Evaluation of the microbial adhesion to hydrocarbon assay applied to a hydrocarbon degrading
bacterium
by Lawrence Otto Schmidt
A thesis submitted in partial fulfillment of the requirements for the degree of Master of Science in
Environmental Engineering
Montana State University
© Copyright by Lawrence Otto Schmidt (1995)
Abstract:
The outer surface of microbial cells contain a variety of chemical compounds which may be involved
in the attachment of cells to surfaces. Hydrophobic/hydrophilic interactions play a large role in
attachment, leading to the development of the concept of cell surface hydrophobic!ty as a measure of
the tendency of a cell to attach to a surface. One of the most popular tests of this tendency is the
Microbial Adhesion to Hydrocarbons (MATH) test. In a typical assay, a cell suspension is contacted
with a small volume of hydrocarbon, and the removal of cells by the hydrocarbon is determined. The
goal of this project was to evaluate the effects of test conditions and bacterial growth conditions on the
results and precision of the MATH test using a single species of hydrocarbon degrading bacteria. Cells
were grown under either carbon limited (C:N = 5:1) or nitrogen limited (C:N = 15:1) conditions, and
MATH contact assays were performed at various agitation intensities, with various hydrocarbon
volumes, and for various mixing times. In addition, the results of the assay were evaluated using light
absorbance and using viable plate counts to determine cell removal due to the hydrocarbon.
Appropriate negative controls were also run in the absence of hydrocarbon to account for wall effects
in the test vessels.
The results of the study were analyzed according to traditional methods, using percent removal and rate
of removal, and compared to the literature. Additional statistical methods were used to evaluate the
variance in results and to determine how much of the variance could be attributed to experimental test
factor values versus how much was due to inherent variability in the assay. In general, reproducibility
was found to be poor, owing perhaps to the strong tendency of the test organism to form clumps in
aqueous solution. Multiple linear regression models and analysis of variance (ANOVA) produced
p-values for the factors that often showed statistical significance, but the actual effects of the test
conditions were found to be negligible from a practical standpoint. Only the agitation rate was found to
be significant according to all test methods: at low agitation rates, significantly lower removals from
aqueous phase were observed than at high rates. - 'I
"I
EVALUATION OF THE MICROBIAL ADHESION TO HYDROCARBON ASSAY
APPLIED TO A HYDROCARBON DEGRADING BACTERIUM
by
Lawrence Otto Schmidt
A thesis submitted in partial fulfillment
of the requirements for the degree
of
Master of Science
in
Environmental Engineering
MONTANA STATE UNIVERSITY
Bozeman, Montana
April 1995
© COPYRIGHT
by
Lawrence Otto Schmidt
1995
All Rights Reserved
U%ni
& A .5 4 7
APPROVAL
of a thesis submitted by
Lawrence Otto Schmidt
This thesis has been read by each member of the thesis committee and has been
found to be satisfactory regarding content, English usage, format, citations, bibliographic
style and consistency, and is ready for submission to the College of Graduate Studies.
A^;/Zi m s
Datey
Chairperson, Graduate Committee
Approved for the Major Department
Z /
Date
Head, Major Department
Approved for the College of Graduate Studies
Date
Graduate Dean
"/
Ul
STATEMENT OF PERMISSION TO USE .
In presenting this thesis in partial fulfillment of the requirements for a master’s
degree at Montana State University, I agree that the Library shall make it available to
borrowers under rules of the Library.
If I have indicated my intention to copyright this thesis by including a Copyright
notice page, copying is allowable only for scholarly purposes, consistent with “fair use” as
prescribed in the U.S. Copyright Law. Requests for permission for extended quotation
from or reproduction of this thesis in whole or in parts may be granted only by the
copyright holder.
Date
iv
TABLE OF CONTENTS
PAGE
LIST O F TA B LES ...........................................................................................
vii
LIST O F FIG U R ES .........................................................................................
viii
A B ST R A C T
ix
.........................................................................................................
IN T R O D U C T IO N ..............................................................................................
Surface Characteristics of Cells ................................... .........................
M A TH te st ...............................................................................................
Consistency of the MATH test ...............................................
Physical Factors .........................................................................
Physiological Factors ...............................................................
Importance of the MATH test . .................................................
Goal of R esearch .................................................................................
H ypotheses ..............................................................................................
I
I
3
3
4
5
5
6
6
B A C K G R O U N D .................................................................................................
Cell Surface H ydrophobicity .......... ........... .......................................
Previous R esearch .............................................................................. .
Measurement of Cell Surface H ydrophobicity .....................................
H ydrophobicity A ssays ..................... ....... .........................................
BATH and MATH Assays ......................... ...........................
Contact A ngle M easurem ent ..................................................
Hydrophobic Interaction Chromatography .............................
Salt A ggregation T est ..............................................................
Tw o Phase P artition ............................... ................................
Binding of Molecular Probes ........ ...........................................
A dhesion to H ydrophobic Surfaces .......................................
D irection o f Spreading ............................................................
Im age A nalysis .........................................................................
Physiological Effects on Cell Surface H ydrophobicity....................... '
G row th Phase .............................................................................
G row th T em perature ................................................................
Cell G eneration .........................................................................
Cell Type and Preparation ......................................................
Physical Factors Effecting the MATH test ........................................
C ontinued R esearch ..............................................................................
Im portance of Research ...........................................................
C om parable T ests .................................................;...................
8
8
11
12
13
13
14
14
15
15
16
16
16
17
18
18
19
19
19
20
21
21
22
V
TABLE OF CONTENTS (continued)
M E T H O D S ..........................................................
P relim inary W ork ..................................................................................
Iso la tio n .......................................................................................
S u b stra te ......................................................................................
C:N ratio calculations .................................................
N utrient M edium ......................................................................
Experim ental Design ...............................................................................
M A TH A ssay ............................................................................
C:N R atio .............
Experim ental Conditions ...........................................
A gitation R ate ............................
H ydrocarbon V olum e ..............................................................
H ydrocarbon Screening .........................................................
Tim e of Mixing ...........................................................................
E xperim ental Procedure .....................................................................
In o cu latio n ................................................................................
W ashing Procedure ...............................................................
R esuspension o f Cells ...........................................................
A ddition to T est Tubes .........................................................
A g ita tio n .......................................
Cell Enum eration Techniques ...............................................
Cell E num eration ...................................................................
A bsorbance M easurem ents.....................................................
24
24
24
25
26
26
27
27
27
28
28
29
29
30
30
30
31
32
32
33
33
34
34
R E SU L T S .............................
T est T ube O bservations ...........................
P ercent Rem oval D ata ..........
Results from Statistical A n aly sis...........................................................
R egression A nalysis ...............................................................
36
36
42
48
48
D IS C U S S IO N ....................................................................................................
A nalysis o f D ata .............
P ercent Rem oval ......................
K inetic A pproach ...................................................................
G raphical A nalysis ................................................................
S tatistical A nalysis ..............................................................
R egression ................................................................................
D ata Inclusion ...........................................................
S ignificant Factors ....................................................
p V alues ..........................................
Significant Coefficients ...............
V ariation Between Experim ents ..........................................
Means and Standard E rror ...................................
50
50
50
51
57
59
59
60
61
61
63
64
65
vi
TABLE OF CONTENTS (continued)
C O N C L U S IO N
.................................................................................................
70
R E FE R E N C E S
.................................................................................................
71
A P P E N D IC E S ..................................................................................................
A ppendix A ..............................................................................................
Experim ental Results ............................................................
Appendix B ............................................................................................
Im age A nalysis Research .......................................................
j
75
76
77
95
96
LIST OF TABLES
Table
1.
2.
3
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
C om position o f nutrient m edia ............................................................
Settings for Agitaiton rate, hydrocarbon volume and mixing tim e .........
Absorbance and Percent Removal for Three H ydrocarbons.................
Composition of buffer solution ..................................................................
Exam ple of a Typical Experim ent .......................................................
Absorbance Responses for Cells Cultured at a C:N ratio of 5 :1 ............
Cell Count Responses for Cells Cultured at a C:N ratio of 1 5 :1 ...........
Absorbance Responses for Cells Cultured at a C:N ratio of 5 :1 ............
Cell Count Responses for Cells Cultured at a O N ratio of 1 5 :1 ...........
Absorbance and Cell Count Data for Cells Cultured at a C:N
ratio of 5:1 for Experiments with Interference........................ .........
Constant Terms and Coefficients for Responses ...............................
Percent absorbance data showing the change from the lowest
settings to the highest settings for absorbance and cell counts.............
T he p values for six responses .........................................................
Constant terms and coefficients for responses ................. ..................
Values for Interexperimental and Intraexperimental Error....................
Page
27
28
29
32
33
43
44
45
46
47
49
51
62
64
65
viii
LIST OF HGURES
Figure
Page
1. Cells Within the Circle are at the Hydrocarbon-water Interface....................
2. Test tube containing unmixed hydrocarbon and cell suspension
3. Control test tube after agitation. (Agitation at a setting of 60 for 40 sec.)......
4. T est tube agitated w ith carbon .................................................................
5. Control test tube. (Agitated at a setting of 60 for 20 s e c .) ........................
6. Control test tube. (Agitated at a setting of 140 for 20 sec.) ..:.................
7. Test tube containing .2 ml of hydrocarbon. (Agitated at a setting of 60 for
20 seconds) ...................................................................................................
8. Test tube containing .2 ml of hydrocarbon. (Agitated at a setting of 140
fo r 20 seco n d s)...............................................................................................
9. Test tube containing I ml of hydrocarbon. (Agitated at a setting of 60 for
20 seconds) ...................................................................................................
10. Plots used to determine the removal rate (k) and removal coefficient (K)
for cells grown in a culture containing a 5:1 C:N ratio.............................
11. Plots used to determine the removal rate (k) and removal coefficient (K)
for cells raised with a 15:1 C:N ratio .....................................................
12. Effects of agitation setting, hydrocarbon volume and time of mixing on
the percent removal (ideal) for isolate growth at a 15:1 C:N ratio for
a b so rb an ce ......................................................................................................
13. Percent removal (ideal) at two mixing times with variation of agitation
rate and hydrocarbon volume ...................................................................
14. Initial cell counts for cultures raised at 5:1 and 15:1 C:N ratio including
the cell counts for interference d a t a .............................................................
15. Responses In (Ao/A) and In(CoZC) vs hydrocarbon volume including
the standard error o f the averages ............................................................
16. Responses In(AoZA) and In(CoZC) vs agitation setting including the
standard error of the averages ...................................................................
17. Responses in (AoZA) and In(CoZC) vs time of mixing including the
standard error of the averages ...................................................................
37
38
38
39
39
40
40
41
41
52
54
57
59
61
67
68
69
IX
ABSTRACT
The outer surface of microbial cells contain a variety of chemical compounds which
may be involved in the attachment of cells to surfaces. Hydrophobic/hydrophilic
interactions play a large role in attachment, leading to the development of the concept of cell
surface hydrophobicity as a measure of the tendency of a cell to attach to a surface. One of
the most popular tests of this tendency is the Microbial Adhesion to Hydrocarbons
(MATH) test. In a typical assay, a cell suspension is contacted with a small volume of
hydrocarbon, and the removal of cells by the hydrocarbon is determined. The goal of this
project was to evaluate the effects of test conditions and bacterial growth conditions on the
results and precision of the MATH test using a single species of hydrocarbon degrading
bacteria. Cells were grown under either carbon limited (C:N = 5 :1) or nitrogen limited
(C:N = 15:1) conditions, and MATH contact assays were performed at various agitation
intensities, with various hydrocarbon volumes, and for various mixing times. In addition,
the results of the assay were evaluated using light absorbance and using viable plate counts
to determine cell removal due to the hydrocarbon. Appropriate negative controls were also
run in the absence of hydrocarbon to account for wall effects in the test vessels.
The results of the study were analyzed according to traditional methods, using
percent removal and rate of removal, and compared to the literature. Additional statistical
methods were used to evaluate the variance in results and to determine how much of the
variance could be attributed to experimental test factor values versus how much was due to
inherent variability in the assay. In general, reproducibility was found to be poor, owing
perhaps to the strong tendency of the test organism to form clumps in aqueous solution.
Multiple linear regression models and analysis of variance (ANOVA) produced p-values for
the factors that often showed statistical significance, but the actual effects of the test
conditions were found to be negligible from a practical standpoint. Only the agitation rate
was found to be significant according to all test methods: at low agitation rates,
significantly lower removals from aqueous phase were observed than at high rates.
I
INTRODUCTION
Measurements of cell surface hydrophobicity are made by a variety of methods. The
MATH (microbial adhesion to hydrocarbon) test is potentially the most useful of the
techniques. The MATH test is simple, fast and reasonably reproducible. The main
drawback with the method is its lack of quantification. The MATH test needs to be tested
for reliability and further the factors which influence the test the most should be
determined.
Surface Characteristics of Cells
The biochemical composition of the outer surfaces of microorganisms strongly
affects the nature and extent of their interaction with and attachment to substrata.
Once a cell contacts a surface, a variety of processes maybe involved in microbial
attachment (Kjellberg 1984).
Structure-mediated processes can be initiated by extracellular polymers which
project from the cell surface. These extracellular polymers attach to surfaces by two
mechanisms: specific and nonspecific binding. Specific binding requires a complementary
site on the binding surface to the extracellular polymer of the bacteria. Nonspecific binding
involves macromolecules or extracellular polymers that interact with the surface of the
substratum or macromolecules present on the surface (Group Report 1984). A number of
possible nonspecific binding processes can occur including: ionic, dipolar, hydrogen
bonding, and hydrophobic/hydrophilic surface interactions.
Cell surface hydrophobicity describes cells with hydrophobic surface
characteristics. The hydrophobic nature of these cells promotes attachment to substrates
which may provide access to essential nutrients. Attachment of bacteria at the hydrocarbon-
2
water interface is a good example of a hydrophobic interaction. These bacteria are able to
colonize surfaces that are unavailable to other cells, thus providing a competitive advantage
in nutrient availability. The hydrocarbon-water interface is a unique environment which
may offer protection for the bacteria from predators (Marshall 1976). These bacteria need to
be able to detach from oil once all the long-chain n-alkanes they grow on have been
utilized. Acinobactercalcoaceticus RAG -1 releases encapsulated emulsans which cause the
bacteria to desorb from the hydrocarbon. The emulsan forms a hydrophilic polymeric film
around the depleted oil droplet that hydrophobic bacteria cannot attach to (Rosenberg and
Kaplan 1987).
Hydrophobic organisms have been found in a variety of environments. In the
medical field, studies on adhesion of bacteria to host tissues indicate that hydrophobicity is
an important factor for the infection process. (Enzer and Douglas 1992). Another example
in medicine is the persistence of bacteria which are able to adhere to plastics used in medical
devices (Rosenberg and Doyle 1992). In the petroleum industry methods are needed for
quick and cost effective removal of insoluble hydrocarbons. Bacteria that are able to adhere
to the hydrocarbon-water interface may play an integral part in development of these
methods (Goswami and Singh 1990). Chemical surfactants are used extensively in
industry, agriculture and medicine. Surface-active compounds that aid in emulsification of
the oil phase into the aqueous phase increase the interfacial area available for microbial
contact (Atlas 1984,672); Advantages of natural surfactants over manufactured surfactants
are their ease of biodegradation and their selectivity for a specific surface. Emulsan, for
instance, is specific for a mixture of aliphatic and aromatic hydrocarbons. The use of
Acinobaetercalcoaceticus RAG-1 to produce emulsan and other extracellular
polysaccharides by fermentation processes is another area of research incorporating bacteria
with hydrophobic surfaces (Shabtai and Wang 1989).
3
MATHTest
Many researchers are working with bacteria that have hydrophobic character and
would like to be able to compare the cell surface hydrophobicity of their organism to
others. A quantitative measure of hydrophobicity should be able to predict the tendency of
bacteria to attach to a hydrophobic surface. The MATH (microbial adhesion to
hydrocarbon) test has been used extensively to screen bacteria for hydrophobic character by
measuring the tendency of organisms to attach to a hydrophobic surface. This test was first
developed in 1980 by Rosenberg as a possible quick and easy method to determine the
hydrophobic character of cells. The test has developed to the point where some researchers
believe that it is a quantitative measure of cell surface hydrophobicity (Lichtenberg et al.
1985).
In a typical experiment, test tubes or cuvettes containing a cell suspension and
hydrocarbon are agitated at a fixed rate for a given time period. The agitation of the solution
allows the microorganisms to come into contact with the hydrocarbon. The phases are then
allowed to separate, thfe lower aqueous phase is separated from the upper hydrocarbon
phase and absorbance measurements are made on the aqueous phase. Removal of cells
from the aqueous phase (as measured by change in absorbance) is then used as a measure
of hydrophobicity.
Consistency of the MATH Test
In order for the MATH test to be useful, results and comparisons throughout a
study must not be affected significantly by the results of error inherent in the MATH test.
Correlations between the MATH test and other methods for determining hydrophobicity
have been observed. A linear correlation was observed for Pasteurellamultiocida when
4
percent adhesion to hydrocarbon was compared to percent retained on octyl-Sepharose
(Darnell et al. 1987). The result of another method comparison was a direct correlation
between cell-surface hydrophobic!ty and adhesion to epithelial cells (Ener and Douglas
1986). Epithelial cells make up the epithelium, the cellular tissue covering surfaces,
forming glands and lining most of the cavities in the body. Evaluation of the
hydrophobicity of the following organisms, Enterobacteraerogenes, Klebsiella oxytoca,
Saccharomyces carsbergensis and Kluyveromyces fragilis varied widely depending on the
measurement technique (Moses and Rouxhet 1987). The MATH test proved to be the most
inconsistent of those techniques.
The MATH test does not always yield internally consistent results. For example.
Sweet et al. (1987) state, “current methods of assessing bacterial hydrophobicity as a
function of adherence to liquid hydrocarbons (especially hexadecane) do not always yield
reproducible results.” Sweet etal. determined in their study that p-xylene should be used
instead of hexadecane. Factors which affect the internal consistency of the MATH test are
both physical and physiological.
Physical Factors
Physical factors which affect the MATH test include agitation rate, hydrocarbon
volume, time of mixing and test tube size. The hydrocarbon volume will affect the surface
area available for attachment, as will the agitation rate. More vigorous agitation may also
increase the transport of cells to the interface. Time may also affect the results since the
longer the cell suspension is in motion the more likely it is that a microbe will come into
contact with the hydrocarbon. If the MATH test is to be internally consistent, the effects of
these factors on response should be minimized.
5
Physiological Factors
Physiological factors affect the growth of cells, subsequently causing changes in
the chemistry of the cell surface. The carbon to nitrogen ratio is one such factor. A C:N
ratio of 7.7:1 has been determined to support balanced growth and product synthesis
(emulsan production) fox Acinetobactercalcoaceticus RAG-1 (Shabtai and Wang 1989).
Carbon limited media will typically limit the ability of organisms to produce extracellular,
materials particularly exopolysaccharide. Conversely, a high C:N ratio should produce
extracellular materials which could lower the hydrophobic!ty of the cells. Other nutrient
conditions that affect the MATH test are phosphate limitation, inorganic ions and dissolved
organic matter. Culture conditions such as temperature, pH, and ionic strength can also
affect the results of the MATH test. Last, generation variation between cultures may affect
the surface properties and thus the cell surface hydrophobicity of cells.
Importance of the MATH test
For the MATH test to be useful, results and comparisons throughout a study must
not be affected significantly by errors inherent in the MATH test. As the number of
comparisons between data from the MATH test increases, it is important to recognize the
limitations of this test. This study will evaluate the effects of mixing time, agitation
intensity, hydrocarbon volume and growth conditions on the performance and precision of
the MATH test.
6
Goal of Research
The goal of this project is to evaluate the effects of test conditions and bacterial
growth conditions oh the results and precision of the MATH test. The specific objectives
are:
1)
Evaluate the reproducibility of the MATH test with a single hydrocarbon­
degrading bacterial species.
2)
Evaluate the effects of mixing intensity, mixing time and hydrocarbon volume on
the test
3)
Evaluate the effect of C:N ratio during growth of bacteria on the results of the
MATH test.
The following null hypotheses were established as a basis for statistical
evaluation of the experimental results. Each of these hypotheses were tested using viable
cell counts and light absorbance as measures of the cell concentration in the suspension.
Hypotheses
I)
For a given set of experimental and bacterial growth conditions, the MATH test
yields reproducible results.
2)
The time of mixing in the assay does not significantly affect the results.
7
3)
The mixing intensity does not significantly affect the results.
4)
The hydrocarbon volume does not significantly affect the results.
5)
The C:N ratio during growth of the bacteria does not significantly affect the results.
6)
Cells are not removed from suspension in the test when hydrocarbon is not present.
8
BACKGROUND
Cell Surface Hydrophobicitv
The surface of many microbes can have a hydrophobic nature. This phenomenon
promotes interactions such as partitioning of bacteria at liquiddiquid and liquid: air interfaces
and adhesion of bacteria either to host tissue or to nonwettable solid surfaces (Marshall
1976). This property allows some bacteria to attach to hydrocarbons and use it as substrate,
increasing cellular mass and population at the hydrocarbon: water interface (Rosenberg et
al. 1982). Many other examples of microbial adhesion to hydrophobic surfaces exist,
including adhesion to contact lenses, elemental sulfur, mineral particles, biomaterials and
others thus emphasizing the importance of this phenomenon (Rosenberg and Doyle 1990).
The term hydrophobicity is misleading, as “hydrophobicity” literally translates as “water
aversion” and therefore contradicts the ability of many microbial cells to disperse readily in
water. Dispersion in aqueous solutions allows microbes to come into contact with the
desired surface, whether hydrophobic or hydrophilic in nature. Many hydrocarbon­
degrading organisms have specifically adapted to surviving and interacting with the
insoluble substrates they use as their carbon source.
The nature of bacterial attachment to hydrophobic surfaces or interfaces is a subject
that has proved to be a source of dispute between physical chemists and microbiologists.
At the Dahlem Konferenzen in 1984 a group of scientists made up of microbiologists and
physical chemists formed a consensus on a number of definitions relating to mechanisms of
attachment. The group agreed upon two basic mechanisms for bacterial attachment to
hydrophobic surfaces: non-specific and specific binding (Marshall 1984). Non-specific
binding involves electrostatic or hydrophobic interactions, while specific binding involves
9
complementary sites on the bacterial surface and the colonizable surface with high affinities
for each other (Sweet et al. 1987). The complexity of defining and understanding the nature
of microbial adhesion to surfaces demands further research.
Microbial adhesion to hydrocarbons at the oil-water interface is a special case of
microbial attachment to hydrophobic surfaces. In nature, microbial growth on waterinsoluble hydrocarbons requires direct contact between bacterial cells and the oil phase. The
first step in hydrocarbon degradation typically involves oxygenase an intracellular enzyme.
With hydrocarbons being very nonpolar, and thus insoluble in water, it follows that the
microbe must come into direct contact with the nonaqueous phase. Microorganisms cannot
grow inside the nonaqueous phase as they require water as well as soluble nitrogen and
phosphorus. As a result, hydrocarbon-degrading bacteria grow at the hydrocarbon-water
interface.
There are four processes involved in biodegradation of hydrocarbons: adhesion,
growth, desorption and surface renewal. Adhesion is the process by which the bacteria
become attached to the water-petroleum interface. Hydrocarbon-degrading bacteria are able
to adhere to the hydrocarbon through hydrophobic interactions. Thin fimbriae which
extrude from the cell membrane allow hydrocarbon-degrading bacteria to adhere to the
hydrocarbon. Three observations support this. First, mutants that lack thin fimbriae fail to
adhere to the hydrocarbons, and grow poorly on the hydrocarbon substrates. Secondly,
subjecting cells with thin fimbriae to high shear causes loss of cell surface hydrophobicity
and the ability to adhere to hydrocarbons. Finally, revertants of nonadhering species regain
the ability to adhere to hydrophobic surfaces (Rosenberg and Kaplan 1987).
Following adhesion/adsorption, growth occurs at the hydrocarbon-water interface.
In this stage there are strong intercellular interactions, as supported by the observation that
desorption of cells from oil yields cell clumps. It is important to note that in initial growth,
cells multiply at an exponential rate until the surface is covered. When the surface cell
10
population reaches a monolayer, the surface area becomes the limiting factor, and the
biomass increases at an arithmetic rate. This indicates why the amount of hydrocarbon
surface area to which microbes can attach limits degradation of hydrocarbons.
There are two factors involved in extending the exponential growth phase:
hydrocarbon emulsification, and the transfer to a new substrate. Hydrocarbon
emulsification increases the surface area available for cell growth. When a 2 mm oil droplet
is broken down to droplets of 10 jm \ diameter, a typical value for bacterial-induced
emulsions, the surface area increases by 200 times. The interfacial tension between the
bacterial-coated hydrocarbons and water is lower than between pure hydrocarbon and
water. It is thought that emulsifying agents utilized by bacteria play a role in oil degradation
at the microscopic level. Bacterial cells produce extracellular emulsans (polysaccharidecontaining emulsifiers) which adhere to hydrophobic compounds. These encapsulated
emulsans are present on cell surface. When the hydrocarbon droplets useful substrates have
been utilized the encapsulated emulsan is released. The released emulsan forms a polymeric
film which plays a key role in desorption and surface renewal. The depleted oil droplet is
covered by a film which now has hydrophilic character restricting attachment of hyrophobic
bacteria (Rosenberg and Kaplan 1987).
During the growth phase of the bacteria cells at the hydrocarbon-water interface, the
emulsan is tightly bound to the cell. The emulsan capsule is released from the cell surface
when starvation conditions occur. Then emulsan adsorbs avidly to the oil droplet, thereby
displacing the cells to the aqueous phase. These “used” droplets are covered with a very
stable monomolecular film of emulsan that prevents bacteria from reattaching to the
depleted droplet (Inouye 1985). The fatty-acid ends are oriented towards the hydrophobic
organic phase while the polar hydroxyl and carboxyl groups are faced towards the aqueous
phase. The released capsule-deficient bacteria are free to attach to fresh substrate.
11
Previous Research
The earliest reference to cell surface hydrophobicity was in Mudd and Mudd’s
classic paper in 1924 in which experiments of bacterial attachment at the oil-water interface
were studied. This research opened up a new field of research in microbial interaction with
surfaces and interfaces. The study attempted to understand the physical-chemical factors
involved in penetration of bacteria through epithelia cells of the animal body by first
investigating the mechanisms of transport of a simpler system. Mudd and Mudd
investigated the kinetic mechanism of bacterial transport at the interface between two
imiscible fluids and postulated that the mechanism was dependent primarily upon the
interfacial surface tension forces at the interface. Marshall and Cruickshank studied cell
surface hydrophobicity and the orientation of certain bacteria (1972). Photographs portray
Flexibacter CW7 cells at the hydrocarbon-water interface oriented perpendicular to the
hydrocarbon surface (Marshall 1976). Another focus of the research was microbial
oxidation of oil products, at first for development as a potential protein source, and later as
a means of cleaning up after oil spills. One such study measured the interfacial area of
hydrocarbon and its relationships to specific growth rates in yeast fermenters. This study
(Wang and Ochoa 1972) determined that the specific growth rate is directly related to the
specific hydrocarbon interfacial area.
The MATH assay was developed in 1980 and proposed bacterial adhesion to
hydrocarbon as a simple, general method for measuring cell surface hydrophobicity. The
method was based on the percentage of adherent cells to various liquid hydrocarbons
subjected to agitation for brief periods. It was also discovered that hydrocarbon degraders
were not the only bacteria with the ability to adhere to the bulk hydrocarbon (Rosenberg,
Gutnick, and Rosenberg 1980). During this time many different hydrophobic interactions
12
were studied, including adhesion to mineral surfaces, fish surfaces, oral tissue and to inert
surfaces such as polystyrene (Rosenberg and Doyle1990).
Research in the late 1980s and early 1990s continues to look at the adhesion of
bacteria to different surfaces and the further development of quantitative measurement
methods. During this time the MATH assay has been redesigned a number of times in an
attempt to increase the reproducibility of the test results. A method that has proved to be
successful in capturing the kinetics of surface colonization in dynamic systems is image
analysis. Using the image analysis system Escher determined that sorption-related
processes are a function of the bulk cell concentration and interface dynamics (Escher
1986). In 1990, Mueller looked at the adsorption process and the effects of different
substrata on the sticking efficiencies of Pseudomonas aeruginosa and Pseudomonas
fluorescens bacteria.
The importance of the MATH test in determining cell surface hydrophobicities of
mutant and nonmutant strains of Acinetobactercalcoaceticus RAG-1 and the role of
emulsan in surface renewal cannot be understated. Through the use of MATH, Rosenberg
was able to determine the roles played by the extracellar thin fimbreae and encapsulated
emulsan. From this work a novel method for comparing cell surface hydrophobicities was
produced and has proved to be among the most useful methods to date.
Measurement of Cell Surface Hydrophobicitv
Historically, measurements of cell surface hydrophobicity have been used to
compare the relative affinities of bacteria for different surfaces or to compare the influence
of different experimental conditions such as nutrient, media or growth phase effects. Some
of these methods can be very time-consuming and labor-intensive, prohibiting their
usefulness in many laboratory situations. The development of the MATH assay was a
13
breakthrough in regard to the relative ease of collecting data indicating a certain affinity of
the bacteria toward the hydrocarbon-water interface. Unfortunately, this assay does not
always agree with some of the other methods. Even when comparing different experiments
using the MATH assay, the results are often hard to interpret. One such example is the
case of A. viscous cells that adhere to hexadecane only after vortexing and not when mixed
with gentle agitation (Rosenberg 1991). Rosenberg’s research in developing a simple test
for cell surface hydrophobicity has undergone many alterations since he first published the
method in 1980.
Hydrophobicity Assays
BATH and MATH Assays
In his first paper Rosenberg described a simple method for determining cell surface
hydrophobicity using BATH (bacterial attachment to hydrocarbon). This was later changed
to MATH (microbial attachment to hydrocarbon) in part because of extensive use of the
assay in studies of eucaryotic organisms (Rosenberg and Doyle 1990). The MATH assay
utilizes the fact that many hydrophobic microbes suspended in buffer media, when agitated,
will adhere to hydrocarbon and form emulsions at the hydrocarbon-water interface.
Microbes in a suspension of buffer solution are placed in a test tube to which the
hydrocarbon (hexadecane, octane or xylene) is added in varying amounts. The test tubes
are vortexed for 120 seconds, after which the mixture is allowed to separate and the
absorbance of the aqueous layer is measured. The absorbance is plotted against the
hydrocarbon volume, thus revealing the affinity the bacteria have towards the hydrocarbon
as a function of hydrocarbon volume (Rosenberg etal. 1980). Lichtenberg etal. (1985)
proposed a kinetic approach for the MATH assay. In this procedure varying amounts of
14
hydrocarbon are added to the cell suspensions and are agitated for fixed consecutive time
periods. Following each agitation the phases are allowed to separate and the absorbance of
the aqueous layer is measured. By plotting the logarithm of the percentage of cells
remaining in the aqueous phase as a function of time, the results yielded a linear
relationship. The authors determined, “the slope of the curve is a quantitative expression of
the affinity of the cells tested for the water: oil interface, and denote it as the removal
coefficient (K) of the cells from the bulk aqueous phase by the hydrocarbon” (Lichtenberg
etal. 1985). The test was further developed with the use of polystyrene cuvettes as the
experimental vessel instead of test tubes (Sharon etal. 1986). It is noteworthy that in a
scientific citation in 1988, a paper in which E. Rosenberg is a co-author, the use of the
kinetic method is not employed (Pines etal. 1988).
Contact Angle Measurement
Contact angle measurement (CAM) is a standard method for studying hydrophobic
surfaces. The method measures the surface free-energy of solid surfaces. CAM was first
used in measuring the hydrophobic surface properties of bacteria in 1972 (van Oss and
Gillman). In order to measure the contact area the surface needs to be homogeneous, flat
and dry. Therefore, conditions for using CAM on microbial studies requires that the cells
be flat and dry. Microbial mats on agar plates is one method used for these studies. This
method could prove useful in its own way, but it takes special equipment and technical
experience to measure the contact angles.
Hydrophobic Interaction Chromatography
Hydrophobic interaction chromatography (HIC) measures microbial adsorption to
octyl-Sepharose or phenyl-Sepharose beads (Rosenberg and Doylel990). A column
15
packed with an aqueous suspension of beads bearing either octyl or phenyl groups is
treated with suspension of cells. As the cells pass through the column some of them stick
to the beads. The percentage adhering is determined from the loss of turbidity or
radioactivity in the elute compared to the initial level. Controls filled with untreated beads
have had cells adhere to them . Another problem is that some researchers have used
salting-out agents to increase the amount of adherence to the beads (Rosenberg and
Doylel990). The nature of suspension solution often plays a role in the results in this test
method. The method has potential if all tests are performed under the same conditions.
Salt Aggregation T est
The salt aggregation test (SAT) is an extremely easy technique for studying the
aggregate behavior of cells (Rozgonyi etal. 1985). The cells are challenged with
increasing amounts of salting-out agents (ammonium sulfate), and as the salt concentration
increases, cells start to clump together. Measurements are compared by observing at what
concentration clumping began. This method will not work with bacteria that clump readily
without added salt and also may be dependent on initial cell concentrations.
Two Phase Partition
Two phase partition (TPP) measures the distribution of cells between two
immiscible fluids (Rosenberg and Doyle 1990). Usually one contains dextran and the other
polyethylene glycol. The bacteria are added to the system and thoroughly mixed. After the
mixture separates, samples are taken from each phase and the amount of cells in each phase
is determined. In some cases cells may bind at the interface, causing problems for recovery
and interpretation of the results. The test is also sensitive to changes in concentration,
batches and molecular weight distribution of the two polymers.
16
Binding of Molecular Probes
This method uses hydrophobic molecular probes that bind at the outermost surface
of cells to determine hydrophobicity. Kjellberg, Lagaercrantz, and Larson studied binding
of radiolabled dodecenoic acid to bacterial cells to quantitatively determine the cell surface
hydrophobicity (1980). Molecular probes often will aggregate without the addition of polar
or charged groups, which can interfere with the results. Another potential problem is the
internalization of some of the probes, which may cause cell damage (Rosenberg and Doyle
1990).
Adhesion to Hydrophobic Surfaces
Microbial adhesion to hydrophobic surfaces can be another very useful method.
Study of bacterial adhesion to hydrophobic surfaces such as polystyrene employs a variety
of techniques (Rosenberg and Doyle 1990). Some examples are measurements of
radiolabeled dodecanoic acid to bacterial cells and microspheres of polystyrene in a aqueous
suspension directly observed via microscope or dynamic observations using image
analysis. One advantage to this method is the ability to use contact angle measurements to
determine the hydrophobicity of the surface. One advantage of polystyrene is the
transparency of the material, which allows direct observation of the surface with
conventional microscopy.
Direction of Spreading
Direction of spreading (DOS) was developed to observe differences between the
hydrophobicity of individual cells and the surfaces of bacterial colonies (Sar 1986). The
direction in which a drop of water spreads when placed at the border between the bacterial
17
lawn and another surface determines the nature of the bacterial lawn surface. The border
surfaces used to make the measurements are: agar/bacteria, glass/bacteria, polystyrene/
bacteria. The droplet tends to move away from the border of the microbial lawn towards
the material surface if the microbial lawn is more hydrophobic. Therefore, the more
hydrophobic the bacteria is the more the water droplet will move towards the hydrophobic
material. Sar found that the method, when tested against the MATH assay, will sometimes
gives higher values for cell surface hydrophobicity.
Image Analysis
Image analysis enables dynamic systems to be studied. Bacterial adhesion to
stationary surfaces can be studied by passing cell solutions through a flow cell and
measuring directly the rates of attachment, detachment and growth on the surface. This
method allows direct observation of the overall processes in a dynamic system through the
use of microscopy and video techniques. The adhesion of S. sanguis 12 to FEP
(fluoroethylenepropylene) and glass surfaces in real time using image analysis is described
by Busscher et al. (1987). The adhesion kinetics of S. sanguis 12 during the first four to
six hours of exposure to both substrata under various conditions yielded the initial rate of
adhesion and the number of cells adhering in a stationary state. Escher in 1986 and Mueller
in 1990 also used image analysis to determine the kinetics of bacterial attachment to
surfaces. Escher’s goal was to determine the influences of independent parameters such as
the fluid dynamics, biomass concentration in the bulk fluid, surface characteristics and cell
physiology on early colonization of surfaces. Mueller’s goal was to determine the effects of
different surfaces, environmental conditions, and microbial species on early bacterial
colonization in continuous flow systems. Mueller used the MATH assay to determine the
hydrophobicity of the bacteria in his study, but did not determine the hydrophobic character
18
of the surfaces tested. Comparisons between the hydrophobic surface character and the cell
surface hydrophobicity may have helped to determine if a correlation between the two
methods could be found.
Physiological Effects on Cell Surface Hydrophobicity
Growth Phase
Microbial populations in batch culture follow a typical growth cycle with three
distinct phases: lag, exponential, and stationary. Lag phase is a time period just after
innoculation of the media during which little or no growth occurs. During the exponential
phase cells grow and replicate exponentially as long as nutrients are available in the culture
media. Once the microbes have utilized most of the available nutrients they enter the
stationary phase and growth of the cells slows down. Cells in an environment in which
nutrients are limited begin to slow down the synthesis of cellular components. Stationary
phase is not static. During this time, for example, E. coli cells will continue to produce
DNA after protein synthesis has been reduced (Neidhardt etal. 1990). Physiological
differences between cells harvested from different phases of growth have also affected the
results of the MATH assay. Lichtenberg et al. (1985). report that S. pyogenes cells
harvested during the logarithmic growth phase had a removal coefficient (K) several orders
of magnitude lower then those harvested during stationary growth. Rosenberg reported
similar results in 1980 when he compared the lag phase and stationary phase
hydrophobicities of S. marcescens cells. For this reason, care has to be taken that the
time of harvesting is constant, so that physiological characteristics of the cells do not
change substantially between experiments.
19
Growth T emperature
Cultures incubated at different temperatures do not always have the same physical
or chemical characteristics. These differences will often affect the MATH assay and other
hydrophobicity tests. Sharon etal.1986 grew S. marcescens at 30°C and 35°C, and a
greater removal coefficient (more hydrophobic) for the cells grown at 30°C was observed.
The temperature of culture solutions needs to be kept constant between experiments for
valid results.
Cell Generation
Isolation of microbial cells for experiments may require cells to be removed from
the original batch culture onto a more ‘friendly’ medium such as agar plates. In some cases,
cells grown and subsequently transferred to another medium may lose some of their
original characteristics (Rosenberg and Doyle 1990). Cells taken from cultures and then
suspended should be of the same generation for each experiment to keep errors due to loss
of genetic material at a minimum.
Cell Type and Preparation
Microbial cells when washed may lose some of their hydrophobic character. This
could result from a loss of surface components when the bacteria are rinsed, agitated and
resuspended. Microbes exist in a number of states in nature, such as in biofilms, microbial
mats and cells in suspension. The microbes in the biofilm will not have the same
characteristics as microbes in suspension or in a microbial mat. Sar found that bacteria in
microbial mats had higher DOS and contact angle measurements correlating to higher
hydrophobicity than readings on suspended cells from the MATH test (1987). Washed
20
cells and unwashed cells were prepared, measurements of cell surface hydrophobic!ty by
DOS and MATH methods were made. The results for both methods were similar except
that the washed cells in some cases had lower values for the MATH test (Sar and
Rosenberg 1987). The authors suspect that removal of the hydrophobic slime from the cells
decreases the hydrophobicity values obtained by the MATH test.
Physical Factors Effecting the MATH Test
The MATH test can be influenced by factors that do not depend on the
physiological characteristics of the microbial cell as affected by growth conditions. Some
outside influences such as adsorption to water films, properties of the aqueous phase,
presence of amphipathic contaminants, cell concentration and cell clumping can affect
bacterial adhesion to the hydrocarbon-water interface. These factors can have pronounced
effects on the results by either increasing or decreasing the MATH measurement. Adhesion
to the sides of glass vessels or other containers will likely cause higher MATH readings.
Cells normally in the bulk liquid which adhere to the agitating vessel wall lower the final
absorbance reading in the MATH test. The cells on the wall have not interacted with the
hydrocarbon and should not be included in the test. A control vessel, a container treated the
same without hydrocarbon, can be used to subtract the amount of cells adhering to the sides
of the mixing vessel after agitation. Bacteria which are extremely hydrophobic have a
tendency to adhere to one another. This tendency makes it difficult to get a well-mixed
suspension of cells and consistent readings for initial absorbance of the cell suspension.
Cell concentration may also prove to influence microbial adherence to hydrocarbons, so
that cell concentrations must be constant between single experiments.
The ionic strength of the buffer solution affects the extent of adhesion of bacteria to
hydrocarbons. Salting-out agents sometimes cause bacteria to aggregate. This causes more
21
of the bacteria to adhere to the hydrocarbon. Equipment used during the experiment must
be clean and free of amphipathic contaminants. Glassware should be treated to remove all
surfactants from the surface, otherwise interference with the MATH test will occur.
Surfactants will interact with the cells in suspension and keep them from adhering to the
hydrocarbon, reducing the hydrophobicity measurement. Even though there are a number
of factors that influence the results, standard procedures and cleanliness can limit the effects
on any one experiment.
Continued Research
Importance of Research
Hydrophobicity of hydrocarbon degraders is not the only area where the MATH
test and other methods are used. Cell surface properties of bacteria isolated from fish have
been studied using MATH and DOS methods Sar and Rosenberg 1987). This research was
used to determine the surface character of the bacteria found on fish scales. The researchers
found that the some of the bacteria have hydrophobic characteristics though the values for
DOS were often greater then those for MATH. It is thought that the hydrophobic surfaces
of the bacteria may aid the fish in motility by reducing the frictional force of water against
the fish.
The development of mouthwash is another case where the MATH test has proved
useful. In 1982 Weiss et al. found that a high proportion of oral bacteria isolated from
extracted teeth and steel bands adhered to hexadecane. Other studies had earlier indicated
that microbes which adhere to hydrocarbons will also behave similarly when treated with
various nontoxic oils. Mouthwashes with an oil phase were tested against commercial
22
mouthwashes. The result was a significant reduction of oral microbial activity and bad
breath when mouthwashes containing oil were used (Rosenberg 1991).
The previous examples illustrate the importance of cell surface hydrophobicity and
the need for different adhesion tests. Hydrophobicity tests are used to elucidate the surface
properties of microbial cells to help determine the possible interactions between the cells
and a surface. Knowledge of this sort is useful, especially when one considers the vast
number of systems that have colonizable surfaces. Understanding the mechanisms of
bacterial degradation of hydrocarbons at the interface would also give valuable information
about how bacteria are able to utilize this hydrophobic carbon source. If the kinetics of
bacterial attachment, desorption, growth and surface renewal can be quantified the
information obtained could help researchers develop methods to maximize degradation of
oil in natural or man-made systems.
Comparable Tests
One of the flaws of the MATH test is its inability to allow for comparison between
different experiments. Many of the factors involved in reproducibility have less to do with
the overall experimental method than with standardization of factors influencing the test.
Another factor that may prove to influence the results of the MATH test are “user errors”.
Some researchers may interpret a method differently then the author intended. For example,
the method used to vortex the cells may have a pronounced effect on the results of the test.
Even though there are better tests which allow comparisons between cell surface
hydrophobicity, few of them are as fast and simple as MATH.
To make comparisons between experiments from many laboratories using MATH
is, at this time, rather inconclusive. There are cases where MATH was compared to other
methods of determining cell surface hydrophobicity, the bacteria came from one generation
23
of bacteria which were scraped from agar plates (Dillon and Koohmaraie 1986). The ability
to compare a variety of microbes from a simple test would allow investigators to screen
bacteria with unknown surface properties for desired or undesired traits with some
assurance that their results are valid. Such a method would make it easier to distinguish
nonadhering hydrocarbon degraders from adhering hydrocarbon degraders. A MATH test
that is quantifiable would also allow results to be easily compared so that influencing
factors can be more easily studied.
24
METHODS
The methods section is divided into four parts: preliminary work, experimental
design, and experimental procedure. The preliminary work describes the isolation of the
bacteria and background experiments and procedures. The next section describes
experimental design followed by the experimental procedure portion which describes how
each experiment was carried out.
Preliminary Work
Isolation
The bacteria used for this research project were isolated from a contaminated soil
sample using a modification of technique designed to isolate acetonitrile-utilizing bacteria
(Chapatwala et al. 1990). These isolates were transferred to silica slant tubes into which
the,hydrocarbon hexadecane was added as substrate (Funk and Krulwich 1964). The
colonies on the slant tube that were able to degrade the hexadecane were further isolated to
obtain single colonies. From the slant-tubes, colonies were selected for transfer to tryptone
glucose extract agar plates. Transferred colonies were then streaked on the agar plates to
isolate individual colonies once more. All of the slant tubes and agar plates were incubated
at 35° C until enough growth occurred to allow for individual isolation of colonies. The
colonies were transferred to 250 ml flasks containing PUM buffer nutrient media with
hexadecane as the carbon source. These flasks were then placed on a shaker table that
controlled the temperature and agitation rate for increased oxygen availability. The bacteria
grew slowly, breaking down the hydrocarbon while creating a murky solution with a great
25
amount of bacterial clumping and emulsion production. From this procedure two isolates
with the desired characteristics were cultivated for use in all experiments.
The characteristics of the two isolates were similar, although one isolate was pink in
color while the other was clear. Both bacterial isolates degraded hexadecane, produced
emulsions and collected at the hydrocarbon-water interface. Only the pink colonies were
used in the MATH experiment. The pink isolates were chosen for two reasons; they were
easier to spot at the hydrocarbon-water interface and individual colonies were easier to
discern from groups of colonies. The isolates grew slowly on hexadecane and the
suspension remained viable for over a month after which it became necessary tosubculture.
Substrate
The substrate desired for growth of cells in batch cultures had to have certain
characteristics in order for the desired results to be achieved. Using an insoluble substrate
such as hexadecane was not practical because the amount of substrate available to the cells
is not easily quantified. Further, residual hydrocarbon adherence to cell surfaces could
interfere with the MATH test results. It was also required that the substrate contain no
nitrogen in the molecular formula. The first choice was fatty acids, such as formate, nbutyrate, iso-butyrate, propionate or acetate. These substrates dissolve readily in aqueous
solutions and are often intermediates in the metabolism of straight-chain alkanes such as
hexadecane. The isolate was able to metabolize all of the fatty acids. However, the inability
to easily sterilize fatty acids decreased their desirability as the substrate. Ethanol was tried
as a substrate because it is simple in chemical make-up and resembles the end of a fatty acid
or alkane chain. Also, the alcohol did not have to be sterilized before addition to the nutrient
media. The isolates were able to use ethanol, although the growth rates were extremely
slow.
26
C:N ratio calculations Calculations for relative amounts of carbon in the form of
ethanol and nitrogen as ammonium chloride were based on an article by Sykes (1975).
Growth yield (Ybioh ) and values for the carbon-to-nitrogen ratios for balanced growth
were determined. The range varied from 6.26:1 to 4.94:1. The C:N ratio is important
because it has been determined that the relative availability of carbon, nitrogen and
phosphorus determines the cells reproduction and production of extracellular polymeric
substances. When both carbon and nitrogen are plentiful, the cells grow and reproduce. As
nitrogen is depleted, some cells will then switch to recycling of nitrogen. In this case
carbon may be excreted from the cells as extracellular polymeric material (EPS). By picking
a C:N ratio above that for balanced growth the cells should have enough carbon to grow
readily. A C:N ratio of 15:1 is high enough above the minimum requirement that an excess
of carbon will be available to the cells in solution. The physical and chemical surface
properties should therefore be different for cells grown at a C:N ratio of 15:1 than for those
grown at 5 :1.
NutrientMedium
The choice of medium in the growth experiments was determined from adjustments
made to the media used to isolate hydrocarbon-degrading organisms. The components of
the nutrient media passed through various changes until one that did not interfere with the
absorbance experiments was found. Initial choices of media formed precipitates that were
carried through the cell suspension preparation and contributed to the absorbance. Initial
cell counts of these experiments were usually lower than those for experiments that did not
have a precipitate left over from the nutrient medium. This interference was not noticed
initially because the interference from the precipitate did not occur until low concentrations
of cells in the solution occurred. When the precipitates were present, an initial drop in
27
absorbance was observed at low agitation rates and mixing times. When the higher
agitation rates were used no further drop in absorbance was observed. At the point where
the absorbance no longer changed cell numbers continued to decrease. The absorbance of
the precipitate in the solution became a background that did not change with the continued
removal of cells. For this reason a modified Buschnell-Haas broth was used for the
culturing of the bacteria (Table I).
Table I. Composition of nutrient media.
Magnesium Sulfate
Calcium Chloride '
Monopotassium Phosphate
Dipotassium Phosphate
Ferric Chloride
(M gS04*7H20)
(CaCl2^ H 2O)
(KH2PO4)
(K2HPO4)
(FeCl3^riH2O)
Ammonium Chloride
for C:N 15:1
Ammonium Chloride
for C:N 5:1
(NH4Cl)
(NH4Cl)
0.2 g/1
0.02 g/1
1.0 g/1
1.0 g/1
0.625 ml of O.Olg/ml stock solution
added to each 125 ml flask.
0.319 ml of 10 mg/ml stock solution
added to each 125 ml flask.
0.955 ml of 10 mg/ml stock solution
added to each 125 ml flask.
Experimental Design
M ATHAssay
C:N Ratio
The MATH test was used to determine if the carbon to nitrogen ratio
(C:N) would affect the hydrophobicity of a hydrocarbon-degrading bacterium. C:N ratios
of 5 :1 and 15:1 were used, representing carbon-limited and carbon-rich environments
respectively. In the breakdown of hydrocarbons by bacteria in sea-water the limiting
nutrient is nitrogen (Atlas 1984). For carbon-limited cells the C:N ratio is 5:1, although
there.can be some variation (Characklis and Marshall 1990,145). The carbon-rich C:N
28
ratio at 15:1 is higher then the ratio (7.7:1) needed for maximum production of emulsan for
Acinetobacter calcoaceticus RAG-1 (Shabtai and Wang 1989).
Experimental Conditions
Concurrent with testing the stated hypothesis on the
effects of the C:N ratio, the experiments were designed to obtain statistical data on the
accuracy and precision of the MATH test itself. The design of this part of the experiment
involved three factors of the MATH test: agitation rate, hydrocarbon volume, and time of
agitation (Table 2). Two different methods were used to obtain this data: absorbance
measurements and viable cell counts. These methods were also compared to determine if
there were significant differences in the results of the two cell enumeration techniques.
Table 2. Settings for agitation rate, hydrocarbon volume and mixing time.
AgitationRate
60
100
140
Hydrocarbon Volume
in ml.
0.2
0.6
1.0
Tim eofM ixing
in seconds.
10
20
40
AgitationRate
Three agitation rates were used in the test. A Maxi Mix II (vortexer source) was
used at the highest setting and the rates were controlled using a Powerstat type 116B
(voltage regulator) on three settings; 60,100 and 140 (Rosenberg 1980). The rate of
rotation was measured by strobe light for each of the settings. The rotation rate of the
mixing plate and the fluid inside the test tubes for each agitation rate measured as rpm
(rotations per minute) were: 1470 at 60, 2810 at 100, and 3100 at 140.
29
Hydrocarbon Volume
The amount of cell suspension in the test tubes was held constant at 5 ml. Three
different hydrocarbon amounts (0.2 ml, 0.6 ml and 1.0 ml) were used in the experiments
In the initial MATH test, 1.2 ml of cell suspension was subjected to varying volumes of
hexadecane, xylene and octane. The volume of hydrocarbon ranged from 0 to 0.2 ml. A
later testing procedure used 5 ml of cell suspension to I ml of hydrocarbon. For the
reported data, the larger volume technique was used because cell counts and absorbance
measurements after agitation required the full 5 ml of cell suspension.
Hydrocarbon Screening
To determine which hydrocarbon would be used for the research the three
hydrocarbons (hexadecane, xylene and octane) were tested. The experimental design
compared the results of each hydrocarbon’s efficiency and variability for removal of
bacterial cells from the media. Initial results indicated the three methods had similar percent
removals and reproducibility (Table 3). The experimental error associated with the three
methods seemed to be similar, and in some cases the experimental error was large.
Table 3. Absorbance and Percent Removal for Three Hydrocarbons
Hydrocarbon
Xylene
Octane
Hexadecane
I nit. Abs.
.301
.302
.290
10 sec.
.209
.213
.215
Time of Mixing
20 sec.
40 sec.
.196
.192
.204
.198
.209
.204
80 sec.
.191
.196
.191
% Removal
3 6 .5 %
3 5 .0 %
3 4 .1 %
This finding was similar to results obtained by Sweet et al( 1987), where it was determined
that residual hydrocarbon can adhere to bacteria and increase the refractive index of the
30
bacteria. The residual could be removed by aerating the solution for one minute causing
desorption from the bacterial surface. The results using this method were less reliable for
hexadecane than for xylene.
One potential problem that can occur when using xylene as the hydrocarbon is lysis
of the cells. This would cause lower absorbance measurements to be observed in the
experiments. Sweet etal. determined that lysis of cells was not occurring in the presence of
xylene (1987). By aerating a suspension of cells and xylene overnight, the xylene was
removed and the original absorbance was recovered. Viable cell counts confirmed that the
cells were not being lysed. This research followed the procedure of Sweet etal. by using
xylene as the hydrocarbon and aerating the cell suspensions to remove xylene attached to
the cells or dissolved in the water.
Tim eofM ixing
Researchers using the MATH test have used agitation times ranging between 5-120
seconds. To minimize the number of mixing periods the point where continued mixing no
longer changed the absorbance was determined. After using three different hydrocarbons
and mixing times running to 120 seconds, it was determined that no significant changes
were taking place after 40 seconds. Table 3 also shows this trend, with absorbance values
changing little after 40 seconds of mixing.
Experimental Procedure
Inoculation
Numerous authors cite growth phase as affecting the MATH test. It was for this
reason that all cells were harvested during stationary phase at the same interval after
31
inoculation of the culture. The soil isolate had a slow growth rate and did not reach the
stationary phase until the seventh day of incubation. To determine when stationary phase
was reached, 5 ml samples were taken twice a day from the culture for nine days.
Stationary phase was reached when the absorbance measurements no longer increased.
Stationary phase was also confirmed from cell count data. Once the stationary phase was
determined, a harvest time was picked that would be constant in all the experiments. All
cells for each experiment were therefore harvested between 190-194 hours, or about 8 days
after inoculation. The culture flasks were agitated at 110 rpm on a New Brunswick
Scientific water bath shaker table at 37° C.
First generation cells, scraped from R2A agar plates, were suspended in 125 ml of
Bacto Bushnell-Haas Broth nutrient medium (Table 3). For a 5:1 C:N, 0.955 ml of stock
solution was added to the nutrient medium, while for a 15:1 C:N, 0.319 ml of solution was
added. Filter sterilized ferric chloride (0.625 ml) was added separately because of its
tendency to precipitate when autoclaved. 3 1 3 ]A of ethanol, the substrate, was added to the
nutrient medium. After 8 days, the cell suspension was transferred to centrifuge bottles and
placed in the centrifuge.
Washing Procedure
The cell suspension was added to centrifuge tubes and placed in a RC5C .Sorvall
Instruments centrifuge set to run for 25 minutes at 12,000 xg. After centrifuging, the liquid
was poured off and the pellet was washed with 50 ml of buffer solution (Table 4) which
had been filtered through a .2 pim filter to remove any precipitate formed by the autoclaving
process. The cells were vortexed to break up the pellet and placed in the centrifuge again.
The same procedure was then performed twice more.
32
Table 4. Composition of buffer solution.
Monopotassium Phosphate
(KH2PO4)
Dipotassium Phosphate_______ (K7HPO4)
22.2 g/1
7.26 g/1
Resuspension of Cells
The cells were resuspended in sufficient phosphate buffer solution to produce the
desired absorbance of 0.2 at 500 nm. This usually gave a total test volume of between 100150 ml. The initial cell counts were determined by plate counts.
Addition to T est tubes
Experiments were initially carried out in polystyrene cuvettes following the
procedure of Sharon etal. (1986). The isolate tended to stick to the surface of the
polystyrene cuvette after agitation. Controls that lacked the hydrocarbon had substantially
lowerabsorbance values after agitation than the initial absorbance measurements. Therefore,
the MATH tests were carried out in 16 * 150 mm Kimex test tubes even though the
procedure was more tedious. To remove potentially interfering surfactants, test tubes were
washed by hand and rinsed three times in distilled water, placed in a muffle furnace set at
500 °C for four hours. The test tubes were then capped and autoclaved for twenty minutes.
This process kept the test tubes free of biological and chemical contaminants that could
interfere with the experiments. Five milliliter portions of the cell suspension were divided
among test tubes, generally using between 12 to 15 test tubes. Hydrocarbon was added to
the cell suspension in three different volumes: 0.2 ml, 0.6 ml and 1.0 ml. Each portion of
hydrocarbon would be in triplicate and three test tubes Would be free of hydrocarbon as
controls. In some cases, when a greater volume of cell suspension was available, a larger
experiment could be run. Table 5 shows a typical experimental array of test tubes with the
33
relative amounts of hydrocarbon, agitation rate and time of mixing. The test tubes were then
placed at a constant temperature of 30° C for 20 minutes.
T able 5.
Example of a Typical Experiment
Agitationrate
HC volume
.60 / 0.2
100 / 0.6
140 / 1.0
Control
10s
V
V
V .
V
20s
V
V
V
V
40s
V
V
V
V
Agitation
After the 20 minute equilibrium period one of the test tubes was placed on the
mixer plate and agitated at the desired setting and time period. The test tubes were randomly
selected to help average out any error in the transfer of cell suspension to the test tubes. The
setting on the Powerstat would be changed only after all the tests for that particular agitation
rate were finished. The experiment progressed until all samples and controls were mixed at
their prescribed agitation rate for a given time period.
Cell Enumeration Techniques
Two methods were used to determine the concentration of cells in the suspension
media; viable plate counts and absorbance measurements. Enumeration of viable cells using
PC (plate Count) methods gives information on the number of cells able to reproduce and
form colonies on agar plates. It is thought, however, that in some cases injured cells that
are still performing some biological functions will not be able to grow on the agar plates
(Yu, Pyle, and McFeters 1992). Absorbance measurements provide information on the
numbers of particles without regard to viability or culturability. Further, absorbance
34
measurements are not as adversely affected by clumping as plate counts. Cell counts were
initially to be determined using direct microscopic counts. Direct microscopic counts did
not work however, due to the tendency of bacteria to clump when passed through the (0.2
jUm pore size) black polycarbonate membrane filter (Nuclepore).
Cell Enumeration
A modified drop plate method was used for the PC (plate count) cell enumeration
(Miles and Misra 1938). The cell suspension was separated from the hydrocarbon layer by
carefully pipeting the lower aqueous layer to clean test tubes. From the aqueous
suspensions of some of the test tubes, I ml of supernatant was removed and transferred to
a dilution tube containing 9 ml of autoclaved buffer and I /d of homogenization fluid and
homogenized to disperse the cells (Camper etal 1985). Serial ten-fold dilutions of the
culture were then performed. Petri dishes containing R2A agar were divided into three
equal portions, one for each of the ten-fold dilutions (e.g., IO"4, IO"5, IO"5). Ten 10 ]A
drops of a diluted cell suspension were placed in each section of the p e tri. The plates were
incubated at 35° C for eight days before colonies became large enough to count. The cell
counts were determined by counting the number of colonies at a dilution where no fewer
than 10 and no more then 50 colonies were observed for each drop of sample at a particular
dilution. The remaining cell suspension was set aside for absorbance measurements.
Absorbance Measurements
Absorbance was measured using a Milton Roy Spectrpnic 601 absorbance
spectrophotometer at 500 nm using buffer media alone as the blank. An initial cell
absorbance of approximately 0.200 was used. For absorbance measurements, the
remaining cell suspension transferred from the original test tubes were aerated for one
35
minute (air flow 3 ml per second) to remove any excess hydrocarbon from the liquid and
the small amounts that stuck to suspended cells. Absorbance measurements were then taken
of all the test tubes containing the cell suspension, including controls. Measurements were
made after transferring 2 ml of sample to a I cm square (4 ml) polystyrene cuvette which
could be placed in the spectrophotometer.
36
RESULTS
The results are divided into three sections; observations, percent removal
measurements and statistical analysis. The first section includes photographs depicting test
tubes used in the experiments. The second section describes different analyses used to
describe and interpret the absorbance data. The statistical analysis section describes the
general design of the experiment and the statistical results.
Test Tube Observations
Agitation at all settings and times produced an emulsion of xylene and water, which
separated into layers after about 30 minutes of rest. The adhering cells collected at the
hydrocarbon-water interface and could be observed as clumps causing deformation
(Figure I). As shown in figure 2, some clumping of cells occurred even before agitation of
the solution. Following agitation, the cell suspension in the control test tubes behaved
differently than in the hydrocarbon-filled test tubes. In the controls, more of the suspended
cells adhered to the glass surface in clumps than in the hydrocarbon-containing test tubes.
Figure 3 shows a typical control test tube after agitation and Figure 4 shows a test tube with
hydrocarbon agitated at the same rate and time period as the control. Figure 5 shows a
control test tube which had been vortexed for 20 seconds with an agitation setting of 60,
while Figure 6 depicts a control vortexed for 20 seconds at a setting of 140. Figure 7
depicts a sample with a hydrocarbon volume of 0.2 ml agitated at 60 for 20 seconds while
figure 8 depicts a test tube containing the same volume, but agitated at 140 and Figure 9
depicts an agitation setting of 140. The emulsion at the interface can be observed in the
photographs, as well as the differences in adherence to the surface of the glass in the
37
Figure I
Cells within the white circle are at the hydrocarbon-water interface.
38
Figure 2
Figure 3
Test tube containing unmixed hydrocarbon
and cell suspension.
Control test tube after agitation.
(Agitated at a setting o f 60 for 40 seconds.)
39
Figure 4
Figure 5
Test tube agitated with hydrocarbon.
(I ml hydrocarbon with agitation setting
of 60 and mixing time of 40 seconds.)
Control test tube
(Agitation setting o f 60 and mixing time
o f 20 seconds.)
40
Figure 6
Figure 7
Control test tube.
(Agitated at a setting of
140 for 20 seconds.)
Test tube containing .2 ml of hydrocarbon.
(Agitated at a setting of 60 for 20 seconds.)
41
Figure 8
Figure 9
Test tube containing .2 ml of hydrocarbon.
(Agitated at a setting of
140 for 20 seconds.)
Test tube containing I ml of hydrocarbon.
(Agitated at a setting of 60 for 20 seconds.)
42
controls (Figures 3,5,6) and hydrocarbon (Figures 4,7-9) and at the different rates of
agitation.
Percent Removal Data
Results from the MATH produced six types of data. First, the initial absorbance
(Ao) and cell counts (Co) were determined for the cell suspension. After the agitation step
was complete and the cell suspension was allowed to settle, absorbance (Ac) and cell
counts (Ce) measurements for the control volumes were taken. The final absorbance (A)
and cell counts (C) for the test tubes containing hydrocarbon were also m easured. Several
combinations of these parameters were developed, and the results were analyzed
arithmetically and statistically.
For the arithmetic method the responses were, percent removal (ideal), percent
removal (control) and percent removal (adjusted), defined as follows:
1)
Percentremoval (control) = ((control-final)/control)* 100
2)
Percent removal (ideal) = ((initial-final)/initial)* 100
3)
Percent removal (adjusted) = ((control-final)/initial)* 100
where percent removal = percentage of cells removed from the suspension, initial = initial
absorbance or cell count measurements, control = control absorbance or cell count
measurements, and final = final absorbance or cell count measurements. The percent
removal (control) was compared to the percent removal (ideal) to determine which method
of analysis was a better fit. The last method, percent removal (adjusted), used all three the
components from the MATH test for the response. The adjusted response was used to
43
measure the percent removal after accounting for the bacteria lost to the surface of the glass
test tube as determined in the controls.
Table 6.
Absorbance Responses for Cells Cultured at a C:N Ratio of 5 :1.
% Removal
(control)
10
20
40
Mean
50.3
66.7
Std. Dev. 13.7
4.1
n
2
2
Mean
63.8
Std. Dev.
n
Mean
56.9
76.3
Std. Dev. 9.1
0.7
n
2
3
Mean
80.0
Std. Dev.
8.5
n
5
Mean
86.5
Std. Dev.
n
Mean
89.5
78.0
Std. Dev. 0.9
14.4
n
2
2
Mean
72.2
Std. Dev. 1.6
n
2
Mean
79.1
81.9
Std. Dev. 8.1
4.3
n
2
2
Absorbance 5 :1
60/0.2
a
60/0.6
60/1.0
100/0.6
100/1.0
140/0.2
140/0.6
140/1.0
% Removal
(ideal)
10
20
40
79.7 84.7 85.7
7.7
6.6
5.3
4
4
4
78.9 81.1 83.9
% Removal
(adjusted)
10
20
40
21.0
32.5
0.3
4.0
2
2
33.3
82.5
5.3
5
86.8
6.8
4
85.5
7.0
4
90.6
4.1
5
98.8
90.0
4.9
5
91.4
4.2
4
26.6
1.1
2
89.5
7.4
4
87.6
0.7
2
89.6
2.7
4
90.7
5.1
3
86.6
0.7
2
90.5
3.6
3
92.3
4.7
4
89.3
0.4
2
91.5
1.9
4
53.7
19.1
2
32.3
0.7
2
42.7
13.9
2
22.7
10.5
3
39.7
16.9
5
7.7
38.8
11.1
2
35.2
1.2
2
a agitation setting / hydrocarbon volume.
b n=l for means with no standard deviation listed.
Absorbance response data from cell suspensions cultured with a C:N ratio of 5 :1 are
documented in Table 6. The responses for percent removal (ideal) and percent removal
(control) behave in a similar manner. The response for percent removal (control) varies
44
from 50.3 % to 81.9 % compared with 79.7 % to 91.5 % for percent removal (ideal). The
values for the percent removal (adjusted) show no real tendencies although the errors are
about the same as the other methods. There is a general upward trend for the responses as
the agitation rate, hydrocarbon volume and time of mixing increase. The results from the
cell count data show a greater percent removal then the absorbance data at the low factor
settings (Table 7). The response for the percent removal (ideal) at high hydrocarbon
Table 7.
Cell Count Responses for Cells Cultured at a C:N Ratio of 5:1.
% Removal
% Removal
Cell counts 5 :1
(control)
(ideal)
10
20
40
a
10
20
40
43.9
81.1
60/0.2
Mean
97.8
98.7
Std. Dev.
14.1
0.2
0.3
n
2
2
2
60/1.0
Mean
33.4
35.0 97.4
98.7
Std. Dev.
30.2
0.7
0.3
n
2
2
2
100/0.6 Mean
90.6
99.6
Std. Dev.
12.1
0.1
n
4
4
"W 8
W T T 99.3
140/0.2 Mean
99.8
Std. Dev.
0.2
0.0
n
2
2
140/0.6 Mean
99.4
Std. Dev.
n
97.8
140/1.0 Mean
99.7
99.7
Std. Dev. 2.5
0.1
0.3
n
2
2
2
% Removal
(adjusted)
10
20
40
7.4
5.1
2
0.8
7.2
5.4
2
8.5
5.9
4
6.6
11.2
25.0
22.6
2
a agitation setting / hydrocarbon volume.
b n=l for means with no standard deviation listed.
volumes, agitation rate and mixing times is almost 100 %. The responses for percent
removal (ideal) and percent removal (control) are consistently greater using cell counts than
those using absorbance A somewhat different trend is found in the data for cell suspensions
45
raised at a C:N ratio of 15:1 (Tables 8 and 9). In this case the percent removal (adjusted)
responses are almost the same for both cell counts and absorbance. The percent removal
(ideal) with responses are greater for the cell counts than for absorbance, as was seen with
the results obtained for cells cultured at 5 :1. In general, comparing Tables 6 and 8 or Tables
Table 8.
Absorbance Responses for Cells Cultured at a C:N Ratio of 15:1.
Absorbance 15:1
a
60/0.2 Mean
Std. Dev.
n
60/0.6 Mean
Std. Dev.
n
60/1.0 Mean
Std. Dev.
n
100/0.2 Mean
Std. Dev.
n
100/0.6 Mean
Std. Dev.
n
100/1.0 Mean
Std. Dev.
n
140/0.2 Mean
Std. Dev.
n
140/0.6 Mean
Std. Dev.
n
140/1.0 Mean
Std. Dev.
n
% Removal
(control)
10
20
40
47.7
45.0
62.4
14.4
13.0
6
5
68.63
60.6
19.0
5
83.95
68.5
3.6
2
73.91
68.4
8.4
5
80
75.3
80.2
10.7
6
71.03
81.2
77.3
8.5
4
80.8
1.1
2
91.21
79.5
4.5
6
81.19
76.2
8.8
5
76.6
3.9
2
78.7
7.5
6
% Removal
(ideal)
10.0
20.0
40.0
69.2
73.1
80.9
8.3
7.8
6.0
7
7
7
81.8
83.3
83.9
% Removal
(adjusted)
10.0
20.0
40.0
31.1
41.8
34.0
12.1
15.7
11.3
6
3
5
36.5
76.9
9.0
7
91.8
1.8
2
86.4
2.1
7
90.0
79.2
8.4
7
91.8
1.8
2
88.6
4.5
7
89.5
84.3
3.9
7
92.3
LI
2
89.2
3.0
7
88.4
43.7
18.3
6
36.2
40.9
11.0
3
26.8
36.1
9.2
6
27.7
40.6
3.6
2
40.4
7.0
7
25.8
35.3
1.0
2
87.7
4.0
7
91.3
2.5
2
88.0
3.9
7
89.8
3.5
7
93.2
3.6
2
89.8
2.2
7
91.2
2.6
7
90.3
0.2
2
91.7
2.8
7
42.1
10.3
5
37.7
10.4
3
44.1
33.4
4.5
7
42.7
39.4
10.2
5
42.2
8.8
2
33.9
9.2
7
a agitation setting / hydrocarbon volume.
b n=l for means with no standard deviation listed.
46
7 and 9, percent removals (ideal and control) were similar between the two C:N ratios
using both absorbance and cell counts.
The results obtained from some of the experiments were not incorporated into the
data set. A precipitate that was not observable by the naked eye formed in some of the cell
suspensions. Since initial cell suspensions were diluted to an absorbance of 0.2, the
precipitate’s contribution necessitated a corresponding decrease in cell concentration. The
Table 9.
Cell Count Responses for Cells Cultured at a C:N Ratio of 15:1.
% Removal
(control)
10
20
40
6 0 / 0 . 2 Mean
-22.6
54.4
a
S t d . Dev 4 6 . 4
31.0
n
4
2
6 0 / 0 . 6 Mean
14.4
S t d . Dev
n
6 0 / 1 . 0 Mean
1 1 2.6
39.0
S t d . Dev 1 3 4 . 3
4 4.9
n
4
2
1 0 0 / 0 . 2 Mean
7 4.4
S t d . Dev
n
1 0 0 / 0 . 6 Mean
89.3
S td . Dev.
3.7
n
5
I 0 0 / 1 . C Mean
86.5
S t d . Dev
n
1 4 0 / 0 . 2 Mean
7 9.9
90.6
96.2
S t d . Dev
8.9
0.9
n
3
3
1 4 0 / 0 . 6 Mean
91.2
S td . Dev.
n
1 4 0 / 1 . C Mean
8 8.8
97.6
98.8
S t d . Dev
4.3
0.4
n
3
4
% Removal
(ideal)
Cell c o u n t s 15:1
10
52.7
31.4
4
20
95.3
40
74.9
22.1
4
83.5
78.2
21.2
3
92.4
5.4
4
95.6
3.6
4
a agitation setting / hydrocarbon volume.
b n =l for means with no standard deviation listed.
9 4.4
91.2
8.8
4
% Removal
(adjusted)
10
20
40
6.0
31.1
17.5
3 9.6
4
4
2.8
3 8.2
37.7
4
49.4
48.1
3
97.7
6.7
94.9
2.8
6
98.8
51.4
40.6
6
7.7
9 6.8
2.1
2
95.7
9 8.4
0.9
4
99.2
0.3
3
99.5
0.4
4
47.3
3 1.6
3
44.4
4 0.3
3 8.7
3
44.7
5 1.2
33.0
3
4 7.8
33.0
36.1
4
47
effect of the particulate matter then, was to lower the initial number of cells in the solution
for the given absorbance of 0.200. The effect of the precipitate carried through the entire
experiment. The absorbance measurements would drop initially until they reached a critical
point. This point was where the particulate matter masked the cells’ effect on the
absorbance. For the 5 :1 cultures with precipitate present, the average absorbance response
for percent removal (ideal) varied from about 60 % to 80 % (Table 10). The percent
removal (ideal) for cell counts was typically 20 % (Table 10) higher since the cell counts
were not affected by the absorbance. Since species of bacteria behave differently at low cell
Table 10.
Absorbance and Cell Count Data for Cells Cultured at a C:N Ratio of 5:1 for
Experiments with Interference.
C:N
5:1
interference
a
60/0.2 Mean
Std. Dev.
n
60/0.6 Mean
Std. Dev.
n
60/1.0 Mean
Std. Dev.
n
100/0.1 Mean
Std. Dev.
n
140/0.: Mean
Std. Dev.
n
140/1.1 Mean
Std. Dev.
n
absorbance
% removal (ideal)
10.0 20.0 40.0
62.5 64.2 69.5
11.7 11.8 13.4
7.0
6.0
7.0
78.9 80.0 82.2
0.0
1.6
2.4
2.0
2.0
2.0
64.8 71.5 72.8
6.7
1.9
3.0
4.0
3.0
4.0
70.3 70.5 71.6
13.7 14.6 12.2
5.0
8.0
5.0
72.2 74.4 76.0
7.6
7.3
8.1
4.0
4.0
4.0
73.3 69.1 75.5
13.6 16.4 15.6
7.0
4.0
6.0
cell counts
% removal (ideal)
10.0 20.0 40.0
83.2
91.7
10.3
5.8
6.0
5.0
88.6
97.4
11.0
2.0
95.2
96.9
5.2
3.8
3.0
3.0
95.8 96.5
4.2
7.0
9 8 .1
98.9
2.3
1.5
3.0
3.0
98.3
98.9
2.6
1.1
5
4.0
a agitation setting / hydrocarbon volume.
b n=l for means with no standard deviation listed.
48
concentrations than they do for high cell concentrations with respect to bacterial attachment
to hydrocarbon the data in Table 10 were not used in further analyses.
Results From Statistical Analysis
The experimental design was governed by the statistical methods used to analyze
the data. The design required that data points from the three factors (time of mixing,
agitation rate, hydrocarbon volume) form the comers and the center point of a cube. The
data collected for the absorbance measurements included other points within the cube. A
linear regression model suing the software Minitab, was used to determine coefficients for
each of the factors. ANOVA (analysis of variance), was performed using the software SAS
to determine p values for each factor and the interexperiment and the intraexperiment errors.
Regression Analysis
Six response variables were determined from the experimental absorbance (A, Ac
and Ao) and cell count (C, Ce and Co) data. The six transformed variables are all
logarithms of ratios, as shown in the natural first column of Table 11. These responses
were used to develop multiple linear regression models using the experimental factors rpm,
time, C:N ratio and hydrocarbon volume. For the absorbance data when hydrocarbon was
present, for example, the equation takes the form:
In (A0ZA) = B0 + B1(rpm) + B2 (time) + B3 (C:N) + B4 (HG volume)
where B0 through B4 are regression coefficients determined through least square error
minimization. For each response, p-values and correlation coefficients R 2 were also
49
determined. Values for the regression coefficients (6 values) are shown in Table 11. The
above linear regression model was expanded to include products of the factors to look for
Table 11.
Constant terms and Coefficients for Responses.
Response
Constant
rpm
Time
C:N
HC volume
InAo/A
1.38
0.70
0.243
-0.625
0.362
In Co/C
1.98
2.62
0.801
-1.711
0.44
InAc/A
0.003
1.36
0.094
-0.123
0.413
In Cc/C
-0.073
2.12
-0.616
-0.701
0.446
InAo/Ac
1.2
-0.617
0.177
-0.560
0.039
In Co/Cc
3.50
-0.054
-0.105
-0.813
0.088
interactions and quadratic terms for each factor to allow for nonlinear behavior.
Step-wise linear regression using Minitab was used to determine which of the
factors, cross factors and squared terms were important to the model. Interaction plots were
analyzed first to determine which factors were important to the model. The interaction plot
of the four factors at a given response helped to determine if a factor was likely to affect
another. The first regression included all of the terms. The p values for the regression were
analyzed and insignificant terms (p > 0.05) were removed. The method proceeded until
only factors and cross factors with significant p values remained. ANOVA was used in the
determination of the p values while Minitab was used to determine the regression
coefficients.
50
DISCUSSION
The discussion is divided into three sections. The first section discusses the
classical approach of comparing MATH test results by percent removal. This section
continues with a discussion on removal rate and removal coefficients as quantitative
measurements of cell surface hydrophobicity. Analysis of the percent removal data is also
approached arithmetically using graphs and standard deviation methods. The last section
discusses the statistical approaches used to analyze the data.
Analysis of Data
Percent Removal
The percent removal approach to analyzing the MATH test allows quick
comparisons of the results but these results are not definitive. The information gained is
limited to comparing the percent removal of bacteria by the hydrocarbon under various
conditions but with no standard measure for comparison. Table 12 illustrates the variation
in the data set by representing the data at the lowest settings to the highest for all three
experimental factors. The variation in percent removal ideal and control are illustrated. For
both absorbance and cell counts the response for the percent removal ideal is greater to start
with. The percent error is larger for the control than for the ideal analysis method for the
absorbance data. The cell count data is similar but with some exceptions had a lower
standard deviation. The cell count data also did not have as many data points so the
standard deviation is probably higher than its true value. From this representation of the
data the following observations can be stated:
51
1)
Error decreases as the factor settings increase from (60, 10, 0.2)* to the highest
settings of (140, 40, 1.0)*.
2)
Percent removal (control) error values are larger than those for percent removal
(ideal).
3)
For percent removal (ideal), the values at the highest settings show greater than
90% removal of the suspended cells.
4)
Standard deviations decrease as settings increase for all data types. This is due
largely to the near-100 % removals.
*Settings are agitation, hydrocarbon volume (ml).
Table 12. Percent absorbance data showing the change from the lowest settings to the
highest settings for absorbance and cell counts.
50.3 ± 1 3 .7 -8 1 .9 ± 0 .4
percent removal (control)
79.7 ± 7 .7 - 9 1 .5 ± 1.9
percent removal (ideal)
47.7 ± 14.4 - 78.7 ± 7.5
percent removal (control)
69.2± 8.3 -9 1 .7 ± 2 .8
43.9 ± NA - 97.8 ± 2.5
percent removal (ideal)
percent removal (control)
97.8 ± 0.2 - 99.7 ± 0.3
percent removal (ideal)
-22.6 ± 46.4 - 98.8 ± 0.4
percent removal (control)
52.7 ± 3 1 .4 - 9 9 .5 ± 0 .4
percent removal (ideal)
5:1
A b sorb an ce
15:1
5:1
C e ll c o u n t s
15:1
KineticApproach
In the MATH test the removal of bacteria by hydrocarbon is known to be a function
of time and hydrocarbon to water ratio. Lichtenberg et al. proposed a kinetic approach to
52
Removal Rate (k) for an absorbance setting of
140
y 4= -0.1457): + 1.14
1. 1 0 -2 1.05 -
y = -0.1757X + 1.0
e 1.00
Linear (0.2)
Linear (I)
Linear (0.6)
0.8 5 - -
........ ^1..^.5 .^(.l.)L...L.).
R; = 0.996 7
time
(min.)
Removal Coefficient (K) for Absorbance Data 5:1
at agitation settings of 60 and 140.
1.4 1.2
0 .8
...— •-3 .8688x + 0.5826
2 = 0.5147
■
---------------- '■■■■
...I
y = i-]V6456x ^ 0 5 f4 5 :“
0.4 ■
0.2
♦
«
60
140
-Linear (60)
- Linear (140)
1Linear (60)
■
Figure 10.
Plots used to determine the removal rate (k) and removal coefficient (K) for
cells grown in a culture containing a 5 :1 C:N ratio.
53
adherence of bacteria to hydrocarbon in 1985. The removal rate (k) was determined by
dividing initial absorbance by experimental absorbance. The log of the quotient, multiplied
by 100, was plotted against the time of agitation. The slope obtained was labeled k. Tb
determine the removal coefficient (K), removal rates for different volumes of hydrocarbon
(0.2,0.6, 1.0 in ml) multiplied by a factor of 2.3 were plotted against the V hAZw. The
slopes for removal coefficients obtained by Lichtenberg etal. (1985) were consistent with
those of Sharon etal. (1986).
The results for k in this study are similar to those determined by Lichtenberg and
Sharon. Two values for removal rate (k) from the Lichtenbefg data are: slope of 0.94 min'1
for 20 ]A and 0.6 min"1for 10 ]A volumes of hydrocarbon. Removal rates were calculated
for both C:N ratios for agitation settings of 60 and 140. The log of cells remaining was
plotted against the time of agitation in minutes. Values for the slope (k) in figure 10 are .26
min'1, . 15 m in'1 and . 17 min"1 for .2 ml, .6 ml and 1.0 ml of hydrocarbon respectively, at
an agitation rate of 140. Most of the results for k yield positive slopes with R2 values of .90
or better. One example, where the slope deviated from the norm, occurred for a C:N of
15:1 , 1.0 ml of hydrocarbon and an agitation setting of 140 (Figure 11). The slope was
positive, so that the value for k was negative and the linear regression produced an R2 value
of 0.23. For this case, the data obtained are not represented by the model and therefore k
should not be used to calculate the removal coefficient (K).
Lichtenberg calculated a K of 376 min"1 for S . pyogenes, while Sharon etal. cited
values of 4.7 min "1 for Streptococcus pyogenes and 72 min"1 for Acinobacter
calcoaceticus. These values are very different from those obtained from the absorbance
data for the isolated hydrocarbon-degrader. For cells grown at 5 :1 with an agitation setting
of 60, K was determined to be 3.8 min
Both Lichtenberg e ta l and Sharon etal. used
substantially lower hydrocarbon volumes, from 5 ^ l - 20 p \ of hydrocarbon to 3 ml of
54
Removal Rate (k) for absorbance data with an
agitaion setting of 140.
1.1
y = -0.2657 x + 1.12
R2 = 0 I? 75
L
1 .05
0
1
0. 2
1
< 0.95
<
^
% 0.9
y —-0.311 8x + 1.125
R2 = 0.9861
&
0.6
4
I
-Linear (I)
Linear (0.6)
"Linear (0.2)
y = 0.1543x - 0.8£
R2 = 0.23 )4
O
0.85
A
0.8
0. 2
0.4
time
0.6
0.8
(min.)
Removal Coefficient (K) for absorbance data at
agitation settings of 60, 100 and 140.
1. 3-
100
140
.............Linear (100)
'""""“ '"Linear (60)
------ -Linear (140)
*
A
M
0.5
0.3
Figure 11.
Plots used to determine the removal rate (k) and removal coefficient (K) for
cells raised with a 15:1 C:N ratio. Values for K were negative for the agitation rates of 60
100 and 140.
55
water. The reason for the low values compared to the results of Lichtenberg etal. may be
due to the hydrocarbon volumes chosen. A t low hydrocarbon-water ratios the response
would be influenced by the effects of a limited surface area for the bacteria to adhere to. On
the other hand, when a large volume of hydrocarbon is used, the effects of the hydrocarbon
surface area should be minimal. If this is true, then statistical analysis of the data should
indicate that hydrocarbon volume has little effect on the numbers of cells adhering to the
interface.
The experiments of Lichtenberg etal. and Sharon etal. used different vessels for
agitation of the hydrocarbon-water suspensions. Lichtenberg etal. used the conventional
test tube method, whereas Sharon etal. investigated the possibility of using polystyrene
cuvettes for agitation and absorbance measurements. They were interested in using cuvettes
because it would simplify the amount of time needed to run an experiment. Cuvettes did not
work for this research because the bacteria adhered to the sides of the cuvette. Another
problem was loss of hydrocarbon and water during the agitation step. The hydrocarbonwater mixture would leak out from under the cuvette cap invalidating the method.
Another important factor of the MATH assay is agitation rate. Lichtenberg etal.
repeated their experiments using a higher agitation rate which resulted in a similar removal
coefficient. The removal coefficient was 414 min ^ for S. pyogenes when subjected to a
higher agitation rate while maintaining all other parameters identical to previous
experiments. They state that it is possible that the removal coefficient is independent of
agitation rate (Lichtenberg 1985). This is not supported by results of this study. The effects
of agitation rate on the removal coefficients for both C:N ratios does not support that
conclusion.
In Figure 10 it is clear that K values for the two agitation rates are not similar. In
fact, they have opposite slopes. The same conclusion can be drawn from Figure 11. The
slopes for the three agitation rates are not the same. The negative slopes and low
56
coefficients could be the result of saturation with respect to hydrocarbon. The removal
coefficients at greater hydrocarbon-to-water ratios should all have similar slopes because
surface area is no longer limited with respect to the amount needed for a monolayer of cell
attachment. The slopes for the removal coefficients of Figure 11 range from -1.05 m in _1to
3.8 min
If the error bars were included, the slopes may not be statistically significantly
different. It is possible that the curve is actually flat for high hydrocarbon-to-water ratios.
In the case of low volumes of hydrocarbon there is a significant change in the surface area
available to the cells for increasing volumes of hydrocarbon. Removal coefficients for this
part of the curve should be much steeper than those for high hydrocarbon volumes. The
statistical analysis should indicate the relative importance of agitation rate and hydrocarbon
volume for a given response.
Clumping of the bacteria may possibly be affecting the results of the removal rate
and removal coefficient analysis. The tendency of these bacteria to aggregate may somehow
influence the removal of the bacteria. The affect of the clumping may cause increased cell
removal that is dependent on concentration. If the removal of cells was exponential then
most of the removal would take place in the first seconds of agitation. This would cause the
slopes to change little with increased mixing times. In most cases, percent removals were
greater then 70% after 10 seconds of mixing and did not increase by more the 20% with
continued mixing (Tables 6-9). It also looks like the C:N ratio may affect this part of the
experiment. For hydrocarbon volumes of 1.0 ml the change in absorbance (AR=% removal
ideal at 40 sec - % removal ideal at 10 sec) over the change in time (AT=40-10 sec). This
gives rates of .24, .17, .05 min"1 for C:N of 5:1 and .37, .35, .25 min"1 for C:N of 15:1.
The rate of change for the 5 :1 data is greater than the rate of change for 15:1 although the
variance in the data may not support this conclusion.
57
Graphical Analysis
A graphical representation including all the factors allows a comparison of the relative
importance of each factor on the response (Figure 12). The importance of all three
3D plot for absorbance of C:N 15:1
% Removal
(id ea l)
S
s
*
S
9
Time in
second s
agit rate / HC
Figure 12.
A three dimensional graph showing the effects of agitation setting,
hydrocarbon volume and time of mixing on the percent removal (ideal) for isolate grown at
a C:N ratio of 15:1 for absorbance.
of the factors is illustrated in Figure 12. The percent removal is lowest at low hydrocarbon
volumes, agitation settings and mixing time. As the agitation rate and volume of
hydrocarbon increase the percent removal (ideal) increases. The percent removal (ideal)
also increases as the mixing time changes from 10 seconds to 40 seconds. The effect of
time of mixing can be observed especially well for the hydrocarbon volume of 0.2 ml and
58
an agitation rate of 60. The other relationships are not nearly as clear. The graph shows
how little percent removal (ideal) changes at the higher settings.
Cell counts for percent removal (ideal) increase as the agitation rate and
hydrocarbon volume increase. A graph for percent removal (ideal) at two mixing times for
different agitation rates and hydrocarbon volumes illustrate how agitation rate and
hydrocarbon volume may influence adhesion of bacteria to hydrocarbon (Figure 13). The
error bars for this graph also show how error decreases with increased mixing time,
agitation setting and hydrocarbon volume. The two combined factors will influence the
percent adhesion to different degrees. Hydrocarbon volume and agitation setting are paired
in this representation and cannot be distinguished from each other. To determine the relative
influence each factor has on the experiment, a different method of analysis has to be used.
The percent removal increases the most as the hydrocarbon volume changes from 0.2 to
1.0. This may not be a real trend, it is likely that the agitation rate is affecting the removal
rate more then the change in the hydrocarbon volume. Also, the error bar for (60/.2) at 10
seconds is greater then the percent removal (ideal) for that point (Figure 13). This is also
true for the mixing time of 40 seconds, indicating that the data collected for small
hydrocarbon volumes and low agitation settings is not very consistent.
The methods described so far were not designed for statistical analysis, although
the experiment was designed with a particular statistical method in mind.
59
no
Percent removal for two mixing times
(10 and 40 sec.) for cell counts
at a C:N of 15:1.
100
■o 90
>
e 80
E
ft 70
L
K
60
50
40
00# . . . . . . . . . . . . I
§n.I . . . . IJIIj. . . . . . . . . . . . . h
i_l- - - h M -i- - 1- - - 1- - - h i Iffl—
110 sec.
I 40 sec.
I------- ,
M
5
O
kO
O
O
's.
VD
O
'■Jj
O
O
agitation
S
S
s e t tin g /H C
(N
5
O
O
Tt
O
Tt
volum e
Figure 13.
Graphical representation for percent removal (ideal) at two mixing times
with variation of agitation rate and hydrocarbon volume.
Statistical Analysis
The experiment was set up so the method of least squares (regression analysis)
could be applied to the results. Four factors: agitation rate, hydrocarbon volume, time of
mixing and C:N ratio influenced the responses, In (Ao/A), In (Co/C), In (Ac/A), In (Cc/C),
In (Ao/Ac) and In (Co/Cc). ANOVA (analysis of variance) was used to determine the
variation between experiments and the adjusted R2 and p values.
Regression
Step wise regressions for each of the factors, cross products and squared terms
were used to determine factors and cross factors important to the model. The data were
60
initially split up into three different groups: interference, agitation setting and agitation rate
(rpm) data. The regression results were used initially to determine whether rpm or agitation
setting as an indicator of mixing intensity, gave a better fit. The method of least squares
was used to determine if interference data should be included with the rest of the data set.
DataInclusion
First the regression data for agitation rate (rpm) and agitation settings were
compared. Analysis of the results of the regression determined that using rpm as the factor
for agitation rate produced a better fit than agitation settings. This is supported by the
higher R2 value (44%) calculated using rpm compared to R2 =39% if agitation is used.
Next, the response for the interference data were analyzed. The regression analysis
for the data with interference determined that it should not be included in the main data set.
The p values for time, HCrH2O, and C:N were all greater than 0.05 for the absorbance data
compared to the significant p values determined for the experiments without interference.
The R2 value of 9% indicated that very few of the points of the interference could be
explained by the model.
It is interesting that R2 for the cell counts in the interference data was significantly
higher than for absorbance. Absorbance may be more easily influenced by changing
conditions of the media and test tubes. Cell counts are not influenced by impurities which
affect the absorbance. If impurities are present which do not affect the removal of cells, but
cause skewed absorbance readings, the more likely it is that the regression will have
outliers. Cell counts for interference would however, only be affected by the initial
absorbance reading, which was set around 0.02 for all experiments. In the cases with
interference the initial cell counts were lower then those in normal experiments (Figure 14).
The error bars for interference for cell counts of 5 :1 indicate that the data for the inter­
ference should be included. As stated earlier, the statistical analysis did not support the
inclusion of interference results in the data pool. Another interesting observation is the
61
difference between the initial cell counts for the two culture types. The overlap of the error
bars indicates that the two data sets are not statistically different (Figure 14).
Initial Cell Count Comparison
2.50E+07
^ 2.00E+07
g 1.50E+07
S Ave. Init. cell counts
B Ave. Init. cell counts (int.)
w
I 1.00E+07
o
O
E 5.00E+06
*
o
0.00E+00
5:1
cu ltu re
1 5:1
co n d itio n s
Figure 14.
Initial cell counts for cultures raised at 5 :1 and 15:1 carbon to nitrogen ratios
including the cell counts for interference data.
Significant Factors
Regression analysis was used to fit data to a model. To determine the relative
importance of a factor to the model, the magnitude of the p value is compared to 0.05. For
values of p > 0.05, there is a very small chance that the factor is important to the model
(Table 13). The coefficients of the linear regression also indicate, indirectly, the relative
importance of each term.
p Values
The first response In (Ao/A) has no values of p that are insignificant to the
model. The factors time of mixing, agitation, hydrocarbon volume and C:N ratio are all
62
important to the model. For the response In (Co/C) the same is true. This is not the case of
the response In (Ac/A) since the coefficients for the factors of hydrocarbon volume and
C:N ratio are statistically significant (p > 0.05). This difference could be related to wall
effects in the control test tubes. If C:N influences the amount of cells which stick to the
glass the controls could have a different response to changes in C:N ratio. One explanation
for the lack of correlation with hydrocarbon volume is that the control absorbance (Ac)
were uneffected by hydrocarbon while the test values (A) were effected. This explanation is
not substantiated by the corresponding cell count regression where p < 0.05 for both
hydrocarbon and C:N.
Table 13.
The p values for the six responses.
response
time
agitation
HC vol.
C:N
Exp (CN)
In (Ao/A)
0.0001
0.0001
0.0001
0.0001
0.0001
0.0001
0.0041
0.0001
0.0001
0.0001
Z
0.0001
0.0001
0.0001
Sg H *
In (Co/C)
In(AcZA)
In Cc/C)
In (Ao/Ac)
In (Co/Cc)
ooooi
%k49&$
i e
F
0.0001
0.0007
*
"
”
0 ,46 }
:m
mm
0.0001
0.0001
0.0001
At different C:N ratios, the production of extracellular materials is likely to be different.
The 5 :1 raised cells could be producing a chemical that affects the absorbance of the
solution but not cell counts. This may account for the differences in significance of the p
value between the cell counts and the absorbance model responses. The last cases to be
examined are those of the initial over the control measurements. The p-values for the
hydrocarbon should be large because no hydrocarbon is actually involved in the
experiment. For the absorbance response only the time of mixing affects the response
63
significantly. This may be because as the mixing time increases, additional cell clumping
takes place. Absorbance values would decrease, but cell counts - which were performed
after dispersing the clumps - would not be. Agitation rate may cause more clumping as
more cells will collide as the rate increases. The more interactions the more likely it
becomes that the cells will stick together. C:N ratio may also play a role in clumping of
cells.
Significant Coefficients The relative importance of factors to the model can be compared
to the magnitude of the coefficients for factors in the model. The mean value of a factor
times its coefficient is one way of determining a coefficient’s effect on the model. The
results in Table 14 were calculated from the regression coefficients and the average value
for a given factor. For example, the agitation rate coefficient for In (Ao/A) is 0.0002877
and the average value for rpm is 2433. The adjusted value is equal to the product of the
terms or 0.70.
For the response In (Ao/A) all the p values are significant and all of the adjusted
coefficients are important to this model, so no terms can be removed. For In (Co/C) all of
the factors are important to the model. The next case, In (Ac/A), has high p values for the
HC volume, and C:N ratio. The adjusted coefficient contradicts the p value data for
hydrocarbon and C:N. Both of these factors have significant coefficients, but the p values
are greater than 0.05. One possible interpretation is that these factors contribute extensively
to the model responses, but that scatter in the data (leading to a low p value) produce low
confidence in the result. On the other hand, the coefficient/factor mean product for time is
only 0.094, while the corresponding p-value was 0.0001, indicating good fit, but a low net
affect of the time as a factor in this response. The results are inconclusive; some factors
may not differ significantly from zero (high p-value) and yet the regression coefficient
64
predicts a significant change in response for a change in that factor. For the response In
(C dC ) all the coefficients are important. This agrees with the results from the p value
Table 14.
Constant terms and Coefficients for Responses.
Response
Constant
rpm
Time
C:N
HC volume
In Ao/A
1.31
0.70
0.243
-0.625
0.362
In Co/C
1.98
2.62
0.801
-1.711
0.44
InAc/A
0.003
1.36
0.094
-0.123
0.413
In C dC
-0.73
2.12
-0.616
-0.701
0.446
In Ao/Ac
1.2
-0.617
0.177
-0.560
0.039
In Co/Cc
3.50
-0.054
-0.105
-0.813
0.088
analysis. The HC coefficients for In Ao/Ac and In Co/Cc are negligible and should not
influence the model. This is consistent with the previous finding from the p values. The
agitation coefficient for In Co/Cc is small enough to neglect, but this is not supported by the
p value.
Variation Between Experiments
The inter and intra experimental errors were determined using the statistical program
SAS for analysis of variance. The values for R2 were determined for the ANOVA model
based on each C:N ratio level as a separate experiment. Values for the relative interexperimental and intraexperimental error and the adjusted R2 values can be found in Table 15.
The results indicate that the responses not including the controls (In Ao/A), In(CoZC) have a
higher interexperiment error than intraexperiment error. The reverse is true in the responses
which include the control data. The worst case, control cell counts, has a value for inter
experimental error which is 26 times larger than the intra experimental error. One possible
explanation for this result is generation effects or possibly the age of the culture used.
65
Table 15.
Response
In (Ao/A)
In (Co/C)
In (A d A )
In (CdC)
In (Ao/Ac)
In (Co/Cc)
Values for Interexperimental and Intraexperimental Error.
Inter exp. error
0.105
0.780
0.069
0.230
0.016
1.390
Intra exp. error
0.060
0.260
0.074
0.530
0.026
0.052
R2
0.78
0.91
0.79
0.88
0.69
0.92
The R2 values are significantly higher when the experiment-to-experiment variation
is accounted for. Cell counts fit the model better than the absorbance measurements.
Conversely, the values for inter-experiment error are of the same order of magnitude as the
coefficient/factor products in Table 14. Thus, one could conclude that the effects of the test
and growth factors are no more significant than variations between experiments.
Means and Standard Error
Determination of the local and global variance and means with respect to test factors
was carried out using the following procedure: I) The data were first divided into two
groups; one for each C:N ratio. 2) The data were then sorted according to a single test
factor ( e.g., agitation setting) and then according to experiment (each batch of cells). 3)
Local means and variances were then determined within each experiment without regard to
the other factor settings (for example, the local mean and variance with respect to agitation
setting 60 in experiment I were determined without further segregation into time and
hydrocarbon volume). 4) A group mean and variance were then computed for all the
experiments at each C:N ratio and test factor value. 5) At each factor value, a mean and
variance was computed for the two C:N ratios combined. The result was a mean response
at that factor setting and a variance for all of the data at that factor setting. 6) Steps 2
through 5 of the above procedure were repeated for each of the other two test factors.
66
The results of this analysis are shown in the following figures. The first
two graphs show the results for the responses In (Ao/A) and In (Co/C) vs. hydrocarbon
volume (Figure 15). The graph shows that the average responses do not change much with
increasing hydrocarbon volumes. The error associated with the responses indicates that the
variation between experiments is greater than variation induced by changing hydrocarbon
volumes. Results of this analysis on the other four responses (which include the controls)
give essentially the same overall picture. In figure 16, the trends of the responses with
respect to agitation are more significant. The error bars do not overlap between the middle
and low settings but the high setting overlaps the middle setting. This indicates that at the
lowest setting, the effects of the error are smaller than the effect of agitation setting. The
agitation setting graphs would be similar to the agitation rate graphs (using rpm) except for
more overlap for the high agitation rates (2800 and 3100).
The last series is the response to time of mixing (Figure 17) . The error
bars again overlap the means of the average responses. There does seem to be a slight trend
toward increasing removal with longer mixing, but with the large overlap and small
changes in the response it can probably be safely assumed that the time of mixing does not
significantly change the results.
67
Average In (Ao/A) Response vs. Hydrocarbon Volume
values for standard error of the averages
1.2
I
1.1
?
1
<| 0.9
^
<
>
...................
^
j
f
< I
♦In Ao/A
I
I
_ _ _ _ _ I_ _ _ _ _
0.8
0.7
0.6
0
0.2
0.4
0.6
0.8
I
1.2
hydrocarbon volume (ml)
Average In (Co/C) response vs Hydrocarbon Volume
values for standard error of the averages
6 1.8
♦In Co/C
hydrocarbon volume (ml)
Figure 15. Graphs of the responses In (Ao/A) and In (Co/C) vs. hydrocarbon volume
including the standard error of the averages.
68
Average In (Co/C) Response vs Agitation Setting
values for standard error of the averages
9
a
2 2-1
2
_Ib_
-
jI
1.8 O
U
w I6 I4 -
Aln Co/C
3k
I2 I 4t)
63
80
100
120
140
160
agitation setting
Average In (Ao/A) Response vs Agitation Setting
values for standard error of the averages
19
_. .............
I I 4►
I .
i
I0
w
id
I
09 8 -
4I
07 -
06 05 4<3
6(3
80
100
agitation setting
120
140
160
♦InAo/A
Figure 16. Graphs o f the responses In (Ao/A) and In (Co/C) vs. agitation setting including
the standard error of the averages.
69
Average In (Co/C) Response vs Mixing Time
values for standard error of the averages
2.4 2.2 ■............. ............. ............ ............ ..............
2 U
^ lB U
............
F#-I1 11 6 - ...... ..... ^►
4►
............ 4 >•............
♦In Co/C
I4 I2 I
5
K)
U
---------- 1---------2(
25
3(
tinie of mixing (se c)
3/5
4()
4i
Average In (Ao/A) Response vs Mixing Time
values for the standard error of the averages
♦In Ao/A
time of mixing (sec)
Figure 17. Graphs of the responses In (Ao/A) and In (Co/C) vs. agitation setting including
the standard error of the averages.
70
CONCLUSION
Evaluation of the MATH test for reproducibility proved to be a complicated task.
The results were evaluated using two statistical approaches. ANOVA was used to
determine the interexperimental and intraexperimental errors. A regression analysis was
used to compare the significance of each of the test and growth factors on the response.
The results from these analyses, when compiled with the results from the conventional
methods (percent removal, removal rate and removal coefficients) were enough to draw
some conclusions. The test is not reproducible for the results obtained using the (ideal)
data. The interexperimental error was two times greater than the intraexperimental error.
The effects of the factors varied and in some cases there was a statistical significance but
not a practical significance.
For the given hypotheses in the introduction the following statements can be made:
1)
The MATH test does not yield reproducible results for the given set of test
conditions and growth conditions for the bacteria used in this study.
2)
The time of mixing will affect the results statistically although it is not of practical
significance.
3)
Mixing intensity has a practical and statistical significance to the results.
4)
The hydrocarbon volume is neither statistically significant or practically significant
for the hydrocarbon volumes used in this research.
5)
The C: N ratio is not a significant factor in the removal of cells from solution.
6)
Cells are removed when agitated without the presence of hydrocarbon.
71
REFERENCES
Brown, W.A., Cooper, D.G. (1992) Hydrocarbon Degradation by Acinetobacter
caJcoaceticus RAG-1 Using the Self-cycling Fermentation Technique. Biotechnology and
Bioengineering, 40, 797-805.
Busscher, H.J., Sjollema, J., van der Mei, H.C. (1990) Relative Importance of Surface
Free Energy as Measure of Hydrophobicity in Bacterial Adhesion to Solid Surfaces, in
Microbial C ellSurfaceHydropkobieity, Doyle, R.J. and Rosenberg, M., Eds., ASM
Publications, Washington, D C., pp 335.
Camper, A.K., LeChevallier, M.W., Broadway, S.C., McFeters, G.A. (1985) Evaluation
of procedures to desorb bacteria from granular activated carbon. Journalof
MircrobiologicalMethods, 3, 187-198.
Chapatwala, K.D., Newaz, M.S., Richardson, J.D., Walfram J.H. (1990) Isolation and
Characterization of Acetenitrile Utilizing Bacteria. Journal o f Industrial Microbiology.
Courtney, H.S., Hasty, D.L., Ofek, I. (1990) Hydrophobicity of Group A Streptococci
and Its Relationship to Adhesion of Streptococci to Host Cells, in Microbial CellSurface
Hydrophobicity, Doyle, R.J. and Rosenberg, M., Eds., ASM Publications, Washington,
D.C., pp 361.
Duncan-Hewitt, W.C. (1990) Nature of the Hydrophobic Effect. inM icrobial CellSurface
Hydrophobicity, Doyle, R.J. and Rosenberg, M., Eds., ASM Publications, Washington,
D.C., pp 39.
Ener, B., Douglas, L.J. (1992) Correlation between cell-surface hydrophobicity of
Candidaalbicans and adhesion to buccal epithelial cells. FEMSMicrobiology Letters, 99,
Escher, A. R. (1986) Colonization of a Smooth Surface by Pseudomonas aeruginosa:
Image Analysis Methods. PH.D. Thesis, Montana State University.
Fattom, A., Shilo, M. (1985) Production of emulcyan by Phormidium J-l: its activity and
function. FEMS Microbiology Ecology, 31, 3-9.
Funk, H.B., Krulwich T.A. (1964) Preparation of Clear Silica Gels than Can Be Streaked.
J Bacteriology, 88, 1200-1201.
Goswami, PI, Singh, H. (1991) Different Modes of Hydrocargon Uptake by Two
Psuedomonas Species. Biotechnology and Bioengineering, 37, 1-11.
Grcmp Report (1984) Mechanisms of Adhesion. In Microbial Adhesion and Aggregation,
Inouye, M. (1987) Bacterial OuterM embranesAsM odelSystems. John Wiley & Sons,
72
Kjelleberg, S. (1984) Adhesion to Inanimate Surfaces. InM icrobialAdhesionand
Aggregation, Ed K.C. Marshall, 51-70.
Klotz, S. ( 1990) Role of Hydrophobic Interactions in Microbial Adhesion to Plastics Used
in Medical Devices. In Microbial Cell SurfaceHydrophobicity, Doyle, R.J. and
Rosenberg, M., Eds., ASM Publications, Washington, D.C., 107.
Lachica, R.V. (1990) Significance of Hydrophobicity in the Adhesiveness of Pathogenic
Gram Negative Bacteria, in Microbial Cell Surface Hydrophobicity, Doyle, R.J. and
Rosenberg, M., Eds., ASM Publications, Washington, D.C., 297.
Lichtenberg, D., Rosenberg, M., Sharfman, N., Ofek, I. (1985) A Kinetic approach to
bacterial adherence to hydrocarbon. JournalMicrobiologyMethods, 4, 141-146.
Marshall K. C. (1976) Interfaces in Microbial Ecology. Harvard University Press,
Cambridge MA, 11-21.
Marshall, K. C., Cruickshank, R. H. (1973) Cell Surface Hydrophobicity and Orientation
of Certain Bacteria at Interfaces. Arch. Mikrobiol., 91, 29-40.
Mudd, S., Mudd, E. B. H. (1924 a) The penetration of bacteria through capillary spaces.
IV. A kinetic mechanism in interfaces. J. Exp. Med. 40, 633-645.
Mudd, S., Mudd, E. B. H. (1924 b) Certain interfacial tension relations and the behavior
of bacteria in films. J. Exp. Med., 40, 647-660.
Mueller, R. F. (1990) Characterization of Initial Events of Bacterial Colonization at SolidWater Interfaces Using Image Analysis.
M.S. Thesis, Montana State University.
Pines, O., Shoham, Y., et. al. (1988) Unmasking of surface components by removal of
cell-associated emulsan from Acinobacter_ Sp. RA G-1. Appl. Microbiol. Biotechnol., 28,
Prokok, A., Ludvik, M., Erickson, L.E. (1972) Growth Models of Cultures with Two
Liquid Phases. VIII. Experimental Observations of Droplet Size and Interfacial Area.
Biotechnology and Bioengineering, Vol XIV, 587-608.
Rittman, B.E. ( 1989) Detachmentfrom Biofilms, in Structure and Function o f Biofilms,
Characklis, W.G., and Wilderer, R A ., Eds., 49-58.
Rosenberg, E., Legmann, R., et. al. (1992) Petroleum Bioremediation - a Multiphase
Problem. Biodegradation, 3, 337-350.
Rosenberg, E. (1992) The Hydrocarbon-Oxidizing Bacteria. In Balows, A. (Ed) The
Prokaryotes (pp. 446-459) Springer-Verlag, Berlin.
Rosenberg, E.,Shabtai, Y. (1987) In Bacterial Outer Membranes as Model Systems.
Masayori Inouye, Ed. (John Wiley & Sons, New York), pp. 311-342.
73
Rosenberg, E , Kaplan, N., Pines, O., Rosenberg, M., Gutnick, D. (1983) Capsular
Polysaccharides Interference with Adherence of Acinetobacter calcoaceticus to
Hydrocarbon. FEMS Microbiology Letters , 17, 157-160.
Rosenberg, M. (1991) Basic and Applied Aspects of Microbial Adhesion at the
Hydrocarbon: Water Interface. CriticalReviewsinMicrobiology, 18, No. 2,159-173.
Rosenberg, M., Doyle, R.J. (1990) Microbial Cell Surface Hydrophobicity: History,
measurement, and significance. In Microbial Cell Surface Hydrophobicity, Doyle, R.J. and
Rosenberg, M., Eds., ASM Publications, Washington, D.C., pp I.
Rosenberg, M. (1984) Ammonium Sulphate Enhances Adherence of Escherechiacoli J-5
to Hydrocarbon and Polystyrene., FEM SMicrobiology Letters, 25, 41-45.
Rosenberg, M., Bayerm, E., Delarea, J., Rosenberg, E. (1982) Role of Thin Fimbriae in
Adherence and Growth of Acinetobactercalcoaceticus RAG-1 on Hexadecane. A p p l and
EnvironmentalMicrobiology., 44, No. 4, 929-937.
Rosenberg, M., Gutnick, D., Rosenberg, E. (1980) Adherence of Bacteria to
Hydrocarbons: A Simple Method for Measuring Cell-Surface Hydrophobicity. FEMS
Microbiology Letters, 9, 29-33.
Sar, N. (1987) Direction of spreading(DOS): a Simple Method for Measuring the
Hydrophobicity of Bacterial Lawns. J. Microbiology Methods, 6, 211-219.
Sar, N., Rosenberg, E. (1987) Fish Skin Bacteria: Colonial and Cellular Hydrophobicity.
Microbial Ecology, 13, 193-202.
Shabtai, Y ., Wang, D. I. C. (1990) Production of Emulsan in a Fermentation Process
Using Soybean Oil (SBO) in a Carbon-Nitrogen Coordinated Feed. Biotechnology and
Bioengineering, 35, 753-765.
Sharon, D., Bar-Ness, R., Rosenberg, M. (1986) Measurement of the Kinetics of
Bacterml Adherence to Hexadecane in Polystyrene Cuvettes. FEMSMicrobiology Letters,
Stolzenback, K.D. (1989) Particle Transport and Attachment, in Structure and Function o f
Biofilms, Characklis, W.G., and Wilderer, R A ., Eds., pp 33-47.
Sweet, S.P., MacFarlane, T.W., Samaranayake, L.P. (1987) Determination of the Cell
Surface Hydrophobicity of Oral Bacteria Using a Modified Hydrocarbon Adherence
Method. FEMS Microbiology Letters, 48, 159-163.
Wadstrom, T. (1990) Hydrophobic Characteristics of Staphylococci: Role of Surface
Structures and Role in Adhesion and Host Colonization, in Microbial CellSurface
Hydrophobicity, Doyle, R.J. and Rosenberg, M., Eds., ASM Publications, Washington,
D.C., pp 315.
Wang, Sy-Dar, Wang, D. I. C. (1990) Mechanisms for Biopolymer Accumulation in
Immobilized Acinetobacter calcoaceticus System. Biotechnology and Bioengineering, 36,
402-410.
74
Yu, F.P., Pyle, B.H., McFeters, G.A. (1993) A Direct Viable Count Method for the
Enumeration of Attached Bacteria and Assessment of Biofilm Disinfection. Journal o f
Microbiological Methods, 17, 167-180.
APPENDICES
76
APPENDIX A
EXPERIMENTAL RESULTS
77
Experiment I
Date: 7/13/94
C:N ratio: 15:1
V
Results
Time in Seconds
Agit rat HC vol.
10
20
60
0.2
0.071 0.055
60
I
0.029 0.029
100
0.6
0.02
0.011
140
0.2
0.017 0.013
140
I
0.017 0.017
plate
I
2
3
4
5
6
7
8
9
10
11
40
60
0.032 0.032
0.03
0.012 0.015
0.013
0.013 0.014
Control
Time in seconds
Agitatic
10
20
40
60 0.127
100
0.08
140
0.07
Dil. fact. ToL CC Std. dev.
0.00E+00 0.000
1.00E+05 5.26E+06 0.148
1.00E+05 3.76E+06 0.107
1.00E+05 4.62E+06 0.069
1.00E+05 3.48E+06 0.261
1.00E+05 1.46E+06 0.092
1.00E+05 1.00E+06 0.339
1.00E+04 1.06E+05 0.398
1.00E+06 1.24E+07 0354
0.00E+00 0.000
1.00E+05 6.22E+06 0.194
Cell Count data
agit. rt,
Plate counts, where each count represents a one ]a\. sample,
he vol, time
1
2
3
4
5
0
I Ink. abs. .181
2 .097 (0.5)
55
62
55
50
41
3 .048 (0.25)
39
32
42
35
40
4 60/1/10
47
48
50
42
44
5 60/1/40
40
21
41
42
30
6 100/.6/20
14
16
14
16
13
7 140/1/10
15
10
11
6
8
8 140/1/40
14
8
16
6
9
9 60/10
14
17
15
10
6
10
11 140/40
69
51
61
79
51
Experiment 2
Date: 7/14/94
C:N ratio: 5:1
Low initial abs.
Agitatic
60
100
140
I
2
3
4
5
6
7
8
52.6
37.6
46.2
34.8
14.6
10
10.6
12.4
Time in Seconds
10
20
40
0.024 0.012 0.014
0.022 0.012
0.01
0
0
0
0.001 0.005 0.004
0
0
0
Time in seconds
10
20
0.095
0.083
0.063
Plate counts, where
I
2
10
6
12
12
21
19
38
33
21
19
65
66
15
17
24
1
2
3
4
5
6
7
8
Dil. fact.
1.00E+06
1.00E+06
1.00E+05
1.00E+05
1.00E+05
1.00E+05
1.00E+04
1.00E+04
T o tC C
Std. dev.
8.20E+06 0.304
8.00E+06 0.468
2.20E+06 0.227
3.34E+06 0.168
2.04E+06 0.285
6.52E+06 0.076
O.OOE+OO 0.000
2.04E+05 0.265
40
each count represents
3
4
5
5
10
10
6
6
4
30
17
23
39
32
25
30
17
15
70
68
57
18
7.765
4.037
3.194
9.094
1342
3391
4.219
4393
62.2 12.050
plate
Agitatio HC volt
60
0.2
60
I
100
0.6
140
0.2
140
I
Control
agit. rate/hc
Initial .165
.06 (0.5)
.041 (0.25)
60/.2/10
60/1/40
100/. 6/20
140/.2/10
140/.2/40
Averag Std. dev
28
10 Averag'
8.2
8
22
33.4
20.4
65.2
20.4
Std. dev
2.490
3.742
5.000
5.595
5.814
4.970
5.413
78
Experiment 3
Results
Date: 7/11/94
Agitatio HC voli
60
0.2
100
0.6
140
I
C:N ratio: 5:1
Time in Seconds
10
20
40
0.045 0.024 0.023
0.022 0.016 0.022
0.026 0.015 0.019
Low initial abs.
plate
80
0.024
0.022
0.021
160
0.026
0.02
0.022
Control
Agitatic
60
Time in seconds
10
20
I
2
3
4
5
6
7
8
Dil. fact.
1.00E+06
1.00E+05
1.00E+05
1.00E+04
1.00E+04
1.00E+04
1.00E+04
1.00E+04
ToL CC Std. dev.
2.16E+07 0.862
9.60E+05 0.519
7.60E+05 0.498
1.36E+05 0.377
1.90E+05 0.368
1.92E+05 0.251
7.40E+04 0.365
3.80E+04 0.220
100
140
Cell Count data
I
2
3
4
5
6
7
8
Plate counts,
agit. rate/he ■
I
initial abs .0.
47
Dil. (0.5) 0.0
6
60/.2/10
14
60/.2/160
19
100/.6/40
15
140/1/10
22
140/1/40
12
140/1/160
3
where each count represents
2
3
4
5
35
6
15
5
6
6
16
14
6
5
8
5
7
10
18
14
26
26
10
18
16
26
14
18
6
7
5
7
4
3
4
5
Experiment 4
Results
Date 7/15/94
Agitatio HC voh
60
0.2
60
I
100
0.2
100
I
140
I
C:N ratio: 15:1
V
Time in Seconds
10
20
0.063
0.05
0.022 0.026
0.018 0.018
0.018 0.018
0.019
0.02
plate
40
0.033
0.025
0.016
0.022
0.022
Control
Agitatic
60
100
140
Time in seconds
10
20
0.111
0.069
0.095
10 Averag Std. dev
21.6 18.623
9.6 4.980
7.6 3.782
13.6 5.128
19 7.000
19.2 4.817
7.4 2.702
3.8 0.837
40
0.074
0.057
1
2
3
4
5
6
7
8
9
10
11
12
Dil. fact.
1.00E+06
1.00E+05
1.00E+05
1.00E+04
1.00E+05
1.00E+04
1.00E+04
1.00E+04
1.00E+03
1.00E+05
1.00E+05
1.00E+05
ToL CC Std. dev.
7.60E+06
0.996
2.60E+06
0341
8.20E+05
0380
4.50E+05
0.236
1.08E+06
0.327
1.00E+05
0.430
1.74E+05
0.242
9.20E+04
0.142
4.02E+04
0.164
6.60E+05
0.173
6.80E+05
0.263
6.20E+05
0.629
Cell Count data
I
2
3
4
5
6
7
8
9
10
11
12
Plate counts, where each count represents
agit. rate/hc ■
I
2
3
4
5
IniL Abs. 0.1
6
4
21
4
3
abs. .130 dil.
19
18
22
34
37
abs. .072 dil.
6
5
8
13
9
abs. .042 dil.
33
36
47
50
59
60/.2/10
8
8
9
13
10
60/1/40
6
12
15
12
5
100/. 2/20
22
20
18
16
11
100/1/20
11
9
8
8
10
140/1/20
33
34
41
45
48
60/10
5
6
7
7
8
100/20
9
8
7
5
5
140/40
2
2
8
9
10
10 Averag
0
7.6
26
8.2
45
10.83
10
17.4
9.2
40.2
6.6
6.8
6.2
Std. dev
7.570
8.860
3.114
10.607
3.545
4.301
4.219
1.304
6.611
1.140
1.789
3.899
79
Experiment 5
Date: 7/16/95
C:N ratio 5:1
Low initial abs.
Init abs. Aggit./I IOs
20s
40s
Oil.
60/.2
0.033 0.033
0.03
1/2
-I 100/.2
0.024
0.02 0.013
1/4
i( 100/1
0.015 0.008 0.008
140/.2
0.019 0.011
0.01
Control
Aggit. 1 10s
20s
40s
60
0.092
100
0.088
140 0.086
Experiment 6
Results
Date: 7/18/94
Agitatio HC volt
60
0.6
100
0.6
140
0.2
140
0.6
140
I
C:N ratio: 15:1
V
Time in Seconds
10
20
0.035 0.032
0.024 0.017
0.018 0.016
0.02 0.018
0.016 0.018
plate
40
0.031
0.016
0.016
0.018
0.016
Control
Agitatic
60
100
140
Time in seconds
10
20
0.102
0.097
0.087
40
0.085
I
2
3
4
5
6
7
8
9
10
11
12
Dil. fact.
1.00E+05
1.00E+05
1.00E+05
1.00E+04
1.00E+04
1.00E+04
1.00E+04
1.00E+04
1.00E+03
1.00E+04
1.00E+05
1.00E+05
Cell Count data
I
2
3
4
5
6
7
8
9
10
11
12
agit. rate/hc'
Init abs. 0.1!
0.75
0.5
0.25
60/.6/20
100/. 6/20
140/.2/20
140/.6/20
140/1/20
60/20
100/20
140/20
Plate counts, where each count represents a one u\. sample.
I
2
3
4
5
6
7
21
18
22
26
28
20
32
12
12
20
16
25
9
17
9
13
19
11
17
19
6
50
46
55
33
54
63
38
38
42
36
37
23
44
44
14
12
6
15
17
21
11
5
22
12
10
6
11
10
10
6
13
10
13
9
10
25
27
32
24
28
27
30
38
41
39
45
57
59
54
11
15
11
9
8
I6
9
13
9
14
10
14
14
12
8
19
15
11
53
62
16
15
8
33
53
8
8
T o tC C Std. dev.
2.45E+06
0.209
1.59E+06
0.312
1.24E+06
0.401
4.87E+05
0.195
4.05E+05
0.240
1.27E+05
0382
1.13E+05
0.419
1.05E+05
0.239
2.84E+04
0.101
4.73E+05
0.163
9.50E+05
0.273
1.20E+06
0.219
0
0
0
0
0
0
9
28
12
5
39
42
9
10
14
28
43
7
16
10 Averag Std. dev
31
24.5 5.126
21
15.9 4.954
14
12.4 4.971
56
48.7 9.476
37
40.5 9.710
6
12.7 4.855
12
113 4.739
12
10.5 2.506
30
28.4 2.875
44
473 7.732
11
9.5 2.593
10
12 2625
80
Experiment 7
Date: 7/19/94
C:N ratio: 5:1
Results
Agitatio
60
60
100
plate
Time in Seconds
HC volt
10
20
40
0.6 0.038 0.032 0.034
0.2 0.048 0.034 0.029
0.6 0.032 0.038 0.038
I
2
3
4
5
6
7
8
9
10
11
12
Interference
Control
Agitatic
60
100
140
Time in seconds
10
20
0.094
0.099
0.054
40
Dil. fact.
1.00E+05
1.00E+04
1.00E+05
1.00E+04
1.00E+04
1.00E+04
1.00E+03
1.00E+03
1.00E+03
1.00E+04
1.00E+04
1.00E+04
T o tC C Std. dev.
1.27E+06
0.324
4.46E+05
0.182
3.85E+06
0.314
2.43E+05
0.254
1.89E+05
0.077
1.10E+05
0.249
4.50E+04
0.093
4.54E+04
0.084
3.29E+04
0.083
3.00E+05
0.193
3.80E+05
0.181
2.38E+05
0345
Cell Count data
I
2
3
4
5
6
7
8
9
10
11
12
agit. rate/hc'
Inif abs. .181
abs. .112 dil
abs. .044 dil
60/.6/10
60/.2/10
100/.6/20
100/. 6/40
140/.6/10
140/.6/20
60/20
100/20
140/20
Experiment 8
Date: 7/21/1994
C:N ratio: 15:1
V
Plate counts, where
I
2
12
17
43
39
37
38
16
23
22
18
12
8
42
42
43
50
32
36
35
29
43
33
23
24
each count represents a one /d. sample.
3
4
5
6
7
13
16
19
8
15
53
34
49
40
32
45
58
57
37
24
34
17
25
18
31
19
20
19
18
17
12
13
16
10
8
51
45
41
41
47
41
45
45
51
50
35
36
36
31
28
35
23
39
23
26
31
34
37
35
35
15
20
16
19
17
Results
Agitatio
60
60
100
140
140
plate
Time in Seconds
HC volt
10
20
40
0.2 0.038 0.024 0.025
I 0.022 0.025 0.014
0.6 0.024 0.012 0.022
0.2 0.013 0.011 0.012
I 0.016 0.019 0.012
1
2
3
4
5
6
7
8
9
10
11
12
Control
Agitatic
60
100
140
Time in seconds
10
20
0 11
0.078
0.102
40
8
12
51
26
28
19
12
52
40
32
28
54
39
9
8
55
38
29
18
8
44
44
32
26
43
33
10
7
50
25
22
45
31
36
35
32
Averag'
12.700
44.600
38.500
24.300
18.889
11.000
45.000
45.400
32.900
30.000
38.000
23.800
Std. dev
4.111
8.127
12.104
6.183
1.453
2.739
4.183
3.806
2.726
5.793
6.864
8.203
Dil. fact. T o tC C Std. dev.
1.00E+06 1.35E+07
0305
1.00E+05 1.39E+06
0.259
0.00E+00
0.000
1.00E+04 2.23E+05
0.238
1.00E+04 6.33E+05
0.125
1.00E+04 7.62E+05
0.102
1.00E+04 2.23E+05
0.146
1.00E+04 2.29E+05
0.133
1.00E+04 9.25E+04
0.257
1.00E+05 1.21E+06
0.281
1.00E+05 1.18E+06
0.211
1.00E+05 1.14E+06
0.284
0.075
Cell Count data
I
2
3
4
5
6
7
8
9
10
11
12
Plate counts,
agit. rate/hc ■
I
Init dil. .2
14
0.75
11
abs. .094 dil. 1/2
abs. .048 dil.
17
60/.2/20
61
60/1/20
74
100/. 6/20
20
140/. 2/20
21
140/1/20
10
60/40
13
100/20
13
140/10
15
where each count represents
2
3
4
5
7
11
17
21
12
10
17
19
20
62
85
24
26
8
12
11
6
26
63
83
29
22
13
10
11
7
20
68
69
20
23
10
17
10
12
31
78
64
24
23
11
6
11
14
6
11
10
7
12
25
67
83
20
18
9
15
18
48
73
11
8
10
8
17
17
9
15
16
28
60
70
20
28
16
63
74
5
8
12
15
16
14
9
21
22
15
12
10 Averag' Std. dev
10 13.50 4.12
17 13.90 3.60
87
12.11
8
14
22.33
63.33
76.20
22.25
22.88
9.25
12
11.78
11.40
5.32
7.94
7.79
3.24
3.04
2.38
3.41
2.49
3.24
81
Experiment 9
7/26/94
Init abs. .188
Dil. Abs.
1/2
0.076
1/4
0.033
C:N 15:1
V
Control
Experiment 10
Date: 8/28/94
C:N ratio: 5:1
V
Aggit. E 10s
60
100 0.081
140
AggiUl60/.2
60/1
100/.2
140/.6
10s
0.055
0.041
0.013
0.013
20s
0.091
40s
20s
0.031
0.031
0.013
0.008
0.065
0.091
Results
Agitatio
60
140
140
140
40s
0.031
0.022
0.013
0.018
plate
Time in Seconds
HC volt
10
20
40
0.2 0.029 0.026 0.031
0.6 0.026 0.026 0.021
0.6
0.028
I 0.024 0.029 0.021
Control
Agitatic
60
Time in seconds
10
20
100
140
0.11
40
0.102
I
2
3
4
5
6
7
8
9
10
11
12
Dil. fact.
1.00E+06
1.00E+05
1.00E+04
1.00E+04
1.00E+04
1.00E+04
1.00E+04
1.00E+04
1.00E+05
1.00E+05
1.00E+05
T ot CC Std. dev.
1.49E+07
0.303
2.08E+06
0.262
3.84E+05
0.101
3.50E+05
0.139
2.28E+05
0.193
0.00E+00
0.000
8.20E+04
0.249
5.50E+04
0324
7.60E+04
0.299
7.90E+05
0.295
1.50E+06
0.290
1.39E+06
0.291
0.09
Cell Count data
7
8
9
10
11
12
Plate counts, where each count represents a one pi. sample.
agit. rate/hc 1
1
2
3
4
5
6
7
Init abs. .20
12
19
13
12
13
21
15
abs. .088 dil.
25
28
18
17
26
13
27
abs. .042 dil.
43
46
37
33
39
40
38
60/.2/10
26
34
37
31
38
33
41
60/.2/40
16
17
21
25
28
22
20
140/.6/10
140/.6/40
5
6
7
10
11
7
9
140/1/10
3
7
4
9
5
5
7
140/1/40
5
10
8
7
10
4
10
60/40
10
6
11
9
8
10
9
100/20
17
10
10
16
19
19
15
140/10
8
20
10
13
21
12
15
Experiment 11
9/3/94
Init abs. .197 AggiVP
Dil.
Abs 60/0.2
C:N 5:1 1/2
0.083 60/1.0
1/4
0.027100/0.6
V
140/0.2
140/1
10s
20s
40s
0.06 0.043 0.036
0.05 0.048 0.032
0.034 0.027 0.022
0.042 0.029 0.029
0.026 0.019 0.019
Control
AggiL E 10s
20s
40s
60 0.101
100
0.1
140
0.09
8
21
21
37
34
26
8
9
8
19
34
41
24
10 Averag Std. dev
14.89
4.51
14 20.80
5.45
37 38.40
3.89
35.00
4.85
22.80
4.39
10
8
4
8
4
11
5
9
7
12
14
14
22
12
6
5
5
8.20
5.50
7.60
7.90
15.00
13.90
2.04
1.78
2.27
2.33
4.35
4.04
82
Experiment 12
Date: 9/8/94
C:N ratio: 5:1
Interference
Results
Time in Seconds
10
20
40
0.094 0.094 0.093
0.087
0.08 0.078
0.076 0.078 0.071
Agitatio HC volt
60
0.2
100
0.6
140
I
Control
Time in seconds
Agitatio
10
20
40
60 0.127
100
0 11
140
0.121
Cell Count data
I
2
3
4
5
6
7
8
9
10
11
12
agit. rate/hc'
Init. abs. .19!
abs. .094 dil.
abs. .058 dil.
60/.2/10
60/.2/40
100/. 6/20
140/1/10
140/1/40
60/10
100/20
140/40
100/. 6/20
Experiment 13
Date: 9/10/94
C:N ratio: 5:1
V
Plate counts, where
I
2
11
10
24
31
15
27
29
30
54
53
14
16
6
4
7
5
19
37
32
21
25
22
13
12
Time in Seconds
HC volt
10
20
40
0.2 0.028 0.014 0.013
0.6 0.026 0.016 0.014
I 0.014 0.012 0.012
Control
Agitatic
60
100
140
I
2
3
4
5
6
7
8
9
10
11
12
Dil. fact.
1.00E+05
1.00E+04
1.00E+04
1.00E+04
1.00E+03
1.00E+03
1.00E+03
1.00E+03
1.00E+04
1.00E+04
1.00E+04
1.00E+03
each count represents a one /<1. sample.
3
4
5
6
7
11
13
12
7
12
25
31
29
23
26
18
22
20
21
16
30
29
29
36
28
44
58
45
44
55
16
10
10
13
13
4
4
4
4
8
5
5
5
7
6
18
15
31
35
35
20
13
27
16
17
25
24
24
23
22
13
11
15
15
18
Results
Agitatio
60
10
140
plate
Time in seconds
10
20
0.07
0.072
40
plate
I
2
3
4
5
6
7
8
9
10
11
8
18
29
18
25
55
13
5
6
23
20
25
10
Dil. fact.
1.00E+05
1.00E+04
1.00E+04
1.00E+04
1.00E+04
1.00E+03
1.00E+03
1.00E+04
1.00E+04
1.00E+05
T o tC C Std. dev.
1.21E+06
0.235
2.78E+05
0.138
1.97E+05
0.173
2.95E+05
0.104
4.94E+04
0.126
1.33E+04
0.170
5.11E+03
0.301
5.67E+03
0.153
2.76E+05
0.292
2.06E+05
0.279
2.33E+05
0.128
1.35E+04
0.219
9
14
25
19
40
16
7
5
31
24
27
18
10 Averag Std. dev
13 12.10
2.85
35 27.80
3.82
21 19.70
3.40
29.50
3.07
46 49.40
6.22
12 13.30
2.26
5.11
1.54
5.67
0.87
32 27.60
8.07
16 20.60
5.74
16 23.30
2.98
10 13.50
2.95
T o tC C Std. dev.
0.00E+0O
0.000
9.70E+05
0.296
2.78E+05
0.159
3.40E+05
0.157
2.13E+05
0.147
1.14E+05
0.212
8.44E+03
0.272
1.06E+04
0.393
3.26E+05
0.138
3.87E+05
0.131
1.04E+06
0.334
0.08
Cell Count data
1
2
3
4
5
6
7
8
9
10
11
Plate counts, where each count represents a one ]A. sample
agit. rate/hc
I
2
3
4
5
6
7
Init. abs. .198
abs. .098 dil.
13
7
6
15
10
11
8
abs. .033 dil.
30
24
26
21
33
26
26
60/.2/10
36
26
39
36
40
30
25
60/.2/40
18
16
19
21
23
21
22
100/ . 6/20
8
8
11
12
10
12
11
140/1/10
7
7
5
7
10
11
7
140/1/40
6
8
10
9
8
7
10
60/10
35
36
33
37
32
28
33
100/20
46
41
31
40
44
40
39
140/40
9
10
11
5
15
13
10
10 Averag Std. dev
11
29
35
27
16
11
16
23
33
7
7
35
39
22
13
11
13
36
41
16
9
34
24
13
19
32
8
9.70
27.78
34.00
21.30
11.40
8.44
10.60
32.56
38.70
10.40
2.87
4.41
5.33
3.13
2.41
2.30
4.17
4.50
5.08
3.47
83
Experiment 14
9/11/94
C :N 15:1
Interference
Init abs. Aggit./! IOs
20s
40s
DiL
60/0.2
0.048 0.044 0.036
1/2
(60/1.0
0.052 0.046
1/4
( 100/0.6 0.049 0.044 0.033
140/1
0.044 0.036 0.036
Control
Aggit. E IOs
20s
40s
60 0.132
100
0.095
140
0.092
Experiment 15
Date: 9/13/94
C:N ratio: 5:1
Results
Agitatio
60
100
140
plate
Time in Seconds
HC volt
10
20
40
0.2 0.093 0.088 0.081
0.6 0.084 0.079 0.081
I 0.079 0.083 0.082
Interference
Control
Agitatic
60
100
140
Time in seconds
10
20
0.183
0.189
0.143
1
2
3
4
5
6
7
8
Dil. fact.
5.00E+04
1.00E+04
I .OOE+04
1.00E+04
1.00E+04
1.00E+04
1.00E+03
1.00E+03
ToL CC Std. dev.
7.70E+05
0.144
4.39E+05
0.148
3.46E+05
0.133
2.46E+05
0.248
1.26E+05
0.160
6.70E+04
0.282
4.86E+04
0.126
1.95E+04
0.194
7
15
44
31
8
15
44
36
22
22
9
15
32
28
14
13
6
55
25
5
45
15
46
23
40
0.192
Cell Count data
agit. rate/hc
I abs. .190
1/2 abs .075
1/4 abs .027
60/.2/10
60/.2/40
100/.6/20
140/1/10
140/1/10
Plate counts, where each count represents
I
2
3
4
5
16
20
15
15
11
40
41
40
48
50
29
36
36
39
34
26
29
33
32
20
11
14
13
11
14
5
6
6
8
7
43
41
54
51
47
20
23
18
13
17
6
17
56
42
23
14
8
44
20
10
10
11
10 Averag Std. dev
15 15.40
2.22
44 43.90
6.51
34.56
4.59
24.56
6.09
16 12.60
2.01
5
6.70
1.89
60 48.60
6.13
21 19.50
3.78
84
Experiment 16
Date: 9/14/94
C:N ratio: 5:1
Interference
Results
plate
Time in Seconds
Agitatio HC volt
10
20
40
60
I
0.06 0.054
0.05
100 0.6 0.054 0.055 0.05
140
0.2 0.053 0.053 0.052
140
I 0.054
Control
Agitatic
60
100
140
Time in seconds
10
20
0.14
0.104
40
DiI. fact. T o t CC Std. dev.
1.00E+06 1.94E+07
0.152
1.00E+04 5.12E+05
0.123
I
2
3
4
5
6
7
8
9
10
11
12
1.00E+04
1.00E+04
1.00E+04
1.00E+04
1.00E+03
1.00E+03
1.00E+05
1.00E+05
1.00E+05
2.51E+05
1.80E+05
7.25E+04
1.21E+05
3.15E+04
1.80E+04
2.73E+06
1.44E+06
1.04E+06
0.105
0.185
0.293
0.215
0.153
0.296
0.111
0.240
0.271
0.108
Cell Count data
I
2
3
4
5
6
7
8
9
10
11
12
Plate counts, where each count represents a one n\. sample.
agit. rate/hc ■
I
2
3
4
5
6
7
I abs. .204
18
17
26
19
18
21
21
1/2 abs. 0.04
39
59
55
50
52
58
46
1/4 abs. .028
60/1/10
27
22
24
28
26
26
22
60/1/40
15
22
15
17
24
22
17
100/. 6/20
11
5
8
9
6
6
5
140/.2/10
16
12
11
12
13
16
13
140/.2/40
29
21
29
35
34
31
29
140/1/10
26
24
22
20
13
13
15
60/10
22
32
29
30
28
24
26
100/20
13
10
13
13
21
11
15
140/40
16
10
10
8
8
14
11
Experiment 17
Date: 9/15/94
C:N ratio: 5:1
V
Results
plate
Time in Seconds
Agitatio HC volt
10
20
40
60
I 0.041
0.023
100
0.6
0.014
140
0.2 0.015
0.011
Control
Agitatic
60
100
140
Time in seconds
10
20
40
0.094
I
2
3
4
5
6
7
8
9
10
11
8
15
49
9
20
47
10 Averag' Std. dev
19 19.40
2.95
57 51.20
6.29
28
16
8
8
34
21
27
16
8
27
16
21
16
11
35
16
28
18
9
9
38
10
Oil. fact.
1.00E+06
1.00E+05
1.00E+05
1.00E+05
1.00E+04
1.00E+04
1.00E+04
1.00E+03
1.00E+05
1.00E+05
1.00E+05
25.10
18.00
7.25
12.10
31.50
18.00
27.33
14.44
10.44
2.64
3.33
2.12
2.60
4.81
5.33
3.04
3.47
2.83
T o t CC Std. dev.
2.63E+07
0.193
2.41 E+06
0.190
1.67E+06
0.281
8.00E+05
0.328
4.00E+05
0.190
1.38E+05
0.213
1.60E+05
0.265
3.95E+04
0.086
1.27E+06
0.285
1.53E+06
0.200
1.90E+06
0.211
0.144
0.152
Cell Count data
agit. rate/hc
I abs. 0.204
dll 1/2
dll 1/4
60/1/10
60/1/40
100/.6/20
140/.2/10
8 140/.2/40
9 60/40
10 100/20
11 140/10
Plate counts, where each count represents a one ]*\. sample.
I
2
3
4
5
6
7
26
35
23
25
30
22
32
30
26
17
20
27
28
25
20
10
10
18
19
19
14
7
7
5
6
6
7
12
34
56
39
36
27
42
42
17
17
16
13
17
9
12
17
21
13
15
8
20
13
37
39
39
33
39
44
43
20
16
16
10
10
12
11
18
16
13
11
13
19
16
19
15
20
13
13
20
24
8
29
20
20
8
37
13
15
44
8
11
21
9
22
24
9
43
10
22
39
11
18
23
10 Averag' Std. dev
19 26.30
5.08
4.58
24.13
13 16.70
4.69
13
8.00
2.62
44 40.00
7.60
14 13.80
2.94
16 16.00
4.24
38 39.50
3.41
13 12.70
3.62
18 15.30
3.06
22 19.00
4.00
85
Experiment 18
Results
Date: 9/16/94
plate
Agitatio
60
100
140
C:N ratio: 5:1
Dil. fact. T o tC C SId. dev.
I contaminated
2 1.00E+04 3.89E+05
0.074
3 1.00E+04 3.02E+05
0.176
4 1.00E+04 1.46E+05
0.165
5 1.00E+04 8.22E+04
0.190
6 contaminated
7 1.00E+03 6.80E+03
0.316
8 contaminated
9 1.00E+05 1.84E+06
0.251
10 1.00E+05 1.40E+06
0.208
11 contaminated
Time in Seconds
HC volt
10
20
40
0.2
0.04
0.025
0.6
0.012
I 0.011
0.013
Interference
Control
Agitatic
60
100
140
Time in seconds
10
20
0.17
0.191
40
I
2
3
4
5
6
7
8
9
10
0.195
Cell Count data
I
2
3
4
5
6
7
8
9
10
agit. rate/hc'
dil I abs .204
dil 1/2 abs .1
dil 1/4 abs .(
60/.2/10
60/.2/40
100/.6/20
140/1/10
140/1/40
60/10
100/20
Experiment 19
Date: 9/17/94
C:N ratio: 5:1
V
Plate counts, where each count represents a one /d. sample.
I
2
3
4
5
6
7
8
9
10 Averag Std. dev
35
35
16
6
38
31
19
8
41
22
12
8
39
27
12
9
34
30
13
9
40
25
14
i6
41
28
14
9
37
38
13
11
43
36
15
8
41
9
3
5
9
10
7
7
7
5
18
16
20
18
16
13
14
15
13
13
Z3
10
14
18
18
12
20
11
Results
Agitatio
60
100
140
100
38.90
30.22
14.60
8.22
2.88
5.33
2.41
1.56
6
6.80
2.15
28
18.40
14.00
4.62
2.92
18
plate
Time in Seconds
HC volt
10
20
40
0.2 0.045 0.039 0.034
0.6
0.03 0.024
0.02
I 0.019 0.016 0.016
0.6
0.023
Control
Agitatic
60
100
140
Time in seconds
10
20
40
0.084
Dil. fact. T o tC C Std. dev.
1.00E+06 7.70E+06
0.318
0.00E+00
0.000
1.00E+04 9.00E+04
0.319
1.00E+04 1.63E+05
0.188
1.00E+04 8.33E+04
0312
1.00E+03 2.01E+04
0.181
1.00E+03 1.48E+04
0.208
1.00E+03 8.60E+03
0.199
1.00E+03 2.66E+04
0.112
1.00E+05 9.30E+05
0.273
11 1.00E+05 1.06E+06
0.425
12 1.00E+05 3.17E+06
0.119
I
2
3
4
5
6
7
8
9
10
0.091
0.125
Cell Count data
I
2
3
4
5
6
7
8
9
10
11
12
Plate counts, where each count represents a one /d. sample.
agit. rate/hc
I
2
3
4
5
6
7
init abs .202
9
5
5
5
12
10
9
abs. .090 dil 1/2
abs. .044 dil
12
10
12
10
13
5
8
60/.2/10
14
17
13
16
14
18
18
60/.2/40
11
11
12
7
7
8
5
100/. 6/20
22
26
16
19
22
15
18
140/1/10
15
18
14
12
9
18
16
140/1/40
10
10
8
7
8
8
8
100/.6/20
24
30
28
31
25
26
30
60/40
9
5
9
7
8
14
12
100/20
5
7
5
10
8
9
16
140/10
41
30
28
34
33
3I
30
a*3RR*S&**&*S&*S
8
6
9
7
5
17
5
22
19
6
26
11
16
29
8
13
9
17
14
9
23
9
15
32
10 Averag Std. dev
9
7.70
2.45
7
23
24
13
12
23
9
15
29
9.00
16.30
8.33
20.10
14.80
8.60
26.60
9.30
10.60
31.70
2.87
3.06
2.60
3.63
3.08
1.71
2.99
2.54
4.50
3.77
8 # #
86
Experiment 20
9/ 18/94
Init abs. .200
Dil.
Abs.
1/2
0.063
1/4
0.041
AggiUl
60/0.2
100/0.6
140/0.2
IOs
20s
40s
0.043
0.046
0.04
0.026
0.029
0.03
0.026
0.02 0.017
C:N 15:1
Control
V
Aggit. I IOs
20s
40s
60 0.146
100
0.131
140
0.101
Experiment 21
Results
Date: 9/19/94
Agitatio HC voli
60
I
100
0.6
140
0.2
C:N ratio: 5:1
V
plate
Time in Seconds
10
20
40
0.033
0.03 0.028
0.016 0.013 0.013
0.017 0.018 0.013
1.00E+04
1.00E+04
1.00E+04
1.00E+04
1.00E+04
1.00E+03
1.00E+04
1.00E+04
1.00E+05
2.79E+05
2.13E+05
1.80E+05
7.40E+04
1.18E+05
2.97E+04
3.20E+05
3.20E+05
1.62E+06
6
21
7
10
8
14
9
12
10 Averag Std. dev
14.22
3.73
32
17
21
6
9
49
29
33
9
25
27
18
5
16
49
29
39
13
18
19
20
5
16
44
33
27
17
23
22
15
7
14
41
25
27
13
32
Control
Time in seconds
10
20
0.09
100
0.088
140
Agitatic
60
Dil. fact. T o tC C Std. dev.
1.00E+06 1.42E+07
0.263
I
2
3
4
5
6
7
8
9
10
11
40
0.202
0.251
0.196
0.409
0.270
0.160
0.159
0.159
0312
0.11
Cell Count data
1
2
3
4
5
6
7
8
9
10
11
Plate counts, where each count represents i
agit. rate/hc 1
I
2
3
4
5
init abs ..208
18
10
13
13
17
abs. 108dil 1/2
abs. .036 dil
36
27
30
33
23
60/1/10
21
21
32
19
14
60/1/40
16
22
12
23
15
100/.6/20
14
6
6
11
9
140/.2/10
9
8
14
13
11
140/.2/40
46
46
38
43
37
60/10
31
33
37
20
28
100/20
28
33
31
42
28
140/40
14
18
19
28
16
18
5
8
36
32
32
15
27.90
21.33
18.00
7.40
11.80
29.70
32.00
32.00
16.20
5.63
5.36
3.53
3.03
3.19
4.74
5.10
5.10
5.05
87
Experiment 22
Results
Date: 9/22/94
Agitatio HC voh
60
I
60
I
140
0.2
C:N ratio: 5:1
Time in Seconds
10
20
0.027 0.018
0.025
0.02
0.01 0.009
plate
40
0.01
0.01
0.01
V
Control
Agitatk
60
100
140
Time in seconds
10
20
40
0.043
I
2
3
4
5
6
7
8
9
10
11
Dil. fact.
1.00E+04
1.00E+04
1.00E+04
1.00E+04
1.00E+04
1.00E+04
1.00E+04
1.00E+04
1.00E+03
1.00E+04
1.00E+05
T o t CC Std. dev.
0.00E+00
0.000
0.00E+00
0.000
1.78E+05
0.238
2.44E+05
0.177
2.15E+05
0333
2.53E+05
0.251
1.76E+05
0.235
1.07E+05
0.245
3.30E+04
0.189
2.39E+05
0.264
9.00E+05
0.210
0.09
Cell Count data
I
2
3
4
5
6
7
8
9
10
11
Plate counts, where each count represents a one u\. sample.
agit. rate/hc
I
2
3
4
5
6
7
initabs. 0.199 dil I
abs. .062 dil 1/2
abs. .024 dil
19
19
18
26
20
17
10
60/1/10
22
24
30
25
18
22
20
60/1/40
17
21
35
15
20
12
23
60/1/10
25
34
29
32
18
20
18
60/1/40
17
17
15
18
26
18
14
140/.2/10
7
8
12
10
10
15
12
140/.2/40
20
38
31
40
37
26
33
60/40
18
14
17
23
28
30
33
140/10
13
10
8
8
9
9
7
tax p crim cn t 23
C :N 5:1
Init abs. AggiVI 10s
20s
40s
Dil.
60/1
0.066
0.06 0.058
1/2
(100/0.6 0.034
0.03
0.03
1/4
( 140/0.2
0.04 0.033
0.03
140/1
0.025 0.015 0.017
Interference
Control
10/4/94
Aggit. I IOs
20s
40s
60
0.075
100
0.129
140 0.132
8
9
19
32
26
18
22
11
38
30
7
13
27
30
28
11
8
36
24
8
10 Averag' Std. dev
17
24
16
31
18
14
31
22
11
17.80
24.40
21.50
25.30
17.60
10.70
33.00
23.90
900
4.24
4.33
7.17
6.34
4.14
2.63
6.24
6.31
I RQ
88
Experiment 24
Date: 10/7/94
C:N ratio: 15:1
V
Results
Time in Seconds
Agitatio HC volt
10
20
60
0.2
0.07 0.068
60
I 0.056 0.063
100
0.6 0.029 0.026
140
0.2 0.033 0.023
140
I
0.03 0.016
40
0.028
0.029
0.022
0.02
0.017
Control
Agitatic
60
100
140
Time in seconds
10
20
0.118
0.111
40
0.081
2
3
4
5
6
7
8
9
10
11
12
13
14
15
1.00E+05
1.00E+05
1.00E+05
1.00E+04
1.00E+05
1.00E+04
1.00E+04
1.00E+04
1.00E+04
1.00E+04
1.00E+03
1.00E+05
1.00E+05
1.00E+05
1.39E+06
1.10E+06
1.41E+06
5.11E+05
7.00E+05
4.14E+05
2.21E+05
2.56E+05
6.80E+04
1.30E+05
2.01E+04
1.07E+06
1.48E+06
1.80E+06
0.289
0.113
0.235
0.140
0.261
0.116
0.248
0.143
0.167
0.202
0.221
0.165
0.188
0.126
Cell Count data
I
2
3
4
5
6
7
8
9
10
11
12
13
14
15
agit. rate/hc initabs. 0.19
dil 1/2
d ill/4
60/.2/10
60/.2/40
60/1/10
60/1/40
100/. 6/20
140/.2/10
140/. 2/40
140/1/10
140/1/40
60/10
100/20
140/40
Plate counts, where
I
2
14
14
20
13
12
10
10
14
49
51
8
7
43
37
18
16
32
29
5
8
13
14
18
20
13
12
17
10
15
19
each count represents a one pi. sample.
3
4
5
6
7
11
11
15
15
16
17
11
12
7
12
12
11
9
10
10
9
11
15
16
17
62
59
45
54
36
8
6
6
7
8
39
42
39
34
50
30
21
26
21
24
27
26
24
27
27
7
7
6
6
7
11
18
14
14
13
21
27
24
17
14
11
10
13
10
9
15
12
14
14
18
20
21
17
14
19
8
13
18
13
13
51
7
41
16
23
7
14
19
8
16
20
9
10
15
12
17
51
10
41
18
21
6
11
15
12
13
18
10 Averag. Std. dev
7 12.60
2.80
13.89
4.01
11 11.00
1.25
19 14.10
3.31
53 51.10
7.17
3
7.00
1.83
48 41.40
4.79
31 22.10
5.49
20 25.60
3.66
9
6.80
1.14
8 13.00
2.62
26 20.10
4.43
9 10.70
1.77
19 14.80
2.78
17 18.00
2.26
89
Experiment 25
Date: 10/8/94
C:N ratio: 5:1
Interference
Results
plate
Time in Seconds
Agitatio HC volt
10
20
60
0.2 0.084 0.078
60
0.2
0.09 0.079
60
I 0.087
140
0.2 0.073 0.065
140
I 0.069 0.065
40
0.078
0.071
0.061
0.06
0.068
Control
Agitatic
60
100
140
Time in seconds
10
20
0.118
0.124
40
0.105
0.148
I
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
Dil. fact. ToLCC Std. dev.
1.00E+05 1.32E+06
0.257
0.00E+00
0.000
1.00E+04 2.20E+05
0.233
1.00E+04 2.09E+05
0.202
1.00E+04 1.29E+05
0.202
1.00E+04 1.60E+05
0.179
1.00E+04 1.33E+05
0.279
1.00E+04 1.43E+05
0.218
1.00E+04 1.00E+05
0.271
1.00E+04 6.10E+04
0.196
1.00E+03 3.79E+04
0.078
1.00E+03 1.70E+04
0.232
1.00E+03 1.69E+04
0.237
1.00E+04 1.97E+05
0.225
1.00E+05 1.74E+06
0.128
1.00E+04 3.94E+05
0.090
1.00E+05 1.37E+06
0.220
Cell Count data
I
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
Plate counts, where each count represents a one /d. sample.
agit. rate/hc ■
I
2
3
4
5
6
7
init abs .204
14
15
19
13
11
7
10
dil 1/2
d ill/4
24
29
28
23
25
20
14
60/.2/10
27
17
24
18
18
23
28
60/.2/40
15
16
10
11
9
11
12
60/.2/10
14
14
18
20
20
13
12
60/.2/40
13
13
8
19
16
15
12
60/1/10
20
10
14
15
14
16
12
60/1/40
13
14
9
10
5
7
10
140/.2/10
7
8
6
6
6
5
7
140/. 2/40
39
32
37
35
42
37
41
140/1/10
15
25
19
16
16
21
14
140/1/40
17
20
17
22
10
23
16
60/10
23
17
14
20
13
25
21
60/40
20
17
13
17
17
17
21
140/10
39
41
43
35
44
43
39
140/40
18
16
16
16
13
14
8
8
16
9
12
10 Averag Std. dev
15 13.20
3.39
23
17
16
15
10
17
12
7
40
14
17
24
16
39
13
14
18
14
18
18
11
9
4
37
13
13
16
17
33
13
20
19
15
16
9
11
5
39
14
24
19
38
10
22.00
20.90
12.90
16.00
13.30
14.33
10.00
6.10
37.90
17.00
16.90
19.70
17.40
39.40
13.70
5.12
4.23
2.60
2.87
3.71
3.12
2.71
1.20
2.96
3.94
4.01
4.42
2.22
3.53
3.02
90
Experiment 26
Date: 10/11/94
C:N ratio: 5:1
Interference
Results
Agitatio
60
60
140
140
plate
Time in Seconds
HC volt
10
20
0.2 0.061 0.048
I 0.071 0.061
0.2 0.058 0.055
I 0.056
I
2
3
4
5
6
7
8
9
10
11
12
13
14
15
40
0.047
0.051
0.045
0.038
Control
Agitatic
60
100
140
Time in seconds
10
20
0.098
0.101
40
0.091
Dil. fact.
1.00E+06
1.00E+05
1.00E+04
1.00E+04
1.00E+04
1.00E+04
1.00E+04
1.00E+03
1.00E+03
1.00E+03
1.00E+03
1.00E+04
1.00E+04
1.00E+04
1.00E+03
T ot CC Std. dev.
8.60E+06
0.221
9.40E+05
0.355
1.78E+05
0.187
1.68E+05
0.255
1.14E+05
0326
2.08E+05
0.193
7.70E+04
0.260
4.32E+04
0.133
1.86E+04
0.218
2.42E+04
0.188
1.61E+04
0.148
3.10E+05
0.156
3.66E+05
0.126
2.38E+05
0.163
4.77E+04
0.131
Cell Count data
I
2
3
4
5
6
7
8
9
10
11
12
13
14
15
Plate counts, where
agit. rate/hc ■
I
2
init abs .201
12
8
abs .075 dll I
11
14
abs .042 dil I
17
17
60/.2/10
13
15
60/.2/40
12
7
60/1/10
20
19
60/1/40
11
5
140/.2/10
41
49
140/.2/40
23
26
140/1/10
24
28
140/1/40
13
14
60/10
28
26
100/20
36
38
140/40
29
28
100/. 6/20
40
41
each count represents a one ]
3
4
5
6
8
11
7
7
4
9
9
10
22
24
18
17
13
20
17
13
6
12
11
11
22
22
24
23
7
6
8
9
35
53
38
45
20
22
15
14
24
27
20
24
16
20
14
18
24
40
28
34
38
36
44
29
26
25
25
24
43
49
57
53
a \.
sample.
7
8
11
18
15
8
18
9
43
14
31
19
35
40
24
54
8
6
14
12
15
14
13
10
38
18
28
14
33
32
21
40
9
9
6
15
21
18
28
6
41
17
16
17
33
20
49
10 Averag Std. dev
10
8.60
1.90
6
9.40
3.34
18 17.80
3.33
26 16.80
4.29
15 11.40
3.72
19 20.80
4.02
6
7.70
2.00
49 43.20
5.75
17 18.60
4.06
20 24.20
4.54
16 16.10
2.38
29 31.00
4.83
36.63
4.63
16 23.80
3.88
51 47.70
6.27
91
Experiment 27
Date: 10/16/94
C:N ratio: 15:1
Interference
Results
plate
Time in Seconds
Agitatio HC volt
10
20
60
0.2 0.085
60
I 0.076
100
0.6
0.061
140
0.2 0.065
140
I
0.06
40
0.075
0.074
0.064
0.062
Control
Agitatic
60
100
140
Time in seconds
10
20
0.148
0.144
40
0.1
I
2
3
4
5
6
7
8
9
10
11
12
13
14
15
Dil. fact.
1.00E+05
1.00E+05
1.00E+04
1.00E+04
1.00E+04
1.00E+04
1.00E+04
1.00E+04
1.00E+04
1.00E+03
I.OOE+04
1.00E+03
1.00E+05
1.00E+05
1.00E+04
T o t CC Std. dev.
1.27E+06
0.292
7.70E+05
0.324
1.98E+05
0.136
3.10E+05
0.210
2.90E+05
0.156
2.18E+05
0.221
2.04E+05
0.154
5.12E+05
0.064
1.56E+05
0.199
5.19E+04
0.072
1.79E+05
0.245
2.02E+04
0.177
7.44E+05
0.214
1.34E+06
0.173
3.99E+05
0.152
Cell Count data
I
2
3
4
5
6
7
8
9
10
11
12
13
14
15
agit. rate/hc •
init abs .200
abs .086 dil I
abs .038 dil I
60/.2/10
60/.2/40
60/1/10
60/1/40
100/. 6/20
140/.2/10
140/.2/40
140/1/10
140/1/40
60/10
100/20
140/40
Plate counts, where each count represents a one ]A. sample.
I
2
3
4
5
6
7
8
16
18
9
11
14
7
4
8
6
9
7
8
13
18
18
20
23
18
23
24
40
26
21
36
34
29
41
25
30
29
29
38
28
26
28
21
27
23
29
16
18
25
19
19
16
18
25
23
48
52
56
57
47
49
50
10
16
14
12
19
19
19
48
52
57
57
47
49
50
13
14
12
21
18
18
24
17
21
16
14
23
22
24
9
9
5
8
7
7
9
15
14
11
14
10
16
11
48
30
38
43
42
45
41
8
14
5
20
27
27
16
18
53
14
53
25
20
8
14
45
9
15
8
18
29
23
19
22
50
16
56
17
25
5
17
36
10 Averag Std. dev
15 12.70
3.71
9
7.70
2.50
16 19.80
2.70
27 31.00
6.50
35 29.00
4.52
21 21.80
4.83
19 20.40
3.13
50 51.20
3.29
17 15.60
3.10
50 51.90
3.73
17 17.90
4.38
20 20.20
3.58
7.44
1.59
12 13.40
2.32
31 39.90
6.05
92
Experiment 28
Date: 10/18/94
C:N ratio: 15:1
V
Results
Time in Seconds
Agitatio HC volt
10
20
40
60
0.2 0.083
0.07 0.063
60
I 0.078 0.062
0.04
100
0.6 0.032 0.037
0.02
140
0.2
0.03 0.035 0.016
140
I 0.035 0.027
0.01
100
0.6
0.038
Control
Time in seconds
Agitatic
10
20
40
60 0.118
0.118
100
0.105
140
0.08
0.08
plate
Dil. fact.
I 1.00E+05
2 1.00E+05
3 1.00E+05
4 1.00E+05
5 1.00E+05
6 1.00E+05
7 1.00E+04
8 1.00E+04
9 1.00E+04
10 1.00E+03
11 1.00E+04
12 1.00E+03
13 1.00E+04
14 1.00E+05
15 1.00E+05
16 1.00E+05
17 1.00E+05
18 1.00E+05
T o tC C Std. dev.
1.57E+06
0.235
8.80E+05
0.292
1.12E+06
0.275
1.07E+06
0.265
7.56E+05
0.188
7.20E+05
0.205
3.28E+05
0.223
1.30E+05
0.154
2.24E+05
0.128
4.02E+04
0.152
1.45E+05
0.265
1.14E+04
0.323
1.15E+05
0.193
1.48E+06
0.208
1.12E+06
0.222
1.26E+06
0.288
1.29E+06
0315
1.37E+06
0.176
Cell Count data
I
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
Plate counts, where each count represents
agit. rate/hc ■
I
2
3
4
5
init abs .204
19
11
17
17
13
abs .065 dil I
13
11
5
7
8
abs .04 dil I
9
14
10
9
9
60/.2/10
10
16
10
13
13
60/.2/40
6
8
6
7
9
60/1/10
5
8
7
8
5
60/1/40
32
28
31
37
41
100/. 6/20
13
13
17
14
10
140/.2/10
26
17
20
21
25
140/.2/40
35
41
45
43
30
140/1/10
11
14
23
18
16
140/1/40
8
12
6
19
13
100/. 6/20
14
11
12
12
8
60/10
19
18
11
16
16
60/40
9
10
15
16
11
100/20
18
16
9
16
8
140/10
14
16
7
11
13
140/40
15
11
14
11
16
one pi\. sample.
6
7
18
22
12
9
6
12
9
9
8
6
9
9
34
22
13
14
25
21
47
50
15
12
8
12
11
11
14
14
12
11
9
10
10
22
16
14
8
15
6
13
12
8
6
27
13
21
37
10
10
10
17
10
12
13
17
9
15
9
14
6
10
7
29
13
23
37
12
12
16
14
9
16
10
10
10 Averag Std. dev
10 15.70
3.68
8
8.80
2.57
16 11.20
3.08
9 10.70
2.83
7.56
1.42
8
7.20
1.48
47 32.80
7.30
10 13.00
2.00
25 22.40
2.88
37 40.20
6.11
14 14.50
3.84
14 11.40
3.69
10 11.50
2.22
9 14.80
3.08
9 11.20
2.49
12 12.60
3.63
13 12.90
4.07
13 13.70
2.41
93
Experiment 29
Date: 10/19/94
C:N ratio: 15:1
V
Results
plate
Time in Seconds
Agitatio HC volt
10
20
60
0.2 0.042 0.052
100
0.6 0.051 0.046
140
0.2 0.044 0.044
140
I 0.043 0.036
40
0.051
0.047
0.103
0.034
Control
Agitatic
60
100
140
Time in seconds
10
20
0.101
40
0.085
I
2
3
4
5
6
7
8
9
10
11
12
13
Dil. fact.
1.00E+05
1.00E+05
1.00E+04
1.00E+05
1.00E+04
1.00E+04
1.00E+04
1.00E+03
1.00E+04
1.00E+03
1.00E+05
1.00E+05
1.00E+05
0.104
ToLCC Std. dev.
1.34E+06
0.193
1.21E+06
0.402
4.54E+05
0.102
8.40E+05
0.219
3.08E+05
0.160
8.70E+04
0.372
1.25E+05
0.197
3.01E+04
0.130
6.70E+04
0.234
8.40E+03
0.298
1.30E+06
0.249
1.57E+06
0.238
9.70E+05
0.266
Cell Count data
Plate counts, where each count represents a one pi\. sample.
agit. rate/hc ■
I
2
3
4
5
6
7
I init abs .204
12
10
14
11
16
17
17
2 abs .075 dil I
8
16
21
17
15
9
8
3 abs .042 dil I
47
36
48
48
45
45
40
4 60/.2/10
5
9
7
9
10
11
6
5 60/.2/40
34
23
36
36
36
26
28
6 100/. 6/20
13
7
5
14
8
7
7
7 140/.2/10
10
10
13
16
12
14
14
8 140/.2/40
34
24
36
29
27
27
34
9 140/1/10
9
6
9
7
5
5
7
10 140/1/40
10
11
9
6
10
12
7
11 60/40
15
15
13
10
19
II
14
12 100/20
10
14
19
20
12
20
14
13 140/10
11
14
8
11
6
10
11
8
14
12
52
9
26
5
10
28
5
9
7
12
12
9
12
7
49
9
34
12
16
29
6
5
13
19
7
10 Averag Std. dev
11 13.40
2.59
8 12.10
4.86
44 45.40
4.62
9
8.40
1.84
29 30.80
4.94
9
8.70
3.23
10 12.50
2.46
33 30.10
3.90
8
6.70
1.57
5
8.40
2.50
13 13.00
3.23
17 15.70
3.74
7
9.70
2.58
94
Experiment 30
Results
Time in Seconds
Agitatio HC volt
10
20
60
0.2 0.055 0.058
60
I 0.054 0.051
100
0.6 0.033 0.029
140
0.2 0.033 0.031
140
I 0.031 0.027
Date: 10/24/94
C:N ratio: 15:1
V
plate
40
0.043
0.045
0.028
0.028
0.27
Control
Agitatic
60
100
140
Time in seconds
10
20
0.133
0.125
0.14
40
0.104
0.103
Cell Count data
agit. rate/hc'
1 init abs .204
2 abs .065 dil I
3 abs .04 dil I
4 60/.2/10
5 60/.2/40
6 60/1/10
7 60/1/40
8 100/.6/20
9 140/.2/10
10 140/. 2/40
11 140/1/10
12 140/1/40
13 60/10
14 60/40
15 100/20
16 140/10
17 140/40
Experiment 31
Date: 10/26/94
C:N ratio: 15:1
Interference
Plate counts, where
I
2
5
5
14
18
10
6
12
19
6
14
13
12
7
6
24
24
27
37
5
8
17
16
13
13
22
16
11
9
9
8
16
10
14
23
Dil. fact.
I 1.00E+06
2 1.00E+05
3 1.00E+05
4 1.00E+05
5 1.00E+05
6 1.00E+05
7 1.00E+05
8 1.00E+05
9 1.00E+04
10 1.00E+04
11 1.00E+04
12 1.00E+03
13 1.00E+05
14 1.00E+05
15 1.00E+05
16 1.00E+05
17 1.00E+05
each count represents a one ftl. sample.
3
4
5
6
7
9
9
10
6
6
21
16
14
20
17
5
8
8
8
6
16
15
19
23
28
11
9
18
11
13
14
8
10
13
10
10
6
9
7
7
24
19
21
22
24
34
39
31
41
38
12
5
9
12
10
22
18
27
22
15
13
15
12
9
22
24
22
19
27
25
6
6
7
8
8
11
16
15
9
12
16
8
9
15
8
28
17
17
15
18
Results
plate
Time in Seconds
Agitatio HC volt
10
20
40
60
I 0.066 0.066
0.06
140
I 0.061 0.056 0.053
Control
Time in seconds
Agitatic
10
20
60
100
0.104
140 0.101
40
0.111
1
2
3
4
5
6
7
8
9
10
8
7
14
5
16
11
8
6
19
43
13
21
15
18
9
8
11
18
Dil. fact.
1.00E+05
1.00E+05
1.00E+04
1.00E+04
1.00E+04
1.00E+04
1.00E+03
1.00E+05
1.00E+04
1.00E+05
ToLCC Std. dev.
7.80E+06
0.289
1.73E+06
0.170
6.80E+05
0.238
1.81E+06
0.268
1.05E+06
0376
1.10E+06
0.212
7.70E+05
0.229
2.09E+06
0.162
3.66E+05
0.139
8.70E+04
0.351
1.89E+05
0.218
1.37E+04
0.251
2.24E+06
0.173
8.30E+05
0.241
1.15E+06
0.263
1.22E+06
0.323
1.83E+06
0.228
9
11
17
6
15
7
8
18
42
5
13
11
23
7
12
10
17
10 Averag' Std. dev
10
7.80
2.25
22 17.30
2.95
6
6.80
1.62
18.11
4.86
5 10.50
3.95
11.00
2.33
11
7.70
1.77
14 20.90
3.38
34 36.60
5.10
8
8.70
3.06
18 18.90
4.12
14 13.70
3.43
28 22.40
3.86
12
8.30
2.00
15 11.50
3.03
19 12.20
3.94
16 18.30
4.16
ToL CC Std. dev.
8.10E+05
0.299
6.60E+05
0.228
2.27E+05
0.168
2.85E+05
0.162
1.75E+05
0.158
1.09E+05
0.261
7.80E+03
0.240
6.11E+05
0.172
3.66E+05
0.077
6.44E+05
0.175
Cell Count data
agit. rate/hc'
1 init abs .200
2 abs .086 dil I
3 abs .038 dil I
4 60/1/10
5 60/1/40
6 140/1/10
7 140/1/40
8 60/40
9 100/20
10 140/10
Plate counts, where each count represents a one fil. sample.
I
2
3
4
5
6
7
6
6
12
9
9
11
7
4
7
8
9
6
5
6
25
20
18
26
16
24
25
31
25
26
23
25
29
32
17
22
16
15
20
17
13
7
8
16
10
8
12
10
6
6
7
9
8
10
11
6
5
6
7
8
7
6
30
40
37
38
39
35
36
6
5
7
8
6
5
7
8
23
33
17
9
5
8
27
37
17
12
12
8
5
39
8
8
5
36
6
6
6
10 Averag Std. dev
10
8.10
2.42
7
6.60
1.51
22.67
3.81
24 28.50
4.62
21 17.50
2.76
14 10.90
2.85
5
7.80
1.87
6.11
1.05
36.60
2.84
6.44
1.13
95
APPENDIX B
IMAGE ANALYSIS RESEARCH
96
Image Analysis System
Experimental Design
The initial experimental design made use of the image analysis system at the Center
for Biofilm Engineering (CBE). Initially the proposed project would determine the kinetics
of bacterial attachment at the hydrocarbon-water interface. Design of the experimental
system went through a series of modifications as different methods were tried and
subsequently rejected. The scope of research was altered after it became obvious that
finding a working system using image analysis was unattainable with materials available
for the research project. The original project design incorporated cell surface hydrophobicity as measured by adhesion of bacteria cells to hydrocarbon, which was to be compared to
results from the kinetics of attachment as determined from the image analysis results.
Setup
The experimental system was similar to that of R. Mueller (1990). The image
analysis system was not changed. In order to obtain slow flow rates a syringe pump was
utilized. A flow rate of 4.8 ml per hour was used in all the experiments. This flow rate
corresponded to a Renolds number of 0.1077 and a mean bulk fluid velocity of 6.15 * IO"5
meters per second. The flow cell dimensions were: depth 0.180 cm, width 1.204 cm and
length 4.054 cm. The flow cell consisted of a shallow groove cut into polycarbonate which
was covered by a glass slide that was pressed against a rubber seal by the compression of a
stainless steel plate. The plate was screwed onto the base of the polycarbonate. The
hydrocarbon-water interface could be observed with a light microscope connected to the
image analysis system.
97
Design for containment of the hydrocarbon, which was to be observed by the image
analysis system, proved to be troublesome. Methods were tried and discarded and it
became apparent that the design of the system was very important to the final success of the
project. One of the first prototypes for containment was a simple hole punched through a
thin Teflon plate. The hydrophobic surface of the Teflon should have kept the hydrocarbon
intact in the hole while water flowed past. Observation of the interface of this system was
easy. The microscope focused on the hole and the interface was found under higher
magnification. It became apparent that the hydrocarbon was not to be contained by the
hydrophobic attraction to the Teflon. Water replaced the hydrocarbon soon after the syringe
pump was turned on. Other hydrophobic materials (polystyrene, polycarbonate and glass)
were also tested, and produced similar results.
In another method a thin layer (monolayer) of hydrocarbon was placed on the
surface of the water in the flow cell and the coverslip replaced. Flow was started and the
interface was observed. In this experiment the oil film slipped between the surface of the
coverslip and the polycarbonate flow cell. Not all of the hydrocarbon became dispersed
with the flow and droplets of hydrocarbon could be observed adhering to the glass
coverslip. The last method used the observation that small droplets of hydrocarbon will
adhere to the glass coverslip. After cleaning the surface of the glass to remove any
surfactants small droplets of hydrocarbon were placed on the coverslip. The coverslip was
inverted and placed on top of the flow cell.
Procedure
In a typical experiment the cell suspension was removed from the water bath during
the stationary growth phase. The cells were prepared as described earlier in the methods
section. The autoclaved elements essential to the experiment were prepared. Two systems
were initially employed, one in which the cell suspension flowed directly to the flow cell.
98
The second design decreased the pulsation of the fluid flow observed in the first system. In
the second design two bottles were placed in series with the syringe pump before the cell
suspension entered the flow cell. Sterilized water from the syringe pump entered the first
bottle forcing air from the head space into a tube connected to the next bottle. The air from
the first bottle entered the second bottle at the top and subsequently forced the cell
suspension out of the bottle through a glass tube. This system had another advantage over
the first because it allowed mixing of the cell suspension in the bottle. From the glass tube
the cell suspension flowed through tybex tubing and into the flow cell. The flow cell was
prepared as described above. After carefully turning the cover slip over it was placed onto
water that was contained by the groove in the flow cell. The flow cell was covered by the
metal clamp and tightened down. The connections to the rest of the system were made and
any trapped air was purged from the system through a relief valve. The syringe pump was
turned on and the interface was observed under the microscope.
Observations
The results of the image analysis experiments indicated that further exploration of
the problem is needed. Finding the hydrocarbon-water interface before the bacterial
suspension had entered the flow cell proved to be difficult. The interface was transparent
and unless something adhered between the phases the exact location of the hydrocarbonwater interface was not easy to find. In some cases a little dust or some other outside object
could be observed and in such cases observation of the interface could begin.
All of the hydrophobic surfaces that were tested failed to hold the hydrocarbon in
place when subjected to fluid flow. In some cases the hydrocarbon would spread as soon
as it came in contact with the water. Droplets of hydrocarbon on the surface of hydrophobic
materials other than glass did not work because of interference of the light entering the
objective lens.
99
Once it was determined that hydrocarbon droplets on the glass slide could be
observed, an experiment was initiated to determine the kinetics of bacteria attachment to
hydrocarbon. Initially the experiment progressed as desired. The location of the
hydrocarbon-water interface was determined. The flow was started after which images
were taken at fifteen minute intervals.
The suspended bacteria could be observed flowing through the cell. Some of the
bacteria could be observed interacting with the hydrocarbon. Cells that came in contact with
the hydrocarbon would have a slower rate of flow than those which did not. The flow rate
nearest the glass surface was slower then fluid flow towards the center of the groove.
After the cells came into contact with the hydrocarbon another problem developed.
The surface of the hydrocarbon began to rotate. The direction of the movement was hard to
determine. Cells that were observed at the interface moved slowly along with the flow of
the hydrocarbon. In some cases cells rotated towards the glass surface only to disappear
from view as they came into contact with the glass. Once the cells contacted the glass
surface they could not be observed any longer. Two different processes could be taking
place. In one case the cell would desorb from the hydrocarbon and enter the bulk fluid only
to exit the flow cell. The other possibility is that the cell moves with the interface as it
rotates, and if the hydrocarbon flows inward and under the glass the cell moves with it. In
this case the cell disappears from view as it gets covered by the hydrocarbon. The exact
rotational movement of the hydrocarbon was not determined. No observation of an attached
cell desorbing into the bulk fluid was observed. Therefore it could not be determined if the
glass forced the cells into the bulk liquid or if the bacteria remained attached only to
resurface else where on the spherical hydrocarbon droplet.
100
As the experiment progressed it became apparent that the tendency of the cells to
clump was interfering with the experiment. The cells were entering the flow cell in small
visible clumps which moved along the lower surface of the channel. Attachment kinetics
could be influenced by the tendency of the cells to clump together. Clumped cells were
never observed attaching to the interface further complicating the determination of the
kinetics of attachment at the interface.
The kinetics of bacterial attachment at the hydrocarbon-water interface could not be
determined using this experimental design. The system was fraught with flaws and any
data obtained from the experiments would be very difficult to interpret. A method for
maintaining a hydrocarbon-water interface stationary is needed.
MONTANA STATE UNIVERSITY LIBRARIES
»