Citation
Optimization of Methods for Phage Display Using Single-Chain Variable Fragment Phagemid Libraries

Material Information

Title:
Optimization of Methods for Phage Display Using Single-Chain Variable Fragment Phagemid Libraries
Creator:
Harpley, Crystal
Publisher:
University of Florida
Publication Date:
Language:
English

Thesis/Dissertation Information

Degree:
Master's ( M.S.)
Degree Grantor:
University of Florida
Degree Disciplines:
Medical Sciences
Medicine
Committee Chair:
Gulig, Paul A.
Committee Members:
Wright, Anita C.
Jin, Shouguang
Graduation Date:
8/9/2008

Subjects

Subjects / Keywords:
Antibodies ( jstor )
Antigens ( jstor )
Bacteriophages ( jstor )
Cellulose nitrate ( jstor )
Escherichia coli ( jstor )
Infections ( jstor )
Libraries ( jstor )
Particle production ( jstor )
Plasmids ( jstor )
Toxins ( jstor )
helper, hyperphage, mazur, phage, phagemid, scfv, shiga, stx2, tomlinson

Notes

Abstract:
Detection assays for biological agents and their products are important to identify disease-causing organisms in water, air, food, and patient samples to enable prevention or treatment of the disease. A commonality in almost all detection assays are proteins that specifically bind to a target molecule. While monoclonal antibodies are the most common detection reagents used in detection assays, phage display reagents are becoming more prevalent. Phage display involves the display of recombinant peptides on bacteriophages. These recombinant phages can be panned against a target antigen to select recombinant peptides that bind specifically to the antigen. Optimization of protocols and reagents used in phage display was the goal of this thesis. Optimization of whole cell panning with the Tomlinson J human synthetic VH + VL (scFv) phagemid library revealed that a trypsin elution between 10 and 30 minutes eluted the highest number of phagemid particles; this may be because the phagemid particles were degraded by trypsin treatment after 45 minutes. A high throughput ELISA for the production and screening of scFv proteins was developed and demonstrated that scFv proteins produced in a microtiter well could be screened by ELISA and produce a detectable S:N even when diluted 1:10. A high throughput ELISA for the production and screening of phagemid particles was attempted, but phagemid particles could not be produced in a microtiter well in titers high enough to make a high throughput ELISA acceptable for screening of phagemid particles. Various phagemid particle amplification tools were analyzed. The hyperphage genome was incorporated into E. coli and maintained as a plasmid to enable phagemid particle amplification. However, these strains had low infection efficiencies with phagemid particles. Helper plasmids were incorporated into E. coli to enable phagemid particle amplification. However, these strains produced low yields of phagemid particles. The titer of laboratory-produced hyperphage was improved by use of E. coli MG1655 (pGTR203), expressing the M13 gIII gene. The improved high titer hyperphage was used to successfully amplify the Tomlinson J scFv phagemid library to ensure that the redundancy of phagemid clones was maintained. Using the newly amplified library to pan against an E. coli O157:H7 Stx2 toxin preparation resulted in 89% of phagemid particles from the third round of panning recognizing the Stx2 toxin preparation in an ELISA. Previous panning procedures used in this laboratory resulted in no better than 1% of selected phagemids being specific to the target antigen, and often no usable specific clones were isolated. The optimization of panning procedures and reagents developed in this thesis greatly increased the efficiency of selection of phagemid particles specific to the target antigen. Using the improved reagents and procedures developed in this theis yielded over 80% positive phagemid clones selected against E. coli O157:H7 flagella by another lab member. The phagemids recognized the major flagellin protein by Western blot (data not shown). Therefore, the improvements on phage display techniques and tools described in this thesis are, in fact, useful and offer promise of success for continuing studies in the laboratory.

Record Information

Source Institution:
UFRGP
Holding Location:
University of Florida
Rights Management:
Copyright Crystal Harpley. Permission granted to the University of Florida to digitize, archive and distribute this item for non-profit research and educational purposes. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder.
Embargo Date:
8/31/2010

Downloads

This item has the following downloads:


Full Text






B 12


10


Ratio of Dilution, Dilution
z I/I 10.01, 1:100
v, /I 10.005, 1:200
0.003, 1:400
4 ~0.001, l:soo
2


0 0.002 0.004 0.006 0.008 0.01

Ratio of Dilution


Figure 3-4. Continued.










LIST OF REFERENCES


1. Lim, D.V., Simpson, J.M., Kearns, E.A., and Kramer, M.F. 2005. Current and developing
technologies for monitoring agents of bioterrorism and biowarfare. Clin. M~icrobiol. Rev.
18:583-607.

2. Peruski, L.F., Jr. and Peruski, A.H. 2003. Rapid diagnostic assays in the genomic biology
era: detection and identification of infectious disease and biological weapon agents.
Biotechniques 35:840-846.

3. Cieslak, T.J., Rowe, J.R., Kortepeter, M.G., Madsen, J.M., Newmark, J., Christopher,
G.W., Culpepper, R.C., and Eitzen, E.M., Jr. 2000. A field-expedient algorithmic
approach to the clinical management of chemical and biological casualties. M~il. M~ed.
165:659-662.

4. Franz, D.R., Jahrling, P.B., Friedlander, A.M., McClain, D.J., Hoover, D.L., Bryne,
W.R., Pavlin, J.A., Christopher, G.W., and Eitzen, E.M., Jr. 1997. Clinical recognition
and management of patients exposed to biological warfare agents. JAM~A 278:3 99-411i.

5. Franz, D.R., Jahrling, P.B., McClain, D.J., Hoover, D.L., Byrne, W.R., Pavlin, J.A.,
Christopher, G.W., Cieslak, T.J., Friedlander, A.M., and Eitzen, E.M., Jr. 2001. Clinical
recognition and management of patients exposed to biological warfare agents. Clin. Lab
M~ed. 21:435-473.

6. YALOW, R. S. and BERSON, S.A. 1959. Assay of plasma insulin in human subj ects by
immunological methods. Nature 184 (Suppl 21):1648-1649.

7. Saiki, R.K., Scharf, S., Faloona, F., Mullis, K.B., Horn, G.T., Erlich, H.A., and Arnheim,
N. 1992. Enzymatic amplification ofbeta-globin genomic sequences and restriction site
analysis for diagnosis of sickle cell anemia. 1985. Biotechnology 24:476-480.

8. Kaper, J.B., Morris, J.G., Jr., and Levine, M.M. 1995. Cholera. Clin. M~icrobiol. Rev.
8:48-86.

9. Sack, D.A., Sack, R.B., Nair, G.B., and Siddique, A.K. 2004. Cholera. Lancet 363:223-
233.

10. Gaffga, N.H., Tauxe, R.V., and Mintz, E.D. 2007. Cholera: a new homeland in Africa?
Am. J. Trop. M~ed. Hyg. 77:705-713.

11. Broeck, D.V., Horvath, C., and De Wolf, M. 2007. Pathogens in focus Vibrio Cholerae:
Cholera toxin. Thze International Journal ofBiochemistry & Cell Biology 39: 1771-1775.

12. Sanchez, J. and Holmgren, J. 2005. Virulence factors, pathogenesis and vaccine
protection in cholera and ETEC diarrhea. Curr. Opin. Immunol. 17:388-398.

13. 2001. Cholera vaccines. Wkly. Epideemiol. Rec. 76: 117-124.









rest resulted in colonies that had a diameter between 0.5 and 2 mm. A small and large of each

strain were further analyzed to determine if the different colony phenotypes also had different

functionality.

The six helper plasmid-containing E. coli TG1 strains were analyzed for their infection

effciencies with two phagemid particles in relation to the infection effciencies of E coli TG1

that did not contain any helper plasmids. Escherichia coli TG1 (M13cp-dg3-sm) and E. coli

TG1 (M13cp-CT-sm) were the only strains that yielded infection effciencies greater than 0.3%

of the infection effciency of E coli TG1 (Table 3-1). Escherichia coli TG1 (M13cp-dg3-sm)

had infection effciencies of 175% and 97%, while E. coli TG1 (M13cp-CT-sm) had infection

effciencies of 7% and 13% compared to the infection effciency of E coli TGl. The differences

in the infection effciencies between the small and large strains of helper plasmid-containing

E. coli revealed that they had different functionalities. Except for E. coli TG1 (M13cp-dg3-sm),

all of the other helper plasmid strains had reduced infection effciencies compared to E. coli

TGl. Therefore, the cost for the E. coli cells maintaining the helper plasmids resulted in a

decrease in their infection effciencies. Because E. coli TG1 (M13cp-dg3-sm) and E. coli TG1

(M13cp-CT-sm) had higher infection effciencies compared to their large phenotypes, it

suggested that the large strains had a defect in some component of the cell involved in

transduction.

Escherichia coli TG1 (M13cp-dg3-sm) had the highest infection effciencies of the helper

plasmid strains; therefore, it was further tested to determine if the high infection effciencies

could be maintained when tested with additional phagemid particles. Escherichia coli TG1

(M13cp-dg3-sm) had infection effciencies of 92%, 113%, and 120% compared to the infection

effciencies ofE. coli TG1 with Vc86, clonel8, and a-AV20N3 (Table 3-2). The infection









the most common elutions used in phage display, elutions by antigen are not uncommon and

have even been noted to be more successful than pH elutions because they decrease the elution

of phages that are not specific to the antigen (89-91).

Panning the Ph.D. 12mer library against phenol-water-extracted y. cholerae 01 LPS that

was immobilized onto nitrocellulose paper and eluted with either glycine (pH 2.2) or y. cholerae

01 LPS was performed. Three rounds of panning were performed with amplification of phages

after rounds one and two. A subtractive panning was performed after round one. This

"negative" panning was performed to decrease the amount of nonspecific phages in the library.

The negative panning involved panning the library against blocked nitrocellulose paper;

therefore, clones specific to the blocker, nitrocellulose paper, or polystyrene should be removed

from the library. Two hundred clones from each panning were individually amplified and

analyzed by ELISA. The 400 phage clones and a-Vc 01 LPS mAb were screened in an ELISA

against Y. cholerae 569B LPS. Anti-Vc 01 LPS mAb generated a S:N of 10, proving that the

LPS successfully bound to the microtiter plate. Anti-BSA phagemid particles were used in the

ELISA against BSA (10Cpg/mL) and generated a S:N of 15, proving that the ELISA was a

successful assay for the detection of phages. All 400 clones gave a S:N less than 1.7; therefore,

all of the clones screened were negative for binding to y. cholerae 01 LPS.

Previous attempts to select recombinant phages to LPS have failed to generate highly

specific recombinant phages that generally lack a consensus sequence (92-94). Also, the

isolation of recombinant phages that bind to LPS is less common than for the isolation of

recombinant phages that bind to protein. This is most likely because protein-protein interactions

are usually stronger than protein-carbohydrate interactions. The antibody-antigen complex is

held together by noncovalent interactions such as hydrogen bonds, van der Waals forces,









the lipid A-core plus repeating O-antigens (80,81). The non-staining bands were not the lipid A

because the lipid A contains two glucosamines that are stainable by silver stain. Therefore, the

TRlzol-extraction appears to have also extracted cellular lipids. Therefore, the TRIzol-extracted

LPS appeared to have more lipid contamination than the phenol-water-extracted LPS.

When the resolved LPS extractions were compared by silver stain to the commercial LPS,

the phenol-water-extracted LPS most closely resembled the commercial LPS. In the silver stain

for TRIzol-extracted LPS and commercial LPS, the TRlzol-extracted LPS had bands at 14 and

20 kDa with non-staining bands between 20 and 30 kDa, the commercial LPS had bands at 14,

20, and 23 kDa. In the silver stain for phenol-water-extracted LPS and commercial LPS, both

LP S prep arati on s had b and s at 1 3, 2 0, and 2 3 kD a. The phenol -water-extracte d LP S had

additional banding from 23-50 kDa, while the commercial LPS had more of a smear from 23-50

kDa. Previously published silver stains of y. cholerae 01 LPS have O-antigen banding/smears

from 20-50 kDa (80,81). Vibrio cholerae 01 LPS has 12 to 18 O-antigen groups (83); therefore,

the additional banding in the phenol-water-extracted LPS may be due to the ability of the phenol-

water method to extract LPS with more O-antigens than that of the method used for the

extraction of the commercial LPS which was phenol extracted and purified by gel-filtration

chromatography. Because the phenol-water-extracted LPS most closely resembled the

commercial LPS and appeared to have less lipid contamination than the TRlzol-extracted LPS,

the phenol -water-extracte d LP S was u se d for later exp erim ents.

Before y. cholerae 01 LPS was panned, a solid support capable of binding a high quantity

of yK cholerae 01 LPS was analyzed. In the past the Ph.D. 12mer phage library was panned

against antigens that were immobilized onto a microtiter well. Previous experiments from this

laboratory suggested that nitrocellulose paper could bind more LPS than a microtiter well. It was










dialyzed (MWCO: 3,500) for 24-48 hours at 40C. The final extraction was used in a Tsai-Frasch

silver stain to analyze the LPS.

TR~zol Reagent Extraction of LPS

TRlzol Reagent (Invitrogen), a mono-phasic solution of phenol and guanidine

isothiocyanate (77), was used to extract y. cholerae N16961 LPS from whole cells. Ten

milliliters of log phase (OD600 ~ 0.4-0.6) y. cholerae N16961 were centrifuged at 13,776 x g for

10 minutes at 40C. The supernatant was decanted, and the cells were suspended in 200 CIL of

TRlzol Reagent and incubated at room temperature for 10 tol5 minutes. Twenty microliters of

chloroform per mg of cells was added to the mixture, vortexed, and incubated at room

temperature for 10 minutes. The mixture was centrifuged at 12,000 x g for 10 minutes at room

temperature. The resulting aqueous phase was transferred to a new microcentrifuge tube and

stored on ice. One hundred microliters of ddH20 was added to the TRlzol-Reagent tube, and a

back extraction was performed. Two more back extractions were performed following the first

back extraction to give a total of four extractions. The extractions were dried with a Savant

SpeedVac Concentrator (Global Medical Instrumentation, Ramsey, MN) for 2.5 hours. The

pellet was suspended in 0.5 mL of 0.375 M MgCl in 95% (v/v) EtOH that was chilled to 40C.

The suspension was centrifuged at 12,000 g for 15 minutes at room temperature. The resulting

pellet was suspended in 200 CIL of ddH20. The final LPS extraction was analyzed in a Tsai-

Frasch silver stain (see below).

Extraction of Periplasmic Proteins

Periplasmic proteins from E. coli were extracted using a Tris-EDTA-Sucrose (TES)

extraction. A standing-overnight culture of an E. coli culture was diluted 1:400 in 100 mL of LB

medium and incubated overnight in a 370C shaking incubator. The next day the cells were

centrifuged at 13,776 x g for 10 minutes at 40C. The supernatant was stored at 40C. The pellet










agents. Because there are numerous detection environments, conditions, and agents, the

development and improvement of detection assays will continue to progress.

Vibrio cholerae

Vibrio cholerae is a gram-negative, curved rod-shaped bacterium with a single polar

flagellum found primarily in estuarine and marine environments. It is a facultative human

pathogen causing the pandemic diarrheal disease cholera (8). Cholera is characterized by

profuse watery diarrhea, vomiting, and leg cramps leading to dehydration and shock (9).

Cholera infection occurs through the ingestion of food or water contaminated with the bacterium.

Because cholera can be prevented by proper sanitation and hygiene, it is uncommon in

industrialized countries and is most prevalent in the Indian subcontinent and sub-Saharan Africa

(10).

Vibrio cholerae uses its toxin co-regulated pilus (TCP) to colonize the small intestine;

once attached to the small intestine, the bacterium secretes cholera toxin (CT) (11). Cholera

toxin binds to the epithelial cell receptor, Ghln, and is transported into the cell. Cholera toxin is

an AB toxin composed of a catalytically active A-subunit surrounded by a homopentameric B-

subunit. Once the CT is internalized, it is transported in a retrograde pathway through the Golgi

to the endoplasmic reticulum. In the endoplasmic reticulum it is retrotranslocated to the cytosol.

In the cytosol CT catalyzes ADP-ribosylation of the GTP-binding protein Gs (8) causing

adenylate cyclase to become constitutively activated. The increase in cAMP levels leads to

secretion of Cl~, HCO3-, and water from epithelial cells into the intestinal lumen causing diarrhea.

The loss of water can amount to 30 liters per day, and without proper rehydration treatment

can lead to a 30% mortality rate (12). Under-reporting of cholera infections is a great problem.

It is estimated that 3 to 5 million cases and 120,000-200,000 deaths occur worldwide annually

(13). However, this number could easily be multiplied by a factor of ten due to unreported cases.









fluorescence, enzymatic activity, chemiluminescence, electrochemiluminescence, metallic beads,

and many more. The detection of these complexes can be assayed on a variety of platforms such

as biosensors, flow cytometry, microarray, and lateral flow diffusion devices (1). Perhaps the

most common immunological assays are Enzyme-Linked ImmunoSorbent Assays (ELISAs).

Such assays either have antibody or antigen bound to a solid support that enables specific,

sensitive, and quantitative detection of antigen or antibody by optical-density detection of a

colorimetric signal.

Nucleic acid-based tools for detection have advanced in the past few decades. Of main

interest is Polymerase Chain Reaction (PCR), which was invented in 1983 by Kary Mullis and

coworkers (7). Polymerase chain reaction involves the amplification ofDNA by use of

oligonucleotide primers, heat stable DNA polymerase, and nucleotides in an exponential

capacity. The original PCR has since been altered to give quantitative real time-PCR (q-PCR or

kinetic PCR) and reverse transcription (RT)-PCR. Quantitative real time-PCR involves the

amplification along with quantification of the DNA, while RT-PCR involves reverse transcribing

a piece of RNA into DNA followed by PCR amplification of the DNA. Advances in PCR

chemistry and thermocyclers have shortened the length of DNA amplification from a few hours

to minutes. With small sample volumes in the amount of a few microliters and containing as

little as one bacterial cell, PCR is one of the most sensitive assays available. However, this

sensitivity increases the risk of contamination generating false positives. Field-based PCR

amplification and identification is not common due to the complexity of the system and highly

trained personnel required to operate and interpret the system. Perhaps the greatest constraint of

nucleic acid-based detection assays is the availability of genomic sequence data of biological

agents (2). Detection assays are continually being invented and improved to detect biological









in blocking buffer and analyzed by ELISA (Fig. 3-11). There was not a significant difference in

the S:N (p=0.07-0.48) of phagemid particles produced in microtiter wells with and without 0. 1%

(w/v) glucose. Therefore, the 0.1% (w/v) glucose in the medium was probably not enough

glucose to cause catabolite repression of the lac promoter. Vc86 phagemid particles produced in

microtiter wells with or without glucose at 300C and 370C did not produce a S:N greater than 1,

whereas Vc86 phagemid particles produced in a flask (S:N ~4) yielded a significantly higher S:N

(p=0.05). Observation of the turbidities of the bacterial cultures that were grown in a microtiter

well revealed only slightly turbid cultures compared to bacterial cultures that were grown in a

flask. The growth conditions may not have been optimal for the bacteria to grow when they had

the added stress of producing phagemid particles. It was possible that the bacterial cultures that

were grown to produce phagemid particles in a microtiter well experienced less aeration than

bacterial cultures grown in a flask or culture tube. The decreased aeration could decrease

bacterial growth and reduce phagemid particle production. Escherichia coli is a facultative

anaerobe which means it makes ATP, which drives biosynthesis, by aerobic respiration or

fermentation. Aerobic respiration generates 36 ATP molecules from one molecule of glucose,

while fermentation only generates 2 ATP molecules from one molecule of glucose. Because

biosynthesis is fueled by ATP, aerobic respiration is more effective for bacterial growth.

When a-BSA phagemid particles were produced at 300C, the S:N was significantly higher

(p<0.04) than when a-BSA phagemid particles were produced at 370C at either glucose

concentration in a microtiter well. It appeared that phagemid particle production was more

permissive at 300C opposed to 370C. The a-B SA phagemid particles produced in a flask (S:N

~12) yielded a significantly higher (p=0.02) S:N compared to a-BSA phagemid particles

produced in microtiter wells (S:N~7) when analyzed at a 1:2 dilution in an ELISA. The reduced










enough. The effort required to troubleshoot the colony blot was determined to not be worth the

time. Therefore, a way to improve the old method for producing scFvs was developed.

To enable high throughput production of scFv proteins, a clone was grown in a 96-well

plate instead of a culture tube or flask, and the scFv-containing supernatant was examined in an

ELISA. When a-BSA clone was grown in a 96-well plate, its scFv-containing supernatant

diluted 1:2 yielded a S:N of 11.5 against BSA in an ELISA. Anti-BSA scFv produced in a flask

yielded a S:N between 10 and 15; therefore, producing scFv proteins in a microtiter well may not

be optimal, but it is acceptable for the purpose of a high throughput screen. Next was to

determine how far the microtiter well-produced scFvs could be diluted and still generate a

sufficient signal to be detected in an ELISA. This test was to determine how many clones could

be mixed together (i.e., diluted) and screened in a single microtiter well to enable even greater

high throughput screening of scFvs. The a-BSA clone was grown in a microtiter well to produce

scFvs. The supernatants were diluted 1:2, 1:4, 1:6, and 1:10 in casein blocking buffer and

analyzed by ELISA (Fig. 3-10).

When a-BSA scFv-containing supernatant was diluted 1:10 it still yielded a S:N of 5.6 in

an ELISA. This S:N is high enough for detection of positive clones. However, the a-BSA clone

was a very strong clone; therefore, the maj ority of clones selected will have a weaker signal than

the a-BSA clone. To lessen the chance of not detecting scFvs with weak affinity or low titers, it

was decided to only dilute scFv clones 1:4 to 1:6 for high throughput screening by ELISA. This

would still enable screening of hundreds of clones per day, a vast improvement to the previous

method of screening clones in culture tubes or flasks which only enabled screening of a couple

dozen clones per day.

























V V V V V B B B B B

30 30 37 37 30 30 30 37 37 30
0 0.1 0 0.1 0.1 0 0.1 0 0.1 0.1
W W W W F W W W W F


Phagemid
Particle

Temp. (oC)
% Glucose
Grown in


Figure 3-11. Analysis of a-BSA and Vc86 phagemid particles produced in microtiter wells by
ELISA. Phagemid particles Vc86 (V) and a-BSA (B) were produced at either 300C
or 370C in medium containing either 0% or 0. 1% glucose in a microtiter well (W) or
in a flask (F). Microtiter wells were coated with 10Cpg/mL BSA, 3 x 10s CFU/mL
V. cholerae whole cells, or PBS. Phagemid particles were diluted 1:2 in blocking
buffer and incubated with the coated wells. Bound phagemid particles were detected
with HRP-conjugated anti-M13 secondary antibody. Wells were developed with
TMB substrate, and their absorbances were read at 630 nm. (*) represents p<0.05 and
(**) represents p<0.01. (n=2 wells)


































Table 3-4. Comparison of transducing units to particles per milliliter of amplified phagemid
particles. The spot-titer method was used to determine tu/mL. Particles/mL was
determined via a sandwich ELISA with anti-M13 antibodies. (n = 2 wells)
Phagemid
particle Amplification (Particles/mL)/
amplified tool used tu/mL Particles/mL (tu/mL)
Vc86 M13cp-CT-sm 8.7 x 105 2.0 x 1010 23000
Vc86 M13cp-CT-sm 4.6 x 106 3.4 x 109 740
clone 18 M13 cp-dg3 -sm 2.4 x 10s 2.2 x 10s 1
a-AV20N3 M13cp-dg3-sm 4.4 x 107 7.0 x 107 2
a-BSA M13cp-dg3-sm 1.3 x 10s 1.8 x 10s 1
Vc86 M13cp-dg3-sm 8.7 x 106 1.7 x 109 190
clone 18 Hyperphage 7.1 x10s 1.1 x1010 15
a-AV20N3 Hyperphage 2.0 x 109 1.1 x 1010 6
a-BSA Hyperphage 5.8 x 10s 1.4 x 1010 24


Table 3-3. Infection efficiencies of E coli TG1 (M13cp-CT-sm) and E. coli JM109 (M13cp-CT-
sm) with phagemid particles. Bacterial strains were transduced with phagemid
particles at a MOI of 0. 1, and transduced bacteria were enumerated. Infection
efficiency is the number of transductions with a phagemid particle into a helper
plasmid-containing bacterial strain divided by the number of transductions with the
same phagemid particle into the bacterial strain without the helper plasmid.
Infection
Strain Phagemid Transductions efficiency %


E. coli TG1
E. coli TG1 (M13cp-CT-sm)
E. coli TG1
E. coli TG1 (M13cp-CT-sm)
E. coli TG1
E. coli TG1 (M13cp-CT-sm)
E. coli JM109
E. coli JM109 (M13cp-CT-sm)
E. coli JM109
E. coli JM109 (M13cp-CT-sm)
E. coli JM109
E. coli JM109 (M13cp-CT-sm)


Vc86
Vc86
Clone 18
Clone 18
a-AV20N3
a-AV20N3
Vc86
Vc86
Clone 18
Clone 18
a-AV20N3
a-AV20N3


2.4 x 10
1.3 x 10
1.4 x 10
5.5 x 104
9.3 x 104
4.4 x 104
6.8 x 10
1.5 x 10
1.7 x10
6.1 x104
1.4 x 10
3.7 x 104


100
54
100
39
100
47
100
22
100
36
100
26










47. Liu, Y., Gilchrist, A., Zhang, J., and Li, X.F. 2008. Detection of viable but nonculturable
Escherichia coli 0157:H7 bacteria in drinking water and river water. Appl. Environ.
M~icrobiol. 74:1502-1507.

48. Kalnauwakul, S., Phengmak, M., Kongmuang, U., Nakaguchi, Y., and Nishibuchi, M.
2007. Examination of diarrheal stools in Hat Yai City, South Thailand, for Escherichia
coli 0157 and other diarrheagenic Escherichia coli using immunomagnetic separation
and PCR method. .Sambeastl~ Asian J. Trop. M~ed. Public Health 38:871-880.

49. Stefan, A., Scaramagli, S., Bergami, R., Mazzini, C., Barbanera, M., Perelle, S., and
Fach, P. 2007. Real-time PCR and enzyme-linked fluorescent assay methods for
detecting Shiga-toxin-producing Escherichia coli in mincemeat samples. Can. J
M~icrobiol. 53:337-342.

50. Zhang, W., Bielaszewska, M., Pulz, M., Becker, K., Friedrich, A.W., Karch, H., and
Kuczius, T. 2008. New immuno-PCR assay for detection of low concentrations of shiga
toxin 2 and its variants. J. Clin. M~icrobiol. 46:1292-1297.

51. Auvray, F., Lecureuil, C., Tache, J., Leclerc, V., Deperrois, V., and Lombard, B. 2007.
Detection, isolation and characterization of Shiga toxin-producing Escherichia coli in
retail-minced beef using PCR-based techniques, immunoassays and colony hybridization.
Lett. Appl. M~icrobiol. 45:646-651.

52. Smith, G.P. 1985. Filamentous fusion phage: novel expression vectors that display cloned
antigens on the virion surface. Science 228:1315-1317.

53. Carmen, S. and Jermutus, L. 2002. Concepts in antibody phage display. Brief: Funct.
G~enomic. Proteomic. 1: 189-203.

54. Endemann, H. and Model, P. 1995. Location of filamentous phage minor coat proteins in
phage and in infected cells. Journal of2~olecular Biology 250:496-506.

55. Henry, T.J. and Pratt, D. 1969. The proteins of bacteriophage M13. Proc. Natl. Acad'. Sci.
U. S. A 62:800-807.

56. Russel, M. and Model, P. 1989. Genetic analysis of the filamentous bacteriophage
packaging signal and of the proteins that interact with it. J. Virol. 63:3284-3295.

57. Messing, J. 1991. Cloning in M13 phage or how to use biology at its best. Gene 100:3-
12.

58. Kwasnikowski, P., Kristensen, P., and Markiewicz, W.T. 2005. Multivalent display
system on filamentous bacteriophage pVII minor coat protein. J. Immunol. Methods
307:135-143.

59. Gao, C., Mao, S., Lo, C.H., Wirsching, P., Lerner, R.A., and Janda, K.D. 1999. Making
artificial antibodies: a format for phage display of combinatorial heterodimeric arrays.
Proc. Natl. Acad'. Sci. U. S. A 96:6025-6030.









CHAPTER 2
MATERIALS AND METHODS

Bacterial Strains, Media, and Growth Methods

The bacterial strains used and their genotypes are listed in Table 2-1. All E. coli strains

were grown in LB broth (10 g tryptone, 5 g yeast extract, 5 g NaC1, and 3 mL of 1 M NaOH in 1

L water) or on LB agar plates containing 1.5% (w/v) agar. Exceptions are E. coli TGl, E. coli

DH~a (pNR100), and E. coli ER2738. Escherichia coli TG1 was grown in 2xTY broth (16 g

tryptone, 10 g yeast extract, 5 g NaC1, and 3 mL of 1 M NaOH in 1 L water) or on 2xTY agar

plates containing 1.5% (w/v) agar. Escherichia coli DH~a (pNR100) was grown in LB broth

containing 100 Cpg/mL ampicillin or on LB agar plates containing 100 Cpg/mL ampicillin and

1.5% (w/v) agar. Escherichia coli ER273 8 was grown in LB broth containing 20 Cpg/mL

tetracycline or on LB agar plates containing 20 Cpg/mL tetracycline and 1.5% (w/v) agar. All

strains of y. cholerae were grown in Luria Bertani broth containing with physiological saline

(LB-N) (10 g tryptone, 5 g yeast extract, 8.5 g NaC1, and 3 mL of 1 M NaOH in 1 L water) or on

LB-N agar plates containing 1.5% (w/v) agar. The most common medium used was 2xTY

containing 100 Cpg/mL ampicillin and 1% (w/v) glucose; therefore, it was abbreviated as 2xTY

AG. Escherichia coli strains harboring hyperphage were grown in broth or on plates as

specified, with the addition of 40 Cpg/mL kanamycin. Escherichia coli strains harboring helper

plasmids were grown in broth or on plates as specified above, with the addition of 30 Cpg/mL

chl orampheni col .

All bacterial cultures were initially grown as a standing-overnight culture. A standing-

overnight culture was made from 10 mL of specified medium that was inoculated with bacteria

from an agar plate. The standing-overnight culture was grown overnight (~16 h) in a 370C

incubator. A log-phase culture was obtained by diluting a standing-overnight culture 1:40 in









(pGTR203) (Hye) were between 0.3 and 1.0. Therefore the amount of particles and transducing

units of hyperphage produced from E. coli EC 100D (pGTR203) (Hye) and E. coli MGl655

(pGTR203) (Hye) were approximately the same, meaning that the hyperphages produced were

infectious. Both F- E. coli strains generated higher titers of hyperphage than did the two F

E. coli strains. Concentration of hyperphage produced from E. coli MGl655 (pGTR203) (Hye)

should generate a high enough titer of hyperphage to enable high quality and high quantity

production of the Tomlinson scFv libraries.

Escherichia coli MGl655 (pGTR203) (Hye) isolate 5 yielded the highest titer of

hyperphage (5.4 x 109 tu/mL); therefore, it was used to produce a large quantity of hyperphage to

be used to amplify the Tomlinson J scFv phagemid library. One liter of hyperphage was

produced from E. coli MGl655 (pGTR203) (Hye) 5 with a titer of 2.0 x 109 tu/mL, and the

phages were concentrated by PEG precipitation (see Materials and Methods) to a volume of 2.8

mL with a final titer of 3 x 1011 tu/mL. This concentrated hyperphage was used to amplify the

Tomlinson J scFv phagemid library.

The original Tomlinson J scFv library was amplified to ensure high quality and high

quantity of phagemid particles produced. The Tomlinson J scFv phagemid library was originally

received in E. coli TGl. This stock was grown in batch culture and frozen. Therefore, clones

were not lost due to poor phagemid particle amplification with the low titer hyperphage because

we grew the original bacterial stock of the library and not the previously made phagemid particle

stock of the library that most likely lost redundancy of clones because it was amplified with low

titer commercial hyperphage. A frozen stock of E coli TG1 (Tomlinson J scFv phagemid

library) was thawed on ice and inoculated into one liter of 2xTY AG broth. This culture was

grown in a 370C shaking incubator until the culture reached log phase. Thirty minutes into the









mAb Fab scFv


VVH



CH2~ Linker
cH3,

Figre1-. trctresofanibdis.m~ (mnoloalanibdy,Fa(fgmnatie






Figrligh Srutu chain).Thies heav andightchainsomAb antbdy) Fab fantiodes arehedtogethe

by disulfide bonds, and the heavy and light chains of the scFv antibody are fused
together by a flexible glycine-serine linker.









phagemid particles could not be produced in a microtiter well in titers high enough to make a

high throughput ELISA acceptable for screening of phagemid particles.

Various phagemid particle amplification tools were analyzed. The hyperphage genome

was incorporated into E. coli and maintained as a plasmid to enable phagemid particle

amplification. However, these strains had low infection efficiencies with phagemid particles.

Helper plasmids were incorporated into E. coli to enable phagemid particle amplification.

However, these strains produced low yields of phagemid particles. The titer of laboratory-

produced hyperphage was improved by use of E coli MGl655 (pGTR203), expressing the M13

giIl gene. The improved high titer hyperphage was used to successfully amplify the Tomlinson J

scFv phagemid library to ensure that the redundancy of phagemid clones was maintained. Using

the newly amplified library to pan against an E. coli 0157:H7 Stx2 toxin preparation resulted in

89% of phagemid particles from the third round of panning recognizing the Stx2 toxin

preparation in an ELISA.

Previous panning procedures used in this laboratory resulted in no better than 1% of

selected phagemids being specific to the target antigen, and often no usable specific clones were

isolated. The optimization of panning procedures and reagents developed in this thesis greatly

increased the efficiency of selection of phagemid particles specific to the target antigen. Using

the improved reagents and procedures developed in this this yielded over 80% positive

phagemid clones selected against E. coli 0157:H7 flagella by another lab member. The

phagemids recognized the maj or flagellin protein by Western blot (data not shown). Therefore,

the improvements on phage display techniques and tools described in this thesis are, in fact,

useful and offer promise of success for continuing studies in the laboratory.










LIST OF FIGURES


Finr IPg e

1-1 Structure of M 13 phage. ............. ...............32.....

1-2 Structures of antib odi es ................. ...............33...............

1-3 Genetic map of plT2 phagemid vector from the Tomlinson scFv library. .................. .....34

3-1 Analysis of TRIzol Reagent-extracted y. cholerae N16961 LPS by SDS-PAGE. .........100

3-2 Analysis of phenol-water-extracted y. cholerae 569B LPS by SDS-PAGE. ................101

3-3 Analysis of the saturation limit of y. cholerae 569B LPS to microtiter wells by
ELI SA ............. ...............102....

3-4 Analysis of the saturation limits of primary antibodies to y. cholerae 569B LPS-
coated microtiter wells by ELISA................ ...............103

3-5 Saturation of y. cholerae 569B LPS to nitrocellulose paper. .............. ....................10

3-6 Titers of eluted phagemid particles from the first one-round panning optimization
experiment. ........... ......__ ...............106..

3-7 Titers of eluted phagemid particles from the second one-round panning optimization
experiment. ........... ......__ ...............107..

3-8 Titers of eluted phagemid particles from the third one-round panning optimization
experiment. ........... ......__ ...............108..

3-9 Analysis of eluted phagemid particles from the third one-round panning optimization
experiment by ELISA. ............. ...............109....

3-10 Analysis of anti-B SA (a-BSA) scFv proteins produced in microtiter wells by ELISA. .1 10

3-11 Analysis of a-BSA and Vc86 phagemid particles produced in microtiter wells by
ELISA ........... ..... .. ...............111...

3-12 Analysis of a-BSA phagemid particles produced in microtiter wells by ELISA. ...........1 12

3-13 Analysis of a-B SA phagemid particles produced by hyperphage by ELISA. ................. 113

3-14 Analysis of phagemid particles (1-22, 43-64) selected from panning against Stx2
toxin preparation by ELISA ........._.___..... .___ ...............114...

3-15 Analysis of phagemid particles (1, 3-5, 23-42, 65-84) selected from panning against
Stx2 toxin preparation by ELISA. ........._._._ ....__ ...............115..









CHAPTER 4
DISCUSSION

Detection of biological agents and their products has become a prime importance in the

United States (85) since the anthrax attacks of 2001 (86). Early detection of biological agents is

important to prevent or inhibit the spread of disease. Detecting biological agents in water, air,

and food enables the proper course of action to be implemented to lessen or prevent the effects of

the organism. Detection of biological agents or their products is also important for diagnosis of a

disease. Detection of the disease-causing agent in patient samples enables the proper treatment

to given.

There are vast numbers of detection assays, and a common component to most of them is a

protein that binds specifically to a target antigen. Monoclonal antibodies are the default tool that

scientists use today to detect target molecules; however, monoclonal antibodies have some

drawbacks that are causing the use of phage display reagents to become more prevalent and take

the place of monoclonal antibodies (2). First, the generation of monoclonal antibodies requires

killing animals and takes months. Also, monoclonal antibodies are rarely isolated to pathogenic

or self antigens, causing limitations in the scope of the target antigen for the generation of

monoclonal antibodies. Phage display is advantageous because it lacks many of the drawbacks

of monoclonal antibodies. In phage display, if using a non-immunized library, the killing of

animals is not required and the isolation of specific recombinant phages can be done in a couple

weeks instead of months. Also, the scope of antigen used for phage display is far greater than

that used for monoclonal antibodies because phage display enables the selection of recombinant

proteins to pathogenic and self antigens. Phage display was developed in 1985 by George P.

Smith as a means to display foreign proteins on a filamentous phage (52). The sequence for the

foreign peptide is encoded in frame with a coat protein gene of the phage genome. This enables









from the panning with the library that contained 1% Vc86 phagemid particles, 10s whole cells

yK cholerae, and an elution time of 30 minutes generated a significantly higher S:N (p=0.01-

0.04) in an ELISA compared to the amplified phagemid particles from the other panning

variations. This suggests that panning with 10"s whole cells with a trypsin elution time of 30

minutes is optimal for selecting phagemid particles that are specific to the panned whole cells

compared to using 109 whole cells or a trypsin elution of 10 minutes. There might have been an

excess of whole cells when panning against 109 whole cells. This excess may have promoted the

binding of nonspecific phages; therefore, fewer whole cells present may have been more efficient

for panning because with fewer whole cells present the specific phagemid particles should

preferentially bind the whole cells and the number of nonspecific phagemid particles would be

lessened. Eluting for 10 minutes may not have been long enough for the trypsin to cleave all the

specific phagemid particles from the whole cells; therefore, eluting for 30 minutes was optimal

because it may have given the trypsin enough time to elute the phagemid particles from the

whole cells.

After the biopanning process is complete, the selected scFvs or phages are screened for

their specifieity to the target molecule. The previous method to screen scFv proteins was to

grow amber suppressor-free phagemid-containing bacteria under conditions to secrete scFv

proteins in culture tubes or flasks. However, this process was very time consuming; therefore, a

way to screen scFv proteins in a high throughput manor was investigated. High throughput

screening of scFv proteins is very common. A Biacore A 100 array system is capable of

screening hundreds of scFv proteins per day (97). This system screens scFv proteins from crude

bacterial extracts on a sensor chip surface. The antibodies that are captured onto the chip are

analyzed and ranked by the percentage of bound scFv proteins remaining on the chip after









unless they are used at concentrations of at least 1 x 106 particles/mL (data not shown). Because

the spot titer method measures the number of infectious phagemid particles, it is possible that the

total number of phagemid particles far exceeded the number of infectious phagemid particles.

To determine if the total number of phagemid particles is greater than the number of infectious

particles an anti-M13 sandwich ELISA was used to enumerate the particles/mL of phagemid

particle-containing supernatants (Table 3-4). Phagemid particles amplified by E. coli TG1

(M13cp-CT-sm) had particle/mL to tu/mL ratios between 740 and 23,000, E. coli TG1 (M13cp-

dg3-sm) had particle/mL to tu/mL ratios between 1 and 190, while phagemid particles amplified

by E. coli TG1 that was superinfected with hyperphage had particle/mL to tu/mL ratios between

6 and 24. Phagemid particles produced by helper plasmid-containing E. coli had particle/mL to

tu/mL ratios that were much higher than those for phagemid particles produced by E. coli that

was superinfected hyperphage. In biopanning the number of phagemid particles is not important,

the number of infectious phagemid particles is what matters because if a phagemid particle is not

infectious then it can not enter a bacterial cell and be produced to eventually be selected and

screened. Therefore, even though the number of particles produced by helper plasmids was high,

the number of infectious particles produced was low, making the helper plasmids an inadequate

amplification tool for the library.

The helper plasmids yielded phagemid particle progeny at titers between 105 tol0s tu/mL.

In a last attempt to get the helper plasmids to generate high titers of phagemid particles, the

M13 cp-dg3 -sm helper plasmid was incorporated into E. coli MGl65 5 to determine if a less

complex E. coli strain could generate higher titers of phagemid particles. Escherichia coli

MGl655 was chosen because it has few known mutations and does not contain the F-plasmid.

Four different phagemid particles were amplified with E. coli MGl655 (M13cp-dg3-sm), and the









particle amplification by helper plasmids was not an optimal amplification tool because it

produced phagemid particle titers too low to be used in biopanning. Producing homemade

hyperphage in E. coli MGl655 (pGTR203) enabled high yield of hyperphage particles. The

improved high titer homemade hyperphage was used to successfully amplify the Tomlinson J

scFv phagemid library to ensure that the redundancy of phagemid clones was maintained.

Specific Aim 4: Isolation of Specific Recombinant Phage to Stx2 Toxin of E. coli 0157:H17

The Tomlinson J scFv phagemid library that was amplified with homemade hyperphage

was panned against Stx2 toxin of E coli 0157:H7. Shiga toxins (Stx) kill mammalian cells and

cause severe disease from infection with E. coli 0157:H7 cells. Escherichia coli 0157:H7 can

produce Shiga toxins 1 and/or 2. Escherichia coli harboring the Stx2 toxin appears to be more

virulent than E. coli harboring Stxl or both Stxl and 2 toxins (36-39). Therefore, Stx2 toxin was

the target molecule chosen for panning. The Stx2 toxin preparation that was used to pan against

was obtained from Toxin Technology (Sarasota, FL) and was stated to contain 50% Stx2 toxin.

The other 50% of proteins in the Stx2 toxin preparation was most likely composed of

contaminating E. coli HB 101 proteins because the Stx2 toxin preparation was prepared from

E. coli HB 101 (pMJ100) lysates. pMJ100 encodes Stx2 toxoid.

A three round biopanning process was performed to pan the Tomlinson J scFv phagemid

library against a Stx2 toxin preparation-coated immunotube (see Materials and Methods) with

amplification between panning rounds to amplify the selected phagemid particles. Phagemid

particles were eluted using trypsin treatment for 10 minutes. Phagemid particles contain a

trypsin cleavage site between the scFv and the pIll. Therefore, phagemid particles that were

bound to Stx2 toxin preparation by their scFv could be eluted from the Stx2 toxin preparation to

be amplified for further rounds of panning. Approximately 1 x 1012 phagemid particles were

added to an immunotube that was coated with 100 Cpg of Stx2 toxin preparation. Approximately











lac promoter
colE1 ori RBS




V r a l evpe/B le ad e r

ScVariable Heavy
ba plT2 Linker

rnyc tag
amber stop codon

M113 ori
glnt



Figure 1-3. Genetic map of plT2 phagemid vector from the Tomlinson scFv library. RBS-
ribosome binding site. pelB leader peptide sequence promotes export of the scFv
protein. Variable Heavy and Variable Light peptide sequences are fused together by
a glycine-serine linker. An amber stop codon is at the junction of the c-myc tag and
the gIII gene to enable conditional expression of the scFv-pIII fusion in an amber
suppressor strain. The M13 origin of replication enables packaging into M13 phage
particles, the bla gene encodes ampicillin resistance, and the colEl origin of
replication enables maintenance as a plasmid in E. coli.














15


S10







Superinfected with (Hye-s m) (Hye-lg)
Hye
a-BSA Phagemid Particles


Figure 3-13. Analysis of anti-BSA (a-BSA) phagemid particles produced by hyperphage by
ELISA. Microtiter wells were coated with 10 Cpg/mL B SA or PB S and incubated with
1 x 10s tu/mL a-BSA phagemid particles produced from E. coli TG1 strains already
infected with hyperphage (Hye-sm, Hye-lg) or from E. coli TG1 that was
superinfected with hyperphage. Bound phagemid particles were detected with HRP-
conjugated anti-M13 secondary antibody. Wells were developed with TMB
substrate, and their absorbances were read at 630 nm (n=3 wells). The S:N from a-
BSA phagemid particles produced from (Hye-sm) was significantly lower (*p<0.01)
than the S:N from a-BSA phagemid particles produced from (Hye-lg) or from
superinfection with hyperphage.









the foreign peptide to be fused to the coat protein, which effectively links the phenotype and

genotype of the phage. Therefore, every phage that displays a foreign protein also encodes its

peptide sequence in the genome.

This thesis describes efforts to optimize phage display methods to increase the chances of

isolating recombinant antibodies or peptides for the potential use in detection assays.

Specific Aim 1: Panning to K cholerae LPS

Vibrio cholerae is a human pathogen that is the causative agent of cholera. Cholera is a

diarrheal disease that affects millions of people worldwide annually (9). Detecting K. cholerae

in the environment and patient samples is the first step to preventing and treating the disease. To

detect Y. cholerae we need to isolate a recombinant phage that binds specifically and sensitively

to Y. cholerae. There are over 200 serogroups of y. cholerae that are classified on the basis of

the O-antigen of their LPS; however, only serogroups 01 and 0139 cause endemic acute

diarrhea. Therefore, the LPS of these two serogroups would be excellent targets for detection.

Vibrio cholerae 01 LPS was chosen as a target for biopanning. Because y. cholerae 01

LPS was no longer sold commercially, it had to be extracted in the laboratory before any panning

experiments could be performed. The previous method to extract LPS in the laboratory was a

phenol-water extraction (76). However, a new LPS extraction was developed using a

commercial RNA isolation reagent, TRlzol Reagent, that could extract LPS in a fraction of the

time of the phenol-water extraction (77). The TRlzol method was even claimed to be able to

extract LPS with less degradation and contamination than the phenol-water method. To

determine which method was better, both methods were used to extract y. cholerae 01 LPS and

the extracted LPS was analyzed and compared to commercially acquired y. cholerae 01 LPS

(Sigma-Aldrich) that had been phenol extracted and purified by gel-filtration chromatography.









100
90
E 80
3 70
S60 -- a lo cells ve
O ~109 cells Vc
50 aC18e16cl~s
4-5 109 cells Vc, 1% Vc86
c, 40- ocelVc1%V8

m, 20
10


15 30 45 60

Trypsin Elution (min)

Figure 3-6. Titers of eluted phagemid particles from the first one-round panning optimization
experiment. The Tomlinson J library +/- 1% Vc86 phagemid particles was panned
against 10s and 109 whole cells of y. cholerae N16961 in suspension. Phagemid
particles that were bound to the y. cholerae whole cells were eluted with 10% (v/v)
trypsin stock (10 mg/mL trypsin, 50 mM Tris-HCI (pH 7.4), 1 mM CaCl2 in water) in
PBS for 15, 30, 45, and 60 minutes. Eluted phagemid particles were enumerated via
the spot-titer method.









particles, which only encode the giII gene, require phage amplification tools to provide phage

genes I-XI to enable phagemid particle production. We chose the Ph.D. 12mer phage library

(NEB) because the Tomlinson scFv phagemid library was disadvantageous due to the poor

quality of current methods to amplify phagemid particles. Because the Ph.D. 12mer phage

library is M13 phage, it does not require the aid of tools to amplify phages like a phagemid

library requires. Before the Ph.D. 12mer phage display library was to be panned against

yK cholerae 01 LPS, an extraction method for y. cholerae 01 LPS and the best solid support for

LPS binding had to be determined. Two LPS extraction methods were compared to determine

which method generated the best quality of LPS to pan against. Also, two binding supports

(microtiter well and nitrocellulose) paper were analyzed to determine which binding support

could bind the most LPS.

The Phenol-Water Method Extracted E cholerae LPS Most Closely Resembled the
Commercially Acquired E cholerae LPS

The phenol-water method for the extraction of LPS (76) was the previous lab procedure

used to extract LPS. However, a new method for LPS extraction involving TRlzol Reagent

(Invitrogen) was described by Yi and Hackett (77) that was stated to be a more efficient method

for LPS extraction than the phenol-water method. The TRlzol Reagent LPS extraction can be

executed in one day, while the phenol-water LPS extraction requires nearly a week. The TRlzol

Reagent method for LPS extraction is also stated to be a cleaner method than the conventional

phenol-water method for LPS extraction by extracting LPS with less degradation and

contamination. TRlzol Reagent is composed of phenol and guanidinium thiocyanate in aqueous

phase, and upon addition of chloroform the TRIzol Reagent can be used to extract LPS.

Vibrio cholerae N16961 LPS was extracted by the TRlzol Reagent extraction of LPS (see

Materials and Methods). The TRlzol Reagent-extracted y. cholerae LPS was resolved by SDS-


















10









Figure 3-15. Analysis of phagemid particles (1, 3-5, 23-42, 65-84) selected from panning against
Stx2 toxin preparation by ELISA. Microtiter wells were coated with 10 Cpg/mL of
Stx2 toxin or carbonate coating buffer and incubated with clones 1, 3 through 5, 23
through 42, and 65 through 84 diluted 1:2 in blocking buffer. Bound phagemid
particles were detected with HRP-conjugated anti-M13 monoclonal antibody. Wells
were developed with TMB substrate, and their absorbances were read at 630 nm (n=2
wells).










0.8


0.7

0.6


0.3


0.2

0.1

0.0
30oC, 0% 30oC, 0.1% 37oC, 0% 37oC, 0.1% 30oC, 0.1%
glucose glucose glucose glucose glucose, in
flask

ar-BSA Phagemid Particles


Figure 3-12. Analysis of a-B SA phagemid particles produced in microtiter wells by ELISA.
Microtiter wells were coated with 10Cpg/mL BSA. Phagemid particles were diluted
1:2 in blocking buffer and incubated with the coated wells. Bound phagemid
particles were detected with HRP-conjugated anti-M13 secondary antibody. Wells
were developed with TMB substrate, and their absorbances were read at 630 nm (n=2
wells). Phagemid particles produced in a microtiter well at 300C produced
significantly higher (*p=0.03) S:N values than phagemid particles produced in a
microtiter well at 370C at the same glucose concentration.








I
E
II,
CR

Cf)
PL
I
Q)

O
~
O
3j


32


10


Figure 3-5. Saturation of y. cholerae 569B LPS to nitrocellulose paper. Vibrio cholerae 569B
LPS was incubated with nitrocellulose paper at concentrations from 1-100 Cpg/mL.
The primary antibody used to detect the LPS was rabbit anti-Y. cholerae 01 LPS
polyclonal antibody (1:400). Bound primary antibody was detected with HRP-
conjugated goat anti-rabbit antiserum and detected with ECL substrate. The
nitrocellulose paper became saturated with y. cholerae 569B LPS at a LPS
concentration of 100 Cpg/mL. The strips incubated with 0 Cpg/mL LPS failed to stain
significantly .


II



II




lil


100


CII




I!111~


3 1









There are many different serogroups of y. cholerae, most of which do not cause acute

diarrhea. Vibrio cholerae is classified into serovars or serogroups on the basis of its

lipopolysaccharide (LPS) O-antigen (14). There are at least 200 known serogroups, of which

serogroups 01 and 0139 are the only ones that cause epidemic or endemic cholera. The 01

serogroup can be further distinguished into three serotypes. Ogawa and Inaba are the most

common serotypes, with Hikojima being rarely reported. These serotypes can be further

classified into two biotypes, El Tor and classical, that differ in their biochemical properties and

phage susceptibilities (8).

With the enormous numbers of cases of cholera every year, there is a need for an effective

diagnosis tool for patient and environmental samples. Detecting endemic serogroups of

yK cholerae early in an outbreak is extremely important for control of an epidemic. A problem

with detection of y. cholerae is that cholera is a disease of developing countries. Outbreaks

normally occur around water-ravaged areas where laboratories are not prevalent. Therefore,

Hield-based assays are the most effective tools for early detection of y. cholerae. Because there

are many serogroups of y. cholerae, most of which present mild symptoms, it is important to

distinguish the epidemic strains from the non-epidemic strains. Vibrio cholerae 01 and 0139

are the epidemic strains. Determining features of these serogroups are the O-antigen of their

LPS and their ability to produce CT. Therefore, Hield tests usually detect these specific antigens.

There are many rapid diagnostic tests for cholera. Some of these detect CT by passive-

latex agglutination (15,16), while others detect the O-antigens of LPS from 01 (17-21) and 0139

(22-24) strains of y. cholerae. A commonly used rapid y. cholerae diagnostic tool is the

multistep colloidal gold-based colorimetric immunoassay known as SMART. This monoclonal

antibody-based test was developed for the detection of y. cholerae 01 (25,26) and 0139 (24)










84. Adey, N.B., Mataragnon, A.H., Rider, J.E., Carter, J.M., and Kay, B.K. 1995.
Characterization of phage that bind plastic from phage-displayed random peptide
libraries. Gene 156:27-31.

85. Persell, D.J. and Robinson, C.H. 2008. Detection and early identification in bioterrorism
events. Fant. Conanunity Health 31:4-16.

86. 2001. Update: Investigation of bioterrorism-related anthrax and interim guidelines for
clinical evaluation of persons with possible anthrax. MWdWR M~orb. Mortal. Wkly. Rep.
50:941-948.

87. Nakamura, M., Watanabe, H., Nishimiya, Y., Tsumoto, K., Ishimura, K., and Kumagai, I.
2001. Panning of a phage VH library using nitrocellulose membranes: application to
selection of a human VH library. J. Biochent. 129:209-212.

88. Yarmush, M.L., Antonsen, K.P., Sundaram, S., and Yarmush, D.M. 1992.
Immunoadsorption: strategies for antigen elution and production of reusable adsorbents.
Biotechnol. Prog. 8: 168-178.

89. Nagumo, Y., Oguri, H., Tsumoto, K., Shindo, Y., Hirama, M., Tsumuraya, T., Fujii, I.,
Tomioka, Y., Mizugaki, M., and Kumagai, I. 2004. Phage-display selection of antibodies
to the left end of CTX3C using synthetic fragments. J. Inanunol. Methods 289: 137-146.

90. Goletz, S., Christensen, P.A., Kristensen, P., Blohm, D., Tomlinson, I., Winter, G., and
Karsten, U. 2002. Selection of large diversities of antiidiotypic antibody fragments by
phage display. J. Mol1. Biol. 315:1087-1097.

91. Meulemans, E.V., Slobbe, R., Wasterval, P., Ramaekers, F.C., and van Eys, G.J. 1994.
Selection of phage-displayed antibodies specific for a cytoskeletal antigen by competitive
elution with a monoclonal antibody. J. Mol1. Biol. 244:353-360.

92. Zhu, Y., Ho, B., and Ding, J.L. 2003. Sequence and structural diversity in endotoxin-
binding dodecapeptides. Biochina. Biophys. Acta 1611:234-242.

93. Thomas, C.J., Sharma, S., Kumar, G., Visweswariah, S.S., and Surolia, A. 2003.
Biopanning of endotoxin-specific phage displayed peptides. Biochens. Biophys. Res.
Conanun. 307:133-138.

94. Noda, K., Yamasaki, R., Hironaka, Y., and Kitagawa, A. 2001. Selection of peptides that
bind to the core oligosaccharide of R-form LPS from a phage-displayed heptapeptide
library. FEM~S2icrobiol. Lett. 205:349-354.

95. Harlow,E. and Lane,D. 1988. Antibodies a Laboratory Manual. Cold Spring Harbor
Laboratory, Cold Spring Harbor, NY. 23-34 pp.

96. Paoli, G.C., Chen, C.Y., and Brewster, J.D. 2004. Single-chain Fv antibody with
specificity for Listeria nzonocytogenes. J. Inanunol. Methods 289: 147-155.


























To my parents, siblings, and Matt. They, never doubted my abilities and always pushed me
harder when I doubted myself.










Analysis of clones by ELISA proved that the maj ority of the selected clones from panning

bound to the Stx2 toxin preparation but failed to determine if the clones bound to Stx2 toxin

because the anti-Stx2 monoclonal antibodies and comparative analysis of +/- Stx2 bacterial

preparations failed to generate positive signals in an ELISA. However, when both of the anti-

Stx2 toxin monoclonal antibodies were examined in a Western blot against SDS-PAGE resolved

Stx-2 toxin preparation, the anti-Stx2 A subunit toxin monoclonal antibody successfully

recognized the Stx2 toxin A subunit in the Stx2 toxin preparation. Therefore, analysis by

Western blot could at least determine if any of the clones recognized the A subunit of Stx2 toxin.

Six clones were screened in a Westemn blot against SDS-PAGE resolved Stx2 toxin

preparation, E. coli 0157:H7 933 (expresses Stxl and 2 toxins) supernatant, and E. coli 0157:H7

87-23 (Stx phage cured) supernatant (Fig. 3-17). Because the Stx2 toxin is composed of A and B

subunits that are approximately 32 and 10-kDa proteins, a clone that recognized Stx2 toxin was

expected to recognize a protein band at 32 kDa or 10 kDa on the Stx2 toxin preparation and the

E. coli 0157:H7 933 supernatant but not on the E. coli 0157:H7 87-23 supernatant. If a clone

recognized bacterial impurities then it would be expected to recognize proteins in all three

antigens. The anti-Stx2 A subunit monoclonal antibody recognized a 32-kDa protein in the

commercial Stx2 toxin preparation and in the E. coli 0157:H7 933 supernatant proving that Stx2

A subunit was present in both preparations and not in E. coli 0157:H7 87-23 supernatant, as was

expected. Anti-BSA was also analyzed against the three antigens as a negative control to ensure

that phagemid particles did not generate artifact bands. Anti-BSA failed to recognize any of the

antigens, as was expected. Clone 46 failed to recognize any of the three antigens. It is possible

that the binding site of the target that it recognized in the Stx2 toxin preparation was denatured

by SDS-PAGE. Clone 48 recognized 35, 45, and 55-kDa proteins in the Stx2 toxin preparation










24. Qadri, F., Hasan, J.A., Hossain, J., Chowdhury, A., Begum, Y.A., Azim, T., Loomis, L.,
Sack, R.B., and Albert, M.J. 1995. Evaluation of the monoclonal antibody-based kit
Bengal SMART for rapid detection of Vibrio cholerae 0139 synonym Bengal in stool
samples. J. Clin. M~icrobiol. 33:732-734.

25. Hasan, J.A., Bernstein, D., Huq, A., Loomis, L., Tamplin, M.L., and Colwell, R.R. 1994.
Cholera DFA: an improved direct fluorescent monoclonal antibody staining kit for rapid
detection and enumeration of Vibrio cholerae Ol. FEM~S2icrobiol. Lett. 120:143-148.

26. Hasan, J.A., Huq, A., Tamplin, M.L., Siebeling, R.J., and Colwell, R.R. 1994. A novel kit
for rapid detection of Vibrio cholerae Ol. J. Clin. M~icrobiol. 32:249-252.

27. Nato, F., Boutonnier, A., Rajerison, A., Grosjean, P., Dartevelle, S., Guenole, A.,
Bhuiyan, N.A., Sack, D.A., Nair, G.B., Fournier, J.A., and Chanteau, S. 2003. One-step
immunochromatographic dipstick tests for rapid detection of Vibrio cholerae 01 and
0139 in stool samples. Clinical and Diagnostic Laboratory Immunology 10:476-478.

28. Nataro, J.P. and Kaper, J.B. 1998. Diarrheagenic Escherichia coli. Clin. M~icrobiol. Rev.
11:142-201.

29. Hussein, H.S. 2007. Prevalence and pathogenicity of Shiga toxin-producing Escherichia
coli in beef cattle and their products. J. Anim Sci. 85:E63-E72.

30. Mead, P.S., Slutsker, L., Dietz, V., McCaig, L.F., Bresee, J.S., Shapiro, C., Griffin, P.M.,
and Tauxe, R.V. 1999. Food-related illness and death in the United States. Emerg. Infect.
Dis. 5:607-625.

31. Boyd, B. and Lingwood, C. 1989. Verotoxin receptor glycolipid in human renal tissue.
Nephron 51:207-210.

32. Robinson, L.A., Hurley, R.M., Lingwood, C., and Matsell, D.G. 1995. Escherichia coli
verotoxin binding to human paediatric glomerular mesangial cells. Ped'iatr. Nephrol.
9:700-704.

33. Takeda, T., Dohi, S., Igarashi, T., Yamanaka, T., Yoshiya, K., and Kobayashi, N. 1993.
Impairment by verotoxin of tubular function contributes to the renal damage seen in
haemolytic uraemic syndrome. 1. Infect. 27:339-341.

34. O'Brien, A.D., Tesh, V.L., Donohue-Rolfe, A., Jackson, M.P., Olsnes, S., Sandvig, K.,
Lindberg, A.A., and Keusch, G.T. 1992. Shiga toxin: biochemistry, genetics, mode of
action, and role in pathogenesis. Curr. Top. M~icrobiol. Immunol. 180:65-94.

35. Strockbine, N.A., Marques, L.R., Newland, J.W., Smith, H.W., Holmes, R.K., and
O'Brien, A.D. 1986. Two toxin-converting phages from Escherichia coli 0157:H7 strain
933 encode antigenically distinct toxins with similar biologic activities. Infect. Immun.
53:135-140.









responsible for approximately 73,000 illnesses and more than 60 deaths per year in the United

States (30).

Escherichia coli 0157:H7 produces Shiga toxins that are responsible for human disease.

These Shiga toxins produce severe cytopathic effects and have a high degree of homology with

Shiga-toxin (Stx) of \1hige//a~ dysenteriae type 1 (28). The Stx toxin is a member of the AB toxin

family. It is composed of a catalytically active A-subunit surrounded by a pentameric B-subunit.

The B subunits specifically bind to glycosphingolipid globotriosylceramide (Gb3) receptors (31)

of the renal glomerular endothelial, mesangial (32) and tubular epithelial cells (33). Upon entry

into the cell, the catalytically active Al subunit cleaves ribosomal RNA leading to the cessation

of protein synthesis and cell death (34). Not only does Stx toxin damage host cells, but it may

also increase the adherence of E coli 0157:H7 to epithelial cells leading to increased risk of

colonization. Tissue culture experiments showed that Stx toxin evoked an increase in a

eukaryotic receptor, nucleolin, that binds to the E. coli 0157 attachment factor, intimin, leading

to increased cell adherence (32).

There are two distinct toxin-converting bacteriophages (phages), 933J and 933W, in E. coli

0157:H7 that generate two genetically related toxins that are antigenically distinct but create

similar biologic effects (35). These two toxins are called Shiga toxins I and II (Stxl and 2).

Experimental studies suggest that E. coli that produces the Stx2 toxin is more virulent than

E. coli that produces Stxl or both Stxl and Stx2 toxins (36). Escherichia coli strains producing

Stx2 toxin are more frequently linked with the development of HUS than Stxl toxin-producing

strains (37,38). In mouse studies, the lethal dose of purified Stx2 toxin is 400 times lower than

that of Stxl toxin (39). In piglet studies, Stx2 toxin-producing E. coli strains caused more severe

neurologic symptoms than strains producing both Stxl and 2 toxins or only Stxl toxin (40).









nitrocellulose paper coated with y. cholerae whole cells, PBS, or 10 Cpg/mL BSA. The colony

blot was determined to be neither selective nor specific. Because the colony blot was not

functioning, a way to improve the culture tube method for the production of scFvs was

performed.

A high throughput ELISA was contrived to enable screening of hundreds of clones per

day. This high throughput procedure involved picking bacterial colonies containing a phagemid

and growing them overnight in a 96-well plate to produce scFvs (see Materials and Methods).

The next day the 96-well plate containing the turbid cultures was centrifuged to remove the

bacterial cells. These scFvs in the supernatant were examined in an ELISA for screening of their

specificity. To determine how well scFvs could be produced in a microtiter well, as opposed to a

culture tube or flask, a bacterial colony containing a phagemid that encodes an scFv specific to

BSA was grown in a microtiter well. When analyzed by ELISA, the anti-BSA (a-B SA) clone

was diluted 1:2 in casein blocking buffer and produced a S:N of 11.5 in an ELISA against BSA.

Anti-BSA scFv produced in a flask produced an average S:N value between 10 andl5; therefore,

producing scFv proteins in a microtiter well may not be optimal, but it was acceptable for the

purpose of a high throughput screen.

We were successful in producing scFvs in a microtiter well. The next test was to

determine how far the scFv can be diluted and still produce a positive signal. The a-BSA clone

was grown in a microtiter well to produce scFvs. The produced scFvs were diluted 1:2, 1:4, 1:6,

and 1:10 in casein blocking buffer and analyzed by ELISA (Fig. 3-10). Diluting the scFv

supernatant 1:10 in casein blocking buffer yielded a S:N of 5.6. This S:N was high enough for

detection of a positive clone and was not significantly different (p=0. 13) than the S:N produced

when using a-BSA phagemid particles that were diluted 1:6. Therefore, diluting scFv









Phage Display

Phage display was initially described in 1985 by George P. Smith (52) as a means to

display foreign proteins on filamentous bacteriophage. Filamentous phages are non-lytic phages

with circular ssDNA genomes. Of the filamentous phages, the Ff family (fl, fd, and M13)

phages infect F+ E. coli through binding with their F pilus. These phages are useful tools to link

genotype and phenotype of select recombinant proteins. This link was created by encoding a

foreign polypeptide in-frame with a coat protein gene of the M13 phage. Phage display could

theoretically be implemented with any phage, but filamentous phages have been the most widely

used. Of the filamentous phages, M13 is the most commonly used.

The M13 phage (Fig. 1-1) replicates in E. coli, turning the bacterium into a phage-

production factory (53). The bacteria harboring these phages do not lyse, but undergo reduced

cell growth due to the stress of phage production. The M13 phage contains a 6.4-kb, circular,

single-stranded DNA genome that encodes phage proteins I to XI. Five of these proteins are coat

proteins. The maj or coat protein (pXIII) is present in approximately 2,700 copies and protects

the genome in a cylindrical manner. The minor coat proteins pVII and plX are necessary for

efficient particle assembly, while the minor coat proteins pIII and pVI are necessary for particle

stability and infectivity (54-56). The pIII protein mediates the binding of the phage to the F pilus

and is necessary for viral uncoating and phage DNA transfer to the cytoplasm of the bacterium

(55). Host enzymes then convert the ssDNA into supercoiled dsDNA, known as the replicative

form (RF) (55). The RF is essential to the phage display system because it can be purified and

manipulated just like a plasmid. Through the manipulation of the RF of M13 came some of the

earliest cloning vectors (57). Through some of these vectors came the development of

recombinant phage libraries.









suspended in 2 mL of PBS and centrifuged at 13,776 x g for 5 minutes at room temperature to

remove bacterial debris. The supernatant was transferred to a new tube to be titered and stored at

40C until used in further rounds of panning or screening.

Panning in Suspension

Vibrio cholerae N16961 was grown to log phase, and 10' to 109 CellS were centrifuged at

10,621 x g for 5 minutes at room temperature. The bacterial pellet was suspended in 1 mL of

PBS and transferred to a microcentrifuge tube. The Tomlinson J library (for round one of

panning) or amplified eluted phages (for subsequent rounds of panning) was added to the

bacterial suspension at a concentration of 1 x 10l phages in 0.5 mL ofPBS. The library and

bacterial cells were incubated for 3 hours at 40C while rotating on a labquake. The cell

suspension was centrifuged at 10,621 x g for 5 minutes at room temperature. The supernatant,

containing the unbound phage library, was stored at 40C. The bacterial pellet containing the

bound phages was suspended in 0.5 mL of PBS to wash off phages that were weakly bound to

the bacteria. The suspension was centrifuged at 10,621 x g for 5 minutes at room temperature.

The supernatant was discarded, and the pellet was suspended in 0.5 mL of PB S for a second

wash. The suspended cells were transferred to a new microcentrifuge tube and centrifuged at

10,621 x g for 5 minutes at room temperature. The supernatant was aspirated, and phages bound

to the bacteria were eluted with 0.5 mL of trypsin-PBS for 10 to 60 minutes at room temperature

on a labquake. Ten microliters of the elution was immediately titered. The rest of the eluted

phages were infected into 5 mL of E coli TG1 at an OD600 Of 0.4. The infected E. coli were

incubated for 30 minutes in a 370C standing water bath. Following incubation, the infected cells

were isolated by centrifugation at 13,776 x g for 10 minutes at room temperature. The cells were

suspended in 0.6 mL of 2xTY medium and plated on three 2xTY AG plates to amplify the

phagemid-containing bacteria. The plates were incubated overnight at 370C. The next day 1.5









that recognized the anti-Stx2 toxin monoclonal antibodies was not recognizing the anti-Stx2

monoclonal antibodies. However, when another HRP-conjugated anti-mouse secondary

antibody was used to recognize the anti-Stx2 toxin monoclonal antibodies it also failed to

produce a positive signal in an ELISA. Both anti-mouse antibodies had been tested within three

days of these experiments against murine monoclonal antibodies and produced a positive signal.

However, positive controls for the anti-mouse antibodies were not used in the ELISA that

analyzed the two anti-mouse antibodies against Stx2 toxin preparation to determine whether or

not they were functional. It was not determined why the monoclonal antibodies did not

recognize the Stx2 toxin preparation in an ELISA. One reason might be that there was not

enough Stx2 toxin with a correctly folded binding site in the Stx2 toxin preparation for the

monoclonal antibodies to recognize. Also, it is possible that the batches of monoclonal

antibodies that we received were somehow defective.

Because the anti-Stx2 toxin monoclonal antibodies could not be used to determine the

binding specificity of the selected panning clones in an ELISA, other means of determining the

binding specificities of the clones were analyzed. Three clones that bound to Stx2 toxin

preparation were screened against Stx2 toxin, E. coli 0157:H7 933 (expresses Stx 1 and 2

toxins) periplasmic break fraction, E. coli 0157:H7 933 supernatant, E. coli DH~a (pNR100)

(expresses Stx2 toxoid) periplasmic break fraction, and E. coli DH~a periplasmic break fraction.

A clone that bound to the E. coli bacterial impurities in the Stx2 toxin preparation would be

expected to bind to all of the antigens, but a clone that bound to Stx2 toxin would be expected to

bind to all of the antigens except the E. coli DH~a periplasmic break fraction. However, the

three selected clones only recognized the Stx2 toxin preparation and none of the other antigens;

therefore, it was not determined if the clones were actually binding to Stx2 toxin.









that E. coli TG1 (Hye-sm) and E. coli TG1 (Hye-lg) had infection efficiencies that were too

low for them to be effective amplification tools to amplify the Tomlinson J scFv library.

However, these strains could be used as tools to amplify single phagemid particles with

acceptable phagemid particle production and quality.

Helper Plasmids Were Not an Optimal Phagemid Particle Amplification Tool

Helper plasmids (Los Alamos National Laboratory, Los Alamos, NM) (64) are M13-based

plasmids that are maintained in E. coli to provide phagemid-containing E. coli with all of the

genes necessary to produce phagemid particles. The helper plasmids come in three forms that

either have a full-length (M13cp), a deleted (M13cp-dg3), or a truncated (M13cp-CT) giIl gene.

The helper plasmid-containing E. coli strains were analyzed for their infection efficiencies with

phagemid particles, production yield of phagemid particles, and the quality of phagemid particles

produced.

The three helper plasmids were electroporated (see Materials and Methods) into

electrocompetent E. coli TGl. All three transformations yielded colonies between 0.5-2 mm in

diameter. A small (0.5 mm) and a large (2 mm) colony from E. coli TG1 (M13-cp), E. coli TG1

(M13cp-CT), and E. coli TG1 (M13cp-dg3) were single colony passage twice. Escherichia coli

TG1 (M13cp-sm) was the only strain that maintained a consistent phenotype of 0.5 mm colonies.

The other 5 strains yielded colonies with diameters of 0.5-2 mm after each passage. To

determine if the different colony size phenotypes had an effect on the functionality of the helper

plasmids, a small and large isolate from each strain were analyzed for their infection efficiencies

with phagemid particles.

To analyze the infection efficiencies of phagemid particles into helper plasmid-containing

E. coli TGl, an infection efficiency (see Materials and Methods) experiment was performed that

compared the infection efficiencies of helper plasmid-containing E. coli with E. coli that did not









dilution of 1:400 (p=0.01) (Fig. 3-4B). Therefore, the saturation limit of the a-Vc 01 LPS pAb

was reached at a dilution of 1:400.

Because a microtiter well coated with 1 Cpg/mL LPS was never saturated by the a-Vc 01

LPS mAb, the a-Vc 01 LPS pAb was used to characterize the saturation limits of nitrocellulose

paper. Pieces of nitrocellulose paper were cut to dimensions (5 mm x 20 mm) that were

approximately equivalent to the surface area of a microtiter well. The same phenol-water-

extracted Y. cholerae LPS used for the microtiter well saturation experiment was also used to

characterize the saturation of nitrocellulose with LPS. Serially diluted LPS from 100-1 Cpg/mL

was used to saturate strips of nitrocellulose paper. These strips of LPS-coated nitrocellulose

paper were examined in an altered Western blot protocol (see Materials and Methods for LPS

Saturation to Nitrocellulose Paper) using the a-Vc 01 LPS pAb at a dilution of 1:400 (Fig. 3-5).

The nitrocellulose paper was not saturated with LPS until 100 Cpg/mL LPS (250 Cpg) was used to

coat the nitrocellulose paper. Therefore, nitrocellulose paper can bind approximately 250 to

2,500 times more LPS than a microtiter well.

Panning to V. cholerae LPS Failed to Yield Phages that Were Specific to V. cholerae LPS

Vibrio cholerae 569B LPS (Sigma-Aldrich) was immobilized onto a polystyrene

Maxisorp microtiter well and panned against the NEB Ph.D. 12mer phage display library (see

Materials and Methods). Five rounds of panning were performed with amplification of phages

after the first four rounds of panning. Each round of panning promotes selection of phages that

specifically bind to the target molecule. Amplifying the selected phages should increase the

percentage of specific phages present in the library to pan with. Panning and amplifying for five

rounds should greatly improve the chances of isolating phages that specifically bind to the target

molecule.









BIOGRAPHICAL SKETCH

Crystal Harpley (Mazur) was born in Columbus, OH; and shortly after moved to

Wellington, FL, where she completed grade school. In 2006 she graduated with high honors

from the University of Florida, with a major in microbiology and cell sciences and a minor in

chemistry. During the process of finishing her master' s thesis, Crystal Harpley got married and

is now Crystal Mazur.









strains in stool specimens. SMART is 95% sensitive and 100% specific to K. cholerae 01

strains (26) and 100% sensitive and 97% specific to 0139 strains (24). A one-step

immunochromatographic dipstick test for the detection of y. cholerae 01 and 0139 LPS in stool

samples was invented by the Institute Pasteur in Paris. This assay requires minimal technical

skill and rapidly detects thresholds of purified LPS at 10 ng/mL for y. cholerae 01 and at 50

ng/mL for y. cholerae 0139 strains in approximately 10 minutes (27). With continuing

improvement and use of methods for detection of y. cholerae, outbreaks could be lessened or

prevented with proper treatment and containment.

Escherichia coli 0157:H7

Escherichia coli 0157:H7 is an enterohemorrhagic serotype of E coli that causes

hemorrhagic colitis (HC) and hemolytic uremic syndrome (HUS). The bacterium is primarily

found on cattle farms and colonizes cattle, swine, and deer intestines with subclinical effects to

the animals (28). Disease in humans comes from the ingestion of beef, milk, vegetables, and

other products that are contaminated with E. coli 0157:H7. Symptoms from an E. coli 0157:H7

infection include mild diarrhea, abdominal pain, vomiting, bloody diarrhea, HC, strokes, and

HUS (28).

Over the past 23 years, 146 Shiga-like toxin producing E. coli (STEC) outbreaks and

sporadic cases of human illnesses have been traced to consumption of beef contaminated with

various E. coli 0157 strains (29). Most of these illnesses were caused by infection with E. coli

0157:H7. Of the 146 outbreaks and sporadic cases 89% occurred in the United States (29). The

large number of cases in the Unites States compared with the rest of the world can be explained

by high levels of beef consumption and availability of E. coli 1 57 diagnostic methods in the

United States. The Center for Disease Control and Prevention estimated that E. coli 0157:H7 is









clones (81%) having a S:N from 10 to 23 against the Stx2 toxin preparation. Therefore, the

panning was successful in isolating clones that bound to Stx2 toxin preparation. Not all of the

clones that were selected bound to Stx2 toxin preparation; 4/59 (7%) of the clones bound to PBS-

treated wells as well as Stx2-coated wells. The four clones had a signal in an ELISA between

0.31 and 0.87 against PBS-treated wells. These clones most likely bound to polystyrene. Even

though the immunotube was blocked with blocking buffer, the unintentional isolation of clones

that bound to the panning vessel is common (84). Of the clones screened, 2/59 (3%) did not bind

to Stx2 toxin preparation or a microtiter well. These clones have unknown binding specificity.

It is common to select clones that have unknown binding specificity. The fact that only

approximately 3% of my clones had unknown binding specificity proves that the panning was

successful in isolating specific recombinant phagemid particles.

Most of the clones selected from panning recognized the Stx2 toxin preparation; however,

the Stx2 toxin preparation was only 50% pure for Stx2 toxin. Therefore, it was unknown if the

clones bound to Stx2 toxin or bacterial impurities in the Stx2 toxin preparation. To determine if

the selected clones bound to Stx2 toxin, two anti-Stx2 toxin monoclonal antibodies were

examined. One antibody recognized the A-subunit of Stx2 toxin, while the other antibody

recognized the B-subunit of Stx2 toxin. However, these monoclonal antibodies never generated

a positive signal in an ELISA against the Stx2 toxin preparation. Using the monoclonal

antibodies at the manufacturer recommended concentration of 1 Cpg/mL failed to generate a

positive signal in an ELISA. It was initially reasoned that the monoclonal antibodies may need

to be used at a higher concentration to increase the S:N values. Increasing the concentration of

monoclonal antibodies in an ELISA to 100 Cpg/mL still failed to produce a positive signal in an

ELISA. Therefore, it was hypothesized that the HRP-conjugated anti-mouse secondary antibody









Methods) was performed (Fig. 3-9). Of the different panning variations, the amplified phagemid

particles from the panning with the library that contained 1% Vc86 phagemid particles, 1 x 10s

whole cells y. cholerae, and an elution time of 30 minutes generated a significantly higher S:N

(p=0.01-0.04) in an ELISA compared to the amplified phagemid particles from the other panning

variations. This suggested that panning with 10s whole cells with a trypsin elution time of 30

minutes was optimal for selecting phagemid particles that were specific to the panned whole

cells compared to using 109 whole cells or a trypsin elution of 10 minutes.

Screening of scFv Proteins by a High Throughput ELISA Was Acceptable

Once the panning procedure is complete, screening of phages or scFvs is performed. The

previous method of screening scFvs involved inducing phagemid-containing bacteria to produce

scFvs in culture tubes or flasks and screening the scFvs by ELISA. However, this method was

very time consuming. Optimizing the screening process to enable high throughput screening of

clones should increase the chances of finding clones that specifically bind to the target molecule.

A colony blot was attempted to try to enable high throughput screening of scFv proteins. The

colony blot (see Materials and Methods) involved the transfer of bacterial colonies that were

grown to secrete scFvs onto a piece of nitrocellulose paper that was coated with a target antigen

and blocked with casein blocking buffer. Following the transfer, the piece of nitrocellulose

paper was washed with TBS to remove unbound scFv proteins, incubated with Protein-L

peroxidase (Sigma-Aldrich) to bind to the kappa light chains of the scFv, washed with TBS to

remove excess Protein-L peroxidase, and developed with ECL (Pierce). Upon development with

ECL, scFv-producing clones specific to BSA failed to produce a positive signal against a piece

of nitrocellulose paper coated in 10 Cpg/mL BSA but did produce a positive signal against pieces

of nitrocellulose paper coated in PBS or y. cholerae whole cells. Also, clones that were not

specific to y. cholerae whole cells in an ELISA produced positive signals against pieces of









PAGE and stained with Coomassie blue for visualization of proteins (Fig. 3-1A) and with Tsai-

Frasch silver stain for visualization of carbohydrates (see Materials and Methods) (Fig. 3-1B).

The Coomassie blue stain of the TRlzol Reagent-extracted LPS, which was diluted 1:2 in

Laemmli sample buffer, showed no detectable proteins. The limit of detection for proteins with

Coomassie blue is 0.3-1 Cpg of protein/band (79); therefore, there was less than 0.3 Cpg of any

single protein loaded onto the gel that was stained by Coomassie blue. The Tsai-Frasch silver

stain for the TRlzol Reagent LPS extraction showed non-staining bands between 20 and 30 kDa.

The Tsai-Frasch silver stain primarily stains carbohydrates and poorly stains proteins and lipids

(78); therefore, the non-staining bands may be lipids or proteins. However, because the

Coomassie blue stain for the TRlzol-extracted LPS showed no detectable proteins and due to the

thickness of the non-staining bands, which correlates to concentration, then the non-staining

bands were most likely due to lipid contamination. The LPS banding pattern of the TRlzol

reagent LPS extraction in the silver stain was very similar to the commercial K. cholerae 569B

LPS (Sigma-Aldrich). On the silver stain, both LPS preparations had bands at approximately 14

and 20 kDa and no bands larger than 50 kDa. Previously published silver stains of y. cholerae

01 LPS had bands at ~10 and 14 kDa (which was stated to be the lipid A-core of LPS), and ~20-

50 kDa (which was stated to be the lipid A-core plus repeating O-antigens) (80,81).

To determine the concentration of the TRlzol-Reagent-extracted LPS, the intensities of

the bands were compared to those of the commercial LPS in the silver stained gel (Fig. 3-1B).

The intensity of the band of the commercial LPS (1 mg/mL) that was diluted 1:5 was

approximately 1.5 times stronger than the intensity of the band of the TRIzol-Reagent-extracted

LPS that was diluted 1:2. Therefore, we estimated the concentration of the TRlzol-Reagent-












Specific Aim 4: Isolation of Specific Recombinant Phage to Stx2 Toxin ofE. coli
0157: H7 ........._.___..... ._ __ ...............146..


5 EPILOGUE ........._.___..... ._ __ ...............153....


LIST OF REFERENCES ........._.___......___ ...............154....


BIOGRAPHICAL SKETCH ........._.___..... .__. ...............163....









produced in a flask (S:N~-4) yielded a significantly higher S:N (p=0.05) compared to Vc86

phagemid particles produced in microtiter wells. When a-BSA phagemid particles were

produced at 300C, the S:N was significantly higher (p<0.04) than when a-BSA phagemid

particles were produced at 370C at either glucose concentration. The a-BSA phagemid particles

produced in a flask (S:N~-12) yielded a significantly higher (p=0.02) S:N compared to a-BSA

phagemid particles produced in microtiter wells (S:N~7).

The experiment was repeated a second time with only the a-B SA phagemid particles

(Fig. 3-12). Anti-BSA phagemid particles produced in microtiter wells yielded significantly

higher signals (p<0.03) when produced at 300C than at 370C at either glucose concentration.

There was not a significant difference (p=0.34) in the signals produced by phagemid particles

that were produced with or without glucose at 300C. When phagemid particles were produced in

microtiter wells at 370C, there was a significantly higher (p=0.02) signal when phagemids were

produced with no glucose compared to 0.1% (w/v) glucose. There was not a significant

difference in the signal (p=0.09) produced by phagemid particles produced in microtiter wells

compared to phagemid particles produced in a flask.

When phagemid particles were diluted 1:2, phagemid particles yielded significantly higher

signals in an ELISA when produced at 300C compared to 370C. Whether the growth medium

included 0.1% (w/v) glucose or not when phagemid particles were incubated at 300C did not

significantly affect the signal of the phagemid particles; therefore, 0. 1% (w/v) glucose was not

enough to cause catabolite repression of the lac promoter. Also, phages produced in flasks

yielded higher signals than phages produced in microtiter wells. To determine if low titers of

phagemid particles produced in a microtiter well were a factor in the lowered ELISA signals for

phagemid particles, the phagemid particles were titered. When a-BSA and Vc86 phagemid









this strain was used for further analysis to determine the quantity and the quality of phagemid

particles produced. However, before E. coli TG1 (M13cp-dg3-sm) was analyzed further,

M13cp-CT-sm was examined.

To test if the helper plasmid (M13-CT-sm) could yield a higher infection efficiency in

another E. coli strain, it was extracted (see Materials and Methods) from E. coli TG1 (M13cp-

CT-sm) and electroporated (see Materials and Methods) into E. coli JM109. Escherichia coli

JM109 was chosen because it was genotypically more similar than E. coli TG1 to the bacterial

strain that Los Alamos National Laboratory used with the helper plasmids (E. coli DH~a F').

Escherichia coli JM109 and E. coli DH~a F' contain relA l and recA1i, while E. coli TG1 does

not.

Escherichia coli TG1 (M13cp-CT-sm) and E. coli JM109 (M13cp-CT-sm) were

transduced with clone 18, a-AV20N3, and Vc86 phagemid particles and were analyzed for their

infection efficiencies compared to E. coli TG1 and E. coli JM109 (Table 3-3). The average

infection efficiency of E coli TG1 (M13cp-CT-sm) was 47%, while the average infection

efficiency of E coli JM109 (M13cp-CT-sm) was 28%. Therefore, using E. coli JM109 to

maintain M13cp-CT-sm did not increase the infection efficiency.

In analyzing the infection efficiencies, E coli TG1 (M13cp-dg3-sm) had an average

infection efficiency of 120% with a standard deviation of 33% and E. coli TG1 (M13cp-CT-sm)

had an average infection efficiency of 30% with a standard deviation of 16%. Escherichia coli

TG1 (M13cp-dg3-sm) had a significantly higher infection efficiency (p=0.01) than E. coli TG1

(M13cp-CT-sm).

Escherichia coli TG1 (M13cp-dg3-sm), E. coli TG1 (M13cp-CT-sm), and E. coli TG1 that

was superinfected with hyperphage were analyzed for their ability to produce phagemid









between 15 and 30 minutes. The phagemid particle titer decreased by approximately 80-93%

from 30 to 45 minutes for the four pannings. More phagemid particles were eluted when panned

against 1 x 109 whole cells. However, there were not enough data to do statistical analysis to

determine if using 1 x 109 whole cells eluted significantly more phagemid particles than using 1

x 10s whole cells in the biopanning process.

Because the best whole cell concentration and trypsin elution time was not definitively

obtained in the first experiment, a second one-round panning was done to analyze these factors.

The purpose of the experiment was to compare the differences between using 1 x 10 1 x 10s,

and 1 x 109 whole cells of y. cholerae N16961 and trypsin elution times of 10, 20, and 30

minutes in the biopanning process. The same protocol was used as the first one-round panning

protocol above, except an additional concentration of whole cells was added, so there were six

parallel pannings instead of four. When the trypsin was added to the bound-phagemid particles,

an aliquot of the eluted phagemid particles was removed at 10, 20, and 30 minutes after the start

of the elution and was titered immediately (Fig. 3-7). There were no significant differences in

the titers of eluted phagemid particles when comparing the concentrations of whole cells and

trypsin elution times used in the second one-round panning experiment. The panning was

expected to yield distinct differences between eluted titers of phagemid particles from the

libraries with and without 1% Vc86 phagemid particles. A certain parameter was supposed to be

optimal and yield higher elution titers from the library with 1% Vc86 phagemid particles.

Because there were not any significant differences in the titers of the eluted phagemid particles at

the different parameters an optimal panning condition could not be determined. To determine

which panning parameter elutes the most phagemid particles that recognize y. cholerae whole









particles produced, and the quality of phagemid particles produced. Hyperphage-containing

E. coli resulted in unacceptable infection efficiencies and helper plasmid-containing E. coli

resulted in unacceptable phagemid particle production titers. However, improvement of

hyperphage that was produced in our laboratory was performed by use of E coli MGl655

(pGTR203) which generated high titer and high quality hyperphage. This high quality

hyperphage was used to successfully amplify the Tomlinson J scFv phagemid library to ensure

that the redundancy of phagemid clones was maintained. This library was then panned against

an Stx2 toxin preparation and resulted in 89% of phagemid particles from the third round of

panning recognizing the Stx2 toxin preparation in an ELISA. The efficiency of this panning was

greatly improved from previous pannings from this laboratory, which usually resulted in no

better than 1% of selected phagemids being specific to the target antigen, and often no isolation

of usable specific clones. Therefore, the optimization of methods and reagents in this thesis will

greatly improve our laboratory and other laboratories chances of selecting recombinant phages

specific to a target molecule.









Escherichia coli TG1 Harboring the Hyperphage Genome Was Not an Optimal Phagemid
Particle Amplification Tool

To overcome the low titers of hyperphage, the hyperphage genome was incorporated into

E. coli and maintained in the strain as a plasmid to ensure that every phagemid-containing E. coli

also contained the hyperphage genome to enable phagemid particle production. Log phase

(OD600 ~0.4) E. coli TG1 was transduced with hyperphage at a MOI of 5, incubated in a 370p

standing water bath for 30 minutes, serially diluted in BSG, and then plated on 2xTY plates

containing 40 Cpg/mL kanamycin. The plates were incubated overnight at 370C. The next day

the plates contained small (0.5 mm diameter) and large (1-1.5 mm diameter) colonies. Two

small and large colonies were single colony passage twice. The small colonies maintained a

small phenotype (0.5 mm diameter), and the large colonies maintained a large phenotype (1-1.5

mm diameter). The two strains were named E. coli TG1 (Hye-sm) and E. coli (Hye-lg).

The infection efficiencies of phagemid particles into E. coli TGl, E. coli TG1 (Hye-sm),

and E. coli (Hye-lg) were analyzed. The three E. coli strains were grown to log phase and

transduced with serially diluted a-B SA and Vc86 phagemid particles in B SG at a MOI of 0.001.

Transductions were incubated in a 370C standing water bath for 30 minutes and plated on 2xTY

AG plates containing 40 Cpg/mL kanamycin for transduced strains ofE. coli TG1 containing

hyperphage and on 2xTY AG plates for transduced E. coli TGl. The numbers of E coli cells

transduced with a-B SA phagemid particles were as follows: 1,200 transductions into E. coli

TGl, 340 transductions into E. coli TG1 (Hye-sm), and 100 transductions into E. coli TG1

(Hye-lg). When the number of a-BSA phagemid particles transduced into E. coli TG1 was set

at an infection efficiency of 100%, the number of a-B SA phagemid particles transduced into

E. coli TG1 (Hye-sm) was 28% and the number of a-B SA phagemid particles transduced into









blot was incubated in 8 mL of ECL substrate (Pierce) for three minutes at room temperature and

developed on CL-XPosure Film (Pierce) by an X-omat 2000 processor (Kodak, Rochester, NY).

Lipopolysaccharide Saturation to Nitrocellulose Paper

The LPS saturation experiment to nitrocellulose paper was performed to determine the

concentration of y. cholerae 569B LPS needed to saturate a piece of nitrocellulose paper.

Phenol-water-extracted y. cholerae 569B LPS was serially diluted from 100 to 1 Cpg/mL in PBS.

Twelve pieces of nitrocellulose paper were cut to dimensions of 5 mm x 20 mm. Nitrocellulose

strips were put in a microcentrifuge tube containing 256 CLL of the LPS dilutions or PB S for a

control. LPS and PBS coated strips were done in duplicates in separate tubes. The strips were

coated overnight at 40C on a labquake. Coated nitrocellulose strips were transferred to new

microcentrifuge tubes. Strips were washed 3 times with 1 mL of PBS on a labquake with 5

minutes per wash. Strips were blocked with 1 mL of casein blocking buffer (Pierce) for one hour

at room temperature on a labquake. Strips were incubated with 0.5 mL of rabbit anti-Y. cholerae

01 LPS polyclonal antibody (Accurate Chemical, Westbury, NY) at a dilution of 1:400 in casein

blocking buffer (Pierce) for one hour at room temperature on a labquake. Strips were washed 3

times with 1 mL of PBS on a labquake with 5 minutes per wash. Strips were incubated with 0.5

mL of goat anti-rabbit peroxidase conjugated monoclonal antibody (Jackson ImmunoResearch

Laboratories, West Grove, PA) at a concentration of 1:1000 in casein blocking buffer (Pierce) for

30 minutes at room temperature on a labquake. Strips were washed 3 times with 1 mL of PBS

on a labquake with 5 minutes per wash. Strips were incubated in 5 mL of ECL substrate (Pierce)

for three minutes at room temperature and developed on CL-XPosure Film (Pierce) by an X-

omat 2000 processor (Kodak).












TABLE OF CONTENTS


page

ACKNOWLEDGMENTS .............. ...............4.....


LIST OF TABLES ........._._ ...... .__ ...............8....


LIST OF FIGURES .............. ...............9.....


LI ST OF AB BREVIAT IONS ........._._ ...... .... ............... 11..


AB S TRAC T ............._. .......... ..............._ 13...


CHAPTER


1 INTRODUCTION ................. ...............15.......... ......


Methods for Bacteriological Detection and Analysis ................. ..............................15
Vibrio cholerae ................ ...............18...
Escherichia coli 0157:H7 .............. ...............20....
Phage Display ............... ...............23...
Recombinant Phage Libraries......................... ...................2
Tomlinson I + J Human Synthetic VH + VL Phagemid Libraries ................ ................ .25
New England Biolabs 12mer Peptide Phage Library ............_......__ ..............26
Tools for Production of Phages from Phagemids ................. ...............27........... ..
H elper Phage .......................... ..............2
Helper Phage with giII Mutations ....._____ .....__......._ ......__ ...........28
Helper Plasmids............... ...............29

2 MATERIALS AND METHODS .............. ...............35....


Bacterial Strains, Media, and Growth Methods............... ...............35
Biopanning of Phage Display Libraries. ............ .....__ ...............36.
Panning on Immunotubes ................. ...............36..............
Panning in Suspension............... ..............3
Panning on Microtiter Wells .............. ...............39....
Panning on Nitrocellulose Paper ................ ....__ ...............41. ....
Spot Titer of Phages ................. ...............43....... .....
Spread Titer of Phages ............ ............ ...............43...
Am plification of Phages ............................ ..__ ... ... ....... ....... .............4
High Throughput Production of Soluble Antibody Fragments (scFv antibodies) ..........44
Deoxyribonucleic Acid Manipulations ........._.. ...._._........._._. ............ 4
Plasmid Extractions ........._..... ...._... ...............44.....

Ag arose Gel El ectrophoresis .............. ...............44....
Electroporation of Plasmids .............. ...............45....
Electrocompetent Cells............... ... ........................4
Enzyme-Linked ImmunoSorbent Assays (ELISAs) ...._.. ................. .....................46









selection for the E. coli cells then bacteria containing a phage from the library, which encodes

lacZa, will have a complemented lacZ gene to enable distinction from bacteria containing a

phage from the environment. The NEB 12mer library consists of approximately 2.7 x 109

electroporated sequences that were amplified once to yield approximately 55 copies of each

sequence. Sequencing of 104 clones from the library yielded six clones (5.8%) that did not

contain a displayed peptide insert. Sequencing from the 98 other clones revealed a wide

diversity of sequences with no obvious positional biases.

Tools for Production of Phages from Phagemids

To propagate phagemid particles, an amplification tool must be supplied that encodes the

rest of the M13 genes necessary for phagemid particle production. These tools may be phages or

plasmids and are referred to as "helpers" because they help the phagemid particles propagate by

supplying the necessary phage genes in trans. Some often used phage amplification tools include

helper phage, hyperphage, phaberge phage, ex-phage, and helper plasmids.

Helper Phage

Helper phage is the most common helper tool used to produce phagemid particles. There

are many variations of helper phages, R408, VCSM13, and M13KO7 (63) that differ slightly. Of

the various helper phages M13KO7 is the most commonly used. It is a derivative of M13 that

has a couple differences including a kanamycin resistance gene and the Pl5A origin of

replication, which allows the genome to be replicated as a plasmid in E. coli.

Because helper phages are basically M13 phages, they supply all of the genes necessary for

production of phagemid particles. Just like M13, helper phages infect F E. coli through the

binding of the pIII protein to the F pilus. Helper phage amplifies phagemid particles to yield

average titers of 2 x 1010-12 phagemids/mL (64). However, because helper phage encodes a wild

type gHI gene, the phage particles produced will contain a mixture of wild-type pIII proteins and









coulombic interactions, and hydrophobic bonds (95). The strength of the antibody-antigen

complex depends on the strength of the bonds used to hold it together. Protein-protein

interactions make use of all noncovalent bonds, but carbohydrate-protein interactions primarily

make use of hydrogen bonds and rarely make use of hydrophobic or coulombic interactions.

Also, protein-protein interactions can be held together by numerous bonds due to the

conformational structure of the protein-protein interaction. The carbohydrate-protein interaction

is usually only bound by a single domain due to the linearity of the carbohydrate structure.

Therefore, the low density and the limited variety of bonds that occur in carbohydrate-protein

interactions make LPS a difficult target to select recombinant proteins that bind to it with high

affinity.

Specific Aim 2: Improve Panning and Screening Process of Biopanning

Biopanning is performed to select phagemid particles that specifically bind to a target. To

increase the chances of selecting phagemid particles that are specific to a target, the optimization

of the biopanning and screening process was performed. Optimization of the panning process

was performed with y. cholerae 01 whole cells as a target, and the concentration of whole cells

and the trypsin elution time were analyzed. The concentration of whole cells used in biopanning

usually ranges from 10' to 109 whole cells (96). The goal was to use enough whole cells to

enable capture of specific recombinant phagemid particles and keep the capture of phagemid

particles not specific to the whole cells to a minimum. The trypsin elution time was also

analyzed. The Tomlinson libraries have a trypsin cleavage site between the pIII protein and the

scFv particle. Therefore, trypsin is used to cleave the phagemid particle from the scFv peptide

that is bound to the target. The Tomlinson protocol recommends a trypsin elution of 10 minutes.

The trypsin elution time was analyzed to maximize the number of eluted phagemid particles that

were specific to the target. Eluting for too short of a time may fail to elute specific phagemid









supernatants as far as 1:10 with other clones can enable detection of positive clones. However,

the a-BSA clone was the strongest clone that we had; therefore, weaker clones might fail to be

detected if analyzed at a dilution of 1:10. For the purpose of a high throughput ELISA screen for

scFvs I would recommend diluting the scFv proteins only 1:4 to 1:6 to lessen the chances of not

detecting weaker binding scFvs.

Screening of Phagemid Particles by a High Throughput ELISA Was Not Optimal

Once the panning procedure is complete, screening of phagemid particles or scFvs is

performed. Previous methods of screening phagemid particles involved the amplification of

eluted phagemid particles in culture tubes or flasks in 2xTY containing 100 Cpg/mL ampicillin, 40

Cpg/mL kanamycin, and 0.1% (w/v) glucose at 300C overnight, followed by screening by ELISA.

This process was time-consuming; therefore, it was tested if phagemid particle production could

be performed in a high throughput manor as with scFvs. Vc86 and a-BSA phagemids in E. coli

TG1 (Hye-sm) (E. coli TG1 that was previously infected with hyperphage) were produced in a

96-well plate in 2xTYcontaining 100 Cpg/mL ampicillin, 40 Cpg/mL kanamycin, and with or

without 0. 1% (w/v) glucose at 300C and 370C overnight. Protocols recommend producing

phagemid particles in medium that has 0.1% (w/v) glucose, but due to catabolite repression of

the lac promoter, which drives transcription of the giII gene, phagemid particle production might

increase without the presence of glucose in the growth medium. The next day, the bacterial cells

were removed by centrifugation, and the phage-containing supernatants were examined in an

ELISA at a dilution of 1:2 in casein blocking buffer (Fig. 3-11). There was not a significant

difference in the S:N (p=0.07-0.48) of phagemid particles produced in microtiter wells with and

without 0. 1% (w/v) glucose. Vc86 phagemid particles produced in microtiter wells with or

without glucose at 300C and 370C did not produce a S:N higher than one. It was also noted that

the cultures in the microtiter plates were only slightly turbid. The Vc86 phagemid particles









phagemid particle yields were as follows: E. coli TG1 that was superinfected with hyperphage

yielded 3 x 10 tu/mL, E. coli TG1 (Hye-sm) yielded 1 x 109 tu/mL, and E. coli TG1 (Hye-lg)

yielded 3 x 10s tu/mL. These results showed that E. coli TG1 that contained hyperphage

produced high titers of phagemid particles.

To determine the quality of a-B SA phagemid particles produced by the three E. coli

strains, the produced a-B SA phagemid particles were analyzed in an ELISA. Microtiter plates

were coated with 10 Cpg/mL BSA or only PBS. Standard ELISA procedures were followed (see

Materials and Methods) to analyze 1 x 10s tu/mL of a-B SA phagemid particles as the primary

antibody (Fig. 3-13). Anti-BSA phagemid particles produced a significantly higher S:N (p=0.01)

when produced from E. coli TG1 (Hye-sm) compared to E. coli TG1 (Hye-lg). There was not a

significant difference in the S:N (p=0.50) from a-BSA phagemid particles produced from

superinfected E. coli TG1 and E. coli TG1 (Hye-lg). Because superinfected E. coli TG1 and

E. coli TG1 (Hye-lg) yielded titers (~3 x 10s tu/mL) of a-B SA phagemid particles that were

approximately a third of that of E coli TG1 (Hye-sm) titers (~1 x 109 tu/mL), E. coli TG1

(Hye-sm)-produced a-BSA phagemid particles were analyzed in the ELISA with a third less

phage-containing supernatant volume. The amplified a-BSA phagemid particles were from the

same clone; therefore, the differences in S:N from the produced a-BSA phagemid particles was

not expected. A possible reason for the S:N differences was that a different amount of

particles/mL was used in the ELISA. The ELISA compared equal tu/mL titers of the a-B SA

phagemid particles produced. However, tu/mL is a measure of infectious particles and not total

particles. Because an ELISA measures particles with binding activity, it is possible that

superinfected E. coli TG1 and E. coli TG1 (Hye-lg) yielded similar particles/mL yields as

E. coli TG1 (Hye-sm), which would account for the differences in the S:N. The results showed









The anti-Stx monoclonal antibodies (a-Stx2 subunit A and a-Stx2 subunit B) were further

analyzed to determine the reason why they did not react in the ELISA. Microtiter wells coated

with 10 Cpg/mL Stx2 toxin preparation or carbonate bicarbonate buffer were incubated with a-

Stx2 subunit A and a-Stx2 subunit B monoclonal antibodies at concentrations of 1, 10, or 100

Cpg/mL. Bound antibodies were detected with HRP-conjugated anti-mouse antibody (Cappel) or

HRP-conjugated anti-mouse antibody (Sigma-Aldrich). Standard ELISA procedures followed

(see Materials and Methods). All parameters generated S:N values less than two. The Stx2 toxin

preparation appeared to be binding to the well because the amplified panning clones were

binding to the Stx2 with high S:N values in microtiter wells. The concentrations of anti-Stx2

monoclonal antibodies were not an issue because using the monoclonal antibodies at a

concentration as high as 100 Cpg/mL did not improve the signal of the monoclonal antibodies in

an ELISA. Two different HRP-conjugated anti-mouse antibodies were used to determine if the

secondary antibody was not recognizing the primary antibodies. Neither HRP-conjugated anti-

mouse antibody improved the signal of the monoclonal antibodies in an ELISA. Positive

controls for the anti-mouse antibodies were not performed in the ELISA to determine whether or

not the anti-mouse antibodies were functional. However, the two HRP-conjugated anti-mouse

antibodies were used within three days of these experiments and generated positive signals in an

ELISA against murine antibodies. The reason why the anti-Stx monoclonal antibodies gave low

signals with Stx2 toxin preparation in an ELISA was not determined.

Analysis by ELISA shows that the maj ority of clones isolated from the Stx2 panning

bound to the Stx2 toxin preparation in an ELISA. Because the Stx2 toxin preparation was only

50% pure for Stx2 toxin, the panning clones could have been binding to Stx2 toxin or the 50% of

impurities in the preparation. To analyze if the Stx2 toxin preparation panning clones recognized









ACKNOWLEDGMENTS

I would like to thank my mentor, Paul A. Gulig, for teaching, guiding, and pushing me to

be a better scientist. I would also like to thank my co-workers in the Gulig lab for their

encouragement, help, and making me smile throughout the work day. Lastly, I would like to

thank my committee members, Shouguang Jin and Anita Wright, for their help and guidance.









A 12


10




2 6








0.0 0.5 1.0 1.5

a-Vc 01 LPS mAb (1 pg/mL)


Figure 3-4. Analysis of the saturation limits of primary antibodies to K. cholerae 569B LPS-
coated microtiter wells by ELISA. Primary antibodies: A) mouse anti-Y. cholerae
01 LPS monoclonal antibody (a-Vc 01 LPS mAb) (0.04-1.3 Cpg/mL) and B) rabbit
anti-K cholerae 01 LPS polyclonal antibody (a-Vc 01 LPS pAb) (1:100-1:1,600)
were reacted with K. cholerae 569B LPS (1 Cpg/mL)-coated microtiter wells. Bound
primary antibodies were detected with either HRP-conjugated goat anti-rabbit
antibody or HRP-conjugated goat anti-mouse antibody. Wells were developed with
TIVB substrate and their absorbances read at 630 nm. The S:N were calculated for
every primary antibody concentration or dilution (n=3 wells). In (A) comparisons of
S:N between every monoclonal antibody concentration used yielded significant
differences (p<0.03) in S:N; therefore, the saturation limit of the monoclonal antibody
could not be determined. In (B) a significant decrease (*p=0.01) in the S:N was not
reached until the polyclonal antibody dilution decreased from 1:400 (ratio dilution of
0.003) to 1:800(ratio dilution of 0.001); therefore, the saturation limit of the
polyclonal antibody was reached at a dilution of 1:400.










3-16 Analysis of anti-Stx2 monoclonal antibodies and clones 46, 48, and 49 by ELISA.......116

3-17 Western blot analysis of phagemid clones from panning on Stx2 toxin preparation. .....1 17

3-18 Analy si s of Stx2 toxin preparati on (Toxin Technol ogi es) by SD S-PAGE ....................11 8









Table 2-1. Bacterial strains al
Strain
E. coli DH~a




E. coli DH~a (pNR100)




E. coli EC100D





E. coli EC100D (pGTR203)
(Hye)





E. coli ER2738




E. coli HB 101 (pMJ100)



E. coli HB2151

E. coli JM109




E. coli MGl655
E. coli MGl655 (pGTR203)
(Hye)
E. coli 0157:H7 EDL933
E. coli 0157:H7 87-23


nd plasmids used.
Genotype / Description
F-9p80dlacZAM15
A(lacZYA-argF)Ul69 deoR,
recA1, endA1, hsdR17(rk
mk/), phoA, supE44, h-, thi-
1, gyrA96, relA1
F-9p80dlacZAM15
A(lacZYA-ar gF)U l69 deoR,
recA1, endA1, hsdR1 7(rk
mi), phoA, supE44, h- thi-
1, gyrA96, relA1 (pNR100)
F- mcrA A(mrr-hsdRMS-
mcrBC) (p80dlacZAM15
AlacX74 recAl1 endAl1
araD139 A(ara, leu)7697
galU galK h- rpsL nupG
pir (DHFR)
F- mcrA A(mrr-hsdRMS-
mcrBC) (p80dlacZAM15
AlacX74 recAl1 endAl1
araD139 A(ara, leu)7697
galU galK h- rpsL nupG
pir (DHFR) (pGTR203)
(Hy Q)
F' lac14, A(lacZ)M 15
proA B zzf: :Tnl0(TetR) /
jhuA2, supE, thi, a(lac-
proAB), A(hsdldSmcrB)5,
(r m McrBC )
F-, hsdS20(rB- mB), recAB3,
ara-14, proA2 lacYI, galK2,
rpsL20(SmR), xyl-5 mtl-1,
supE44, h-, (pMJ100)
ara, A(lac-pro), thi/
F'proA B lacP4ZAM15
endA1, recA1, gyrA96, thi,
hsdR17 (rk mk ), relA1,
supE44, A( lac-proAB) /F'
traD36, proAB,
laqI9ZAM15
F ilvG rfb-50, rph-1
F ilv(T, rfb-50, rph-1
(pGTR203) (Hye)
Stx1 Stx2 gyrA
Stxf~, Stx2-, gyrA


Source / Reference
Bethesda Research
Laboratories, Rockville,
MD


(69)




Epicentre, Madison, WI





This paper






New England Biolabs,
Ipswich, MA



(70)



Amersham Pharmacia,
Piscataway, NJ
Invitrogen, Carlsbad, CA




(71)
This paper

(72)
(73)









(see Materials and Methods). Analysis of ELISA data showed that S:N values as high as 7 when

less than 5 x 104 tu/mL of phagemid particles were used as a primary antibody. Previous data in

the laboratory showed that phagemid particles were not detectable in an ELISA unless they were

used at concentrations of at least 1 x 106 particles/mL. Spot titering is a measure of transducing

units (number of infectious units), while an ELISA is a measure of the number of particles that

possess antibody activity. To determine if the measured tu/mL were underestimating the

phagemid particles/mL, an anti-M13 sandwich ELISA was performed to quantify the number of

phagemid particles produced by E. coli TG1 (M13cp-dg3-sm), E. coli TG1 (M13cp-CT-sm), and

E. coli TG1 superinfected by hyperphage. A microtiter plate was coated with anti-M13 pVIII

monoclonal antibody at a concentration of 10 Cpg/mL. Phagemid particles were incubated with

the coated wells at dilutions of 1:2, 1:20, or 1:200. The phagemid particles were then detected

using HRP-conjugated anti-M13 monoclonal antibody at a dilution of 1:2,500, followed by

standard ELISA procedures (see Materials and Methods). A concentrated phagemid preparation,

which had been amplified by hyperphage, was serially diluted and used as a standard for

particles/mL. The values for particles/mL were linearly extrapolated from the standard curve of

the standard phagemid particle (Table 3-4). Phagemid particles amplified by E. coli TG1

(M13cp-CT-sm) had particles/mL to tu/mL ratios between 740 and 23,000. Therefore, phagemid

particles were being produced in high quantities; however, less than 0. 1% of the phagemid

particles produced were infectious. Phagemid particles amplified by E. coli TG1 (M13cp-dg3-

sm) had particles/mL to tu/mL ratios between 1 and 190. Phagemid particles amplified by

E. coli TG1 that was superinfected with hyperphage had particles/mL to tu/mL ratios between 6

and 24. Of the three amplification tools (M13cp-dg3-sm, M13cp-CT-sm, and commercial









have improved immensely. Particularly, the reagents and detection equipment have improved to

allow for the detection and identification of agents in as little as a few minutes (2). An ideal

detection system would be rapid, sensitive, selective, inexpensive, and would not require

extensive training of personnel to operate the system. While there is no single optimal-detection

method, there are numerous methods for detecting a variety of agents in a variety of

environments.

Immunological tools are one of the most widely used and successful methods for detection

of biological agents. Since the first radioimmunoassay was developed by Yalow and Berson in

1959 for the detection of human insulin (6), immunoassays have expanded and been used for the

detection of a variety of agents. The main components of an immunoassay are summarized by

Andreotti et al. as follows: "Immunoassays rely upon four basic components regardless of the

application and underlying technology: (i) the antigen to be detected; (ii) the antibody or

antiserum used for detection; (iii) the method to separate bound antigen and antibody complexes

from unbound reactants; and (iv) the detection method. The efficacy of any given immunoassay

is dependent on two maj or factors: the efficiency of antigen-antibody complex formation and the

ability to detect these complexes."

The most important component of an immunoassay is the antibody. The discovery of

different types of antibodies has altered the range and scope of immunoassays. Polyclonal

antibodies have largely been supplanted by monoclonal antibodies. In the current age,

recombinantly engineered antibodies are supplanting monoclonal antibodies (2). The demand

for a variety of immunological assays reflects the growing number of assays developed to

optimize their use by increasing their sensitivity, speed, handling, and cost. The specific binding

of antibody to antigen has been coupled with a variety of detection applications including









One hundred phage clones from the fifth round of panning were amplified and screened by

ELISA against y. cholerae 569B LPS, carbonate coating buffer, and PBS-treated microtiter

wells. The a-Vc 01 LPS mAb was also tested against the three coating antigens and generated a

S:N of 15.5, proving that the LPS successfully bound to the microtiter plate. Anti-BSA

phagemid particles were used in the ELISA against BSA (10Cpg/mL), carbonate coating buffer,

and PBS-treated microtiter wells and generated a S:N of 22, proving that the ELISA was a

successful assay for the detection of phages. All 100 clones gave a S:N <1.4. A positive signal

was classified as a S:N >2. Therefore, all of the clones screened by ELISA were negative,

meaning that they failed to specifically bind to K. cholerae 569B LPS.

Panning the Ph.D. 12mer phage library against y. cholerae 01 LPS immobilized onto a

microtiter well failed to yield phage clones specific to K. cholerae 01 LPS. It was therefore

attempted to improve the panning procedure. Instead of using a microtiter well as a binding

support for LPS, nitrocellulose paper was used as a binding support. Nitrocellulose paper can

bind approximately 250 to 2,500 times more LPS than a microtiter well; therefore, if more

antigen is present to pan against then the chances of isolating clones that specifically bind to LPS

may improve. Also, the elution method was tested. The previous elution method was with

glycine (pH 2.2). To determine if a glycine elution or an elution with a high concentration of the

target antigen is more successful in eluting phages, both elution methods were tested in parallel.

The theory behind an antigen elution is that a high concentration of antigen in solution will

rapidly bind phages bound to the antigen on the nitrocellulose paper when the phages are

transiently released. The end result is that these phages will elutee" off of the antigen-coated

nitrocellulose paper and bind to antigen in solution. Antigen elution is supposed to be a less

harsh treatment than acidic or basic elutions, which may denature the peptides.









During assembly of M13, the foreign protein is fused to a coat protein and displayed on the

surface of the phage. The minor coat protein III (pIll) is the most common protein for fusions,

but the maj or coat protein XIII (pXIII) and the other minor coat proteins of M13 have also been

used for recombinant fusions (58,59). Phage displays using the pIII and pXIII proteins have

different advantages. Using the pXIII protein as the fusion protein enables high copy display of

the recombinant protein because there are over 2,700 copies of the pXIII protein on the surface

of the phage. However, a drawback in using the pXIII protein as the fusion protein is its

limitations in the size of the displayed protein (60). The pXIII protein can only display peptides

less than six amino acids in length before the function of the coat protein becomes compromised

and the number of infectious particles plummets. If the size of the display peptide increases to

eight amino acids, only 40% of the phages are infective; if the display peptide increases to 16

amino acids, less than 1% of the phages are infective. Recombinant fusions using the pIII

protein are not as restricted in the size of the display peptide (53). The pIll-fusion protein can

display peptides of 100 amino acids or greater before the ability of the pIII protein to bind to the

F pilus of E coli becomes compromised (52). Also, since there are only five copies of the pIII

protein on the surface of phage particles, the ability to select high affinity binding phage particles

is greater than that of the ability of the pXIII-fusion protein.

Recombinant Phage Libraries

Recombinant phage libraries are composed of phages that display a fused protein to a coat

protein of the phage. There are two main phage display libraries, phage and phagemid libraries.

Phage libraries are M13 with the addition of a recombinant fusion to the giII gene, while

phagemid libraries are just plasmids that contain a recombinant fusion to the giII gene. There

are advantages and disadvantages to both systems. Phagemid libraries are advantageous due to

their greater library diversity because of their higher ligation-transformation efficiency, and their









Washer (Bio-Tek). One hundred microliters horseradish peroxidase (HRP) conjugated

secondary antibodies were diluted in PBS casein blocking buffer (Pierce) containing 0.05% (v/v)

Tween-20 and added to each well and incubated 30 minutes at room temperature. The secondary

antibody was aspirated and washed three times with 300 CLL of PBS-0.05T by the ELx 800 Strip

Washer (Bio-Tek). Substrate was prepared by dissolving one capsule of phosphate-citrate (PC)

buffer (0.05 M phosphate-citrate buffer (pH 5.0), 0.03% (w/v) sodium perborate) (Sigma-

Al dri ch) in 100 mL of water. A ten milligram tablet of 3, 3', 5, 5'-tetramethylb enzi dine

substrate (Sigma-Aldrich) was added to 10 mL of the PC buffer to give a final concentration of 1

mg/mL. Two hundred microliters of substrate was added to each well and incubated for 30

minutes at room temperature. The plate was read in an ELx 800 UV plate reader (Bio-Tek) at

630 nm. The data were analyzed with KcJunior (Bio-Tek) software.

Infection Efficiency

Infection efficiency experiments were done to compare the infection efficiency of

phagemid particles to bacterial strains harboring helper plasmids to the same strains not

harboring helper plasmids. Escherichia coli standing-overnight cultures were diluted 1:40 in 3

mL of medium containing 1% (w/v) glucose for strains containing no helper plasmids and in

medium containing 1% (w/v) glucose and 30 Cpg/mL chloramphenicol for strains containing

helper plasmids. The diluted cultures were grown in a 370C shaking incubator till the OD600 WAS

between 0.4 and 0.6. Once the desired OD600 WAS obtained, 3 x 107 bacteria were infected with

phage at a MOI of 0. 1 and incubated for 20 minutes in a 370C standing water bath. The

transductions were titered using the spread-titer method. Transduced E. coli strains containing

helper plasmids were titered on 2xTY AG plates containing 30 Cpg/mL chloramphenicol.

Transduced E. coli strains containing no helper plasmids were titered on 2xTY AG plates.









Table 2-1. Continued
Strain
E. coli TG1


E. coli TG1 (pGTR203)



Hye


M13cp

M13cp-CT

M13cp-dg3

pGTR203

pMJ100

pNR100
K. cholerae 569B
K. cholerae N16961


Genotype / Description
A(lac-proAB), supE, thi,
hsdD5/F' traD36, proA B,
lac4, lacZAM15
A(lac-proAB), supE, thi,
hsdD5/F' traD36, proA B,
lacP4, lacZAM 15
(pGTR203)
Hyperphage genome
(contains all M13 genes
except gIll)
M13mpl9; CamR; full
length gIII
M13mpl9; CamR; truncated
8III
M13mpl9; CamR; deleted
8III
pACYC184 with M13 gIll;
CamR
pBluescript with slt-II;
AmpR
Stx2 toxoid; AmpR
Classical, Inaba
El Tor, Inaba


Source / Reference
MRC, Cambridge, UK


Gopal Sapparapu



(66)


(64)

(64)

(64)

Gopal Sapparapu

(70)

(69)
(74)
(75)










36. Yamagami, S., Motoki, M., Kimura, T., Izumi, H., Takeda, T., Katsuura, Y., and
Matsumoto, Y. 2001. Efficacy of postinfection treatment with anti-Shiga toxin (Stx) 2
humanized monoclonal antibody TMA-15 in mice lethally challenged with Stx-producing
Escherichia coli. J. Infect. Dis. 184:738-742.

37. Ostroff, S.M., Tarr, P.I., Neill, M.A., Lewis, J.H., Hargrett-Bean, N., and Kobayashi,
J.M. 1989. Toxin genotypes and plasmid profiles as determinants of systemic sequelae in
Escherichia coli 0157:H7 infections. J. Infect. Dis. 160:994-998.

38. Boerlin, P., McEwen, S.A., Boerlin-Petzold, F., Wilson, J.B., Johnson, R.P., and Gyles,
C.L. 1999. Associations between virulence factors of Shiga toxin-producing Escherichia
coli and disease in humans. J. Clin. M~icrobiol. 37:497-503.

39. Tesh, V.L., Burris, J.A., Owens, J.W., Gordon, V.M., Wadolkowski, E.A., O'Brien, A.D.,
and Samuel, J.E. 1993. Comparison of the relative toxicities of Shiga-like toxins type I
and type II for mice. Infect. Immun. 61:3392-3402.

40. Donohue-Rolfe, A., Kondova, I., Oswald, S., Hutto, D., and Tzipori, S. 2000.
Escherichia coli 0157:H7 strains that express Shiga toxin (Stx) 2 alone are more
neurotropic for gnotobiotic piglets than are isotypes producing only Stxl or both Stxl
and Stx2. 1 Infect. Dis. 181:1825-1829.

41. Magliulo, M., Simoni, P., Guardigli, M., Michelini, E., Luciani, M., Lelli, R., and Roda,
A. 2007. A rapid multiplexed chemiluminescent immunoassay for the detection of
Escherichia coli 0157:H7, Yersinia enterocolitica, Salmonella tyiphimurium, and Listeria
monocytogenes pathogen bacteria. J. Agric. Food Chem. 5 5:493 3-493 9.

42. Wang, L.J., Wu, C.S., Hu, Z.Y., Zhang, Y.F., Li, R., and Wang, P. 2008. Sensing
Escherichia coli 0157:H7 via frequency shift through a self-assembled monolayer based
QCM immunosensor. J. Zhejiang. Univ Sci. B 9:121-131.

43. Johnston, L.M., Elhanafi, D., Drake, M., and Jaykus, L.A. 2005. A simple method for the
direct detection of Salmonella and Escherichia coli 0157:H7 from raw alfalfa sprouts
and spent irrigation water using PCR. J. FoodProt. 68:2256-2263.

44. Jin, H.Y., Tao, K.H., Li, Y.X., Li, F.Q., and Li, S.Q. 2005. Microarray analysis of
Escherichia coli 0157:H7. World J. Gastroenterol. 11:5811-5815.

45. Thevenot, D.R., Toth, K., Durst, R.A., and Wilson, G.S. 2001. Electrochemical
biosensors: recommended definitions and classification. Biosens. Bioelectron. 16:121-
131.

46. Theegala, C.S., Small, D.D., and Monroe, W.T. 2008. Oxygen electrode-based single
antibody amperometric biosensor for qualitative detection of E coli and bacteria in water.
J. Environ. Sci. Health A Tox. Hazard. Subst. Environ. Eng 43:478-487.


































O 2008 Crystal J. Harpley









S:N of phagemid particles used in an ELISA could have been due to less phagemid particles

present in the phagemid particle-containing supernatant that was produced in a microtiter well as

opposed to the phagemid particle-containing supernatant that was produced in a flask. This

could have been due to reduced aeration in the microtiter well causing decreased bacterial

growth and thus reduced quantities of phagemid particles produced in a microtiter well.

Phagemid particle production in microtiter wells was repeated a second time with a-BSA

phagemid particles (Fig. 3-12). Anti-BSA phagemid particles yielded significantly higher

signals (p<0.03) when produced at 300C as opposed to 370C at either glucose concentration.

These data support the previous data that 300C is a more permissive temperature than 370C for

phagemid particle production. There was not a significant difference (p=0.34) in the signals

produced by phagemid particles that were produced with or without glucose at 300C. These data

also support the previous data that 0. 1% (w/v) glucose in the medium does not provide an

advantage or hindrance to phagemid particle production. While the second experiment did not

yield a significant difference (p=0.09) in signals produced by a-B SA phagemid particles

produced in a flask opposed to in a microtiter well, the microtiter well-produced phagemid

particles yielded lower signals than did phagemid particles produced in a flask.

Because phagemid particles that were produced from the same clone resulted in various

ELISA signals when phage-containing supernatant were analyzed at a dilution of 1:2 it was

possible that the reason for the differences in ELISA signals from phage-containing supernatant

were due to differences in phagemid particle titers produced under the various conditions.

Therefore, the titers of a-B SA and Vc86 phagemid particles produced in a microtiter well and in

a flask were analyzed. Phagemid particles produced at 300C in a microtiter well yielded

approximately 107 tu/mL while phagemid particles produced at 370C in a microtiter well yielded









K~ cholerae 01 LPS that was immobilized onto nitrocellulose paper or a microtiter well
failed to yield phages that recognized y. cholerae 01 LPS.


2. To optimize panning and screening procedures to allow for a more efficient
biopanning process that will be more likely to isolate and detect specific recombinant
phagemid particles.

* In biopanning with the Tomlinson scFv phagemid library, a trypsin elution time of 10-30
minutes eluted the most phagemid particles. Screening scFv proteins by a high throughput
ELISA was an acceptable screening process and enabled the possible screening of
hundreds of clones per day. Screening phagemid particles by a high throughput ELISA
was not an acceptable screening process because phagemid particles could not be produced
in high enough titers in a microtiter well to make a high throughput ELISA screen
effective.


3. To optimize the production of phagemid particles to ensure high quality and high
quantity of phagemid particles.

* Phagemid particle amplification by E. coli harboring hyperphage or helper plasmids was
not an effective method to produce phagemid particles. Homemade hyperphage that was
produced in E. coli MGl655 (pGTR203) enabled high quality and high quantity
production of hyperphage. With homemade hyperphage, the Tomlinson J library was
effectively amplified to ensure that the redundancy of the clone population in the library
was maintained.


4. To isolate specific recombinant phagemid particles to E. coli 0157:H7 Stx2 toxin
using the optimization techniques discovered in previous aims.

* Using the improved homemade hyperphage and the homemade hyperphage-amplified
Tomlinson J scFv library to pan against a commercial Stx2 toxin preparation resulted in
isolation of phagemid particles that recognized the Stx2 toxin preparation.









extracted LPS was approximately 260 Cpg/mL. The purity of this LPS extraction was less than

0.2 Cpg of protein/yg of LPS.

The phenol-water LPS extraction (76) was performed to compare results with the TRIzol

Reagent LPS extraction. The phenol-water LPS extraction method differs little from the original

phenol-water LPS extraction described by Westphal, Luderitz, and Bister in 1952 (82). The

phenol-water LPS extraction method utilizes the property that most proteins but not LPS are

soluble in phenol, and that LPS is soluble in water. At temperatures above 680C, phenol and

water are miscible. Upon cooling to 5-100C and centrifugation, three phases result: an aqueous

phase containing LPS, a phenol phase containing proteins, and a solid phase containing water-

and phenol-insoluble compounds. Removal and purification of the LPS-containing aqueous

phase follows the separation of the phases.

One and a half liters of log phase (OD600 ~0.4-0.6) K. cholerae 569B was used for

extraction of LPS by the phenol-water extraction protocol (see Materials and Methods). The

four aqueous extractions were dialyzed and digested by DNase (Qiagen) and RNase (Qiagen)

because nucleic acids are also extracted into the aqueous phase. Prior to digestion the dialyzed

extraction was diluted 1:100 in ddH20, and the absorbance was measured at 260 nm and 280 nm

(A260 and A280). The A260, which correlates to DNA concentration, predigestion was 0.277 and

post-digestion was 0.148. The A280, which correlates to protein concentration, predigestion was

0.127 and post-digestion was 0.059. The final phenol-water extraction was examined for

proteins by Coomassie blue stain (Fig. 3-2A) and for carbohydrates by Tsai-Frasch silver stain

(see Materi als and Method s) (F ig. 3 -2B). The C oomas si e b lue stai n for phenol -water- extracted

yK cholerae LPS showed a protein band at approximately 15 kDa. This could possibly have been

due to the 0.04 mg/mL RNase, 13.7 kDa, present in the extraction. The band intensity correlated









Based on the results described above, we decided to use the phenol-water method for the

extract on of y. cholerae 0 1 LP S over the TRlzol Reagent method. The phenol-water- extracted

LPS most closely resembled y. cholerae 569B LPS (Sigma-Aldrich) when examined by silver

stain. The phenol-water extraction method even appears to be a more sensitive and delicate

extraction method than that used to extract the commercial LPS, because more precise banding

from O-antigens was detected in the phenol-water-extracted LPS than in the commercial LPS by

silver stain. The phenol-water LPS extraction extracted more protein than the TRlzol Reagent

LPS extraction; however, for using the LPS extraction for panning it is more important to have

intact O-antigens than minor protein contamination.

More V. cholerae LPS Can Be Bound to Nitrocellulose Paper than to a Microtiter Well

In most panning procedures the target molecule to be panned against is immobilized onto a

solid support. We wanted to compare the saturation limits for LPS binding to a microtiter well

and to a piece of nitrocellulose paper of equal dimensions. Enzyme-linked immunosorbent

assays were performed to determine the saturation limits of the amount of phenol-water-

extracted LPS that could be bound to a microtiter well. A 96-well Maxisorp plate (Nunc) was

coated with 0. 1-200 Cpg/mL of y. cholerae 569B LPS that was extracted by the phenol-water

method. Two different primary antibodies were used to detect the LPS: a mouse anti-

K~ cholerae 01 LPS monoclonal antibody, a-Vc 01 LPS mAb, (Austral Biologicals, San Ramon,

CA) and a rabbit anti-Y. cholerae 01 polyclonal antibody, a-Vc 01 LPS pAb (Accurate

Chemical, Westbury, NY). The primary antibodies were used at their manufacturer

recommended concentration or dilution: a concentration of 1 Cpg/mL for a-Vc 01 LPS mAb and

a dilution of 1:100 for the a-Vc 01 LPS pAb. Standard ELISA procedures were used for the

assay (see Materials and Methods) (Fig. 3-3). When using the a-Vc 01 LPS mAb a significant

difference (p<0.05) between the signal to noise (S:N) values (signal on antigen-coated








A


B


250
150
100
ye


-~r


20


Figure 3-1. Analysis of TRlzol Reagent-extracted y. cholerae N16961 LPS by SDS-PAGE.
Samples were resolved on a 12% (w/v) polyacrylamide gel and stained with A)
Coomassie blue stain or B) Tsai-Frasch silver stain. (1) TRIzol-extracted
yl cholerae N16961 LPS. (2) y. cholerae 569B LPS (Sigma-Aldrich).


MWV


MWV "1


250
150
100
75

50
37
25;


so -'
37


25
20 & 1



15 -
10









4 mL of PBS containing 0.05% (v/v) Tween-20 (PBS-0.05T). The excess PBS-0.05T was

aspirated, and bound phages were eluted with 0.5 mL of trypsin-PB S (10% (v/v) trypsin stock

(10 mg/mL trypsin (Type XIII from Bovine Pancreas) (Sigma-Aldrich, St. Louis, MO), 50 mM

Tris-HCI (pH 7.4), 1 mM CaCl2 in water) in PBS) for 10 minutes at room temperature on a

labquake. Ten microliters of the eluted phages were immediately titered, while the rest of the

eluted phages were infected into 5 mL of E coli TG1 at an OD600 Of 0.4. The infected E. coli

was incubated for 30 minutes at 370C in a standing water bath. Following incubation, the

infected cells were isolated by centrifugation at 13,776 x g for 10 minutes at room temperature.

Cells were suspended in 0.6 mL of 2xTY medium and plated on three 2xTY AG plates to

amplify the phagemid-containing E. coli. The plates were incubated overnight at 370C. The

next day 1.5 mL of 2xTY was added to the lawns of bacteria, and the bacterial lawns were

scraped from the plate in the medium and pooled together. Of the pooled bacteria, 50 CLL was

inoculated into 100 mL of 2xTY AG medium and grown in a 370C shaking incubator until the

OD600 WAS 0.4. Once the OD was reached, 10 mL of the culture was superinfected with

homemade hyperphage at a MOI of 20 and incubated for 30 minutes in a standing 370C water

bath. The superinfection was centrifuged at 13,776 x g for 10 minutes at room temperature. The

resulting pellet was suspended in 100 mL of 2xTY containing 100 Cpg/mL ampicillin, 40 Cpg/mL

kanamycin, and 0.1% (w/v) glucose and incubated overnight in a shaking 300C water bath. The

next day the culture was centrifuged at 13,776 x g for 10 minutes at 40C. Phage-containing

supernatant was precipitated with final concentrations of 13 mM Polyethylene Glycol (PEG) and

0.55 M NaC1. PEG and NaCl were dissolved in the supernatant at room temperature and then

the PEG-containing supernatant was incubated overnight at 40C. The next day the PEG-

precipitated supernatant was centrifuged at 13,776 x g for 20 minutes at 40C. The pellet was









cells, another one-round panning was performed with analysis of phagemid particle specifieity

via ELISA.

A third one-round panning was performed to analyze whole cell concentration and trypsin

elution time further. The purpose of the experiment was to compare the differences between

using 1 x 10s and 1 x 109 whole cells of y. cholerae N16961 and trypsin elution times of 10 and

30 minutes in the biopanning process. The same protocol was used as for the first one-round

panning protocol above, except that only two elution time points were taken. When the trypsin

was added to the bound phagemid particles, an aliquot of the eluted phagemid particles was

removed at 10 and 30 minutes after the start of the elution and titered immediately (Fig. 3 -8).

When panning with the Tomlinson J scFv phagemid library that contains 1% (~2.5 x 10s tu)

Vc86 phagemid particles, the amount of eluted phagemid particles did not significantly change

(p=0.50) from 10 to 30 minutes when panning with 1 x 109 whole cells. When panning with the

library that contains 1% (~2.5 x 10s tu) Vc86 phagemid particles, the amount of eluted phagemid

particles significantly decreased (p=0.04) from 10 to 30 minutes when panning with 1 x 10s

whole cells. Because the amount of eluted phagemid particles either stayed the same or

decreased from 10 to 30 minutes in the pannings with 1% Vc86 phagemid particles, it suggested

that eluting for 30 minutes had no advantage over eluting for 10 minutes.

To test if the eluted phagemid particles at the various time points were specific to

K~ cholerae, an ELISA was performed. The eluted phagemid particles from the third one-round

panning experiment were amplified in E. coli TG1 (Hye-lg), which is E. coli TG1 that was

already infected with hyperphage, and used at a concentration of 1 x 10s tu/mL in an ELISA.

The phagemid particles were analyzed by ELISA against y. cholerae whole cells (3 x 10s

CFU/mL) and PBS-treated microtiter wells. A standard ELISA protocol (see Materials and









particles were produced in microtiter wells at 300C, the titers were approximately 10' tu/mL.

When a-BSA and Vc86 phagemid particles were produced in microtiter wells at 370C, the titers

were approximately106 tu/mL. Growing phagemid particles at 300C in microtiter wells yielded

approximately 10-fold higher tu/mL titers than phagemid particles that were produced in

microtiter wells at 370C. Whether the growth medium contained 0.1% (w/v) glucose or not did

not affect the titers of phagemid particles. All phages produced in microtiter wells were

amplified by E. coli TG1 (Hye-sm). Escherichia coli TG1 (Hye-sm) is E. coli TG1 that was

already infected with hyperphage. Having hyperphage already present in E. coli removes an

additional step in phagemid particle production to enable a more high throughput screen When

a-BSA and Vc86 phagemid particles were produced in flasks at 300C with 0.1% (w/v) glucose in

the medium, the titers were approximately 10s to109 tu/mL. The phagemid particles produced in

the flasks were amplified by superinfection with hyperphage, instead of by E. coli TG1 harboring

hyperphage. Growing phagemid particles by superinfection with hyperphage may have resulted

in higher titers of phagemid particles compared to phagemid particles produced from E. coli that

already contained hyperphage. The high throughput phagemid particle ELISA was not optimal

for screening phagemid particles because phagemid particles could not be produced in high

enough titers in microtiter wells to make screening of phagemid particles effective in a high

throughput manor.

Conclusion of Specific Aim 2

When panning with the Tomlinson J human synthetic VH + VL phagemid library, trypsin

elution was best between 10 and 30 minutes. Phagemid particles could possibly be degraded by

trypsin treatment of 45 minutes or greater. When screening scFvs, a high throughput ELISA was

acceptable. The scFv proteins can be diluted 1:10 in an ELISA with other scFv proteins and still

produce high enough signals to be detected. When screening phages, screening in a high









simplicity enables easier genetic manipulation than with phage vectors. Also, phagemid particles

isolated from phagemid libraries are able to be produced to generate more phagemid particles or

secrete the recombinant protein, which is usually the ultimate end product. A disadvantage of

phagemid particles is their dependence on the aid of a helper phage to provide the rest of the

M13 proteins in trans to enable phagemid particle production and assembly. Phage libraries are

advantageous due to their lack of dependence on helper phages. Because phage libraries are

M13 phage with some alterations, they are capable of propagating themselves by simple

infection ofE. coli. A disadvantage to phage libraries is that they are less stable than phagemid

libraries (61). Also, phage particles are only able to produce phages and have to be further

manipulated to be able to produce the recombinant protein alone.

Tomlinson I + J Human Synthetic VH + VL Phagemid Libraries

The Tomlinson libraries are naivev" libraries comprised of approximately 1 x 10s random

phagemids derived from non-immunized human donors (62). Naive libraries enable greater

diversification of antibody genes, increasing the probability of isolating phagemids specific to a

wide variety of targets. The Tomlinson libraries also encode greater diversity through random

side-chain diversification. These phagemid libraries encode a single chain F variable (scFv)

gene fusion to the giII gene of a library vector plasmid contained in a M13 phage. The scFv

(Fig. 1-2) is composed of a single polypeptide with VH and VL domains that are joined by a

flexible glycine-serine linker. The Tomlinson libraries were constructed by use of reverse

transcription and PCR to amplify the VH and VL antibody genes from B-lymphocytes of human

donors. Universal degenerate primers were then used to anneal to the 5' end of the exons

encoding the antibody V-gene, which is conserved in humans. The mRNA from the B

lymphocytes was converted to cDNA, which represents the VH and VL antibody genes. The

cDNA was then PCR assembled using an overlap extension technique and contained the




























108 108 108 108 109 109 109 109 a-Vc Vc86
10 10 30 30 10 10 30 30 01
min min min min min min min min LPS
+Vc86 +Vc86 +Vc86 +Vc86 mAb


Figure 3-9. Analysis of eluted phagemid particles from the third one-round panning optimization
experiment by ELISA. Microtiter wells coated with 3 x 10s CFU/mL K. cholerae
whole cells or PB S and were incubated with 1 x 10s tu/mL of amplified eluted
phagemid particles or 10 Cpg/mL of mouse anti-Y. cholerae 01 LPS monoclonal
antibody. Bound phagemid particles were detected with HRP-conjugated anti-M13
secondary antibody, and the monoclonal antibody was detected with HRP-conjugated
goat anti-mouse antibody. Wells were developed with TMB substrate, and their
absorbances were read at 630 nm (n=3 wells). There was a significantly higher S:N
(*p=0.01-0.04) from the amplified eluted phagemid particles from the panning with
10s whole cells with a 30 minute elution using the library containing 1% Vc86 when
compared to the amplified eluted phagemid particles from the other pannings.









phagemid particles will express non-recombinant pIII proteins along with recombinant pIII

proteins. To overcome the drawback of helper phages encoding a wild type giII gene, many giII

gene mutated variations of helper phages were engineered. Many of these strains either have a

deletion of the giII gene or encode amber stop codons within the giII gene to lessen the

expression of wild type pIII proteins (65-67). Of the giII gene mutated helper phages,

hyperphage is the most effective for producing high titers of phagemid particles that have

multivalent display of pIll-fusion proteins (66). Hyperphage was the previous amplification tool

used by our laboratory to amplify phagemid particles; however, the titer of the hyperphage stock

was too low to effectively amplify every phagemid particle in the library. Therefore, a way to

improve phagemid particle production was investigated.

To overcome the problem of the low titer of the stock of hyperphage from effectively

infecting every phagemid-containing bacterium, the hyperphage genome was transduced into

E. coli to be maintained as a plasmid to ensure that every phagemid-containing bacterium also

contained hyperphage to enable amplification of phagemid particles. When hyperphage was

transduced into E. coli TGl, small (0.5 mm diameter) and large (1-1.5 mm diameter) colonies

resulted. When the small and large colonies were passage, the small colonies maintained a

small phenotype (0.5 mm diameter) and the large colonies maintained a large phenotype (1-1.5

mm diameter). The two strains were named E. coli TG1 (Hye-sm) and E. coli (Hye-lg). The

reason for the variation in the colony size was not known. Because colonies were grown under

kanamycin resistance for hyperphage, both strains contained hyperphage, or at least they

contained the kanamycin resistance gene that hyperphage carries. The differences in the colony

sizes suggested that the bacteria that formed the colonies were not replicating at the same rate.

The smaller colonies might be small because they are under stress from replicating the









Abstract of Thesis Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Master of Science

OPTIMIZATION OF METHODS FOR PHAGE DISPLAY USING SINGLE-CHAIN
VARIABLE FRAGMENT PHAGEMID LIBRARIES

By

Crystal J. Harpley

August 2008

Chair: Paul A. Gulig
Major: Medical Sciences

Detection assays for biological agents and their products are important to identify disease-

causing organisms in water, air, food, and patient samples to enable prevention or treatment of

the disease. A commonality in almost all detection assays are proteins that specifically bind to a

target molecule. While monoclonal antibodies are the most common detection reagents used in

detection assays, phage display reagents are becoming more prevalent. Phage display involves

the display of recombinant peptides on bacteriophages. These recombinant phages can be

panned against a target antigen to select recombinant peptides that bind specifically to the

antigen. Optimization of protocols and reagents used in phage display was the goal of this thesis.

Optimization of whole cell panning with the Tomlinson J human synthetic VH + VL (ScFv)

phagemid library revealed that a trypsin elution between 10 and 30 minutes eluted the highest

number of phagemid particles; this may be because the phagemid particles were degraded by

trypsin treatment after 45 minutes. A high throughput ELISA for the production and screening

of scFv proteins was developed and demonstrated that scFv proteins produced in a microtiter

well could be screened by ELISA and produce a detectable S:N even when diluted 1:10. A high

throughput ELISA for the production and screening of phagemid particles was attempted, but









(81 mM Tris-HC1, 20 mM Tris-Base, 154 mM NaC1, pH 7.5) containing 0.1% (v/v) Tween-20

(TB S-0. 1T) for the first round of panning and with 200 CLL of 0. 1 M TBS containing 0.5% (v/v)

Tween-20 (TB S-0.05T) for subsequent rounds of panning. The washed wells were incubated

with 2 x 1011 PFU of the Ph.D.-12mer peptide library (NEB) (for round one of panning) or

amplified phages (for subsequent rounds of panning) diluted in 100 CLL of TBS-0. 1T for one hour

at room temperature with gentle rocking. The unbound phages were aspirated and stored at 40C.

The wells were washed 10 times with TBS-0. 1T for the first round of panning and with 200 CLL

of TBS-0.05T for subsequent rounds of panning. The phages were eluted with 100 CLL of 0. 1 M

glycine (pH 2.2) for 10 minutes at room temperature with gentle agitation. The glycine

containing the eluted phages was transferred to a microcentrifuge tube and neutralized with 15

CLL of Tris-HCI (pH 9.0). Five microliters of the eluted phages were immediately titered. One

hundred and ten microliters of the remaining elution was added to 20 mL of LB diluted 1:100

with a standing-ovemnight culture of E coli ER2738. The culture was grown for 4.5 hours in a

370C shaking incubator. The turbid culture was centrifuged at 13,776 x g for 10 minutes at 40C.

The phages were precipitated by slowly adding 3.3 mL of PEG solution (20% (w/v) PEG and 2.5

M NaCl in water) to the phage-containing supernatant with continuous swirling. The PEG-

containing supernatant was incubated overnight at 40C. The next day the PEG-containing

supernatant was centrifuged at 13,776 x g for 20 minutes at 40C. The pellet was suspended in 1

mL of 0. 1 M TBS and centrifuged at 10,621 x g for 5 minutes at room temperature to remove

bacterial debris. The phages were precipitated a second time by slowly adding 167 CLL of PEG

solution to the supernatant with continuous swirling. The PEG-containing supernatant was

incubated one hour at 40C and centrifuged at 13,776 x g for 15 minutes at 40C. The pellet was

suspended in 200 CLL of 0. 1 M TBS to be titered and stored at 40C until used in further rounds of









2.2 x 106 phagemid particles were eluted after round one of panning. The round-one-eluted

phages were amplified with homemade hyperphage to 2 x 1010 tu/mL in 100 mL ofbroth. The

amplified phagemid particles were then PEG precipitated to 9 x 10l tu/mL in 2.5 mL PBS.

Approximately 1 x 1012 phagemid particles from the PEG-precipitated phagemid particles from

round one were panned against Stx2 for a second round of panning. Approximately 2.0 x 107

phagemid particles were eluted after round two of panning. The round-two-eluted phages were

amplified with homemade hyperphage to 2.5 x 1010 tu/mL in 100 mL broth. The amplified

phagemid particles were then PEG precipitated to 7.8 x 10l tu/mL in 2.5 mL PBS.

Approximately 1 x 1012 phagemid particles from the PEG-precipitated phagemid particles from

round two were panned against Stx2 for a third round of panning. Approximately 1.5 x 1010

phagemid particles were eluted after the third round of panning. The increase in the number of

phagemid particles eluted (2 x 106 for round one, 2 x 10' for round two, 2 x 10'0 for round three)

after every round suggests that phagemid particles that specifically bound to Stx2 toxin

preparation were being selected and amplified.

Eighty-four phagemid particles from the third round of panning were individually

amplified with homemade hyperphage (see Materials and Methods) to be analyzed for specificity

by ELISA and Western blot experiments. Forty-four clones were initially analyzed by ELISA.

Microtiter wells coated with 10 Cpg/mL Stx2 toxin preparation were incubated with phagemid

particle-containing supernatants that were diluted 1:2 in blocking buffer, 1 Cpg/mL anti-

Stx211E10 (a-Stx2 subunit A) monoclonal antibody (Toxin Technology), 1 Cpg/mL anti-Stx2-

BBl12 (a-Stx2 subunit B) monoclonal antibody (Toxin Technology), or 1 x 10s tu/mL a-BSA

phagemid particles (negative control for determining the noise for S:N calculations) (Fig. 3-14).

Of the 44 clones screened, 40 had a S:N between 9 and 18. The two anti-Stx toxin monoclonal









CHAPTER 1
INTTRODUCTION

Methods for Bacteriological Detection and Analysis

There has always been a need for detection of biological agents. One basic reason for this

need is for prevention of disease by identifying biological agents in food, water, or air. Of key

interest are organisms that have potential as bioterrorism agents. With resent bioterrorist events,

the need for rapid detection of bioterrorism agents has increased (1). Another reason for

detection of biological agents is for the purpose of diagnosing disease by detecting agents in

patient samples so that suitable treatment can be given. Detection is also of interest to

epidemiology. By detecting agents that cause disease and analyzing them through epidemiology,

measures can be implemented to control or prevent outbreaks. Various methods have been used

to detect organisms or their products.

Since the discovery of microscopic organisms, methods for detecting the organisms have

evolved. One of the first methods of detection was culturing an organism in enrichment

medium, followed by selective medium. Afterwards, the organism would go through various

biochemical and metabolic tests for identification. However, culture-based identification of

organisms can take days or even weeks. As seen in the anthrax attacks in the fall of 2001 and the

SARS virus outbreaks of 2002-2003, symptoms can occur within days of exposure and infection

can spread rapidly. These reasons enforce the need for rapid detection of biological agents (2).

The rapid detection of biological agents is necessary to treat individuals at risk, limit

transmission of the disease, improve public health surveillance and epidemiology, and monitor

environmental impact (3-5).

Methods have progressed to optimize detection of agents and their products in clinical and

environmental settings. Over the last 25 years, assays for detection and identification of agents









particles, but eluting for too long of a time may only increase the elution of nonspecific

phagemid particles or even degrade the phagemid particles.

Comparative pannings were performed to determine the optimal concentration of

yK cholerae whole cells and optimal trypsin elution time. Parallel pannings were performed with

the Tomlinson J scFv library with and without 1% Vc86 phagemid particles spiked into the

library. Vc86 is a clone isolated from the Tomlinson I scFv library that recognizes an unknown

antigen on Y. cholerae whole cells. Because ~1011 phagemids are being used in the pannings

there only exists a 100-fold redundancy of clones in the non-spiked library. Therefore,

enumeration of eluted phagemid particles after one round of panning may not yield significant

differences in eluted titers under various test parameters. Spiking the library with 1% phagemid

particles that are specific to y. cholerae whole cells should enable distinction of an optimal

panning parameter that selects and elutes the highest number of specific recombinant phagemid

particles. We were looking for a combination of the concentration of whole cells and trypsin

elution time to elute more phagemid particles from the Vc86-spiked library than the non-spiked

library. Because the only difference between the pannings was that the Vc86-spiked library had

a higher initial concentration of specific phagemid particles to K. cholerae whole cells, an

increase in eluted phagemid particles with a certain parameter should correlate to an increase in

specific phagemid particles being eluted.

The first one-round panning used 1 x 10s or 1 x 109 whole cells of y. cholerae with trypsin

elutions of 15, 30, 45, and 60 minutes (Fig. 3-6). Elution of phagemid particles with trypsin was

optimal between 15 and 30 minutes. The phagemid particle titer decreased by 80 to 93% from

30 to 45 minutes for the four pannings. The decrease in the titer of phagemid particles was most

likely due to degradation of the phagemid particles by trypsin. Therefore, trypsin elution should










72. Riley, L.W., Remis, R.S., Helgerson, S.D., McGee, H.B., Wells, J.G., Davis, B.R.,
Hebert, R.J., Olcott, E.S., Johnson, L.M., Hargrett, N.T., Blake, P.A., and Cohen, M.L.
1983. Hemorrhagic colitis associated with a rare Escherichia coli serotype. N. Engl. J
M~ed. 308:681-685.

73. Karpman, D., Connell, H., Svensson, M., Scheutz, F., Alm, P., and Svanborg, C. 1997.
The role of lipopolysaccharide and Shiga-like toxin in a mouse model ofEscherichia coli
0157:H7 infection. J. Infect. Dis. 175:611-620.

74. Bik, E.M., Gouw, R.D., and Mooi, F.R. 1996. DNA fingerprinting of Vibrio cholerae
strains with a novel insertion sequence element: a tool to identify epidemic strains. J
Clin. M~icrobiol. 34:1453-1461.

75. Heidelberg, J.F., Eisen, J.A., Nelson, W.C., Clayton, R.A., Gwinn, M.L., Dodson, R.J.,
Haft, D.H., Hickey, E.K., Peterson, J.D., Umayam, L., Gill, S.R., Nelson, K.E., Read,
T.D., Tettelin, H., Richardson, D., Ermolaeva, M.D., Vamathevan, J., Bass, S., Qin, H.,
Dragoi, I., Sellers, P., McDonald, L., Utterback, T., Fleishmann, R.D., Nierman, W.C.,
and White, O. 2000. DNA sequence of both chromosomes of the cholera pathogen Vibrio
cholerae. Nature 406:477-483.

76. Evans,T.J. 2000. Septic Shock Methods and Protocols. In Methods in Molecular
Medicine. Humana Press, 19-23.

77. Yi, E.C. and Hackett, M. 2000. Rapid isolation method for lipopolysaccharide and lipid
A from gram-negative bacteria. Analyst 125:651-656.

78. Tsai, C.M. and Frasch, C.E. 1982. A sensitive silver stain for detecting
lipopolysaccharides in polyacrylamide gels. Anal. Biochem. 119:115-119.

79. Sasse, J. and Gallagher, S.R. 2003. Detection of proteins on blot transfer membranes.
Curr. Protoc. Mol1. Biol. Chapter 10:Unitl0.

80. Waldor, M.K., Colwell, R., and Mekalanos, J.J. 1994. The Vibrio cholerae 0139
serogroup antigen includes an O-antigen capsule and lipopolysaccharide virulence
determinants. Proc. Natl. Acad'. Sci. U. S. A 91:11388-11392.

81. Villeneuve, S., Boutonnier, A., Mulard, L.A., and Fournier, J.M. 1999. Immunochemical
characterization of an Ogawa-Inaba common antigenic determinant of Vibrio cholerae
01. Microbiology 145 ( Pt 9):2477-2484.

82. Westphal, O., Luderitz, O., and Bister, F. 1952. Uber Die Extraktion Von Bakterien Mit
Phenol Wasser. Zeitschrift fur Naturforschung Section B-A Journal of Chemical Sciences
7:148-155.

83. Villeneuve, S., Souchon, H., Riottot, M.M., Mazie, J.C., Lei, P., Glaudemans, C.P.,
Kovac, P., Fournier, J.M., and Alzari, P.M. 2000. Crystal structure of an anti-
carbohydrate antibody directed against Vibrio cholerae 01 in complex with antigen:
molecular basis for serotype specificity. Proc. Natl. Acad'. Sci. U. S. A 97:8433-8438.









incubation an aliquot of the growing culture was removed and titered to ensure that the library

was initially in a concentration high enough that redundancy of clones was present. The titer of

the E. coli TG1 (Tomlinson J scFv phagemid library) 30 minutes into incubation was 1.3 x 107

CFU/mL, and because the culture volume was one liter there were 1.3 x 1010 CFU. The

Tomlinson J phagemid library contains ~1 x 10s phagemid clones; therefore, there was initially a

100-fold redundancy of clones in the library. Approximately 2.3 x 1010 CFU of log phase E. coli

TG1 (Tomlinson J scFv phagemid library) was superinfected with homemade hyperphage at a

MOI of 10 and incubated for 30 minutes in a 370C standing incubation. The superinfected

bacteria were then titered on 2xTY AG plates containing 40 Cpg/mL kanamycin to ensure that

enough bacteria were superinfected so that the redundancy of the phagemid clones was

maintained. Approximately 2.9 x 1010 bacteria containing a phagemid were superinfected, which

represents about 300-fold redundancy of clones. The superinfected bacteria were then grown in

1 liter of broth under conditions that promoted phagemid particle production (see Materials and

Methods). The final yield Tomlinson J scFv phagemid particles in the amplified culture was

6.6 x 109 tl/mL. The liter of phagemid particles was then PEG precipitated to 15 mL of

4.4 x 1011 tu/mL, which yields 6.6 x 1012 tu. Titering throughout the Tomlinson J scFv library

amplification ensured that the redundancy of the library clones was maintained at each step. The

Tomlinson J scFv library that was amplified with homemade hyperphage was frozen and also

used for biopanning.

Conclusion of Specific Aim 3

Escherichia coli TG1 harboring hyperphage was not an optimal phagemid particle

amplification tool because of the low infection efficiency with phagemid particles. The infection

efficiencies of phagemid particles into E. coli TG1 (Hye-sm) and E. coli TG1 (Hye-lg) were

28% or less than the infection efficiency of phagemid particles into E. coli TGl. Phagemid









Because the high throughput screen for scFv proteins was successful, it was examined if

a similar high throughout ELISA could be applied to the screening of phagemid particles.

Previous methods of screening phages involved producing phagemid particles in a culture tube

or flask and screening the supernatants by ELISA. This method was time consuming; therefore,

a high throughput screen was investigated. Previous high throughput ELISAs have been

developed for the screening of phages; however, while their screening is rapid, the production of

the phages is time consuming. These high throughput ELISAs require overnight production of

phages in a culture tube or flask followed by PEG precipitation of the phages for the

concentrated phages to be screened by ELISA (100,101). We hoped to develop an even more

high throughput ELISA by not only screening phagemid particles in a microtiter well but also

growing the phagemid particles in a microtiter well.

Phagemid particles from the a-BSA and Vc86 clones were transduced into E. coli TG1

(Hye-sm), which harbors the hyperphage genome as a plasmid. Producing phagemid particles

in E. coli TG1 (Hye-sm) removes the additional step of superinfection with hyperphage;

therefore, reducing the time required to produce phagemid particles. Colonies from the

transduction were picked and grown in a microtiter well to produce phagemid particles. Growth

temperature and glucose concentration of the medium were analyzed. The Tomlinson protocol

recommends producing phagemid particles in medium that has 0.1% (w/v) glucose. However,

this concentration of glucose could promote catabolite repression of the lac promoter, which

promotes transcription of the gHI gene. Therefore, it was examined if phagemid particles would

be produced better in medium with or without 0. 1% (w/v) glucose.

The a-B SA and Vc86 phagemid particles were produced in a microtiter well at 300C or

370C with and without 0. 1% (w/v) glucose. The phage-containing supernatants were diluted 1:2









throughput ELISA was not optimal. Phagemid particles cannot be produced in high enough

titers when produced in microtiter wells to be useful in a high throughput manor.

Specific Aim 3: Improve Phagemid Particle Production

Phagemid particle production is one of the most important steps in biopanning. Because

the only M13 gene in a phagemid is the giII gene, use of phagemids requires an amplification

tool to be provided in trans to supply the additional M13 genes to enable phagemid particle

production. There needs to be high quality and high quantity phagemid particle production

during amplification of phagemid particles to ensure that redundancy of clones in the library is

maintained and clones are produced in high enough titers to be useful in assays that use

phagemid particles. Hyperphage (Progen) was the previous phagemid particle amplification

tool. Hyperphage is an M13-based phage that contains a partial deletion of the giII gene to

enable production of phagemid particles that display recombinant pIII proteins but no wild-type

pIII proteins. However, the commercially obtained hyperphage stock was of poor quality. The

hyperphage stock was stated to contain 1012 particles/mL; however, it only had 10s tu/mL, as

determined in an infection assay. Because there was not enough Tomlinson scFv phagemid

library stock for panning experiments, the library stock needed to be amplified. The Tomlinson I

and J scFv libraries contain 1 x 10s phagemid clones. To ensure that the Tomlinson scFv

libraries are amplified to ensure that the redundancy of clones is maintained, an amplification

tool needs to enable high quality and high quantity phagemid particle production. Because

commercially obtained hyperphage had too low of titers to be an effective phagemid particle

amplification tool, other ways of phagemid particle production were analyzed.









CHAPTER 5
EPILOGUE

The homemade hyperphage-amplified Tomlinson J library produced in these studies was

used to pan against E. coli 0157:H7 flagella and yielded over 80% positive phagemid clones.

The phagemids recognized the maj or flagellin protein by Western blot. The high throughput

scFv screen was used in the E. coli 0157:H7 flagellum proj ect and proved useful in identifying

rare scFv-secreting clones. Therefore, the improvements on phage display techniques and tools

described in this thesis are, in fact, useful and offer promise of success for continuing studies in

the laboratory.









was suspended in 25 mL of PBS by vortexing. The suspension was centrifuged at 13,776 x g for

10 minutes at 40C. The supernatant was discarded, and the pellet was suspended in 10 mL of

TES (20% (w/v) D-sucrose, 30 mM Tris-HC1, 1 mM EDTA) in water. The suspended pellet was

incubated for 15 minutes at room temperature with occasional swirling. The suspension was

centrifuged at 13,776 x g for 10 minutes at 40C. The supernatant was discarded, and the pellet

was gently washed with 10 mL of 0.5 mM MgCl2 without dislodging the pellet. The 10 mL of

the 0.5 M MgCl2 WASh was discarded, and the pellet was suspended by vortexing in 10 mL of 0.5

mM MgCl2. The suspension was incubated in an ice bath for 15 minutes with occasional

swirling. The suspension was centrifuged at 21,525 x g for 30 minutes at 40C. The supernatant

was collected and concentrated with 10,000 MWCO Amicon (Millipore, Billerica, MA) filters

and stored at 40C. The pellet was suspended in 10 mL of PBS and stored at 40C.

Determination of Protein Concentration

Protein concentration was determined by DC-Protein Assay (Bio-Rad). Protein samples

were diluted in PBS. Five microliters of each sample was added to a microtiter well followed by

25 pIL of Reagent A and 200 pIL of Reagent B. All wells were performed in triplicate. Protein

standards used were Bovine Serum Albumin (B SA) in concentrations from 0-1.4 mg/mL. The

reaction took place for 15 minutes at room temperature and was read at 750 nm with an ELx 800

UV plate reader (Bio-Tek). Protein concentrations were analyzed using KC Junior (Bio-Tek)

software.

Sodium Dodecyl Sulfate-Polyacrylamide Gel Electrophoresis (SDS-PAGE)

Sodium dodecyl sulfate-polyacrylamide gel electrophoresis was performed to resolve

yK cholerae LPS and protein preparations. ReadyGel Tris-glycine 12% (w/v), 15% (w/v), and 4-

20% (w/v) PAGE gels (Bio-Rad) were used with the MiniProtean-Electrophoresis system (Bio-

Rad) for SDS-PAGE analysis of samples. The LPS samples were sonicated by a Vibra Cell





I I I I I I 1 111111 1 I I I 1 111111 1 I I I I 11111 1 I I I I I II,,


20

15

10


Figure 3-14. Analysis of phagemid particles (1-22, 43-64) selected from panning against Stx2
toxin preparation by ELISA. Microtiter wells were coated with 10 Cpg/mL of Stx2
toxin preparation (Toxin Technologies) and incubated with clones 1 through 22 and
43 through 64 diluted 1:2 in blocking buffer, 1 Cpg/mL anti-Stx2 subunit A
monoclonal antibody (11E10), 1 Cpg/mL anti-Stx2 subunit B monoclonal antibody
(BBl12), or anti-BSA phagemid particles diluted 1:2 in blocking buffer. Microtiter
wells coated with 10 Cpg/mL BSA were incubated with anti-BSA phagemid particles
that were diluted 1:2 in blocking buffer. Bound phagemid particles were detected
with HRP-conjugated anti-M13 monoclonal antibody. Bound monoclonal antibodies
were detected with HRP-conjugated goat affinity purified anti-mouse antibody.
Wells were developed with TMB substrate, and their absorbances were read at 630
nm (n=2 wells). The noise used to calculate S:N was the signal of anti-BSA
phagemid particles that were incubated with Stx2-coated wells.









produced phagemid particles have over 5 to 20-fold increased activity (67). Of the three giII

gene mutated helper phages, hyperphage has the highest display of recombinant fusions. This is

most likely due to partial read through of the amber stop codons in non-suppressor strains with

Ex-phage or Phaberge phage. This partial readthrough generates display of wild-type and

fusion-pIII proteins. Hyperphage does not have any readthrough of the wild-type giII gene

because of the partial deletion of the giII gene, ensuring that all pIII proteins are recombinant

fusions. Hyperphage not only has a higher display level of fusion-pIII proteins, but it also

packages phagemid particles over 100 times more efficiently than Ex-phage or Phaberge phage

(68).

Of the giIl gene-mutated helper phages, hyperphage is the most advantageous for use in

phage display systems. Hyperphage enables multivalent display of pIII fusions, which makes it

ideal in phagemid particle production. However, an issue with hyperphage is that the phagemid

particle stocks generated using hyperphage have titers of 109-10 phagemids/mL (64). This is a log

or two lower than phagemid particles produced using helper phage. However, many of the

phagemid particles generated by helper phage are useless because they bear no recombinant pIII

proteins. Therefore, even though the quantity of phagemid particles produced with hyperphage

is lower than those produced with helper phage, the quality of phagemid particles produced is

much higher.

Helper Plasmids

Helper plasmids (64) are M13-based plasmids used for phagemid particle amplification

that are engineered in three forms to overcome many disadvantages of helper phage and

hyperphage. The plasmids are M13mpl9 with a chloramphenicol resistance gene cloned in from

pBSL121 to allow for selection ofE. coli containing the helper plasmids. The M13 origin of

replication was deleted and replaced with the pl5a origin of replication from pMPM-K3. This









Protein and Lipopolysaccharide (LPS) Manipulations

Phenol-Water Extraction of LPS

A phenol-water extraction (76) was performed to extract y. cholerae 569B LPS from

whole cells. One and a half liters of y. cholerae 569B was grown to log phase in a 370C shaking

incubator. Bacteria were pelleted at 13,776 x g for 10 minutes at 40C and suspended in 10 mL of

1% (w/v) NaCl in water. The bacteria were centrifuged and suspended in another 10 mL of 1%

(w/v) NaCl in water. Next, the bacteria were centrifuged and suspended in 10 mL of ddH20.

This suspension was warmed to 700C, along with 24 mL of 90% (w/v) phenol (made directly

before extraction). Ten milliliters of warmed 90% (w/v) phenol was added to the 10 mL cell

suspension and vortexed. The twenty milliliter suspension was warmed in a 700C water bath for

20 minutes while vortexing frequently. The phenol suspension was transferred to an ice bath

where it was swirled in ice water for 5 minutes. The phenol suspension was centrifuged at 2,284

x g for 25 minutes at 40C. The aqueous phase of the phenol extraction was transferred to a new

centrifuge tube and stored on ice. The same volume of water as that extracted was added back

into the phenol suspension tube and heated to 700C for 20 minutes with frequent vortexing. The

extraction step was repeated, and the two aqueous phase extractions were combined into the

same tube. Aqueous extractions were placed in a 700C water bath for 15 minutes. Twelve

milliliters of warmed 90% (w/v) phenol was added to the warmed aqueous extraction and

incubated at 700C for 10-15 minutes with frequent vortexing. Two back extractions were

performed on the aqueous extractions. The final extraction was dialyzed (MWCO: 3,500) for

24-48 hours at 40C. The dialyzed extraction was digested with DNase (Qiagen) (20 Clg

DNase/mL dialysis volume) and RNase (Qiagen) (40 Clg RNase/mL dialysis volume) in MgCl (1

CIL 20% (w/v) MgCl/mL dialysis volume) overnight at 370C. The digested extraction was









Specific Aim 4: Isolation of Specific Recombinant Phage to Stx2 Toxin of E. coli 0157:H7

Once the Tomlinson J scFv library was amplified and homemade hyperphage was

improved to enable high quality and high quantity amplification of the library, a target molecule

had to be selected for panning. Escherichia coli 0157:H7 is an enterohemorrhagic serotype of

E. coli that is responsible for approximately 73,000 illnesses and more than 60 deaths per year in

the United States (30). Stx toxins kill mammalian cells and are the maj or causes of damage in

E. coli 0157:H7-infected individuals. Escherichia coli 0157:H7 can produce Stx toxins 1

and/or 2. Escherichia coli that produces Stx2 toxin appears to be more virulent than E. coli that

produces Stxl or both Stxl and 2 toxins (36-40). Therefore, Stx2 toxin would be a good target

for detection in animal and patient samples.

A Stx2 toxin preparation was obtained that was stated to be 50% pure for Stx2 toxin. The

other 50% of the preparation most likely contained bacterial impurities from E. coli HB 101

because the Stx2 toxin preparation was made from E. coli HB 101 (pMJ100-encodes Stx2 toxin).

A 100% pure Stx2 preparation would have been ideal; however, if the preparation contains 50%

Stx2 toxin then the most of the clones selected should theoretically be specific to Stx2 toxin.

The Tomlinson J scFv library that was amplified with homemade hyperphage was panned

against the Stx2 toxin preparation. Three rounds of panning were performed with amplification

of eluted phagemid particles by homemade hyperphage between each panning. Every panning

used 1 x 1012 phagemid particles. The number of eluted phagemid particles for the three

pannings were as follows: 2 x 106 for round one, 2 x 10' for round two, and 2 x 10'0 for round

three. The increase in the numbers of eluted phagemid particles suggested that phagemid

particles specific to Stx2 toxin preparation were being selected and amplified at each round.

Clones from the third round of panning were individually amplified and screened by ELISA.

Analysis of 59 clones resulted in 53/59 clones (89%) having a S:N from 2 to 23 with 48/59









resulting titers were: 2.5 x 10 6.3 x 106, 2.5 x 10 and 3.9 x 107 tu/mL. Producing phagemid

particles from E. coli MGl655 that contained a helper plasmid did not increase the titer of

phagemid particles produced. The helper plasmids were extensively analyzed and determined to

be poor amplification tools. Therefore, hyperphage was the best amplification tool, but it was

available in titers too low to be used effectively.

To overcome the low titers of commercial hyperphage, we made homemade hyperphage.

Homemade hyperphage was previously made by transducing F E. coli cells that harbored a

plasmid that encoded the giIl gene, pGTR203, with hyperphage and growing these transduced

cells under conditions that promoted phage production. In the past, hyperphage amplified from

E. coli TG1 (pGTR203) yielded titers between 107 to 10s tu/mL and E. coli HB2151 (pGTR203)

yielded titers between 106 to 107 tu/mL. Many liters of hyperphage had to be produced and

concentrated to generate hyperphage in a titer high enough to be used for a single panning

experiment. Therefore, a way to increase the titer of homemade hyperphage was investigated to

generate enough hyperphage to be used in multiple experiments. Escherichia coli TG1 and

E. coli HB2151 are both F+ strains. Because hyperphage infects bacteria by the F pilus, it was

hypothesized that some of the hyperphage produced were being taken back into the bacteria

which would decrease the titers of the hyperphage produced. Therefore, it was investigated if

hyperphage could be produced in higher titers when produced from a F- E. coli strain.

Escherichia coli MGl655 and E. coli EC 100D were the two F- strains of E. coli chosen to

amplify hyperphage. These strains were transformed with pGTR203 encoding the giII gene and

the hyperphage genome, and Hyve isolates from each strain were grown under conditions to

promote phage production. The particles/mL and the tu/mL were analyzed (Table 3-5). The

average titer of hyperphage produced from the five E. coli EC 100D (pGTR203) (Hy@) isolates




Full Text

PAGE 1

1 OPTIMIZATION OF METHODS FOR PH AGE DISPLAY USI NG SINGLE-CHAIN VARIABLE FRAGMENT PHAGEMID LIBRARIES By CRYSTAL J. HARPLEY A THESIS PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLOR IDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE UNIVERSITY OF FLORIDA 2008

PAGE 2

2 2008 Crystal J. Harpley

PAGE 3

3 To my parents, siblings, and Matt. They, never doubted my abilities and always pushed me harder when I doubted myself.

PAGE 4

4 ACKNOWLEDGMENTS I would like to thank m y mentor, Paul A. Gulig, for teaching, guiding, and pushing me to be a better scientist. I would also like to thank my co-workers in the Gulig lab for their encouragement, help, and making me smile thro ughout the work day. Lastly, I would like to thank my committee members, Shouguang Jin and Anita Wright, for their help and guidance.

PAGE 5

5 TABLE OF CONTENTS page ACKNOWLEDGMENTS...............................................................................................................4 LIST OF TABLES................................................................................................................. ..........8 LIST OF FIGURES.........................................................................................................................9 LIST OF ABBREVIATIONS........................................................................................................ 11 ABSTRACT...................................................................................................................................13 CHAP TER 1 INTRODUCTION..................................................................................................................15 Methods for Bacteriological Detection and Analysis.............................................................15 Vibrio cholerae .......................................................................................................................18 Escherichia coli O157:H 7......................................................................................................20 Phage Display.........................................................................................................................23 Recombinant Phage Libraries................................................................................................. 24 Tomlinson I + J Human Synthetic VH + VL Phagemid Libraries.................................... 25 New England Biolabs 12mer Peptide Phage Library...................................................... 26 Tools for Production of Phages from Phagemids................................................................... 27 Helper Phage...................................................................................................................27 Helper Phage with gIII Mutations ................................................................................... 28 Helper Plasmids...............................................................................................................29 2 MATERIALS AND METHODS........................................................................................... 35 Bacterial Strains, Media, and Growth Methods...................................................................... 35 Biopanning of Phage Display Libraries.................................................................................. 36 Panning on Immunotubes................................................................................................36 Panning in Suspension..................................................................................................... 38 Panning on Microtiter Wells........................................................................................... 39 Panning on Nitrocellulose Paper.....................................................................................41 Spot Titer of Phages........................................................................................................43 Spread Titer of Phages.....................................................................................................43 Amplification of Phages.................................................................................................. 43 High Throughput Production of Soluble An tibody Fragm ents (scFv antibodies).......... 44 Deoxyribonucleic Acid Manipulations................................................................................... 44 Plasmid Extractions......................................................................................................... 44 Agarose Gel Electrophoresis...........................................................................................44 Electroporation of Plasmids............................................................................................45 Electrocompetent Cells......................................................................................................... ..45 Enzyme-Linked ImmunoSorbent Assays (ELISAs)............................................................... 46

PAGE 6

6 Infection Efficiency........................................................................................................... .....47 Protein and Lipopolysacchar ide (LP S) Manipulations........................................................... 48 Phenol-Water Extraction of LPS.....................................................................................48 TRIzol Reagent Extraction of LPS..................................................................................49 Extraction of Periplasmic Proteins.................................................................................. 49 Determination of Protein Concentration......................................................................... 50 Sodium Dodecyl Sulfate-Polyacrylam ide Gel Electrophoresis (SDS-PAGE) ................ 50 Coomassie Blue Staining................................................................................................. 51 Tsai-Frasch Silver Staining............................................................................................. 51 Western Blot....................................................................................................................52 Lipopolysaccharide Saturation to Nitrocellulose Paper .................................................. 53 Colony Blot with scFv.....................................................................................................54 3 RESULTS...............................................................................................................................57 Rationale for Study.................................................................................................................57 Specific Aim 1: Panning to V. cholerae LPS.........................................................................57 The Phenol-Water Method Extracted V. cholerae LPS Most Closely Resembled the Comm ercially Acquired V. cholerae LPS...................................................................58 More V. cholerae LPS Can Be Bound to N itrocellulose Paper th an to a Microtiter Well ..............................................................................................................................62 Panning to V. cholerae LPS Failed to Yi eld Phages that W ere Specific to V. cholerae LPS........................................................................................................... 64 Conclusion of Specific Aim 1......................................................................................... 66 Specific Aim 2: Improve Panning and Screening Process of Biopanning .............................67 Elution of Bound Phages by Trypsin duri ng Panning W as Optimal between 10 and 30 Minutes................................................................................................................... 68 Screening of scFv Proteins by a High Throughput ELISA W as Acceptable.................. 71 Screening of Phagemid Particles by a High Throughput ELISA W as Not Optimal....... 73 Conclusion of Specific Aim 2......................................................................................... 75 Specific Aim 3: Improve Phag em id Particle Production........................................................ 76 Escherichia coli TG1 Harboring the Hyperphage Geno me Was Not an Optimal Phagemid Particle Amplification Tool........................................................................ 77 Helper Plasmids Were Not an Optimal Phage mid Particle Amplification Tool.............80 Homemade Hyperphage Titers Were Incre ased with Amplification in E. coli MG1655 (pGTR203)...................................................................................................85 Conclusion of Specific Aim 3......................................................................................... 88 Specific Aim 4: Isolation of Specific Reco m binant Phage to Stx2 Toxin of E. coli O157:H7..............................................................................................................................89 Conclusion of Specific Aim 4......................................................................................... 95 4 DISCUSSION.......................................................................................................................119 Specific Aim 1: Panning to V. cholerae LPS .......................................................................120 Specific Aim 2: Improve Panning and Screening Process of Biopanning ...........................126 Specific Aim 3: Improve Phag em id Particle Production...................................................... 135

PAGE 7

7 Specific Aim 4: Isolation of Specific Recom binant Phage to Stx2 Toxin of E. coli O157:H7............................................................................................................................146 5 EPILOGUE...........................................................................................................................153 LIST OF REFERENCES.............................................................................................................154 BIOGRAPHICAL SKETCH.......................................................................................................163

PAGE 8

8 LIST OF TABLES Table Page 2-1 Bacterial strains and plasmids used................................................................................... 55 3-1 Infection efficiencies of E. coli TG1 containing various helper plasm ids with phagemid particles............................................................................................................. 97 3-2 Infection efficiencies of E. coli TG1 (M13cp-dg3-sm ) with phagemid particles.............. 97 3-3 Infection efficiencies of E. coli TG1 (M13cp-CT-sm ) and E. coli JM109 (M13cpCT-sm) with phagemid particles........................................................................................ 98 3-4 Comparison of transducing units to pa rticle s per milliliter of amplified phagemid particles..............................................................................................................................98 3-5 Comparison of transducing units to pa rticle s per milliliter of amplified homemade hyperphage produced from FE. coli strains...................................................................... 99

PAGE 9

9 LIST OF FIGURES Figure Page 1-1 Structure of M13 phage..................................................................................................... 32 1-2 Structures of antibodies......................................................................................................33 1-3 Genetic map of pIT2 phagemid vector from the Tomlinson scFv library i .......................34 3-1 Analysis of TRIzol Reagent-extracted V. cholerae N16961 LPS by SDS-PAGE. .........100 3-2 Analysis of phenol-water-extracted V. cholerae 569B LPS by SDS-PAGE. .................. 101 3-3 Analysis of the saturation limit of V. cholerae 56 9B LPS to microtiter wells by ELISA..............................................................................................................................102 3-4 Analysis of the saturation limits of primary antibodies to V. cholerae 569B LPScoated m icrotiter wells by ELISA.................................................................................... 103 3-5 Saturation of V. cholerae 569B LPS to nitrocellulose paper. .......................................... 105 3-6 Titers of eluted phagemid particles from the first one-round panning optimization experiment..................................................................................................................... ...106 3-7 Titers of eluted phagemid particles from the second one-round panning optimization experiment..................................................................................................................... ...107 3-8 Titers of eluted phagemid particles from the third one-round panning optimization experiment..................................................................................................................... ...108 3-9 Analysis of eluted phagemid particle s from the third one-round panning optimization experiment by ELISA......................................................................................................109 3-10 Analysis of anti-BSA ( -BSA) scFv proteins produced in m icrotiter wells by ELISA.. 110 3-11 Analysis of -BSA and Vc86 phagem id particles produced in microtiter wells by ELISA..............................................................................................................................111 3-12 Analysis of -BSA phagem id particles produced in microtiter wells by ELISA............ 112 3-13 Analysis of -BSA phagem id particles produced by hyperphage by ELISA..................113 3-14 Analysis of phagemid particles (1-22, 43-64) selected from panning against Stx2 toxin preparation by ELISA.............................................................................................114 3-15 Analysis of phagemid particles (1, 3-5, 23-42, 65-84) selected from panning against Stx2 toxin preparation by ELISA....................................................................................115

PAGE 10

10 3-16 Analysis of anti-Stx2 monoclonal an tibodies and clones 46, 48, and 49 by ELISA.. ..... 116 3-17 Western blot analysis of phagem id clones from panni ng on Stx2 toxin preparation...... 117 3-18 Analysis of Stx2 toxin preparation (Toxin Technologies ) by SDS-PAGE......................118

PAGE 11

11 LIST OF ABBREVIATIONS 2xTY AG 2xTY m edium containing 100 g/mL and 1% (w/v) glucose BSA Bovine serum albumin BSG Phosphate-buffered saline c ontaining 0.1% (w/v) gelatin CFU Colony forming unit CT Cholera toxin ddH2O Deionized distilled water dH2O Deionized water DNA Deoxyribonucleic acid ECL Enhanced chemiluminescence ELISA Enzyme-linked immunosorbent assay HC Hemorrhagic colitis HRP Horseradish peroxidase HUS Hemolytic uremic syndrome Hy Hyperphage LPS Lipopolysaccharide mAb Monoclonal antibody MWCO Molecular weight cut off OD Optical density pAb Polyclonal antibody PBS Phosphate-buffered saline PBST-0.01 PBS containing 0.01% (v/v) Tween-20 PBST-0.05 PBS containing 0.05% (v/v) Tween-20 PC Phosphate-citrate PCR Polymerase chain reaction

PAGE 12

12 PEG Polyethylene glycol PFU Plaque forming unit RF Replication factor RNA Ribonucleic acid S:N Signal to noise scFv Single chain F variable SDS-PAGE Sodium dodecyl sulfatepolyacrylamide gel electrophoresis Stx-2 Shiga-like toxin 2 TMB 3,3 ,5,5 -tetramethylbenzidine tu Transducing unit MOI Multiplicity of infection Vc Vibrio cholerae v/v Volume per volume w/v Weight per volume

PAGE 13

13 Abstract of Thesis Presen ted to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Master of Science OPTIMIZATION OF METHODS FOR PH AGE DISPLAY USI NG SINGLE-CHAIN VARIABLE FRAGMENT PHAGEMID LIBRARIES By Crystal J. Harpley August 2008 Chair: Paul A. Gulig Major: Medical Sciences Detection assays for biological agents and their products are im portant to identify diseasecausing organisms in water, air, food, and patient samples to enable prevention or treatment of the disease. A commonality in almo st all detection assays are proteins that specifically bind to a target molecule. While monoclonal antibodies are the most common detection reagents used in detection assays, phage display reagents are becoming more prev alent. Phage display involves the display of recombinant peptides on bact eriophages. These recombinant phages can be panned against a target antigen to select reco mbinant peptides that bind specifically to the antigen. Optimization of protocols and reagents used in phage displa y was the goal of this thesis. Optimization of whole cell panning with the Tomlinson J human synthetic VH + VL (scFv) phagemid library revealed that a trypsin eluti on between 10 and 30 minutes eluted the highest number of phagemid particles; this may be because the phagemid particles were degraded by trypsin treatment after 45 minutes. A high th roughput ELISA for the production and screening of scFv proteins was developed and demonstrated that scFv proteins produced in a microtiter well could be screened by ELISA and produce a de tectable S:N even when diluted 1:10. A high throughput ELISA for the production and screeni ng of phagemid particles was attempted, but

PAGE 14

14 phagemid particles could not be produced in a microtiter well in titers high enough to make a high throughput ELISA acceptable for screening of phagemid particles. Various phagemid particle amplification t ools were analyzed. The hyperphage genome was incorporated into E. coli and maintained as a plasmid to enable phagemid particle amplification. However, these strains had low in fection efficiencies with phagemid particles. Helper plasmids were incorporated into E. coli to enable phagemid particle amplification. However, these strains produced low yields of phagemid particles. The titer of laboratoryproduced hyperphage was improved by use of E. coli MG1655 (pGTR203), expressing the M13 gIII gene. The improved high titer hyperphage was used to successfully amplify the Tomlinson J scFv phagemid library to ensure that the redund ancy of phagemid clones was maintained. Using the newly amplified libra ry to pan against an E. coli O157:H7 Stx2 toxin preparation resulted in 89% of phagemid particles from the thir d round of panning recognizing the Stx2 toxin preparation in an ELISA. Previous panning procedures used in this la boratory resulted in no better than 1% of selected phagemids being specific to the target antigen, and often no usable specific clones were isolated. The optimization of panning procedures and reagents developed in this thesis greatly increased the efficiency of sel ection of phagemid partic les specific to the ta rget antigen. Using the improved reagents and procedures develope d in this theis yielded over 80% positive phagemid clones selected against E. coli O157:H7 flagella by another lab member. The phagemids recognized the major flagellin protein by Western blot (data no t shown). Therefore, the improvements on phage display techniques and t ools described in this thesis are, in fact, useful and offer promise of success for continuing studies in the laboratory.

PAGE 15

15 CHAPTER 1 INTRODUCTION Methods for Bacteriological Detection and Analysis There has always been a need for detection of biological agents. One basic reason for this need is for prevention of disease by identifying biol ogical agents in food, water, or air. Of key interest are organism s that have po tential as bioterrorism agents. With resent bioterrorist events, the need for rapid detection of bioterrorism agents has increased (1). Another reason for detection of biological agents is for the purpos e of diagnosing disease by detecting agents in patient samples so that suitable treatment can be given. Detection is also of interest to epidemiology. By detecting agents that cause disease and analyzing them through epidemiology, measures can be implemented to control or prev ent outbreaks. Various methods have been used to detect organisms or their products. Since the discovery of microscopic organisms, methods for detecting the organisms have evolved. One of the first methods of detecti on was culturing an organism in enrichment medium, followed by selective medium. Afterwards, the organism would go through various biochemical and metabolic tests for identificatio n. However, culture-based identification of organisms can take days or even weeks. As seen in the anthrax attacks in the fall of 2001 and the SARS virus outbreaks of 2002-2003, symptoms can occur within days of exposure and infection can spread rapidly. These reasons enforce the need for rapid detection of biological agents (2). The rapid detection of biological agents is necessary to treat individuals at risk, limit transmission of the disease, improve public he alth surveillance and epidemiology, and monitor environmental impact (3-5). Methods have progressed to optimize detection of agents and their produc ts in clinical and environmental settings. Over the last 25 years, as says for detection and identification of agents

PAGE 16

16 have improved immensely. Particularly, the reag ents and detection equipment have improved to allow for the detection and identification of agents in as little as a few minutes (2). An ideal detection system would be rapid, sensitive, selective, inexpensive, and would not require extensive training of personnel to operate the system. While th ere is no single optimal-detection method, there are numerous methods for detectin g a variety of agents in a variety of environments. Immunological tools are one of th e most widely used and successful methods for detection of biological agents. Since the first radioimmunoassay was developed by Yalow and Berson in 1959 for the detection of human insulin (6), imm unoassays have expanded and been used for the detection of a variety of agents. The main components of an immunoassay are summarized by Andreotti et al. as follows: Immunoassays rely upon four basic components regardless of the application and underlying technolo gy: (i) the antigen to be de tected; (ii) the antibody or antiserum used for detection; (iii) the method to separate bound antigen and antibody complexes from unbound reactants; and (iv) the detection method. The e fficacy of any given immunoassay is dependent on two major factors: the efficien cy of antigen-antibody complex formation and the ability to detect these complexes. The most important component of an imm unoassay is the antibody. The discovery of different types of antibodies has altered the range and scope of immunoassays. Polyclonal antibodies have largely been supplanted by mo noclonal antibodies. In the current age, recombinantly engineered antibodies are suppl anting monoclonal antibodies (2). The demand for a variety of immunological a ssays reflects the growing number of assays developed to optimize their use by increasing their sensitivity, speed, handling, and cost. The specific binding of antibody to antigen has been coupled with a variety of detection ap plications including

PAGE 17

17 fluorescence, enzymatic activity, chemiluminescence, electrochemiluminescence, metallic beads, and many more. The detection of these complexes can be assayed on a variety of platforms such as biosensors, flow cytometry, microarray, and lateral flow diffusion devices (1). Perhaps the most common immunological assa ys are Enzyme-Linked ImmunoSorbent Assays (ELISAs). Such assays either have antibody or antigen bound to a solid support that enables specific, sensitive, and quantitative dete ction of antigen or antibody by op tical-density detection of a colorimetric signal. Nucleic acid-based tools for detection have a dvanced in the past few decades. Of main interest is Polymerase Chain Reaction (PCR), which was invented in 1983 by Kary Mullis and coworkers (7). Polymerase chain reaction involves the amplification of DNA by use of oligonucleotide primers, heat stable DNA polym erase, and nucleotides in an exponential capacity. The original PCR has since been altered to give quantitative real time-PCR (q-PCR or kinetic PCR) and reverse transcription (RT )-PCR. Quantitative real time-PCR involves the amplification along with quantif ication of the DNA, while RT-P CR involves reverse transcribing a piece of RNA into DNA followed by PCR amplification of the DNA. Advances in PCR chemistry and thermocyclers have shortened th e length of DNA amplification from a few hours to minutes. With small sample volumes in th e amount of a few microliters and containing as little as one bacterial cell, PCR is one of the most sensitive assays available. However, this sensitivity increases the risk of contaminati on generating false positiv es. Field-based PCR amplification and identification is not common du e to the complexity of the system and highly trained personnel required to operate and interpret the system. Perhaps the greatest constraint of nucleic acid-based detection assays is the availability of genomic sequence data of biological agents (2). Detection assays are continually being invented and improve d to detect biological

PAGE 18

18 agents. Because there are numerous detecti on environments, conditions, and agents, the development and improvement of detection assays will continue to progress. Vibrio cholerae Vibrio cholerae is a g ram-negative, curved rod-sh aped bacterium with a single polar flagellum found primarily in estuarine and marine environments. It is a facultative human pathogen causing the pandemic diarrheal disease cholera (8). Cholera is characterized by profuse watery diarrhea, vomiting, and leg cram ps leading to dehydration and shock (9). Cholera infection occurs through th e ingestion of food or water contaminated with the bacterium. Because cholera can be prevented by proper sa nitation and hygiene, it is uncommon in industrialized countries and is most prevalent in the Indian subcontinent and sub-Saharan Africa (10). Vibrio cholerae uses its toxin co-regulated pilus (TCP ) to colonize the small intestine; once attached to the small intestine, the bacter ium secretes cholera toxi n (CT) (11). Cholera toxin binds to the epit helial cell receptor, GM1, and is transported into the cell. Cholera toxin is an AB toxin composed of a catalytically active A-subunit surrounded by a homopentameric Bsubunit. Once the CT is internal ized, it is transported in a retr ograde pathway through the Golgi to the endoplasmic reticulum. In the endoplasmic re ticulum it is retrotranslocated to the cytosol. In the cytosol CT catalyzes ADP-ribosy lation of the GTP-binding protein Gs (8) causing adenylate cyclase to become constitutively activ ated. The increase in cAMP levels leads to secretion of Cl-, HCO3 -, and water from epithelial cells into th e intestinal lumen causing diarrhea. The loss of water can amount to 30 liters per day, and without proper rehydration treatment can lead to a 30% mortality rate (12). Under-repo rting of cholera infections is a great problem. It is estimated that 3 to 5 million cases and 120,000-200,000 deaths occur worldwide annually (13). However, this number could easily be multip lied by a factor of ten due to unreported cases.

PAGE 19

19 There are many different serogroups of V. cholerae, most of which do not cause acute diarrhea. Vibrio cholerae is classified into serovars or serogroups on the basis of its lipopolysaccharide (LPS) O-antigen (14). There are at least 200 known serogroups, of which serogroups O1 and O139 are the only ones that ca use epidemic or endemic cholera. The O1 serogroup can be further distinguished into thr ee serotypes. Ogawa and Inaba are the most common serotypes, with Hikojima being rarely reported. These sero types can be further classified into two biotypes, El Tor and classica l, that differ in their biochemical properties and phage susceptibilities (8). With the enormous numbers of cases of cholera every year, there is a need for an effective diagnosis tool for patient a nd environmental samples. De tecting endemic serogroups of V. cholerae early in an outbreak is extremely importan t for control of an epidemic. A problem with detection of V. cholerae is that cholera is a disease of developing countries. Outbreaks normally occur around water-ravaged areas where la boratories are not prevalent. Therefore, field-based assays are the most eff ective tools for early detection of V. cholerae. Because there are many serogroups of V. cholerae most of which present mild symptoms, it is important to distinguish the epidemic strains from the non-epidem ic strains. Vibrio cholerae O1 and O139 are the epidemic strains. Determining features of these serogroups are the O-antigen of their LPS and their ability to produce CT. Therefore, field tests usually detect these specific antigens. There are many rapid diagnostic tests for choler a. Some of these detect CT by passivelatex agglutination (15,16), while others detect the O-antigens of LPS from O1 (17-21) and O139 (22-24) strains of V. cholerae A commonly used rapid V. cholerae diagnostic tool is the multistep colloidal gold-based colorimetric im munoassay known as SMART. This monoclonal antibody-based test was developed for the detection of V. cholerae O1 (25,26) and O139 (24)

PAGE 20

20 strains in stool specimens. SMART is 95% sensitive and 100% specific to V. cholerae O1 strains (26) and 100% sensitive and 97% sp ecific to O139 strains (24). A one-step immunochromatographic dipstick test for the detection of V. cholerae O1 and O139 LPS in stool samples was invented by the Institute Pasteur in Paris. This assay requires minimal technical skill and rapidly detects thresholds of purified LPS at 10 ng/mL for V. cholerae O1 and at 50 ng/mL for V. cholerae O139 strains in approximately 10 minutes (27). With continuing improvement and use of methods for detection of V. cholerae, outbreaks could be lessened or prevented with proper treatment and containment. Escherichia coli O157:H7 Escherich ia coli O157:H7 is an enterohemorrhagic serotype of E. coli that causes hemorrhagic colitis (HC) and hemolytic uremic syndrome (HUS). The bacterium is primarily found on cattle farms and colonizes cattle, swine, a nd deer intestines with subclinical effects to the animals (28). Disease in humans comes from the ingestion of beef milk, vegetables, and other products that are contaminated with E. coli O157:H7. Symptoms from an E. coli O157:H7 infection include mild diarrhea abdominal pain, vomiting, bloody diarrhea, HC, strokes, and HUS (28). Over the past 23 years, 146 Shiga-like toxin producing E. coli (STEC) outbreaks and sporadic cases of human illnesses have been traced to consumption of beef contaminated with various E. coli O157 strains (29). Most of these illn esses were caused by infection with E. coli O157:H7. Of the 146 outbreaks and sporadic cases 89% occurred in the United States (29). The large number of cases in the Unites States compared with the rest of the world can be explained by high levels of beef consumption and availability of E. coli O157 diagnostic methods in the United States. The Center for Disease Control and Prevention estimated that E. coli O157:H7 is

PAGE 21

21 responsible for approximately 73,000 illnesses and more than 60 deaths per year in the United States (30). Escherichia coli O157:H7 produces Shiga toxins that ar e responsible for human disease. These Shiga toxins produce severe cytopathic effects and have a high degree of homology with Shiga-toxin (Stx) of Shigella dysenteriae type 1 (28). The Stx toxin is a member of the AB toxin family. It is composed of a catalytically active A-subunit surrounded by a pentameric B-subunit. The B subunits specifically bind to glycosphingolip id globotriosylceramide (Gb3) receptors (31) of the renal glomerular endotheli al, mesangial (32) and tubular ep ithelial cells (33). Upon entry into the cell, the catalytically active A1 subunit cleaves ri bosomal RNA leading to the cessation of protein synthesis and cell d eath (34). Not only does Stx toxin damage host cells, but it may also increase the adherence of E. coli O157:H7 to epithelial cells l eading to increased risk of colonization. Tissue culture experiments show ed that Stx toxin evoked an increase in a eukaryotic receptor, nucleolin, that binds to the E. coli O157 attachment factor, intimin, leading to increased cell adherence (32). There are two distinct toxinconverting bacteriophages ( phages), 933J and 933W, in E. coli O157:H7 that generate two genetic ally related toxins that are an tigenically distinct but create similar biologic effects (35). These two toxins are called Shiga toxins I and II (Stx1 and 2). Experimental studies suggest that E. coli that produces the Stx2 toxin is more virulent than E. coli that produces Stx1 or both Stx1 and Stx2 toxins (36). Escherichia coli strains producing Stx2 toxin are more frequently linked with th e development of HUS than Stx1 toxin-producing strains (37,38). In mouse studies, the lethal dose of purified St x2 toxin is 400 times lower than that of Stx1 toxin (39). In pi glet studies, Stx2 toxin-producing E. coli strains caused more severe neurologic symptoms than strains producing bot h Stx1 and 2 toxins or only Stx1 toxin (40).

PAGE 22

22 Detection of E. coli O157:H7 has become of increasing im portance to the food industry in the United States as outbreaks continue to o ccur. Traditional met hods of detection of E. coli O157:H7 involve plating and cu lturing, enumeration methods, biochemical testing, microscopy, and flow cytometry. Other methods have been developed, including immunoassays (41), immunomagnetic separations ( 42), nucleic acid probe-based me thods based on hybridization and polymerase chain reaction (PCR) (43), and DNA mi croarrays (44). However, many of these assays are time-consuming and not suitable for rapid detection of E. coli O157:H7. Therefore, biosensors have been developed for the rapid detection of E. coli O157:H7 cells and Stx toxins. An electrochemical biosensor is a self-contai ned integrated device, which is capable of providing specific quantitative or semi-quantitative analytical information using a biological recognition element (biochemical receptor) which is retained in direct spatial contact with an electrochemical transduction elem ent (45). Some advantages of biosensors include their continuous data acquisition ability, target specifi city, fast response, mass produce feasibility, and the simplicity of sample preparation. A quartzcrystal microbalance (QCM ) has been developed for detection of E. coli O157:H7 cells. Upon specific binding of E. coli O157:H7, the QCM uses its ultra sensitive mass-measuring sensor to detect decreases in the crystal-resonance frequency to enable detection of 2.0 x 102 CFU/mL of E. coli O157:H7 (42). An amperometric biosensor for E. coli O157:H7 cells made use of a dissolvedoxygen probe to enable detection of 50 CFU/mL in as little as 20 minutes of prep aration and processing time. Upon binding of the bacterial cells, a decrease in en zyme activity results in a change in oxygen concentration that was detected with a Clark-type oxygen electrode probe (46). Polymerase chain reaction-based assays are also very common for detection of E. coli O157:H7 cells (47-51) and Stx toxins (50).

PAGE 23

23 Phage Display Phage display was initially described in 1985 by George P. Sm ith (52) as a means to display foreign proteins on fila mentous bacteriophage. Filamentous phages are non-lytic phages with circular ssDNA genomes. Of the filament ous phages, the Ff family (f1, fd, and M13) phages infect F+ E. coli through binding with their F pilus. These phages are useful tools to link genotype and phenotype of select recombinant proteins. This link was created by encoding a foreign polypeptide in-frame with a coat protei n gene of the M13 phage. Phage display could theoretically be implemented with any phage, but filamentous phages have been the most widely used. Of the filamentous phages, M13 is the most commonly used. The M13 phage (Fig. 1-1) replicates in E. coli, turning the bacterium into a phageproduction factory (53). The b acteria harboring these phages do not lyse, but undergo reduced cell growth due to the stress of phage producti on. The M13 phage contains a 6.4-kb, circular, single-stranded DNA genome that enc odes phage proteins I to XI. Fi ve of these proteins are coat proteins. The major coat protein (pXIII) is pr esent in approximately 2,700 copies and protects the genome in a cylindrical manner. The minor coat proteins pVII and pIX are necessary for efficient particle assembly, while the minor coat proteins pIII and pVI are necessary for particle stability and infectivity (54-56). The pIII protein mediates the binding of the phage to the F pilus and is necessary for viral uncoa ting and phage DNA transfer to the cytoplasm of the bacterium (55). Host enzymes then convert the ssDNA into supercoiled dsDNA, known as the replicative form (RF) (55). The RF is essential to the pha ge display system because it can be purified and manipulated just like a plasmid. Through the ma nipulation of the RF of M13 came some of the earliest cloning vectors (57). Through some of these vectors came the development of recombinant phage libraries.

PAGE 24

24 During assembly of M13, the foreign protein is fused to a coat protein and displayed on the surface of the phage. The minor coat protein II I (pIII) is the most common protein for fusions, but the major coat protein XIII (pXIII) and the other minor coat proteins of M13 have also been used for recombinant fusions (58,59). Phage di splays using the pIII and pXIII proteins have different advantages. Using the pXIII protein as the fusion protein enab les high copy display of the recombinant protein because there are over 2,700 copies of the pXIII protein on the surface of the phage. However, a drawback in using the pXIII protein as the fusion protein is its limitations in the size of the displayed protein (6 0). The pXIII protein can only display peptides less than six amino acids in length before the function of the coat protein becomes compromised and the number of infectious partic les plummets. If the size of the display peptide increases to eight amino acids, only 40% of the phages are inf ective; if the display peptide increases to 16 amino acids, less than 1% of the phages are in fective. Recombinant fusions using the pIII protein are not as restricted in the size of the display peptide (53). Th e pIII-fusion protein can display peptides of 100 amino acids or greater before the ability of the pIII protein to bind to the F pilus of E. coli becomes compromised (52). Also, since there are only five copies of the pIII protein on the surface of phage particles, the abilit y to select high affinity binding phage particles is greater than that of the abil ity of the pXIII-fusion protein. Recombinant Phage Libraries Recom binant phage libraries are composed of pha ges that display a fused protein to a coat protein of the phage. There are two main phage display libraries, phage and phagemid libraries. Phage libraries are M13 with the addition of a recombinant fusion to the gIII gene, while phagemid libraries are just plasmids that contain a recombinant fusion to the gIII gene. There are advantages and disadvantages to both systems. Phagemid libra ries are advantageous due to their greater library diversity be cause of their higher ligation-transformation efficiency, and their

PAGE 25

25 simplicity enables easier genetic manipulation than with phage vectors. Also, phagemid particles isolated from phagemid libraries are able to be produced to generate more phagemid particles or secrete the recombinant protein, which is usually the ultimate end product. A disadvantage of phagemid particles is their depende nce on the aid of a helper phage to provide the rest of the M13 proteins in trans to enable phagemid partic le production and assembly. Phage libraries are advantageous due to their lack of dependence on helper phages. Because phage libraries are M13 phage with some alterati ons, they are capable of propa gating themselves by simple infection of E. coli. A disadvantage to phage lib raries is that they are less stable than phagemid libraries (61). Also, phage particles are only able to produce phages and have to be further manipulated to be able to produce the recombinant protein alone. Tomlinson I + J Human Synthetic VH + VL Phagemid Libraries The Tomlinson libraries are nave libra ries comprised of approximately 1 x 108 random phagemids derived from non-immunized human d onors (62). Nave libraries enable greater diversification of antibody genes, increasing the probability of isolating phagemids specific to a wide variety of targets. The Tomlinson libraries also encode greater diversity through random side-chain diversification. These phagemid libra ries encode a single chain F variable (scFv) gene fusion to the gIII gene of a library vector plasmid contained in a M13 phage. The scFv (Fig. 1-2) is composed of a single polypeptide with VH and VL domains that are joined by a flexible glycine-serine linker. The Tomlinson libraries were construc ted by use of reverse transcription and PCR to amplify the VH and VL antibody genes from B-lymphocytes of human donors. Universal degenerate primers were then used to anneal to the 5 end of the exons encoding the antibody V-gene, which is conserved in humans. The mRNA from the B lymphocytes was converted to cDNA, which represents the VH and VL antibody genes. The cDNA was then PCR assembled using an overl ap extension technique and contained the

PAGE 26

26 restriction enzyme sites for subcloning into the pIT2 library vector. The library vector, pIT2, encodes an M13 origin of replication, an ampicillin resistance gene ( bla ), and both His6 and myc tags (Fig. 1-3). The pIT2 phagemids were electroporated into E. coli, where they were superinfected with helper phage to generate phagemid particles displaying a scFv-pIII fusion protein. The phagemid particles display the sc Fv on the surface of the phage particles and encode the scFv gene in the phagemid genome, linking phenotype with genotype. New England Biolabs 12mer Peptide Phage Library The New England Biolabs 12m er Peptide Ph age Library (Ph.D. System, New England Biolabs) is a combinatorial phage library that encodes a random sequence of twelve amino acids fused to the gIII gene of the M13 genome. The 12 random amino acids are fused to the Nterminus of the pIII protein, which is displaye d on the surface of the M13 phage. The first residue of the mature protein is the first random ized position. The peptide is followed by a (GlyGly-Gly-Ser) spacer linked to the wild type pI II protein. The library is constructed in M13 phage with an insertion of the lacZ gene fragment into the genome. The insertion of the lacZ gene fragment enables distinction of E. coli that harbor phages from the library opposed to environmental phages that do not contain the lacZ gene. This is done by blue/white screening of phages on agar that contains X-gal and IPTG. When an E. coli strain has a functional lacZ gene it will produce -galactosidase, which is a he terodimer composed of an and an peptide. Neither of these peptides have enzymatic activity on their own; therefore, if one component is missing then -galactosidase enzymatic activity is lost and if they are complemented they spontaneously combine to generate an active structure. Beta-galactosidase cleaves X-gal into a product that becomes oxidized to generate an insol uble blue product. IPTG is an inducer of the lac promoter which drives transcription of the lacZ gene. Therefore, if phages are produced in E. coli cells with a nonfunctional lacZ gene and a functional lacZ gene and grown under

PAGE 27

27 selection for the E. coli cells then bacteria containing a pha ge from the library, which encodes lacZ will have a complemented lacZ gene to enable distinction from bacteria containing a phage from the environment. The NEB 12me r library consists of approximately 2.7 x 109 electroporated sequences that were amplified onc e to yield approximately 55 copies of each sequence. Sequencing of 104 clones from the lib rary yielded six clones (5.8%) that did not contain a displayed peptide insert. Sequenci ng from the 98 other clones revealed a wide diversity of sequences with no obvious positional biases. Tools for Production of Phages from Phagemids To propagate phagem id particles, an amplification tool must be supplied that encodes the rest of the M13 genes necessary for phagemid part icle production. These tools may be phages or plasmids and are referred to as helpers because they help the phagemid particles propagate by supplying the necessary phage genes in trans. Some often used phage amplification tools include helper phage, hyperphage, phaberge pha ge, ex-phage, and helper plasmids. Helper Phage Helper phage is the m ost common helper tool used to produce phagemid particles. There are many variations of helper phages, R408, VCSM13, and M13KO7 ( 63) that differ slightly. Of the various helper phages M13KO7 is the most commonly used. It is a derivative of M13 that has a couple differences including a kanamyci n resistance gene and the P15A origin of replication, which allows the genome to be replicated as a plasmid in E. coli. Because helper phages are basically M13 phages, they supply all of the genes necessary for production of phagemid particles. Just like M13, helper phages infect F+ E. coli through the binding of the pIII protein to the F pilus. Help er phage amplifies phagemid particles to yield average titers of 2 x 1010-12 phagemids/mL (64). However, because helper phage encodes a wild type gIII gene, the phage particles produced will contai n a mixture of wild-type pIII proteins and

PAGE 28

28 pIII-fusion proteins. This heterogeneity in di splay of fusion proteins can result in progeny phagemids bearing all wild type pIII proteins or only a monovalent display of the recombinant pIII protein (65). The low level of display of re combinant pIII proteins results in low efficiency of selection of recombinant phage mid particles. Because the help er phages will also be produced and packaged, amplification of phagemids with helper phage generates a heterogeneous mix of phage particles encoding phagemids and helper phage genomes. The number of helper phages produced can sometimes be greater than or e qual the number of phagemid particles generated (64). To get around the problems of helper phag e, many alterations have been implemented to improve helper phage. Helper Phage with gIII Mutations There are many variations to help er phage. One of the key varia tions to helper phage is the deletion or mutation of the gIII gene in the helper phage genome, yielding hyperphage, Exphage, and Phaberge phage. All three gIII -mutated helper phages ar e derivatives of M13KO7 that lack full gIII gene functionality but sti ll possess pIII proteins. Thes e pIII proteins enable the helper phages to bi nd to the F pilus of E. coli and transfer their helper phage genomes into the E. coli. Hyperphage contains a partial deletion of the gIII gene, while Ex-phage and Phaberge phage have amber stop codons within the gIII gene. Ex-phage contains two amber stop codons, and Phaberge phage contains only one. The mutations in the gIII gene of the helper phages promote multivalent display of fusion proteins, which enhances the avidity of binding of phagemid particles to the target molecule. This increased avidity is desirable because it increases the chances of selecting positive clones. In comparing the antigen-binding activity of phagemid particles produced using M13KO7, hyperphage-produced phagemid particles have ov er 400-fold increased activity (66), Ex-phageproduced phagemid particles have over 100-fold increased activity (65), and Phaberge-phage-

PAGE 29

29 produced phagemid particles have over 5 to 20fold increased activity (67). Of the three gIII gene mutated helper phages, hyperphage has the highe st display of recombinant fusions. This is most likely due to partial read through of the amber stop codons in non-suppressor strains with Ex-phage or Phaberge phage. This partial readthrough generates display of wild-type and fusion-pIII proteins. Hyperphage does not have any readthrough of the wild-type gIII gene because of the partial deletion of the gIII gene, ensuring that all pIII proteins are recombinant fusions. Hyperphage not only ha s a higher display level of fu sion-pIII proteins, but it also packages phagemid particles over 100 times more efficiently than Ex-phage or Phaberge phage (68). Of the gIII gene-mutated helper phages, hyperphage is the most advantageous for use in phage display systems. Hyperphage enables multivalent display of pIII fusions, which makes it ideal in phagemid particle production. However, an issue with hyperphage is that the phagemid particle stocks generated usi ng hyperphage have titers of 109-10 phagemids/mL (64). This is a log or two lower than phagemid particles produced using helper phage. However, many of the phagemid particles generated by helper phage are useless because they bear no recombinant pIII proteins. Therefore, even though the quantity of phagemid particles prod uced with hyperphage is lower than those produced with helper phage the quality of phagemid particles produced is much higher. Helper Plasmids Helper plasm ids (64) are M13-based plasmids used for phagemid particle amplification that are engineered in three forms to overc ome many disadvantages of helper phage and hyperphage. The plasmids are M13mp19 with a chloramphenicol resistance gene cloned in from pBSL121 to allow for selection of E. coli containing the helper plasmids. The M13 origin of replication was deleted and replaced with the p15a origin of replication from pMPM-K3. This

PAGE 30

30 eliminates the ability of helper plasmids to be packaged in progeny phage particles, resulting in progeny phages that contain the phagemid s but not the helper plasmids. The gIII gene of the helper plasmid was partially delete d or fully deleted to yield thr ee helper plasmids with varying lengths of their gIII gene These three helper plasmids are maintained in E. coli and provide phagemid particles with all of the necessary structural proteins for phagemid particle amplification. One of the advantages of the helper plasmids is that they have a full (M13cp), deleted (M13cp-dg3), or truncated (M13cp-CT) gIII gene, which enables monovalent to multivalent display of recombinant proteins. Multivalent display phagemid particles possess high avidity binding ability, while monovalent display phagemid partic les possess high affinity binding ability. Thus, if high affinity monovalent display phagemid pa rticles are desired, M13cp would be used. If high avidity multivalent display phagemid partic les are desired, M13cp-dg3 or M13cp-CT might be used. An additional advantage is that the helper plasmids are maintained in E. coli. This negates the need for superinfection, removing the limiting factor of the number of helper phages or hyperphages needed to amplify phagemid particles. This thesis describes the efforts to optimi ze biopanning processes and associated reagents to improve the likelihood of is olating recombinant phages that specifically bind to biological agents or their products. The recombinant phages that specifically bind to biological agents or their products will eventually be used to enable detection of the agents. The specific aims for this study are: 1. To optimize extraction methods a nd determine saturation limits of V. cholerae O1 LPS and use these methods to pa n phage display libraries to V. cholerae LPS. A phenol-water method was a more effectiv e extraction method than TRIzol Reagent method for the extraction of V. cholerae O1 LPS and 10-100 times more LPS could be bound to nitrocellulose paper as opposed to a microtiter well. Biopannings against

PAGE 31

31 V. cholerae O1 LPS that was immobilized onto nitrocellulose paper or a microtiter well failed to yield phages that recognized V. cholerae O1 LPS. 2. To optimize panning and screening proce dures to allow for a more efficient biopanning process that will be more likely to isolate and detect specific recombinant phagemid particles. In biopanning with the Tomlinson scFv phage mid library, a trypsin elution time of 10-30 minutes eluted the most phagemid particles. Screening scFv protei ns by a high throughput ELISA was an acceptable screening process and enabled the possible screening of hundreds of clones per day. Screening phagemid particles by a high throughput ELISA was not an acceptable screening process because phagemid particles could not be produced in high enough titers in a mi crotiter well to make a high throughput ELISA screen effective. 3. To optimize the production of phagemid par ticles to ensure high quality and high quantity of phagemid particles. Phagemid particle amplification by E. coli harboring hyperphage or helper plasmids was not an effective method to produce phagemid particles. Homemade hyperphage that was produced in E. coli MG1655 (pGTR203) enabled high quality and high quantity production of hyperphage. With homemade hyperphage, the Tomlinson J library was effectively amplified to ensure that the re dundancy of the clone population in the library was maintained. 4. To isolate specific recombinant phagemid particles to E. coli O157:H7 Stx2 toxin using the optimization techniques discovered in previous aims. Using the improved homemade hyperphage and the homemade hyperphage-amplified Tomlinson J scFv library to pan against a commercial Stx2 toxin preparation resulted in isolation of phagemid particles that recognized the Stx2 toxin preparation.

PAGE 32

32 Figure 1-1. Structure of M13 phage. The M13 phage particle c ontains a ssDNA, circular, 6.4-kb genome that encodes genes I-XI and the M13 ori The M13 phage has five coat proteins: the major coat protein (pXIII) that is present in 2,700 copies and the minor coat proteins (pIII, pVI, pVII, and pI X) that are present in 5 copies each. pIII pVI pVIII pVII pIX M13 oriIV XI I VI X V VII IX XIII III II ~6.4-kb ssDNApIII pVI pVIII pVII pIX M13 oriIV XI I VI X V VII IX XIII III II ~6.4-kb ssDNA

PAGE 33

33 Figure 1-2. Structures of an tibodies. mAb (monoclonal an tibody), Fab (fragment antigen binding), scFv (single-chai n variable fragment), VH (variable heavy chain), VL (variable light chain), CH1-3 (constant heavy chain domains 1-3), and CL (constant light chain). The heavy and light chains of mAb and Fab antibodies are held together by disulfide bonds, and the heavy and light chains of the scFv antibody are fused together by a flexible glycine-serine linker. mAb FabscFvVHVHVHVLVLVLLinker CLCLCH1CH1CH3CH2

PAGE 34

34 Figure 1-3. Genetic map of pIT2 phagemid vect or from the Tomlinson scFv library. RBSribosome binding site. pelB leader peptide sequence promotes export of the scFv protein. Variable Heavy and Variable Light peptide seque nces are fused together by a glycine-serine linker. An amber stop codon is at the junction of the c-myc tag and the gIII gene to enable conditional expression of the scFv-pIII fusion in an amber suppressor strain. The M13 or igin of replication enable s packaging into M13 phage particles, the bla gene encodes ampicill in resistance, and th e colE1 origin of replication enables maintenance as a plasmid in E. coli. M13 ori pelB leader lac promoter colE1 ori

PAGE 35

35 CHAPTER 2 MATERIALS AND METHODS Bacterial Strains, Media, and Growth Methods The bacte rial strains used and their genotypes are listed in Table 2-1. All E. coli strains were grown in LB broth (10 g tryptone, 5 g yeast extract, 5 g NaCl, and 3 mL of 1 M NaOH in 1 L water) or on LB agar plates contai ning 1.5% (w/v) agar. Exceptions are E. coli TG1, E. coli DH5 (pNR100), and E. coli ER2738. Escherichia coli TG1 was grown in 2xTY broth (16 g tryptone, 10 g yeast extract, 5 g NaCl, and 3 mL of 1 M NaOH in 1 L water) or on 2xTY agar plates containing 1.5% (w/v) agar. Escherichia coli DH5 (pNR100) was grown in LB broth containing 100 g/mL ampicillin or on LB agar plates containing 100 g/mL ampicillin and 1.5% (w/v) agar. Escherichia coli ER2738 was grown in LB broth containing 20 g/mL tetracycline or on LB ag ar plates containing 20 g/mL tetracycline and 1 .5% (w/v) agar. All strains of V. cholerae were grown in Luria Bertani broth containing with physiological saline (LB-N) (10 g tryptone, 5 g yeast extract, 8.5 g Na Cl, and 3 mL of 1 M NaOH in 1 L water) or on LB-N agar plates containing 1.5% (w/v) agar The most common medium used was 2xTY containing 100 g/mL ampicillin and 1% (w/v) glucose; therefore, it was abbreviated as 2xTY AG. Escherichia coli strains harboring hyperphage were grown in broth or on plates as specified, with the addition of 40 g/mL kanamycin. Escherichia coli strains harboring helper plasmids were grown in brot h or on plates as specified above, with the addition of 30 g/mL chloramphenicol. All bacterial cultures were initially grown as a standing-overnight culture. A standingovernight culture was made from 10 mL of specifi ed medium that was inoculated with bacteria from an agar plate. The standing-overnight culture was grown overnight (~16 h) in a 37C incubator. A log-phase culture was obtained by diluting a sta nding-overnight culture 1:40 in

PAGE 36

36 medium and incubating the culture in a 37C incu bator with shaking until the optical density at 600 nm (OD600) was between 0.4 to 0.6. The Tomlinson J Human Synthetic VH + VL phagemid library was constructed by Medical Research Council, Cambridge, U.K. (Human Single Fold scFv Libraries I + J (Tomlinson I + J). 2002. Cambridge, UK, MRC La boratory of Molecula r Biology, MRC Centre for Protein Engineering.) and was obtained from the Interdisciplinary Center for Biotechnology Research Hybridoma Core, University of Flor ida. The Ph. D. 12mer peptide library was obtained from New England Biolabs (NEB, Ipswich, MA). The hyperphage (66) used to amplify phagemid particles was obtained from Progen Biotechnik (Heidelberg, Germany). Biopanning of Phage Display Libraries Panning on Immunotubes A polystyrene imm unotube (Nunc, Rochester, NY) was coated with 2 mL of 50 g/mL Shiga-like toxin 2 (Stx2 (Toxin Technology, Sara sota, FL)) in phosphate-buffered saline (PBS) (Cellgro, Manassas, VA) (137 mM NaCl, 2.7 mM KCl, 4.3 mM Na2HPO4, 1.4 mM KH2PO4, pH 7.3). The immunotube was rotated overnight on a la bquake at 4C. The next day the tube was washed three times with 4 mL of PBS to remove non-immobilized toxin. The tube was filled with casein blocker (Pierce, Rockford, IL) containing 0.05% (v /v) Tween-20 and incubated for two hours at room temperature (~25C). Excess blocker was removed by rinsing the tube three times with 4 mL of PBS. The Tomlinson J library (for round one of panning) or amplified eluted phages (for subsequent rounds of panning) was added to the immunot ube at a concentration of 1 x 1012 phages in 4 mL of casein blocker (Pierce) containing 0.05% (v/v) Tween-20. The library was incubated for one hour at room temperat ure on a labquake, followed by a one hour standing incubation. The unbound library was removed by aspiration, and the weakly bound phages were removed by washing the tube 10 times for round one and 20 times for rounds two and three with

PAGE 37

37 4 mL of PBS containing 0.05% (v/v) Tween-20 (PBS-0.05T). The excess PBS-0.05T was aspirated, and bound phages were eluted with 0.5 mL of trypsin-PBS (10% (v/v) trypsin stock (10 mg/mL trypsin (Type XIII from Bovine Panc reas) (Sigma-Aldrich, St. Louis, MO), 50 mM Tris-HCl (pH 7.4), 1 mM CaCl2 in water) in PBS) for 10 minut es at room temperature on a labquake. Ten microliters of the eluted phages were immediately titered, while the rest of the eluted phages were infected into 5 mL of E. coli TG1 at an OD600 of 0.4. The infected E. coli was incubated for 30 minutes at 37C in a stan ding water bath. Following incubation, the infected cells were isolated by centrifugation at 13,776 x g for 10 minutes at room temperature. Cells were suspended in 0.6 mL of 2xTY medi um and plated on three 2xTY AG plates to amplify the phagemid-containing E. coli. The plates were incubate d overnight at 37C. The next day 1.5 mL of 2xTY was added to the lawn s of bacteria, and the bacterial lawns were scraped from the plate in the medium and pool ed together. Of the pooled bacteria, 50 L was inoculated into 100 mL of 2xT Y AG medium and grown in a 37C shaking incubator until the OD600 was 0.4. Once the OD was reached, 10 mL of the culture was superinfected with homemade hyperphage at a MOI of 20 and incuba ted for 30 minutes in a standing 37C water bath. The superinfection was cen trifuged at 13,776 x g for 10 minutes at room temperature. The resulting pellet was suspended in 100 mL of 2xTY containing 100 g/mL ampicillin, 40 g/mL kanamycin, and 0.1% (w/v) glucose and incubated overnight in a shaking 30C water bath. The next day the culture was centr ifuged at 13,776 x g for 10 minutes at 4C. Phage-containing supernatant was precipitated with final concentra tions of 13 mM Polyethylene Glycol (PEG) and 0.55 M NaCl. PEG and NaCl were dissolved in th e supernatant at room temperature and then the PEG-containing supernatant was incubated overnight at 4C. The next day the PEGprecipitated supernatant was cen trifuged at 13,776 x g for 20 minutes at 4C. The pellet was

PAGE 38

38 suspended in 2 mL of PBS and centrifuged at 13 ,776 x g for 5 minutes at room temperature to remove bacterial debris. The supe rnatant was transferred to a new tube to be titered and stored at 4C until used in further rounds of panning or screening. Panning in Suspension Vibrio cholerae N16961 was grown to log phase, and 107 to 109 cells were centrifuged at 10,621 x g for 5 minutes at room temperature. Th e bacterial pellet was su spended in 1 mL of PBS and transferred to a microcentrifuge tube The Tomlinson J library (for round one of panning) or amplified eluted phages (for subs equent rounds of panning) was added to the bacterial suspension at a concentration of 1 x 1011 phages in 0.5 mL of PB S. The library and bacterial cells were incubated for 3 hours at 4C while rota ting on a labquake. The cell suspension was centrifuged at 10,621 x g for 5 minutes at room temperature. The supernatant, containing the unbound phage library, was stored at 4C. The bacteria l pellet containing the bound phages was suspended in 0.5 mL of PBS to wa sh off phages that were weakly bound to the bacteria. The suspension was centrifuged at 10,621 x g for 5 minutes at room temperature. The supernatant was discarded, and the pellet was suspended in 0.5 mL of PBS for a second wash. The suspended cells were transferred to a new microcentrifuge tu be and centrifuged at 10,621 x g for 5 minutes at room temperature. The supernatant was aspirated, and phages bound to the bacteria were eluted with 0.5 mL of tryps in-PBS for 10 to 60 minutes at room temperature on a labquake. Ten microliters of the elution was immediately titered. The rest of the eluted phages were infected into 5 mL of E. coli TG1 at an OD600 of 0.4. The infected E. coli were incubated for 30 minutes in a 37C standing water bath. Following incubation, the infected cells were isolated by centrifugation at 13,776 x g for 10 mi nutes at room temperature. The cells were suspended in 0.6 mL of 2xTY medium and plat ed on three 2xTY AG plates to amplify the phagemid-containing bacteria. Th e plates were incubated overni ght at 37C. The next day 1.5

PAGE 39

39 mL of 2xTY was added to the lawns of bacteria, and the bacterial lawns were scraped from the plate in the medium and pooled toge ther. Of the pooled bacteria, 50 L were inoculated into 50 mL of 2xTY AG medium and grown in a 37C shaking incubator till the OD600 was 0.4. Once the OD was reached, 10 mL of the culture was superinfected with homemade hyperphage at a MOI of 20 and incubated for 30 minutes in a 37 C standing water bath. The superinfection was centrifuged at 13,776 x g for 10 minutes at room temperature. The resulting pellet was suspended in 50 mL of 2xTY containing 100 g/mL ampicillin, 40 g/mL kanamycin, and 0.1% (w/v) glucose and incubated overnight in a 30C shaking water bath. The next day the culture was centrifuged at 13,776 x g for 10 minutes at 4C. The phage-containing supernatant was precipitated by slowly adding 12.5 mL of PEG solution (20% (w/v) PEG and 2.5 M NaCl in water) to the supernatant with continuous swirling. The PEG-containing supernatant was incubated overnight at 4C. The next day th e PEG-containing supernatant was centrifuged at 13,776 x g for 20 minutes at 4C. The pellet was suspended in 1 mL of PBS and centrifuged at 10,261 x g for 5 minutes at room temperature to re move bacterial debris. The supernatant was transferred to a new tube to be titered and stored at 4C until used in further rounds of panning or screening. Panning on Microtiter Wells A polysorp m icrotiter well (Nunc) was coated with 150 L of 100 g/mL V. cholerae 569 lipopolysaccharide (LPS) (Sigma-Aldri ch) in carbonate coating buffer ( 0.1 M NaHCO3, 0.02% (w/v) NaN3 (pH 8.6) ). Prior to coating, the LPS was sonicated by a Vibra Cell (Sonics & Materials Inc., Danbury, CT) for one minute. The microtiter well was coated overnight at 4C in a humid chamber. The next day the microtiter we ll was equilibrated to room temperature for 15 minutes. The LPS was aspirated and blocked with 200 L of carbonate blocking buffer for one hour at 4C. The well was washed six times with 200 L of 0.1 M Tris-buffered saline (TBS)

PAGE 40

40 (81 mM Tris-HCl, 20 mM Tris-Base, 154 mM Na Cl, pH 7.5) containi ng 0.1% (v/v) Tween-20 (TBS-0.1T) for the first round of panning and with 200 L of 0.1 M TBS cont aining 0.5% (v/v) Tween-20 (TBS-0.05T) for subsequent rounds of panning. The washed wells were incubated with 2 x 1011 PFU of the Ph.D.-12mer peptide librar y (NEB) (for round one of panning) or amplified phages (for subsequent rounds of panning) diluted in 100 L of TBS-0.1T for one hour at room temperature with gentle rocking. The unbound phages were aspirated and stored at 4C. The wells were washed 10 times with TBS-0.1T for the first round of panning and with 200 L of TBS-0.05T for subsequent rounds of pa nning. The phages were eluted with 100 L of 0.1 M glycine (pH 2.2) for 10 minutes at room temper ature with gentle agitation. The glycine containing the eluted phages was transferred to a microcentrifuge tube and neutralized with 15 L of Tris-HCl (pH 9.0). Five microliters of the eluted phages were immediately titered. One hundred and ten microliters of the remaining el ution was added to 20 mL of LB diluted 1:100 with a standing-overnight culture of E. coli ER2738. The culture was grown for 4.5 hours in a 37C shaking incubator. The turbid culture wa s centrifuged at 13,776 x g for 10 minutes at 4C. The phages were precipitated by slowly adding 3.3 mL of PEG so lution (20% (w/v) PEG and 2.5 M NaCl in water) to the phage-containing su pernatant with continuous swirling. The PEGcontaining supernatant was incubated overnight at 4C. The next day the PEG-containing supernatant was centrifuged at 13,776 x g for 20 minut es at 4C. The pellet was suspended in 1 mL of 0.1 M TBS and centrifuged at 10,621 x g for 5 minutes at room temperature to remove bacterial debris. The phages were preci pitated a second time by slowly adding 167 L of PEG solution to the supernatant with continuous swirling. The PEG-containing supernatant was incubated one hour at 4C and centrifuged at 13, 776 x g for 15 minutes at 4C. The pellet was suspended in 200 L of 0.1 M TBS to be titered and stored at 4C until used in further rounds of

PAGE 41

41 panning or screening. Five rounds of panning were done with screening of clones after the fifth round of panning. Panning on Nitrocellulose Paper Six pieces o f nitrocellulose paper (Bio-Rad) were cut to the surface-area dimensions of a microtiter well (5 mm x 20 mm). One a nd a half milliliters of approximately 340 g/mL V. cholerae 569 LPS (phenol-water-extracted ) in PBS was added to si x microcentrifuge tubes, each containing a strip of nitrocellulose paper. Prior to coating, the LPS was sonicated by a Vibra Cell (Sonics & Materials Inc.) for one minut e. The nitrocellulose papers were coated overnight at 4C on a rotating labquake. The next day two pieces of LPScoated nitrocellulose paper were each transferred to a new microcentrifuge tube and washed five times with 1 mL of PBS on a labquake with five minutes per wash. Wash ed nitrocellulose strips were transferred to new microcentrifuge tubes and bl ocked for one hour in PBS casein blocking blocker (Pierce) on a labquake at room temperature. Blocked nitrocellulose stri ps were transferred to new microcentrifuge tubes and washed five times with 1 mL of PBS on a labquake with five minutes per wash. Each of the washed pieces of nitrocellulose papers was transferred to a new microcentrifuge tube containing 1.5 x 1011 PFU of the Ph.D.-12mer peptide library (NEB) diluted in 1 mL of PBS. Nitroc ellulose strips were incubated with the library for one hour at room temperature on a labquake. After one hour, the panned nitrocellulose strips were transferred to new microcentrifuge tubes and washed five times with 1 mL of PBS on a labquake with five minutes per wash to remove the unbound library. Washed nitrocellulose strips were transferred to new microcentrifuge tubes to be eluted. One pi ece of nitrocellulose was acid eluted by adding 250 L of 0.1 M glycine (pH 2.2) to the stri p and incubating it for 10 minutes at room temperature on a labquake. The acid eluted nitrocellulose paper was removed to a new microcentrifuge tube, and the glyc ine solution containing the eluted phages was neutralized with

PAGE 42

42 37 L of 1 M Tris-HCl (pH 9.0). The acid eluted phages were stored at 4C until used for amplification. The other piece of panned nitroc ellulose paper was antigen eluted by adding 250 L of approximately 340 g/mL of V. cholerae 569B LPS diluted in PBS to the piece of nitrocellulose paper and incubated one hour at room temperature on a labquake. The piece of nitrocellulose paper was removed, and the rema ining LPS solution containing the eluted phages was acid eluted as described above. Five microl iters of each eluted phage solution was titered, and the rest of the two elutions were each adde d to 20 mL of LB diluted 1:100 with a standingovernight culture of E. coli ER2738. The cultures were grown for 4.5 hours in a 37C shaking incubator. The turbid cultures were centrifuged at 13,776 x g for 10 minutes at 4C. Five microliters of the amplified phages were titered. The rest of the phages were PEG precipitated by slowly adding 3.3 mL of PEG solution to the phage-containing supern atant with continuous swirling. The PEG-containing supern atant was incubated overnight at 4C. The next day the PEG-containing supernatant was ce ntrifuged at 13,776 x g for 20 minutes at 4C. The pellet was suspended in 1 mL of PBS and centrifuged at 13 ,776 x g for 5 minutes at room temperature to remove bacterial debris. The supernatants were transferred to new microc entrifuge tubes, titered, and stored at 4C. After the first round of panning, the amplifie d eluted phages were negatively panned on. Two pieces of nitrocellulose paper (5 mm x 20 mm) were blocked for one hour in 1 mL of PBS casein blocking buffer (Pierce) on a labquake at room temperature. The blocked nitrocellulose strips were washed three times with 1 mL of PBS on a labquake at room temperature with five minutes per wash. One piece of blocked nitroc ellulose paper was added to the acid-elutedamplified phages from the round one panning an d incubated one hour on a labquake at room temperature. The other strip of blocked nitrocellulose paper was added to the LPS and acid-

PAGE 43

43 eluted-amplified phages from the round one panni ng and incubated one hour on a labquake at room temperature. The negatively panned pieces of nitrocellulose paper were removed, and the remaining amplified-eluted phages were used for two more rounds of panning. After the third round of panning the eluted phages were screen by ELISA. Spot Titer of Phages Escherich ia coli TG1 was grown to log phase an d centrifuged at 13,776 x g for 10 minutes at 4C. The resulting pellet was suspended in PBS to yield an E. coli TG1 concentration of 1 x 1010 CFU/mL. One hundred microlit ers of the concentrated E. coli TG1 was spread on a 2xTY AG plate and set to dry for 30 seconds. Serially diluted phages in phosphate-buffered saline containing 0.1% (w/v) gelatin (BSG) was dropped onto the plate in 10 L drops. Once the drops dried, the plate was incubate d overnight at 37C. The next day the colonies were counted, and the approximate titer was calculated. Spread Titer of Phages Phages were serially diluted in BSG. One hundred m icroliters of the diluted phages were added to 0.9 mL of log phase E. coli TG1. The infected E. coli TG1 was incubated 20 minutes in a 37C standing water bath. One hundred microlit ers of the infections were plated on 2xTY AG plates. The plates were incubated overnight at 37C. The next day the colonies were counted, and the approximate titer was calculated. Amplification of Phages A colony of E. coli TG1 containing a phagem id was picked from a plate with a sterile toothpick and swirled in 3 mL of 2xTY AG medium. The culture was grown overnight at 37C. The overnight culture was diluted 1:20 in 3 mL of 2xTY AG medium and grown to log phase in a 37C shaking incubator. Three hundred microlit ers of the log-phase cult ure was infected with hyperphage at a MOI of 10 and incubated in a 37 C standing water bath for 30 minutes. The

PAGE 44

44 culture was added to 30 mL of 2xTY medium containing 100 g/mL ampicillin, 40 g/mL kanamycin, and 0.1% (w/v) glucose. The cultu re was grown overnight in a 30C shaking incubator. The turbid culture was centrifuge d at 13,776 x g for 10 minutes at 4C. The phagecontaining supernatant was tit ered and stored at 4C. High Throughput Production of Soluble An tibody Fragments (scFv antibodies) A colony of E. coli HB2151 containing a phagemid was picked from a plate with a sterile toothpick and swirled in 200 L of 2xTY containing 100 g/mL ampicillin, 0.1% (w/v) glucose, and 1 mM Isopropyl -D-1-thiogalactopyranoside (IPTG) in a 96-well polystyr ene microtiter plate (Corning, Corning, NY). The plate was incubated overnight at 37C. The next day the plate was centrifuged at 4,667 x g for 10 minutes at 20C. Supernatants were analyzed by ELISA. Deoxyribonucleic Acid Manipulations Plasmid Extractions Plasm id extractions for cultures of 3 mL were performed with the QIAprep Spin Miniprep kit (Qiagen, Germantown, MD), while cultures of 100 mL or greater were performed with the Plasmid Midi kit (Qiagen). Extrac tion procedures were performed as directed in the instruction manual. Agarose Gel Electrophoresis Deoxyribonucleic acid sam ples from 1-5 L were added to 2 L of 10 x Gel Loading Buffer (Invitrogen, Carlsbad, CA). The samples we re loaded onto a 0.7% (w /v) agarose gel using Tris-borate-EDTA buffer (89 mM Tris, 89 mM boric acid, 2 mM EDTA) containing 10 g/mL ethidium bromide. The gel was electrophoresed at 100 V until the loadin g dye was two-thirds down the gel. The DNA bands were analyzed on a Gel Doc XR (Bio-Rad, Hercules, CA).

PAGE 45

45 Electroporation of Plasmids Electroporation was perform ed to transform plasmids into E. coli. DNA (30-1000 ng/ L) was added to 50 L of electrocompetent cells (see below) in a microcentrifuge tube. The tube was flicked to mix the DNA with the cells. Once the DNA and cells were mixed, they were transferred to a chilled 0.1 cm electroporation cuvette and elec troporated at 1.25 kV/cm with a MicroPulser (Bio-Rad). Nine hundred and fift y microliters of 2xTY medium was immediately added to the transformation and tr ansferred to a small culture tube where it was incubated in a 37C static water bath for 1 hour. The electropora tion was serially diluted in BSG and plated on appropriate plates. The pl ates were incubated overnight at 37C The next day the colonies were counted, and the number of transformations was calculated. Electrocompetent Cells Electrocom petent cells were made by diluting a standing overnight culture 1:100 in 1 L of medium containing 1% (w/v) glucose and appropriat e antibiotics. The culture was incubated in a 37C shaking incubator until the OD600 was between 0.4 and 0.6. All centrifuge rotors, wash buffers, and centrifuge tubes that were used were pre-chilled to 4C. Once the OD600 was between 0.4 and 0.6, the culture was chilled on ice in a 4C cold room for 45 minutes or more until the culture was ~4C. The culture was centr ifuged at 13,776 x g for 10 minutes at 4C, and the supernatant was decanted. The pelle ts were suspended in 1.2 L chilled ddH2O by vortexing or pipeting until the cell suspension was homogeneous. The suspended pellets were centrifuged to pellet the bacterial cells. The supernatant was decanted, and the pellets were suspended in 400 mL of chilled ddH2O. The bacteria were pelleted by centrifugation, and the supernatant was decanted. The pellets were suspended in 200 mL of chilled ddH2O into one centrifuge tube. The bacteria were pelleted by centrifugation, and the supernatant was d ecanted. The pellet was suspended in 25 mL of chilled ddH2O and transferred to a 35 mL ce ntrifuge tube. The bacteria

PAGE 46

46 were pelleted by centrifugation, and the supernatan t was decanted. The pellet was suspended in 25 mL of chilled 10% (w/v) gluc ose in water and centrifuged to pellet the bacteria. The supernatant was decanted, and the bacteria were suspended in 0.5-1 mL of chilled 10% (w/v) glucose in water. The suspended bacteria were either used directly for transformation or aliquoted and frozen on dry ice in 95 % (v/v) EtOH and stored at -80C. Enzyme-Linked ImmunoSorbent Assays (ELISAs) Enzym e-linked immunosorbent assays we re performed for the screening and characterization of phage partic les, scFvs, LPS, bacterial pr otein preparations, monoclonal antibodies, and polyclonal antibodies. The co ating antigen or antibody was diluted in PBS (Cellgro) or carbonate/bicarbonate buffer (15 mM Na2CO3 and 35 mM NaHCO3 in water), and 100 L was loaded onto a 96-well polystyrene flatbottomed microtiter plate (Becton-Dickinson, Franklin Lakes, NJ). Concentr ations used of whole cells (108 tu/mL), LPS (1 g/mL), toxin (10 g/mL), and proteins (10 g/mL) were kept relatively consiste nt. The coated microtiter plate was incubated overnight in a humid chamber at 4C. The next day the microtiter plate was equilibrated to room temp erature for 30 minutes. Wells were aspirated with an ELX 800 Strip Washer (Bio-Tek, Winooski, VT) and washed once with 300 L of PBS. If the coating antigen was whole cells, LPS, or toxin, then the coati ng antigen was removed by vacuum or pipet and sterilized. Two hundred microliters of PBS casein blocking buffer (Pierce) containing 0.05% (v/v) Tween-20 was added to each well and incubate d two hours at room temperature. Blocker was aspirated by the ELX 800 Strip Washer (Bio-Tek). On e hundred microliters of primary antibody diluted in PBS casein blocking buffer (Pierce) containing 0.05% (v/v) Tween-20 was added to each well and incubated one hour at ro om temperature. Concentrations used of phagemid particles were 107 to 109 tu/mL and antibodies were 1 to 10 g/mL. The primary antibody was aspirated and washed three times with 300 L of PBS-0.05T by the ELX 800 Strip

PAGE 47

47 Washer (Bio-Tek). One hundred microliter s horseradish peroxidase (HRP) conjugated secondary antibodies were diluted in PBS casein blocking buffer (Pierce) containing 0.05% (v/v) Tween-20 and added to each well an d incubated 30 minutes at room temperature. The secondary antibody was aspirated and washed three times with 300 L of PBS-0.05T by the ELX 800 Strip Washer (Bio-Tek). Substrate was prepared by dissolving one capsule of phosphate-citrate (PC) buffer (0.05 M phosphate-citrate buffer (pH 5.0) 0.03% (w/v) sodium perborate) (SigmaAldrich) in 100 mL of water. A ten milligram tablet of 3, 3, 5, 5-tetramethylbenzidine substrate (Sigma-Aldrich) was added to 10 mL of th e PC buffer to give a final concentration of 1 mg/mL. Two hundred microliters of substrate was added to each well and incubated for 30 minutes at room temperature. The plate was re ad in an ELx 800 UV plate reader (Bio-Tek) at 630 nm. The data were analyzed with KcJunior (Bio-Tek) software. Infection Efficiency Infection efficiency experim ents were done to compare the infection efficiency of phagemid particles to bacterial strains harboring helper plasmids to the same strains not harboring helper plasmids. Escherichia coli standing-overnight cultures were diluted 1:40 in 3 mL of medium containing 1% (w/v) glucose for strains containing no helper plasmids and in medium containing 1% (w/v) glucose and 30 g/mL chloramphenicol for strains containing helper plasmids. The diluted cultures were grown in a 37C shaking incubator till the OD600 was between 0.4 and 0.6. Once the desired OD600 was obtained, 3 x 107 bacteria were infected with phage at a MOI of 0.1 and incubated for 20 mi nutes in a 37C standi ng water bath. The transductions were titered using th e spread-titer method. Transduced E. coli strains containing helper plasmids were titered on 2xTY AG plates containing 30 g/mL chloramphenicol. Transduced E. coli strains containing no helper plasmi ds were titered on 2xTY AG plates.

PAGE 48

48 Protein and Lipopolysaccharide (LPS) Manipulations Phenol-Water Extraction of LPS A phenol-water extraction (76) was perform ed to extract V. cholerae 569B LPS from whole cells. One and a half liters of V. cholerae 569B was grown to log phase in a 37C shaking incubator. Bacteria were pell eted at 13,776 x g for 10 minutes at 4C and suspended in 10 mL of 1% (w/v) NaCl in water. The bacteria were ce ntrifuged and suspended in another 10 mL of 1% (w/v) NaCl in water. Next, the bacteria were centrifuged and susp ended in 10 mL of ddH2O. This suspension was warmed to 70C, along with 24 mL of 90% (w/v) ph enol (made directly before extraction). Ten milliliters of warmed 90% (w/v) phenol was added to the 10 mL cell suspension and vortexed. The twenty milliliter suspension was warmed in a 70C water bath for 20 minutes while vortexing frequently. The phenol suspension was transferred to an ice bath where it was swirled in ice water for 5 minutes The phenol suspension was centrifuged at 2,284 x g for 25 minutes at 4C. The aqueous phase of the phenol extraction was transferred to a new centrifuge tube and stored on ice. The same vol ume of water as that extracted was added back into the phenol suspension tube and heated to 70C for 20 minutes with frequent vortexing. The extraction step was repeated, and the two aqueous phase extractions were combined into the same tube. Aqueous extractions were placed in a 70C water bath for 15 minutes. Twelve milliliters of warmed 90% (w/v) phenol was added to the warmed aqueous extraction and incubated at 70C for 10-15 minut es with frequent vortexing. Two back extractions were performed on the aqueous extractions. The final extraction was dialyzed (MWCO: 3,500) for 24-48 hours at 4C. The dialyzed extraction was digested with DNase (Qiagen) (20 g DNase/mL dialysis volume) and RN ase (Qiagen) (40 g RNase/mL dialysis volume) in MgCl (1 L 20% (w/v) MgCl/mL dialysis volume) overni ght at 37C. The di gested extraction was

PAGE 49

49 dialyzed (MWCO: 3,500) for 24-48 h ours at 4C. The final extrac tion was used in a Tsai-Frasch silver stain to analyze the LPS. TRIzol Reagent Extraction of LPS TRIzol Rea gent (Invitrogen), a monophasic solution of phenol and guanidine isothiocyanate (77), was used to extract V. cholerae N16961 LPS from whole cells. Ten milliliters of log phase (OD600 ~ 0.4-0.6) V. cholerae N16961 were centrifuged at 13,776 x g for 10 minutes at 4C. The supernatant was decante d, and the cells were suspended in 200 L of TRIzol Reagent and incubated at room temperat ure for 10 to15 minutes. Twenty microliters of chloroform per mg of cells was added to th e mixture, vortexed, and incubated at room temperature for 10 minutes. The mixture was ce ntrifuged at 12,000 x g for 10 minutes at room temperature. The resulting aqueous phase was transferred to a new mi crocentrifuge tube and stored on ice. One hundred microliters of ddH2O was added to the TRIzol-Reagent tube, and a back extraction was performed. Two more back extractions were performed following the first back extraction to give a total of four extractio ns. The extractions were dried with a Savant SpeedVac Concentrator (Globa l Medical Instrumentation, Ramsey, MN) for 2.5 hours. The pellet was suspended in 0.5 mL of 0.375 M MgCl in 95% (v/v) EtOH that was chilled to 4C. The suspension was centrifuged at 12,000 g for 15 mi nutes at room temperature. The resulting pellet was suspended in 200 L of ddH2O. The final LPS extracti on was analyzed in a TsaiFrasch silver stain (see below). Extraction of Periplasmic Proteins Periplasmic proteins from E. coli were extracted using a Tris-EDTA-Sucrose (TES) extraction. A standing-overnight culture of an E. coli culture was diluted 1: 400 in 100 mL of LB medium and incubated overnight in a 37C shaki ng incubator. The next day the cells were centrifuged at 13,776 x g for 10 minutes at 4C. Th e supernatant was stored at 4C. The pellet

PAGE 50

50 was suspended in 25 mL of PBS by vortexing. The suspension was centrifuged at 13,776 x g for 10 minutes at 4C. The supernatant was discar ded, and the pellet was suspended in 10 mL of TES (20% (w/v) D-sucrose, 30 mM Tris-HCl, 1 mM EDTA) in water. Th e suspended pellet was incubated for 15 minutes at room temperature with occasional swirling. The suspension was centrifuged at 13,776 x g for 10 minutes at 4C. The supernatant was discarded, and the pellet was gently washed with 10 mL of 0.5 mM MgCl2 without dislodging the pe llet. The 10 mL of the 0.5 M MgCl2 wash was discarded, and the pellet wa s suspended by vortexing in 10 mL of 0.5 mM MgCl2. The suspension was incubated in an ice bath for 15 minutes with occasional swirling. The suspension was centrifuged at 21,525 x g for 30 minutes at 4C. The supernatant was collected and concentrated with 10,000 MWCO Amicon (Millipore, B illerica, MA) filters and stored at 4C. The pellet was suspende d in 10 mL of PBS and stored at 4C. Determination of Protein Concentration Protein concentration was determ ined by DC-Protein Assay (Bio-Rad). Protein samples were diluted in PBS. Five microliters of each sample was added to a microtiter well followed by 25 L of Reagent A and 200 L of Reagent B. All wells were performed in triplicate. Protein standards used were Bovine Serum Albumin (BSA ) in concentrations from 0-1.4 mg/mL. The reaction took place for 15 minutes at room temper ature and was read at 750 nm with an ELx 800 UV plate reader (Bio-Tek). Protein concentratio ns were analyzed using KC Junior (Bio-Tek) software. Sodium Dodecyl Sulfate-Polyacrylami de Gel Electropho resis (SDS-PAGE) Sodium dodecyl sulfate-polyacrylamide gel electrophoresis was performed to resolve V. cholerae LPS and protein preparations ReadyGel Tris-glycine 12% (w/v), 15% (w/v), and 420% (w/v) PAGE gels (Bio-Rad) were used with the MiniProtean-Electr ophoresis system (BioRad) for SDS-PAGE analysis of samples. The LPS samples were sonicated by a Vibra Cell

PAGE 51

51 (Sonics & Materials Inc.) for 2 minutes and d iluted in Laemmli sample buffer (Bio-Rad). Protein samples were vortexed and diluted in La emmli sample buffer. After dilution in Laemmli sample buffer, the samples were boiled for 10 minut es and loaded in the gels. The samples were electrophoresed for 1 hour at 100 Volts in elec trode buffer (25 mM Tris, 0.17 M glycine, and 0.1% (w/v) SDS). Coomassie Blue Staining Proteins resolved by SDS-PAGE we re stained with 0.05% (w/v) Coomassie Blue R, 50% (v/v) methanol, and 10% (v/v) glaci al acetic acid in water and ag itated gently for 10 minutes. The gel was rinsed with ddH2O to remove excess stain. The ge l was destained with 5% (v/v) methanol and 7% (v/v) glacial ac etic acid in water overnight while gently shaking. The next day a picture was taken of the gel, and the gel was dried. Tsai-Frasch Silver Staining Tsai-Frasch silver staining ( 78) was performed on SDS-PAGE resolved LPS that was extracted from V. cholerae. Carbohydrates resolved by SDS-P AGE were fixed in 40% (v/v) ethanol and 5% (v/v) glacial acetic acid in water overnight. Carbohydrates were oxidized by incubating the fixed gel in 0.7% (w/v) periodic acid, 40% (v/v) ethanol, and 5% (v/v) glacial acetic acid in water for five minutes. The ge l was washed three times in 500 mL of ddH2O with 15 minutes per wash. Reagent A (1 % (v/v) 10 M NaOH and 6.7% (v/v) NH4OH in water) and Reagent B (20% (w/v) AgNO3 in water) were made immediatel y before staining. The stain was made by drop wise titration of Reagent B into Re agent A until a hazy brown color to the solution appeared and remained. The titrated r eagent was diluted in 115 mL of ddH2O to make the final stain solution. The gel was placed in the stai ning solution and stained for 10 minutes with agitation. The gel was washed three times in 500 mL of ddH2O with 15 minutes per wash. The gel was developed with 1 x Developer (5 x Tsai -Frasch developer: 1.3 mM citric acid and 0.25%

PAGE 52

52 (v/v) of 37% (v/v) formaldehyde in ddH2O) for 3 to 10 minutes with constant gentle shaking. The developed gel was soaked in 1% (v/v) acet ic acid in water for 10 minutes and transferred into 50 mL of ddH2O. A picture was taken of the gel, and then the gel was dried. Western Blot Proteins resolved by SDS-PAGE we re tran sferred onto a nitrocellulose membrane, followed by reaction with antibodies and develo pment with enhanced chemiluminescent (ECL) (Pierce). A Mini-TransBlotting cell (Bio-Rad) wa s used for the transfer of proteins to the nitrocellulose membrane. The transfer buffer (25 mM Tris, 192 mM glycine, and 20% (v/v) methanol) was pre-chilled to 4C. The transf er occurred at 100 V for one hour at 4C. Transferred proteins were analyzed by Ponceau S staining. Fifteen milliliters of 0.3% (w/v) Ponceau S dissolved in 3% (v/v) trichloroace tic acid (TCA) was adde d to the transferred membrane and incubated 10 minutes at room te mperature with gentle agitation. The stained membrane was gently washed in 25 mL of dH2O to remove excess stain. The membrane was photographed to view the stained proteins. The membrane was blocked in 25 mL of PBS casein blocking buffer (Pierce) contai ning 1% (v/v) Tween-20 for two hours at room temperature or overnight at 4C with gentle shaking. Primary antibodies were diluted in PBS casein blocking buffer (Pierce) containing 0.05% (v/v) Tween-20 to give a final volume of 10 mL and were added to the blocked membrane for one hour at room temperature with gentle shaking. The membrane was washed three times with 20 mL of PBS-0.05T while shaking for 15 minutes for each wash. Secondary antibodies conjugated w ith HRP were diluted in PBS casein blocking buffer (Pierce) containing 1% (v /v) Tween-20 to give a final volume of 10 mL and were added to the membrane for 30 minutes at room temperat ure with gentle shaking. The membrane was washed three times with 20 mL of PBS-0.05T while shaking for 10 minutes for each wash. The

PAGE 53

53 blot was incubated in 8 mL of ECL substrate (Pierce) for three minutes at room temperature and developed on CL-XPosure Film (Pierce) by an X-omat 2000 processor (Kodak, Rochester, NY). Lipopolysaccharide Saturation to Nitroce llulose Paper The LPS saturation experiment to nitrocellu lose paper was performed to determine the concentration of V. cholerae 569B LPS needed to saturate a piece of nitrocellulose paper. Phenol-water-extracted V. cholerae 569B LPS was serially diluted from 100 to 1 g/mL in PBS. Twelve pieces of nitrocellulose paper were cut to dimensions of 5 mm x 20 mm. Nitrocellulose strips were put in a microcentrifuge tube containing 256 L of the LPS dilutions or PBS for a control. LPS and PBS coated stri ps were done in duplicates in sepa rate tubes. The strips were coated overnight at 4C on a labquake. Coated ni trocellulose strips we re transferred to new microcentrifuge tubes. Strips were washed 3 times with 1 mL of PBS on a labquake with 5 minutes per wash. Strips were blocked with 1 mL of casein blocking buffer (Pierce) for one hour at room temperature on a labqua ke. Strips were incubated with 0.5 mL of rabbit anti-V. cholerae O1 LPS polyclonal antibody (Accurate Chemical, We stbury, NY) at a dilution of 1:400 in casein blocking buffer (Pierce) for one hour at room te mperature on a labquake. Strips were washed 3 times with 1 mL of PBS on a labqua ke with 5 minutes per wash. Strips were incubated with 0.5 mL of goat anti-rabbit peroxi dase conjugated monoclonal an tibody (Jackson ImmunoResearch Laboratories, West Grove, PA) at a concentratio n of 1:1000 in casein blocking buffer (Pierce) for 30 minutes at room temperature on a labquake. St rips were washed 3 times with 1 mL of PBS on a labquake with 5 minutes per wash. Strips were incubated in 5 mL of ECL substrate (Pierce) for three minutes at room temperature and de veloped on CL-XPosure Film (Pierce) by an Xomat 2000 processor (Kodak).

PAGE 54

54 Colony Blot with scFv To test whether bacteria produce scF v that ar e specific to a target molecule, a colony blot was developed to screen scFv in a high thr oughput manor. A Petri dish size piece of nitrocellulose paper (Bio-Rad) wa s coated overnight at 4C with 10 mL of target antigen at a concentration that the target an tigen was panned. The next day th e coated nitrocellulose paper was washed three times with 15 mL of 0.1 M TB S with five minutes per wash on a labquake. The nitrocellulose paper was then blocked for one hour at room temperature with 10 mL of casein blocker (Sigma-Aldrich) on a labquake. Afte r being blocked, the nitrocellulose paper was overlaid onto a plate containing bacterial colonies and stamped with a metal-plate stamp. The bacterial-colony-containing nitr ocellulose was washed three times with 15 mL of 0.1 M TBS with five minutes per wash on a labquake. The nitrocellulose paper was incubated with Protein L-peroxidase (Sigma-Aldrich) that was dilute d 1:2,000 in 10 mL of casein blocker (SigmaAldrich) for one hour at room temperature on a labquake. The nitr ocellulose paper was washed three times with 15 mL of 0.1 M TBS with five minutes per wash on a labquake. The nitrocellulose paper was incubate d in 5 mL of ECL substrate (Pie rce) for three minutes at room temperature and developed on CL-XPosure Film (Pierce) by an X-omat 2000 processor (Kodak).

PAGE 55

55 Table 2-1. Bacterial strains and plasmids used. Strain Genotype / Description Source / Reference E. coli DH5 F80dlacZ M15 ( lacZYA-argF )U169 deoR, recA1, endA1 hsdR 17(rk mk +), phoA, supE44 -, thi1, gyrA96, relA1 Bethesda Research Laboratories, Rockville, MD E. coli DH5 (pNR100) F80dlacZ M15 ( lacZYA-argF )U169 deoR, recA1, endA1, hsdR17 (rk mk +), phoA supE44, thi1, gyrA96, relA1 (pNR100) (69) E. coli EC100D Fmcr A (mrr-hsd RMS mcr BC) 80dlacZ M15 lacX74 recA1 end A1 araD139 ( ara, leu )7697 gal U gal K rps L nupG pir+(DHFR) Epicentre, Madison, WI E. coli EC100D (pGTR203) (Hy ) Fmcr A (mrr-hsd RMS mcr BC) 80dlacZ M15 lacX74 recA1 end A1 araD139 ( ara, leu )7697 gal U gal K rps L nupG pir+(DHFR) (pGTR203) (Hy ) This paper E. coli ER2738 F lacIq, ( lacZ )M15 proA+B+, zzf::Tn10( TetR ) / fhuA2, supE, thi, ( lacproAB ), ( hsdMS-mcrB )5, (r m McrBC) New England Biolabs, Ipswich, MA E. coli HB101 (pMJ100) F-, hsdS20(rBmB), recAB 3, ara-14, proA 2 lacYl galK 2, rpsL 20(SmR), xyl-5 mtl -l, supE44, -, (pMJ100) (70) E. coli HB2151 ara, ( lac-pro), thi / F' proA+B+, lacIqZ M15 Amersham Pharmacia, Piscataway, NJ E. coli JM109 endA1, recA1, gyrA96, thi, hsdR17 (rk mk +), relA 1, supE44, ( lac-proAB ) /F traD36, proAB laq IqZ M15 Invitrogen, Carlsbad, CA E. coli MG1655 F, -, ilvG, rfb -50, rph -1 (71) E. coli MG1655 (pGTR203) (Hy ) F, -, ilvG, rfb -50, rph -1 (pGTR203) (Hy ) This paper E. coli O157:H7 EDL933 Stx1+, Stx2+, gyrA (72) E. coli O157:H7 87-23 Stx1-, Stx2-, gyrA (73)

PAGE 56

56 Table 2-1. Continued Strain Genotype / Description Source / Reference E. coli TG1 (lac proAB ), supE, thi, hsdD 5/F traD36, proA+B, lacIq, lacZ M15 MRC, Cambridge, UK E. coli TG1 (pGTR203) (lac proAB ), supE, thi, hsdD 5/F traD36, proA+B, lacIq, lacZ M15 (pGTR203) Gopal Sapparapu Hy Hyperphage genome (contains all M13 genes except gIII ) (66) M13cp M13mp19; CamR; full length gIII (64) M13cp-CT M13mp19; CamR; truncated gIII (64) M13cp-dg3 M13mp19; CamR; deleted gIII (64) pGTR203 pACYC184 with M13 gIII; CamR Gopal Sapparapu pMJ100 pBluescript with slt-II; AmpR (70) pNR100 Stx2 toxoid; AmpR (69) V. cholerae 569B Classical, Inaba (74) V. cholerae N16961 El Tor, Inaba (75)

PAGE 57

57 CHAPTER 3 RESULTS Rationale for Study There exists a strong need for rapid, sensitiv e, and selective m ethods of detection for biological agents. Detecting bi ological organisms and their products has many uses. One reason is for diagnosis of patient samples. Detecti on and identification of a disease-causing agent in patient samples allows for the pr oper treatment to be given. De tecting biological agents in food, water, and air allows for preventative measures to be enforced to prevent or inhibit the spread of disease. A common component in almost all detection assays is a protein that specifically binds to a target molecule. This thesis describes the optimization of protocols us ed to isolate specific proteins that bind to target mo lecules by using phage display. The specific aims of this study are: 1. To optimize extraction methods and determine saturation limits of V. cholerae LPS and use these methods to pan phage display libraries to V. cholerae LPS. 2. To optimize panning and screening proce dures to allow for a more efficient biopanning process that will be more likely to isolate and detect specific recombinant phagemid particles. 3. To optimize the production of phagemid par ticles to ensure high quality and high quantity of phagemid particles. 4. To isolate specific recombinant phagemid particles to E. coli O157:H7 Stx2 toxin using the optimization techniques discovered in previous aims. Specific Aim 1: Panning to V. cholerae LPS The f irst goal of this work was to isolate phage display reagents that specifically recognized V. cholerae O1 LPS. There are many serogroups of V. cholerae but only O1 and O139 cause epidemic cholera. Therefore, O1 LP S serves as a useful target because it is a distinguishing feature of V. cholerae O1 strains. Previous attemp ts in the laboratory to obtain phage display reagents that specifically recognized V. cholerae O1 LPS may have failed because the V. cholerae O1 LPS was panned with the Tomlinson scFv phagemid library. Phagemid

PAGE 58

58 particles, which only encode the gIII gene, require phage amplifica tion tools to provide phage genes I-XI to enable phagemid particle production. We chose the Ph.D. 12mer phage library (NEB) because the Tomlinson scFv phagemid lib rary was disadvantageous due to the poor quality of current methods to amplify phagemid particles. Because the Ph.D. 12mer phage library is M13 phage, it does not require the ai d of tools to amplif y phages like a phagemid library requires. Before the Ph.D. 12mer pha ge display library was to be panned against V. cholerae O1 LPS, an extraction method for V. cholerae O1 LPS and the best solid support for LPS binding had to be determined. Two LPS extraction methods were compared to determine which method generated the best quality of L PS to pan against. Also, two binding supports (microtiter well and nitrocellulo se) paper were analyzed to determine which binding support could bind the most LPS. The Phenol-Water Method Extracted V. cholerae LPS Most Closely Resembled the Commerc ially Acquired V. cholerae LPS The phenol-water method for the extraction of LPS (76) was the previous lab procedure used to extract LPS. However, a new met hod for LPS extraction involving TRIzol Reagent (Invitrogen) was described by Yi a nd Hackett (77) that was stated to be a more efficient method for LPS extraction than the phenol-water method. The TRIzol Reagent LPS extraction can be executed in one day, while the phenol-water LPS ex traction requires nearly a week. The TRIzol Reagent method for LPS extraction is also stated to be a cleaner method than the conventional phenol-water method for LPS extraction by ex tracting LPS with less degradation and contamination. TRIzol Reagent is composed of phenol and guanidinium thiocyanate in aqueous phase, and upon addition of chloroform the TRIz ol Reagent can be used to extract LPS. Vibrio cholerae N16961 LPS was extracted by the TRIzol Reagent extraction of LPS (see Materials and Methods). Th e TRIzol Reagent-extracted V. cholerae LPS was resolved by SDS-

PAGE 59

59 PAGE and stained with Coomassie blue for visua lization of proteins (Fi g. 3-1A) and with TsaiFrasch silver stain for visualization of carbohydr ates (see Materials and Methods) (Fig. 3-1B). The Coomassie blue stain of the TRIzol Reagent-extracted LPS, which was diluted 1:2 in Laemmli sample buffer, showed no detectable proteins. The limit of detection for proteins with Coomassie blue is 0.3-1 g of protein/band (79); theref ore, there was less than 0.3 g of any single protein loaded onto the gel that was stai ned by Coomassie blue. The Tsai-Frasch silver stain for the TRIzol Reagent LPS extraction sh owed non-staining bands be tween 20 and 30 kDa. The Tsai-Frasch silver stain prim arily stains carbohydrates and poor ly stains proteins and lipids (78); therefore, the non-staini ng bands may be lipids or proteins. However, because the Coomassie blue stain for the TRIzol-extracted L PS showed no detectable proteins and due to the thickness of the non-staining bands, which correl ates to concentrati on, then the non-staining bands were most likely due to lipid contamination. The LPS banding pattern of the TRIzol reagent LPS extraction in the silver stai n was very similar to the commercial V. cholerae 569B LPS (Sigma-Aldrich). On the s ilver stain, both LPS preparations had bands at approximately 14 and 20 kDa and no bands larger than 50 kDa. Previously published silver stains of V. cholerae O1 LPS had bands at ~10 and 14 kDa (which was stated to be the lipid A-core of LPS), and ~2050 kDa (which was stated to be the lipid Acore plus repeating O-antigens) (80,81). To determine the concentration of the TRIzol-Reagent-extract ed LPS, the intensities of the bands were compared to those of the commerci al LPS in the silver stained gel (Fig. 3-1B). The intensity of the band of the commercial LPS (1 mg/mL) that was diluted 1:5 was approximately 1.5 times stronger than the intensit y of the band of the TRIzol-Reagent-extracted LPS that was diluted 1:2. Therefore, we estim ated the concentration of the TRIzol-Reagent-

PAGE 60

60 extracted LPS was approximately 260 g/mL. The purity of this LPS extraction was less than 0.2 g of protein/ g of LPS. The phenol-water LPS extraction (76) was perf ormed to compare results with the TRIzol Reagent LPS extraction. The phe nol-water LPS extraction method di ffers little from the original phenol-water LPS extraction desc ribed by Westphal, Luderitz, a nd Bister in 1952 (82). The phenol-water LPS extraction method utilizes the property that most prot eins but not LPS are soluble in phenol, and that LPS is soluble in water. At temp eratures above 68C, phenol and water are miscible. Upon cooling to 5-10C and centrifugation, thre e phases result : an aqueous phase containing LPS, a phenol phase containing proteins, and a solid phase containing waterand phenol-insoluble compounds. Removal and pur ification of the LP S-containing aqueous phase follows the separation of the phases. One and a half liters of log phase (OD600 ~0.4-0.6) V. cholerae 569B was used for extraction of LPS by the phenol-water extraction protocol (see Material s and Methods). The four aqueous extractions were dialyzed and di gested by DNase (Qiagen) and RNase (Qiagen) because nucleic acids are also extracted into the aqueous phase. Prior to digestion the dialyzed extraction was diluted 1:100 in ddH2O, and the absorbance was measured at 260 nm and 280 nm (A260 and A280). The A260, which correlates to DNA concentration, predigestion was 0.277 and post-digestion was 0.148. The A280, which correlates to protein concentration, predigestion was 0.127 and post-digestion was 0.059. The final phenol-water extraction was examined for proteins by Coomassie blue stain (Fig. 3-2A) an d for carbohydrates by Tsai-Frasch silver stain (see Materials and Methods) (Fig. 3-2B). The Coomassie blue st ain for phenol-water-extracted V. cholerae LPS showed a protein band at approximately 15 kDa. This could possibly have been due to the 0.04 mg/mL RNase, 13.7 kDa, present in the extraction. The ba nd intensity correlated

PAGE 61

61 with the concentration of RNase present in the sa mple. Therefore, the protein contamination was most likely not from bacterial prot eins that were co-extracted. The Tsai-Frasch silver stain for phenol-water-extracted V. cholerae LPS showed similar banding patterns to that of the V. cholerae 569B LPS (Sigma-Aldrich). On the silver stain, both LPS preparations had bands at approximately 13, 20, and 23 kDa. The phenol-wat er extraction showed distinct banding around 20-50 kDa that correlates with published data of where the banding of the O-antigen of LPS occurs. The commercially obtained LPS had smearing around 20-50 kDa. Vibrio cholerae O1 LPS has 12-18 O-antigen groups (83); therefore, the additional banding in the phenol-waterextracted LPS may be due to the ability of our phenol-water method to extract LPS with higher numbers of O-antigen side chains than that of the method used for the extraction of the commercial LPS that was phenol extracted and purified by gel-filtration chromatography. To determine the concentration of the phenol-water-extracted LPS, the intensities of the bands were compared to those of the commercial L PS in the silver stained gel (Fig. 3-2B). The intensities of the bands of the commercial LPS (1 mg/mL) that was diluted 1:8 were between the intensities of the bands of th e phenol-water-extracted LPS th at was diluted 1:2 and 1:4. Therefore, we estimated the concentration of the commercial LPS at a dilution of 1:8 was comparable to the concentration of the phenol-wat er-extracted LPS at a dilution of 1:3. Because the commercial LPS was at a concentration of 1 mg/mL, the concentration of the phenol-waterextracted LPS was approximately 375 g/mL. The purity of this LPS extraction was approximately 1 g of protein/ g of LPS. However, if the protein band from the Coomassie blue stain was from RNase, the phenol-water-e xtracted LPS would have less than 0.3 g of bacterial protein contamination/ g of LPS, which is the same prot ein concentration in the TRIzolextracted LPS.

PAGE 62

62 Based on the results described above, we d ecided to use the phenol-water method for the extraction of V. cholerae O1 LPS over the TRIzol Reagent method. The phenol-water-extracted LPS most closely resembled V. cholerae 569B LPS (Sigma-Aldrich) when examined by silver stain. The phenol-water extraction method even appears to be a more sensitive and delicate extraction method than that used to extract th e commercial LPS, because more precise banding from O-antigens was detected in the phenol-wate r-extracted LPS than in the commercial LPS by silver stain. The phenol-water LPS extraction ex tracted more protein than the TRIzol Reagent LPS extraction; however, for using the LPS extrac tion for panning it is more important to have intact O-antigens than minor protein contamination. More V. cholerae LPS Can Be Bound to Nitrocellulose Paper than to a Microtiter Well In m ost panning procedures the target molecule to be panne d against is immobilized onto a solid support. We wanted to compare the satu ration limits for LPS binding to a microtiter well and to a piece of nitrocellulo se paper of equal dimensions. Enzyme-linked immunosorbent assays were performed to determine the sa turation limits of the amount of phenol-waterextracted LPS that could be bound to a microtite r well. A 96-well Maxisorp plate (Nunc) was coated with 0.1-200 g/mL of V. cholerae 569B LPS that was extracted by the phenol-water method. Two different primary antibodies were used to detect the LPS: a mouse antiV. cholerae O1 LPS monoclonal antibody, Vc O1 LPS mAb (Austral Biologi cals, San Ramon, CA) and a rabbit antiV. cholerae O1 polyclonal antibody, Vc O1 LPS pAb (Accurate Chemical, Westbury, NY). The primary antibodies were used at their manufacturer recommended concentration or d ilution: a concen tration of 1 g/mL for Vc O1 LPS mAb and a dilution of 1:100 for the Vc O1 LPS pAb. Standard ELISA procedures were used for the assay (see Materials and Methods) (Fig. 3-3). When using the Vc O1 LPS mAb a significant difference (p 0.05) between the signal to noise (S :N) values (signal on antigen-coated

PAGE 63

63 well/signal on PBS-treated well) of LPS bound to -Vc O1 LPS mAb reacting to HRPconjugated anti-mouse antibody was not reached until the concentration of the LPS used to coat the microtiter plate decreased from 1 g/mL to 0.1 g/mL (0.1 to 0.01 g LPS) (p = 0.01). When using the Vc O1 LPS pAb, a significant difference in th e S:N was not reached until the concentration of the LPS decreased from 10 g/mL to 1 g/mL (p = 0.01). Both the -Vc O1 LPS mAb and the Vc O1 LPS pAb when used at the manufacturer recommended concentrations saturated LPS-coated microtiter we lls when the microtiter wells were coated with LPS at a concentration of 10 g/mL. For the sake of conservi ng the LPS, the concentration of LPS used to coat microtiter wells was decided to be 1 g/mL. While there was a significant difference in the S:N when using 10 g/mL of LPS as opposed to 1 g/mL of LPS with the Vc O1 LPS pAb, the S:N difference was only 0.5. This d ecrease in the S:N by 0.5 did not merit the need to use 10-fold more LPS; therefore, for further experiments V. cholerae LPS was used at a concentration of 1 g/mL to coat microtiter wells. The saturation limits for the primary antibodies were analyzed using 1 g/mL V. cholerae LPS to coat a microtiter plate in an ELISA experiment. The -Vc O1 LPS mAb was serially diluted to concentrations from 1.3 to 0.04 g/mL by 2-fold dilutions. The Vc O1 LPS pAb was serially diluted from 1:100 to 1:3 ,200 by 2-fold dilutions. When using 1 g/mL LPS to coat a microtiter well, the saturation limit of the Vc O1 LPS mAb was not reached (Fig. 3-4A). There were significant differences (p=0.01-0.03) between the S:N of each dilution step when using the Vc O1 LPS mAb at concentrations from 1.3-0.04 g/mL. Therefore, when using 1 g/mL LPS the sa turation limit for the Vc O1 LPS mAb was greater than 1.3 g/mL. When using 1 g/mL LPS to coat a microtiter well, a significant difference in the measured S:N between dilution steps when using the serially diluted Vc O1 LPS pAb was not reached until a

PAGE 64

64 dilution of 1:400 (p=0.01) (Fig. 3-4B). Therefore, the saturation limit of the Vc O1 LPS pAb was reached at a dilution of 1:400. Because a microtiter well coated with 1 g/mL LPS was never saturated by the Vc O1 LPS mAb, the Vc O1 LPS pAb was used to characterize the saturation limits of nitrocellulose paper. Pieces of nitrocellulose paper were cut to dimensions (5 mm x 20 mm) that were approximately equivalent to the surface area of a microtiter well. The same phenol-waterextracted V. cholerae LPS used for the microtiter well saturation experiment was also used to characterize the saturatio n of nitrocellulose with LPS. Serially diluted LPS from 100-1 g/mL was used to saturate strips of nitrocellulose pa per. These strips of L PS-coated nitrocellulose paper were examined in an altered Western blot protocol (see Materi als and Methods for LPS Saturation to Nitrocellulose Paper) using the Vc O1 LPS pAb at a dilution of 1:400 (Fig. 3-5). The nitrocellulose paper was not saturated with LPS until 100 g/mL LPS (250 g) was used to coat the nitrocellulose paper. Therefore, n itrocellulose paper can bind approximately 250 to 2,500 times more LPS than a microtiter well. Panning to V. cholerae LPS Failed to Yield Phages that W ere Specific to V. cholerae LPS Vibrio cholerae 569B LPS (Sigma-Aldrich) was i mmobilized onto a polystyrene Maxisorp microtiter well and panned against the NEB Ph.D. 12mer phage display library (see Materials and Methods). Five r ounds of panning were performed with amplification of phages after the first four round s of panning. Each round of panning promotes selection of phages that specifically bind to the target molecule. Amp lifying the selected phages should increase the percentage of specific phages present in the libra ry to pan with. Panning and amplifying for five rounds should greatly improve the chances of isolating phages that specifically bind to the target molecule.

PAGE 65

65 One hundred phage clones from the fifth round of panning were amplified and screened by ELISA against V. cholerae 569B LPS, carbonate coating buffer, and PBS-treated microtiter wells. The -Vc O1 LPS mAb was also tested against the three coating antig ens and generated a S:N of 15.5, proving that the LPS successfully bound to the microtiter plate. Anti-BSA phagemid particles were used in the ELISA against BSA (10 g/mL), carbonate coating buffer, and PBS-treated microtiter wells and generated a S:N of 22, proving that the ELISA was a successful assay for the detecti on of phages. All 100 clones gave a S:N <1.4. A positive signal was classified as a S:N 2. Therefore, all of the clones screened by ELISA were negative, meaning that they failed to specifically bind to V. cholerae 569B LPS. Panning the Ph.D. 12mer phage library against V. cholerae O1 LPS immobilized onto a microtiter well failed to yield phage clones specific to V. cholerae O1 LPS. It was therefore attempted to improve the panning procedure. Instead of using a microtiter well as a binding support for LPS, nitrocellulose paper was used as a binding support. Nitrocellulose paper can bind approximately 250 to 2,500 times more LPS than a microtiter well; therefore, if more antigen is present to pan against then the chances of isolating clone s that specifically bind to LPS may improve. Also, the elution method was te sted. The previous elution method was with glycine (pH 2.2). To determine if a glycine elutio n or an elution with a high concentration of the target antigen is more successful in eluting phages both elution methods were tested in parallel. The theory behind an antigen elution is that a high concentration of antigen in solution will rapidly bind phages bound to the antigen on the nitrocellulose paper when the phages are transiently released. The end result is that th ese phages will elute o ff of the antigen-coated nitrocellulose paper and bind to antigen in solution. Antigen elution is supposed to be a less harsh treatment than acidic or basic elu tions, which may denature the peptides.

PAGE 66

66 Vibrio cholerae 569B LPS that was phenol-water-extract ed was immobilized onto a strip of nitrocellulose paper and panned with the Ph.D. 12mer phage display library (NEB) (see Materials and Methods). Two pa rallel panning experiments were pe rformed; one panning used a glycine (pH 2.2) elution, while the other panning used a V. cholerae 569B LPS elution. Three rounds of panning were performed with a negati ve panning performed between rounds one and two. The negative panning involved panning the libra ry against a strip of nitrocellulose paper that was blocked in casein blocking buffer (Sigma -Aldrich) to remove phages that were specific to nitrocellulose paper or casein blocking buffer. Amplification of phages was performed after rounds one and two to increase the number of selected phages to improve the chances of isolating phages specific to the target mol ecule. Two hundred clones for each panning were selected from the round 3 elution and amplif ied for screening of phages by ELISA. The 400 phage clones were screened in an ELISA against V. cholerae 569B LPS and PBS. The -Vc O1 LPS mAb was also tested against LPS and PBS-tr eated wells and generate d a S:N of 10, proving that the LPS successfully bound to the microtiter pl ate. Anti-BSA phagemid particles were used in the ELISA against BSA (10 g/mL) and PBS-treated microtiter wells and generated a S:N of 15, proving that the ELISA was a successful assay for the detection of phages. All 400 clones gave a S:N less than 1.7. Therefore, all of the clones screened were negative and failed to specifically bind to V. cholerae O1 LPS. Conclusion of Specific Aim 1 In com paring the phenol-water and the TR Izol Reagent methods for extraction of V. cholerae O1 LPS, the phenol-water method was bett er. The phenol-water extraction for LPS yielded LPS that most closely resembled commercially obtained V. cholerae O1 LPS by silver stain of SDS-PAGE-resolved LPS. The phenol-wat er method even extracted LPS that appeared to have more O-antigens than the commercially obtained V. cholerae O1 LPS. In comparing the

PAGE 67

67 saturation limits of V. cholerae O1 LPS bound to microtiter wells a nd to nitrocellulose paper, the nitrocellulose paper bound approxi mately 250 to 2,500 times more LPS than did microtiter wells. Panning the Ph.D. 12mer phage display library against V. cholerae O1 LPS immobilized onto nitrocellulose and onto microtiter wells failed to yield phages that were specific to V. cholerae O1 LPS in an ELISA. Specific Aim 2: Improve Panning a nd Screening Process of Biopanning There are many steps in the bi opanning and screening process. By analyzing certain key steps and optim izing them, the chances of isola ting phagemid particles th at are specific to the target molecule panned against should increa se. Optimization of the panning process was performed with V. cholerae O1 whole cells as a target. Whole cells were chosen as the target because they are the target antigen for the bios ensors under development by our collaborators. Optimization of the concentration of whole cells were analyzed to determine if higher or lower concentrations of the antigen were more likely to select for specific phagemid particles. Increasing the concentration of antigen shoul d increase the amount of phagemid particles selected. This could mean an increase in phagemid particles that are specif ic and/or non-specific to the target molecule. Decreasing the concentr ation of antigen should decrease the amount of phagemid particles selected. This could mean a decrease in phagemi d particles that are specific and/or non-specific to the target molecule. An id eal concentration would se lect a high number of phagemid particles that are specific to the target molecule and a low number of phagemid particles that are not specific to the target molecule. The Tomlinson scFv libraries have a trypsin cleavable site between the scFv peptide and the pIII peptide of the phagemid particle. Cleaving this site separates the phagemid particle from the antigen, leaving the scFv bound to the antigen. Optimizing the time of trypsin elution may improve the chances of isolating specific phagemid particles. Eluting with trypsin for too short

PAGE 68

68 of a time may fail to elute specific phagemid pa rticles, but eluting for too long of a time may increase the amount of nonspecific phagemid par ticles eluted or may even cause degradation of the phagemid particles. Elution of Bound Phages by Trypsin durin g Panning Was Optimal betw een 10 and 30 Minutes A one-round panning procedure was done to dete rmine the optimal concentration of whole cells in suspension and the optimal trypsin elution time for biopanning. The purpose of the experiment was to compare the differences between using 1 x 108 or 1 x 109 whole cells of V. cholerae N16961 and trypsin elution times of 15, 30, 45, and 60 minutes in the biopanning process. Four parallel pannings were performed. Two were panned with the Tomlinson J library (2.5 x 1010 tu), one against 1 x 108 and the other against 1 x 109 whole cells of V. cholerae in suspension. The two other pannings us ed the Tomlinson J library (2.5 x 1010 tu) containing 1% (~2.5 x 108 tu) of Vc86 phagemid particles (an scFv-p roducing phagemid particle previously isolated in this lab by Dr. Rebecca Moose-Clem ente from the Tomlinson scFv library that recognizes V. cholerae 1019 and E. coli O157:H7 whole cells). Panning against a library that is already enriched with phagemid particles th at recognize the target molecule should help determine the optimal panning parameters. If more phagemid particles are eluted with the enriched library than the non-enri ched library at a certain panning parameter then it would lead to the conclusion that the phagemid particles that recognized the target molecule were eluted at that parameter. One panning used 1 x 108, and the other used 1 x 109 whole cells of V. cholerae in suspension. Standard suspension panning pr ocedures (see Materi als and Methods) were performed. When the trypsin was added to th e bound-phagemid particles, an aliquot of the eluted phagemid particles was removed at 15, 30, 45, and 60 minutes after the start of the elution and was titered immediately (Fig. 3-6). Elution of phagemid particles w ith trypsin was optimal

PAGE 69

69 between 15 and 30 minutes. The phagemid par ticle titer decreased by approximately 80-93% from 30 to 45 minutes for the four pannings. Mo re phagemid particles were eluted when panned against 1 x 109 whole cells. However, there were not enough data to do statis tical analysis to determine if using 1 x 109 whole cells eluted significantly more phagemid particles than using 1 x 108 whole cells in the biopanning process. Because the best whole cell concentration a nd trypsin elution time was not definitively obtained in the first experiment, a second one-rou nd panning was done to analyze these factors. The purpose of the experiment was to compare the differences between using 1 x 107, 1 x 108, and 1 x 109 whole cells of V. cholerae N16961 and trypsin elution times of 10, 20, and 30 minutes in the biopanning process. The same pr otocol was used as the first one-round panning protocol above, except an additional concentratio n of whole cells was added, so there were six parallel pannings instead of four. When the trypsin was added to the bound-phagemid particles, an aliquot of the eluted phagemid particles was removed at 10, 20, and 30 minutes after the start of the elution and was titered im mediately (Fig. 3-7). There were no significant differences in the titers of eluted phagemid particles when comparing the con centrations of whole cells and trypsin elution times used in the second one -round panning experiment. The panning was expected to yield distinct differences between eluted titers of phagemid particles from the libraries with and without 1% Vc86 phagemid partic les. A certain parameter was supposed to be optimal and yield higher elution titers from th e library with 1% Vc86 phagemid particles. Because there were not any significa nt differences in the titers of the eluted phagemid particles at the different parameters an optimal panning cond ition could not be determined. To determine which panning parameter elutes the most phagemid particles that recognize V. cholerae whole

PAGE 70

70 cells, another one-round panning was performed with analysis of phagemid particle specificity via ELISA. A third one-round panning was performed to anal yze whole cell concentration and trypsin elution time further. The purpose of the experi ment was to compare the differences between using 1 x 108 and 1 x 109 whole cells of V. cholerae N16961 and trypsin elution times of 10 and 30 minutes in the biopanning process. The same protocol was used as for the first one-round panning protocol above, except that only two elution time points were taken. When the trypsin was added to the bound phagemid particles, an a liquot of the eluted phagemid particles was removed at 10 and 30 minutes after the start of th e elution and titered im mediately (Fig. 3-8). When panning with the Tomlinson J scFv pha gemid library that contains 1% (~2.5 x 108 tu) Vc86 phagemid particles, the amount of eluted phagemid particles did not significantly change (p=0.50) from 10 to 30 minutes when panning with 1 x 109 whole cells. When panning with the library that contains 1% (~2.5 x 108 tu) Vc86 phagemid particles, the amount of eluted phagemid particles significantly decreas ed (p=0.04) from 10 to 30 minutes when panning with 1 x 108 whole cells. Because the amount of eluted phagemid particles either stayed the same or decreased from 10 to 30 minutes in the pannings w ith 1% Vc86 phagemid particles, it suggested that eluting for 30 minutes had no adva ntage over eluting for 10 minutes. To test if the eluted phagemid particles at the various time points were specific to V. cholerae an ELISA was performed. The eluted phagemid particles from the third one-round panning experiment were amplified in E. coli TG1 (Hy -lg), which is E. coli TG1 that was already infected with hyperphage, an d used at a concentration of 1 x 108 tu/mL in an ELISA. The phagemid particles were analyzed by ELISA against V. cholerae whole cells (3 x 108 CFU/mL) and PBS-treated microtiter wells. A standard ELISA protocol (see Materials and

PAGE 71

71 Methods) was performed (Fig. 3-9). Of the diffe rent panning variations, the amplified phagemid particles from the panning with the library that contained 1% Vc86 phagemid particles, 1 x 108 whole cells V. cholerae, and an elution time of 30 minutes generated a significantly higher S:N (p=0.01-0.04) in an ELISA compared to the amp lified phagemid particles from the other panning variations. This suggested that panning with 108 whole cells with a trypsin elution time of 30 minutes was optimal for selecting phagemid part icles that were specifi c to the panned whole cells compared to using 109 whole cells or a trypsin elution of 10 minutes. Screening of scFv Proteins by a Hig h Throughput ELISA Was Acceptable Once the panning procedure is com plete, screen ing of phages or scFvs is performed. The previous method of screening scFvs involved in ducing phagemid-containi ng bacteria to produce scFvs in culture tubes or flasks and screeni ng the scFvs by ELISA. However, this method was very time consuming. Optimizing the screening process to enable high throughput screening of clones should increase the chances of finding clones that specifically bind to the target molecule. A colony blot was attempted to try to enable hi gh throughput screening of scFv proteins. The colony blot (see Materials and Meth ods) involved the transfer of bacterial colonies that were grown to secrete scFvs onto a piece of nitrocellulo se paper that was coated with a target antigen and blocked with casein blocking buffer. Following the transfer, the piece of nitrocellulose paper was washed with TBS to remove unbound scFv proteins, incubated with Protein-L peroxidase (Sigma-Aldrich) to bind to the kappa light chains of the scF v, washed with TBS to remove excess Protein-L peroxidase, and developed with ECL (Pierce). Upon development with ECL, scFv-producing clones specific to BSA fa iled to produce a positive signal against a piece of nitrocellulose paper coated in 10 g/mL BSA but did produce a positive signal against pieces of nitrocellulose paper coated in PBS or V. cholerae whole cells. Also, clones that were not specific to V. cholerae whole cells in an ELISA produced positive signals against pieces of

PAGE 72

72 nitrocellulose paper coated with V. cholerae whole cells, PBS, or 10 g/mL BSA. The colony blot was determined to be ne ither selective nor specific. Because the colo ny blot was not functioning, a way to improve the culture tube method for the production of scFvs was performed. A high throughput ELISA was contrived to enable screening of hundreds of clones per day. This high throughput procedure involved pi cking bacterial colonies containing a phagemid and growing them overnight in a 96-well plate to produce scFvs (see Materials and Methods). The next day the 96-well plate containing the tu rbid cultures was centrifuged to remove the bacterial cells. These scFvs in the supernatant we re examined in an ELISA for screening of their specificity. To determine how well scFvs could be produced in a microtiter well, as opposed to a culture tube or flask, a bacteria l colony containing a phagemid that encodes an scFv specific to BSA was grown in a microtiter well. Wh en analyzed by ELISA, the anti-BSA ( -BSA) clone was diluted 1:2 in casein blocking buffer and pr oduced a S:N of 11.5 in an ELISA against BSA. Anti-BSA scFv produced in a flask produced an average S:N value between 10 and15; therefore, producing scFv proteins in a mi crotiter well may not be optimal, but it was acceptable for the purpose of a high throughput screen. We were successful in producing scFvs in a microtiter well. The next test was to determine how far the scFv can be diluted and still produce a positive signal. The -BSA clone was grown in a microtiter well to produce scFvs. The produced scFvs were diluted 1:2, 1:4, 1:6, and 1:10 in casein blocking buffer and analyzed by ELISA (Fig. 3-10). Diluting the scFv supernatant 1:10 in casein blocking buffer yielde d a S:N of 5.6. This S:N was high enough for detection of a positive clone and was not significantly different (p=0.13) than the S:N produced when using -BSA phagemid particles that were d iluted 1:6. Therefore, diluting scFv

PAGE 73

73 supernatants as far as 1:10 with other clones can enable detection of positive clones. However, the -BSA clone was the strongest cl one that we had; therefore, weaker clones might fail to be detected if analyzed at a dilution of 1:10. Fo r the purpose of a high throughput ELISA screen for scFvs I would recommend diluting the scFv proteins only 1:4 to 1:6 to less en the chances of not detecting weaker binding scFvs. Screening of Phagemid Particles by a Hig h Throughput ELISA Was Not Optimal Once the panning procedure is com plete, scre ening of phagemid particles or scFvs is performed. Previous methods of screening pha gemid particles involved the amplification of eluted phagemid particles in culture tubes or flasks in 2xTY containing 100 g/mL ampicillin, 40 g/mL kanamycin, and 0.1% (w/v) glucose at 30 C overnight, followed by screening by ELISA. This process was time-consuming; therefore, it was tested if phagemid particle production could be performed in a high throughput manor as with scFvs. Vc86 and -BSA phagemids in E. coli TG1 (Hy -sm) ( E. coli TG1 that was previously infected wi th hyperphage) were produced in a 96-well plate in 2xTYcontaining 100 g/mL ampicillin, 40 g/mL kanamycin, and with or without 0.1% (w/v) glucose at 30C and 37 C overnight. Protoc ols recommend producing phagemid particles in medium that has 0.1% (w/v ) glucose, but due to catabolite repression of the lac promoter, which drives transcription of the gIII gene, phagemid particle production might increase without the presence of glucose in the gr owth medium. The next day, the bacterial cells were removed by centrifugation, and the phage-cont aining supernatants were examined in an ELISA at a dilution of 1:2 in casein blocking buf fer (Fig. 3-11). There was not a significant difference in the S:N (p=0.07-0.48) of phagemid part icles produced in micr otiter wells with and without 0.1% (w/v) glucose. Vc86 phagemid particles produced in micro titer wells with or without glucose at 30C and 37C di d not produce a S:N higher than on e. It was also noted that the cultures in the microtiter plates were onl y slightly turbid. Th e Vc86 phagemid particles

PAGE 74

74 produced in a flask (S:N ~4) yielded a significantly higher S:N (p=0.05) compared to Vc86 phagemid particles produced in microtiter wells. When -BSA phagemid particles were produced at 30C, the S:N was signifi cantly higher (p<0.04) than when -BSA phagemid particles were produced at 37C at either glucose concentration. The -BSA phagemid particles produced in a flask (S:N ~12) yielded a significantly high er (p=0.02) S:N compared to -BSA phagemid particles produced in microtiter wells (S:N~7). The experiment was repeated a second time with only the -BSA phagemid particles (Fig. 3-12). Anti-BSA phagemid particles produced in microtiter wells yielded significantly higher signals (p<0.03) when produced at 30C than at 37C at either glucose concentration. There was not a significant difference (p=0.34) in the signals produced by phagemid particles that were produced with or wit hout glucose at 30C. When phage mid particles were produced in microtiter wells at 37C, there was a significan tly higher (p=0.02) signal when phagemids were produced with no glucose compared to 0.1% (w /v) glucose. There was not a significant difference in the signal (p=0.09) produced by pha gemid particles produced in microtiter wells compared to phagemid particles produced in a flask. When phagemid particles were diluted 1:2, pha gemid particles yielded significantly higher signals in an ELISA when produ ced at 30C compared to 37C. Whether the growth medium included 0.1% (w/v) glucose or not when phagemi d particles were incubated at 30C did not significantly affect the signal of the phagemid particles; theref ore, 0.1% (w/v) glucose was not enough to cause catabolite repression of the lac promoter. Also, phages produced in flasks yielded higher signals than phage s produced in microtiter wells. To determine if low titers of phagemid particles produced in a microtiter well were a factor in the lowered ELISA signals for phagemid particles, the phagemid particles were titered. When -BSA and Vc86 phagemid

PAGE 75

75 particles were produced in mi crotiter wells at 30C, the titers were approximately107 tu/mL. When -BSA and Vc86 phagemid particles were produced in microtiter wells at 37C, the titers were approximately106 tu/mL. Growing phagemid particles at 30C in microtiter wells yielded approximately 10-fold higher tu/mL titers than phagemid particles that were produced in microtiter wells at 37C. Whether the growth medium contained 0.1% (w/v) glucose or not did not affect the titers of phagem id particles. All phages produced in microtiter wells were amplified by E. coli TG1 (Hy -sm). Escherichia coli TG1 (Hy -sm) is E. coli TG1 that was already infected with hyperphage. Having hyperphage already present in E. coli removes an additional step in phagemid par ticle production to enable a more high throughput screen When -BSA and Vc86 phagemid particles were produced in flasks at 30C with 0.1% (w/v) glucose in the medium, the titers were approximately108 to109 tu/mL. The phagemid particles produced in the flasks were amplified by superinf ection with hyperphage, instead of by E. coli TG1 harboring hyperphage. Growing phagemid particles by superi nfection with hyperphage may have resulted in higher titers of phagemid particles comp ared to phagemid particles produced from E. coli that already contained hyperphage. The high throug hput phagemid particle ELISA was not optimal for screening phagemid particle s because phagemid particles could not be produced in high enough titers in microtiter wells to make scr eening of phagemid particles effective in a high throughput manor. Conclusion of Specific Aim 2 When panning with the Tom linson J human synthetic VH + VL phagemid library, trypsin elution was best between 10 and 30 minutes. Phag emid particles could possibly be degraded by trypsin treatment of 45 minutes or greater. Wh en screening scFvs, a high throughput ELISA was acceptable. The scFv proteins can be diluted 1:10 in an ELISA with other scFv proteins and still produce high enough signals to be detected. When screening phages, screening in a high

PAGE 76

76 throughput ELISA was not optimal. Phagemid particles cannot be produced in high enough titers when produced in microtiter wells to be useful in a high throughput manor. Specific Aim 3: Improve Phagemid Particle Production Phagem id particle production is one of the most important steps in biopanning. Because the only M13 gene in a phagemid is the gIII gene, use of phagemids requires an amplification tool to be provided in trans to supply the additional M13 genes to enable phagemid particle production. There needs to be high quality and high quantity phagemid particle production during amplification of phagemid pa rticles to ensure that redundanc y of clones in the library is maintained and clones are produced in high enou gh titers to be useful in assays that use phagemid particles. Hyperphage (Progen) was the previous phagemid particle amplification tool. Hyperphage is an M13-based phage that contains a par tial deletion of the gIII gene to enable production of phagemid particles that display recombinant pIII proteins but no wild-type pIII proteins. However, the commercially obt ained hyperphage stock was of poor quality. The hyperphage stock was st ated to contain 1012 particles/mL; however, it only had 108 tu/mL, as determined in an infection assay. Because there was not enough Tomlinson scFv phagemid library stock for panning experiment s, the library stock needed to be amplified. The Tomlinson I and J scFv libraries contain 1 x 108 phagemid clones. To ensure that the Tomlinson scFv libraries are amplified to ensure that the redundancy of clones is maintained, an amplification tool needs to enable high qua lity and high quantity phagemid particle production. Because commercially obtained hyperphage had too low of titers to be an effective phagemid particle amplification tool, other ways of phage mid particle production were analyzed.

PAGE 77

77 Escherichia coli TG1 Harboring th e Hyperphage Geno me Was Not an Optimal Phagemid Particle Amplification Tool To overcome the low titers of hyperphage, the hyperphage genome was incorporated into E. coli and maintained in the strain as a plasmi d to ensure that every phagemid-containing E. coli also contained the hyperphage genome to enab le phagemid particle production. Log phase (OD600 ~0.4) E. coli TG1 was transduced with hyperphage at a MOI of 5, incubated in a 37C standing water bath for 30 minutes, serially d iluted in BSG, and then plated on 2xTY plates containing 40 g/mL kanamycin. The plates were incubated overnight at 37C. The next day the plates contained small (0.5 mm diameter) and large (1-1.5 mm diameter) colonies. Two small and large colonies were si ngle colony passaged twice. Th e small colonies maintained a small phenotype (0.5 mm diameter), and the large colonies maintained a large phenotype (1-1.5 mm diameter). The two strains were named E. coli TG1 (Hy -sm) and E. coli (Hy -lg). The infection efficiencies of phagemid particles into E. coli TG1, E. coli TG1 (Hy -sm), and E. coli (Hy -lg) were analyzed. The three E. coli strains were grown to log phase and transduced with serially diluted -BSA and Vc86 phagemid particle s in BSG at a MOI of 0.001. Transductions were incubated in a 37C standi ng water bath for 30 minutes and plated on 2xTY AG plates containing 40 g/mL kanamycin for transduced strains of E. coli TG1 containing hyperphage and on 2xTY AG plates for transduced E. coli TG1. The numbers of E. coli cells transduced with -BSA phagemid particles were as follows: 1,200 transductions into E. coli TG1, 340 transductions into E. coli TG1 (Hy -sm), and 100 transductions into E. coli TG1 (Hy -lg). When the number of -BSA phagemid particles transduced into E. coli TG1 was set at an infection efficiency of 100%, the number of -BSA phagemid particles transduced into E. coli TG1 (Hy -sm) was 28% and the number of -BSA phagemid particles transduced into

PAGE 78

78 E. coli TG1 (Hy -lg) was 0.1% of the number of tr ansduced phagemid particles into E. coli TG1. The numbers of E. coli transduced with Vc86 phagemid particles were as follows: 4,300 transductions into E. coli TG1, 950 transductions into E. coli TG1 (Hy -sm), and 80 transductions into E. coli TG1 (Hy -lg). When the number of Vc86 phagemid particles transduced into E. coli TG1 was set at an infection efficiency of 100%, the number of Vc86 phagemid particles transduced into E. coli TG1 (Hy -sm) was 22% and the number of Vc86 phagemid particles transduced into E. coli TG1 (Hy -lg) was 2% of the number of transduced phagemid particles into E. coli TG1. Escherichia coli TG1 (Hy -sm) had an infection efficiency that was approximately 25% of the infection efficiency of E. coli TG1, and E. coli TG1 (Hy -lg) had an infection efficiency that was approxi mately 1% of the infection efficiency of E. coli TG1. The infection efficiencies of E. coli TG1 (Hy -sm/lg) were too low to for the E. coli strains to be useful to amplify the Tomlinson scFv libraries. However, they could still be useful as an amplification tool to amplify single clones when the transducti on efficiency is not critical. Therefore, the quality and quantit y of phagemid particles produced by E. coli TG1 (Hy -sm/lg) were analyzed. Escherichia coli TG1 (Hy -sm) and E. coli TG1 (Hy -lg) were analyzed for their ability to produce phagemid particles. Escherichia coli TG1 (Hy -sm) and E. coli TG1 (Hy -lg) were transduced with -BSA phagemid particles (as described above) and grown under conditions to produce phagemid particles (see Materials and Methods). Escherichia coli TG1 containing BSA phagemids was superinfected and grown under conditions to produce phagemid particles (see Materials and Methods ). The superinfected E. coli TG1 was used as a comparison for strains of E. coli TG1 containing hyperphage as a phagemid particle production tool. Anti-BSA

PAGE 79

79 phagemid particle yields were as follows: E. coli TG1 that was superinfected with hyperphage yielded 3 x 10 8 tu/mL, E. coli TG1 (Hy -sm) yielded 1 x 109 tu/mL, and E. coli TG1 (Hy -lg) yielded 3 x 108 tu/mL. These results showed that E. coli TG1 that contained hyperphage produced high titers of phagemid particles. To determine the quality of -BSA phagemid particle s produced by the three E. coli strains, the produced -BSA phagemid particles were analyzed in an ELISA. Microtiter plates were coated with 10 g/mL BSA or only PBS. Standard ELI SA procedures were followed (see Materials and Methods) to analyze 1 x 108 tu/mL of -BSA phagemid particles as the primary antibody (Fig. 3-13). Anti-BSA phagemid particles produced a significantly higher S:N (p=0.01) when produced from E. coli TG1 (Hy -sm) compared to E. coli TG1 (Hy -lg). There was not a significant difference in the S:N (p=0.50) from -BSA phagemid particles produced from superinfected E. coli TG1 and E. coli TG1 (Hy -lg). Because superinfected E. coli TG1 and E. coli TG1 (Hy -lg) yielded titers (~3 x 108 tu/mL) of -BSA phagemid particles that were approximately a third of that of E. coli TG1 (Hy -sm) titers (~1 x 109 tu/mL), E. coli TG1 (Hy -sm)-produced -BSA phagemid particles were analy zed in the ELISA with a third less phage-containing supernatan t volume. The amplified -BSA phagemid particles were from the same clone; therefore, the differe nces in S:N from the produced -BSA phagemid particles was not expected. A possible reason for the S:N differences was that a different amount of particles/mL was used in the ELISA. The ELISA compared equal tu/mL titers of the -BSA phagemid particles produced. However, tu/mL is a measure of infectious particles and not total particles. Because an ELISA measures partic les with binding activity, it is possible that superinfected E. coli TG1 and E. coli TG1 (Hy -lg) yielded similar particles/mL yields as E. coli TG1 (Hy -sm), which would account for the differences in the S:N. The results showed

PAGE 80

80 that E. coli TG1 (Hy -sm) and E. coli TG1 (Hy -lg) had infection efficiencies that were too low for them to be effective amplification t ools to amplify the Tomlinson J scFv library. However, these strains could be used as t ools to amplify single pha gemid particles with acceptable phagemid particle production and quality. Helper Plasmids Were Not an Optimal Ph agemid Particle Amplifica tion Tool Helper plasmids (Los Alamos National Labor atory, Los Alamos, NM) (64) are M13-based plasmids that are maintained in E. coli to provide phagemid-containing E. coli with all of the genes necessary to produce phagemid particles. The helper plasmids come in three forms that either have a full-length (M13cp), a dele ted (M13cp-dg3), or a truncated (M13cp-CT) gIII gene. The helper plasmid-containing E. coli strains were analyzed for their infection efficiencies with phagemid particles, production yield of phagemid pa rticles, and the quality of phagemid particles produced. The three helper plasmids were electropor ated (see Materials and Methods) into electrocompetent E. coli TG1. All three transformations yiel ded colonies between 0.5-2 mm in diameter. A small (0.5 mm) and a large (2 mm) colony from E. coli TG1 (M13-cp), E. coli TG1 (M13cp-CT), and E. coli TG1 (M13cp-dg3) were single colony passaged twice. Escherichia coli TG1 (M13cp-sm) was the only strain that maintain ed a consistent phenotype of 0.5 mm colonies. The other 5 strains yielded co lonies with diameters of 0.52 mm after each passage. To determine if the different colony size phenotypes ha d an effect on the functionality of the helper plasmids, a small and large isolate from each strain were analyzed for their infection efficiencies with phagemid particles. To analyze the infection efficiencies of phage mid particles into helper plasmid-containing E. coli TG1, an infection efficiency (see Materials and Methods) experiment was performed that compared the infection efficienci es of helper plasmid-containing E. coli with E. coli that did not

PAGE 81

81 contain helper plasmids. The bacterial strains were grown to log phase and transduced with BSA or Vc86 phagemid particles at a MOI of 0.1 (Table 3-1). The number of E. coli TG1cells that were transduced with phagemid particles was set at an infection efficiency of 100%, and the number of E. coli TG1 cells harboring helpers that were transduced with phagemid particles was compared to the infection efficiency of E. coli TG1. Of the six E. coli TG1 strains harboring helper plasmids, E. coli TG1 (M13cp-dg3-sm) and E. coli TG1 (M13cp-CT-sm) were the only strains that yielded inf ection efficiencies grater than 0.3% of the infection efficiency of E. coli TG1. Escherichia coli TG1 (M13cp-dg3-sm) had infection efficiencies of 175% and 97%, while E. coli TG1 (M13cp-CT-sm) had infection efficiencies of 7% and 13% compared to the infection efficiency of E. coli TG1. Because the small and large isolates of helper plasmid-containing E. coli had different infection effi ciencies, the colony phenotype had an effect of the functionality of the he lper plasmid-containing E. coli. Because initial tests with E. coli TG1 (M13cp-dg3-sm) had an infection efficiency th at was as high or highe r than the infection efficiency of E. coli TG1, its infection efficiency was further analyzed with the transduction of additional phagemid particles. Escherichia coli TG1 (M13cp-dg3-sm) was transduced w ith clone 18 (a phagemid particle isolated from the Tomlinson I scFv library that recognizes the A27L protein of vaccinia virus), -AV20N3 (a phagemid particle isolated from the Tomlinson I scFv library that recognizes rAuto (a recombinant L. monocytogenes murein hydrolase)), and Vc86 phagemid particles. The various strains were analyzed further for their infection efficiencies compared to E. coli TG1 (Table 3-2). Escherichia coli TG1 (M13cp-dg3-sm) had infection efficiencies of 92%, 113%, and 120% compared to the infection efficiency of E. coli TG1. Escherichia coli TG1 (M13cpdg3-sm) had infection efficiencies that were approximately the same as E. coli TG1; therefore,

PAGE 82

82 this strain was used for further analysis to determine the quantity and the quality of phagemid particles produced. However, before E. coli TG1 (M13cp-dg3-sm) was analyzed further, M13cp-CT-sm was examined. To test if the helper plasmid (M13-CT-sm) c ould yield a higher inf ection efficiency in another E. coli strain, it was extracted (see Materials and Methods) from E. coli TG1 (M13cpCT-sm) and electroporated (see Materials and Methods) into E. coli JM109. Escherichia coli JM109 was chosen because it was genotypically more similar than E. coli TG1 to the bacterial strain that Los Alamos National Laborat ory used with the helper plasmids ( E. coli DH5 F). Escherichia coli JM109 and E. coli DH5 F contain relA 1 and recA 1, while E. coli TG1 does not. Escherichia coli TG1 (M13cp-CT-sm) and E. coli JM109 (M13cp-CT-sm) were transduced with clone 18, -AV20N3, and Vc86 phagemid particle s and were analyzed for their infection efficiencies compared to E. coli TG1 and E. coli JM109 (Table 3-3). The average infection efficiency of E. coli TG1 (M13cp-CT-sm) was 47%, while the average infection efficiency of E. coli JM109 (M13cp-CT-sm) was 28%. Therefore, using E. coli JM109 to maintain M13cp-CT-sm did not increase the infection efficiency. In analyzing the infection efficiencies, E coli TG1 (M13cp-dg3-sm) had an average infection efficiency of 120% with a standard deviation of 33% and E. coli TG1 (M13cp-CT-sm) had an average infection efficiency of 30% with a standard deviation of 16%. Escherichia coli TG1 (M13cp-dg3-sm) had a significantly highe r infection efficiency (p=0.01) than E. coli TG1 (M13cp-CT-sm). Escherichia coli TG1 (M13cp-dg3-sm), E. coli TG1 (M13cp-CT-sm), and E. coli TG1 that was superinfected with hyperphage were an alyzed for their abil ity to produce phagemid

PAGE 83

83 particles. To determine if th e addition of 1 mM IPTG to th e growth medium would increase phagemid particle production, phagemid particle amplification was performed in parallel with and without the addition of 1 mM IPTG to the growth medium. Isopropyl-beta-Dthiogalactopyranoside is a molecular mimic of a lactose metabolite that induces transcription from the lac promoter. Because the lac promoter drives transcription of the gIII -scFv fusion, it was hypothesized that increased e xpression of pIII-scFv fusions would result in higher yields of phagemid particles. Phagemid particles were am plified (see Materials a nd Methods) and titered by the spot-titer method (see Mate rials and Methods) (Table 3-4) Phagemid particles were amplified to titers of 106-108 tu/mL by E. coli TG1 (M13cp-dg3-sm), 105 -106 tu/mL by E. coli TG1 (M13cp-CT-sm), and 108 -109 tu/mL by E. coli TG1 superinfected with commercial hyperphage. Titers of phagemid particles produced with 1 mM IPTG were either the same or lower than titers produced without IPTG. Theref ore, it was determined that the addition of 1 mM IPTG did not improve phagemid particle production. Phagemid particles produced by E. coli TG1 (M13cp-dg3-sm) and E. coli TG1 (M13cp-CT-sm) yielded low titers of phagemid particles. An overnight shaki ng culture for the amplification of phagemid particles contains approximately 109 CFU/mL. Therefore, for the helper pl asmids to yield titers as low as 105 tu/mL means that only 1 in 10,000 bacteria produ ced an infectious phagemid particle. These titers were too low for either helper plasmid to be used as an effective phagemid particle production tool. Even though helper plasmids produced phagemid particle titers too low to be used as an amplification tool, the phag emid particles produced by E. coli TG1 (M13cp-dg3-sm), E. coli TG1 (M13cp-CT-sm), and E. coli TG1 superinfected by hyperphage were further analyzed by ELISA. Phagemid particles were serially diluted and analyzed in a standard ELISA procedure

PAGE 84

84 (see Materials and Methods). Anal ysis of ELISA data showed that S:N values as high as 7 when less than 5 x 104 tu/mL of phagemid particles were used as a primary antibody. Previous data in the laboratory showed that phagemid particles were not detectable in an ELISA unless they were used at concentrations of at least 1 x 106 particles/mL. Spot titering is a measure of transducing units (number of infectious units), while an ELISA is a measure of the number of particles that possess antibody activity. To determine if the measured tu/mL were underestimating the phagemid particles/mL, an anti-M 13 sandwich ELISA was performed to quantify the number of phagemid particles produced by E. coli TG1 (M13cp-dg3-sm), E. coli TG1 (M13cp-CT-sm), and E. coli TG1 superinfected by hyperphage. A microtit er plate was coated with anti-M13 pVIII monoclonal antibody at a concentration of 10 g/mL. Phagemid particles were incubated with the coated wells at dilutions of 1:2, 1:20, or 1: 200. The phagemid particles were then detected using HRP-conjugated anti-M13 monoclonal an tibody at a dilution of 1:2,500, followed by standard ELISA procedures (see Materials and Me thods). A concentrated phagemid preparation, which had been amplified by hyperphage, was seri ally diluted and used as a standard for particles/mL. The values for particles/mL were li nearly extrapolated from the standard curve of the standard phagemid particle (Table 3-4). Phagemid particles amplified by E. coli TG1 (M13cp-CT-sm) had particles/mL to tu/mL ra tios between 740 and 23,000. Therefore, phagemid particles were being produced in high quantities; however, less than 0.1% of the phagemid particles produced were infectious Phagemid particles amplified by E. coli TG1 (M13cp-dg3sm) had particles/mL to tu/mL ratios between 1 and 190. Phagemid particles amplified by E. coli TG1 that was superinfected with hyperphage had particles/mL to tu/mL ratios between 6 and 24. Of the three amplification tools (M13cp-dg3-sm, M13cp-CT-sm, and commercial

PAGE 85

85 hyperphage), commercial hyperphage was the only t ool that could amplify phagemid particles to titers that were high enough to be us ed in phage display applications. Phagemid particles amplifie d by helper plasmids in E. coli TG1 yielded phagemid particle titers between 105 and 108 tu/mL. Helper plasmids amplified phagemid particles to titers too low to be useful in biopanning. To determine if another E. coli strain could improve the titers of helper plasmid-amplified phagemid particles, E. coli MG1655 containing helper plasmid (M13cp-dg3-sm) was used to analyze phagemid particle production. Escherichia coli MG1655 was chosen because it has few known mutations and lacks the F plasmid. Therefore, it was hypothesized that the stress upon the E. coli for maintaining the helper plasmids would be lessened if it did not already have to maintain the F plasmid. Four different phagemid particles were amplified with E. coli MG1655 (M13cp-dg3-sm), and the resulting titers were: 2.5 x 107, 6.3 x 106, 2.5 x 107, and 3.9 x 107 tu/mL for an average of 2.4 x 107 tu/mL, which was still too low to be useful in biopanning. Homemade Hyperphage Titers Were Increas ed with Amplification in E. coli MG1655 (pGTR203) Commercial hyperphage was the only phagemid particle amplif ication tool tested that produced phagemid particle progeny in yields high enough to be used in biopanning. However, the titer of the commercial hyperphage stock was not high enough to be used to amplify the Tomlinson scFv libraries. Therefore, a way to improve homemade hyperphage production was analyzed. Because hyperphage has a partial deletion of its gIII gene, it needs the gIII gene to be provided in trans in the E. coli strain used to produce hyperp hage particles. Homemade hyperphage was previously produced by using pGTR203, a plasmid that encodes gIII. Escherichia coli TG1 (pGTR203) or E. coli HB2151 (pGTR203) were infected with hyperphage, and the transduced bacteria were grown under conditions that promoted hyperphage production.

PAGE 86

86 In the past, hyperphage amplified from E. coli TG1 (pGTR203) yielded titers between 107 to 108 tu/mL and E. coli HB2151 (pGTR203) yielded titers between 106 to 107 tu/mL. Both of these strains are F+ E. coli. Because hyperphage infects bacteria by the F pilus, it was hypothesized that some of the produced hyperphages could have been taken up by the bacteria. Therefore, it was tested if hyperphage could be produced in higher titers when produced in an FE. coli. Two FE. coli strains, E. coli MG1655 and E. coli EC100D, were used to amplify hyperphage. The strains were made electr ocompetent (see Materials and Methods) and transformed by electroporation (see Ma terials and Methods) with pGTR203. Escherichia coli MG1655 (pGTR203) and E. coli EC100D (pGTR203) were made electrocompetent and transformed by electroporation with the hyperphage genome that had been extracted from E. coli TG1 (hyperphage) with a QIAprep Spin Miniprep kit (Qiagen) (see Materials and Methods). Five colonies from each transformation with hyperphage, E. coli MG1655 (pGTR203) (Hy ) 15 and E. coli EC100D (pGTR203) (Hy ) 1-5, were single colony passaged twice and analyzed for their ability to produce hyperphage. Standing overnight cultures of all ten isolates were diluted 1:400 in broth and grown under conditions that promote phage production (s ee Materials and Methods). The hyperphage produced were analyzed for tu/mL by the whole plate titer method (see Materials and Methods) and for particles/mL by an anti-M13 sandwich ELI SA (described above) (T able 3-5). The titers of hyperphage produced from the five E. coli EC100D (pGTR203) (Hy ) isolates 2.1 x 108 2.5 x 108 tu/mL were significantly lower than the hyperphage produced from the five E. coli MG1655 (pGTR203) (Hy ), 3.0 x 109.0 x 109 tu/mL (p=0.02). The ratios of particles/mL to tu/mL of hyperphages produced from E. coli EC100D (pGTR203) (Hy ) were between 0.4 and 7.8, and the ratios of particles/mL to tu/mL of hyperphages produced from E. coli MG1655

PAGE 87

87 (pGTR203) (Hy ) were between 0.3 and 1.0. Therefore th e amount of particles and transducing units of hyperphage produced from E. coli EC100D (pGTR203) (Hy ) and E. coli MG1655 (pGTR203) (Hy ) were approximately the same, meani ng that the hyperphages produced were infectious. Both FE. coli strains generated higher titers of hyperphage than did the two F+ E. coli strains. Concentration of hyperphage produced from E. coli MG1655 (pGTR203) (Hy ) should generate a high enough titer of hyperphage to enable high quality and high quantity production of the Tomlinson scFv libraries. Escherichia coli MG1655 (pGTR203) (Hy ) isolate 5 yielded the highest titer of hyperphage (5.4 x 109 tu/mL); therefore, it was used to pr oduce a large quantity of hyperphage to be used to amplify the Tomlinson J scFv phagemid library. One liter of hyperphage was produced from E. coli MG1655 (pGTR203) (Hy ) 5 with a titer of 2.0 x 109 tu/mL, and the phages were concentrated by PEG precipitation (see Materials and Methods) to a volume of 2.8 mL with a final titer of 3 x 1011 tu/mL. This concentrated hyperphage was used to amplify the Tomlinson J scFv phagemid library. The original Tomlinson J scFv library was amplified to ensure high quality and high quantity of phagemid particles produced. The Tom linson J scFv phagemid library was originally received in E. coli TG1. This stock was grown in batch cu lture and frozen. Therefore, clones were not lost due to poor phagemid particle amp lification with the low titer hyperphage because we grew the original bacterial stock of the library and not the previously made phagemid particle stock of the library that most likely lost redundancy of clones because it was amplified with low titer commercial hyperphage. A frozen stock of E. coli TG1 (Tomlinson J scFv phagemid library) was thawed on ice and inoculated into one liter of 2xTY AG broth. This culture was grown in a 37C shaking incubator until the culture reached log phase. Thirty minutes into the

PAGE 88

88 incubation an aliquot of the grow ing culture was removed and titered to ensure that the library was initially in a concentration high enough that redundancy of clones was present. The titer of the E. coli TG1 (Tomlinson J scFv phagemid library ) 30 minutes into incubation was 1.3 x 107 CFU/mL, and because the culture volum e was one liter there were 1.3 x 1010 CFU. The Tomlinson J phagemid library contains ~1 x 108 phagemid clones; therefore, there was initially a 100-fold redundancy of clones in th e library. Approximately 2.3 x 1010 CFU of log phase E. coli TG1 (Tomlinson J scFv phagemid library) was s uperinfected with homemade hyperphage at a MOI of 10 and incubated for 30 minutes in a 37C standing incubation. The superinfected bacteria were then titered on 2xTY AG plates containing 40 g/mL kanamycin to ensure that enough bacteria were superinf ected so that the redundancy of the phagemid clones was maintained. Approximately 2.9 x 1010 bacteria containing a phagemid were superinfected, which represents about 300-fold redundancy of clones. Th e superinfected bacteria were then grown in 1 liter of broth under conditions that promoted phagemid particle production (see Materials and Methods). The final yield Tomlinson J scFv pha gemid particles in the amplified culture was 6.6 x 109 tu/mL. The liter of phagemid particles was then PEG precipitated to 15 mL of 4.4 x 1011 tu/mL, which yields 6.6 x 1012 tu. Titering throughout the Tomlinson J scFv library amplification ensured that the redun dancy of the library clones was maintained at each step. The Tomlinson J scFv library that was amplified w ith homemade hyperphage was frozen and also used for biopanning. Conclusion of Specific Aim 3 Escherich ia coli TG1 harboring hyperphage was not an optimal phagemid particle amplification tool because of the low infection e fficiency with phagemid particles. The infection efficiencies of phagemid particles into E. coli TG1 (Hy -sm) and E. coli TG1 (Hy -lg) were 28% or less than the infection effi ciency of phagemid particles into E. coli TG1. Phagemid

PAGE 89

89 particle amplification by helper plasmids was not an optimal amplification tool because it produced phagemid particle titers too low to be used in biopanning. Producing homemade hyperphage in E. coli MG1655 (pGTR203) enabled high yield of hyperphage particles. The improved high titer homemade hyperphage was used to successfully amplify the Tomlinson J scFv phagemid library to ensure that the redundancy of phagemid clones was maintained. Specific Aim 4: Isolation of Sp ecific Recombinant Phage to Stx2 Toxin of E. coli O157:H7 The Tom linson J scFv phagemid library th at was amplified with homemade hyperphage was panned against Stx2 toxin of E. coli O157:H7. Shiga toxins (Stx) kill mammalian cells and cause severe disease from infection with E. coli O157:H7 cells. Escherichia coli O157:H7 can produce Shiga toxins 1 and/or 2. Escherichia coli harboring the Stx2 toxin appears to be more virulent than E. coli harboring Stx1 or both Stx1 and 2 toxins (36-39). Therefore, Stx2 toxin was the target molecule chosen for panning. The Stx2 toxin preparation that wa s used to pan against was obtained from Toxin Technolog y (Sarasota, FL) and was stated to contain 50% Stx2 toxin. The other 50% of proteins in the Stx2 toxin preparation was most likely composed of contaminating E. coli HB101 proteins because the Stx2 toxi n preparation was prepared from E. coli HB101 (pMJ100) lysates. p MJ100 encodes Stx2 toxoid. A three round biopanning process was perfor med to pan the Tomlinson J scFv phagemid library against a Stx2 toxin preparation-coated immunotube (see Materials and Methods) with amplification between panning roun ds to amplify the selected pha gemid particles. Phagemid particles were eluted using tr ypsin treatment for 10 minutes. Phagemid particles contain a trypsin cleavage site between the scFv and the p III. Therefore, phagemid particles that were bound to Stx2 toxin preparation by their scFv could be eluted from the Stx2 toxin preparation to be amplified for further rounds of panning. Approximately 1 x 1012 phagemid particles were added to an immunotube th at was coated with 100 g of Stx2 toxin preparation. Approximately

PAGE 90

90 2.2 x 106 phagemid particles were eluted after round one of panning. The round-one-eluted phages were amplified with homemade hyperphage to 2 x 1010 tu/mL in 100 mL of broth. The amplified phagemid particles were then PEG precipitated to 9 x 1011 tu/mL in 2.5 mL PBS. Approximately 1 x 1012 phagemid particles from the PEG-preci pitated phagemid particles from round one were panned against Stx2 for a second round of panning. Approximately 2.0 x 107 phagemid particles were eluted after round two of panning. The round-two-eluted phages were amplified with homemade hyperphage to 2.5 x 1010 tu/mL in 100 mL broth. The amplified phagemid particles were then PEG precipitated to 7.8 x 1011 tu/mL in 2.5 mL PBS. Approximately 1 x 1012 phagemid particles from the PEG-preci pitated phagemid particles from round two were panned against Stx2 for a th ird round of panning. Approximately 1.5 x 1010 phagemid particles were eluted af ter the third round of panning. Th e increase in the number of phagemid particles eluted (2 x 106 for round one, 2 x 107 for round two, 2 x 1010 for round three) after every round suggests that phagemid par ticles that specifica lly bound to Stx2 toxin preparation were being se lected and amplified. Eighty-four phagemid particles from the third round of panning were individually amplified with homemade hyperphage (see Materials and Methods) to be analyzed for specificity by ELISA and Western blot experime nts. Forty-four clones were initially analyzed by ELISA. Microtiter wells coated with 10 g/mL Stx2 toxin preparation were incubated with phagemid particle-containing supernatants that we re diluted 1:2 in blocking buffer, 1 g/mL antiStx211E10 ( -Stx2 subunit A) monoclona l antibody (Toxin Technology), 1 g/mL anti-Stx2BB12 ( -Stx2 subunit B) monoclonal anti body (Toxin Technology), or 1 x 108 tu/mL -BSA phagemid particles (negative cont rol for determining the noise for S:N calculations) (Fig. 3-14). Of the 44 clones screened, 40 had a S:N between 9 and 18. The two anti-Stx toxin monoclonal

PAGE 91

91 antibodies had a S:N less than 2 against Stx2 to xin preparation. The Stx2 toxin preparation may not have bound to the microtiter wells very e fficiently, or the anti-S tx2 monoclonal antibodies may not have worked very well. Conversely, th e phagemid clones that were positive may have bound to the microtiter well instead of the Stx2 toxin preparation. To determine if the clones were binding to Stx2 or to the microt iter well, 4 of the previously screened clones along with 44 new cl ones were analyzed via ELISA using the same procedure as above with the addi tion of PBS-treated microtiter wells as a negative control (Fig. 3-15). Of the 48 clones, 44 had signals of 0.5 or higher against Stx2 toxi n preparation-coated microtiter wells. Four clones (clones 1, 23, 35, a nd 39) had signals as high or higher against PBS-treated microtiter wells compared to Stx2 toxi n preparation-coated microtiter wells. These four clones could possibly recognize polystyrene. Because the library was panned against Stx2 toxin preparation-coated immunot ubes, it was possible that some clones were selected that bound to the polystyrene immunotube. Fifty-three of 59 clones (89%) had a S:N from 2 to 23, and 48 of 59 clones (81%) had a S:N from 10 to 23 agai nst Stx2 toxin preparation. Therefore, the majority of the clones screened bound to the Stx2 toxin preparation. Ho wever, 4/59 (7%) of the clones bound to PBS-treated wells as well as Stx2 -coated wells. The four clones had a signal in an ELISA between 0.31-0.87 agains t PBS-treated wells. These clones most likely bound to polystyrene. Isolation of plastic binding phages is fairly common (84). No matter how well the panning vessel is blocked with blocking buffer, phages can still be selected that bind to the panning vessel. Of the 59 clones screened, 2 (3 %) did not generate a positive signal against Stx2-coated microtiter wells or non-coated microtiter wells. These clones had unknown binding specificity.

PAGE 92

92 The anti-Stx monoclonal antibodies ( -Stx2 subunit A and -Stx2 subunit B) were further analyzed to determine the reason why they did no t react in the ELISA. Microtiter wells coated with 10 g/mL Stx2 toxin preparation or carbonate bicarbonate buffer were incubated with Stx2 subunit A and -Stx2 subunit B monoclonal antibodies at concentrations of 1, 10, or 100 g/mL. Bound antibodies were detected with HR P-conjugated anti-mouse antibody (Cappel) or HRP-conjugated anti-mouse antibody (Sigma-Aldri ch). Standard ELISA procedures followed (see Materials and Methods). All parameters generated S:N values less than two. The Stx2 toxin preparation appeared to be binding to the we ll because the amplified panning clones were binding to the Stx2 with high S:N va lues in microtiter wells. The concentrations of anti-Stx2 monoclonal antibodies were not an issue b ecause using the monoclonal antibodies at a concentration as high as 100 g/mL did not improve the signal of the monoclonal antibodies in an ELISA. Two different HRP-conjugated anti-mous e antibodies were used to determine if the secondary antibody was not recogn izing the primary antibodies. Neither HRP-conjugated antimouse antibody improved the signal of the m onoclonal antibodies in an ELISA. Positive controls for the anti-mouse antibodies were not pe rformed in the ELISA to determine whether or not the anti-mouse antibodies we re functional. However, the two HRP-conjugated anti-mouse antibodies were used within three days of these experiments and generated positive signals in an ELISA against murine antibodies. The reason why the anti-Stx monoclonal antibodies gave low signals with Stx2 toxin preparation in an ELISA was not determined. Analysis by ELISA shows that the majority of clones isolated from the Stx2 panning bound to the Stx2 toxin preparati on in an ELISA. Because the Stx2 toxin preparation was only 50% pure for Stx2 toxin, the panning clones could have been binding to St x2 toxin or the 50% of impurities in the preparation. To analyze if th e Stx2 toxin preparation panning clones recognized

PAGE 93

93 the Stx2 toxin or the impurities, three clones were selected and examined for reactivity with the Stx2 toxin preparation and E. coli supernatants or periplasmic break fractions. The strains examined were E. coli O157:H7 933 (expresses Stx1 and 2 toxins), E. coli DH5 (pNR100) (expresses Stx2 toxoid), and E. coli DH5 Microtiter wells were coated with 0.01 mg/mL Stx2 toxin preparation, 0.07 mg/mL E. coli O157:H7 933 periplasmic break fraction, 2 mg/mL E. coli O157:H7 933 supernatant, 1 mg/mL E. coli DH5 (pNR100) periplasmic break fraction, 1 mg/mL E. coli DH5 periplasmic break fraction, or carbon ate coating buffer. If the clones bound to Stx2 toxin, they were expected to genera te positive signals to a ll the antigens except for E. coli DH5 and carbonate coating buffer. If the clones bound to impurities in the Stx2 toxin preparation, the clones were expected to gene rate positive signals for all the antigens except carbonate coating buffer. Coated wells were incubated with 10 g/mL of -Stx2 subunit A monoclonal antibody (11E10), 10 g/mL of -Stx2 subunit B monoclonal antibody (BB12), 1 x 109 tu/mL of phagemid particles of Stx2 clones (46, 48, and 49), or -BSA phagemids. Standard ELISA procedures were followed (see Materials and Methods) (F ig. 3-16). All parameters generated S:N values of less than 1.5 except fo r clones 46, 48, and 49 that were analyzed with Stx2 toxin preparation-coated microtiter wells. Because the monoclonal antibodies to Stx2 toxin preparation did not generate posit ive signals in an ELISA against any of the parameters, it was not determined if the clones from the Stx2 toxi n preparation panning were reacting to Stx2 toxin or impurities in the commercially obtained Stx2 toxin preparation. Clones that were selected from the Stx2 toxin preparation panning and had a positive signal by ELISA to Stx2 toxin preparation were furt her analyzed by Western blot to determine if the clones were binding to Stx2 toxin or to im purities in the Stx2 toxin preparation. Six clones were screened in a Western blot against Stx2 toxin preparation, E. coli O157:H7 933 (expresses

PAGE 94

94 Stx1 and 2 toxins) supernatant, and E. coli O157:H7 87-23 (Stx phage cu red). Because the Stx2 toxin is composed of A and B subunits that are approximately 32 and 10-kDa proteins, a clone that bound to Stx2 toxin was expect ed to recognize a 32 or 10-kDa protein in the Stx2 toxin preparation and E. coli O157:H7 933 supernatant but not in E. coli O157:H7 87-23 supernatant. A clone that bound to impurities was expected to recognize proteins in all three antigens. A 4-20% polyacrylamide gel was loaded with 1.5 g/well of Stx2 toxin preparation, 120 mg/well of E. coli O157:H7 933 supernatant, and 120 mg/well of E. coli O157:H7 87-23 supernatant. The resolved proteins were transf erred to nitrocellulose and analyzed by Western blot (see Materials and Methods ). Phagemid particles ( -BSA and Stx2 toxin preparation panning clones 46, 48, 80, 81, and 83) were used as a primary antibody at approximately 109 tu/mL, and -Stx2 subunit A monoclonal antibody was used at 4 g/mL. Anti-Stx2 subunit B monoclonal antibody failed to generate a positive si gnal by Western blot. Standard Western blot procedures were performed (see Mate rials and Methods) (Fig. 3-17). The -Stx2 subunit A monoclonal antibody recognized a 32 -kDa (the size of the A subunit) protein in the Stx2 toxin preparation and in the E. coli O157:H7 933 supernatant, there by proving that Stx2 toxin subunit A was present in both preparations but not in the E. coli O157:H7 87-23 supernatant as was expected. Anti-BSA failed to recognize any of th e antigens, as was expected. Clone 46 failed to recognize any of the three antig ens. Clones 48 recognized 35, 45, and 55-kDa proteins in the Stx2 toxin preparation and a 35-kDa protein in both E. coli O157:H7 supernatants. This result means that clone 48 probably bound to bacterial protein impurities in the Stx2 toxin preparation. Clones 49 and 80 bound to a 35-kDa protein in all th ree antigens and a 20-kDa protein in the Stx2 toxin preparation and in E. coli O157:H7 87-23 supernatant. Therefore, clones 49 and 80 probably bound to bacterial protein impurities in the Stx2 toxin preparation. Clones 81 and 83

PAGE 95

95 appeared to recognize the same bacterial protein impurities as cl ones 49 and 80, except that their band intensities were much lighter. Western blot and ELISA analyses showed that clones obtained from panning against the Stx2 toxin preparation recognized the Stx2 toxin preparation. The clones that have been screened so far appeared to bind to impurities in the Stx2 toxin preparation. To determine if the Stx2 toxin preparation contained 50% Stx2 toxin as stated by th e producer, the toxin preparation was resolved and analyzed by SDS-PAGE (Fi g. 3-18). If the Stx2 toxin preparation was comprised of 50% Stx2 toxin, 10 and 32-kDa prot ein bands should have been the predominant protein bands. However, a protein band at a pproximately 32 kDa was a minor protein band. A protein band at 10 kDa was a major protein ba nd; however, whether thes e protein bands were actually the Stx2 toxin subunits was not determined. The 10 and 32-kDa protein bands comprised less than 10% of the Stx2 toxin prepar ation. Because the concentration of Stx2 toxin was so low in the commercially obtained St x2 toxin preparation, it was probable that the majority of the phagemid particles isolated from the Stx2 panning would recognize impurities instead of Stx2 toxin. Conclusion of Specific Aim 4 Panning against comm ercially obtained Stx2 toxi n preparation was successful in isolating phagemid particles that bound to the Stx2 toxin pr eparation. Initial sc reening of clones by Western blot showed that the clones bound to impur ities in the commercially obtained Stx2 toxin preparation and not Stx2 toxin. The commercially obtained Stx2 toxin preparation was supposed to contain 50% Stx2 toxin; however SDS-PAGE analysis showed th at less no more than 10% of the proteins present in the prepar ation may have been Stx2 toxin. Therefore, it was likely that the majority of clones selected from the panni ng would recognize impurities in the Stx2 toxin preparation and not Stx2 toxin. However, this panning exercise showed that the Tomlinson J

PAGE 96

96 library amplified with the homemade hyperphage was a more effective tool than the previously amplified libraries.

PAGE 97

97 Table 3-1. Infection efficiencies of E. coli TG1 containing various helper plasmids with phagemid particles. Bacterial strains were transduced with anti-BSA ( -BSA) and Vc86 phagemid particles at a MOI of 0.1, and transduced bacteria were enumerated. Infection efficiency is the number of trans ductions into a bacterial strain divided by the number of transductions into E. coli TG1 and is represented as a percentage. Strain Transductions with -BSA Infection efficiency with -BSA (%) Transductions with Vc86 Infection efficiency with Vc86 (%) Average infection efficiency (%) E. coli TG1 1.2 x 106 1003.6 x 106 100 100 E. coli TG1 (M13cp-sm) 9.3 x 102 0.16.9 x 103 0.3 0.2 E. coli TG1 (M13cp-lg) 8.3 x 102 0.19.6 x 103 0.3 0.2 E. coli TG1 (M13cp-CT-sm) 7.9 x 104 74.8 x 105 13 10 E. coli TG1 (M13cp-CT-lg) <1.3 x 102 <0.01<1.3 x 102<0.01 <0.01 E. coli TG1 (M13cp-dg3-sm) 2.1 x 106 1753.6 x 106 97 136 E. coli TG1 (M13cp-dg3-lg) <1.3 x 102 <0.01<1.3 x 102<0.01 <0.01 Table 3-2. Infecti on efficiencies of E. coli TG1 (M13cp-dg3-sm) with phagemid particles. Bacterial strains were transduced with phagemid particles at a MOI of 0.1, and transduced bacteria were enumerated. Infection efficiency is the number of transductions with a phagemid into a b acterial strain divided by the number of transductions with the same phagemid into E. coli TG1. Strain Phagemid Transductions Infection efficiency (%) E. coli TG1 Vc86 5.3 x 106 100 E. coli TG1 (M13cp-dg3-sm) Vc86 4.9 x 106 92 E. coli TG1 Clone 18 2.4 x 106 100 E. coli TG1 (M13cp-dg3-sm) Clone 18 2.7 x 106 113 E. coli TG1 -AV20N3 1.5 x 106 100 E. coli TG1 (M13cp-dg3-sm) -AV20N3 1.8 x 106 120

PAGE 98

98 Table 3-3. Infecti on efficiencies of E. coli TG1 (M13cp-CT-sm) and E. coli JM109 (M13cp-CTsm) with phagemid particles. Bacteria l strains were transduced with phagemid particles at a MOI of 0.1, and transduced bacteria were enumerated. Infection efficiency is the number of transductions with a phagemid particle into a helper plasmid-containing bacterial strain divide d by the number of transductions with the same phagemid particle into the bacter ial strain without the helper plasmid. Strain Phagemid Transductions Infection efficiency (%) E. coli TG1 Vc86 2.4 x 105 100 E. coli TG1 (M13cp-CT-sm) Vc86 1.3 x 105 54 E. coli TG1 Clone 18 1.4 x 105 100 E. coli TG1 (M13cp-CT-sm) Clone 18 5.5 x 104 39 E. coli TG1 -AV20N3 9.3 x 104 100 E. coli TG1 (M13cp-CT-sm) -AV20N3 4.4 x 104 47 E. coli JM109 Vc86 6.8 x 105 100 E. coli JM109 (M13cp-CT-sm) Vc86 1.5 x 105 22 E. coli JM109 Clone 18 1.7 x 105 100 E. coli JM109 (M13cp-CT-sm) Clone 18 6.1 x 104 36 E. coli JM109 -AV20N3 1.4 x 105 100 E. coli JM109 (M13cp-CT-sm) -AV20N3 3.7 x 104 26 Table 3-4. Comparison of transducing units to particles per milliliter of amplified phagemid particles. The spot-titer method was used to determine tu/mL. Particles/mL was determined via a sandwich ELISA with anti-M13 antibodies. (n = 2 wells) Phagemid particle amplified Amplification tool used tu/mL Particles/mL (Particles/mL)/ (tu/mL) Vc86 M13cp-CT-sm 8.7 x 105 2.0 x 1010 23000 Vc86 M13cp-CT-sm 4.6 x 106 3.4 x 109 740 clone 18 M13cp-dg3-sm 2.4 x 108 2.2 x 108 1 -AV20N3 M13cp-dg3-sm 4.4 x 107 7.0 x 107 2 -BSA M13cp-dg3-sm 1.3 x 108 1.8 x 108 1 Vc86 M13cp-dg3-sm 8.7 x 106 1.7 x 109 190 clone 18 Hyperphage 7.1 x 108 1.1 x 1010 15 -AV20N3 Hyperphage 2.0 x 109 1.1 x 1010 6 -BSA Hyperphage 5.8 x 108 1.4 x 1010 24

PAGE 99

99 Table 3-5. Comparison of transducing units to particles per milliliter of amplified homemade hyperphage produced from FE. coli strains. The spot-titer method was used to determine tu/mL. Particles/mL was dete rmined via a sandwich ELISA with anti-M13 antibodies. (n = 2 wells). Hyperphage production strain tu/mL Particles/mL (Particles/mL)/ (tu/mL) E.coli EC100D (pGTR203) (Hy ) 1 5.3 x 107 4.2 x 108 7.8 E.coli EC100D (pGTR203) (Hy ) 2 9.4 x 107 3.5 x 108 3.7 E.coli EC100D (pGTR203) (Hy ) 3 3.2 x 108 1.1 x 108 0.4 E.coli EC100D (pGTR203) (Hy ) 4 2.9 x 106 < 1.0 x 107 <3 E.coli EC100D (pGTR203) (Hy ) 5 6.0 x 108 2.1 x 108 0.4 E.coli MG1655 (pGTR203) (Hy ) 1 1.0 x 109 1.0 x 109 1.0 E.coli MG1655 (pGTR203) (Hy ) 2 8.0 x 108 7.3 x 108 0.9 E.coli MG1655 (pGTR203) (Hy ) 3 3.0 x 109 1.3 x 109 0.4 E.coli MG1655 (pGTR203) (Hy ) 4 4.3 x 109 1.4 x 109 0.3 E.coli MG1655 (pGTR203) (Hy ) 5 5.4 x 109 1.5 x 109 0.3

PAGE 100

100 Figure 3-1. Analysis of TRIzol Reagent-extracted V. cholerae N16961 LPS by SDS-PAGE. Samples were resolved on a 12% (w/v) pol yacrylamide gel and stained with A) Coomassie blue stain or B) Tsai-Frasc h silver stain. (1) TRIzol-extracted V. cholerae N16961 LPS. (2) V. cholerae 569B LPS (Sigma-Aldrich).

PAGE 101

101 Figure 3-2. Analysis of phenol-water-extracted V. cholerae 569B LPS by SDS-PAGE. Samples were resolved on a 4-20% (w/v) polyacrylamide gel and stained with A) Coomassie blue stain or B) Tsai-F rasch silver stain. (1-3) V. cholerae 569B LPS (SigmaAldrich) at dilutions of 1:4 (1), 1:8 (2), and 1:16 (3). (4-6) Phenol-water-extracted V. cholerae 569B LPS at dilutions of 1:2 (4), 1:4 (5), and 1:8 (6).

PAGE 102

102 Figure 3-3. Analysis of the saturation limit of V. cholerae 569B LPS to microtiter wells by ELISA. Signal to noise (S:N) ratios were calculated for the react ivity of mouse antiV. cholerae O1 LPS monoclonal antibody ( -Vc O1 LPS mAb) (1 g/mL) and rabbit antiV. cholerae O1 LPS polyclonal antibody ( -Vc O1 LPS pAb) (1:100) to V. cholerae 569B LPS (0.1-200 g/mL)-coated microtiter wells. Bound primary antibodies were detected with either HRP-conjugate d goat anti-rabbit antibody or HRP-conjugated goat anti-mouse antibody. Wells were developed with TMB substrate, and their absorbances read at 630 nm. (n=3 wells). There was a significant decrease (*p=0.01) in S:N when the concentration of LPS decreased from 1 to 0.1 g/mL when using the -Vc O1 LPS mAb. There was a significant decrease (*p=0.01) in the S:N when the concentration of LPS decreased from 10 to 1 g/mL when using the -Vc O1 LPS pAb. 0 2 4 6 8 10 0.11101001000 V. cholerae LPS ( g/mL)S:N -Vc O1 LPS mAb (1 g/mL) -Vc O1 LPS pAb (1:100)**

PAGE 103

103 A Figure 3-4. Analysis of the saturati on limits of primary antibodies to V. cholerae 569B LPScoated microtiter wells by ELISA. Primary antibodies: A) mouse antiV. cholerae O1 LPS monoclonal antibody ( -Vc O1 LPS mAb) (0.04-1.3 g/mL) and B) rabbit antiV. cholerae O1 LPS polyclonal antibody ( -Vc O1 LPS pAb) (1:100-1:1,600) were reacted with V. cholerae 569B LPS (1 g/mL)-coated microtiter wells. Bound primary antibodies were detected with either HRP-conjugated goat anti-rabbit antibody or HRP-conjugated goat anti-mouse antibody. Wells were developed with TMB substrate and their absorbances read at 630 nm. The S:N were calculated for every primary antibody concentrat ion or dilution (n=3 wells). In (A) comparisons of S:N between every monoclonal antibody con centration used yi elded significant differences (p<0.03) in S:N; therefore, th e saturation limit of the monoclonal antibody could not be determined. In (B) a significant decrease (*p=0.01) in the S:N was not reached until the polyclonal antibody dilu tion decreased from 1:400 (ratio dilution of 0.003) to 1:800(ratio dilution of 0.001); therefore, the satu ration limit of the polyclonal antibody was reached at a dilution of 1:400. 0 2 4 6 8 10 12 0.00.51.01.5 -Vc O1 LPS mAb (1 g/mL)S:N

PAGE 104

104 Ratio of Dilution, Dilution 0.01, 1:100 0.005, 1:200 0.003, 1:400 0.001, 1:800 <0.001, 1:1,600 B Figure 3-4. Continued. 0 2 4 6 8 10 12 00.0020.0040.0060.0080.01 Ratio of DilutionS:N *

PAGE 105

105 Figure 3-5. Saturation of V. cholerae 569B LPS to nitrocellulose paper. Vibrio cholerae 569B LPS was incubated with nitrocellulose paper at concentrations from 1-100 g/mL. The primary antibody used to detect the LPS was rabbit antiV. cholerae O1 LPS polyclonal antibody (1:400). Bound prim ary antibody was detected with HRPconjugated goat anti-rabbit antiserum and detected with ECL substrate. The nitrocellulose paper became saturated with V. cholerae 569B LPS at a LPS concentration of 100 g/mL. The strips incubated with 0 g/mL LPS failed to stain significantly.

PAGE 106

106 0 10 20 30 40 50 60 70 80 90 100 15304560 Trypsin Elution (min)Phages Eluted /107 (tu/mL) Figure 3-6. Titers of eluted phagemid partic les from the first one-round panning optimization experiment. The Tomlinson J library +/ 1% Vc86 phagemid particles was panned against 108 and 109 whole cells of V. cholerae N16961 in suspension. Phagemid particles that were bound to the V. cholerae whole cells were eluted with 10% (v/v) trypsin stock (10 mg/mL trypsin, 50 mM Tris-HCl (pH 7.4), 1 mM CaCl2 in water) in PBS for 15, 30, 45, and 60 minutes. Eluted pha gemid particles were enumerated via the spot-titer method. 108 cells Vc 109 cells Vc 108 cells Vc, 1% Vc86 109 cells Vc, 1% Vc86

PAGE 107

107 A B Figure 3-7. Titers of eluted phagemid partic les from the second one-round panning optimization experiment. The Tomlinson J library +/ 1% Vc86 phagemid particles was panned against 107, 108, and 109 whole cells of V. cholerae N16961 in suspension. Phagemid particles that were bound to the V. cholerae whole cells were eluted in 10% (v/v) trypsin stock (10 mg/mL trypsin, 50 mM Tris-HCl (pH 7.4), 1 mM CaCl2 in water) in PBS for 10, 20, and 30 minutes. Eluted phage mid particles were enumerated via the spot-titer method. A) Panning with the Tomlinson J library B) Panning with the Tomlinson J library (+) 1% Vc86 phagemid particles. 0 20 40 60 80 100 120 1015202530 Trypsin Elution (min)Phages Eluted/107 (tu/mL) 0 20 40 60 80 100 120 102030 Trypsin Elution (min)Phages Eluted/106 (tu/mL) 107 cells Vc 108 cells Vc 109 cells Vc 107 cells Vc 108 cells Vc 109 cells Vc

PAGE 108

108 0 2 4 6 8 10 12 010203040 Trypsin Elution (min)Phages Eluted/107 (tu/mL) Figure 3-8. Titers of eluted phagemid partic les from the third one-round panning optimization experiment. The Tomlinson J library +/ 1% Vc86 phagemid particles was panned against 108 and 109 whole cells of V. cholerae N16961 in suspension. Phagemid particles that were bound to the V. cholerae whole cells were eluted in 10% (v/v) trypsin stock (10 mg/mL trypsin, 50 mM Tris-HCl (pH 7.4), 1 mM CaCl2 in water) in PBS for 10 and 30 minutes. Eluted phagemid particles were enumer ated via the spottiter method. 10 8 cells Vc 109 cells Vc 108 cells Vc, 1% Vc86 109 cells Vc, 1% Vc86

PAGE 109

109 Figure 3-9. Analysis of eluted phagemid par ticles from the third one-round panning optimization experiment by ELISA. Microtit er wells coated with 3 x 108 CFU/mL V. cholerae whole cells or PBS and we re incubated with 1 x 108 tu/mL of amplified eluted phagemid particles or 10 g/mL of mouse antiV. cholerae O1 LPS monoclonal antibody. Bound phagemid particles were de tected with HRP-conjugated anti-M13 secondary antibody, and the monoclonal anti body was detected with HRP-conjugated goat anti-mouse antibody. Wells were deve loped with TMB substrate, and their absorbances were read at 630 nm (n=3 we lls). There was a si gnificantly higher S:N (*p=0.01-0.04) from the amplified eluted phagemid particles from the panning with 108 whole cells with a 30 minute elution us ing the library containing 1% Vc86 when compared to the amplified eluted phagem id particles from the other pannings. 0 2 4 6 8 10 10 min 10 min +Vc86 30 min 30 min +Vc86 10 min 10 min +Vc86 30 min 30 min +Vc86 -Vc O1 LPS mAb Vc86 S:N 10 8 10 8 10 8 10 8 10 9 10 9 10 9 10 9

PAGE 110

110 Figure 3-10. Analysis of anti-BSA ( -BSA) scFv proteins produced in microtiter wells by ELISA. Microtiter wells were coated with 10 g/mL BSA or PBS. Anti-BSA scFv particles were produced in a microtiter well and isolated by centrifugation and the scFv-containing supernatants were diluted 1:2, 1:4, 1:6, or 1:8 in blocking buffer and incubated with the coated wells. Bound sc Fv particles were detected with HRPconjugated Protein L. We lls were developed with TMB substrate, and their absorbances were read at 630 nm. (n=2 wells). 0 2 4 6 8 10 12 14 1:21:41:61:10 Dilution of -BSA scFv Supernatant in Blocking BufferS:N

PAGE 111

111 Figure 3-11. Analysis of -BSA and Vc86 phagemid particles produced in microtiter wells by ELISA. Phagemid particles Vc86 (V) and -BSA (B) were produced at either 30C or 37C in medium containing either 0% or 0.1% glucose in a microtiter well (W) or in a flask (F). Microtiter wells were coated with 10 g/mL BSA, 3 x 108 CFU/mL V. cholerae whole cells, or PBS. Phagemid particles were dilu ted 1:2 in blocking buffer and incubated with the coated wells Bound phagemid particles were detected with HRP-conjugated anti-M13 secondary antibody. Wells were developed with TMB substrate, and their ab sorbances were read at 630 nm. (*) represents p<0.05 and (**) represents p<0.01. (n=2 wells) Phagemid Particle V V V V V B B B B B Temp. (C) 30 30 37 37 30 30 30 37 37 30 % Glucose 0 0.1 0 0.1 0.1 0 0.1 0 0.1 0.1 Grown in W W W W F W W W W F 0 2 4 6 8 10 12 14S:N ** **

PAGE 112

112 Figure 3-12. Analysis of -BSA phagemid particles produced in microtiter wells by ELISA. Microtiter wells were coated with 10 g/mL BSA. Phagemid particles were diluted 1:2 in blocking buffer and incubated w ith the coated wells. Bound phagemid particles were detected with HRP-conj ugated anti-M13 secondary antibody. Wells were developed with TMB substrate, and th eir absorbances were read at 630 nm (n=2 wells). Phagemid particles produced in a microtiter well at 30C produced significantly higher (*p=0.03) S:N values th an phagemid particles produced in a microtiter well at 37C at the same glucose concentration. 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.830C, 0% glucose 30C, 0.1% glucose 37C, 0% glucose 37C, 0.1% glucose 30C, 0.1% glucose, in flask -BSA Phagemid ParticlesSignal

PAGE 113

113 Figure 3-13. Analysis of anti-BSA ( -BSA) phagemid particles produced by hyperphage by ELISA. Microtiter wells were coated with 10 g/mL BSA or PBS and incubated with 1 x 108 tu/mL -BSA phagemid particles produced from E. coli TG1 strains already infected with hyperphage (Hy -sm, Hy -lg) or from E. coli TG1 that was superinfected with hyperphage. Bound phagemid particles were detected with HRPconjugated anti-M13 secondary antibody. Wells were developed with TMB substrate, and their absorbances were read at 630 nm (n=3 wells). The S:N from BSA phagemid particles produced from (Hy -sm) was significantly lower (*p<0.01) than the S:N from -BSA phagemid particles produced from (Hy -lg) or from superinfection with hyperphage. 0 5 10 15 20 Superinfected with Hy (Hy -sm)(Hy -lg) a-BSA Phagemid ParticlesS:N

PAGE 114

114 Figure 3-14. Analysis of phagemid particles (1-22, 43-64) selected from panning against Stx2 toxin preparation by ELISA. Microt iter wells were coated with 10 g/mL of Stx2 toxin preparation (Toxin Technologies) a nd incubated with cl ones 1 through 22 and 43 through 64 diluted 1:2 in blocking buffer, 1 g/mL anti-Stx2 subunit A monoclonal antibody (11E10), 1 g/mL anti-Stx2 subunit B monoclonal antibody (BB12), or anti-BSA phagemid particles dilu ted 1:2 in blocking buffer. Microtiter wells coated with 10 g/mL BSA were incubated with anti-BSA phagemid particles that were diluted 1:2 in blocking buffer. Bound phagemid particles were detected with HRP-conjugated anti-M13 monoclona l antibody. Bound monoclonal antibodies were detected with HRP-conjugated goa t affinity purified anti-mouse antibody. Wells were developed with TMB substrate, and their absorbances were read at 630 nm (n=2 wells). The noise used to calculate S:N was the signal of anti-BSA phagemid particles that were incubated with Stx2-coated wells. 0 5 10 15 20 251 2 3 4 5 6 7 8 9 1 0 1 1 1 2 1 3 1 4 1 5 1 6 1 7 1 8 1 9 2 0 2 1 2 2 4 3 4 4 4 5 4 6 4 7 4 8 4 9 5 0 5 1 5 2 5 3 5 4 5 5 5 6 5 7 5 8 5 9 6 0 6 1 6 2 6 3 6 4 B B 1 2 1 1 E 1 0 a B SAS:N

PAGE 115

115 Figure 3-15. Analysis of phagemid particles (1, 3-5, 23-42, 65-84) selected from panning against Stx2 toxin preparation by ELISA. Micr otiter wells were coated with 10 g/mL of Stx2 toxin or carbonate coating buffer a nd incubated with clones 1, 3 through 5, 23 through 42, and 65 through 84 diluted 1:2 in blocking buffer. Bound phagemid particles were detected with HRP-conj ugated anti-M13 monoclonal antibody. Wells were developed with TMB substrate, and th eir absorbances were read at 630 nm (n=2 wells). 0 5 10 15 20 251 3 4 5 2 3 2 4 2 5 2 6 2 7 2 8 2 9 3 0 3 1 3 2 3 3 3 4 3 5 3 6 3 7 3 8 3 9 4 0 4 1 4 2 6 5 6 6 6 7 6 8 6 9 7 0 7 1 7 2 7 3 7 4 7 5 7 6 7 7 7 8 7 9 8 0 8 1 8 2 8 3 8 4S:N

PAGE 116

116 Figure 3-16. Analysis of anti-Stx2 monoclona l antibodies and clones 46, 48, and 49 by ELISA. Microtiter wells were coated with 10 g/mL of Stx2 toxin, 70 g/mL of E. coli O157:H7 933 periplasmic break fraction, 2 mg/mL of E. coli O157:H7 933 supernatant (O157 Sup.), 1mg/mL of E. coli DH5 (pNR100) periplasmic break fraction, and 1mg/mL of E. coli DH5 periplasmic break fraction. (PP) represents periplasmic break fractions. Bound phagemi d particles were detected with HRPconjugated anti-M13 monoclonal antibody. Bound monoclonal antibodies were detected with HRP-conjugate d goat affinity purified an ti-mouse antibody. Wells were developed with TMB substrate, and th eir absorbances were read at 630 nm (n=2 wells). mAb 11E10 mAb BB12 clone 46 clone 48 clone 49 -BSA DH5 PP DH5 (pNR100) PP O157 Sup. O157 PP Stx2 0 2 4 6 8 10 S:N

PAGE 117

117 Figure 3-17. Western blot analys is of phagemid clones from pa nning on Stx2 toxin preparation. Samples were resolved by SDS-PAGE on a 4-20% (w/v) polyacrylamide gel and analyzed by Western blot. Samples: (T) -1.5 g/well of Stx2 toxin, (+) 120 g/well E. coli O157:H7 933 (contains Stx1 and 2 t oxins) supernatant, and (-) -120 g/well E. coli O157:H7 87-23 (does not produce Stx toxi n) supernatant. Primary antibodies: Anti-Stx2 toxin subunit A monoclonal an tibody (Mab); Stx2 phagemid clones 46, 48, 80, 81, 83. A) Western blot was developed with TMB and exposed for 30 seconds. B) Western blot was developed with TMB and exposed for 5 minutes.

PAGE 118

118 Figure 3-18. Analysis of Stx2 toxin prepara tion (Toxin Technologies) by SDS-PAGE. Five micrograms of Stx2 toxin preparation was resolved on a 4-20% (w/v) polyacrylamide gel and stained with Coomassie blue stain. Stx2 toxin is comprised of a 32-kDa A subunit and five 10-kDa B subunits.

PAGE 119

119 CHAPTER 4 DISCUSSION Detection of biological agents and their produ cts has becom e a prime importance in the United States (85) since the anth rax attacks of 2001 (86). Early de tection of biological agents is important to prevent or inhibit th e spread of disease. Detecting biological agents in water, air, and food enables the proper course of action to be implemented to lessen or prevent the effects of the organism. Detection of biologi cal agents or their products is al so important for diagnosis of a disease. Detection of the disease-causing agent in patient samples enab les the proper treatment to given. There are vast numbers of detection assays, a nd a common component to most of them is a protein that binds specifically to a target antigen. Monoclonal anti bodies are the default tool that scientists use today to detect target molecules; however, monoclonal antibodies have some drawbacks that are causing the use of phage displa y reagents to become more prevalent and take the place of monoclonal antibodies (2). First, the genera tion of monoclonal antibodies requires killing animals and takes months. Also, monoclona l antibodies are rarely isolated to pathogenic or self antigens, causing limitations in the scope of the target antigen for the generation of monoclonal antibodies. Phage display is advant ageous because it lacks many of the drawbacks of monoclonal antibodies. In phage display, if using a non-immunized library, the killing of animals is not required and the isolation of sp ecific recombinant phages can be done in a couple weeks instead of months. Also, the scope of anti gen used for phage display is far greater than that used for monoclonal antibodies because phage display enables the se lection of recombinant proteins to pathogenic and self antigens. Ph age display was developed in 1985 by George P. Smith as a means to display foreign proteins on a filamentous phage (52). The sequence for the foreign peptide is encoded in frame with a coat protein gene of the phage genome. This enables

PAGE 120

120 the foreign peptide to be fused to the coat pr otein, which effectively links the phenotype and genotype of the phage. Therefore, every phage that displays a fo reign protein also encodes its peptide sequence in the genome. This thesis describes efforts to optimize phage display methods to increase the chances of isolating recombinant antibodies or peptides for the potential use in detection assays. Specific Aim 1: Panning to V. cholerae LPS Vibrio cholerae is a hum an pathogen that is the causa tive agent of cholera. Cholera is a diarrheal disease that affects millions of people worldwide annually (9). Detecting V. cholerae in the environment and patient samples is the first step to preventing and treating the disease. To detect V. cholerae we need to isolate a recombinant phage that binds specifically and sensitively to V. cholerae. There are over 200 serogroups of V. cholerae that are clas sified on the basis of the O-antigen of their LPS; however, only se rogroups O1 and O139 cause endemic acute diarrhea. Therefore, the LPS of these two serogr oups would be excellent targets for detection. Vibrio cholerae O1 LPS was chosen as a target for biopanning. Because V. cholerae O1 LPS was no longer sold commercially, it had to be extracted in the laboratory before any panning experiments could be performed. The previous method to extract LPS in the laboratory was a phenol-water extraction (76). However, a new LPS extraction was developed using a commercial RNA isolation reagent, TRIzol Reagent, that could extract LPS in a fraction of the time of the phenol-water extraction (77). The TRIzol method was even claimed to be able to extract LPS with less degradation and contamination than the phenol-water method. To determine which method was better, both methods were used to extract V. cholerae O1 LPS and the extracted LPS was analyzed and compared to commercially acquired V. cholerae O1 LPS (Sigma-Aldrich) that had been phenol extracted and purified by gel-filtration chromatography.

PAGE 121

121 The criteria for the optimal LPS extraction we re to have little or no protein and lipid contamination and to have similar banding patterns to the commercial V. cholerae O1 LPS. Vibrio cholerae O1 LPS was extracted using the phenol -water and TRIzol methods. The LPS was resolved by SDS-PAGE and analyzed by Coomassie blue staining and Tsai-Frasch silver staining (Figs. 3-1 and 3-2). Vibrio cholerae O1 LPS that was extracted by the TRIzol reagent had less protein contamination than did the LPS that was extracted by the phenol-water method. The relative purity of the TR Izol-extracted LPS was less than 0.3 g of protein/ g of LPS, while the relative purity of the phenol -water-extracted LPS was approximately 1 g of protein/ g of LPS. However, the protein contamin ation in the phenol-water LPS extraction was most likely due to the RNase that was used in th e extraction. The Coomassie blue stain for the phenol-water-extracted LPS show ed a protein band at approxim ately 15 kDa. The molecular weight of RNase is 13.7 kDa. The phenol-w ater LPS extraction contained 0.04 mg/mL of RNase, which correlates to the band intensity of the 15-kDa protein band. Because the phenolwater LPS extraction only cont ained approximately 0.02 mg/mL of DNase, MW of 31kDa, the DNase might not have been concen trated enough to be seen by Coom assie blue stain. Therefore, both LPS extractions probably had less than 0.3 g of bacterial protein contamination for every g of LPS extracted. The TRIzol extracted LPS showed non-staining bands between 20 to 30 kDa on the silver stain (Fig. 3-1B). The Tsai-Fra sch silver stain primarily stains carbohydrates and poorly stains proteins and lipids (78). Because the Coomassie blue stain for the TRIzolextracted LPS showed no detectable proteins a nd due to the thickness of the non-staining bands, which correlates to concentr ation, the non-staining bands ar e most likely due to lipid contamination. Previously published silver stains of V. cholerae O1 LPS had bands at ~10 and 14 kDa which were stated to be the lipid A-core of LPS, and ~20-50 kDa which were stated to be

PAGE 122

122 the lipid A-core plus repeating O-antigens (80,81). The non-stai ning bands were not the lipid A because the lipid A contains two glucosamines that are stainable by silver stain. Therefore, the TRIzol-extraction appears to have also extracted cellular lipids. Therefore, the TRIzol-extracted LPS appeared to have more lipid contamin ation than the phenol-water-extracted LPS. When the resolved LPS extractions were comp ared by silver stain to the commercial LPS, the phenol-water-extracted LPS most closely resembled the commercial LPS. In the silver stain for TRIzol-extracted LPS and commercial LPS, the TRIzol-extracted LPS had bands at 14 and 20 kDa with non-staining bands between 20 and 30 kDa, the commercial LPS had bands at 14, 20, and 23 kDa. In the silver stain for phenol -water-extracted LPS and commercial LPS, both LPS preparations had bands at 13, 20, and 23 kDa. The phenol-water-extracted LPS had additional banding from 23-50 kDa, while the co mmercial LPS had more of a smear from 23-50 kDa. Previously published silver stains of V. cholerae O1 LPS have O-antigen banding/smears from 20-50 kDa (80,81). Vibrio cholerae O1 LPS has 12 to 18 O-antigen groups (83); therefore, the additional banding in the phenol -water-extracted LPS may be due to the ability of the phenolwater method to extract LPS with more O-antige ns than that of the method used for the extraction of the commercial LPS which was phenol extracted and purified by gel-filtration chromatography. Because the phenol-water-ext racted LPS most closely resembled the commercial LPS and appeared to have less lipid contamination than the TRIzol-extracted LPS, the phenol-water-extracted LPS was used for later experiments. Before V. cholerae O1 LPS was panned, a solid support capable of binding a high quantity of V. cholerae O1 LPS was analyzed. In the past th e Ph.D. 12mer phage library was panned against antigens that were immobilized onto a microtiter well. Previous experiments from this laboratory suggested that nitroce llulose paper could bind more LPS than a microtiter well. It was

PAGE 123

123 hypothesized that having more LPS present to bind phages would enable more phages that specifically bound to the LPS to be selected. However, increasing the amount of LPS present may also increase the amount of nonspecific phage s selected as well as the amount of specific phages selected. Therefore, more antigen presen t to pan against may not necessarily improve the chances of selecting phages specific to the antigen. To determine which solid support could bind more LPS, the saturation limit of LPS was analyzed on a microtiter well and a piece of nitroce llulose paper with the same surface area as a microtiter well. The microtiter well became saturated with LPS when coated with 0.1-1 g LPS, and the nitrocellulose paper became satu rated with LPS when coated with ~250 g LPS. Therefore, nitrocellulose paper could bind approximately 250-2,500 times more LPS than a microtiter well of equal area. Nitrocellulose pa per has previously been used as a solid support for proteins in panning proce dures involving glutathione S-tr ansferase and hen egg white lysozyme (87). Nitrocellulose paper was chosen as a binding support because it had a high capacity to bind proteins (~100 g/cm2). To determine if panning against more antigen increases the sel ection of phages specific for the antigen, both microtiter well s and nitrocellulose paper were used as solid supports for panning against V. cholerae O1 LPS. Vibrio cholerae O1 LPS (Sigma-Aldrich) was immobilized onto a polystyrene Maxisorp microtiter well and panned with the NEB Ph.D. 12mer phage display library. Five rounds of panning were performed with amplification of phages after the first four rounds of panning. Panning promotes the selection of phages that specifically bind to the target antigen over phage s that are not specific to the an tigen. Amplifying pools of eluted phages between pannings enriches the amount of specific phages. Therefore, each additional round of panning should increase the amount of specific phages in the pool of eluted phages.

PAGE 124

124 One hundred phages from the fifth round of panning were individually amplified and analyzed in an ELISA against V. cholerae O1 LPS, carbonate coating buffer, and PBS. Anti-Vc O1 LPS mAb generated a S:N of 15.5, proving that the LPS successfully bound to the microtiter plate. Anti-BSA phagemid pa rticles were used in the ELISA against BSA (10 g/mL) and generated a S:N of 22, proving that the ELISA successfully detected phages. All 100 clones gave S:N <1.4, whereas a positive signal was classified as a S:N 2. Therefore, all of the clones screened by ELISA were negative for binding to V. cholerae 569B LPS. Because panning the Ph.D. 12mer library against V. cholerae O1 LPS that was immobilized onto a microtiter well failed to yield phages that were positive to V. cholerae O1 LPS, the panning procedure was altered in an attempt to improve the chances of selecting phages that specifically bind to the antigen. The first alteration was to change the antigen binding support from a microtiter well to nitrocellulose. It was determined that nitrocellulose paper could bind more LPS than a microtiter well; ther efore, with more LPS present to pan against more specific phages may be selected. Also, the mechanism of phage elution was analyzed. Previous panning procedures with the Ph.D. library used a glycine (pH 2. 2) elution. There are numerous elution methods, pH, salt, pressure, te mperature, antigen, none of which are optimal for every antigen-antibody complex. The key is to find a method that effectively dissociates the antigen-antibody complex without causing harm to either the antibody or the antigen (88). An acidic elution inte rferes with the electrostatic and hydrophobic interactions of the antigenantibody complex and causes them to dissociate. Acidic elutions do not work for every antigenantibody complex; sometimes the low pH of the elutant can denature the antibodies or antigens. Elution with antigen elutes the antibody from the immobili zed antigen-antibody complex by competing with the immobilized antigen for bindi ng with the antibody. While pH elutions are

PAGE 125

125 the most common elutions used in phage displa y, elutions by antigen are not uncommon and have even been noted to be more successful than pH elutions because th ey decrease the elution of phages that are not speci fic to the antigen (89-91). Panning the Ph.D. 12mer library against phenol-water-extracted V. cholerae O1 LPS that was immobilized onto nitroce llulose paper and eluted with either glycine (pH 2.2) or V. cholerae O1 LPS was performed. Three rounds of panning were performed with amplification of phages after rounds one and two. A subtractive pa nning was performed after round one. This negative panning was performed to decrease th e amount of nonspecific phages in the library. The negative panning involved pann ing the library against bloc ked nitrocellulose paper; therefore, clones specific to the blocker, nitrocellulose paper, or polystyrene should be removed from the library. Two hundred clones from each panning were individually amplified and analyzed by ELISA. Th e 400 phage clones and -Vc O1 LPS mAb were screened in an ELISA against V. cholerae 569B LPS. Anti-Vc O1 LPS mAb gene rated a S:N of 10, proving that the LPS successfully bound to the microtiter plate. Anti-BSA phagemid particles were used in the ELISA against BSA (10 g/mL) and generated a S:N of 15, proving that the ELISA was a successful assay for the detection of phages. All 400 clones gave a S:N less than 1.7; therefore, all of the clones screened were negative for binding to V. cholerae O1 LPS. Previous attempts to select recombinant pha ges to LPS have failed to generate highly specific recombinant phages that generally la ck a consensus sequence (92-94). Also, the isolation of recombinant phages that bind to LPS is less common than for the isolation of recombinant phages that bind to protein. This is most likely because protei n-protein interactions are usually stronger than protei n-carbohydrate interactions. Th e antibody-antigen complex is held together by noncovalent in teractions such as hydrogen bonds, van der Waals forces,

PAGE 126

126 coulombic interactions, and hydrophobic bonds ( 95). The strength of the antibody-antigen complex depends on the strength of the bonds used to hold it togeth er. Protein-protein interactions make use of all noncovalent bonds, but carbohydrate-pro tein interactions primarily make use of hydrogen bonds and rarely make use of hydrophobic or coulombic interactions. Also, protein-protein interac tions can be held together by numerous bonds due to the conformational structure of the protein-protein interaction. Th e carbohydrate-prote in interaction is usually only bound by a single domain due to th e linearity of the car bohydrate structure. Therefore, the low density and the limited va riety of bonds that occu r in carbohydrate-protein interactions make LPS a difficult target to select recombinant proteins that bind to it with high affinity. Specific Aim 2: Improve Panning a nd Screening Process of Biopanning Biopanning is perform ed to select phagemid partic les that specifically bind to a target. To increase the chances of selecti ng phagemid particles that are speci fic to a target, the optimization of the biopanning and screening process was pe rformed. Optimization of the panning process was performed with V. cholerae O1 whole cells as a ta rget, and the concentr ation of whole cells and the trypsin elution time were analyzed. Th e concentration of whole cells used in biopanning usually ranges from 107 to 109 whole cells (96). The goal was to use enough whole cells to enable capture of specific recombinant phagemid particles and keep the capture of phagemid particles not specific to the w hole cells to a minimum. The trypsin elution time was also analyzed. The Tomlinson libraries have a trypsin cleavage site between the pIII protein and the scFv particle. Therefore, trypsin is used to cl eave the phagemid particle from the scFv peptide that is bound to the target. Th e Tomlinson protocol recommends a trypsin elution of 10 minutes. The trypsin elution time was analyzed to maximi ze the number of eluted phagemid particles that were specific to the target. Eluting for too short of a time may fail to elute specific phagemid

PAGE 127

127 particles, but eluting for too long of a time may only in crease the elution of nonspecific phagemid particles or even de grade the phagemid particles. Comparative pannings were performed to determine the optimal concentration of V. cholerae whole cells and optimal trypsin elution time. Parallel pannings were performed with the Tomlinson J scFv library w ith and without 1% Vc86 phagemi d particles spiked into the library. Vc86 is a clone isolated from the To mlinson I scFv library th at recognizes an unknown antigen on V. cholerae whole cells. Because ~1011 phagemids are being used in the pannings there only exists a 100-fold redundancy of clones in the non-spiked library. Therefore, enumeration of eluted phagemid particles afte r one round of panning may not yield significant differences in eluted titers unde r various test parameters. Spiking the library with 1% phagemid particles that are specific to V. cholerae whole cells should enable di stinction of an optimal panning parameter that selects and elutes the highest number of specific recombinant phagemid particles. We were looking for a combination of the concentration of whole cells and trypsin elution time to elute more phagemid particles from the Vc86-spiked library than the non-spiked library. Because the only difference between th e pannings was that the Vc86-spiked library had a higher initial concentration of specific phagemid particles to V. cholerae whole cells, an increase in eluted phagemid particles with a cert ain parameter should correlate to an increase in specific phagemid particles being eluted. The first one-round panning used 1 x 108 or 1 x 109 whole cells of V. cholerae with trypsin elutions of 15, 30, 45, and 60 minutes (Fig. 3-6). Elution of phagem id particles with trypsin was optimal between 15 and 30 minutes. The phagemid particle titer decreased by 80 to 93% from 30 to 45 minutes for the four pannings. The decrea se in the titer of phagemid particles was most likely due to degradation of the phagemid particle s by trypsin. Therefore, trypsin elution should

PAGE 128

128 not exceed 30 minutes. More phagemid partic les were eluted when panned against 1 x 109 whole cells for both the spiked and nonspiked libraries However, there were not enough data to do statistical analysis to determine if using 1 x 109 whole cells eluted sign ificantly more phagemid particles than using 1 x 108 whole cells in the biopanning process. The best whole cell concentra tion and trypsin elution time were not definitively obtained from the first one-round panning; therefore, a second one-round panning was performed to further analyze the trypsin elution time and whole cell concentration. The second one-round panning used 107, 108, or 109 whole cells of V. cholerae with a trypsin elution of 10, 20, and 30 minutes (Fig. 3-7). There were no significant di fferences in the titers of eluted phagemid particles when comparing the con centrations of whole ce lls and trypsin elution times used in the second one-round panning experiment for both library pannings. This could mean that there may not be a significant advantage in using the diffe rent elution times or concentrations of whole cells for selecting specific recombinant phages. Because using titering to analyze the differe nces in the amount of eluted phagemid particles under various parameters failed to definitively determine a panning condition that eluted the highest number of specific recombinant phagemid particles, another one-round panning was performed with analysis of phagemi d particle specificity by ELISA. Following elution of the phagemid particles, the phagemid particles were amplified in batch and analyzed by ELISA to determine which panning parameter el uted the most phagemid particles that were specific to V. cholerae whole cells. The third one-round panning used 108 or 109 whole cells of V. cholerae with trypsin elutions of 10 and 30 minutes (Fig. 3-8). The el uted phagemid particles were amplified in batch and analyzed by ELISA. Of the different panning variations, the amplified phagemid particles

PAGE 129

129 from the panning with the library that contained 1% Vc86 phagemid particles, 108 whole cells V. cholerae, and an elution time of 30 minutes gene rated a significantly higher S:N (p=0.010.04) in an ELISA compared to the amplifie d phagemid particles from the other panning variations. This suggests that panning with 108 whole cells with a tryps in elution time of 30 minutes is optimal for selecting phagemid partic les that are specific to the panned whole cells compared to using 109 whole cells or a trypsin elution of 10 minutes. There might have been an excess of whole cells wh en panning against 109 whole cells. This excess may have promoted the binding of nonspecific phages; therefore, fewer whol e cells present may have been more efficient for panning because with fewer whole cells pr esent the specific phage mid particles should preferentially bind the whole cells and the number of nonspecific phagemid particles would be lessened. Eluting for 10 minutes may not have been long enough for the trypsin to cleave all the specific phagemid particles from the whole cells; therefore, eluting for 30 minutes was optimal because it may have given the trypsin enough tim e to elute the phagemid particles from the whole cells. After the biopanning process is complete, the selected scFvs or phages are screened for their specificity to the target molecule. The previous met hod to screen scFv proteins was to grow amber suppressor-free phagemid-containing bacteria under conditions to secrete scFv proteins in culture tubes or flasks. However, th is process was very time consuming; therefore, a way to screen scFv proteins in a high th roughput manor was investig ated. High throughput screening of scFv proteins is very common. A Biacore A 100 array system is capable of screening hundreds of scFv proteins per day (97). This system sc reens scFv proteins from crude bacterial extracts on a sensor chip surface. Th e antibodies that are captured onto the chip are analyzed and ranked by the per centage of bound scFv proteins remaining on the chip after

PAGE 130

130 dissociation of nonspecific scFv proteins in buffer. His-tagged scFvs can be screened in an automated high throughput approach by measure of surface plasmon resonance using a Qiagen BioRobot 3000 LS (98). A chromatography press is capable of purifying and screening hundreds of scFv proteins per day (99). This sy stem involves His-tagged scFvs being purified by a cation exchange column and immobilized meta l ion affinity chromatography. The proteins were then quantified by a bicinchoninic acid ( BCA) assay and analyzed by SDS-PAGE. While there are many high throughput screens for scFv prot eins, most of the methods involve expensive equipment that is not readily av ailable to our laboratory. Therefore, our goal was to develop a high throughput screen for scFv proteins. A colony blot was attempted to screen scFv proteins. The colony blot involved the transfer of bacterial colonies that were grown to secr ete scFv proteins onto an antigen-coated piece of nitrocellulose paper. The scFv proteins that bou nd to the nitrocellulose via the antigen were then detected with Protein-L peroxidase and identified with ECL. However, when this system was attempted, the colony blot was determined to be neither specific nor selective. When the -BSA clone was screened by this method, it did not generate a positive signal on BSA-coated nitrocellulose but did generate a positive signal on PBS and V. cholerae -coated nitrocellulose paper. Also, when clones se lected in a panning against V. cholerae whole cells were screened by this method, they generated posi tive signals to BSA, PBS, and V. cholerae -coated nitrocellulose papers. It appeared as if random clones bound to the nitrocellulose paper despite it being blocked with casein buffer and th at all clones that produced scFv s were detected. For this problem to occur many steps or just one step co uld have gone wrong in the experiment. It is possible that not enough antigen was used to coat the nitrocellulose, the nitrocellulose did not capture the antigen, the nitrocel lulose was poorly blocked, or th e washes were not stringent

PAGE 131

131 enough. The effort required to tr oubleshoot the colony blot was de termined to not be worth the time. Therefore, a way to improve the old method for producing scFvs was developed. To enable high throughput production of scFv proteins, a clone was grown in a 96-well plate instead of a culture tube or flask, and the scFv-containing supernatant was examined in an ELISA. When -BSA clone was grown in a 96-well pl ate, its scFv-containing supernatant diluted 1:2 yielded a S:N of 11.5 against BSA in an ELISA. Anti-BSA scFv produced in a flask yielded a S:N between 10 and 15; th erefore, producing scFv proteins in a microtiter well may not be optimal, but it is acceptable for the purpose of a high thro ughput screen. Next was to determine how far the microtiter well-produced scFvs could be diluted and still generate a sufficient signal to be detected in an ELISA. This test was to determine how many clones could be mixed together (i.e., diluted) and screened in a single microtit er well to enable even greater high throughput screening of scFvs. The -BSA clone was grown in a microtiter well to produce scFvs. The supernatants were diluted 1:2, 1:4, 1:6, and 1:10 in casein blocking buffer and analyzed by ELISA (Fig. 3-10). When -BSA scFv-containing supernatant was dilute d 1:10 it still yielded a S:N of 5.6 in an ELISA. This S:N is high enough for de tection of positive clones. However, the -BSA clone was a very strong clone; therefore, the majority of clones selected will have a weaker signal than the -BSA clone. To lessen the chance of not detecting scFvs with weak affinity or low titers, it was decided to only dilute scFv clones 1:4 to 1:6 for high throughput screening by ELISA. This would still enable screening of hundreds of clones per day, a vast improvement to the previous method of screening clones in culture tubes or flasks which only enable d screening of a couple dozen clones per day.

PAGE 132

132 Because the high throughput screen for scFv proteins was successful, it was examined if a similar high throughout ELISA could be applied to the screening of phagemid particles. Previous methods of screening phages involved pr oducing phagemid particles in a culture tube or flask and screening the supe rnatants by ELISA. This method was time consuming; therefore, a high throughput screen was investigated. Previous high throughput ELISAs have been developed for the screening of pha ges; however, while their screen ing is rapid, the production of the phages is time consuming. These high thro ughput ELISAs require ov ernight production of phages in a culture tube or flask followed by PEG precipitation of the phages for the concentrated phages to be screened by ELISA (100,101). We hoped to develop an even more high throughput ELISA by not only screening phagemid particles in a microtiter well but also growing the phagemid particles in a microtiter well. Phagemid particles from the -BSA and Vc86 clones were transduced into E. coli TG1 (Hy -sm), which harbors the hyperphage genome as a plasmid. Producing phagemid particles in E. coli TG1 (Hy -sm) removes the additional step of superinfection with hyperphage; therefore, reducing the time required to produ ce phagemid particles. Colonies from the transduction were picked and grown in a microtiter well to produce phagemid particles. Growth temperature and glucose concentr ation of the medium were analyzed. The Tomlinson protocol recommends producing phagemid particles in medi um that has 0.1% (w/v) glucose. However, this concentration of glucose could promote catabolite repression of the lac promoter, which promotes transcription of the gIII gene. Therefore, it was examin ed if phagemid particles would be produced better in medium with or without 0.1% (w/v) glucose. The -BSA and Vc86 phagemid particles were produced in a microtiter well at 30C or 37C with and without 0.1% (w/v) glucose. The phage-containing supernat ants were diluted 1:2

PAGE 133

133 in blocking buffer and analyzed by ELISA (Fig. 311). There was not a significant difference in the S:N (p=0.07-0.48) of phagemid particles produced in microtiter wells with and without 0.1% (w/v) glucose. Therefore, the 0.1% (w/v) glucose in the medium was probably not enough glucose to cause catabol ite repression of the lac promoter. Vc86 phagemid particles produced in microtiter wells with or without glucose at 30 C and 37C did not produce a S:N greater than 1, whereas Vc86 phagemid particles produced in a flas k (S:N ~4) yielded a significantly higher S:N (p=0.05). Observation of the turbidities of the ba cterial cultures that were grown in a microtiter well revealed only slightly turbid cultures compar ed to bacterial cultures that were grown in a flask. The growth conditions may not have been op timal for the bacteria to grow when they had the added stress of producing phagemid particles. It was possible that the bacterial cultures that were grown to produce phagemid particles in a microtiter well experienced less aeration than bacterial cultures grown in a flask or culture tube. The decreased aeration could decrease bacterial growth and reduce phagemid particle production. Escherichia coli is a facultative anaerobe which means it makes ATP, which driv es biosynthesis, by aerobic respiration or fermentation. Aerobic respiration generates 36 ATP molecules from one molecule of glucose, while fermentation only generates 2 ATP molecule s from one molecule of glucose. Because biosynthesis is fueled by ATP, aerobic respira tion is more effective for bacterial growth. When -BSA phagemid particles were produced at 30C, the S:N was significantly higher (p<0.04) than when -BSA phagemid particles were produ ced at 37C at either glucose concentration in a microtiter we ll. It appeared that phagemi d particle production was more permissive at 30C opposed to 37C. The -BSA phagemid particles pr oduced in a flask (S:N ~12) yielded a significantly hi gher (p=0.02) S:N compared to -BSA phagemid particles produced in microtiter wells (S:N~7) when analyzed at a 1:2 dilution in an ELISA. The reduced

PAGE 134

134 S:N of phagemid particles used in an ELISA c ould have been due to less phagemid particles present in the phagemid particle -containing supernatant that was produced in a microtiter well as opposed to the phagemid particle-containing supern atant that was produced in a flask. This could have been due to reduced aeration in the microtiter well causi ng decreased bacterial growth and thus reduced quantities of phagemi d particles produced in a microtiter well. Phagemid particle production in microtit er wells was repeated a second time with -BSA phagemid particles (Fig. 3-12). Anti-BSA pha gemid particles yielded significantly higher signals (p<0.03) when produced at 30C as opposed to 37C at either glucose concentration. These data support the previous da ta that 30C is a more permissi ve temperature than 37C for phagemid particle production. There was not a significant difference (p=0.34) in the signals produced by phagemid particles that were produced with or without glucose at 30C. These data also support the previous data that 0.1% (w/v) glucose in th e medium does not provide an advantage or hindrance to phagemid particle prod uction. While the second experiment did not yield a significant difference (p= 0.09) in signals produced by -BSA phagemid particles produced in a flask opposed to in a microtiter well, the microtiter well-produced phagemid particles yielded lower signals than di d phagemid particles produced in a flask. Because phagemid particles that were produced from the same clone resulted in various ELISA signals when phage-containing supernatants were analyzed at a dilution of 1:2 it was possible that the reason for the differences in ELISA signals from phage-containing supernatants were due to differences in phagemid particle titers produced under the various conditions. Therefore, the titers of -BSA and Vc86 phagemid particles produced in a microtiter well and in a flask were analyzed. Phagemid particles produced at 30C in a microtiter well yielded approximately 107 tu/mL while phagemid particles produced at 37C in a microtiter well yielded

PAGE 135

135 approximately 106 tu/mL. Producing phagemid particles at 30C in microtiter wells yielded approximately 10-fold higher tu/mL titers than phagemid particles that were produced in microtiter wells at 37C. Whether the growth medium contained 0.1% (w/v) glucose or not did not affect the titers of ph agemid particles. When -BSA and Vc86 phagemid particles were produced in flasks at 30C with 0.1% (w/v ) glucose in the medium, the titers were approximately108 to109 tu/mL. Producing phagemid particles in a microtiter well yielded higher titers and signals in an ELISA when produced at 30C. Whether glucose was in the media or not did not affect phagemid particle titers or ELI SA signals. Phagemid particles could not be produced in a microtiter well in high enough titers to make a high throughput screen acceptable. When clone Vc86, a clone of average strength, was produced in a microtiter well, it was not able to generate a positive signal. Therefore, the high throughput ELISA screen for phagemid particles was not acceptable for screening of phagemid particles. Specific Aim 3: Improve Phagemid Particle Production Am plification of phagemid particles is an essential step in biopanning. A phagemid contains the gIII gene of M13 but lacks the rest of the M13 genes needed for phagemid particle production. Therefore, a phagemid re quires an amplification tool to provide the rest of the M13 genes in trans to enable phagemid particle production. It is esse ntial that there is high quality and high quantity phagemid particle production during amplification to ensure that the redundancy of clones in the library is maintained and that the library is in high enough titers to be useful in assays. The most commonly used amplification tool fo r phagemids is helper phage. There are many variations of helper phages, R408, VCSM13, a nd M13KO7 (63) that differ slightly. Of the helper phages, M13KO7 is the most commonly used Because helper phages contain a wild type gIII gene, they are not optimal for phagemid particle amplification because the produced

PAGE 136

136 phagemid particles will express non-recombinant pIII proteins along with recombinant pIII proteins. To overcome the drawback of helper phages encoding a wild type gIII gene, many gIII gene mutated variations of helper phages were engineered. Many of these strains either have a deletion of the gIII gene or encode amber stop codons within the gIII gene to lessen the expression of wild type pIII proteins (65-67). Of the gIII gene mutated helper phages, hyperphage is the most effective for producing high titers of phagemid particles that have multivalent display of pIII-fusion proteins (66). Hyperphage was the previous amplification tool used by our laboratory to amplify phagemid particles; however, th e titer of the hyperphage stock was too low to effectively amplify every phagemid particle in the library. Therefore, a way to improve phagemid particle production was investigated. To overcome the problem of the low titer of the stock of hyperphage from effectively infecting every phagemid-containing bacterium, the hyperphage genome was transduced into E. coli to be maintained as a plasmid to ensure that every phagemid-cont aining bacterium also contained hyperphage to enable amplification of phagemid particles. When hyperphage was transduced into E. coli TG1, small (0.5 mm diameter) and large (1-1.5 mm diameter) colonies resulted. When the small and large colonies were passaged, the small colonies maintained a small phenotype (0.5 mm diameter) and the large colonies maintained a large phenotype (1-1.5 mm diameter). The two strains were named E. coli TG1 (Hy -sm) and E. coli (Hy -lg). The reason for the variation in the colony size was not known. Because colonies were grown under kanamycin resistance for hyperphage, both strain s contained hyperphage, or at least they contained the kanamycin resistance gene that hyperphage carries. The differences in the colony sizes suggested that the bacteria that formed the colonies were not replicating at the same rate. The smaller colonies might be small because they are under stress from replicating the

PAGE 137

137 hyperphage genome along with the E. coli genome, and the large colonies might be larger because they are capable of growing faster beca use they are under less stress because they may not be replicating everything. To determine if the two strains we re functionally different, their infection efficiencies were compared to E. coli TG1 and their ability to produce phagemid particles was compared to E. coli TG1 containing a phagemid th at was superinfected with hyperphage. The infection efficiencies of -BSA and Vc86 into E. coli TG1, E. coli TG1 (Hy -sm), and E. coli (Hy -lg) were analyzed. When the number of -BSA phagemid particles transduced into E. coli TG1 was set at an infection efficiency of 100%, the number of -BSA phagemid particles transduced into E. coli TG1 (Hy -sm) was 28% and the number of -BSA phagemid particles transduced into E. coli TG1 (Hy -lg) was 0.1% of the number of transduced phagemid particles into E. coli TG1. When the number of Vc86 pha gemid particles transduced into E. coli TG1 was set at an infection efficiency of 100%, the number of Vc86 phagemid particles transduced into E. coli TG1 (Hy -sm) was 22% and the number of Vc86 phagemid particles transduced into E. coli TG1 (Hy -lg) was 2% of the number of transduced phagemid particles into E. coli TG1. Escherichia coli TG1 (Hy -sm) had an infection efficiency that was approximately 25% of the infection efficiency of E. coli TG1, and E. coli TG1 (Hy -lg) had an infection efficiency that was approximate ly 1% of the infection efficiency of E. coli TG1. It appears as if E. coli TG1 has a reduced ability to be transd uced with phagemid particles when it is harboring hyperphage. The differences in the infection efficiencies between E. coli TG1 (Hy -sm) and E. coli TG1 (Hy -lg) show that their functionali ties are d ifferent. The reason for E. coli TG1 (Hy -lg) generating larger colonies may ha ve been because it shed a function required in transduction and theref ore was able to grow faster. For example, the production of

PAGE 138

138 the F pilus could have been compromised by the stress from E. coli TG1 replicating the hyperphage genome. Both hyperphage-containing strains of E. coli TG1 had infection efficiencies that were too low for the strains to be used to effectively amplify clones in a librar y. While the hyperphagecontaining E. coli strains have infection efficiencies too low to be used to amplify the library their infection efficiencies are still acceptable for amplif ying single clones because the transduction efficiency is not critical for the am plification of a single cl one. Therefore, the quality and quantity of phagemid particles produced by E. coli TG1 (Hy -sm/lg) were analyzed. Anti-BSA phagemid particles were produced by E. coli TG1 (Hy -sm), E. coli TG1 (Hy -lg), and E. coli TG1 that was superinfected with hyper phage. Anti-BSA phagemid particle yields were as follows: E. coli TG1 that was superinfected with hyperphage yielded 3 x 10 8 tu/mL, E. coli TG1 (Hy -sm) yielded 1 x 109 tu/mL, and E. coli TG1 (Hy -lg) yielded 3 x 108 tu/mL. While the ability of hyperphage-containing E. coli TG1 to be transduced with phagemid particles was reduced comp ared to the ability of E. coli TG1, its ability to produce phagemid particles was not reduced. Phagemid partic les produced by hyperphage -containing bacteria yielded titers high enough to be us ed as a phagemid particle amplification tool for single clones. The qualities of -BSA phagemid particles produced by the three E. coli strains were analyzed by ELISA (Fig. 3-13). Anti-BSA pha gemid particles produced a significantly higher S:N (p=0.01) when produced from E. coli TG1 (Hy -sm) compared to E. coli TG1 (Hy -lg). There was not a significant difference in the S:N (p=0.50) from -BSA phagemid particles produced from superinfected E. coli TG1 and E. coli TG1 (Hy -lg). Because the titer of phagemid particles produced from E. coli TG1 (Hy -sm), ~1 x 109 tu/mL, was three times higher than the titer of phag emid particles produced from E. coli TG1 (Hy -lg) or E. coli TG1

PAGE 139

139 that was superinfected with hyperphage, ~3 x 108 tu/mL, a third less phage-containing supernatant was used in the ELISA for phagemid particles that were produced from E. coli TG1 (Hy -sm). The decrease in the S:N from -BSA phagemid particles produced from E. coli TG1 (Hy -sm) could have resulted if th e particles/mL of the phagemi d particles produced from all three strains were the same. Because hyperphage-containing E. coli TG1 was not going to be used to amplify the library, furthe r analysis of it was stopped, so that the time used to analyze it could be put to better use finding a more eff ective phagemid particle amplification tool. Hyperphage-containing bacteria was not acceptable as an amplif ication tool for the library because of their low infection efficiencies, but E. coli TG1 (Hy -sm) was acceptable for use as an amplification tool for the am plification of single clones. Helper plasmids are another amplification t ool for the production of phagemid particles (64). They are M13-based plas mids that are maintained in E. coli to provide phagemids with all the genes necessary to produce phage mid particles. Helper plasmids were engineered in three forms that differ in the length of their gIII genes. The helper plasmi ds contain a full-length (M13cp), a deleted (M13cp-dg3) or a truncated (M13cp-CT) gIII gene. These three helper plasmids were transformed into E. coli and analyzed for their infection efficiency with phagemid particles, their ability to produ ce phagemid particles, and the qua lity of the phagemid particles produced. When the helper plasmids were transformed into E. coli TG1, the resulting colony sizes varied from a diameter of 0.5 to 2 mm. The reasons for various colony sizes are hypothesized above. A small (0.5 mm) colony and a large (2 mm) colony from each transformation were passaged to determine if the colony size phenotype would be maintained. Escherichia coli TG1 (M13cp-sm) was the only strain that maintained a consistent phenotype of 0.5 mm colonies, the

PAGE 140

140 rest resulted in colonies that had a diameter between 0.5 and 2 mm. A small and large of each strain were further analyzed to determine if the different colony phenotyp es also had different functionality. The six helper plasmid-containing E. coli TG1 strains were analyzed for their infection efficiencies with two phagemid particles in relation to the infec tion efficiencies of E. coli TG1 that did not contain a ny helper plasmids. Escherichia coli TG1 (M13cp-dg3-sm) and E. coli TG1 (M13cp-CT-sm) were the only strains that yielded infection efficiencies greater than 0.3% of the infection efficiency of E. coli TG1 (Table 3-1). Escherichia coli TG1 (M13cp-dg3-sm) had infection efficiencies of 175% and 97%, while E. coli TG1 (M13cp-CT-sm) had infection efficiencies of 7% and 13% compared to the infection efficiency of E. coli TG1. The differences in the infection efficiencies be tween the small and large strain s of helper plasmid-containing E. coli revealed that they had different functionalities. Except for E. coli TG1 (M13cp-dg3-sm), all of the other helper plasmid strains had reduced infection efficiencies compared to E. coli TG1. Therefore, the cost for the E. coli cells maintaining the helper plasmids resulted in a decrease in their infectio n efficiencies. Because E. coli TG1 (M13cp-dg3-sm) and E. coli TG1 (M13cp-CT-sm) had higher infection efficiencies compared to their large phenotypes, it suggested that the larg e strains had a defect in some component of the cell involved in transduction. Escherichia coli TG1 (M13cp-dg3-sm) had the highest inf ection efficiencies of the helper plasmid strains; therefore, it was further tested to determine if the high infection efficiencies could be maintained when tested with additional phagemid particles. Escherichia coli TG1 (M13cp-dg3-sm) had infection efficiencies of 92%, 113%, and 120% compared to the infection efficiencies of E. coli TG1 with Vc86, clone18, and -AV20N3 (Table 3-2). The infection

PAGE 141

141 efficiency of E. coli TG1 (M13cp-dg3-sm) was approximately the same as E. coli TG1; therefore, E. coli TG1 did not suffer any reduction in inf ection efficiency due to maintaining M13cp-dg3-sm. Escherichia coli TG1 (M13cp-CT-sm) was the only othe r helper plasmid strain besides E. coli TG1 (M13cp-dg3-sm) that had an infec tion efficiency close to that of E. coli TG1. However, its infection efficiency was still too low for it to be used as an amplification tool for the library. Therefore, M13c p-CT-sm was transformed into E. coli JM109 to try and increase its infection efficiency. Escherichia coli DH5 F was the bacterial strain that the engineers of the helper plasmids used to work with the help er plasmids; therefore, we tested if an E. coli strain more similar to E. coli DH5 F would generate higher infecti on efficiencies if it contained M13cp-CT-sm. Escherichia coli JM109 and E. coli DH5 F contain relA1 and recA1 while E. coli TG1 does not. However, infection efficiency experiments with E. coli JM109 (M13cpCT-sm) with three different phagemids yiel ded lower infection efficiencies than did E. coli TG1 (M13cp-CT-sm). The average infection efficiency of E. coli TG1 (M13cp-CT-sm) was 47%, while the average infection efficiency of E. coli JM109 (M13cp-CT-sm) was 28%. Therefore, using E. coli JM109 to maintain M13cp-CT-sm did not increase its infection efficiency. When the infection efficiencies of E coli TG1 (M13cp-dg3-sm) were analyzed, this strain yielded an average infecti on efficiency of 120% and E. coli TG1 (M13cp-CT-sm) had an average infection efficiency of 30%. Escherichia coli TG1 (M13cp-dg3-sm) had a significantly higher infection efficiency (p=0.01) than E. coli TG1 (M13cp-CT-sm). Escherichia coli TG1 (M13cp-dg3-sm), E. coli TG1 (M13cp-CT-sm), and E. coli TG1 that was superinfected with hyperphage were anal yzed for their abilities to produce phagemid particles. Phagemid particles were produced with and without 1 mM IPTG in the growth

PAGE 142

142 medium to analyze if IPTG had an effect on phagemid particle production yield. Because the lac promoter drives transcription of the gIII -scFv fusion it was hypothesi zed that the expressing more pIII-scFv fusions would result in more phagemid particles produced. However, when phagemid particles were titere d, the phagemid particles produ ced in growth medium that contained IPTG had the same or lower titers of phagemid particles than when phagemid particles were produced without IPTG in th e growth medium. Because ther e was not an increase in the phagemid particle yields when they were produced with IPTG, it suggests that the amount of pIII protein was not a limiting factor in phagemid particle production. For the titers that had a decrease in phagemid particle yields when produ ced with IPTG the increase in pIII protein may have been harmful to phage assembly because th e excess of pIII protein may have been toxic. Phagemid particles were amplified to titers of 106-108 tu/mL by E. coli TG1 (M13cp-dg3-sm), 105 -106 tu/mL by E. coli TG1 (M13cp-CT-sm), and 108 -109 tu/mL by E. coli TG1 superinfected with commercial hyperphage. Titers of phagemid particles produced from E. coli TG1 that contained helper plasmids were too low for the he lper plasmids to be an effective amplification tool. Phagemid particle produc tion cultures usually contain ~109 CFU/mL; therefore, for the helper plasmids to only generate 105 tu/mL means that only 1 in every 10,000 bacteria produced an infectious phagemid particle. The use of help er plasmids as an amplification tool did not generate high enough titers of phagemid particle s for them to effectively amplify a library. Even though the helper plasmids were determin ed to be an ineffec tive amplification tool, they were further analyzed for the quality of phagemid particles produced by ELISA. Analysis by ELISA revealed that phagemid particles produced by helper plasmids resulted in a S:N of 7 when less than 5 x 104 tu/mL of phagemid particles were used as a primary antibody. Data obtained from our laboratory show ed that phagemid particles are not detectable in an ELISA

PAGE 143

143 unless they are used at concen trations of at least 1 x 106 particles/mL (data not shown). Because the spot titer method measures the number of infectious phagemid pa rticles, it is possible that the total number of phagemid particles far exceeded the number of infectious phagemid particles. To determine if the total number of phagemid par ticles is greater than the number of infectious particles an anti-M13 sandwich ELI SA was used to enumerate the particles/mL of phagemid particle-containing supernatants (Table 3-4). Phagemid particles amplified by E. coli TG1 (M13cp-CT-sm) had particle/mL to tu/mL ratios between 740 and 23,000, E. coli TG1 (M13cpdg3-sm) had particle/mL to tu/mL ratios between 1 and 190, while phagemid particles amplified by E. coli TG1 that was superinfected w ith hyperphage had particle/mL to tu/mL ratios between 6 and 24. Phagemid particles produced by helper plasmid-containing E. coli had particle/mL to tu/mL ratios that were much higher than those for phagemid particles produced by E. coli that was superinfected hyperphage. In biopanning the number of phagemid partic les is not important, the number of infectious phagemid particles is what matters because if a pha gemid particle is not infectious then it can not enter a bacterial cell and be produced to eventua lly be selected and screened. Therefore, even t hough the number of partic les produced by helper plasmids was high, the number of infectious particle s produced was low, making the helper plasmids an inadequate amplification tool for the library. The helper plasmids yielded phagemid particle progeny at titers between 105 to108 tu/mL. In a last attempt to get the helper plasmids to generate high titers of phagemid particles, the M13cp-dg3-sm helper plasmid was incorporated into E. coli MG1655 to determine if a less complex E. coli strain could generate higher ti ters of phagemid particles. Escherichia coli MG1655 was chosen because it has few known muta tions and does not cont ain the F-plasmid. Four different phagemid particles were amplified with E. coli MG1655 (M13cp-dg3-sm), and the

PAGE 144

144 resulting titers were: 2.5 x 107, 6.3 x 106, 2.5 x 107, and 3.9 x 107 tu/mL. Producing phagemid particles from E. coli MG1655 that contained a he lper plasmid did not in crease the titer of phagemid particles produced. The helper plasmids were extensively analyzed and determined to be poor amplification tools. Therefore, hyperpha ge was the best amplification tool, but it was available in titers too low to be used effectively. To overcome the low titers of commercial hyperphage, we made homemade hyperphage. Homemade hyperphage was previo usly made by transducing F+ E. coli cells that harbored a plasmid that encoded the gIII gene, pGTR203, with hyperphage and growing these transduced cells under conditions that promoted phage produc tion. In the past, hyperphage amplified from E. coli TG1 (pGTR203) yielded titers between 107 to 108 tu/mL and E. coli HB2151 (pGTR203) yielded titers between 106 to 107 tu/mL. Many liters of hyperphage had to be produced and concentrated to generate hyperphage in a tite r high enough to be used for a single panning experiment. Therefore, a way to increase the t iter of homemade hyperphage was investigated to generate enough hyperphage to be used in multiple experiments. Escherichia coli TG1 and E. coli HB2151 are both F+ strains. Because hyperphage infect s bacteria by the F pilus, it was hypothesized that some of the hyperphage produced were being taken back into the bacteria which would decrease the titers of the hyperphage produced. Therefore, it was investigated if hyperphage could be produced in high er titers when produced from a FE. coli strain. Escherichia coli MG1655 and E. coli EC100D were the two Fstrains of E. coli chosen to amplify hyperphage. These strains were transformed with pGTR203 encoding the gIII gene and the hyperphage genome, and five isolates from each strain were grown under conditions to promote phage production. The particles/mL and the tu/mL were analyzed (Table 3-5). The average titer of hyperphage produced from the five E. coli EC100D (pGTR203) (Hy ) isolates

PAGE 145

145 was 2.1 x 108.5 x 108 tu/mL. The average titer of hyperphage produced from the five E. coli MG1655 (pGTR203) (Hy ) isolates was 3.0 x 109 .0 x 109 tu/mL. The titer of hyperphage produced from the E. coli MG1655 (pGTR203) (Hy ) isolates was significantly higher (p=0.02) than the titer of hyperphage produced from the E. coli EC100D (pGTR203) (Hy ) isolates. The higher titers produced from E. coli MG1655 might be because it has less known mutations than E. coli EC100D. The ratios of particles/mL to tu/mL of hyperphages produced from E. coli EC100D (pGTR203) (Hy ) were between 0.4 and 7.8. The ratios of particles/mL to tu/mL of hyperphages produced from E. coli MG1655 (pGTR203) (Hy ) were between 0.3 and 1.0. The particle/mL to tu/mL ratios of hyperphage produced from the Fstrains were approximately the same, which means that approximately all of the phagemid particles produced were infectious. Escherichia coli MG1655 (pGTR203) (Hy ) produced higher titers of hyperphage than did both F+ strains. Whether or not the hi gher titers of hyperphage were due to being produced in an Fstrain was not determined because E. coli EC100D (pGTR203) (Hy ) produced approximately the same titers of hyperphage as E. coli TG1 (pGTR203) (Hy ). What was determined was that the bacterial strain used to produce hyper phage had an effect on hyperphage production. Because E. coli MG1655 (pGTR203) (Hy ) 5 produced the highest titers of hyperphage, it was used to produce a large batc h culture of hyperphage. The hyper phage was then concentrated to approximately 1 x 1012 tu in 3 mL, which was concentrated enough to amplify the Tomlinson J scFv phagemid library. The Tomlinson J scFv ph agemid library was amplified with homemade hyperphage with quality control ch ecks by titering to ensure that the amplification of the library generated high quality and high quantity of phagemid particles in a way that the redundancy of the library was maintained.

PAGE 146

146 Specific Aim 4: Isolation of Specific Recombinant Phage to Stx2 Toxin of E. coli O157:H7 Once the Tom linson J scFv library was amplified and homemade hyperphage was improved to enable high quality and high quantity am plification of the library, a target molecule had to be selected for panning. Escherichia coli O157:H7 is an enterohemorrhagic serotype of E. coli that is responsible for approximately 73,000 illnesses and more than 60 deaths per year in the United States (30). Stx toxins kill mammalia n cells and are the major causes of damage in E. coli O157:H7-infected individuals. Escherichia coli O157:H7 can produce Stx toxins 1 and/or 2. Escherichia coli that produces Stx2 toxin appears to be more virulent than E. coli that produces Stx1 or both Stx1 and 2 toxins (36-40). Therefore, Stx2 toxin would be a good target for detection in animal and patient samples. A Stx2 toxin preparation was obt ained that was stated to be 50% pure for Stx2 toxin. The other 50% of the preparation most likely contained bacterial impurities from E. coli HB101 because the Stx2 toxin preparation was made from E. coli HB101 (pMJ100-encodes Stx2 toxin) A 100% pure Stx2 preparation would have been id eal; however, if the preparation contains 50% Stx2 toxin then the most of the clones selected sh ould theoretically be specific to Stx2 toxin. The Tomlinson J scFv library that was amplified with homemade hyperphage was panned against the Stx2 toxin preparation. Three rounds of panning were performed with amplification of eluted phagemid particles by homemade hyperphage between each panning. Every panning used 1 x 1012 phagemid particles. The number of el uted phagemid particles for the three pannings were as follows: 2 x 106 for round one, 2 x 107 for round two, and 2 x 1010 for round three. The increase in the num bers of eluted phagemid particles suggested that phagemid particles specific to Stx2 toxin preparation we re being selected and amplified at each round. Clones from the third round of panning were i ndividually amplified and screened by ELISA. Analysis of 59 clones resulted in 53/59 clones (89%) having a S:N from 2 to 23 with 48/59

PAGE 147

147 clones (81%) having a S:N from 10 to 23 against the Stx2 toxin preparation. Therefore, the panning was successful in isolating clones that bo und to Stx2 toxin preparation. Not all of the clones that were selected bound to Stx2 toxin preparation; 4/59 (7%) of the clones bound to PBStreated wells as well as Stx2-coated wells. The four clones had a signal in an ELISA between 0.31 and 0.87 against PBS-treated wells. These cl ones most likely bound to polystyrene. Even though the immunotube was blocked w ith blocking buffer, the unintentional isolation of clones that bound to the panning vessel is common (84). Of the clones sc reened, 2/59 (3%) did not bind to Stx2 toxin preparation or a microtiter well These clones have unknown binding specificity. It is common to select clones that have unknown binding specificity. The fact that only approximately 3% of my clones had unknown bindi ng specificity proves that the panning was successful in isolating specific r ecombinant phagemid particles. Most of the clones selected from panning rec ognized the Stx2 toxin preparation; however, the Stx2 toxin preparation was only 50% pure fo r Stx2 toxin. Therefore, it was unknown if the clones bound to Stx2 toxin or bacterial impurities in the Stx2 toxin preparation. To determine if the selected clones bound to Stx2 toxin, two anti-Stx2 toxin monoclonal antibodies were examined. One antibody recognized the A-su bunit of Stx2 toxin, wh ile the other antibody recognized the B-subunit of Stx2 toxin. However, these monoclonal antibodies never generated a positive signal in an ELISA against the St x2 toxin preparation. Using the monoclonal antibodies at the manufacturer re commended concentration of 1 g/mL failed to generate a positive signal in an ELISA. It was initially reasoned that the monoclonal antibodies may need to be used at a higher concentra tion to increase the S:N values. Increasing the concentration of monoclonal antibodies in an ELISA to 100 g/mL still failed to produce a positive signal in an ELISA. Therefore, it was hypothesized that the HRP-conjugated anti-mouse secondary antibody

PAGE 148

148 that recognized the anti-Stx2 toxin monoclonal antibodies was not recognizing the anti-Stx2 monoclonal antibodies. However, when another HRP-conjugated anti-mouse secondary antibody was used to recognize the anti-Stx2 t oxin monoclonal antibodies it also failed to produce a positive signal in an ELISA. Both anti-mouse antibodies had been tested within three days of these experiments against murine monocl onal antibodies and produced a positive signal. However, positive controls for the anti-mouse an tibodies were not used in the ELISA that analyzed the two anti-mouse antibodies against Stx2 toxin preparation to determine whether or not they were functional. It was not dete rmined why the monoclonal antibodies did not recognize the Stx2 toxin preparati on in an ELISA. One reason might be that there was not enough Stx2 toxin with a correctly folded binding site in the Stx2 toxin preparation for the monoclonal antibodies to recogni ze. Also, it is possible that the batches of monoclonal antibodies that we received were somehow defective. Because the anti-Stx2 toxin monoclonal antibodi es could not be used to determine the binding specificity of the select ed panning clones in an ELISA, other means of determining the binding specificities of the clones were analyz ed. Three clones that bound to Stx2 toxin preparation were screen ed against Stx2 toxin, E. coli O157:H7 933 (expresses Stx 1 and 2 toxins) periplasmic break fraction, E. coli O157:H7 933 supernatant, E. coli DH5 (pNR100) (expresses Stx2 toxoid) peri plasmic break fraction, and E. coli DH5 periplasmic break fraction. A clone that bound to the E. coli bacterial impurities in the Stx2 toxin preparation would be expected to bind to all of the an tigens, but a clone that bound to Stx2 toxin would be expected to bind to all of the antigens except the E. coli DH5 periplasmic break fraction. However, the three selected clones only recognized the Stx2 to xin preparation and none of the other antigens; therefore, it was not determined if the cl ones were actually binding to Stx2 toxin.

PAGE 149

149 Analysis of clones by ELISA proved that the ma jority of the selected clones from panning bound to the Stx2 toxin preparation but failed to determine if th e clones bound to Stx2 toxin because the anti-Stx2 monoclona l antibodies and comparative an alysis of +/Stx2 bacterial preparations failed to generate positive signals in an ELISA. However, when both of the antiStx2 toxin monoclonal antibodies were examined in a Western blot against SDS-PAGE resolved Stx-2 toxin preparation, the anti-Stx2 A s ubunit toxin monoclonal antibody successfully recognized the Stx2 toxin A subunit in the Stx2 toxin preparation. Therefore, analysis by Western blot could at least determine if any of the clones recognized the A subunit of Stx2 toxin. Six clones were screened in a Western bl ot against SDS-PAGE resolved Stx2 toxin preparation, E. coli O157:H7 933 (expresses Stx1 and 2 toxins) supernatant, and E. coli O157:H7 87-23 (Stx phage cured) supernatant (Fig. 3-17). Because the Stx2 toxin is composed of A and B subunits that are approximately 32 and 10-kDa prot eins, a clone that rec ognized Stx2 toxin was expected to recognize a protein band at 32 kDa or 10 kDa on the Stx2 toxin preparation and the E. coli O157:H7 933 supernatant but not on the E. coli O157:H7 87-23 supernatant. If a clone recognized bacterial impurities th en it would be expected to re cognize proteins in all three antigens. The anti-Stx2 A subunit monoclonal antibody recognized a 32-kDa protein in the commercial Stx2 toxin preparation and in the E. coli O157:H7 933 supernatant proving that Stx2 A subunit was present in both preparations and not in E. coli O157:H7 87-23 supernatant, as was expected. Anti-BSA was also analyzed against the three antigens as a negative control to ensure that phagemid particles did not ge nerate artifact bands. Anti-BSA failed to recognize any of the antigens, as was expected. Clone 46 failed to rec ognize any of the three antigens. It is possible that the binding site of the targ et that it recognized in the Stx2 toxin preparation was denatured by SDS-PAGE. Clone 48 recognized 35, 45, and 55-kD a proteins in the St x2 toxin preparation

PAGE 150

150 and a 35-kDa protein in both E. coli O157:H7 supernatants, which means that it probably recognizes bacterial protein impuriti es in the commercial Stx2 toxin Clones 49 and 80 bound to a 35-kDa protein in all three antigens and a 20-kD a protein in the Stx2 toxin preparation and in E. coli O157:H7 87-23. Therefore, clones 49 an d 80 probably bound to bacterial protein impurities in the commercial Stx2 toxin. Clones 81 and 83 appeared to recognize the same bacterial protein impurities as clones 49 and 80, except that thei r band intensities were much lighter. Screening by ELISA showed that 89% of the clones bound to the Stx2 toxin preparation. Western blot analysis of six clones revealed that none recognized Stx2 toxin and that five of the clones most likely recognized bacterial impurities in the Stx2 toxin preparation. Because both of the anti-Stx2 monoclonal antibodies did not recognize Stx2 toxin in the Stx2 toxin preparation in an ELISA and the anti-Stx2 B subunit monoclona l antibody did not recognize Stx2 B subunit in a Western blot, we assumed that there was not a high concentration of Stx2 toxin in the Stx2 toxin preparation. The Stx2 toxin preparation was stated to be 50% pur e for Stx2 toxin by SDSPAGE analysis. To determine if this was true the Stx2 toxin preparation was resolved by SDSPAGE and stained by Coomassie blue (Fig. 3-18). If the Stx2 toxin preparation was 50% Stx2 toxin, 32-kDa and 10-kDa protein bands should have been the most prominent bands. However, SDS-PAGE analysis of the Stx2 toxin preparation showed that a 32-kDa protein band was a minor protein band that represente d less than 5% of the protein in the preparation. There was a band at 10 kDa that had a strong band intensity. These bands might not have even represented Stx2 toxin subunits but may have just been protei ns with a similar molecular weight. What is certain is the amount of Stx2 toxin in the Stx2 toxin preparation did not comprise 50% or even 10% of the proteins in the Stx2 t oxin preparation. Theref ore, it is probable th at the majority of

PAGE 151

151 the clones screened from the panning recognized bacterial impurities from E. coli and not Stx2 toxin. Extensive screening of add itional clones may result in isolati on of a clone specific to Stx2 toxin; however, it is re commended to try panning again with a better Stx2 toxin preparation. The panning was successful in selecting phagemid par ticles that bound specifica lly to the Stx2 toxin preparation, the phagemid particles selected just did not have the desired binding specificity. It is recommended to use the newly amplified Tomlin son scFv library with the aid of the improved homemade hyperphage to pan against another target with a high degree of purity. This thesis details the optimization of phage display protocols and r eagents. Analysis of extraction methods for LPS were compared and it was determined that a phenol-water method, opposed to a TRIzol method, extracted V. cholerae O1 LPS that most closely resembled commercially acquired V. cholerae O1 LPS. Also, binding platforms were analyzed to determine which platform could bind the most LPS. It was determined th at nitrocellulose paper could bind approximately 250 to 2,500 times more L PS than that a microtiter well. Optimization of whole cell panning with the Tomlinson J scFv phagemid library revealed that a trypsin elution between 10 and 30 minutes eluted th e highest number of phagemid pa rticles; this may be due to phagemid particle degradation occu rring with trypsin treatment af ter 45 minutes. Screening of phage display reagents was analyzed and a high throughput ELISA was developed for the production and screening of scFv prot eins that was able to generate S:N values of produced scFv proteins from an ELISA when diluted 1:10. A similar high throughput ELISA was developed for the production and screening of phagemid partic les; however, this method was not acceptable due to low titers produced of phagemid particles in a microtiter well. The optimization of phagemid particle production was performed. Multiple phagemid particle amplification tools were analyzed for th eir infection efficiencies yields of phagemid

PAGE 152

152 particles produced, and the quality of phagemi d particles produced. Hyperphage-containing E. coli resulted in unacceptable infection effici encies and helper plasmid-containing E. coli resulted in unacceptable phagemid particle pr oduction titers. However, improvement of hyperphage that was produced in our laboratory was performed by use of E. coli MG1655 (pGTR203) which generated high titer and hi gh quality hyperphage. This high quality hyperphage was used to successfully amplify the Tomlinson J scFv phagemid library to ensure that the redundancy of phagemid clones was mainta ined. This library was then panned against an Stx2 toxin preparation and resulted in 89% of phagemid particles fr om the third round of panning recognizing the Stx2 toxin preparation in an ELISA. Th e efficiency of this panning was greatly improved from previous pannings from this laboratory, which usually resulted in no better than 1% of selected phagemids being specif ic to the target antige n, and often no isolation of usable specific clones. Theref ore, the optimization of methods a nd reagents in this thesis will greatly improve our laboratory and other laboratories chances of selecting recombinant phages specific to a target molecule.

PAGE 153

153 CHAPTER 5 EPILOGUE The hom emade hyperphage-amplified Tomlinson J library produced in these studies was used to pan against E. coli O157:H7 flagella and yielded ov er 80% positive phagemid clones. The phagemids recognized the major flagelli n protein by Western blot. The high throughput scFv screen was used in the E. coli O157:H7 flagellum project and proved useful in identifying rare scFv-secreting clones. Therefore, the im provements on phage display techniques and tools described in this thesis are, in fact, useful and offer promise of success for continuing studies in the laboratory.

PAGE 154

154 LIST OF REFERENCES 1. Lim, D.V., Simpson, J.M ., Kearns, E.A., and Kramer, M. F. 2005. Current and developing technologies for monitoring agents of bioterrorism and biowarfare. Clin. Microbiol. Rev. 18:583-607. 2. Peruski, L.F., Jr. and Peruski, A.H. 2003. Rapid diagnostic assays in the genomic biology era: detection and identificat ion of infectious disease and biological weapon agents. Biotechniques 35:840-846. 3. Cieslak, T.J., Rowe, J.R., Kortepeter, M.G., Madsen, J.M., Newmark, J., Christopher, G.W., Culpepper, R.C., and Eitzen, E.M., Jr. 2000. A field-expedient algorithmic approach to the clinical management of chemical and biol ogical casualties. Mil. Med. 165:659-662. 4. Franz, D.R., Jahrling, P.B., Friedlander, A.M., McClain, D.J., Hoover, D.L., Bryne, W.R., Pavlin, J.A., Christopher, G.W., and Eitzen, E.M., Jr. 1997. Clinical recognition and management of patients exposed to biological warfare agents. JAMA 278:399-411. 5. Franz, D.R., Jahrling, P.B., McClain, D. J., Hoover, D.L., Byrne, W.R., Pavlin, J.A., Christopher, G.W., Cieslak, T.J ., Friedlander, A.M., and Eitzen, E.M., Jr. 2001. Clinical recognition and management of patients exposed to biological warfare agents. Clin. Lab Med. 21:435-473. 6. YALOW, R.S. and BERSON, S.A. 1959. Assay of plasma insulin in human subjects by immunological methods. Nature 184 (Suppl 21):1648-1649. 7. Saiki, R.K., Scharf, S., Faloona, F., Mulli s, K.B., Horn, G.T., Erlich, H.A., and Arnheim, N. 1992. Enzymatic amplification of beta-globi n genomic sequences and restriction site analysis for diagnosis of sickle cell anemia. 1985. Biotechnology 24:476-480. 8. Kaper, J.B., Morris, J.G., Jr., and Levine, M.M. 1995. Cholera. Clin. Microbiol. Rev. 8:48-86. 9. Sack, D.A., Sack, R.B., Nair, G.B., and Siddique, A.K. 2004. Cholera. Lancet 363:223233. 10. Gaffga, N.H., Tauxe, R.V., and Mintz, E.D. 2007. Cholera: a ne w homeland in Africa? Am. J. Trop. Med. Hyg. 77:705-713. 11. Broeck, D.V., Horvath, C., and De Wolf, M. 2007. Pathogens in focus Vibrio Cholerae: Cholera toxin. The International Journal of Biochemistry & Cell Biology 39:1771-1775. 12. Sanchez, J. and Holmgren, J. 2005. Viru lence factors, pathogenesis and vaccine protection in cholera and ETEC diarrhea. Curr. Opin. Immunol. 17:388-398. 13. 2001. Cholera vaccines. Wkly. Epidemiol. Rec. 76:117-124.

PAGE 155

155 14. Ansaruzzaman, M., Bhuiyan, N.A., Safa, A., Sultana, M., McUamule, A., Mondlane, C., Wang, X.Y., Deen, J.L., von Seidlein, L., Cl emens, J.D., Lucas, M., Sack, D.A., and Balakrish, N.G. 2007. Genetic dive rsity of El Tor strains of Vibrio cholerae O1 with hybrid traits isolated from Bangladesh and Mozambique. Int. J. Med. Microbiol. 297:443449. 15. Almeida, R.J., Hickman-Brenner, F.W., So wers, E.G., Puhr, N.D., Farmer, J.J., III, and Wachsmuth, I.K. 1990. Comparison of a latex a gglutination assay and an enzyme-linked immunosorbent assay for detecting cholera toxin. J. Clin. Microbiol. 28:128-130. 16. Yam, W.C., Lung, M.L., and Ng, M.H. 1992. Evaluation and optimization of a latex agglutination assay for dete ction of cholera toxin and Escherichia coli heat-labile toxin. J. Clin. Microbiol. 30:2518-2520. 17. Andrade, J.R., da Silva, J.L., Barbosa, M.F., and Caldas, C.A. 1992. Preliminary report on the application of the coagglutination test for rapid diagnosis of cholera in the upper Solimoes River area in the Brazilian Amazon region. Braz. J. Med. Biol. Res. 25:375378. 18. Carillo, L., Gilman, R.H., Mantle, R.E., N unez, N., Watanabe, J., Moron, J., Quispe, V., and Ramirez-Ramos, A. 1994. Rapid detection of Vibrio cholerae O1 in stools of Peruvian cholera patients by using monoc lonal immunodiagnostic kits. Loyaza Cholera Working Group in Peru. J. Clin. Microbiol. 32:856-857. 19. Colwell, R.R., Hasan, J.A., Huq, A., Loomis L., Siebeling, R.J., Torres, M., Galvez, S., Islam, S., Tamplin, M.T., and Bernstein, D. 1992. Development and evaluation of a rapid, simple, sensitive, monoclonal antibody-based co-a gglutination test for direct detection of Vibrio cholerae 01. FEMS Microbiol. Lett. 76:215-219. 20. Jesudason, M.V., Thangavelu, C.P., and Lalitha, M.K. 1984. Rapid screening of fecal samples for Vibrio cholerae by a coagglutination technique. J. Clin. Microbiol. 19:712713. 21. Rahman, M., Sack, D.A., Wadood, A., Yasmin, M., and Latif, A. 1989. Rapid identification of Vibrio cholerae serotype O1 from primary isolation plates by a coagglutination test. J. Med. Microbiol. 28:39-41. 22. Agarwal, V., Biswas, M., Pathak, A.A., and Saoji, A.M. 1995. Rapid detection of Vibrio cholerae 0139 in faecal specimens by coagglutination. Indian J. Med. Res. 101:55-56. 23. Hasan, J.A., Huq, A., Nair, G.B., Garg, S., Mukhopadhyay, A.K., Loomis, L., Bernstein, D., and Colwell, R.R. 1995. Development and testing of monoclonal antibody-based rapid immunodiagnostic test kits for direct detection of Vibrio cholerae O139 synonym Bengal. J. Clin. Microbiol. 33:2935-2939.

PAGE 156

156 24. Qadri, F., Hasan, J.A., Hossain, J., Chowdhury, A., Begum, Y.A., Azim, T., Loomis, L., Sack, R.B., and Albert, M.J. 1995. Evaluati on of the monoclonal antibody-based kit Bengal SMART for rapid detection of Vibrio cholerae O139 synonym Bengal in stool samples. J. Clin. Microbiol. 33:732-734. 25. Hasan, J.A., Bernstein, D., Huq, A., Loomis, L., Tamplin, M.L., and Colwell, R.R. 1994. Cholera DFA: an improved direct fluorescen t monoclonal antibody staining kit for rapid detection and enumeration of Vibrio cholerae O1. FEMS Microbiol. Lett. 120:143-148. 26. Hasan, J.A., Huq, A., Tamplin, M.L., Siebeling, R.J., and Colwell, R.R. 1994. A novel kit for rapid detection of Vibrio cholerae O1. J. Clin. Microbiol. 32:249-252. 27. Nato, F., Boutonnier, A., Rajerison, A., Grosjean, P., Dartevelle, S., Guenole, A., Bhuiyan, N.A., Sack, D.A., Nair, G.B., Four nier, J.A., and Chanteau, S. 2003. One-step immunochromatographic dipstick tests for rapid detection of Vibrio cholerae O1 and O139 in stool samples. Clinical and Diagnostic Laboratory Immunology 10:476-478. 28. Nataro, J.P. and Kaper, J.B. 1998. Diarrheagenic Escherichia coli Clin. Microbiol. Rev. 11:142-201. 29. Hussein, H.S. 2007. Prevalence and pa thogenicity of Shiga toxin-producing Escherichia coli in beef cattle and their products. J. Anim Sci. 85:E63-E72. 30. Mead, P.S., Slutsker, L., Dietz, V., McCai g, L.F., Bresee, J.S., Shapiro, C., Griffin, P.M., and Tauxe, R.V. 1999. Food-related illne ss and death in the United States. Emerg. Infect. Dis. 5:607-625. 31. Boyd, B. and Lingwood, C. 1989. Verotoxin rece ptor glycolipid in human renal tissue. Nephron 51:207-210. 32. Robinson, L.A., Hurley, R.M., Lingwood, C., and Matsell, D.G. 1995. Escherichia coli verotoxin binding to human paediatric glomerular mesangial cells. Pediatr. Nephrol. 9:700-704. 33. Takeda, T., Dohi, S., Igarashi, T., Yama naka, T., Yoshiya, K., and Kobayashi, N. 1993. Impairment by verotoxin of tubular function contributes to the renal damage seen in haemolytic uraemic syndrome. J. Infect. 27:339-341. 34. O'Brien, A.D., Tesh, V.L., Donohue-Rolfe A., Jackson, M.P., Olsnes, S., Sandvig, K., Lindberg, A.A., and Keusch, G.T. 1992. Shiga to xin: biochemistry, genetics, mode of action, and role in pathogenesis. Curr. Top. Microbiol. Immunol. 180:65-94. 35. Strockbine, N.A., Marques, L.R., Ne wland, J.W., Smith, H.W., Holmes, R.K., and O'Brien, A.D. 1986. Two toxin-converting phages from Escherichia coli O157:H7 strain 933 encode antigenically distinct toxins with similar biologic activities. Infect. Immun. 53:135-140.

PAGE 157

157 36. Yamagami, S., Motoki, M., Kimura, T., Izumi, H., Takeda, T., Katsuura, Y., and Matsumoto, Y. 2001. Efficacy of postinfection tr eatment with anti-Shiga toxin (Stx) 2 humanized monoclonal antibody TM A-15 in mice lethally challenged with Stx-producing Escherichia coli J. Infect. Dis. 184:738-742. 37. Ostroff, S.M., Tarr, P.I., Neill, M.A., Le wis, J.H., Hargrett-Bean, N., and Kobayashi, J.M. 1989. Toxin genotypes and plasmid profiles as determinants of systemic sequelae in Escherichia coli O157:H7 infections. J. Infect. Dis. 160:994-998. 38. Boerlin, P., McEwen, S.A., Boerlin-Petzo ld, F., Wilson, J.B., Johnson, R.P., and Gyles, C.L. 1999. Associations between virule nce factors of Shiga toxin-producing Escherichia coli and disease in humans. J. Clin. Microbiol. 37:497-503. 39. Tesh, V.L., Burris, J.A., Owens, J.W., Gordon, V.M., Wadolkowski, E.A., O'Brien, A.D., and Samuel, J.E. 1993. Comparison of the relative toxicities of Shig a-like toxins type I and type II for mice. Infect. Immun. 61:3392-3402. 40. Donohue-Rolfe, A., Kondova, I., Oswal d, S., Hutto, D., and Tzipori, S. 2000. Escherichia coli O157:H7 strains that express Sh iga toxin (Stx) 2 alone are more neurotropic for gnotobiotic pi glets than are isotypes producing only Stx1 or both Stx1 and Stx2. J. Infect. Dis. 181:1825-1829. 41. Magliulo, M., Simoni, P., Gu ardigli, M., Michelini, E., Luciani, M., Lelli, R., and Roda, A. 2007. A rapid multiplexed chemiluminescent immunoassay for the detection of Escherichia coli O157:H7, Yersinia enterocolitica Salmonella typhimurium and Listeria monocytogenes pathogen bacteria. J. Agric. Food Chem. 55:4933-4939. 42. Wang, L.J., Wu, C.S., Hu, Z.Y., Zha ng, Y.F., Li, R., and Wang, P. 2008. Sensing Escherichia coli O157:H7 via frequency shift through a self-assembled monolayer based QCM immunosensor. J. Zhejiang. Univ Sci. B 9:121-131. 43. Johnston, L.M., Elhanafi, D., Drake, M., and Jaykus, L.A. 2005. A simple method for the direct detection of Salmonella and Escherichia coli O157:H7 from raw alfalfa sprouts and spent irrigation water using PCR. J. Food Prot. 68:2256-2263. 44. Jin, H.Y., Tao, K.H., Li, Y.X., Li, F.Q ., and Li, S.Q. 2005. Microarray analysis of Escherichia coli O157:H7. World J. Gastroenterol. 11:5811-5815. 45. Thevenot, D.R., Toth, K., Durst, R.A., and Wilson, G.S. 2001. Electrochemical biosensors: recommended definitions and classification. Biosens. Bioelectron. 16:121131. 46. Theegala, C.S., Small, D.D., and Monroe, W.T. 2008. Oxygen electrode-based single antibody amperometric biosensor for qualitative detection of E. coli and bacteria in water. J. Environ. Sci. Health A Tox. Hazard. Subst. Environ. Eng 43:478-487.

PAGE 158

158 47. Liu, Y., Gilchrist, A., Zhang, J., and Li, X.F. 2008. Detection of viable but nonculturable Escherichia coli O157:H7 bacteria in drinking water and river water. Appl. Environ. Microbiol. 74:1502-1507. 48. Kalnauwakul, S., Phengmak, M., Kongmu ang, U., Nakaguchi, Y., and Nishibuchi, M. 2007. Examination of diarrheal stools in Hat Yai City, South Thailand, for Escherichia coli O157 and other diarrheagenic Escherichia coli using immunomagnetic separation and PCR method. Southeast Asian J. Trop. Med. Public Health 38:871-880. 49. Stefan, A., Scaramagli, S., Bergami, R., Mazzini, C., Barbanera, M., Perelle, S., and Fach, P. 2007. Real-time PCR and enzyme-linked fluorescent assay methods for detecting Shiga-toxin-producing Escherichia coli in mincemeat samples. Can. J. Microbiol. 53:337-342. 50. Zhang, W., Bielaszewska, M., Pulz, M., Becker, K., Friedrich, A.W., Karch, H., and Kuczius, T. 2008. New immuno-PCR assay for detection of low concentrations of shiga toxin 2 and its variants. J. Clin. Microbiol. 46:1292-1297. 51. Auvray, F., Lecureuil, C., Tache, J., Leclerc, V., Deperrois, V., and Lombard, B. 2007. Detection, isolation and characte rization of Shiga toxin-producing Escherichia coli in retail-minced beef using PCR-based techni ques, immunoassays and colony hybridization. Lett. Appl. Microbiol. 45:646-651. 52. Smith, G.P. 1985. Filamentous fusion phage: novel expression vectors that display cloned antigens on the virion surface. Science 228:1315-1317. 53. Carmen, S. and Jermutus, L. 2002. Concepts in antibody phage display. Brief. Funct. Genomic. Proteomic. 1:189-203. 54. Endemann, H. and Model, P. 1995. Location of filamentous phage minor coat proteins in phage and in infected cells. Journal of Molecular Biology 250:496-506. 55. Henry, T.J. and Pratt, D. 1969. The proteins of bacteriophage M13. Proc. Natl. Acad. Sci. U. S. A 62:800-807. 56. Russel, M. and Model, P. 1989. Genetic analysis of the fila mentous bacteriophage packaging signal and of the prot eins that interact with it. J. Virol. 63:3284-3295. 57. Messing, J. 1991. Cloning in M13 phage or how to use biology at its best. Gene 100:312. 58. Kwasnikowski, P., Kristensen, P., an d Markiewicz, W.T. 2005. Multivalent display system on filamentous bacteriopha ge pVII minor coat protein. J. Immunol. Methods 307:135-143. 59. Gao, C., Mao, S., Lo, C.H., Wirsching, P., Lerner, R.A., and Janda, K.D. 1999. Making artificial antibodies: a format for phage disp lay of combinatorial heterodimeric arrays. Proc. Natl. Acad. Sci. U. S. A 96:6025-6030.

PAGE 159

159 60. Iannolo, G., Minenkova, O., Petruzzelli, R., and Cesareni, G. 1995. Modifying filamentous phage capsid: limits in the size of the major capsid protein. J. Mol. Biol. 248:835-844. 61. Scott,J.K. and Barbas,C.F.I. 2001. Phage display vectors. In: Ph age Display Laboratory Manual. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY. 2.1-2.19 pp. 62. Tomlinson, I.M., Walter, G., Marks, J.D., Llewelyn, M.B., and Winter, G. 1992. The repertoire of human germline VH sequences reveals about fifty groups of VH segments with different hypervariable loops. J. Mol. Biol. 227:776-798. 63. Rakonjac, J., Jovanovic, G., and Model, P. 1997. Filamentous phage infection-mediated gene expression: construction and propagation of the gIII deletion mutant helper phage R408d3. Gene 198:99-103. 64. Chasteen, L., Ayriss, J., Pavlik, P., and Bradbury, A.R. 2006. Eliminating helper phage from phage display. Nucleic Acids Res. 34:e145. 65. Baek, H., Suk, K.H., Kim, Y.H., and Cha, S. 2002. An improved helper phage system for efficient isolation of specific an tibody molecules in phage display. Nucleic Acids Res. 30:e18. 66. Rondot, S., Koch, J., Breitling, F., and Dubel, S. 2001. A helper phage to improve singlechain antibody presentation in phage display. Nat. Biotechnol. 19:75-78. 67. Soltes, G., Barker, H., Marmai, K., P un, E., Yuen, A., and Wiersma, E.J. 2003. A new helper phage and phagemid vector system improves viral display of antibody Fab fragments and avoids propagation of insert-less virions. J. Immunol. Methods 274:233244. 68. Soltes, G., Hust, M., Ng, K.K., Bansal, A., Fi eld, J., Stewart, D.I., Dubel, S., Cha, S., and Wiersma, E.J. 2007. On the influence of vector design on antibody phage display. J. Biotechnol. 127:626-637. 69. Wen, S.X., Teel, L.D., Judge, N.A., and O'Brien, A.D. 2006. Genetic toxoids of Shiga toxin types 1 and 2 protect mice agains t homologous but not heterologous toxin challenge. Vaccine 24:1142-1148. 70. Weinstein, D.L., Jackson, M.P., Perera, L.P., Holmes, R.K., and O'Brien, A.D. 1989. In vivo formation of hybrid toxins comprising Shig a toxin and the Shiga-like toxins and role of the B subunit in localiza tion and cytotoxic activity. Infect. Immun. 57:3743-3750. 71. Guyer, M.S., Reed, R.R., Steitz, J.A., and Low, K.B. 1980. Identification of a sex-factoraffinity site in Escherichia coli as gw. Cold. Spring. Harb. Symp. Quant. Biol. 45:135140.

PAGE 160

160 72. Riley, L.W., Remis, R.S., Helgerson, S.D., McGee, H.B., Wells, J.G., Davis, B.R., Hebert, R.J., Olcott, E.S., Johnson, L.M., Harg rett, N.T., Blake, P.A., and Cohen, M.L. 1983. Hemorrhagic colitis associated with a rare Escherichia coli serotype. N. Engl. J. Med. 308:681-685. 73. Karpman, D., Connell, H., Svensson, M., Scheutz, F., Alm, P., and Svanborg, C. 1997. The role of lipopolysaccharide and Sh iga-like toxin in a mouse model of Escherichia coli O157:H7 infection. J. Infect. Dis. 175:611-620. 74. Bik, E.M., Gouw, R.D., and M ooi, F.R. 1996. DNA fingerprinting of Vibrio cholerae strains with a novel insertion sequence element: a tool to identify epidemic strains. J. Clin. Microbiol. 34:1453-1461. 75. Heidelberg, J.F., Eisen, J.A., Nelson, W. C., Clayton, R.A., Gwinn, M.L., Dodson, R.J., Haft, D.H., Hickey, E.K., Peterson, J.D., Um ayam, L., Gill, S.R., Nelson, K.E., Read, T.D., Tettelin, H., Richardson, D., Ermolaeva, M.D., Vamathevan, J., Bass, S., Qin, H., Dragoi, I., Sellers, P., McDonald, L., Utterback, T., Fleishmann, R.D., Nierman, W.C., and White, O. 2000. DNA sequence of both chromosomes of the cholera pathogen Vibrio cholerae. Nature 406:477-483. 76. Evans,T.J. 2000. Septic Shock Methods and Protocols. In Methods in Molecular Medicine. Humana Press, 19-23. 77. Yi, E.C. and Hackett, M. 2000. Rapid is olation method for lipopolysaccharide and lipid A from gram-negative bacteria. Analyst 125:651-656. 78. Tsai, C.M. and Frasch, C.E. 1982. A sensitive silver stain for detecting lipopolysaccharides in polyacrylamide gels. Anal. Biochem. 119:115-119. 79. Sasse, J. and Gallagher, S.R. 2003. Detecti on of proteins on blot transfer membranes. Curr. Protoc. Mol. Biol. Chapter 10:Unit10. 80. Waldor, M.K., Colwell, R., and Mekalanos, J.J. 1994. The Vibrio cholerae O139 serogroup antigen includes an O-antigen capsule and lipopolysaccharide virulence determinants. Proc. Natl. Acad. Sci. U. S. A 91:11388-11392. 81. Villeneuve, S., Boutonnier, A., Mulard, L.A., and Fournier, J.M. 1999. Immunochemical characterization of an Ogawa-Inab a common antigenic determinant of Vibrio cholerae O1. Microbiology 145 ( Pt 9):2477-2484. 82. Westphal, O., Luderitz, O., and Bister, F. 1952. Uber Die Extraktion Von Bakterien Mit Phenol Wasser. Zeitschrift fur Naturforschung Section B-A Journal of Chemical Sciences 7:148-155. 83. Villeneuve, S., Souchon, H., Riottot, M.M ., Mazie, J.C., Lei, P., Glaudemans, C.P., Kovac, P., Fournier, J.M., and Alzari, P.M. 2000. Crystal structure of an anticarbohydrate antibody directed against Vibrio c holerae O1 in complex with antigen: molecular basis for serotype specificity. Proc. Natl. Acad. Sci. U. S. A 97:8433-8438.

PAGE 161

161 84. Adey, N.B., Mataragnon, A.H., Rider, J.E., Carter, J.M., and Kay, B.K. 1995. Characterization of phage that bind plas tic from phage-displayed random peptide libraries. Gene 156:27-31. 85. Persell, D.J. and Robinson, C.H. 2008. Det ection and early identific ation in bioterrorism events. Fam. Community Health 31:4-16. 86. 2001. Update: Investigation of bioterrorism-related anthrax and interim guidelines for clinical evaluation of pers ons with possible anthrax. MMWR Morb. Mortal. Wkly. Rep. 50:941-948. 87. Nakamura, M., Watanabe, H., Nishimiya, Y., Tsumoto, K., Ishimura, K., and Kumagai, I. 2001. Panning of a phage VH library using nitrocellulose membranes: application to selection of a human VH library. J. Biochem. 129:209-212. 88. Yarmush, M.L., Antonsen, K.P., Sundaram, S., and Yarmush, D.M. 1992. Immunoadsorption: strategies for antigen el ution and production of reusable adsorbents. Biotechnol. Prog. 8:168-178. 89. Nagumo, Y., Oguri, H., Tsumoto, K., Shindo, Y., Hirama, M., Tsumuraya, T., Fujii, I., Tomioka, Y., Mizugaki, M., and Kumagai, I. 2004. Phage-display selection of antibodies to the left end of CTX3C using synthetic fragments. J. Immunol. Methods 289:137-146. 90. Goletz, S., Christensen, P.A., Kristense n, P., Blohm, D., Tomlin son, I., Winter, G., and Karsten, U. 2002. Selection of large divers ities of antiidiotypic antibody fragments by phage display. J. Mol. Biol. 315:1087-1097. 91. Meulemans, E.V., Slobbe, R., Wasterval, P., Ramaekers, F.C., and van Eys, G.J. 1994. Selection of phage-displayed antibodies specifi c for a cytoskeletal antigen by competitive elution with a monoclonal antibody. J. Mol. Biol. 244:353-360. 92. Zhu, Y., Ho, B., and Ding, J.L. 2003. Seque nce and structural diversity in endotoxinbinding dodecapeptides. Biochim. Biophys. Acta 1611:234-242. 93. Thomas, C.J., Sharma, S., Kumar, G. Visweswariah, S.S., and Surolia, A. 2003. Biopanning of endotoxin-specific phage displayed peptides. Biochem. Biophys. Res. Commun. 307:133-138. 94. Noda, K., Yamasaki, R., Hironaka, Y., and Kitagawa, A. 2001. Selection of peptides that bind to the core oligosaccharide of R-fo rm LPS from a phage-displayed heptapeptide library. FEMS Microbiol. Lett. 205:349-354. 95. Harlow,E. and Lane,D. 1988. Antibodies a Laboratory Manual. Cold Spring Harbor Laboratory, Cold Spri ng Harbor, NY. 23-34 pp. 96. Paoli, G.C., Chen, C.Y., and Brewst er, J.D. 2004. Single-chain Fv antibody with specificity for Listeria monocytogenes J. Immunol. Methods 289:147-155.

PAGE 162

162 97. Leonard, P., Safsten, P., Hearty, S., McDonnell, B., Finlay, W., and O'Kennedy, R. 2007. High throughput ranking of recombinant av ian scFv antibody fragments from crude lysates using the Biacore A100. J. Immunol. Methods 323:172-179. 98. Su, B., Hrin, R., Harvey, B.R., Wang, Y. J., Ernst, R.E., Hampton, R.A., Miller, M.D., Strohl, W.R., An, Z., and Montgomery, D.L. 2007. Automated high-throughput purification of antibody fragment s to facilitate evaluation in functional and kinetic based assays. J. Immunol. Methods 322:94-103. 99. Bannister, D., Wilson, A., Prowse, L., Walsh, M., Holgate, R., Jermutus, L., and Wilkinson, T. 2006. Parallel, high-throughput pu rification of recombinant antibodies for in vivo cell assays. Biotechnol. Bioeng. 94:931-937. 100. Lobel, L.I., Morseman, J.P., Zeng, X., Lust bader, J.W., Chen, H., and Allnutt, F.C. 2001. Development of a fluorescence based high thr oughput assay for antagon ists of the human chorionic gonadotropin receptor extracellular domain: analysis of peptide inhibitors. J. Biomol. Screen. 6:151-158. 101. Vanhercke, T., Ampe, C., Tirry, L., and Denol f, P. 2005. Rescue and in situ selection and evaluation (RISE): a method for high-thr oughput panning of phage display libraries. J. Biomol. Screen. 10:108-117.

PAGE 163

163 BIOGRAPHICAL SKETCH Crystal Harpley (Mazur) was born in Colu m bus, OH; and shortly after moved to Wellington, FL, where she completed grade scho ol. In 2006 she graduated with high honors from the University of Florida, with a major in microbiology a nd cell sciences and a minor in chemistry. During the process of finishing her masters thesis, Crystal Harpley got married and is now Crystal Mazur.