D-731b
Species-Specific Bacteriophage Amplification for Enhanced MALDI-TOF MS Bacterial Diagnostics
1310  Maple St., Golden, CO 80401
303-384-2493
[email protected]
C.R. Cox1, C.R. McAlpin1, J.C. Rees2, K.J. Voorhees1
1Colorado
School of Mines, Department of Chemistry and Geochemistry, Golden, CO; 2Centers for Disease Control and Prevention, Atlanta, GA
ABSTRACT
Background. Matrix assisted laser desorption time-of-flight mass spectrometry (MALDI-TOF MS) is emerging as a
sensitive approach for rapid bacterial diagnostics. Most bacterial identification techniques (conventional MALDI-TOF MS
included) rely on tedious culturing practices and/or costly molecular methods, which render them impractical for rapid
bacterial identification. With a view towards addressing some of these shortcomings, we employed the evolutionarily
conserved interactions between a bacteriophage (phage) and its bacterial host to develop new, rapid, and species-specific
diagnostic spectroscopy methods.
Methods. Enterobacteria phage MS2 and plague diagnostic phage φA1122 were used as well-characterized biomarkers
for identification of E. coli and Y. pestis, respectively. Infections were monitored by MALDI-TOF MS over a three-hour
period to assay for an increase in phage major capsid protein concentrations in comparison to bacterial concentrations as
a means of signal amplification. This allowed for production of phage-specific protein mass profiles as a diagnostic
output. Further, phage infections were modeled in silico using a modified Payne and Jansen algorithm to predict the
approximate time during infections that phage concentrations would expand beyond the MALDI-TOF-MS limit of detection
(LOD).
Results. MS2-derived signal amplification resulting from infection of 5x106 cfu/ml E. coli, a bacterial concentration that
was initially below the limit of detection, was detectable within one hour. Similarly, a φA1122 amplification signal resulting
from infection of 3x106 cfu/ml Y. pestis was detectable within two hours. A modified computational model allowed for
prediction of progeny phage concentration as a function of time and significant agreement between mathematically
calculated phage growth curves and those experimentally obtained by MALDI-TOF-MS was observed.
Conclusions. Novel and reproducibly predictable phage-based signal amplification readily allowed for MALDI-TOF MS
bacterial detection and diagnostic identification in as little as one hour for E. coli, and within two hours for Y. pestis,
effectively negating the need for overnight bacterial enrichment required by most conventional diagnostic methods.
RESULTS
RESULTS
dx/dt = ax – bvx
dy/dt = bvx – ky
dv/dt = kLy - bvx
Modified Payne-Jansen phage therapy model
Table 1: Modeling parameters
INTRODUCTION
•  MALDI-TOF MS is a useful as a tool for rapid bacterial detection and identification.
•  Several shortcomings have been identified with regards to its relative sensitivity and need for bacterial samples grown
under exact conditions. Indeed, while a majority of bacterial phylotypes are often detectable by MALDI-TOF MS on their
own, this requires time-intensive overnight culture under controlled growth conditions.
METHODS
Figure 2. A) Phage amplification-based MALDI-TOF MS detection of E. coli. Mass spectra are shown for A) E. coli, B) MS2 phage, C) E. coli-MS2 phage
mixture before phage amplification (bacterial and phage concentrations below the MALDI-TOF MS LOD), and D) the same mixture after 3 hours of phage
amplification.
Phage amplification
•  φA1122: Amplification was performed with initial Y. pestis concentrations ranging from 1x105 to 1x108 cfu/ml and were infected using an initial phage concentration of 3x106
pfu/ml.
•  MS2: Amplification was performed with an initial E. coli concentration of 3x106 cfu/ml and an equal quantity of MS2.
•  Amplification was simultaneously tracked using MALDI-TOF MS and conventional agar plate plaque assays(1) over 5, 10, 15, or 30-minute intervals for a period of 1-4 hours.
•  During the course of amplification experiments, aliquots were prepared for MALDI-TOF-MS analysis by pelleting bacterial debris at 10,000 x g for 5 minutes, followed by
filtration with 0.22 µm PES filters prior to application to MALDI-TOF grids.
•  To address these issues, highly species-specific phage amplification was applied and determined to significantly improve
the overall utility of MALDI-TOF MS for bacterial analysis, effectively shortening detection times by removing the need
for overnight bacterial enrichment, while at the same time lowering the amount of bacteria required for detection.
•  Despite such advances, laborious and time-intensive
methods associated with use of phage-based MALDITOF MS in bacterial diagnostics can be problematic,
particularly when involving extensive time-course
studies requiring multiple simultaneous analyses.
MALDI-TOF-MS analysis
•  Mass spectral analysis was conducted utilizing a ferulic acid matrix (15 mg/ml) in a 17:33:50 mixture of 88% formic acid, acetonitrile and de-ionized water (2). Mass spectra
were obtained with a 337 nm N2 laser in linear mode using a PerSeptive Biosystems Voyager-DE STR+ MALDI-TOF-MS (Applied Biosystems, Inc.).
•  Samples were applied to grids using the dried droplet method in a sandwich fashion (2) as follows: 0.5 µl matrix: 0.5 µl sample: 0.5 µl matrix.
•  Spectra were collected using an accelerating voltage of 25 kV, a grid voltage of 75%, a delayed extraction time of 100 ns, and a 2 kDa low mass ion gate and an average of
150 laser shots taken from 3 individually prepared sample spots. Raw data was exported into Sigma Plot (Systat Software) for spectral analysis.
•  Instrument limit of detection (LOD) was determined for by analyzing a series of 10-fold serial phage dilutions assayed using the 15.8 kDa φA1122, and 13.8 kDa MS2 capsid
monomers as biomarkers..
•  Once the maximum growth curve had been determined using the model, a MALDI-TOF mass spectrum was obtained corresponding to the calculated maximum time from a
given infection reaction.
•  In order to reduce the requirement for repeated
preparation and analysis, a modified phage therapy
model was investigated as a means for predicting the
time during a given phage infection when a detectable
signal would occur.
Modified computational model
•  The Payne-Jansen computational model(3,4) for phage therapy was modified to approximate in vitro infection by removing the variables defining human host effects on the
bacteria [H(t)x] and phage [h(t)y], as well as the replication rate for infected bacteria (ay), and the phage decay rate (m).
•  All computations were performed using Mathematica 8 (Wolfram Software, Champaign, IL).
Literature cited
•  Phage amplification is the process whereby a burst of
progeny phage are produced as a result of speciesspecific infection and replication in a bacterial host.
1.  K. Carlson. Working with Bacteriophages: Common Techniques and Methological Approaches. In: E. Kutter, A. Sulakvelidze, eds. Bacteriophages Biology and Applications. New York CRC Press, 2005, 437."
2. A. J. Madonna, F. Basile, I. Ferrer, M.A. Meetani, J. C. Rees, K. J. Voorhees. On probe sample pretreatment for the detection of proteins above 15 kDa from whole spectrometry. Rapid. Commun. Mass. Spectrom. 2000, "
14, 2220. "
3. R. J. Payne, V. A. Jansen. Understanding bacteriophage therapy as a density-dependent kinetic process. J. Theor. Biol. 2001, 208, 37."
4. B. J. Cairns, A.R. Timms, V. A. Jansen, I.F. Connerton, R. J. Payne. Quantitative models of in vitro bacteriophage-host dynamics and their application to phage therapy. PLoS. Pathog. 2009, 5, e1000253."
•  When applied in an in vitro fashion, this event leads to a
substantial increase in the number of detectable phage
proteins in an infection reaction.
•  This increase in phage proteins in comparison to
bacterial host concentrations can be exploited using
MALDI-TOF MS as a method for signal amplification
during bacterial detection and identification.
Bacterial strains, phages and culture conditions
•  Yersinia pestis A1122 (Centers for Disease Control, Division of Vector-Borne Infectious Diseases [CDC/DVBID], Ft. Collins, CO) was streaked onto brain heart infusion (BHI)
(Difco-BD) agar and incubated for 48 hours at either 28°C or 37°C. E. coli ATCC strain 15597 was streaked onto Luria Bertani (LB) agar and incubated overnight at 37°C.
Single isolated colonies were transferred to 5 ml of BHI or LB broth, and incubated with aeration at 28°C or 37°C to an optical density (OD620) of 0.2.
•  Phage φA1122 was propagated with Y. pestis A1122 at 28ºC in BHI utilizing an MOI of 0.1. MS2 was propagated with E. coli ATCC 15597 at 37°C in LB at an MOI of 1.0.
CONCLUSIONS
Figure 1. Phage amplification-coupled to MALDI-TOF MS for bacterial detection and identification.
Application of species-specific phage (starting concentration below MALDI-TOF MS detection threshold)
to a sample suspected to contain target bacteria leads to phage amplification. The resulting burst
liberates phages for MALDI-TOF MS analysis yielding a phage-specific protein profile (seen here as a
37.8 kDa major capsid protein peak and its 18.9 kDa doubly charged ion), thus indicating the presence
of target bacterial species.
Figure 3. Prediction of MALDI-TOF MS detection threshold for E. coli
phage MS2 phage amplification. (A) A three hour in vitro phage
amplification was performed at 37ºC and monitored by MALDI-TOF-MS.
Phage amplification is indicated by a single 13.8 kDa capsid peak. (B)
Model prediction of phage amplification. MALDI-TOF-MS and modeling
were conducted with an initial bacterial concentration of xi = 5x106 cfu/ml
and an initial phage concentration of vi = 5x106 pfu/ml (MOI=1.0) (solid
line). MALDI-TOF-MS LOD is shown as a dotted line. All other
parameters were held constant with a = 1, b = 10-7, k = 1.5 and L = 1100.
Figure 4. Prediction of MALDI-TOF MS detection threshold for Yersinia
pestis phage φA1122 amplification. (A) A three hour in vitro phage
amplification was performed at 28ºC and monitored by MALDI-TOF-MS.
Phage amplification is indicated by a single 15.8 kDa capsid peak. (B)
Model prediction of phage amplification. MALDI-TOF-MS and modeling
were conducted with an initial bacterial concentration of xi = 3x106 cfu/ml
and an initial phage concentration of vi = 3x106 pfu/ml (MOI=1.0) (solid
line). MALDI-TOF-MS LOD is shown as a dotted line. All other
parameters were held constant with a = 1, b = 10-7, k = 1.5 and L = 180.
•  Phage amplification was observed to increase the utility of MALDI-TOF MS-based bacterial detection by eliminating the
need for extensive bacterial culture prior to analysis.
•  MS2 amplification-mediated detection of E. coli was reproducibly and predictably observed within 1 hour of infection.
• ϕA1122-amplification-mediated detection of Y. pestis was reproducibly and predictably observed within 2 hours.
•  Using a modified phage therapy modeling algorithm, the point in time which phage concentrations exceeded MALDITOF MS LOD could be accurately, and reproducibly predicted prior to experimentation.