSM-DGGE (denaturing gradient gel electropho-resis) and CLSM-FC (flow cytometry).
5. Conclusions
This review has focussed on a unique seven fluorochrome staining/hybridisation and immunohistochemical staining protocol for thesimultaneous allocation of components (proteins, lipids, nucleic acids,α-, β-polysaccharides, antibodies) in bioaggregates such as sludgeflocs and aerobic granules. It discusses the development of fluor-ochromes for both CLSM images and FISH and has laid down a series ofguidelines for their use. The fluorochrome must be delivered to thesite of interest and remain there long enough to acquire theexperimental data. It should not induce significant perturbations ofthe specimen sample structure under study andmolecular associationand metabolic fate of the fluorochromes and probes and fluoro-chrome-fluorochrome interactions and environmental conditions oftemperature, sample pH need to be taken into consideration duringselection and their application. The review also tabulates the spectralproperties and cross references of all of the fluorochrome stainspresently used and recognises which pair of stains cannot be usedtogether because of overlapping emission/excitation spectra.
Finally the review assesses the stereological and spatial correlationsamong the probed substances converting colored images into digitalforms with criteria set by the acquired analysis. With a set of softwareprograms [MATLAB; GAMBIT' FLUENT] an unbiased 3-dimensionaltopology of an image was developed from a 2-dimensional section. Theunderpinning principles were then tested using a classical hydrodynamicquery of how a porous floc moves through an infinitely large Newtonianfluid.
Acknowledgement
This study was partly supported by the State Key Laboratory ofUrban Water Resources and Environments (SKLUWRE), HarbinInstitute of Technology, China.
Appendix A. The MATLAB® program for converting the originalcolor images to grey images
Programming
Interpretation
Clear all
Clear all previous parameters. c=61; Number of untreated images (Total
number of untreated images are 61), andthe number of the first image is 28.
Saving paths and names of sampleimages are decided according to local path.
S3=‘ch00.tif’;
Part of names of colorful images. T1=‘G:\Tutorial_Paper\Grey\Image_’; Saving paths and names of transformed
grey images are decided according tolocal path.
T3=‘.tif’;
Part of names of transformed greyimages.
for n=d:(c+d–1)
S2 is the number of colorful images. S2=int2str(n);if nb100;S2=[‘0’,S2]if nb10;S2=[‘0’,S2]end
endS=[S1,S2,S3];
‘S’ is complete path and name of colorful
image.
[Ind, map]=imread(S); Import colorful images to MATLAB®. I=ind2grey(Ind, map); Convert to grey images. T2=int2str(n–d);
if (n–d)b100;T2=[‘0’,T2]if (n–d)b10;T2=[‘0’,T2]end
endT=[T1,T2,T3];
Complete saving paths and names for
grey images, from 0 to 60. T2 is a numberbetween 0 and 60.
imwrite (I,T, ‘TIFF’);
Saving the grey images. The formats areall ‘tif.’ end
Appendix B. Program to resize pixels on images
The MATLAB programming for resizing.
Programming
Interpretation
clear all
Clear all previous parameters. I0=512 Original image size prior
resizing is 512 by 512 (512*512).
n=61; Numbers of grey images. a=101; Longitudinal pixel size after resizing. b=101; Lateral pixel size after resizing. S1=‘G:\Tutorial_Paper\Grey\Image_’; Saving paths and former part of file
names of untreated grey images(according to the local paths).
S3=‘.tif’;
Saving paths and latter part of filenames of untreated grey images.
T1=‘G:\Tutorial_Paper\Resized\Image_’;
Saving paths and former part of filenames of resized images(according to the local paths).
T3=‘.tif’;
Saving paths and latter part offile names of resized images.
for k=0:(n–1)
Individual treatment of greyimages from number 0 to 60. S2=int2str(k); Saving paths and middle part offile names of grey images.
if kb100S2=[‘0’,S2];if kb10
S2=[‘0’,S2];end
endS=[S1,S2,S3];
Saving paths and names of
untreated grey images
I=imread(S); Import the untreated grey images
into MATLAB®.
C=double(I);la=10/a;lb=10/b;for i=1:a
for j=1:bi1=floor(la*(i–1))+1;j1= floor(lb*(j–1))+1;i2= floor(la*i)+1;j2= floor(lb*j)+1;
A MATLAB program for modifying the bileveled images to make aspherical floc.
Programming
Interpretation
clear all
Clear all previous parameters. a=101; Height of an image (pixels) Longitudinal pixel number of
untreated grey image.
b=101; Width of an image (pixels) Lateral pixel number of
untreated grey image.
c=61; Number of images Number of untreated grey images. a0=50.5; Longitudinal position of the
spherical center.
b0=50.5; Lateral position of the
spherical center.
c0=30.5; Vertical position of the spherical
center (in perpendicular tothe image surface).
R=30.5;
Radius of the sphere.
A A
275S.S. Adav et al. / Biotechnology Advances 28 (2010) 255–280
(continued)ppendix D (continued)
Programming
Interpretation
S1=‘G:\Tutorial_Paper\Resized\Bileveled\Image_’;
Saving paths and former part offile names of untreated images.
S3=‘.tif’; % the last part of the path
Saving paths and latter part offile names of untreated images.
T1=‘G:\Tutorial_Paper\Resized\cut\Image_’;
Saving paths and former part offile names of treated images.
T3=‘.tif’;
Saving paths and latter part offile names of treated images. for n=0:(c−1)
S2=int2str(n);
Saving paths and middle part ofnumber in the file names if nb100;
S2=[‘0’,S2];if nb10
S2=[‘0’,S2];endendS=[S1,S2,S3];
Saving paths and complete file
names of untreated images.
I=imread(S); Import the untreated images
into MATLAB®.
C=double(I);for i=1:afor j=1:b
if C(i,j)b=ich=n+0.5;Rc=sqrt(R^2−(h−c0)^2);If sqrt((i−a0)^2−(j−b0)^2)NRc
C(i,j)=0;end
endend
T2=int2str(n);
The middle part of file names oftreated images. if nb100
T2=[‘0’,T2];if nb10T2=[‘0’,T2];
endendT=[T1,T2,T3];
Saving paths and complete file
names of treated images.
imwrite (C,T, ‘TIFF’);end Saving the treated images.
Appendix E. Program color reversion
A MATLAB® program for converting image colors:
Programming
Interpretation
clear all
Clear all previous parameters. a=101; Longitudinal pixel number
of untreated grey image.
b=101; Lateral pixel number of
untreated grey image.
c=61; Number of untreated grey images. d=10;S1=‘G:\Tutorial_Paper\Resized\Bileveled\Image_’;
Saving paths and former part offile names of untreated images
S3=‘.tif’; % the last part of the path
Saving paths and latter part offile names of untreated images
T1=‘G:\Tutorial_Paper\Resized\cut\Image_’;
Saving paths and former partof file names of treated images
T3=‘.tif’;
Saving paths and latter partof file names of treated images for n=0:(c+2*d−1)
if nbd;C=ones (a,b);
endf nN(c+d−1)
C=ones (a,b);end
if (nN=d && nb=(c+d−1))S2=int2str(n);
Saving paths and middle part of
number in the file names
if (n−d)b100;S2=[‘0’,S2];if (n−d)b10S2=[‘0’,S2];
endend
(continued)ppendix E (continued)
Programming
Interpretation
S=[S1,S2,S3];
Saving path and completefile names of untreated images.
I=imread(S);
Import the untreated greyimages into MATLAB®.
C=255-double(I);
Color convert treatment. T2=int2str(n); The middle part of file
names of images those converted.
if nb100;T2=[‘0’,T2];if nb10;
T2=[‘0’,T2];end
endT=[T1,T2,T3];
Saving paths and complete
file names of treated images.
imwrite (C,T, ‘TIFF’); Saving the treated images.
end
Appendix F. Steps for producing object plots
The following steps were executed to obtain Fig. 14a and b.
(1) Right click the “Image_000.tif” icon on the screen and selectLabelling\LabelVoxel to the “Image_000.tif” icon.
(2) Left click the red icon LableVoxel, select the subvoxel accuracyfor Option in the grey working area; left click DoIt to linka green bar named “Image_000 Lables” to the “Image_000.tif”bar.
(3) Right click the green “Image_000 Labels” bar and select Surfa-ceGen to attach the red icon named SurfaceGen to the green“Image_000 Labels” bar.
(4) Left click the red SurfaceGen icon, and select add border, adjustcords, then left click Triangulate to create the surface with agreen icon named “Image_000.surf.”
(5) Right click “Image_000.surf” and select SurfaceView will yieldthe image of the reconstructed object (Fig. 14a).
(6) In case a simplified version of the grids is needed, left click
“Image_000.surf” and then click to obtained Fig. 14b.
Appendix G. Steps for producing grid files
The following steps are listed to generate grid file from theproduced surface file in Appendix F.
(1) Select button in the “Amira Viewer” window, click Tests\
Intersection test for the intersection check.(2) Click Edit\Flip to see the triangles formed. Click Test\Aspect
ratio and Test\Tetra quality to check the aspect ratios of thegenerated triangles. If all tests are satisfied, we can save thedata file.
(3) To produce the three-dimensional object view, right click“Image_000.sur” and select Compute\TetraGen,Inside for theRegions Option, and Run now to produce grid file.
(4) Deselect SurfaceView, right click “Image_000.grid” and selectGridVolume, select lines for Draw Style and then click Add, onewill get the grid view for the sample images (Fig. 16).
(5) The file can be saved by left click “Image_000.grid,” then clickFile\Save Data.
Appendix H. Steps for producing mesh files
The following steps were executed in GAMBIT to obtained meshfiles.
(1) Start the GAMBIT program.(2) Open File/Import/Mesh, click Accept.
(3) Click Solver\FLUENT_5/6.(4) To eliminate some separated volumes.
(5) Left click and , Left click and select volume
except the “cell_zone.” In this case only one volume named“f_volume.1” should be picked up. Close the “Volume List(Multiple)” window.
(6) Left click Apply in the Delete Volumes window. This deletes thevolume.
(7) To add boundaries.
(8) Click and on the main window. Specify the “x1” in
the Name: box, and choose INTERFACE for the Type: option.
(9) Click bottom on the right of the box under Entity: and left
click the “f_face1386” in the extended menu, click to
add the selected face to the “Picked” box, and click Close toclose the extended menu. Click the Apply button to turn“f_face1386” into an INTERFACE. The name and type will beshown in the box under Name and Type, which indicated thesuccessful conversion of the boundary types.
(10) Similarly, we turn “f_face23,” “f_face.513,” “f_face.660,”“f_face.202,” “f_face.1836” into “x2,” “y1,” “y2,” “z1,” “z2,”respectively. Thus we have the six faces on the cubic boxturned into faces of type INTERFACE.
(11) Export the floc gird to mesh files. Click File\Export\Mesh.Construction of the surrounding mesh.
(12) Left click , and right click the second button onthe
first row under Volume, in the pull down list select .(13) In the Create Real Cylinder window, click Apply.(14) To divide the volumewith height of 3050 and both radii of 152,
click Apply.
(15) Click on the window, click Apply.
(16) Click under Geometry, under “Face”with both width
and height of 4000. Click Apply.(17) Change the Direction to ZX Centered in the Direction while keep
other parameters to make another face.
(18) Click under face.
(19) Click on the right to the Faces Pick box.
(20) In the extendedmenu, select “face.2197” and “face.2198” under
Available and click to add the two faces to the Picked
box, click Close. Select Move and deselect Copy; select Rotateunder Operation: click Define button next to the Axis. In theprompt window, select Positive next to “X” under Direction andclick Apply.
(21) Click Apply in the Move/Copy Faces window.
(22) Click and . In the Volume box, select “volume.4,” and
in the Face box, select “face.2197,” and click Apply.(23) In the Volume box, select “volume.6,” and in the Face box, select
“face.2198,” and click Apply.
(24) Click and . In the Faces box, select “face.2199,”
“face.2208,” “face.2212” and “face.2224,” and then click Apply.
(25) Click set the Direction to be YZ Centered. Click Apply.
(26) Click . Select “face.2227” in the Faces Pick Box, set−100 in
the box next to “x:” under Local, then click Apply and clickClose.
(27) Click set the Direction to be YZ Centered. Click Apply.
(28) Click . Select “face.2228” in the Faces Pick Box, set 100 in
the box next to “x:” under Local, then click Apply.(29) Click Close.
(30) Click and . Select “volume.12” in the Volume Box.
(31) Select “face.2227” in the Face Box. Click Apply.(32) Select “volume.23” in the “Volume” Box. Select “face.2228” in
the Face Box.(33) Click Apply.
(34) Click and .. Select “face 2228.”
(35) Select “face.2229” and “face.2253” in the Faces box. Click Apply.
(36) Click and . A window, namely Move/Copy Volumes.
(37) Pick up all volumes but the “cell_zone.” Select Connectedgeometry. Click Apply.
(38) Click and . A window, namely Create Straight Line, is
prompted.(39) Select “vertex.1341” and “f_vertex.1272” in Vertices option,
select Virtural and deselect Host, then press Apply.(40) In the same way, create straight edges between “f_vertex.531”
and “vertex.1342.”(41) In the sameway, create straight edges between “f_vertex.1274”
and “vertex.1339.”(42) In the same way, create straight edges between “f_vertex.532”
and “vertex.1340.”(43) Then Fig. 17 is revealed.
(44) Click and . A window, namely Create Face From
WireFrame,is prompted.
(45) In the Edges option blank, select “v_edge.3548,” “v_edge.3549,”“f_edge.1341” and “edge.3532,” click Apply.
(46) In the Edges option blank, select “v_edge.3549,” “v_edge.3550,”“f_edge.623” and “edge.3536,” click Apply.
(47) In the Edges List (Multiple) window, select “v_edge.3550,”“v_edge.3551,” “f_edge.1342” and “edge.3528.” Click Apply.
(48) In the Edges List (Multiple) box, select “v_edge.3548,”“v_edge.3551,” “f_edge.1340” and “edge.3531.” Click Apply.
(49) Click and A window, namely Stitch Faces, is
prompted.(50) In the Faces option box, select “v_face.2278,” “v_face.2280,”
“v_face.2282,” “v_face.2284,” “f_face.1386,” “face.2257,”“face.2260,” “face.2255,” “face.2259,” then click Apply.
(51) Volume “v_volume.42” is created. Click Apply.(52) “v_volume.43” is created.(53) Eliminating the extra volumes.
(55) Mesh the volume. Click and . A window, namely
Mesh Edges, is prompted.(56) Select “v_edge.3550,” “v_edge.3551,” “v_edge.3548,” “v_edge.3549,”
click Apply.(57) In the Edges List (Multiple) option box, select “edge.3541,”
“edge.3515,” “edge.3520,” “edge.3543,” “edge.3544,” “edge.3519,”“edge.3542” and “edge.3517,” set “interval size” to be 16, and clickApply.
(58) In the Edges box, select “edge.3445,” select First Length for theType, specify 16 for the Length, and set the Interval count to be20.
277S.S. Adav et al. / Biotechnology Advances 28 (2010) 255–280
(59) The same setting should be made to the following edges:“edge.3446,” “edge.3447,” “edge.3448,” “edge.3449,”“edge.3450,” “edge.3451,” “edge.3452,” “edge.3463,”“edge.3464,” “edge.3465,” “edge.3466,” “edge.3467,”“edge.3472,” “edge.3473,” “edge.3474,” “edge.3475,”“edge.3499,” “edge.3505,” “edge.3510,” “edge.3513,”“edge.3514,” “edge.3516,” “edge.3518,” “edge.3521,”“edge.3530,” “edge.3534,” “edge.3535,” “edge.3537,”“edge.3539,” “edge.3540,” “edge.3545,” “edge.3546,”“edge.3547.”
(60) Get the Fig. 18.
(61) Mesh the volume by left click under Mesh, and then click
to activate the “Mesh Volume” window. Click “Apply”
(66) Then one get Fig. 19.(67) Set the boundary, click Apply, the “inlet” name will appear in
the box under Name blank. Click Apply.(68) Export the mesh. Click File\Export\Mesh…, put down the mesh
file name. Click Accept.
Appendix I. Execution steps for CFD package Fluent™
The following steps were executed with Fluent™ ver. 6.0 forsolving the Eq. (12) with boundary conditions (Eqs. (13a), (13b), and(13c)):
(1) Open FluentTM, selecting 3D double precision (3ddp) andclicking ‘Run.’
(2) Open File\Read\Case. In the prompted window, selecting thetarget mesh file and clicking ‘OK’ to upload the mesh file intoFluent™.
(3) The size of floc in the mesh is not its true physical size. As aresult, a scaling action has been added. Open Grid\Scale…, anewwindowwill prompt as set the Scale Factors as 2.2084e−7for all directions. (Note the exact value depends on the truephysical size of a pixel in CLSM images).
(4) Click “Scale” at the bottom and closing the window. Now themesh is of the true size.
(5) Open Define\Models\Solver… Selecting the “Segregated” solv-er, “Implicit” formulation and “Steady” for time option.
(6) Open Define\Materials… Selecting the “Water-Liquid” from“Database…” Here, the defaulted constant values for waterdensity (998.2 kg m−3) and viscosity (1.003e−3 kg m−1s−1)are used.
(7) Open Define\Boundary Conditions. A new window promptedwith the names for each boundary in it. In the attachedmeshes,“xp,” “xn,” “yp,” “yn,” “zp,” “zn” represent the six surface of thecubic boxwhich contains the spherical floc; “xp1,” “xn1,” “yp1,”“yn1,” “zp1,” “zn1” represent the six surface of the closedhexahedral surface of the outsider domain (note the six surfaceof the cubic containing the spherical flocs geometricallycorrespond to the six surface of the hexahedral surface of theoutsider domain). As noted, types of these boundaries can be allregarded as “wall.” A change of the types of the twelve facets
should be done (from “wall” to “interface”) since fluid wasexpected to flow through these facets.
(8) Open Define\Grid-Interfaces. A new window will prompt.The twelve “interface”-type facets are visible in the win-dow. Geometrically corresponding facets should be ‘really’merged.
(9) Select targeting facets name under “Interface Zone 1,” andselecting its counterpart under the “Interface Zone 2.” Then,specifying a name under “Grid Interface” for the merge face,and clicking “Create.” Now the facets are merged and fluid canfreely flow through themerged facets. The same procedures areapplied to the other five pair facets.
(10) Open Define\Boundary Conditions… Specifying the in-comingflow velocity for both the inlet and sidewall (namely “tubewall”in the mesh file) of the cylinder and set the outlet of thecylinder as an “outflow” type (note the names of inlets in theattached meshes are “x_in,” “y_in,” “z_in” respectively for thex, y, z directions). It is necessary to pay attention on the outletof the cylinder as well.
(11) To set velocity of the tube wall, an option “MovingWall” can beselected in the “Momentum” blank, whereas the same valueequal to the in-coming flow velocity has been set in the “Speed”and consistent the in-coming flow direction with that in the“Direction blank.”
(12) Open Define\Boundary Conditions… Selecting the region of“fluid” (flow regions interior to and adjacent to the sphericalfloc) under “Zone” and clicking “Set.” In the prompt window,selecting water-liquid for the “Material Name.” The same workshould be done for the region of “fluid:0” (flow region exteriorto the floc).
(13) Open Solve\Controls\Solution… In the prompt window (asshown in Figure _), leaving the Under-Relaxation Factors as thedefaulted values and choose “Standard” for Pressure, “SIMPLE”for Pressure–Velocity Coupling, and “Second Order Upwind”method for the momentum discretization.
(14) Figure _ Parameters setting in window “Solution Controls”to give initial values for the computing to start, Open Solving\Initialize\Initialize.
(15) In the prompt window, choose the name of the inlet boundarycondition under “Compute From,” and click “Init.” This will setthe variable values according to those on the inlet face.
(16) Open Solving\Monitors\ Residuals. In the prompt window, setthe Convergence Criterion as 1e−4 (0.01%) for all residuals.This will favor the accuracy of results for the consideringproblem.
(17) Open Solving\Iterate to start computing.
References
Adav SS, Lee DJ. Single-culture aerobic granules with Acinetobacter calcoaceticus. ApplMicrobiol Biotechnol 2008;78:551–7.
Adav SS, Lee DJ, Lai JY. Effects of aeration intensity on formation of phenol-fed aerobicgranules and extracellular polymeric substances. Appl Microbiol Biotechnol2007a;77:175–82.
Adav SS, Lee DJ, Ren NQ. Biodegradation of pyridine using aerobic granules in thepresence of phenol. Water Res 2007b;41:2903–10.
Adav SS, Lee DJ, Tay JH. Extracellular polymeric substances and structural stability ofaerobic granule. Water Res 2007c;42:644–50 2008.
Adav SS, Lee DJ, Tay JH. Activity and structure of stored aerobic granules. EnvironTechnol 2007d;28:1227–35.
Adav SS, Chang CH, Lee DJ. Hydraulic characteristics of aerobic granules using sizeexclusion chromatography. Biotechnol Bioeng 2008a;99:791–9.
Adav SS, Lee DJ, Lai JY. Intergeneric coaggregation of strains isolated fromphenoldegrading aerobic granules. Appl Microbiol Biotechnol 2008b;79:657–61.
Adav SS, Lee DJ, Show KY, Tay JH. Aerobic granular sludge: recent advances. Biotech-nology Advances 2008c;26:411–23.
Adav SS, Lee DJ, Lai JY. Proteolytic activity in stored aerobic granular sludge andstructural integrity. Bioresour Technol 2009;100:68–73.
Allesen-Holm M, Barken KB, Yang L, Klausen M, Webb JS, Kjelleberg S, et al. Acharacterization of DNA release in Pseudomonas aeruginosa cultures and biofilms.Mol Microbiol 2006;59:1114–28.
Amann RI. Fluorescently labelled, rRNA-targeted oligonucleotide probes in the study ofmicrobial ecology. Mol Ecol 1995;4:543–54.
Amann RI, Binder BJ, Olson RJ, Chisholm SW, Devereux R, Stahl DA. Combination of 16SrRNA-targeted oligonucleotide probes with flow cytometry for analysing mixedmicrobial populations. Appl Environ Microbiol 1990;56:1919–25.
Amann R, Snaidr J, Wagner M, Ludwig W, Schleifer KH. In situ visualization of highgenetic diversity in a natural microbial community. J Bacteriol 1996;178:3496–500.
Amos WB, White JG. How the confocal laser scanning microscope entered biologicalresearch. Biol Cell 2003;95:335–42.
Ayol A, Filibeli A, Sir D, Kuzyaka E. Aerobic and anaerobic bioprocessing of activatedsludge: floc disintegration by enzymes. J Environ Sci Health A Tox Hazard SubstEnviron Eng 2008;43:1528–35.
Ballou B, Fisher GW, Deng JS, Hakala TR, SrivastavaM, Farkas DL. Cyanine fluorochrome-labeled antibodies in vivo: assessment of tumor imaging using Cy3, Cy5, Cy5.5, andCy7. Cancer Detect Prev 1998;22:251–7.
Berod A, Hartman BK, Pujol JF. Importance of fixation in immunohistochemistry. Use offormaldehyde solutions at variable pH for the localization of tyrosine hydroxylase. JHistochem Cytochem 1981;29:844–50.
Blenkinsopp SA, Costerton JW. Understanding bacterial biofilms. Trends Biotechnol1991;9:138–43.
Bockelmann U, ManzW, Neu TR, Szewzyk U. Investigation of lotic microbial aggregatesby a combined technique of fluorescent in situ hybridization and lectin-binding-analysis. J Microbial Methods 2002;49:75–87.
Boczar BA, BegleyWM, Larson RJ. Characterization of enzyme-activity in activated-sludgeusing rapid analyses for specific hydrolases. Water Environ Res 1992;64:792–7.
Boessmann M, Staudt C, Neu TR, Horn H, Hempel DC. Investigation and modeling ofgrowth, structure, and oxygen penetration in particle supported biofilms. ChemEng Technol 2003;26:219–22.
Boessmann M, Neu TR, Horn H, Hempel DC. Growth, structure and oxygen penetrationin particle supported autotrophic biofilms. Water Sci Technol 2004;49:371–7.
Brandtzaeg Brandtzaeg P. In: Bullock GR, Petrusz P, editors. Tissue preparation methodsfor immunohistochemistry. Techniques in immunocytochemistryLondon: Aca-demic Press; 1982. p. 1-75.
CammarotaMC, Sant'AnnaGL.Metabolic blocking of exopolysaccharides synthesis: effectson microbial adhesion and biofilm accumulation. Biotechnol Lett 1998;20:1–4.
Cerca N, Martins S, Sillankorva S, Jefferson KK, Pier GB, Oliveira R, et al. Effects of growthin the presence of subinhibitory concentrations of dichloroaxacillin on Staphylo-coccus epidermidis and Staphylococcus haemolyticus biofilms. Appl Environ Micro-biol 2005;71:8677–82.
Chen MJ, Zhang Z, Bott TR. Direct measurement of the adhesive strength of biofilms inpipes by micromanipulation. Biotechnol Tech 1998;12:875–80.
ChenMY, Lee DJ, Yang Z, PengXF. Fluorescent staining for study of extracellular polymericsubstances in membrane biofouling layers. Environ Sci Technol 2006;40:6642–6.
Chen MY, Lee DJ, Tay JH. Distribution of extracellular polymeric substances in aerobicgranules. Appl Microbiol Biotechnol 2007a;73:1463–9.
Chen MY, Lee DJ, Tay JH. Staining of extracellular polymeric substances and cells inbioaggregates. Appl Microbiol Biotechnol 2007b;75:467–74.
Chen H, Liu S, Yang F, Xue Y, Wang T. The development of simultaneous partialnitrification, ANAMMOX and denitrification (SNAD) process in a single reactor fornitrogen removal. Bioresour Technol 2009;100:1548–54.
Chiu ZC, Chen MY, Lee DJ, Tay STL, Tay JH, Show KY. Diffusivity of oxygen in aerobicgranules. Biotechnol Bioeng 2006;94:505–13.
Chiu ZC, Chen MY, Lee DJ, Wang CH, Lai JY. Oxygen diffusion in active layer of aerobicgranule with step change in surrounding oxygen levels. Water Res 2007a;41:884–92.
Chiu ZC, Chen MY, Lee DJ, Wang CH, Lai JY. Oxygen diffusion and consumption in activeaerobic granules of heterogeneous structure. Appl Microbiol Biotechnol 2007b;75:685–91.
Chu CP, Lee DJ. Multiscale structures of biological flocs. Chem Eng Sci 2004;59:1875–83.Chu CP, Lee DJ, Tay JH. Floc model and intrafloc flow. Chem Eng Sci 2005;60:565–75.Confer DR, Logan BE. Location of protein and polysaccharide hydrolytic activity in
suspended and biofilm wastewater cultures. Water Res 1998;32:31–8.Connil N, Dousset X, Onno B, Pilet MF, Breuil MF, Montel MC. Enumeration of Carno-
bacterium divergens V41, Carnobacterium piscicola V1 and Lactobacillus brevis LB62by in situ hybridisation-flow cytometry. Lett Appl Microbiol 1998;27:302–6.
Costerton JW, Irvin RT, Cheng KJ. The bacterial glycocalyx in nature and disease. AnnuRev Microbiol 1981;35:299–324.
Cottrell MT, Kirchman DL. Natural assemblages of marine Proteobacteria and numbersof the Cytophaga-Flavobacter cluster consuming low and high-molecular weightdissolved organic matter. Appl Environ Microbiol 2000;66:1692–7.
Dalton HM, March PE. Molecular genetics of bacterial attachment and biofouling. CurrOpin Biotechnol 1998;9:252–5.
Dar SA, Yao L, van Dongen U, Gijs Kuenen J, Muyzer G. Analysis of diversity andactivity of sulfate-reducing bacterial communities in sulfidogenic bioreactorsusing 16S rRNA and dsrB genes as molecular markers. Appl Environ Microbiol2007;73:594–604.
De Beer D, O'Flaharty V, Thaveesri J, Lens P, Verstraete W. Distribution of extracellularpolysaccharides and flotation of anaerobic sludge. Appl Microbiol Biotechnol1996;46:197–201.
DeLong EF, Wickham GS, Pace NR. Phylogenetic stains: ribosomal RNA-based probes forthe identification of single cells. Science 1989;243:1360–3.
Erlandsen SL, Kristich CJ, Dunny GM, Wells CL. High-resolution visualization of themicrobial glycocalyx with low-voltage scanning electron microscopy: dependenceon cationic dyes. J Histochem Cytochem 2004;52:1427–35.
Farber SA, Pack M, Ho SY, Johnson ID, Wagner DS, Dosch R, et al. Genetic analysis ofdigestive physiology using fluorescent phospholipid receptors. Science 2001;292:1385–8.
Fazi S, Amalfitano S, Pernthaler J, Puddu A. Bacterial communities associated with benthicorganicmatter inheadwater streammicrohabitats. EnvironMicrobiol 2005;7:1633–40.
Ferrari BC, Tujula N, Stoner K, Kjelleberg S. Catalyzed reporter deposition-fluorescencein situ hybridization allows for enrichment-independent detection of microcolony-forming soil bacteria. Appl Environ Microbiol 2006;72:918–22.
Fowler SD, Greenspan P. Application of Nile red, a fluorescent hydrophobic probe, forthe detection of neutral lipid deposits in tissue sections: comparison with oil red O.J Histochem Cytochem 1985;33:833–6.
Friedman L, Kolter R. Two genetic loci produce distinct carbohydrate-rich structuralcomponents of the Pseudomonas aeruginosa biofilm matrix. J Bacteriol 2004;186:4457–65.
Frolund B, Palmgren R, Keiding K, Nielsen P. Extraction of extracellular polymers fromactivated sludge using a cation exchange resin. Water Res 1996;30:1749–58.
Gerhardt P, Murray RGE, Wood WA, Krieg NR. Methods for general and molecularbacteriology. Washington, D.C: American Society for Microbiology; 1994.
Goel R, Mino T, Satoh H, Matsuo T. Comparison of hydrolytic enzyme systems in pureculture and activated sludge under different electron acceptor conditions. WaterSci Technol 1998;37:335–43.
Goldberg J. Biofilms and antibiotic resistance: a genetic linkage. Trends Microbiol2002;10:264.
Gonzalez JE, Maher MP. Ellular fluorescent indicators and voltage/ion probe reader(VIPR™): tools for ion channel and receptor drug discovery. Recept Channels2002;8:283–95.
Gonzalez JE, Tsien RY. Voltage sensing by fluorescence resonance energy transfer insingle cells. Biophys J 1995;69:1272–80.
Gonzalez JE, Tsien RY. Improved indicators of cell membrane potential that usefluorescence resonance energy transfer. Chem Biol 1997;4:269–77.
Griffiths G. Fine structure immunocytochemistry. New York: Springer-Verlag; 1993.Heydorn A, Nielsen AT, Hentzer M, Sternberg C, Givskov M, Ersboll BK, et al.
Quantification of biofilm structures by the novel computer program COMSTAT.Microbiol 2000a;146:2395–407.
Heydorn A, Ersboll BK, Hentzer M, Parsek MR, Givskov M, Molin S. Experimentalreproducibility in flowchamber grown biofilms. Microbiology 2000b;146: 2409–15.
Heydorn A, Ersboll BJ, Kato J, Hentzer M, Parsek MR, Tolker-Nielsen T, et al. Statisticalanalysis of Pseudomonas aeruginosa biofilm development: impact of mutations ingenes involved in twitching motility, cell-to-cell signalling, and stationary-phasesigma factor expression. Appl Environ Microbiol 2002;68:2008–17.
Hogardt M, Trebesius K, Geiger AM, Hornef M, Rosenecker J, Heesemann J. Specific andrapid detection by fluorescent in situ hybridization of bacteria in clinical samplesobtained from cystic fibrosis patients. J Clin Microbiol 2000;38:818–25.
Jackson KD, Starkey M, Kremer S, Parsek MR, Wozniak DJ. Identification of psl, a locusencoding a potential exopolysaccharide that is essential for Pseudomonasaeruginosa PAO1 biofilm formation. J Bacteriol 2004;186:4466–75.
Jang A, Yoon YH, Kim IS, Kim KS, Bishop PL. Character and evaluation of aerobic granulesin sequencing batch reactor. J Biotechnol 2003;105:71–82.
Jang A, Okabe S, Watanabe Y, Kim I, Bishop P. Measurement of growth rate of ammoniaoxidizing bacteria in partially submerged rotating biological contactor byfluorescent in situ hybridization (FISH). J Environ Sci Eng 2005;4:413–20.
Jiang HL, Tay JH, Maszenan AM, Tay STL. Bacterial diversity and function of aerobicgranules engineered in a sequencing batch reactor for phenol degradation. ApplEnviron Microbiol 2004a;70:6767–75.
Jiang HL, Tay JH, Tay ST. Changes in structure, activity and metabolism of aerobicgranules as a microbial response to high phenol loading. Appl Microbiol Biotechnol.2004b;63:602–8.
Kang HC, Fisher PJ, Prendergast FG, Haugland RP. Bodipy: a novel fluorescein and NBDsubstitute. J Cell Biol 1988;107:34a.
Kaufmann P, Pfefferkorn A, Teuber M, Meile L. Identification and quantification of Bifi-dobacterium species isolated from food with genusspecific 16S rRNA-targetedprobes by colony hybridization and PCR. Appl EnvironMicrobiol 1997;63:1268–73.
Kim HY, Yeon KM, Lee CH, Lee S, Swaminathan T. Biofilm structure and extracellularpolymeric substances in low and high dissolved oxygenmembrane bioreactors. SepSci Technol 2006;41:1213–30.
Kjelleberg S, Molin S. Is there a role for quorum sensing signals in bacterial biofilms?Curr Opin Microbiol 2002;5:254–8.
Kloeke FV, GeeseyGG. Localization and identification of populations of phosphatase-activebacterial cells associated with activated sludge flocs. Microb Ecol 1999;38:201–14.
Koch B, Worm J, Jensen LE, Hojberg O, Nybroe O. Carbon limitation induces σs-dependent gene expression in Pseudomonas fluorescens in soil. Appl EnvironMicrobiol 2001;67:3363–70.
Kolloffel B, Meile L, Teuber M. Analysis of brevibacteria on the surface of gruyere cheesedetected by in situ hybridisation and by colony hybridisation. Lett Appl Microbiol1999;29:317–22.
Kute T, Westphal U. Steroid-protein interactions. XXXIV. Chemical modification of α1-acid glycoprotein for characterization of the progesterone binding site. BiochimBiophys Acta 1976;420:195–213.
Lahtinen SJ, Gueimonde M, Ouwehand AC, Reinikainen JP, Salminen SJ. Probiotic bacteriamay become dormant during storage. Appl Environ Microbiol 2005;71:1662–3.
Larsson LI. Immunocytochemistry: theory and practice. Boca Raton: CRC Press; 1988.Larsson LI. Tissue preparation methods for light microscopic immunohistochemistry.
Appl Immunohistochem 1993;1:2-16.Laue H, Schenk A, Li H, Lambertsen L, Neu TR, Molin S, et al. Contribution of alginate and
levan production to biofilm formation by Pseudomonas syringae. Microbiology2006;152:2909–18.
Lawrence TR, Neu TR, Swerhone GDW. Application of multiple parameter imaging forthe quantification of algae, bacterial, and exopolymer components of microbialbiofilms. J Microbial Methods 1998;32:253–61.
279S.S. Adav et al. / Biotechnology Advances 28 (2010) 255–280
Lawrence JR, Swerhone GDW, Leppard GG, Araki T, Zhang X, West MM, et al. ScanningtransmissionX-ray, laser scanning, and transmissionelectronmicroscopymappingof theexopolymeric matrix of microbial biofilms. Appl Environ Microbiol 2003;69:5543–54.
Lawrence JR, Chenier MR, Roy R, Beaumier D, Fortin N, Swerhone GDW, et al. Microscaleand molecular assessment of impacts of nickel, nutrients, and oxygen level onstructure and function of river biofilm communities. Appl Environ Microbiol2004;70:4326–39.
Lawrence JR, Swerhone GDW,Wassenaar LI, Neu TR. Effects of selected pharmaceuticalson riverine biofilm communities. Can J Microbiol 2005;51:655–69.
Lawrence JR, Swerhone GD, Kuhlicke U, Neu TR. In situ evidence for microdomains inthe polymer matrix of bacterial microcolonies. Can J Microbiol 2007;53:450–8.
Lee N, Nielsen PH, Adersen KH, Juretschko S, Nielsen JL, Schleifer KH, et al. Combination offluorescent in situ hybridization andmicroautoradiography—a new tool for structure–function analyses in microbial ecology. Appl Environ Microbiol 1999a;65: 1289–97.
Lee SK, Lee JY, Lee MY, Chung SM, Chung JH. Advantages of calcium green-1 over otherfluorescent dyes in measuring cytosolic calcium in platelets. Anal Biochem1999b;273:186–91.
Lee CC, Lee DJ, Lai JY. Amylase activity in substrate deficiency aerobic granules. ApplMicrobiol Biotechnol 2009a;81:961–7.
Lee CC, Lee DJ, Lai JY. Labeling enzymes and extracellular polymeric substances inaerobic granules. J Taiwan Inst Chem Eng 2009b;40:505–10.
Lehtola MJ, Torvinen E, Miettinen IT, Keevil CW. Fluorescence in situ hybridizationusing peptide nucleic acid probes for rapid detection of Mycobacterium aviumsubsp. avium and Mycobacterium avium subsp. paratuberculosis in potable-waterbiofilms. Appl Environ Microbiol 2006;72:848–53.
Lemaire R, Webb RI, Yuan ZG. Micro-scale observations of the structure of aerobicgranules used for the treatment of nutrient-rich industrial wastewater. ISME J2008;2:528–41.
Li Y, Liu Y. Diffusion of substrate and oxygen in aerobic granule. Biochem Eng J2005;27:45–52.
Lippincott-Schwartz J, Patterson G. Development and use of fluorescent proteinmarkers in living cells. Science 2003;300:87–91.
Losekann T, Knittel K, Nadalig T, Fuchs B, Niemann H, Boetius A, et al. Diversity andabundance of aerobic and anaerobic methane oxidizers at the Haakon Mosby MudVolcano, Barents Sea. Appl Environ Microbiol 2007;73:3348–62.
Ma Y, Peng XF, Lee DJ. Stereological estimation of structural change in floc matrixfollowing weak ultrasonication. J Chin Inst Chem Engrs 2007;38:275–80.
Magnuson TS, Neal AL, Geesey GG. Combining in situ reverse transcriptase polymerasechain reaction, optical microscopy, and X ray photoelectron spectroscopy toinvestigate mineral surface-associated microbial activities. Microb Ecol 2004;48:578–88.
Mao SY. Conjugation of fluorochromes to antibodies. methods in molecular biology. In:Vaces LC, editor. Immunocytochemical Methods and Protocols; 2008. p. 43–7. 34.
Massol-Deya AA,Whallon J, Hickey RF, Tiedje JM. Channel structures in aerobic biofilmsof fixed-film reactors treating contaminated groundwater. Appl Environ Microbiol1995;61:769–77.
McSwain BS, Irvine RL, Hausner M, Wilderer PA. Composition and distribution ofextracellular polymeric substances in aerobic flocs and granular sludge. ApplEnviron Microbiol 2005;71:1051–7.
Michael T, Smith CM. Lectins probe molecular film in biofouling: characterization ofearly films on non-living and living surfaces. Mar Ecol Prog Ser 1995;119:229–36.
Miyawaki A, Sawano A, Kogure T. Lighting up cells: labeling proteins with fluorophores.Nat Cell Biol 2003;5:S1–7.
Moller S, Pedersen AR, Poulsen LK, Arvin E, Molin S. Activity and three-dimensionaldistribution of toluene-degrading Pseudomonas putida in a multispecies biofilmassessed by quantitative in situ hybridization and scanning confocal lasermicroscopy. Appl Environ Microbiol 1996;62:4632–40.
Mook OR, Van Overbeek C, Ackema EG, Van Maldegem F, Frederiks WM. In situlocalization of gelatinolytic activity in the extracellular matrix of metastages ofcolon cancer in rat liver using quenched fluorogenic DQ-gelatin. J HistochemCytochem 2003;51:821–9.
Moreno Y, Botella S, Alonso JL, Ferrus MA, Hernandez M, Hernandez J. Specific detectionof Arcobacter and Campylobacter strains in water and sewage by PCR andfluorescent in situ hybridization. Appl Environ Microbiol 2003;69:1181–6.
Mujumdar RB, Ernst LA, Mujumdar SR, Lewis CJ, Waggoner AS. Cyanine dye labelingreagents: sulfoindocyanine succinimidyl esters. Bioconjug Chem 1993;4:105–11.
Murray J, Marusich MF, Capaldi RA, Aggeler R. Focused proteomics: monoclonalantibody-based isolation of the oxidative phosphorylation machinery anddetection of phosphoproteins using a fluorescent phosphoprotein gel stain.Electrophoresis 2004;25:2520–5.
Neu TR. Significance of bacterial surface-active compounds in interaction of bacteriawith interfaces. Microbiol Rev 1996;60:151–66.
Neu TR. In situ cell and glycoconjugate distribution in river snow studied by confocallaser scanning microscopy. Aquat Microb Ecol 2000;21:85–95.
NeuTR, Lawrence JR.Developmentand structureofmicrobial biofilms in riverwater studiedby confocal laser scanning microscopy. FEMS Microbiol Immunol 1997;24:11–25.
Neu T, Swerhone GD, Lawrence JR. Assessment of lectin-binding analysis for in situdetection of glycoconjugates in biofilm systems. Microbiology 2001;147:299–313.
Neu TR, Kuhlicke U, Lawrence JR. Assessment of fluorochromes for two-photon laserscanning microscopy of biofilms. Appl Environ Microbiol 2002;68:901–9.
Neu TR, Woelfl S, Lawrence JR. Three-dimensional differentiation of photo-autotrophicbiofilm constituents by multi-channel laser scanning microscopy (single-photonand two-photon excitation). J Microbiol Methods 2004;56:161–72.
Nielsen PH, Jahn A. Extraction of EPS, p. 50–72. In: Wingender J, Neu TR, Flemming H-C,editors. Microbial extracellular polymeric substances: characterization, structure,and function. Berlin, Germany: Springer-Verlag; 1999. p. 21–47.
Nielsen PH, Raunkjaer K, Norsker NH, Jensen NA, Hvitved-Jacobsen T. Transformation ofwastewater in sewer systems — a review. Water Sci Technol 1992;25:17–31.
Nielsen PH, Frolund B, Keiding K. Changes in the composition of extracellular polymericsubstances in activated sludge during anaerobic storage. Appl Microbiol Biotechnol1996;44:823–30.
NielsenPH,RoslevP,DueholmTE,Nielsen JL.Microthrix parvicella, a specialized lipid consumerin anaerobic–aerobic activated sludge plants. Water Sci Technol 2002;46:73–80.
Nivens DE, Ohman DE, Williams J, Franklin MJ. Role of alginate and its O acetylation information of Pseudomonas aeruginosa microcolonies and biofilms. J Bacteriol2001;183:1047–57.
Nosyk O, Haseborg ET, Metzger U, Frimmel FH. A standardized pre-treatment method ofbiofilm flocs for fluorescence microscopic characterization. J Microbiol Methods2008;75:449–56.
Nybroe O, Jorgensen PE, Henze M. Enzyme-activities in waste-water and activated-sludge. Water Res 1992;26:579–84.
Okabe S, Satoh H,Watanabe Y. In situ analysis of nitrifying biofilms as determined by in situhybridizationand theuseofmicroelectrodes. Appl EnvironMicrobiol 1999;65:3182–91.
Ouverney CC, Fuhrman JA. Combined microautoradiography-16S rRNA probe tech-nique for determination of radioisotope uptake by specific microbial cell types insitu. Appl Environ Microbiol 1999;65:1746–52.
Panchuk-Voloshina N, Haugland RP, Bishop-Stewart J, Bhalgat MK, Millard PJ, Mao F, et al.Alexa dyes, a series of new fluorescent dyes that yield exceptionally bright,photostable conjugates. J Histochem Cytochem 1999;47:1179–88.
Peng JS, Tsai WC, Chou CC. Inactivation and removal of Bacillus cereus by sanitizer anddetergent. Int J Food Microbiol 2002;77:11–8.
Pernthaler A, Pernthaler J, Amann R. Fluorescence in situ hybridization and catalyzedreporter deposition for the identification of marine bacteria. Appl EnvironMicrobiol 2002;68:3094–101.
Picioreanu C, van Loosdrecht MCM, Heijnen JJ. Mathematical modeling of biofilmstructure with a hybrid differential-discreet cellular automata approach. Biotech-nol Bioeng 1998;58:101–16.
Poppert S, Essig A, Stoehr B, Steingruber A, Wirths B, Juretschko S, et al. Rapid diagnosisof bacterial meningitis by real-time PCR and fluorescence in situ hybridization.J Clin Microbiol 2005;43:3390–7.
Puchtler H, Meloan SN. On the chemistry of formaldehyde fixation and its effects onimmunohistochemical reactions. Histochemistry 1985;82:201–4.
Recktenwald M, Wawrzynczyk J, Dey ES, Norrlow O. Enhanced efficiency of industrial-scale anaerobic digestion by the addition of glycosidic enzymes. J Environ SciHealth 2008;43:1536–40.
Ren TT, Liu L, Sheng GP, Liu XW, Yu HQ, Zhang MC, et al. Calcium spatial distribution inaerobic granules and its effects on granule structure, strength and bioactivity.Water Res 2008;42:3343–52.
Ross WR. The phenomenon of sludge pelletizationin in the anaerobic treatment of amaize processing waste. Water SA 1984;10:197–204.
Sarkisova S, Patrauchan MA, Berglund D, Nivens DE, Franklin MJ. Calcium-inducedvirulence factors associated with the extracellular matrix of mucoid Pseudomonasaeruginosa biofilms. J Bacteriol 2005;187:4327–37.
Sato K, Anzai J. Fluorometric determination of sugars using fluorescein-labeledconcanavalin A–glycogen conjugates. Anal Bioanal Chem 2006;384:1297–301.
Schmid M, Twachtmann U, Klein M, Strous M, Juretschko S, Jetten M, et al. Molecularevidence for genus level diversity of bacteria capable of catalyzing anaerobicammonium oxidation. Syst Appl Microbiol 2000;23:93-106.
Schmid M, Thill A, Purkhold U, Walcher M, Bottero JY, Ginestet P, et al. Characterizationof activated sludge flocs by confocal laser scanning microscopy and image analysis.Water Res 2003;37:2043–52.
Schmidt JE, Ahring BK. Extracellular polymers in granular sludge from different upflowanaerobic sludgeblanket (UASB) reactors. ApplMicrobiol Biotechnol 1994;42:457–62.
Schramm A, Larsen LH, Revsbech NP, Amann RI, Schleifer KH. Structure and function ofa nitrifying biofilm as determined by in situ hybridization and the use ofmicroelectrodes. Appl Environ Microbiol 1996;62:641–7.
Schramm A, de Beer D, Wagner M, Amann RI. Identification and activities in situ ofNitrosospira and Nitrospira spp. as dominant populations in a nitrifying fluidizedbed reactor. Appl Environ Microbiol 1998;64:3480–5.
Sekar R, Griebe T, Flemming HC. Influence of image acquisition parameters onquantitative measurement of biofilms using confocal laser scanning microscopy.Biofouling 2002;18:47–56.
Seveus L, Vaisala M, Hammlia I, Kojola H, Roomans GM, Soini E. Use of fluorescentEuropium chelates as labels in microscopy allows glutaraldehyde fixation andpermanent mounting and leads to reduced autofluorescence and good long-termstability. Microsc Res Techn 1994;28:149–54.
Shen CF, Kosaric N, Blaszczyk R. The effect of selected heavy metals (Ni, Co and Fe) onanaerobic granules and their extracellular polymeric substance (EPS). Water Res1993;27:25–33.
Sohier D, Lonvaud-Funel A. Rapid and sensitive in situ hybridizationmethod for detectingand identifying lactic acid bacteria in wine. Food Microbiol 1998;15:391–7.
Staudt C, Horn H, Hempel DC, Neu TR. Volumeric measurements of bacterial cells andextracellular polymeric substance glycoconjugates in biofilms. Biotechnol Bioeng2004;88:585–92.
Strathmann M, Wingender J, Flemming HC. Application of fluorescently labelled lectinsfor the visualization and biochemical characterization of polysaccharides inbiofilms of Pseudomonas aeruginosa. J. Microbiol. Methods 2002;50:237–48.
Strathmann M, Wingender J, Flemming HC. Application of fluorescently labeled lectinsfor the visualization and biochemical characteristics of polysaccharides in biofilmof Pseudomonas aeruginosa. J Microbiol Methods 2003;50:237–48.
Takeda M, Nakano F, Nagase T, Iohara K, Koizumi JI. Isolation and chemical compositionof the sheath of Sphaerotilus natans. Biosci Biotechnol Biochem 1998;62:1138–43.
Tay JH, Yang SF, Liu Y. Hydraulic selection pressure-induced nitrifying granulation insequencing batch reactors. Appl Microbiol Biotechnol 2002;59:332–7.
Teuber M, Brodisch KEU. Enzymatic-activities of activated-sludge. Eur J Appl Microbiol1977;4:185–94.
Thurnheer T, Gmur R, Guggenheim B. Multiplex FISH analysis of a six-species bacterialbiofilm. J Microbiol Methods 2004;56:37–47.
Tsai MW, Lee DJ, Lai JY. Mass transfer limit of fluorescent dyes during multicolorstaining of aerobic granules. Appl Microbiol Biotechnol 2008;78:907–13.
Tsien RY. Fluorescent indicators of ion concentrations. Methods Cell Biol 1989;30:127–56.Tsien RY. Building and breeding molecules to spy on cells and tumors. FEBS Letts
2005;579:927–32.Vereb G, Jares-Erijman E, Selvin PR, Jovin TM. Temporally and spectrally resolved
imaging microscopy of lanthanide chelates. Biophys J 1998;74:2210–22.Vesey G, Ashbolt N, Fricker EJ, Deere D, Williams KL, Veal DA, et al. The use of ribosomal
Wallner G, Amann RI, Beisker W. Optimising fluorescent in-situ hybridization withrRNA-targeted oligonucleotide probes for flow cytometric identification ofmicroorganisms. Cytometry 1993;14:136–43.
Wallner G, Erhart R, Amann R. Flow cytometric analysis of activated sludge with rRNA-targeted probes. Appl Environ Microbiol 1995;61:1859–66.
Wang ZW, Liu Y, Tay JH. Distribution of EPS and cell surface hydrophobicity in aerobicgranules. Appl Microbiol Biotechnol 2005;69:469–73.
Wanner O, Gujer P. A multispecies biofilm model. Biotechnol Bioeng 1985;47:172–84.Wessendorf MW, Brelje TC. Which fluorophore is brightest? A comparison of staining
Whiteley CG, Lee DJ. Enzymology of sludge-floc degradation using confocal laserscanning microscopy: an enzyme-floc model. Focus on Biotechnology Research,Chap. 7, pp. 213–133. EC HEarns Ed. Nova Pub, USA. ISBN: 1-59454-863-3. (2006).
Wories HJ, Koek JH, Lodder G, Lugtenburg J, Fokkens R, Driessen O, et al. A novel watersoluble fluorescent probe: synthesis, luminescence and biological properties of thesodium salt of the 4-sulfonato-3, 3′5, 5′-tetramethyl-2, 2′-pyrromethene-1, 1′-BF2complex. Recl Trav Chim Pays-Bas. 1985;104:280–91.
Wu X, Liu H, Liu J, Haley KN, Treadway JA, Larson JP, et al. Immonofluorescent labelingof cancer marker Her2 and other cellular targets with semiconductor quantumdots. Nat Biotechnol 2003;21:41–6.
Xia Y, Kong Y, Nielsen PH. In situ detection of protein-hydrolysing microorganisms inactivated sludge. FEMS Microbiol Ecol 2007;60:156–65.
Xia Y, Kong Y, Nielsen PH. In situ detection of starch-hydrolyzing microorganisms inactivated sludge. FEMS Microbiol Ecol 2008;66:462–71.
Xu KD, McFeters GA, Stewart PS. Biofilm resistance to antimicrobial agents.Microbiology 2000;146:547–9.
Yang Z, Peng XF, Lee DJ, Su A. Advective flow in spherical floc. J Colloid Interf Sci2007a;308:451–9.
Yang Z, Peng XF, ChenMY, Lee DJ, Lai JY. Intra-layer flow in fouling layer onmembranes.J Membr Sci 2007b;297:280–6.
Yang Z, Peng XF, Lee DJ, Su A. Reynolds-number-dependent permeability of wastewatersludge flocs. J Chin Inst Chem Engrs 2007c;38:135–41.
Yang Z, Lee DJ, Mujumdar AS, Peng XF, Su A, Hsu C. Probing heterogeneous structure ofaggregates. Drying Technol 2008;26:1018–23.
Yip KP, Kurtz I. Confocal fluorescence microscopy measurements of pH and calcium inliving cells. Methods Cell Biol 2002;70:417–27.
Yoshioka T, Hikita I, Matsutani T, Yoshida R, AsakawaM, Toyosaki-Maeda T, et al. DS-Nhas an experimental model of atopic dermatitis induced by Staphylococcus aureusproducing staphylococcal enterotoxin C. Immunology 2003;108:562–9.
Zhang J, Campbell RE, Ting AY, Tsien RY. Creating new fluorescent probes for cellbiology. Nat Rev Mol Cell Biol 2002;3:906–18.
Zhang ZP,AdavSS, ShowKY, Tay JH, LiangDT, LeeDJ, et al. Characteristics of rapidly formedhydrogen-producing granules and biofilms. Biotechnol Bioeng 2008;101:926–36.
Zimmer M. Green fluorescent protein: applications, structure, and related photo-physical behavior. Chem Rev 2002;102:759–81.
