TABLE OF CONTENTS
TABLE OF CONTENTS ............................................................................................... i
LIST OF FIGURES ..................................................................................................... iii
LIST OF TABLES ....................................................................................................... iv
LIST OF ABREVIATIONS ..........................................................................................v
ACKNOWLEDGEMENTS ......................................................................................... vi
ABSTRACT ................................................................................................................ vii
CHAPTER I – General Introduction on M. marinum and Biofilms ..............................1
Mycobacteria Species ............................................................................................... 1
M. marinum ............................................................................................................... 2
Biofilms..................................................................................................................... 3
NTM Biofilms ........................................................................................................... 5
M. marinum Biofilm ................................................................................................. 5
Relevance of M. marinum ......................................................................................... 6
CHAPTER II – Quantifying Biofilm Formation ...........................................................7
INTRODUCTION .........................................................................................................7
METHODS ....................................................................................................................8
Bacterial strains and growth conditions .................................................................... 8
MBEC™ plate assay ................................................................................................. 8
CDC reactor assay..................................................................................................... 9
Polyvinyl chloride (PVC) plate assay ..................................................................... 10
Biofilm growth in the tube reactor system .............................................................. 11
RESULTS ....................................................................................................................12
Assay of M. marinum biofilm formation using the MBECTM assay system .......... 12
Assay of biofilm formation in PVC 96-well plates ................................................ 14
Growth of M. marinum biofilms on different substrates in the CDC Reactor ........ 16
DISCUSSION ..............................................................................................................19
CHAPTER III – Differential Fluorescence Induction of Mycobacterium marinum
Promoters ..........................................................................................23
INTRODUCTION .......................................................................................................23
METHODS ..................................................................................................................26
Bacterial strains and growth conditions .................................................................. 26
Biofilm growth and harvest in the Tube Reactor system ........................................ 28
Biofilm growth and harvest in the Shaking Flask system ...................................... 29
Planktonic cultures growth and harvest from cell culture flasks ............................ 29
FACS sorting of bacteria ........................................................................................ 30
The Biofilm Fluorescent Library (BFL) sort .......................................................... 30
The Biofilm Fluorescent Pre-sorted Library 3 (BFPSL3) sort ............................... 33
The Biofilm Fluorescent Pre-sorted Library 4 (BFPSL4) sort ............................... 34
i
Flow Cytometry Analysis ....................................................................................... 34
Sequencing of selected clones ................................................................................ 35
RESULTS ....................................................................................................................37
Clones found to be induced during biofilm formation............................................ 37
Clones found to be induced during planktonic growth ........................................... 37
DISCUSSION ..............................................................................................................41
Promoters associated with DNA regulation and general metabolism .................... 41
Promoters associated with membrane processes .................................................... 42
Promoters associated with lipoproteins .................................................................. 43
CHAPTER IV – The In Situ Visualization of Biofilm Induced Promoters .................46
INTRODUCTION .......................................................................................................48
Precedence in the literature ..................................................................................... 48
Rational for our approach ....................................................................................... 51
METHODS ..................................................................................................................57
Bacterial strains and growth conditions .................................................................. 57
Biofilm growth for the in situ visualization ............................................................ 58
In situ visualization by confocal microscopy.......................................................... 58
In situ visualization data analysis ........................................................................... 59
RESULTS ....................................................................................................................61
Time-Course Pictures.............................................................................................. 61
Semi-Quantitative Analysis .................................................................................... 61
DISCUSSION ..............................................................................................................69
The V27R Construct ............................................................................................... 69
Clone OB1-101 ....................................................................................................... 71
Clone OB2-113 ....................................................................................................... 72
Clone OB1-102 ....................................................................................................... 73
The G13R Construct ............................................................................................... 73
Conclusions ............................................................................................................. 74
REFERENCES ............................................................................................................76
ii
LIST OF FIGURES
Figure II.1…………………………..……………………………………………………13
Figure II.2……………………………..…………………………………………………15
Figure II.3……………………………………………..…………………………………17
Figure III.1…...……………………………………..……………………………………25
Figure III.2…...……………………………………..……………………………………27
Figure III.3…...……………………………………..……………………………………32
Figure IV.1…...……………………………………..……………………………………53
Figure IV.2…...…………………………………..………………………………………55
Figure IV.3…...………………………………..…………………………………………63
Figure IV.4…...………………………………..…………………………………………63
Figure IV.5…...………………………………..…………………………………………64
Figure IV.6…...………………………………..…………………………………………68
Figure IV.7…...………………………………..…………………………………………69
iii
LIST OF TABLES
Table III.1…...……………….…………………..………………………………………39
Table III.3…...……………….…………………..………………………………………40
Table III.3.A…..…………….……….…………..………………………………………45
Table III.3.B…..…………….……….…………..………………………………………46
Table III.4…...……………….…………………..………………………………………47
Table IV.1…...……………….………………………………………………..…………65
iv
LIST OF ABREVIATIONS
AHL – N-acyl homoserine lactone
BLAST – Basic alignment search tool
CSLM – Confocal scanning laser microscopy
DFI – Differential fluorescence induction
DPC – Dilution plate count
EPS – Extracellular polymeric substance
KAN – Kanamycin
KEGG – Kyoto encyclopedia of genes and genomes
NTM – Non tuberculosis mycobacteria
MCS – Multiple cloning site
OADC – Oleic acid albumin dextrose complex
oriE – Origin of replication in E. coli
oriM – Origin of replication in M. marinum
PCL2M – Promoter capture library 2 of M. marinum
PDIM – Phthiocerol dimicocerosate
RPI – Relative percent increase
TBSCG – TB structural genomics consortium
v
ACKNOWLEDGEMENTS
All my gratitude goes to Dr. Lucia Barker who gave me the opportunity and
guidance in accomplishing this work; for her invaluable mentorship, patience, and for the
provision of ever stimulating projects in her lab.
I would also like to thank my graduate advising committee members: Dr. Jon
Holy for his training in confocal microscopy and for providing me with much technical
insight; Dr. Mat Andrews for his review and helpful feedback on the project.
Thanks go to The University of Minnesota Duluth Biology and Chemistry
departments for their Graduate Teaching Assistantship support. The opportunity to teach
and interact with students has been my most precious graduate school experience. Thanks
also to the Biology Department and the College of Science and Engineering for travel
grants that allowed me to present parts of this project at international meetings.
Finally thank you to my Grandparents, Helen and Philip Bremmer, who supported
me with encouragements and trust through college and graduate school; to friends and
loved ones, because none of this would have been possible without them.
This work was supported by the N.I.H. grant # AI060862
vi
ABSTRACT
Mycobacterium marinum is a ubiquitous, nontuberculosis mycobacteria which
causes granulomatous infections in poikilothermic animals. This bacterium also causes
“fish tank granuloma” in humans, a disease which presents as granulomatous lesions that
are generally confined to the extremities. Lethal cases have been reported, however, in
immunocompromized individuals. We describe here a study of M. marinum in the most
likely environmental niche of this aquatic organism, the biofilm. We first assayed the
Mycobacterium marinum strain 1218R, a fish outbreak isolate, for biofilm formation
using several different methods and surfaces to determine the most efficient replicable
assay for M. marinum biofilm formation. These data indicated that the silicone was, of all
the material tested, the most effective substratum for the generation of M. marinum
biofilms, and that a silicone tube reactor could be used to generate large masses of
bacterial biofilms for further study. We then implemented a green fluorescent protein
(Gfp) promoter capture library of M. marinum’s genome to screen for differential gene
expression when planktonic bacteria commit to life in a biofilm environment. We
identified biofilm induced genes matching mycobacterial clade members’ sequences
involved in peptide synthesis, gene regulation and metabolic processes. We also
identified a hypothetical adhesion protein broadly conserved among mycobacteria.
Finally, we used confocal scanning laser microscopy in order to detect in situ variances in
Gfp production on selected clones. Using a quantitative analysis method, and comparing
selected clones to a constitutive control, we were able to measure gene induction within
specific sites of the biofilm structure. Clone OB1-101, which contains the promoter from
a putative adhesion protein, was assayed in this system. In transverse sections of an OB1101 biofilm, a gradient of Gfp production was observed that indicated promoter induction
at the loci of biofilm attachment to the glass slide, confirming the predicted function of
this putative adhesion protein. These data indicate that the activity of specific promoters
can be observed in situ to help in identifying the nature of the genes expressed during the
establishment and maintenance of mycobacterial biofilms.
vii
CHAPTER I – General Introduction on M. marinum and Biofilms
In March 2007 the Center for Disease Control (CDC) reported that an estimated
third of the world’s population was infected with tuberculosis (TB). Each year, nine
million new cases of TB are reported and two million people die from the disease. While
most cases are connected with overcrowding and poverty in third world countries, almost
14,000 new cases were reported in the United States in 2006, and after 30 years of
decline, the number of TB cases in the United States increased by 20 percent between
1985 and 1992 (CDC 2007). Most recently a scare was caused by a traveler carrying a
strain of extensively drug-resistant Mycobacterium tuberculosis by plane between the
United States and Europe. Two lessons were underlined by this latest incident. First, that
in spite of authorities efforts, the traveler was able to cross oceans unstopped, showing
the ease and speed at which a diseases can cross borders. Secondly, it has now become
clear that the deadly bacterium is no longer an affliction of the poor; instead, it has
become a real threat across political boundaries and economic status.
Mycobacteria Species
Mycobacterium tuberculosis is probably the most famous of mycobacteria.
Isolated by Robert Koch in 1882, it causes the deadly granulomatous infection of the
lungs known as tuberculosis. Before the Acquired Immunodeficiency Syndrome (AIDS)
virus epidemic, opportunistic, non tuberculosis mycobacteria (NTM) infections were rare
and affected predominately males over sixty. It is now estimated that 25 to 50% of AIDS
patients in Europe and in the United States harbor NTM infections (Falkinham 1996).
1
Pathogenic or innocuous, environmental or host-associated; mycobacteria are
ubiquitous habitants of a wide variety of niches. The Mycobacterium genus is
distinctively acid fast due to characteristic mycolic acids composing the cell wall.
Mycolic acids are long carbohydrate chains responsible for mycobacteria’s
hydrophobicity, their hardiness and the cording phenotype of certain species (Singleton
2001).
M. marinum
Among the NTM, Mycobacterium marinum is a ubiquitous, facultative
intracellular organism. It was originally isolated by Aronson et al. (1926). It grows best at
30-32oC (Davis 2002, Ramakrishnan 1994), although some strains grow well at 25oC and
others up to 37oC (Aronson 1926, Clark 1963). M. marinum is classified as a slow
grower, yet it doubles every 4-6 hours (Rogall 1990, Stahl 1990), This is much faster than
M. tuberculosis which divides every 12-24 hours or M. leprae, which can stay for weeks
without dividing (Mims 1993).
M. marinum can cause granulomatous lesions in fish and poikilothermic animals.
Because of this similarity with how M. tuberculosis infections present in humans, the
disease is also called “fish tuberculosis” (Edelstein 1994, Glukman 1995, Huminer 1986,
Gomez 1993). In fact, the most common infection-causing mycobacterial isolates in fish
are M. marinum, M. fortuitum and M. chelonae (Bercovier 2001). M. marinum outbreaks
can be devastating for piscicultures and can also increase human exposure to the
pathogen (Bercovier 2001, Gomez 1993).
2
Infection in humans is called “fish tank-granuloma” or “swimming poolgranuloma” and is generally traced to environmental water sources. It is usually limited
to the extremities (Ramkrishnan 1997), which is consistent with M. marinum’s lower
optimal growth temperature (Edelstein 1994, Joe 1995, Mollohan 1961). Some cases
however, of more serious, disseminated infections have been reported in
immunocompetent patients (Alloway 1995, Ashford 2003), while infections can be
systemic and lethal in immunocompromized patients (Hanau 1994, Parent 1995). Cases
of mycobacterial infection have been on the increase in the United States as a result of
contact with contaminated water or animals (Dobos 1999) making M. marinum an
emerging zoonotic pathogen.
Biofilms
Biofilms are an organized group of bacteria forming spontaneously on biotic or
abiotic surfaces in aqueous environments. Bacteria in biofilms are typically encased in an
extracellular polymeric substance (EPS) that they secrete (Costerton 1999, Hall-Stoodley
2002). Biofilms are also very dynamic and changing structures. Within the colony,
bacteria exist under different phenotypes allowing microcolonies to form, grow and
detach in order to colonize new surfaces. This optimizes colony survival and contribiutes
to bacterial spread as a whole (Hall-Stoodley 2004). This arrangement is analogous to the
complex and specialized behaviors found among multicellular organisms (Shapiro 1998).
This dynamic behavior is made possible by communication between bacteria within a
biofilm. This communication was shown to be controlled by quorum sensing and
3
autoinducing signals. These include N-Acyl homoserine lactones (AHLs) (Fuqua 1994)
or Rhamnolipids (Epinosa-Urgel 2003, Jensen 2007).
The ability to form biofilms can be an advantage for otherwise free-floating
(planktonic) bacteria. In general, the mere fact that bacteria belong to a structured colony,
even one as heterogeneous as a biofilm, is sufficient to confer protection against a
number of environmental stressors (Barbeau 1998, Costerton 1987). Secreted EPS is yet
another level of protection within the structure, where it can keep in moisture and shield
bacteria from toxic compounds. In this fashion, Biofilms are a bacterial adaptation for
survival in harsh environmental conditions. Protected inside biofilms, bacteria are
resistant to temperature change, desiccation and low nutrient conditions (Stoodley 2002).
They are also notoriously resistant to antibiotics and disinfectants that are otherwise
effective on their planktonic counterparts (Donlan 2002, Costerton 1999).
Finally, the ability to form a biofilm is ubiquitous among bacteria and fungi, and
the nature of the surfaces onto which biofilms form is also similarly varied. Biofouling
and periphyton are forms of biofilms found in the environment. Biofilm can also be
associated with biological surfaces. Examples of biofilms include dental plaque and
growth on medical implants and catheters. Biofilms that present in hospital settings are
often associated with recurrent nosocomial infections (Barbeau 1998).
4
NTM Biofilms
It is believed that the natural niche of environmental NTM species is especially
important given their industrial and medical relevance (Flakinham 1996). NTM are
known to form biofilms on a variety of surfaces such as water pipes and within sewage
sludge (Schulze-Robbeke 1993, Schulze-Robbeke 1995). Mycobacterial biofilms have
also been found in water treatment plants, domestic water supplies (Falkinham 2001,
Schwartz 1998), and dental unit waterlines (Schulze-Robbeke 1995, Schulze-Robbeke
1992, Walker 2000). In a study by Schulze-Robbeke et al. (1995), the researchers found
M. chelonae and M. fortuitum were found in concentrations 3-4 fold higher within
biofilms than growing planktonically in the surrounding aqueous environment. More
recent studies performed with M. fortuitum and M. chelonae have shown that both of
these fast growing mycobacteria are capable of forming biofilms on a wide variety of
surfaces and under a variety of conditions (Hall-Stoodley 1998, Hall-Stoodley 1999,
Hall-Stoodley 2002). These studies suggest the importance of life in a biofilm for
mycobacterial species and indicate that biofilms may be the natural reservoirs for
environmental mycobacteria.
M. marinum Biofilm
M. marinum was also found in water distribution systems (Falkinhmam 2001).
This is key because it implies that the likely natural niche of this organism, like other
environmental mycobacteria, may also be inside biofilms. Isolation techniques from the
natural environment however, typically involve culturing bacteria at 37oC. This could
potentially exclude the detection of slow growing mycobacteria, such as M. marinum,
5
especially those with optimal growth temperatures between 28oC and 32oC. This could
explain why M. marinum had not been detected in some environmental surveys.
Other studies have examined M. marinum biofilm formation on polypropylene
pegs (Bardouniotis 2003). While biofilm formation occurred readily, the colony showed,
however, an increased susceptibility to biocides compared to planktonic bacteria. This
finding is especially surprising because bacteria growing inside a biofilm are typically
several fold more resistant to antibacterial compounds (Chambless 2005).
Relevance of M. marinum
Mycobacterium marinum is an especially relevant organism to study as it causes
infection in humans and economically devastating outbreaks in fisheries. Furthermore,
our group and others believe that M. marinum is an excellent model for the study of M.
tuberculosis (Barker 1998, Pagan-Ramos 1998, Ramakrishan 1997, Talaat 1998). While
M. marinum creates similar granulomatous lesions, belongs to the same genetic clade
(Kaattari 2006, Tønjum 1998), and has a comparable antibiotic susceptibility (Barker
2007), it grows faster and does not require biosafety level 3 lab precautions. In addition,
it is genetically tractable and a searchable genome that has been completely sequenced is
available.
6
CHAPTER II – Quantifying Biofilm Formation
INTRODUCTION
Environmental and pathogenic NTM have been shown to form biofilms on
various surfaces and under varying nutrient conditions. The study of NTM biofilms is
especially important because of the industrial and medical implications for infection by
pathogenic NTM species (Falkinham 1996). Early work by Schulze-Robbeke et al.
indicated that in addition to environmental sources of NTM organisms, artificial surfaces
such as tap water pipes and sewage sludge were colonized by NTM species (SchulzeRobbeke 1993). Other studies of biofilm samples from water treatment plants, domestic
water supplies and dental units by the same group indicated that these biofilms may be an
important site for the replication of NTM bacteria (Schulze-Robbeke 1995, SchulzeRobbeke 1992).
Because of the association of M. marinum with water sources, we sought to more
fully characterize the formation of biofilms by this organism. Various methods were
examined to characterize M. marinum biofilm development on different materials and to
implement simple, replicable techniques to quantify biofilm growth. We investigated
three different procedures and five different materials that supported biofilm growth. We
used both the MBEC™ assay system and a 96-well polyvinylchloride (PVC) plate
colorimetric assay to quantify biofilm formation. In addition, the CDC reactor system
was used to compare M. marinum biofilm growth on High Density Polyethylene,
Polycarbonate, and Silicone coupons. Finally, we have implemented a tube reactor
7
system to generate large amounts of M. marinum biofilms for further biochemical and
molecular studies.
METHODS
Bacterial strains and growth conditions
All biofilm quantification experiments were performed using Mycobacterium
marinum 1218R (ATCC 927), a fish outbreak isolate. Frozen stocks were diluted 1:25 in
7H9 (Difco, BD Biosciences, San Jose, CA) and grown for 10-14 days at 32oC to reach
stationary phase, then stored at 4oC. This stationary phase 4oC stock of M. marinum was
diluted 1:10 and grown to logarithmic phase by incubation for 4 days in 7H9 broth
supplemented with OADC (Oleic acid, Albumin, Dextrose and Catalase; BBL, BD
Biosciences, San Jose, CA) at 32oC. For biofilm formation assays, we inoculated
logarithmic phase organisms into 7H9 without OADC supplement. To determine
bacterial numbers in the inocula, dilution plate counts (DPC) were performed using 7H10
solid media (Difco) supplemented with OADC and 10 g/mL cyclohexamide (SigmaAldrich, St. Louis, MO).
MBEC™ plate assay
Logarithmic phase 1218R was centrifuged 10 min at 14,000 rpm. The supernatant
was discarded and the pellet resuspended in PBS and passed through a 25 gauge needle to
break up bacterial clumps. The number of bacteria in each inoculum was determined by
DPC. Approximately 2.5 x 107 CFU of M. marinum was suspended in 22mL of 7H9 and
added to each MBEC™ assay system (MBEC™ Biofilm Technologies Ltd., Calgary,
8
Alberta). The polypropylene MBEC™ plates were incubated at 32oC on a rocking
platform and the media changed every 3 days. Biofilm formation was determined at each
time point by breaking off 4 pegs with sterile pliers and placing them in individual
Eppendorf tubes. PBS was added to the pegs, the tubes vortexed briefly to remove
planktonic organisms and the PBS removed. Biofilm organisms were detached and
dispersed by adding fresh PBS to each tube and sonicating with a VibraCell sonicator
with cup horn attachment (ModelVCX 500, Sonics, Newtown, CT) for 20 seconds (4
cycles of 5 seconds on/5 seconds off) on ice. The number of organisms was determined
by DPC and expressed as colony forming units (CFU) per peg. The experiment was
repeated 3 times.
CDC reactor assay
Logarithmic phase 1218R was centrifuged 10 min at 14,000 rpm. The supernatant
was discarded and the pellet re-suspended in PBS and passed through a 25 gauge needle
to break up bacterial clumps. The number of bacteria was determined by DPC
(approximately 5.8 x 107 CFU/mL). The CDC Biofilm Reactor (BioSurface
Technologies Corp., Bozeman, MT) consists of 6 polypropylene coupon holders
suspended from a ported lid. Each coupon holder accommodates 3 x 12.7 mm diameter
coupons (High Density Polyethylene (HDP), Polycarbonate (PC) and Silicone). A
volume of 400 mL of 7H9 media in the CDC reactor was inoculated with 1 mL of
organisms and incubated 4 hours at 32oC without stirring to allow for bacterial
attachment. Continuous 7H9 flow was initiated with a peristaltic pump (Ismatec IPC ISM
930, Glattbrugg, Switzerland) at 0.4 mL/min to allow the content of the reactor to be
9
replaced every 4 hours (the minimum division time of M. marinum). At each time point, a
coupon of each material was removed from the reactor, suspended in PBS and vortexed
briefly to remove planktonic organisms. Fresh PBS was added and the coupon sonicated
for 20 seconds (4 cycles of 5 seconds on/5 seconds off). The number of organisms was
determined by DPC and expressed as CFU per mm2. The experiment was repeated 3
times.
Polyvinyl chloride (PVC) plate assay
An ultra violet sterilized 96-well polyvinyl chloride (PVC) plate (BD
Biosciences) was used for an inoculation in double triplicates (6 wells per column) of 1:2
serial dilutions of 1218R in logarithmic growth. The number of bacteria in the undiluted
(1:1) inocula (approximately 2.0 x 106 CFU/mL) was determined by DPC and serial
dilutions performed in 7H9 media. Sample wells were filled with 20 L of diluted
bacteria plus 80 L of 7H9. Wells with 100 L of media alone were used as negative
controls. This assay was a modification of the technique described by Recht et al. (2000).
The PVC plates were incubated at 32oC in a high humidity chamber. After 10 days, each
well was emptied using a multipipetter and 100 L of Carbol Fuchsin (CF) was added to
three of the wells while the remaining three wells were stained with 100 L of Crystal
Violet (CV) stain. The stains were left in the wells for 15 minutes. Each well was then
rinsed with distilled water at least 4 times, until the media-only control wells appeared
clear. The plate was left to dry overnight at room temperature. The CF and CV stains
were dissolved using 100L of DMSO or 100L of 95% EtOH per well, respectively.
Staining was quantified by reading absorbance at λ=580 and λ=570 nm for CF and CV,
10
respectively, on a microplate spectrophotometer (SPECTRAmax®, Molecular Devices
Corporation, Sunnyvale, CA). The experiment was repeated twice.
Biofilm growth in the tube reactor system
An inoculum of approximately 4.5 x 108 CFU of logarithmic phase 1218R was
centrifuged, resuspended in 7H9, and passed through a 25 gauge needle to break up
bacterial clumps. The number of bacteria in the inoculum was determined by DPC. The
tube reactor consists of a continuous sterile silicone line connecting a 7H9 reservoir to a
waste container (Sauer 2002). After 4 hours, flow was initiated at a rate of 0.07 mL/min
allowing for continuous 7H9 media replacement every 4 hours (the minimum division
time of M. marinum). A bubble trap (Flow Break; Cole-Parmer, Vernon Hills, IL) was
used to prevent air bubbles from passing through the tubing and shearing off the biofilm.
1218R biofilm was grown along 1 meter of 4.8 mm (inner diameter) silicone tube (ColeParmer). After 10 days, the tubing was rinsed with PBS to remove unattached planktonic
organisms. The tube was divided into segments and sonicated to remove attached
bacterial cells. The tubing was also scraped with a sterile cotton swab, or a sterile spatula
to remove the attached biofilm. After suspension of the harvested organisms in PBS,
each sample was sonicated for 15 seconds (3 cycles of 5 s on/5 s off). The wet weight of
the bacterial harvest was determined and the number of organisms yielded from each
method was determined by DPC and expressed as CFU per meter. The experiment was
repeated twice.
11
RESULTS
Assay of M. marinum biofilm formation using the MBECTM assay system
The MBECTM assay system (Ceri 1999) is a polystyrene plate with 96 pegs
extending downward into a ridged reservoir. The plate is agitated at a constant speed
after bacterial inoculation. The pegs are broken off at each time point and the peg
washed and sonicated in order to remove and enumerate adherent bacteria. Though not a
continuous culture system, media was replaced every 3 days over the course of each
experiment. As shown in Figure II.1., biofilm formation on the pegs was approximately
3.5 x 104 organisms per peg after 14 days of culture. Although there is some variation
from peg to peg, all experiments exhibited similar growth curves and biofilm formation
appeared to plateau after 7-10 days of biofilm growth in this system.
12
CFU/peg
1.0E+05
1.0E+04
1.0E+03
1.0E+02
0
5
10
15
Time (d)
Figure II.1. Biofilm formation by M. marinum in the MBEC™ plate assay system. Four
pegs were randomly chosen from different sites on the plate. Each time point represents
the average CFU per washed and sonicated peg. Error bars indicate the standard error of
the mean in this representative experiment. Experiment was repeated three times.
13
Assay of biofilm formation in PVC 96-well plates
We chose the method of Recht et al., a 96-well plate assay, to measure biofilm
formation by M. marinum on polyvinylchloride (PVC) after 10 days (Recht 2000). In
addition to the above assay in which wells are stained with crystal violet, washed, and the
stain dissolved in ethanol before quantification with a spectrophotometer, we also
attempted a variation of this assay in which adherent bacteria were stained with Carbol
Fuschin and the stain dissolved with DMSO. As illustrated in Figure II.2., both methods
indicated that an inoculum of approximately 5.0 x 103 M. marinum per well yielded the
most biofilm formation as measured by either assay. Previous experiments indicated that
the sensitivity of the assay required at least 1.0 x 103 CFU/ well for detection of biofilm
formation after 10 days (data not shown). This method, however, gave the greatest
variation both from experiment to experiment and within individual experiments
compared with the MBEC assay system and the CDC reactor method (described above).
14
() OD580, (■) OD570
3.0
2.5
2.0
1.5
1.0
0.5
0.0
1.0E+04
2.0E+04
3.0E+04
4.0E+04
5.0E+04
Number of bacteria inoculated
Figure II.2. Colorimetric quantification of biofilm formation by Mycobacterium
marinum in a 96-well PVC plate. In this representative experiment, the x-axis represents
the initial inocula per well, and the y-axis the optical density reading from the 10 day
biofilm stained with carbol fuchsin (♦) or crystal violet (■).
15
Growth of M. marinum biofilms on different substrates in the CDC Reactor
The CDC Reactor is a continuous culture system where sterile coupons made of
different substrates can be simultaneously assayed for the support of biofilm growth
(Goeres 2005). Coupons made of Silicone, HDP, and PC were attached to holders and
suspended within a batch culture of M. marinum. Each holder contained one coupon of
each material, and the position of the different coupons was varied in each experiment to
insure that differences in biofilm growth were not due to placement within the reactor.
At each time point, coupons were removed, rinsed, and the bacteria enumerated. As
shown in Figure II.3. (a), although all of the materials supported biofilm formation,
silicone exhibited the greatest biofilm growth (3 x 106 cfu/cm2) at the 14 day time point.
This was approximately 6-10 times more adherent bacteria than HDP and approximately
100 times more bacteria than PC after 2 weeks of biofilm growth. Despite differences in
biofilm formation on different surfaces, there was an overall increase in the number of
bacteria on each material over time, and biofilm development appeared to plateau after 7
to 10 days. The amount of biofilm attachment on each substrate was similar from
experiment to experiment and silicone, always supported the most bacterial growth,
followed by HDP and PC.
16
a
7
6.5
Log CFU/cm2
6
5.5
5
4.5
4
3.5
3
0
b
2
4
6
8
10
12
14
16
Time (d)
Figure II.3. (a) Biofilm formation by M. marinum in the CDC reactor assay system. In
this representative experiment, each time point represents the number of CFU per mm2
recovered by harvesting biofilm from washed and sonicated coupons made of silicone
(▲), high density polyethylene (■) and polycarbonate (♦). (b) Gross biofilm formation in
the tube reactor assay system.
17
Growth of M. marinum in a tube reactor system
Based on the finding in the CDC reactor system that silicone supported the most
biofilm growth, a tube reactor system was designed and built based on work previously
published by Sauer, et al. (2002). In these experiments, an inoculum of 4.5 x 108 CFU
was added to 1 meter of sterile silicone tubing and 73 μl/min of 7H9 media flow
maintained throughout the experiment. This flow rate replaced the media in the tubing
every 4 hours. After 14 days of constant flow, the tubing (Figure II.3. (b)) was washed
and scraped. The yield, using a spatula to scrape the tubing before enumeration, was
approximately 1g of wet mass biofilm organisms per meter of 4.8 mm internal-diameter
silicone tubing. These bacteria were then enumerated using DPC and the number of
organisms harvested at 2 weeks was 7.75 x 108 CFU/meter.
18
DISCUSSION
This was the first report of the use of a CDC reactor system to assay
mycobacterial biofilm formation, and the first to compare different biofilm assay systems
with the same strain of mycobacteria (Hall-Stoodley 2006). The characterization of M.
marinum biofilm formation is relevant in that M. marinum is associated with disease in
fish and in humans exposed to fish or to aqueous environments. The advantages of
working with M. marinum are that more and more is being learned about the molecular
and pathogenic characteristics of this organism (Barker 1998, Davis 2002, El-Etr 2004)
and the genomic sequence of M. marinum is being completed. There is also much known
about the interaction of M. marinum with host cells (Barker 1997, El-Etr 2001,
Ramakrishnan 1994) and in animal models (Bouley 2001, Davis 2002, Prouty 2003,
Ruley 2004), but little was known about the ability of this organism to form biofilms.
The collective NTM species are emerging as a problematic and potentially serious
threat to human health (Herdman 2004). These organisms can colonize drinking water
distribution systems (Falkinham 2001, September 2004), dental unit waterlines
(Pankhurst 2003), and even specialized health care equipment (Kressel 2001). The study
of NTM biofilm formation has included, but is not limited to, M. xenopi (Dailloux 2003),
M. phlei (Bardouniotis 2001), M. fortuitum and M. chelonae (Hall-Stoodley 1999) and
recently, M. ulcerans (Marsollier 2004). These potential pathogens likely replicate in
biofilm communities growing on artificial surfaces to which humans are exposed.
Falkingham, et al. (2001) were the first to report the isolation of M. marinum from
biofilms in water distribution systems. Because many studies culture environmental
19
biofilm samples at 37oC, this organism may not be cultured nor identified as frequently as
other pathogenic mycobacteria due to its lower growth temperature.
Bardouniotis et al. (2003) first characterized the growth of M. marinum biofilms
in the MBECTM system with and without exposure to antimicrobial agents, and another
group has quantified M. marinum biofilm formation as compared to other NTM
organisms, including M. ulcerans (Marsollier 2004). The MBEC™ assay system is a
relatively straightforward and highly repeatable assay to measure biofilm formation by
M. marinum. It yielded comparable amounts of biofilm formation on the HDP surface
(as measured by CFU/cm2) within the CDC reactor system. Furthermore, our results
correlate with previous studies using the MBEC™ assay system to study M. marinum
biofilm formation (Bardouniotis 2003). In addition, the Bardouniotis study described a
similar sigmoidal growth curve as seen in Figure II.1. This indicates that M. marinum
biofilm formation is characterized by typical surface growth kinetics that have been
previously demonstrated in studies of other NTM species (Hall-Stoodley 1998). It is
important to note that although the media in the MBEC™ plate is changed regularly, the
assay is not performed under continuous flow as with the CDC reactor system. The
advantages of the MBEC™ assay are consistency and ease of use. This method could
therefore be used to screen a high number of M. marinum strains for the ability to form
biofilms or the formation of M. marinum biofilms in different nutrient or environmental
conditions.
The PVC plate colorimetric assay yielded the least consistent results in
quantifying biofilm growth. Though the advantage of this system is ease of preparation,
data varied from experiment to experiment. The media contained in the assay wells is not
20
replaced over the course of the experiment. In addition, the assay required a large
bacterial inoculum to reach a threshold of sensitivity for visualization of dissolved
bacterial stain. This is most likely due, in part, to the fact that M. marinum is a slow
growing organism with a doubling time of at least 4 hours. This assay is, perhaps, better
suited to bacterial strains with more rapid growth. Other investigators using this assay to
study biofilm formation by M. avium have had success with this system (Carter 2003).
The PVC 96-well plate assay is well suited to measuring initial attachment since later
time points in biofilm development are likely influenced by detachment of mature
biofilms. Our results, however, did clearly indicate that CF was superior to CV for
assessing biofilm formation by M. marinum, consistent with CF binding to lipids rather
than protein moieties in the bacterial cell wall.
The CDC reactor consists of various coupons inserted into holders suspended in a
continuous flow of rotating media. Our laboratory was the first to study M. marinum
biofilm formation on different materials using this assay. We found that all coupons
tested in the CDC reactor yielded detectable biofilm formation, but silicone was
consistently a better surface for growth of M. marinum biofilms. After 14 days, it yielded
six times more biofilm CFU's than HDP, and over one hundred and twenty times more
biofilm CFU's than PC. In each case, we found that after a period of rapid proliferation, a
stable biofilm was formed by 14 days post-inoculum. The CDC reactor was a very
successful method to assay M. marinum biofilm formation under continuous flow
conditions. Another advantage of this apparatus is that the CDC reactor allows the
simultaneous testing of various materials for biofilm formation and accumulation. This
21
method could therefore be implemented to screen specific surface materials for
mycobacterial adherence and biofilm formation in continuous flow.
Based on results from the CDC reactor indicating that, of those tested, silicone is
the best material for biofilm growth, we designed a silicone tubing reactor similar to that
described for other bacterial biofilm systems (Sauer 2002). We were able to harvest large
quantities of M. marinum biofilm from the tube reactor system. When biofilm
development was enumerated for the tube reactor after 10 days of biofilm growth (5.1 x
106 CFU/ cm2), it was comparable to the biofilm formation quantified on the silicone
coupon suspended in the CDC reactor system (3.0 x 106 CFU/ cm2) for 14 days. The
high yields of bacterial mass and numbers from the tube reactor system has enabled
further molecular and biochemical characterization of biofilms formed by mycobacterial
species. These studies were of particular interest in that the ability to grow and sustain
biofilm growth in the laboratory was the preliminary work necessary in order to perform
the genetic and confocal microscopic studies further explained in this work.
22
CHAPTER III – Differential Fluorescence Induction of
Mycobacterium marinum Promoters
INTRODUCTION
We showed in Chapter II that we were able to reproducibly grow Mycobacterium
marinum biofilm using a variety of methods. M. marinum biofilm formed and was
maintained on all surfaces tested; with more hydrophobic surfaces being the most
conducive to biofilm development (Hall-Stoodley 2007). Using these practical findings,
we were interested in better understanding the genetic mechanisms of biofilm formation
by M. marinum. For this purpose we implemented a promoter capture assay in order to
perform a non-exhaustive differential fluorescence induction (DFI) screen for promoters
induced during biofilm formation.
Using a reporter gene assay to measure levels of promoter activity has commonly
been used and reported in the literature (Valdivia 1998). This methods typically involves
the use of a “promoter trap”. That is, genomic DNA from the organism of interest is
digested by restriction enzymes to collect fragments of a desired size range. Fragments
are then cloned upstream of a promoterless reporter gene such as gfp or lacZ (Valdivia
2000). The library formed is called a promoter capture library because some of these
fragments may cause the expression of the reporter gene, indicating a promoter within.
The expression of the reporter gene is contingent to the promoter being both activated
and reading into the promoterless reporter gene. Figure III.1 shows a diagram of the
principle behind the promoter trap assay. This very powerful method allows for promoter
activity testing in a variety of organisms (Slauch 1994).
23
We used the optimized green-fluorescing protein produced by gfpmut2, described
by Cormack and coworkers (1996). This Gfp variant is 20 times brighter than the wild
type Gfp when excited by a 488nm laser beam. It is also conveniently expressed in both
mycobacteria and E. coli via the shuttle vector pFPV27 developed by Valdivia et al.
(2000). Using Gfp as a reporter protein allowed us to use flow cytometry in order to
measure variances in promoter activation (Valdivia 1996). This method is perfectly suited
to work with a large promoter capture library as it has the advantage of providing a rapid
measurement of fluorescence as well as the possibility to sort bacteria according to this
criterion. Furthermore, this screening method has successfully been used in the past to
localize promoters activated in M. marinum during trafficking inside macrophages
(Barker 1997) or during granuloma formation (Ramakrishnan 2000).
24
Figure III.1. Principle behind the promoter trap assay. (A) Genomic DNA is excised at
random using restriction endonucleases. (B) Fragments may contain the promoter region
of a gene. (C) The Genomic DNA fragments are ligated intothe plasmid pFPV27
containing a reported green fluorescent protein (gfp) gene. Gfp is only produced if the
insert contains an activated promoter and the promoter is reading into the gfpmut2 gene.
25
METHODS
Bacterial strains and growth conditions
All experiments were performed using Mycobacterium marinum 1218R (ATCC
927), a fish outbreak isolate. The PCL2M Library used to screen for biofilm-induced
clones was the promoter capture library constructed by Barker et al. (Barker, 1998) using
M. marinum 1218R. A schematic illustrating the making of the PCL2M library can be
found in Figure III.1. A 1mL frozen sample of PCL2M clones was diluted 1:24 in 7H9
(Difco, BD Biosciences, San Jose, CA), supplemented with OADC (BBL, BD
Biosciences, San Jose, CA), and grown without KAN for 5 days at 32oC to allow for
library recovery. The culture was then centrifuged 20min at 1,500xg. The supernatant
was discarded, the pellet was re-suspended in 1mL PBS and diluted 1:24 in 7H9/OADC
supplemented with 30g/mL KAN (FisherBiotech, Fisher Scientific, Fair Lawn, NJ). The
culture was grown 10d to stationary phase, refrigerated at 4oC and used as stock culture
for all other experiments. The negative control clone used for all experiments was V27R
which is the 1218R strain containing an insertless vector. V27R produces a constant
baseline amount of Gfp by a phenomenon of read-through. Clone G13R as described by
Barker et al. (1998, 1999), is always highly fluorescent and was used as a positive control
in the differential fluorescence induction assay. Stationary stock solutions of V27R and
G13R were obtain by incubating the strains 1:24 in 7H9/OADC/KAN for 10 days. For all
assays, 1mL samples of stationary stock cultures were incubated in 24mL of
7H9/OADC/KAN for 4d to reach mid logarithmic phase. To determine bacterial numbers
in the inocula, dilution plate counts (DPC) were performed using 7H10 solid
26
27
Figure III.2. Promoter capture library (PCL2M) construction. (A) The M. marinum genome was extracted and cut with
endonuclease Sau3A. (B) Genomic fragments, 200bp - 1kbp long, were ligated upstream of the promoterless gfpmut2 gene and
transformed into E. coli for amplification on selective media containing KAN. (C) Plasmids were extracted and transformed into
M. marinum to produce the PCL2M library.
media (Difco) supplemented with OADC, 10g/mL cyclohexamide (Sigma-Aldrich, St.
Louis, MO), and 30g/mL KAN.
Biofilm growth and harvest in the Tube Reactor system
The tube reactor assay used for this experiment was a modification of that
described by Sauer et al. (2004). An inoculum of approximately 4.5 x 108 CFU of
logarithmic phase bacteria was centrifuged, re-suspended in 7H9, and passed through a
19ga needle to break up bacterial clumps. The number of bacteria in the inocula was
determined by DPC. The tube reactor consists of a continuous sterile silicone line
connecting a 7H9 reservoir to a waste container. The biofilm was grown along a 1 meter
portion of 4.8mm (inner diameter) silicone tube (Cole-Parmer). Bacteria were inoculated
using a 25ga needle through the proximal end of the 4.8mm ID silicone tube. After a 4
hour attachment period of bacteria without flow, an Ismatec peristaltic pump (ISMATEC,
Glattbrugg, Switzerland ) was used to initiate continuous media replacement
(0.073mL/min) within the line. This allowed the content of the reactor to be replaced
every 4h (minimum division time of M. marinum). A bubble trap (Flow Break; ColeParmer, Vernon Hills, IL) was used to prevent air bubbles from passing through the
tubing. After 10 days, the tubing was rinsed with 5mL PBS to remove unattached
planktonic organisms. Under sterile conditions, the silicone tube was cut lengthwise and
the biofilm scraped off with a metal spatula into 10mL of sterile PBS/0.01%SDS. The
tube containing the biofilm harvest was then vortexed for 30s and sonicated using a
VibraCell sonicator with cup horn attachment (Model VCX 500, Sonics, Newtown, CT)
28
in 5s on/ 5s off cycles until no bacterial clumps were visible. 1mL samples of this
solution were used for flow cytometry experiments.
Biofilm growth and harvest in the Shaking Flask system
A 1mL inocula averaging 6.16x106 CFU of logarithmic phase bacteria was mixed
with 9mL of 7H9 supplemented with KAN and placed in a 25cm2 polystyrene cell culture
flask (Corning, Corning, NY). Each flask was incubated for 14 days on an incubator
shaker at 130rpm (Model C24, New Brunswick Scientific, Edison, NJ). At days 3, 6, and
9, all but 1mL of media was emptied using a sterile suction apparatus. 9mL of fresh 7H9
supplemented with KAN was then added before the flasks were returned to the shaking
incubator. Biofilm harvest was executed by scraping the side of the flasks with a sterile
cotton swab, leaving behind cells accumulated at the air/media interface. The biofilm
samples were eluted from the swab into 0.5mL of PBS/0.01% SDS, spun down at
10,000rpm for 4min, then re-suspended in 0.5mL of PBS/0.01% SDS. Each sample was
sonicated for 30s in 5s on/ 5s off cycles before analysis by flow cytometry.
Planktonic cultures growth and harvest from cell culture flasks
Planktonic cultures were grown in similar conditions as in the shaking flask
system except that each flask was shaken once daily instead of continuously. After 14
days, the planktonic content of the flask was harvested, spun down at 1,500xg for 20min
and re-suspended in 0.5mL of PBS/0.01% SDS. Each sample was sonicated for 30s in 5s
on/ 5s off cycles before analysis by flow cytometry.
29
FACS sorting of bacteria
Bacteria were sorted using a Flow Activated Cell Sorter (FACS) Aria flow
cytometer (BD Biosciences, San Jose, CA). Bacterial samples harvested from planktonic
cultures, the tube reactor, or from the shaking flask system were passed through a 25ga
needle before sorting in order to reduce clump size. Samples were gated by Forward
Scatter and Side Scatter with a photomultiplier to select for appropriate M. marinum size
and complexity. Samples were then graphed with Side Scatter on the x-axis and GFP-A
on a logarithmic scale y-axis using a 530/30BP filter. To determine background
fluorescence, a sample of wild type Mycobacterium marinum 1218R (ATCC 927) or the
vector alone transformant V27R, were used as a negative controls. Fluorescence was
defined as single events ranging between 2 and 500 fold above background fluorescence.
Sorting of mycobacteria was done at 35psi using a 70m nozzle.
The Biofilm Fluorescent Library (BFL) sort
A logarithmic PCL2M culture (inoculum: 2.1x1010 CFU) was grown in a tube reactor
biofilm system for 10d then sorted by FACS (Figure III.3.a). Hyper-fluorescent bacteria
(10-200 fold above 1218R background) and fluorescent bacteria (2-9 fold above
background) were sorted into two separate 15mL tubes containing 1mL of PBS/0.01%
SDS. Cells were re-suspended by adding 4mL of PBS/SDS to each tube and then
vortexing for 30s. Tubes were spun down at 1,500xg for 20min. The pellet was resuspended in PBS/SDS and the content of the tube was plated for single colonies on
7H10/OADC supplemented with Cyclohexamide and KAN. Libraries were labeled
BFL3+ (for biofilm hyper-fluorescent clones) and BFL+ (for biofilm fluorescent clones).
30
After 7 days of incubation, BFL3+ and BFL+ individual colonies which expressed the
lowest amounts of fluorescence on the plates were separated onto individual plates and
incubated for 7 days. Two swab samples were then taken from each plate. The first
samples were mixed with 1mL of freezing media (7H9/10% OADC/20% Glycerol) in
cryogenic tubes, quick frozen in liquid nitrogen and placed in the freezer at -80oC. The
second samples were put in a tissue culture flask containing 10mL of 7H9/OADC
supplemented with KAN, incubated for 10 days, then refrigerated at 4oC and used as
stock cultures for biofilm assay experiments.
31
Figure III.3. Flow chart describing how the three “biofilm fluorescent clone” libraries
were made. PCL2M is the promoter capture library constructed with the pFPV27 vector
and M. marinum total genome. BF stands for biofilm growth. PK stands for planktonic
growth. FACS stands for flow activated cell sorting. BFL and BFPSL stand respectively
for biofilm fluorescent library and biofilm fluorescent presorted library. BFPSL libraries
were presorted by FACS enrichment for non-fluorescent phenotype during planktonic
growth. All three libraries were sorted by FACS for fluorescent phenotype when grown
in a biofilm. Biofilms in (a) and (b) were grown in the silicone tube reactor system. The
biofilm in (c) was grown using the shaking flask system.
32
The Biofilm Fluorescent Pre-sorted Library 3 (BFPSL3) sort
A second set of clones was isolated using a planktonic PCL2M culture pre-sorted
by FACS for no fluorescence above V27R (Figure III.3.b). Approximately 4.4x105
planktonic bacteria expressing no fluorescence above V27R background were collected
into 1mL PBS/SDS and plated on 7H9/OADC/Cyclo/Kan as described above. This new
library was labeled PSL3 (for pre-sorted library 3). After 7 days of incubation, bacterial
lawns were collected from each plate into 40mL of 7H9, sonicated for 20s (4 x 5s on / 5s
off cycles) and vortexed until no clumps were apparent. The tube was spun down at
1,500xg for 20min, the pellet re-suspended in 10mL of freezing media, and divided in
1mL aliquots into freezing tubes. Each tube was quick frozen in liquid nitrogen and
placed in the freezer at -80oC. An aliquot was grown 4d to log phase (inocula: 7.67x109
CFU), cultivated in a Tube Reactor biofilm system for 10d, as described above, then
sorted by FACS. Hyper-fluorescent bacteria (10-200 fold above background) and
fluorescent bacteria (2-9 fold above background) were sorted into two separate 15mL
tubes containing 1mL of PBS/0.01% SDS. 4mL of PBS/SDS was added to each tube then
vortexed for 30s. Tubes were spun down at 2,700rpm for 20min. The pellet was
resuspended in PBS/SDS and the content of the tube was plated for single colonies on
7H10/OADC supplemented with Cyclohexamide and KAN. This new library was labeled
BFPSL3 for Biofilm Fluorescent Presorted Library 3. After 7 days of incubation,
BFPSL3 individual colonies which expressed the lowest amounts of fluorescence on the
plates were separated onto individual plates and incubated for 7 days. Two swab samples
were then taken from each plate. The first samples were mixed with 1mL of freezing
media (7H9/10% OADC/20% Glycerol) in cryogenic tubes, quick frozen in liquid
33
nitrogen and placed in the freezer at -80oC. The second samples were put in a tissue
culture flask containing 10mL of 7H9/OADC supplemented with KAN, incubated for 10
days, then refrigerated at 4oC and used as stock cultures for Biofilm Assay experiments.
The Biofilm Fluorescent Pre-sorted Library 4 (BFPSL4) sort
This sort was a modification of the BFPSL3 sort in that the PCL2M biofilm
culture was grown in the shaking culture flask assay instead of the tube reactor system
(Figure III.3.c). All other steps in the protocol were performed as described in the
previous section.
Flow Cytometry Analysis
BFL, BFPSL3 and BFPSL4 clones were grown in parallel between the shaking
flask biofilm assay and the SAD planktonic flask assay. Flow cytometry experiments
were performed to compare fluorescence levels between biofilm and planktonic cultures
using a FACSCalibur flow cytometer (BD, Franklin Lakes, NJ). Results were analyzed
using the manufacturer’s CellQuestTM software. 10,000 events for each Shaking Flask
biofilm and its parallel planktonic culture were gated by Forward Scatter and Side Scatter
with a photo multiplier to select for appropriate M. marinum size and complexity. Gated
samples were then graphed with Side Scatter on the x-axis and GFP-A on a logarithmic
scale y-axis. Biofilm and planktonic samples of V27R and G13R were used respectively
as negative and positive controls for each experiment. Geometric means were determined
using the analysis software by selecting a sample of the population containing at least 80
percent of the total gated events. The ratio between biofilm and planktonic geometric
34
means for each clone gave an indication of GFP induction. Each clone was tested in
triplicate.
Sequencing of selected clones
Plasmids were extracted from M. marinum clones of interest using a Beadbeater
plasmid mini-prep (Biospec Products, Bartlesville, OK). Samples were beaten at the
maximum setting for one minute, using 22µm beads. Plasmids were isolated using the
Wizard® Plus SV Minipreps DNA Purification System (Promega, Madison, WI).
Plasmids were then transformed and amplified in library efficient DH5α competent cells
following the manufacturer instructions (Invitrogen, Carlsbad, CA). Plasmids were then
extracted from stationary phase E. coli using the Wizard® Plus SV Minipreps DNA
Purification System. Plasmids and DNA primers flanking each side of the insert were
sent for sequencing to the Biomedical Genomic Center, Sequencing and Genotyping
Facility on the University of Minnesota, Twin Cities campus (Minneapolis, MN). The
primers used were LUF15 (TTGTAGTGCTTGTGGTGGCATC), 27bp upstream of the
insert and LUR15 (TAAGCTTGATATCGAATTCCTGC), 9bp downstream of the insert
on pFPV27. When inserts exceeded 800bp in size, new primers were designed and the
samples were re-sent in order to completely sequence the insert. Sequencing results in
ABI file format were aligned and base calling was verified using the freeware version of
SeqAssem (SequentiX, Klein Raden, Germany). We used Basic Alignment Search Tool
(BLAST) against the M. marinum genome (Sanger Institute, Cambridge, United
Kingdom) to determine the nature of DNA sequences directly downstream of each insert.
Using the BLAST results, insert sequences and approximately 1Kbp of M. marinum
35
DNA immediately downstream of the homologous region were analyzed for the presence
of promoters and open reading frames (ORFs). This information allowed us to match
differentially induced promoters and their corresponding genes in M. marinum. The same
fragments were then compared to known protein sequences using BLAST from the
National Center for Biotechnology Information (NCBI, Bethesda, MD). Finally, we
compared our DNA sequence specifically against M. tuberculosis proteome using the TB
Structural Genomics Consortium BLAST (University of California, Oakland, CA).
36
RESULTS
Clones found to be induced during biofilm formation
A total of ten clones out of approximately 500 clones from the BFL, BFPSL3 and
BFPSL4 libraries were found to be reproducibly induced during biofilm formation. A
summary of the amount of induction per run and predicted protein function can be found
in Table III.1. We found that the non-presorted BFL library did not yield any positive
results during this screen and all clones induced during biofilm growth come from the
presorted library, BFPSL3 and BFPSL4. The number of fold increase represents for each
clone the amount of biofilm induction above the V27R control ratio. Clone and control
ratios were calculated by dividing geometric mean fluorescence values during biofilm
growth after 14 days, by geometric mean fluorescence values during planktonic growth
after 14 days. The number of fold increase was calculated by dividing each clone ratio by
the ratio of the control (V27R).
Clones found to be induced during planktonic growth
A total of five clones were found to be reproducibly induced during planktonic
growth. A summary of the amount of induction per run and predicted protein function
can be found in Table III.2. We found in this category clones from all three libraries. The
number of fold decrease represents for each clone the amount of planktonic induction
below the V27R control ratio. The ratios were found for each clone by dividing
geometric mean fluorescence values during biofilm growth after 14 days by geometric
mean fluorescence values during planktonic growth after 14 days. Clone and control
ratios were calculated by dividing each clone ratio by the ratio of the control (V27R). If
37
individual clone ratios were below one, then the control ratio was divided by the inverse
of the ratio for the particular clone to calculate number of fold decrease.
38
Clones Induced During Biofilm Formation
Clone
OB1-1
OB1-57
OB1-75
A
OB1-93
OB1-101
OB1-102
OB2-2
OB2-11
B
OB2-14
OB2-113
Fold Increase
(3 runs)
1.40
1.96
1.36
1.45
1.37
0.6
1.66
1.20
1.46
1.41
1.26
1.21
2.40
28.86
1.40
41.30
65.30
1.29
1.52
1.27
1.29
1.28
1.40
1.17
1.42
1.73
1.42
1.47
1.96
1.79
Predicted function
Carboxylesterase LipT
Class*
Mb Lp
X
X
Re
X
NADH Dehydrogenase
NADH Quinone
oxydoreductase
Sec-independent protein
translocase
Transmembrane protein TatC
Transmembrane
serine/threonine kinase, PknL
Hypothetical protein
Recombination protein, RecR
Hypothetical cell adhesion
protein
Conserved hypothetical
protein among mycobacteria
Transcriptional attenuator
X
X
X
X
Lipoprotein, LprB
X
X
X
X
X
D-amino acid oxidase, Aao-1
X
X
Glucose/Sorbose
dehydrogenase
Conserved lipoprotein, LppZ
X
Hypothetical Transmembrane
prothein
X
X
X
Table III.1. Relative percent increase in fluorescence of clones induced during biofilm
formation. (A) Clones from the BFPSL3. (B) Clones from the BFPSL4. The number of
fold increase represents for each clone amount of biofilm induction above the V27R
control. (Number of fold increase = (clone ratio / V27R ratio)). * Mb = membrane
associated, Lp = Lipoprotein, Re = DNA regulation and general metabolism
39
Clones Induced During Planktonic Growth
Clone
A OB1-64
(PK)
OB2-7
(PK)
OB2-23
B (PK)
OB2-71
(PK)
OB-29
C (PK)
Fold Decrease
(3 runs)
-4.27
-3.70
-1.97
-2.09
1.08
-1.98
-2.16
-2.01
-1.01
1.21
-1.4
-1.97
-3.76
-2.59
-4.18
Predicted function
Class*
Mb Lp Re
Transcriptional Regulator
Transcriptional Regulator MarR
X
Regulatory protein, TetR-N
Transcriptional regulator, AcrR
Conserved hypothetical protein
X
Conserved hypothetical
Lignin peroxidase, LipJ
Adenyl cylclase
Putative protein
Conserved hypothetical protein
X
X
X
Table III.2. Relative percent increase in fluorescence of clones induced during
planktonic formation. (A) Clones from the BFPSL3. (B) Clones from the BFPSL4. (C)
Clones from the BFL. The number of fold increase represents for each clone amount of
planktonic induction below the V27R control. (Number of fold increase = (clone ratio /
V27R ratio). * Mb = membrane associated, Lp = Lipoprotein, Re = DNA regulation and
general metabolism
40
DISCUSSION
The results of this genetic study were non-exhaustive. However, there were clear
patterns observed with the promoters expressing differential induction when bacteria
were grown in a biofilm compared to planktonic growth (Tables III.1 and III.2 last
column). A more detailed description of the clones induced during biofilm formation or
planktonic growth can be found in Tables III.3.A-B and in Table III.4 at the end of this
chapter. In addition to the predicted functions, functional hierarchies (TBSGC 2007,
KEGG 2007) and ProKnow functions predictions (TBSCG 2007) are listed.
Promoters associated with DNA regulation and general metabolism
Mechanisms of metabolism regulation play a key role in biofilm formation,
maintenance and survival. Chambless and coworkers (2005) explore the metabolic
processes accounting for biofilm resistance to microbial agents. They describe those as
influencing growth, diffusion through the structure, detachment of planktonic cells from
the colony, and cell death. It is not surprising, given the importance of metabolism
regulation in biofilms that nine out of ten clones induced during biofilm formation, and
four out of five clones induced during planktonic growth were associated with gene
regulation and general metabolic mechanisms. Particularly interesting in the literature is
the observation that cells in deeper biofilm layers enter a state of reduced metabolic
activity (Sternberg 1999). This phenomenon is associated in part with reduced nutrients
diffusion through the biofilm mass (De Beer, 1994). In this context, it is particularly
interesting to notice that Clone OB1-102, corresponding to a promoter in M. ulcerans, M.
avium and M. tuberculosis associated with a transcriptional attenuator protein, is the
41
clone that showed the highest level of induction during biofilm formation (up to 6,500
percent above the V27R control).
Promoters associated with membrane processes
In our screen we found six promoters corresponding to proteins associated with
the cell membrane; all of which are induced during biofilm formation. Most are also
associated with metabolic functions, some are lipoproteins and one is a hypothetical
protein. Interestingly, while most membrane-bound proteins found during our screen
were specific to M. marinum clade members, M. ulcerans and M. tuberculosis, OB1-101
was conserved broadly across bacterial species.
Molecules associated with the membrane induced during biofilm formation seem
especially relevant in the context of biofilm attachment to surfaces and cell to cell
adhesion within the structure. In fact, the ProKnow prediction method, capable of
inferring protein function by analyzing their structure (Pal 2005), shows by the frequency
of ontologies from three-dimensional folds that the promoter in OB1-101 is associated
with a similar protein in M. tuberculosis which possesses characteristics of cell adhesion
motifs. Furthermore, this promoter is activated almost 3,000 percent more than the V27R
control. The hypothetical protein associated with clone OB2-113 also shows such motifs
with up to 96 percent increased compared to the V27R control.
In the literature, adhesion molecules are also tightly associated with biofilm
formation. It is the case for diarrheagenic and enterotoxigenic E. coli for example
(Sherlock 2004, 2005) and for P. aeruginosa biofilm forming in the lung (Prince 1992).
Interestingly, all those studies make a link between adhesins, biofilm formation and
42
virulence. Such study could therefore be designed to look at the role played by the genes
associated with the promoters in OB1-101 and OB2-113. Knock-out experiments would
explore both the mutant ability to form biofilms and infectivity the animal host.
Promoters associated with lipoproteins
Another important trend noticed in this screen is the recurrence of lipoproteins.
Three in ten clones induced during biofilm formation, and one in five clone induced
during planktonic growth in our study corresponded to a promoter associated with a
lipoprotein.
In the literature, lipids play key roles in mycobacteria. For example, mycolic acids
are responsible for the cording phenotype (Glickman 2000). Importantly, this cording
phenotype is shared by Mycobacterium marinum and is associated with virulence in
Mycobacterium tuberculosis (Middlebrook 1947, Darzin 1956, Pierce 1956). Other lipid
based molecules such phthiocerol dimicocerosate (PDIM), associated with the cell wall
of pathogenic bacteria, have also been involved in M. tuberculosis virulence in the mouse
model (Camacho 1999, Cox 1999). Particularly interesting is the recent review from
Karakousis et al. (2004) indicating that membrane-associated lipids play an
immunosuppressant role in pathogenesis. Since all of the lipids associated with biofilm
growth found in our study are membrane-associated, this warrants the question as to
whether or not a correlation exists between biofilm formation and virulence. Such
correlations have indeed been found. Rhamnolipids for example are quorum sensing
molecules secreted by Pseudomonas aeruginosa. It has been shown that these molecules
play a role in priming surfaces to allow for P. aeruginosa biofilm colonization. Other
43
studies such as the one by Ojha and coworkers (2005) found a link between the inability
to maintain a mature biofilms in Mycobacterium smegmatis and the loss of a gene
involved in cord factor biosynthesis. Since mycolic acids such as cord factor can be
involved in both virulence (M. tuberculosis) and biofilm formation(M. smegmatis), we
can propose that the membrane-bound lipids produced during biofilm formation may also
be associated with the virulence of M. marinum. This is especially relevant considering
that the biofilm is the likely environmental niche of this pathogenic environmental
isolate.
44
45
Conserved
among
mycobacteria
Conserved
across bacterial
species
M. ulcerans
M. ulcerans
M. tuberculosis
M. ulcerans
M. avium
M. tuberculosis
M. ulcerans
M. tuberculosis
Organism
MUL_2613
MAV_4232
Rv3267
YP-907792
Rv3716c
ABL05674
YP_906168
Rv2094c
YP_906290
YP_883197
Rv3157
YP_906112
Rv2045c
Designation
Conserved hypothetical
Transcriptional attenuator
Conserved hypothetical
Recombination protein, RecR
Conserved Hypothetical
Transmembrane
serine/threonine kinase,
PknL
NADH dehydrogenase
NADH Quinone
oxydoreductase
NADH dehydrogenase
Sec-independent protein
translocase
transmembrane protein,
TatA
Carboxylesterase, LipT
Probable carboxylesterase
LipT
Predicted function
Protein binding
Cell adhesion
Cell growth/maintenance
DNA regulation
Proteolysis and peptidolysis
Signal transduction
(1) Prokaryotic transcription
factor**
(2) Other transcript. Factors**
(3) Transcript. attenuation
protein**
N/A***
(1) Transferase**
(2) Protein-serine/threonine
kinase**
(3) Non-specific protein kinase**
(1) Conserved hypotheticals*
N/A***
Oxydoreductase activity
Mitochondrial electron
transport
ATP synthesis
Hydrolase activity
Catalytic activity
Cell Adhesion
ProKnow Function Prediction
(TBSGC)*
(1) Genetic info. Processing**
(2) Folding, sorting, degradation**
(3) Protein export**
Functional Hierarchy
(TBSGC)* (KEGG)**
(1) Macromolecule metabolism*
(2) Degradation of
macromolecules*
(3) Esterase and lipase*
(1) Small-molecule metabolism*
(2) Energy metabolism*
(3) Respiration*
(4) Aerobic*
Table III.3.A. BLAST results for the clones induced during biofilm formation (Part A): Presorted planktonic library screened for biofilm
fluorescent clones after growth in the silicone tube reactor. * TBSGC = Mycobacterium tuberculosis Structural Genomic Consortium
(http://www.doe-mbi.ucla.edu/TB/). ** KEGG = Kyoto Encyclopedia of Genes and Genomes (http://www.genome.jp/kegg/). ***N/A =
Not Available.
OB1-102
OB1-101
OB1-93
OB1-75
OB1-57
OB1-1
Clone
Blast Results for the Clones Induced During Biofilm Formation (Part A)
46
M. ulcerans
M. ulcerans
M. tuberculosis
M. ulcerans
M. ulcerans
M. avium
M. tuberculosis
Organism
YP_905997
YP_905867
Rv3006
YP_905624
YP_907541
YP_880665
Rv1274
Designation
Hypothetical Transmembrane
protein
Conserved lipoprotein LppZ
Probable Conserved
lipoprotein LppZ
D-amino acid oxidase, Aao-1
Lipoprotein, LprB
Lipoprotein, LprB
Possible Lipoprotein LprB
Predicted function
N/A***
(1) Macromolecule metabolism*
(2) Cell Envelope*
(3) Lipoproteins (lppA-lprO)*
(1) Metabolism**
(2) Amino acid metabolism**
(3) D-arginine metabolism**
Functional Hierarchy
(TBSGC)* (KEGG)**
(1) Macromolecule metabolism*
(2) Cell Envelope*
(3) Lipoproteins (lppA-lprO)*
N/A***
N/A***
N/A***
N/A***
ProKnow Function Prediction
(TBSGC)*
Table III.3.B BLAST results for the clones induced during biofilm formation (Part B): Presorted planktonic library screened for biofilm
fluorescent clones after growth in shaking flasks. * TBSGC = Mycobacterium tuberculosis Structural Genomic Consortium
(http://www.doe-mbi.ucla.edu/TB/). ** KEGG = Kyoto Encyclopedia of Genes and Genomes (http://www.genome.jp/kegg/). ***N/A =
Not Available.
OB2-113
OB2-14
OB2-11
OB2-2
Clone
Blast Results for the Clones Induced During Biofilm Formation (Part B)
47
MAV_5125
MUL_2284
YP_001133
Rv3120c
M.vanbaalenii
M. avium
M. ulcerans
M. gilvum
M. tuberculosis
M. ulcerans
M. tuberculosis
OB2
7
OB2
23
OB2
71
OB
29
MUL_2420
Rv3130c
YP_951799
YP_908307
YP_955602
YP_884949
M. ulcerans
M.vanbaalenii
M. smegmatis
OB1
64
Conserved hypothetical
protein
Hypothetical protein
Adenylate/guanylate cyclase
Conserved hypothetical
Hypothetical protein
Regulatory protein, TetR
Transcriptional Regulator
Transcriptional Regulator
MarR family
Predicted function
(1) Conserved hypotheticals*
(1) Oxidoreductase**
(2) Peroxide acceptor**
(3) lignin peroxidase**
N/A***
Functional Hierarchy
(TBSGC)* (KEGG)**
(1) Prokaryotic transcript. factor**
(2) MarR family transcript.
regulators**
(1) Prokaryotic transcript. factor**
(2) Helix-turn-helix TetR/AcrR
family**
Biosynthesis
Transport
metabolism
N/A***
N/A***
N/A***
N/A***
ProKnow Function Prediction
(TBSGC)*
Table III.4. BLAST results for the clones induced during Planktonic Growth: (A) Presorted planktonic library screened for biofilm
fluorescent clones after growth in the silicone tube reactor. (B) Presorted planktonic library screened for biofilm fluorescent clones
after growth in shaking flasks. (C) Planktonic library screened for biofilm fluorescent clones after growth in the silicone tube reactor. *
TBSGC = Mycobacterium tuberculosis Structural Genomic Consortium (http://www.doe-mbi.ucla.edu/TB/). ** KEGG = Kyoto
Encyclopedia of Genes and Genomes (http://www.genome.jp/kegg/). ***N/A = Not Available.
C
B
A
Designation
Organism
Clone
Blast Results for the Clones Induced During Planktonic Growth
CHAPTER IV – The In Situ Visualization of Biofilm Induced Promoters
INTRODUCTION
Differential fluorescence induction (DFI) allowed us to identify promoters
selectively turned on when Mycobacterium marinum was grown in a biofilm. In this next
set of experiments, we were interested in finding out if we could use the same Gfp
reporter construct, this time to witness in situ the loci of promoter activation within the
biofilm structure.
Precedence in the literature
Constitutive Gfp Constructs
Bacterial constructs constitutively producing green fluorescent protein have
commonly been used in optical microscopy. Researchers have applied this method to
show bacterial interactions during multi species biofilm formation (Skillman 1998). The
method yields a straightforward way to tell bacterial species apart in situ and convincing
two-dimensional views of forming biofilms. Others have used Gfp producing bacteria to
distinguish cells from the extracellular polymeric substance (EPS) forming the biofilm
matrix. Kuehn et al. (2001), or more recently, Maeyama and coworkers (2004), have used
this method with Sphingomonas sp. and Escherichia coli respectively to observe biofilm
monoculture. In their studies, Gfp producing bacteria and stained EPS were assayed in a
time lapsed experiment using confocal scanning laser microscopy (CSLM). In CSLM,
light emitted by the excited sample is screened through a pinhole, or detector aperture. As
a result, all out of focus light is excluded and only light coming from the focal plane is
48
detected. This method is especially valuable because it permits the reconstitution and
visualization of 3-dimentional fluorescence-emitting structures from a stack of 2dimentional planes. In both these studies, CSLM gave a view of spatial repartition of
bacteria and EPS throughout the depth of the biofilm structure, as well as a measure of
the kinetics in the biofilm formation process. Lastly, the constitutive Gfp approach has
also been used to quantify antibiotic susceptibility of the fungus Aureobasidium pullulans
biofilm to biocides (Sabev 2006). In this study, the loss of constitutive fluorescence
observed under the microscope and measured by spectrophotometry was indicative of
cell death.
Inducible Gfp Constructs
Experiments in the literature also include selectively producing Gfp by the use of
inducible promoters. With inducible Gfp constructs, the in situ visualization of
fluorescence becomes a dynamic process both spatially and temporally. This work was
accomplished by cloning selected inducible promoters upstream of promoterless gfp
genes (Valdivia 1996). These kinds of constructs have been modified to target the study
of N-acyl homoserine lactone (AHL) controlled gene expression. AHL is a broadly
conserved quorum sensing molecule involved in intra and cross-species bacterial
communication.
Steidle and coworkers (2001) used an AHL sensor plasmid model to monitor the
levels of quorum sensing molecules in the soil. This team also created a reporter
chromosomal cassette in AHL-negative Pseudomonas putida in order to detect AHL
production by indigenous soil bacteria. This study used CSLM to visualize fluorescent
49
bacteria associated with plant roots, illustrating the power of the technique to detect the
fluorescence emission of individual bacteria. Similarly, a study by Burmølle et al (2003)
used a modified strain of E. coli as a biosensor. The gfpmut3 gene was fused with the
Vibrio fischeri luxR-PluxI gene and inserted inside a high copy plasmid. The construct
turned the transformed bacteria into a powerful detector of AHLs in the soil, as detected
by flow cytometry. Another study used a similar AHL sensor to determine the nature of
the cross talk between P. aeruginosa and Burkholderia cepacia, two bacteria capable of
forming mixed biofilms inside the lungs of cystic fibrosis patients (Riedle 2001). Here
the researchers used CSLM to detect AHL-triggered Gfp production at the surface of
forming biofilms. They were also able to determine that, while both bacteria produced
quorum sensing molecules, only P. aeruginosa was capable of eliciting an AHL specific
cross-species response in the mixed-culture biofilm.
Unstable Gpf Constructs
Further modification of the technique became possible with the development of
unstable Gfp variants (Andersen 1998). Here the Gfp protein is synthesized with a tag
that is targeted by intracellular proteases. As a result, unstable Gfp produced by bacteria
is degraded within a few hours. This molecular tool is especially tailored to the use of
CSLM, as it permits a time-lapsed visualization of promoter activation throughout the
biofilm structure. For example, researchers have used inducible promoters coupled with
CSLM visualization to observe bacteria responding to AHL stimulation (Andersen 2000,
Hentzer 2002). Others have employed the technique to monitor biofilm response to
antibiotic challenge (Bagge 2004). Finally, as within P. putida biofilm studies,
50
researchers have looked at promoters that are associated with bacterial growth turning off
when the given layer of bacteria was buried under younger generations of microbes
(Sternberg 1999). In this latter study, using an unstable Gfp reporter construct, the
researchers were able to use CSLM to show the outermost biofilm layer expressing Gfp,
while production was absent in deeper strata. If stable Gfp had been used instead, it
would have been difficult to observe the stratification, for all bacteria inside the biofilm
would still be harboring the Gfp they had produced while actively dividing. This novel
approach to visualize bacterial metabolism was pushed a step further with the work of
Werner and coworkers (2004) when they designed a growth rate dependant unstable Gfp
reporter system for P. aeruginosa. Here, the scientists were able visualize a stratification
of metabolic rates in situ. The ensuing fluorescence intensity gradient was observed by
making frozen sections of biofilm by epifluorescence instead of using confocal
microscopy.
Rational for our approach
Relevance of the Study
Our in situ visualization study is unique in a number of ways. First, this is the first time
mycobacterial biofilms have been used for an in situ study using inducible Gfp
constructs. In this sense, we are studying a biofilm of a unique nature due to the cording
phenotype of Mycobacterium marinum. In contrast with previous work used for in situ
studies, M. marinum does not form the traditional compact and homogenous structure
where tight gaps between bacteria are filled with EPS. Biofilms formed by M. marinum
are, in comparison, widely spaced structures composed of lipid-rich, intertwined,
51
branching, and serpentine cords of bacteria growing in thickness and complexity with age
(figure IV.1).
Finally, the uniqueness of our study relies on the fact that we used the biofilm
inducible promoters we found by DFI to observe biofilms cross-sections by CSLM. As a
result, the promoters used in this study are not clear on/off switches as previously
described, but rather promoters expressing a gradient in activation varying when grown
under different conditions.
52
Figure IV.1. Confocal microscopy picture of Gfp-producing M. marinum biofilm after
14 days. Note the widely spaced structure composed of intertwined, branching, and
serpentine cords of bacteria (bar = 50µm).
53
Preliminary evidence
Preliminary data showed distinct differences in the distribution of fluorescence
between samples (Figure IV.2.). In Figure IV.2.A., Gfp production seemed spread out
evenly throughout the biofilm thickness, whereas in Figure IV.2.B., high levels of Gfp
production were very clearly located at, and adjacent to, the site of biofilm attachment to
the glass slide. This observation became especially relevant when sequencing of the
biofilm-induced promoter revealed its location directly upstream of a gene homologous
to one coding for a putative adhesin protein in Mycobacterium tuberculosis. The
expectation would be that a putative attachment protein would be produced closest to the
locus of bacterial adhesion to the glass surface, and if this is the case, we would see more
Gfp produced in that area.
Dealing with Quenching
One difficulty when using confocal imagery to quantify variations in fluorescence
within a sample is a phenomenon called quenching. Quenching occurs when emitted or
transmitted light is blocked and becomes fainter as it propagates through the sample. In
our case, even though the laser intensity was constant in all experiments, the beam of
light gradually faded away as it traveled into the biofilm. Accordingly, the intensity of
Gfp fluorescence emitted was naturally greater adjacent to the slide and
54
A
B
79µm
Figure IV.2. Confocal microscopy pictures of Gfp-producing M. marinum biofilms. (A)
Clone OB1-102 after 14 days (bar = 75µm), and (B) Clone OB1-101 after 14 days (bar =
50µm) Side views were visualized in pseudo-color, where white indicates the maximum
level of Gfp production. In clone OB1-102 biofilm, Gfp production seemed spread out
evenly throughout the biofilm thickness, whereas in clone OB1-101 biofilm, high levels
of Gfp production were very clearly located at, and adjacent to, the site of biofilm
attachment to the glass slide
55
gradually decreased as the excitation decreased. This phenomenon is well documented in
confocal microscopy and much research effort has been made to quantify it in a number
of different backgrounds (Méallet-Renault 2000, Bowman 2003). The study of a
biological system such as a biofilm however, is so variable by nature that it is impossible
to predict the incidence of quenching on the observed fluorescence. While researchers
typically avoid the difficulty of accounting for quenching by looking at frozen biofilm
cross-sections instead of using CSLM because the technique is prone to artifacts (Werner
2004), we developed instead a semi-quantitative methodology to compensate for the
natural intensity gradient in our samples due to quenching. Consequently, we were able
to obtain a repeatable method to quantify the loci of Gfp production within the biofilm
structure.
56
METHODS
Bacterial strains and growth conditions
All experiments were performed using Mycobacterium marinum 1218R (ATCC
927), a fish outbreak isolate. The negative control for quenching used for all experiments
was V27R (1218R containing only the insertless vector pFPV27). Strain V27R produces
a constant baseline amount of Gfp by a phenomenon of read-through. Clone G13R as
described by Barker et al. (Barker 1998, Barker 1999), is always highly fluorescent and
was used as a positive control in the in situ visualization assay. Stationary stock solutions
of V27R and G13R were obtained by incubating the strains 1:24 in 7H9/OADC/KAN for
10 days. For all assays, 1mL samples of stationary stock cultures were incubated in 24mL
of 7H9/OADC/KAN for 4d to reach mid logarithmic phase. To determine bacterial
numbers in the inocula, dilution plate counts (DPC) were performed using 7H10 solid
media (Difco) supplemented with OADC, 10g/mL cyclohexamide (Sigma-Aldrich, St.
Louis, MO), and 30g/mL KAN. M. marinum transformants OB1-101, OB1-102, and
OB2-113 were chosen for visualization from the list of biofilm induced clones in the
differential fluorescence induction assay. Stationary stock solutions of OB1-101, OB1102, and OB2-113 were obtained by incubating the strains 1:24 in 7H9/OADC/KAN for
10 days. For all assays, 1mL samples of stationary stock cultures were incubated in 24mL
of 7H9/OADC/KAN for 4d to reach mid logarithmic phase. To determine bacterial
numbers in the inocula, dilution plate counts (DPC) were performed using 7H10 solid
media (Difco) supplemented with OADC, 10g/mL cyclohexamide (Sigma-Aldrich,),
and 30g/mL KAN.
57
Biofilm growth for the in situ visualization
Biofilms for in situ visualization were grown in the Stovall Convertible Flow Cell
system (Stovall Life Science, Greensboro, NC). The flow cells are double-sided with
glass cover slips to allow for non-invasive visualization by CLSM of the forming biofilm.
Logarithmic phase bacterial inocula of selected clones were sonicated using a VibraCell
sonicator with cup horn attachment (Model VCX 500, Sonics, Newtown, CT) in 5s on/ 5s
off cycles and passed 5 times through a tuberculin syringe. The number of bacteria in the
inocula was determined by DPC. A volume of 200µL from each sample was inoculated
into the Flow Cells via a self sealing injection port. Flow Cells were connected via an
Ismatec peristaltic pump to a continuous flow of sterile 7H9/KAN media at a rate of
100µL/min. Single and 4-channel bubble traps were used (Biosurface Technologies Corp.
Bozeman, MT) to prevent air bubbles from circulating inside the Flow Cells. The
apparatus was incubated upside-down for 14 days at 32oC. Cells were inverted before
visualization so as to visualize the side of the apparatus facing upwards during
incubation. Using this method, we were able to visualize actively attaching biofilms
structures rather than bacteria sedimenting on the bottom of the flow cell.
In situ visualization by confocal microscopy
Confocal visualization was performed with a Nikon TE2000 Inverted Microscope
(Nikon Instruments Inc., Melville, NY) using a 60X Oil, Nikon objective. Samples were
excited using a Spectra-Physics Laser (Mountain View, CA) model 163-C, emitting at
488nm. Visualization was achieved using a Nikon D-Eclipse C1 camera. The firmware
EZ-C1 Software was used to analyze the biofilm formation pictures taken at days 7, 10
58
and 14. Neutral density filters 8X and 4X were used to reduce the laser light intensity and
thus minimize bleaching of the samples. Confocal stacks were recorded at 19.92µs pixel
dwell using the smallest pinhole (30µm). Sections of 211.7µm by 211.7µm were taken at
2µm depth intervals starting at the site of biofilm attachment to the glass cover slip.
During visualization, the laser intensity was not changed, only the sensitivity (or gain) of
the photomultiplier diodes (PMT) was changed. The PMTs were adjusted each time so
that the sample would reveal a full range of fluorescence intensities. We then used the
“Volume Render” function of the analysis software to obtain a 3-D render of the total
stack seen both from the top (0o tumble), and from the side (90o tumble). Side views
where visualized in pseudo-color, where a gradient going from white to pink and red,
through yellow, green and dark blue represented the range in green fluorescence
intensities. Figure IV.2 (page 8) shows two representative pictures using “Volume
render” seen from top and side.
In situ visualization data analysis
A semi-quantitative analysis was performed on each stack to capture the
differences in intensity gradients observed visually between samples. Fifty points were
chosen at random on a field where bacteria were present. The software measured the
average fluorescence intensity at each point through each section of the stack, along the
z-axis. Data was recorded on a Microsoft Excel spreadsheet. Since points were picked at
random on the field, some did not match areas where bacteria were present. By noting the
fluorescence value at those points, we were able to determine background values where
fluorescence did not corresponded to that of the actual biofilm. These values were
59
considered background noise and were not used for average fluorescence intensity
calculations. Only fluorescence values above background were used for analysis.
Fluorescence data points were averaged at each biofilm section throughout the stack. In
order to normalize the average fluorescence values between stacks, each section’s
average fluorescence was expressed as a percentage of the maximum fluorescence of the
given stack. Normalized fluorescence averages were then graphed as a function of
biofilm depth. A best fit curve was determined for each graph between the first point
above 90 percent fluorescence and the first point bellow 20 percent fluorescence. In some
cases, when fluorescence values jumped back over 20 percent for more than two data
points, the next point bellow 20 percent was included in the best fit curve calculation.
The slope of the best fit curve gave us an indication of the decrease in fluorescence as a
function of distance from the glass slide.
60
RESULTS
Time-Course Pictures
We were able to reproducibly grow biofilms in the Convertible Flow Cell system
over a period of 14 days. Figure IV.3 shows reconstituted 3-dimentional pictures taken at
day 7, 10 and 14 of biofilm formed by clones OB2-113 viewed both from top and from
the side. We observe a gradual increase in biofilm mass over time consistent with normal
biofilm development. We also observe a similarity of appearance at given time points in
biofilms formed by the different clones (data not shown).
Semi-Quantitative Analysis
For the semi-quantitative analysis, we picked 50 points randomly over the field on
areas where bacteria were present. Figure IV.4 shows a representative picture of the
random way in which points were chosen. The software calculated the fluorescence
values at each point, through each plane in the stack. We discarded background
fluorescence values and averaged all points above background in the plane. In order to
obtain a more representative comparison of fluorescence loci between clones, we chose
to express each fluorescence value as a percentage of the maximum fluorescence in the
given stack. Figure IV.5 shows graphs of the percent maximum fluorescence versus zsection depth for a representative time course. On each graph, the best fit curve is
representative of the decrease in fluorescence values over the biofilm thickness. The
slopes of the best fit curves are also represented. The slopes give an indication of the
average variation in fluorescence throughout the thickness of the biofilm for each clone at
61
A
B
C
Figure IV.3. Confocal microscopy pictures of Gfp-producing M. marinum biofilms (bar
= 50µm), at day 7 (A), day 10 (B) and day 14 (C). The reconstituted 3-D images seen
from the top and from the side (in pseudo-color) of a representative time course show a
gradual increase of biofilm mass and thickness over time.
Figure IV.4. Confocal microscopy horizontal
section of Gfp-producing M. marinum biofilms in
pseudo-color (bar = 50µm). Representative picture
of the 50 points randomly chosen over the field on
areas where bacteria were present.
62
A
Percent of Max.
Fluorescence
OB-C101 - Day 7
120.00
100.00
80.00
60.00
40.00
20.00
0.00
0.00
y = -13.457x + 173.96
R2 = 0.9305
5.00
10.00
15.00
20.00
C
25.00
30.00
35.00
40.00
Micron
B
Percent of Max.
Fluorescence
OB-C101 - Day 14
Percent of Max.
Fluorescence
OB-C101 - Day 10
120
100
80
60
40
20
0
y = -3.2757x + 113.52
R2 = 0.904
0
10
20
30
40
120.00
100.00
80.00
60.00
40.00
20.00
0.00
0.00
y = -2.6384x + 129.78
R2 = 0.9414
20.00
40.00
60.00
80.00
100.00
Micron
50
60
Micron
Figure IV.5. Representative confocal microscopy pictures of Gfp-producing M. marinum
biofilms (bar = 50µm). (A) Clone OB-C101 after 7 days, (B) Clone OB-C101 after 10
days, and (C) Clone OB-C101 after 14 days. Side views where visualized in pseudocolor. Each corresponding graph depicts the percent of maximum fluorescence per depth
in micron. The best fit curves are representative of the decrease in fluorescence values
over the biofilm thickness. The equations for the best fit curves are also represented. The
slopes (the negative coefficient in front of the variable x) give an indication of the
average variation in fluorescence throughout the thickness of the biofilm at each time
point.
63
Day
V27R
OB1-101
OB2-113
OB1-102
G13R
7
-1.5 (0.5)
-15.1 (3.6)
-7.4 (0.4)
-6.8 (0.4)
-2.4 (0.1)
10
-1.5 (0.2)
-2.8 (0.5)
-2.4 (0.5)
-2.5 (0.7)
-3.5 (1.5)
14
-1.7 (0.1)
-3.6 (1.4)
-3.6 (2.0)
-0.7 (0.3)
-1.6 (0.3)
Table IV.1. Average slopes for all clone tested at days 7,
10 and 14. The standard deviations are represented in
parenthesis. Values were calculated from two different
fields of view over two separate experiments.
64
days 7, 10 and 14. They are recorded on Table IV.1. It shows constancy in V27R’s slope
between time points averaging -1.6 (Standard deviation (SD) of 0.4). Analysis results are
compiled in Figure IV.6. All clones are compared to the V27R control clone’s average
over the entire time course (Figure IV.6.A). The grey area on the graph represents the
values within one standard deviation of the V27R average. For each other clone, the
average comprised between one standard deviation above and below is compared to the
V27R grey area. This allows us to visualize average slope values significantly above of
below that of the control.
Clone OB1-101 possessed the steepest slope and hence the most pronounced
fluorescence gradient of all clones assayed (Figure IV.6.B). At day 7, the slope within the
biofilm reached -13.8 (SD=0.4). This indicated a fluorescence intensity gradient much
greater than that accountable for quenching alone. In fact, the steep slope indicated that
Gfp was produced principally adjacent to the glass slide where the bacteria adhered. As
bacteria grow further and further from the glass slide, the amount of Gfp produced
sharply decreased. By days 10 and 14, we observed a sharp decrease in the values of the
fluorescent gradient through the biofilm, yet still significantly above that of the V27R
control.
The slope values of clone OB2-113, promoter corresponding to the hypothetical
membrane protein, were similar to those of clone OB1-101 (Figure IV.6. C), except that
the slope at day seven is not as steep (-7.2, SD = 0.2). Slopes of clones OB1-101 and
OB2-113 at days 14 are virtually identical (-2.6 (0.1) and -2.4 (0.1)).
The promoter corresponding to our transcriptional attenuator, clone OB1-102 exhibited a
much different trend over the time course (Figure IV.6. D). The slope is steep at day
65
seven (-7.4, SD = 0.4), then decreases to -2.5 (1.2) at day 10. The difference come in that
at day 14, the slope reaches a value of -0.6 (0.3), which is significantly less steep than the
slope of clone V27R.
The graph of clone G13R shows the most variance over time (Figure IV.6. E).
G13R is the clone containing a strong promoter. At days 7 and 14 the graph has a
biphasic characteristic (figure IV.7.). Over the thickness of the biofilm, average percent
fluorescence seems to decrease only to peak back up deeper into the biofilm mass. Slopes
at days 7 and 14 are similar to that of V27R, respectively, -2.4 (0.1) and -1.4 (0.3). The
slope at day 10 is the steepest and most regular over the clone’s time course with a value
of -4.7 (1.2).
66
Individual Slopes Compared to V27R Control
20
18
Slope Values (-)
16
14
12
10
8
6
4
2
0
d7
d10
V27R
d14
d7
d10
d14
OB-C101
d7
d10
d14
OB-C113
d7
d10
d14
OB-C102
d7
d10
d14
G13R
Figure IV.6. Individual slopes compared to the V27R control. This graph provides a
visual comparison between the average slope value for the V27R control and the other
clones tested. An average slope value is given for each clone at day 7, 10 and 14. The
grey area on the graph represents the slope values comprised within one standard
deviation of the V27R averages. Values were calculated from two different fields of view
over two separate experiments.
67
A
G13 - Day 7
Percent of Max.
Fluorescence
120
100
80
y = -2.5196x + 97.34
R2 = 0.5066
60
40
20
0
0
10
20
30
40
Micron
G13 - Day 10
B
Percent of Max.
Fluorescence
120
100
y = -3.7573x + 127.3
80
2
R = 0.9488
60
40
20
0
0
10
20
30
40
50
60
50
60
Micron
C
G13 - Day 14
Percent of Max.
Fluorescence
120
100
y = -1.5898x + 74.478
R2 = 0.5263
80
60
40
20
0
0
10
20
30
40
Micron
Figure IV.7. Semi-quantitative analysis graphs of
the G13R construct time course at day 7 (A), day
10 (B), and day 14 (C). The percent of maximum
fluorescence is expressed as a function of microns
through the biofilm thickness.
68
DISCUSSION
With the in situ visualization technique, we were able to quantify the behavior of
selected Mycobacterium marinum promoters over time when the bacteria are grown in a
biofilm. We have observed no significant differences between the appearance of biofilms
formed by the different M. marinum transformants at any time point. This observation is
especially important because we based this work on the assumption that biofilms would
be structurally identical between M. marinum 1218R transformants regardless of the
insert they contained. Looking at the reconstituted 3-dimentional biofilm pictures gives
us a visual confirmation of this assumption. The progressive colonization of the
Convertible Flow Cell surface by M. marinum over the 14 days time course also clearly
shows that, even though glass is not the most suitable surface onto which the
hydrophobic M. marinum preferably adheres (Hall-Stoodley 2006), a steadily growing
biofilm can be cultured on glass under continuous flow.
The V27R Construct
Clone V27R time course analysis was a pivotal control for the in situ
quantification of variances in fluorescence within the biofilm structures, because it
indicated baseline values for quenching. Fluorescence in V27R arises by a phenomenon
of ribosomal readthrough into the promoterless gfp gene. There is no evidence showing
that the levels of Gfp production by V27R vary. We hence predicted that fluorescence
levels would be constant throughout the biofilm, and that any gradient in fluorescence
level observed within V27R biofilm would be due principally to quenching. Semiquantitative analysis revealed that between days 7, 10 and 14, fluorescence gradient
69
within V27R biofilm stayed constant (-1.6 average slope; SD = 0.4). The fact that the Gfp
gradient’s slope stayed constant over time allowed us to use V27R as an indicator for
quenching.
Our two assumptions were first that Gfp production would be even throughout the
V27R control biofilm and second that the nature of the biofilms produced by all clones
was independent to their plasmid transformant. These assumptions would allow us to
distinguish between the three possible scenarios when comparing V27R to all other
clones tested. First, if fluorescence was evenly distributed through the biofilm thickness
on a given sample then we would observe a slope value roughly equivalent to that
determined for clone V27R. Second, if fluorescence was more pronounced closest to the
site of attachment to the glass slide, then we would observe a more pronounced gradient
in the decrease in fluorescence. As a result, the absolute slope value calculated would be
significantly greater (i.e. more negative) than that observed for V27R. Third and lastly, if
the bacteria produced a greater amount of fluorescence at the biofilm periphery, or later
in biofilm development, then we would observe a less pronounced gradient in the
decrease in fluorescence. In other words, if the production of Gfp at the periphery was
markedly greater, we would observe an absence or a reversal in the fluorescence gradient.
In both cases, this would result in the absolute slope value calculated to be less than that
observed for V27R.
70
Clone OB1-101
In clone OB1-101, the promoter is associated with a gene homologous to the
Mycobacterium tuberculosis gene, Rv3716c coding for a 133 amino acid protein
(E=4x10-77). Information collected by the TB Structural Genomic Consortium (TBSGC
2007) reveals that Rv3716c is a highly conserved hypothetical protein, not only among
mycobacteria of M. marinum’s clade, but also in a number of unrelated bacterial species.
Furthermore, the ProKnow prediction method, capable of inferring protein function by
analyzing their structure (Pal 2005), shows by the frequency of ontologies from threedimensional folds that Rv3716c possesses characteristics of cell adhesion motifs.
Accordingly, our ISV findings reflected Rv3716c’s predicted function. First, because
they suggested that induction of the promoter for Rv3267 was localized at the site of cell
attachment to the surface and, secondly, because past an initial attachment period, the
activity of the promoter is greatly decreased. This suggests that the M. marinum protein
may be synthesized early to permit anchoring of the biofilm community to surfaces
instead of evenly throughout biofilm development. These findings indicate that the M.
marinum homolog of Rv3716c could be responsible for Mycobacterium marinum
attachment to surfaces. Furthermore, the ubiquity of homology to Rv3716c throughout
the bacterial species is an indication that perhaps we have found a global bacterial
adhesin protein.
71
Clone OB2-113
Clone OB2-113 contains a M. marinum promoter upstream of a gene coding for
the 143 amino acid protein MUL_2097 in Mycobacterium ulcerans strain Agy99. This is
especially interesting because MUL_2090 is unique to M. ulcerans and M. marinum
(Stinear 2007). This was confirmed by performing a BLAST using NCBI. Furthermore,
MUL_2097, like Rv3716c, is also a hypothetical protein likely associated with the
bacterial membrane. These findings led us to hypothesize that the promoter in the clone
OB2-113 would behave like one associated with a protein responsible for biofilm
attachment. By ISV, we showed that OB2-113 reveals similar patterns as that of OB1101. Much like clone OB1-101, the intensity gradient in OB2-113 was most highly
pronounced at day 7 then sharply decreased by day 10 and was unchanged at day 14.
Gradient values between the two clones were virtually identical at days 10 and 14. The
intensity gradient in OB2-113, however, was not as pronounced at day 7 as in OB1-101
(7.2; SD = 0.2 in OB2-113 compared to 13.8; SD = 0.4 in OB1-101). These new data
suggest that we have identified M. marinum and M. ulcerans specific proteins playing a
putative role in biofilm attachment. In light of these findings, it is possible that there may
be a relationship between expression of this protein and the ability to form biofilm in an
aquatic environment. This is consistent with the fact that M. marinum and M. ulcerans
are two environmental isolates whereas other mycobacteria, for which sequences are
available, such as M. tuberculosis and M. leprae, are solely associated with human
pathogenicity. In other words, the expression of the putative adhesin MUL_2097 in M.
marinum and in M. ulcerans may give these bacteria the distinctive ability to attach and
persist outside of a host in their most likely aquatic niche, the biofilm.
72
Clone OB1-102
In clone OB1-102, the promoter activates a gene highly conserved among
mycobacteria species. The M. marinum nucleotide sequence is a perfect match with the
hypothetical Mycobacterium ulcerans protein, MUL_2613 (P=0.0). Furthermore, this
protein shows similar folding characteristics to the transcriptional attenuator MAV_4232
in Mycobacterium avium (P=10-160). In Mycobacterium tuberculosis, it matches Rv3267,
a 498 amino acid long conserved hypothetical protein with tyrosine-phosphatase
properties (CpsA-related protein, P=0.0). Unlike any other clone visualized, the slope of
the intensity gradient indicates that while some Gfp is produced close to the site of
attachment early in biofilm development, Gfp production gradually shifts to the biofilm
periphery by day 14. Those findings show that by semi-quantitative analysis it is possible
to measure a fluorescence intensity gradient opposed to that that due to quenching. While
we measured gradients that were more pronounced than the control with clones OB1-101
and OB2-113, clone OB1-102 on the other hand, shows an gradient opposite to these
previous. This is a confirmation that the quantitative method we developed gives us
indeed the numerical tools to confirm our initial visual observations.
The G13R Construct
Variability was the greatest in the clone possessing the G13 insert. Pervious work
done by Barker and coworkers (1998, 1999) has well characterized this strong sigma 70like putative promoter conserved in E. coli. As shown by means of promoter trap assay, it
is induced during trafficking inside macrophages, and otherwise highly expressed
constitutively. We expected that the G13R clone would show varying levels of activation
73
in situ, in response to stress as in grew within a biofilm. The variations in activation we
observed were both spatially and temporally (figure IV.7.). Unlike other clones tested,
G13R shows at days 7 and 14 (Figure IV.7 A and C respectively), a bimodal nature. This
reflects promoter activation with strong foci both towards the site of attachment to the
glass slide and also towards the periphery of the biofilm exposed to media flow. At day
10 however (figure IV.7. B), the graph shows a typical intensity gradient consistent with
promoter activation adjacent to the glass slide. This series of results lead us to believe
that the G13 promoter activation is not likely correlated to localization within the biofilm.
Regardless, clone G13R showed patterns very different to all other clones tested.
Conclusions
We used in situ visualization and CSLM to localize promoter activation within
Mycobacterium marinum biofilm, both spatially and over time. The semi-quantitative
assay described has also proven to be a useful method in implying or adding to protein
predicted function in association with other molecular techniques of prediction.
Further ameliorations would involve the use of unstable Gfp protein coupled with
the promoters found by DFI. Blokpole et al. (2003) showed that the construct that can be
used with both slow and fast growing mycobacteria. With such a construct, we could
further reduce artifacts in the intensity gradient caused by quenching by eliminating Gfp
accumulation within cells over the length of the time course. It is also possible to add
another level of analysis, by coupling CSLM with NMR (Majors 2005). While such
previous work coupling metabonomics with CSLM has been previously carried out, it has
only focused on monitoring biofilm growth. By combining promoter in situ visualization
74
and metabonomics measurements we would get a better understanding of the molecules
secreted in association with promoter induction by the bacteria living among the
community. In this fashion, we could assay for the production of auto-inducers or other
quorum sensing molecules coupled in real-time with the activation of specific promoters.
These novel techniques open new horizons towards the elucidation of the complex
behavior of bacteria within biofilms.
75
REFERENCES
Alloway, J.A., S. M. Evangelisty, and J. S. Sartin. 1995. Mycobacterium marinum
arthritis. Semin. Arthritis Rheum. 24:382-390.
Andersen, J. B., A. Heydorn, M. Hentzer, L. Eberl, O. Geisenberger, B. B. Christensen,
S. Molin, M. Givskov. 2000. gfp-Based N-Acyl Homoserine-Lactone Sensor
System for Detection of Bacterial Communication. Appl. Environ. Microbiol.
67(2):575-585.
Andersen, J. B., C. Sternberg, L. K. Poulsen, S. P. Bjørn, M. Givskov, S. Molin. 1998.
New Unstable Variants of Green Fluorescent Protein for Transient Gene
Expression in Bacteria. Appl. Environ. Mocrobiol. 64(6):2240-2246.
Aronson, J.D. 1926. Spontaneous tuberculosis in saltwater fish. J. Infect. Dis. 39:315320.
Ashford, R.W. 2003. Invasive Mycobacterium marinum infections. Emerging Infect. Dis.
9:1496-1498
Bagge, N., M. Hentzer, J. B. Andersen, O. Ciofu, M. Givskov, N. Høiby. 2004. Dynamic
and Spatial Distribution of β-Lactamase Expression in Pseudomonas aeruginosa
Biofilm. Antimicrob. Agents Chemoter. 48(4):1168-1174.
Barbeau J, C . Gauthier, P. Payment. 1998. Biofilms, infectious agents, and dental unit
waterlines: a review. Can. J. Microbiol. 44:1019-1028.
Bardouniotis, E., H. Ceri, and M.E. Olson. 2003. Biofilm formation and biocide
susceptibility testing of Mycobacterium fortuitum and Mycobacterium marinum.
Curr. Microbiol. 46: 28-32.
76
Bardouniotis, E., W. Huddleston, H. Ceri, and M. E. Olson. 2001. Characterization of
biofilm growth and biocide susceptibility testing of Mycobacterium phlei using
the MBEC™ assay system. FEMS Microbiol. Lett. 203:263-267.
Barker L.P., D.M. Brooks, P.L.C. Small. 1998. The identification of Mycobacterium
marinum genes differentially expressed in macrophage phagosomes using
promoter fusions to green fluorescent protein. Mol. Microbiol. 29:1167-1177.
Barker, L. P., K. M. George, S. Falkow, and P. L. C. Small. 1997. Differential trafficking
of live and dead Mycobacterium marinum organisms in macrophages. Infect.
Immun. 65(4):1497-1504.
Barker L.P., B.A. Lien, O.S.Brun, D.D. Schaak, K.A. McDonough, L C. Chang. 2007. A
Mycobacterium marinum Zone of Inhibition Assay as a Method for Screening
Potential Antimycobacterial Compounds from Marine Extracts. Planta Medica.
73(6):559-63.
Bercovier, H., and V. Vincent. 2001. Mycobacterial infections in domestic and wild
animals due to Mycobacterium marinum, M. fortuitum, M. chelonae, M.
porcinum, M. farcinogenes, M. smegmatis, M. scrofulaceum, M. xenopi, M.
kansasii, M. simiae, and M. genavense. Rev. Sci. Tech. Off. Int. Epiz. 20(1):265290.
Blokpoel, M. C. J., R. O’Toole, M. J. Smeulders, H. D. Williams. 2003. Development
and Application of Unstable Gfp Variants to Kinetic Studies of Mycobacterial
Gene Expression. J. of Microbiol. Methods. 54:203-211.
Bouley, D. M., N. Ghori, K. L. Mercer, S. Falkow, and L. Ramakrishnan. 2001. Dynamic
nature of host-pathogen interactions in Mycobacterium marinum granulomas.
Infect. Immun. 69(12):7820-7831.
77
Bowman, R. D., K. A. Kneas, J. N. Demas, A, Periasamy. 2003. Conventional, Confocal
and Two-Photon Fluorescence Microscopy Investigations of Polymer-Supported
Oxygen Sensors. Journal of Microscopy. 211:112-120.
Burmølle, M., L. H. Hansen, G. Oregaard, S. J. Sørensen. 2003. Presence of N-Acyl
Homocerine Lactones in Soil Detected bya Whole-Cell Biosensor and Flow
Cytometry. 45:226-236.
Camacho L.R., D. Ensergueix, E. Perez, B. Gicquel, C. Guilhot. 1999. Identification of a
virulence gene cluster of Mycobacterium tuberculosis by signature-tagged
transposon mutagenesis. Mol. Microbiol. 34:257-267.
Carter, G., M. Wu, D. C. Drummond, and L. Bermudez. 2003. Characterization of
biofilm formation by clinical isolates of Mycobacterium avium. J. Med.
Microbiol. 52:747-752.
CDC (Center for Disease Control and Prevention). 2007. Trends in Tuberculosis
Incidence – United States. MMWR. 56(11)245-50.
Ceri, H., M. E. Olson, C. Stremick, R. R. Read, D. Morck, and A. Buret. 1999. The
Calgary biofilm device: New technology for rapid determination of antibiotic
susceptibilities of bacterial biofilms. J. Clin. Microbiol. 37:1771-1776.
Chambless, J. D., S. M. Hunt, P. S. Stewart. 2005. A Three-Dimensional Computer
Model of Four Hypothetical Mechanisms Protecting Biofilms from
Antimicrobials. Appl. Environ. Microbiol. 72(3):2005-2013.
Clark, H.F., and C. C. Shepard. 1963. Effect of environmental temperatures on infection
with Mycobacterium marinum (balnei) of mice and a number of poikilothermic
species. J. Bacteriol. 86:1057-1069.
78
Cormack BP, Valdivia RH, Falkow S. 1996. FACS-optimized mutants of the green
fluorescent protein. Gene. 173:33-38.
Costerton J.W. 1999. Introduction to biofilms. Int. J. Antimicrob. Agents. 11:217-221.
Costerton J.W., Cheng KJ, Geesey GG, Ladd TI, Nickel, JC, Dasgupta M, Marrie TJ.
1987. Bacterial biofilms in nature and disease. Annu. Rev Microbiol. 41:435-464.
Costerton J.W. et al. 1999. Bacterial biofilms: a common cause of persistent infections.
Science. 284:1314-1318.
Cox, J.S., B. Chen, M McNeil, W. R .Jacobs Jr. 1999. Complex lipid determines tissuespecific replication of Mycobacterium tuberculosis in mice. Nature 402:79-83.
Dailloux, M., M. Albert, C. Laurain, S. Andolfatto, A. Lozniewski, P. Hartemann, and L.
Mathieu. 2003. Mycobacterium xenopi and drinking water biofilms. Appl.
Environ. Microbiol. 69(11):6946-6948.
Davis, J.M., H. Clay, J. L. Lewis, N. Ghori, P. Herbomel, and L. Ramakrishnan. 2002.
Real-time visualization of mycobacterium-macrophage interactions leading to
initiation of granuloma formation in zebrafish embryos. Immunity. 17(6):693702.
Darzins, E and Fahr, G. 1956. Cord-forming property, lethality and pathogenicity of
mycobacteria. Dis. Chest. 30:642-648.
De Beer, D., P. Stoodley, F. Roe, and Z. Lewandowski. 1994. Effect of Biofilm Structure
on Oxygen Distribution and Mass Transport. Biotechnol. Bioeng. 43:1131-1138.
Dobos, M.K. et al. 1999. Emergence of a unique group of necrotizing mycobacterial
diseases. Emerg. Infect. Dis. 5: 367-378.
79
Donlan, R. M., 2002. Biofilms: microbial life on surfaces. Emerg. Infect. Dis. 8(9):881890.
Edelstein, H. 1994. Mycobacterium marinum skin infections. Arch. Intern. Med.
154:1359-1364.
El-Etr, S. H., S. Subbian, S. L. G. Cirillo, and J. D. Cirillo. 2004. Identification of two
Mycobacterium marinum loci that affect interactions with macrophages. Infect.
Immun. 72(12):6902-6913.
El-Etr, S. H., L. Yang, and J. D. Cirillo. 2001. Fish monocytes as a model for
mycobacterial host-pathogen interactions. Infect. Immun. 69(12):7310-7317.
Epinosa-Urgel, Manuel. 2003. Resident Parking Only: Rhamnolipids Maintain Fluid
Channels in Biofilms. J. of Bacteriol. 185(3): 699-700.
Falkinham III, J.O., 1996. Epidemiology of infection by non-tuberculosis mycobacteria.
Clin. Microbiol. Rev. 9:177-215.
Falkinham III, JO et al. 2001. Factors influencing numbers of Mycobacterium avium,
Mycobacterium intracellulare, and other mycobacteria in drinking water
distribution systems. Appl. Env. Microbiol. 67:1225-1231.
Fuqua WC, Winans SC, Greenberg EP. 1994. Quorum sensing in bacteria: the luxR-LuxI
family of cell density-responsive transcriptional regulators. J. Bacteriol. 176:269275.
Glickman M. S., J. S. Cox, W. R. Jacobs Jr. 2000. A novel mycolic acid cyclopropane
synthetase is required for cording, persistence and virulence of Mycobacterium
tuberculosis. Mol. Cell 5: 717-727.
80
Gluckman, S. J. 1995. Mycobacterium marinum. Clin. Dermatol. 13:273-276.
Goeres, D. M., L. R. Loetterle, M. A. Hamilton, R. Murga, D. W. Kirby, and R. M.
Donlan. 2005. Statistical assessment of a laboratory method for growing biofilms.
Microbiology 151(3):757-762.
Gómez, S., A. Bernabé, M.A. Gómez, J.A. Navaro, and J. Sánchez. 1993. Fish
mycobacteriosis : morphopathological and immunocytochemical aspects. J. Fish
Dis. 16:137-141.
Hall-Stoodley, L., O. S. Brun, G. Polshyna, L. P. Barker. 2006. Mycobacterium marinum
Biofilm Formation Reveals Cording Morphology. FEMS Microbiol. Lett. 257:4349
Hall-Stoodley, L., J. W. Costerton, and P. Stoodley. 2004. Bacterial biofilms: from the
natural environment to infectious diseases. Nat. Rev. Microbiol. 2:95-108.
Hall-Stoodley L., C.W. Keevil, H.M. Lappin-Scott. 1999. Mycobacterium fortuitum and
Mycobacterium chelonae biofilm formation under high and low nutrient
conditions. J. Appl. Microbiol. Symp. Suppl. 85:60S-69S.
Hall-Stoodley L and H. Lappin-Scott. 1998. Biofilm formation by the rapidly growing
mycobacterial species Mycobacterium fortuitum. FEMS Microbiol. Lett. 168: 7784.
Hall-Stoodley, L., and P. Stoodley. 2002. Developmental regulation of microbial
biofilms. Curr. Opin. Microbiol. 13:228-233.
81
Hanau, L. H., A. Leaf, R. Soeiro, L. M. Weiss, and S. S. Pollack. 1994. Mycobacterium
marinum infection in a patient with the acquired immunodeficiency syndrome.
Cutis 54:103-105.
Hentzer, M., K. Riedel, T. B. Rasmussen, A. Heydorn, J. B. Andersen, M. R. Parsek, S.
A. Rice. L. Eberl, S. Molin, N. Høiby, S. Kjelleberg, M. Kivskov. 2002. Inibition
of Quorum Sensing in Pseudomonas aeruginosa Biofilm Bacteria by a
Halogeneted Furanon Compound. Microbiology. 148:87-102
Herdman, A. V., and J. C. Steele, Jr. 2004. The new mycobacterial species – emerging or
newly distinguished pathogens. Clin. Lab. Med. 24(3):651-690.
Huminer, D., S.D. Pitlik, C. Block, L. Kaufman, S. Amit, and J. B. Rosenfeld. 1986.
Aquarium-borne Mycobacterium marinum skin infection. Dermatology 122:698703.
Jensen, Peter Ø. et al. 2007. Rapid Necrotic Killing of Polymorphonuclear Leukocytes is
Caused by Quorum-Sensing-Controlled Production of Rhamnolipids by
Pseudomonas aeruginosa. Microbiology. 153:1329-1338.
Joe, L., and E. Hall. 1995. Mycobacterium marinum disease in Ann Arundel County:
1995 update. Md. Med. J. 44:1043-1046.
Kaattari, I.M., M.W. Rhodes, S.L. Kaattari, and E.B. Shotts. 2006. The Evolving Story of
Mycobacterium tuberculosis Clade Memebers Detected in Fish. J. of Fish Dis.
29:509-520.
Karakousis P.C., W.R. Bishai, S.E. Dorman. 2004. Mycobacterium tuberculosis cell
envelope lipids and the host immune response. Cell. Microbiol. 6:105-116.
KEGG: Kyoto Encyclopedia of Genes and Genomes. 2007. BRITE Database of
82
functional hierarchies and binary relationships of biological entities.
http://www.genome.jp/kegg/brite.html
Kressel, A. B., and F. Kidd. 2001. Pseudo-outbreak of Mycobacterium chelonae and
Methylobacterium mesophilicum caused by contamination of an automated
endoscopy washer. Infect. Control Hosp. Epidemiol. 22:414-418.
Kuehn, M., M. Mehl, M. Hausner, H. –J. Bungartz, S. Wuertz. 2001. Time-Resolved
Study of Biofilm Architecture and and Transport Processes Using Experimental
and Simulation Techniques: The Role of EPS. Water Sci. Technol. 43(8):143-151.
Maeyama, R., Y. Mizunoe, J. M. Anderson, M. Tanaka, T. Matsuda. 2004. Confocal
Imaging of Biofilm Formation Process Using Fluoropobed Escherichia coli and
Fluoro-Stained Exopolysaccharide. Journal of Biomedical Materials Research.
Part A. 70(2):274-82.
Majors, P. P., J. S. McLean, G. E. Pinchuck, J. K. Frederickson, Y. A. Gorby, K. R.
Minard, R. A. Wind. 2005. NMR Methods for in situ Biofilm Metabolism
Studies. J. Microbiol. Methods. 62:337-344.
Marsollier, L., T. Stinear, J. Aubry, J. P. Saint André, R. Robert, P. Legras, A-L.
Manceau, C. Audrain, S. Bourbon, H. Kouakou, and B. Carbonnelle. 2004.
Aquatic plans stimulate the growth of and biofilm formation by Mycobacterium
ulcerans in axenic culture and harbor these bacteria in the environment. Appl.
Environ. Microbiol. 70:1097-1103.
Méallet-Renault, R., H. Yoshikawa, Y. Tamaki, T. Asahi, R. B. Pansu, H. Masuhara.
2000. Confocal Microscopic Study on Fluorescence Quenching Dynamics of
Single Latex Beads in Poly(vinyl alcohol) Film. Polym. Adv. Technol. 11:772777.
83
Middlebrook G, Dubos R, Pierce C. 1947. Virulence and morphological characteristics of
mammalian tubercle bacilli. J. Exp. Med. 86:175-184.
Mims, C.A., J.H.L. Playfair, I.M. Roitt, D. Wakelin, R. Williams, R.M Anderson. 1993.
Medical Microbiology. Mosby Europe Limited. London. England.
Mollohan, C. S., and M. S. Romer. 1961. Public health significance of swimming pool
granuloma. Am. J. Pub. Health. 51:883-891.
Ojha A., M. Anand, A. Bhatt, L. Kremer, W.R. Jacobs Jr, G.F. Hatfull. 2005. GroEL1: a
dedicated chaperone involved in mycolic acid biosynthesis during biofilm
formation in mycobacteria. Cell. 123:861-873.
Pagan-Ramos E., J. Song, M. McFalone, M.H. Mudd, V. Deretic1998. Oxidative stress
response and characterization of the oxyR-ahpC and furA-katG loci in
Mycobacterium marinum. J. Bacteriol. 180:4856-4864.
Pal, D., D. Eisenberg. 2005. Inference of Protein Function from Protein Structure.
Structure. 13:121-130.
Pankhurst, C. L. 2003. Risk assessment of dental unit waterline contamination. Prim.
Dent. Care 10(1):5-10.
Parent, L.J., M.M Salan, P.C. Applebaum, and J.H. Dosset. 1995. Diseminated
Mycobacterium marinum infection and bacteremia in a child with severe
combined immunodeficiency. Clin. Infect. Dis. 21:1325-1327.
Pierce C, R. J. Dubos, W. B. Schaefer. 1956. Differential characteristics in vitro and in
vivo of several substrains of BCG: multiplication and survival in vivo. Am. Rev.
Tub. Pulmon. Dis. 74:683-698.
84
Prince, A. 1992. Adhesins and Receptors of Pseudomonas aeruginosa Associated with
Infection of the Respiratory Tract. Microb. Pathog. 13(4):251-260.
Prouty, M. G., N. E. Correa, L. P. Barker, P. Jagadeeswaran, and K. E. Klose. 2003.
Zebrafish-Mycobacterium marinum model for mycobacterial pathogenesis. FEMS
Microbiol. Lett. 11066:1-6.
Ramakrishnan L. Images in clinical medicine. 1997. Mycobacterium marinum infection
of the hand. N. Engl. J. Med. 337:612.
Ramakrishnan L, Federspiel N, Falkow S. 2000. Granuloma-specific expression of
Mycobacterium virulence proteins form the glycine-rich PE-PGRS family.
Science. 288:1436-1439.
Ramakrishnan, L. and, S. Falkow. 1994. Mycobacterium marinum persists in cultured
mammalian cells in a temperature-restricted fashion. Infect. Immun. 62:32223229.
Ramakrishnan L, Valdivia RH, McKerrow JH, Falkow S. Mycobacterium marinum
causes both long-term subclinical infection and acute disease in the leopard frog
(Rana pipiens). Infect. Immun. 1997. 65:767-773.
Recht, J., A. Martínez, S. Torello, and R. Kolter. 2000. Genetic analysis of sliding
motility in Mycobacterium smegmatis. J. Bacteriol. 182:4348-4351.
Riedel, K., M. Hentzer, O. Geinsenberg, B. Huber, A. Steidle, H. Wu, N. Høiby, M.
Givskov, S. Molin, L. Eberl. 2001. N-Acylhomocerine-Lactone-Mediated
Communication Between Pseudomonas aeruginosa and Burkholderia cepacia in
Mixed Biofilms. Microbiology. 147:3249-3262.
85
Rogall T, Wolters J, Flohr T, Bottger EC. 1990. Toward a phylogeny and definition of
species at the molecular level within the genus Mycobacterium. Int. J. Syst.
Bacteriol. 40:323-330. ¾
Ruley, K. M., J. H. Ansede, C. L. Pritchett, A. M. Talaat, R. Reimschuessel, and M.
Trucksis. 2004. Identification of Mycobacterium marinum virulence genes using
signature-tagged mutagenesis and the goldfish model of mycobacterial
pathogenesis. FEMS Microbiol. Lett. 232:75-81.
Sabev, H. A., G. D. Robson, P.S. Handley. 2006. Influence of Starvation, Surface
Attachment, and Biofilm Growth on the Biocide Susceptibility of the
Biodeteriogenic yeast Aerobasidium pullulans. J. of Appl. Microbiol. 101:319330.
Sauer, K., A. K. Camper, G. H. Ehrlich, J. W. Costerton, and D. G. Davies. 2002.
Pseudomonas aeruginosa displays multiple phenotypes during development as a
biofilm. J. Bacteriol. 184:1140-1154.
Schulze-Robbecke, R. 1993. Mycobacteria in the Environment. Immun. Infekt.
21(5):126-131.
Schulze-Robbecke, R., C. Feldmann, R. Fischeder, B. Janning, M. Exner, and G. Wahl.
1995. Dental units: an environmental study of sources of potentially pathogenic
mycobacteria. Tubercle Lung Dis. 76(4):318-323.
Schulze-Robbecke, R., B. Janning, and R. Fischeder. 1992. Occurrence of mycobacteria
in biofilm samples. Tuberc Lung Dis. 73(3):141-144.
Schwartz T., S. Kalmbach, S. Hoffmann, U. Szewzyk, U. Obst. 1998. PCR-based
detection of mycobacteria in biofilms from a drinking water distribution system.
J. Microbiol. Meth. 34:113-1237.
86
September, S. M., V. S. Brözel, and S. N. Venter. 2004. Diversity of nontuberculoid
Mycobacterium species in biofilms of urban and semiurban water distribution
systems. Appl. Environ. Microbiol. 70(12):7571-7573.
Shapiro, J.A. 1998. Thinking about bacterial populations as multicellular organisms.
Annu. Rev. Microbiol. 52:81-104.
Sherlock, O., R. M. Vejborg, and P. Klemm. 2005. The TibA Adhesin/Invasin from
Enterotoxigenic Escherichia coli Is Self Recognizing and Induces Bacterial
Aggregation and Biofilm Formation. Infect and Immun. 73(4):1954-1963.
Sherlock, O., M. A. Schembri, A. Reisner, P. Klemm. 2004. Novel Role for the AIDA
Adhesin from Diarrheagenic Escherichia coli: Cell Aggregation and Biofilm
Formation. J. Bactriol. 186(23):8058-8065.
Singleton, P., D. Sainsbury. 2001. Dictionary of Microbiology and Molecular Biology.
3rd Edition. John Wiley and Sons Ltd. West Sussex. England.
Skillman, L. C., I. W. Sutherland, M. V. Jones, A. Goulsbra. 1998. Green Fluorescent
Protein as a Novel Species-Specific Marker in Enteric Dula-Species Biofilms.
Microbiology. 144:2095-2101.
Slauch JM, Mahan MJ, Mekalanos JJ. 1994. In vivo expression technology for selection
of bacterial genes specifically induced in host tissues. Methods Enzymol.
235:481-4.
Stahl, D.A. and Urbance J.W. 1990. The division between fast- and slow-growing species
corresponds to natural relationships among the mycobacteria. J. Bacteriol.
172:116-124.
87
Steidle, A., K. Sigl, R. Schuhegger, A. Ihiring, M. Schmid, S. Gantner, M. Stoffels, K.
Riedle, M. Givskov, A. Hartmann, C. Langerbartels and L. Eberl. 2001.
Visualization of N-acylhomoserine Lactone-Mediated Cell-Cell Communication
between Bacteria Colonizing the Tomato Rhizosphere. Appl. Environ. Microbiol.
67(12):5761-5770.
Sternberg, C., B. B. Christensen, T. Johansen, A. T. Nielsen, J. B. Andersen, M. Givskov
and S. Molen. 1999. Distribution of Bacterial Growth Activity in Flow-Chamber
Biofilms. Appl. Environ. Microbiol. 65(9):4108-4117.
Stinear, T. P. et al. 2007. Reductive Evolution and Niche Adaptation Inferred from the
Genome of Mycobacterium ulcerans, the Causative Agent of the Buruli Ulcer.
Genome Res. 17:192-200.
Stoodley, P. K. Sauer, D. G. Davis, and J. W. Costerton. 2002. Annu. Rev. Microbiol.
56:187-209.
Talaat, A.M., R. Reimschuessel, S.S. Wasserman, M. Trucksis. 1998. Goldfish Carassius
auratus, a novel animal model for the study of Mycobacterium marinum
pathogenesis. Infect. Immun. 66:2938-2942.
TBSGC: TB Structural Genomics Consortium. 2007. http://www.doe-mbi.ucla.edu/TB/.
Tønjum, T., D.B. Welty, E. Jantzen, and P.L. Small. 1998. Differentiation of M.
marinum, and M. haemophilum: Mapping of their Relationships to M.
tuberculosis by Fatty Acid Profile Analysis, DNA-DNA Hybridization, and 16S
rRNA Gene Sequence Analysis. J. of Clin. Microbiol. 36(4):918-925.
Valdivia RH, Cormack BP and Falkow S. 1998. “GFP: Green fluorescent protein
strategies and applications”. M. Chalfie and S. Kain, eds. Wiley-Liss, New York.
88
Valdivia RH, and S. Falkow. 1996. Bacterial Genetics by Flow Cytometry: Rapid
Isolation of Salmonella typhimurium Acid-Inducible Promoters by Differential
Fluorescence Induction. Mol. Microbiol. 22(2):367-78
Valdivia RH, and Ramakrishnan L. 2000. Applications of gene fusions to green
fluorescent protein and flow cytometry to the study of bacterial gene expression in
host cells. Meth. Enzymol. 326:47-73.
Walker J.T., D.J. Bradshaw, A.M. Bennett, M.R. Fulford, M.V. Martin, P.D. Marsh.
2000. Microbial biofilm formation and contamination of dental-unit water
systems in general dental practice. Appl. Environ. Microbiol. 66:3363-3367.
Werner, E., F. Roe, A. Bungnicourt, M. J. Franklin, A. Heydorn, S. Molen, B. Pitts, P. S.
Stewart. 2004. Stratified Growth of Pseudomonas aeruginosa Biofilm.Appl.
Environ. Microbiol. 70(10):6188-6196.
89