Download A BIOSENSOR APPROACH FOR THE DETECTION OF ACTIVE

A BIOSENSOR APPROACH FOR THE DETECTION OF ACTIVE VIRUS USING
FTIR SPECTROSCOPY AND CELL CULTURE
by
Felipe T. Lee Montiel
______________________
A Dissertation Submitted to the Faculty of the
DEPARTMENT OF AGRICULTURAL AND BIOSYSTEMS ENGINEERING
In Partial Fulfillment of the Requirements
for the Degree of
DOCTOR OF PHILOSOPHY
In the Graduate College
THE UNIVERSITY OF ARIZONA
2011
2
THE UNIVERSITY OF ARIZONA
GRADUATE COLLEGE
As members of the Dissertation Committee, we certify that we have read the dissertation
prepared by Felipe T. Lee Montiel
entitled A BIOSENSOR APPROACH FOR THE DETECTION OF ACTIVE VIRUS
USING FTIR SPECTROSCOPY AND CELL CULTURE
and recommend that it be accepted as fulfilling the dissertation requirement for the
Degree of Doctor of Philosophy.
_______________________________________________________________________ Date:
07/29/2011
Dr. Mark Riley
_______________________________________________________________________
Date: 07/29/2011
Dr. Jeong-Yeol Yoon
_______________________________________________________________________
Date: 07/29/2011
Dr. Lingling An
_______________________________________________________________________
Date: 07/29/2011
Dr. Kelly Reynolds
_______________________________________________________________________
Date: 07/29/2011
Dr. Joel Cuello
Final approval and acceptance of this dissertation is contingent upon the candidate’s
submission of the final copies of the dissertation to the Graduate College.
I hereby certify that I have read this dissertation prepared under my direction and
recommend that it be accepted as fulfilling the dissertation requirement.
________________________________________________ Date: 07/29/2011
Dissertation Director: Dr. Mark Riley
3
STATEMENT BY AUTHOR
This dissertation has been submitted in partial fulfillment of requirements for an
advanced degree at the University of Arizona and is deposited in the University Library
to be made available to borrowers under rules of the Library.
Brief quotations from this dissertation are allowable without special permission, provided
that accurate acknowledgment of source is made. Requests for permission for extended
quotation from or reproduction of this manuscript in whole or in part may be granted by
the head of the major department or the Dean of the Graduate College when in his or her
judgment the proposed use of the material is in the interests of scholarship. In all other
instances, however, permission must be obtained from the author.
Signed: Felipe T. Lee Montiel
4
ACKNOWLEDGEMENTS
I would like to express my sincere gratitude to my supervisor Dr. Mark Riley for his
mentoring and guidance. He has encouraged me to reach my full potential as a research
scientist and to be an independent and critical thinker. He provided me with many
academic and vocational opportunities on campus to widen my horizons.
I also acknowledge my committee members, Dr. Jeong-Yeol Yoon, Dr. Lingling An, Dr.
Kelly Reynolds and Dr. Joel Cuello, for their advice and time. These people have
supported me through the transition from a new student with a new project, to a confident
scientist.
I would also like to thank everyone who helped with experiments, talking through ideas,
my classmates and colleagues for making the campus an enjoyable and intellectually
stimulating place to be. In particular the other members of the Riley lab, Dr. Werner
Zimmt, Phat LeTran, Dominic, Brian, Jessi and Federico.
Thanks to Bodil Cass for her support and encouragement. I love you Bo.
Muchas gracias a toda mi familia en especial a mi madre, mi padre y mi hermano, por su
amor y apoyo. Espero poder regresarles pronto un poco de lo mucho que me han dado, y
también quiero decir que siento no haber estado cerca por el tiempo que he estado lejos
de ellos luchando por mis sueños pero espero que les pueda agradecer mucho todo lo que
he hecho por mi porque sin ellos nunca hubiera logrado esta meta.
5
DEDICATION
I would like to dedicate this work to my father for his example of hard work everyday, to
my mother for her love and advice.
6
TABLE OF CONTENTS
LIST OF TABLES ................................................................................................... 9 LIST OF FIGURES ................................................................................................ 10 ABSTRACT ........................................................................................................... 11 INTRODUCTION .................................................................................................. 13 1. Background of Biosensors............................................................................... 13 1.1 Biosensors ................................................................................................. 13 1.2 Cell-Based Biosensors .............................................................................. 14 1.3 FTIR Spectroscopy in CBB ...................................................................... 16 1.4 Virus Detection with FTIR-CBB .............................................................. 17 2. Waterborne Viruses ......................................................................................... 19 2.1 Biology of Enterovirus Infection .............................................................. 20 2.2 Morphological and Biochemical Alterations in Enterovirus Infected Cells
......................................................................................................................... 21 2.3 Globalization, Outbreaks and Drinking Water Problems ......................... 24 2.4 Water monitoring ...................................................................................... 26 3. Current Enterovirus Detection Methods.......................................................... 28 3.1 Problems with Current Detection Methods............................................... 29 4. Research Objectives ........................................................................................ 30 4.1 Specific Aims ............................................................................................ 30 7
PRESENT STUDY ................................................................................................ 31 Overall Conclusions and Recommendations .................................................. 32 REFERENCES ....................................................................................................... 33 APPENDIX A: A NOVEL DETECTION SCHEME FOR POLIOVIRUS
INFECTION USING FTIR SPECTROSCOPY AND CELL CULTURE ............ 37 Abstract ............................................................................................................... 38 Background ......................................................................................................... 40 Results ................................................................................................................. 44 Discussion ........................................................................................................... 49 Conclusions ......................................................................................................... 53 Methods ............................................................................................................... 54 References ........................................................................................................... 61 Figures ................................................................................................................. 65 Tables .................................................................................................................. 71 APPENDIX B: FTIR SPECTROSCOPY AND PCA TO DISCRIMINATE CELLS
WITH DIFFERENT ENTEROVIRUS INFECTIONS .......................................... 75 Abstract ............................................................................................................... 76 Background ......................................................................................................... 78 Results ................................................................................................................. 82 Discussion ........................................................................................................... 85 Conclusions ......................................................................................................... 90 Acknowledgments ............................................................................................... 90 8
Materials and methods......................................................................................... 91 References ......................................................................................................... 109 APPENDIX C: REAL-TIME, CONTINUOUS MEASUREMENT OF CHANGES
IN CELLULAR COMPONENTS FOLLOWING VIRAL INFECTION ............ 112 Abstract ............................................................................................................. 113 Introduction ....................................................................................................... 115 Results ............................................................................................................... 119 Conclusions ....................................................................................................... 127 Figures ............................................................................................................... 128 Materials and methods....................................................................................... 135 Acknowledgments ............................................................................................. 138 References ......................................................................................................... 139
9
LIST OF TABLES
Table 1. Molecular and physiological changes of mammalian cells caused by poliovirus
infection……………………………………………….…………………………… 23
10
LIST OF FIGURES
Figure 1. Diseases contributing to the water, sanitation and hygiene related disease
burden. Adapted from Safer water, better health : costs, benefits and sustainability of
interventions to protect and promote health. WHO 2008 ......................................... 20
Figure 2. Attribution of disease burden from water, sanitation and hygiene to
areas/sectors. Adapted from Safer water, better health : costs, benefits and
sustainability of interventions to protect and promote health. WHO 2008 .............. 25
11
ABSTRACT
Worldwide, 3.575 million people die each year from water-related diseases. The water
and sanitation crisis claims more lives than any warfare and is predicted to be one of the
biggest global challenges of this century. The rapid, accurate detection of viral pathogens
from environmental samples is an ongoing and pertinent challenge in biological
engineering. Currently employed methods are lacking in either efficiency or specificity.
Here we explore a novel method for virus detection and concurrently use this method to
learn more about the very early stages of the virus infection process. The method
combines Fourier transform infrared (FTIR) spectroscopy, a method of visualizing
molecules based on changes in vibration of particles, and mammalian cells as the
biosensor. This method is used to detect and investigate viruses from the family
picornaviridae, chosen due to their public health burden and their widespread presence in
environmental samples, especially water sources. This family includes the Polioviruses,
echoviruses and Coxsackieviruses, among others, many of which are human pathogens.
The research outlined in this dissertation is aimed at developing and implementing a new
cell-based biosensor that combines the advantages of FTIR spectroscopy with the ability
of buffalo green monkey kidney (BGMK) cells to sense diverse stimuli, including
infective enteroviruses. The goal of developing this biosensor is outlined in the first
paper. The second paper focuses on the application of advanced statistical methods to
analyze the spectra to discriminate different viral infections in BGMK cells. Finally, we
designed a non-reactive metal biochamber to use with attenuated total reflectance-FTIR.
12
This allowed near-continuous acquisition of real-time spectral data for the study of
biochemical changes in mammalian cells caused by poliovirus (PV1) infection. This
system is capable of tracking changes in cell biochemistry in minute intervals for many
hours at a time.
This work demonstrates the feasibility of FTIR spectroscopy in combination with the
broad sensitivity of mammalian cells for potential use in the detection of infective viruses
from environmental samples. We envision this method being extended to high
throughput, automated systems to screen for viruses or other toxins in drinking water
systems and medical applications.
Key words: Fourier transform infrared (FTIR) spectroscopy, cell-based biosensor,
attenuated total reflectance, zinc selenide, buffalo green monkey kidney (BGMK) cells.
13
INTRODUCTION
1. Background of Biosensors
1.1 Biosensors
Over the last two decades, biosensor research has been growing tremendously, with a
special interest in environmental, medical, toxicological, drug discovery, single-cell
analysis and defense applications. Biosensors are analytical devices that integrate a
biological sensing element (for example, DNA, enzymes, antibodies, proteins or whole
cells) that converts external stimuli into signals, combined with a transducer element
(electrochemical or optical) for reading the signal and processing [1-3].
Some desirable feature of any new biosensor include the capacity for continuous
monitoring, rapid response times, automation, reusability and portability (point-of-care)
[1]. The concept of point-of-care (POC) testing has been well established especially for
implementation in challenging environments around the world. Results can be obtained in
a much shorter time when the testing facility is conveniently brought to the vicinity of the
sample. There has been a major effort in recent years to develop POC molecular
diagnostic devices for resource-limited regions where well-equipped centralized
laboratories are not readily accessible [10]. POC testing is most useful for applications
that have time-sensitive samples including medical/veterinary diagnostics, environmental
monitoring, and defense services.
14
The following minimum requirements have been proposed in order for a device to be
considered a real POC system: (A) minimal pretreatment to allow the direct introduction
of the sample; (B) portability, that is, small in size, lightweight and autonomous power
supply; (C) user-friendly such that it can be operated by people with minimal training;
and (D) offer qualitative and quantitative results [10].
1.2 Cell-Based Biosensors
Cell-based biosensors (CBB) have two components: a sensitive element (cellular
responses) and a transducer (in the case presented here, an FTIR spectroscopy device)
that uses the cells’ remarkable ability to detect, transduce, and amplify very small
changes in external stimuli. The combination of biologically relevant cellular detection
interfaced with a powerful optical tool provides many advantages over previous
biosensors. CBB can have broad sensitivity to pathogens and have the potential to
provide rapid, sensitive, low-cost options for applications requiring the detection and
monitoring of unknown agents. These characteristics have the potential to revolutionize
toxicology, environmental contamination management and other medical areas [4].
The use of CBB outside of the laboratory, however, has been challenging due to many
requirements including preparation of the sample, maintenance of the biological
environment for the sensitive element, and integration of the electronics for data
collection and analysis. Some aspects of biosensors limit their use outside of controlled
laboratory conditions. For example, it is difficult to have a portable cell culture system
15
that keeps the cells alive and uncontaminated in the field. Nonetheless, long-term
maintenance of cells on biochips has been achieved using enclosed fluid chips and
endothelial cells in a portable bench top toxin sensor. This device monitored changes in
the cells with Electric Cell-substrate Impedance Sensing following exposure to water
samples [5].
An additional issue with CBB is that ‘noise’ can occur when other environmental stimuli
present in the experimental system interfere with the cells in culture. The cellular
responses to changes in temperature, pH or osmolarity can be difficult to distinguish from
responses to the stimulus of interest. These confounding stimuli could create a false
positive result if the unwanted response is similar to the response of interest, or a false
negative if the response of interest is masked by the noise. The most effective CBB
systems will aim to minimize this interference by providing the cells with a tightly
controlled environment in the bioreactor growth chamber [6]. For example, DeBusschere
et al. [7] designed a system that addresses several of these issues with an integrated
silicon–polydimethylsiloxane cell-cartridge.
Cell based biosensors can be of many types, such as optical, thermal, electrical or
electrochemical, depending on the transducing mechanism used. In recent years,
electrical biosensors are becoming a popular approach to be used in CBB. These
biosensors measure current and/or voltage to sense changes in the cell condition, usually
to detect cell binding. They are simple, inexpensive and generally energy efficient, giving
16
them perhaps the best potential for lab-on-a-chip adaptations. Electrical sensors can be
further divided into amperometric (measuring changes in voltage or current with
electrodes), or impedance (measuring electrical impedance) [8]. Impedance biosensors
have been employed to measure changes in the cell cytoskeleton or cell morphology to
indirectly track intracellular signaling and receptor-specific response profiles [9].
Spectroscopy using infrared light may be a good technique to use in biosensors. Infrared
light is well resolved and has a long wavelength that penetrates into the cells further and
with less unwanted scattering than other wavelengths. It can also be targeted towards
specific regions of a sample using a microscope [12]. It has the potential to detect small
changes from a small sample of cells without the need for expensive reagents. Infrared
spectroscopy also has the potential to infer structural changes in particular biomolecules,
even in living samples. These are detected by linking the frequency of absorbance regions
to the different structural features using coordinates on the spectral region [13].
1.3 FTIR Spectroscopy in CBB
Fourier transform infrared (FTIR) spectroscopy is a nondestructive technique based on
vibration spectroscopy that has shown much potential as a tool for use in microorganism
identification [11]. It has the potential to inform about cellular biochemistry without
destroying the sample. The spectroscopic study of biological cells and tissues is a
dynamic area of research with the primary goal elucidating how viruses, bacteria and
different types of cancer function and can be accurately detected. Consequently, FTIR
17
has the potential of providing qualitative understanding of the cellular biochemistry
behind these illnesses. FTIR spectroscopy can be applied for pathogen detection and
simultaneously to provide a spectrum of basic details about cell behavior in response to
infection [4].
1.4 Virus Detection with FTIR-CBB
FTIR spectroscopy can be combined with cell culture for the development of a novel
strategy for virus detection using live cells as biosensors. This approach utilizes
mammalian cell culture and infrared light absorbance to monitor specific absorbance
patterns produced following changes in cell components (such as lipids, proteins, nucleic
acids and sugars) subsequent to the virus infection, effectively using the cells as
biosensors [1, 14].
Mid-infrared spectroscopy can be used to observe changes in cell molecular machineries.
It can track many aspects of the cell, such as nucleic acids, proteins, and lipids that have
characteristic and well-defined IR-active vibrational modes. This type of data has broad
applications in biomedicine, cellular biochemistry, disease monitoring, and observation
of single-celled organisms [15-17]. It has previously been used to detect changes in
cellular components following exposure to toxins [18] and to detect cell apoptosis [19].
Riley et al. [20] developed an infrared spectroscopic monitoring scheme using infrared
transmissive optical fibers composed of TAS (Te, As, and Se)to detect an inhalation
health hazard, Streptolysin O. Another application of FTIR has been to study herpes
18
infection kinetics, for example, Erukhimovitch et. al. [21] found that an IR peak
attributed to carbohydrates disappeared completely in an exponential kinetics as a result
of infection with herpes viruses HSV-I or VZV. This peak might be considered an
important parameter in the study of the kinetics of the herpes virus infection and in the
clinical detection of herpes virus infections [21, 22]. In addition, Vargas et. al. [23] used
electrophoretic deposition on a Germanium ATR crystal to capture and detect poliovirus
and a bacteriophage.
The rapid and non-invasive manner of the cellular spectroscopy approach opens
numerous possibilities for translation of this research to biomedical applications and
biosensors. FTIR virus detection is faster than the currently standard methods required to
detect viable viruses and improves on the shortcomings of the best available methods for
human virus detection, which include cell culture and ICC/PCR (integrated cell culture /
PCR).
19
2. Waterborne Viruses
Viruses have the highest potential infectivity of all waterborne microorganisms, requiring
exposure to only a small number of virions to cause infection. They often replicate in the
gastrointestinal tract and spread through the feces of infected individuals in large
numbers, and generally have the longest survival in the environment. They are not
efficiently removed by conventional filtration and are more resistant to disinfectants than
bacteria. Viruses are known to be the causative agent in 8% of drinking water outbreaks
reported in recent years [24].
Waterborne viruses are particularly important due to their associative illnesses and
effective transmission at low titers. Waterborne viruses include enteroviruses, hepatitis A
virus (hepatitis, liver damage), noroviruses (diarrhea), astrovirus (diarrhea), adenovirus
(diarrhea, respiratory disease, eye infections, heart disease), and rotavirus (diarrhea).
Enteroviruses are small (approximately 30 nm), nonenveloped, single-stranded RNA
viruses with an icosahedral capsid. They are among the most common viral infectious
agents in humans worldwide [25, 26]. Figure 1 shows an estimated percentage of
diarrheal diseases caused by enteroviruses worldwide. Serotypes of human enteroviruses
have traditionally been classified into echoviruses, Coxsackieviruses group A and B, and
polioviruses [25]. Clinical symptoms vary from mild febrile illness (diarrhea, fever) to
more severe diseases, including meningitis, myocarditis, paralytic illness, and
encephalitis. Each year, an estimated 10-15 million symptomatic enterovirus infections
occur in the United States [27]. In addition, four of the human enteric viruses,
20
Coxsackievirus, echovirus, calicivirus, and adenovirus, have been included among the
microorganisms of concern on the Environmental Protection Agency’s (EPA) Drinking
Water Contaminant Candidate List (CCL) [28].
Diseases contributing to the water, sanitation and hygiene related disease burden
)*+,,-./+0$1*2/+2/2$$
(mostly from enteroviruses)
!"#$
&(#$
3+0+,*+$
456/275+0$5/8+6.1/$
*59/:7.52$
;6-/,2$
'#$
%&#$
Figure 1. Diseases contributing to the water, sanitation and hygiene related disease
burden. Modified from Safer water, better health: costs, benefits and sustainability of
interventions to protect and promote health, adapted from W.H.O. 2008 [29].
2.1 Biology of Enterovirus Infection
Enteroviruses replicate primarily in the respiratory or gastrointestinal tracts but can
advance to the circulatory system and thereby to other organs. The virus articles first bind
to receptors on the cell surface. These specific receptors have been characterized
previously; for example, poliovirus binds to CD155 in the immunoglobulin superfamily.
21
Echovirus 1 bind to Alpha-2-beta-1-integrin (VLA-2) and other echoviruses bind to
Decay accelerating factor (DAF; CD55). Some Coxsackieviruses also use similar
receptors to echoviruses, for example, Coxsackie A9 binds to Alpha-v-beta-a-integrin,
some bind to DAF, whereas other Coxsackie A viruses use a rhino-virus receptor
[30](and references therein).
Poliovirus, Coxsackieviruses and echoviruses are cytophatic for most cell types and can
cause changes in cell metabolism and appearance eventually leading to cell lysis.
Changes can include detachment from the cell substrate and shrinking to a round cell
morphology. Changes in cell internal structures can also be seen, including
rearrangement of the cytoskeleton, clustering of the ribosomes, formation of membranous
vesicles and of the nucleus into a crescent shape. Chromatin is also rearranged to localize
at the edges of the nucleus. Host translation and transcription are usually inhibited during
the infection process and the plasma membrane is made more permeable.
2.2 Morphological and Biochemical Alterations in Enterovirus Infected Cells
The major morphological and biochemical steps in the poliovirus infection process are
detailed in Table 1.
Enteroviruses cause many changes to the cell nucleus. The cell’s translation is inhibited
at 2 hours after infection. Subsequently, Polymerase I is inhibited within 2-3h, Pol II at
4h and pol III by 5h, greatly reducing the transcriptional ability of the cell. At
22
approximately 4 hours after infection the cell no longer synthesizes its own DNA and the
chromatin is condensed into the nucleolus.
Enteroviruses also affect the cell cytoskeleton components. Actin microfilaments,
microtubules and intermediate filaments are considerably rearranged during infection and
this affects cell shape, mobility and anchoring. Intermediate filaments accumulate at the
nucleous and possibly serve as sites for viral RNA replication. The cell’s cytoplasmic
membranes are also altered following infection to make many membranous vesicles.
Most enterovirus infections end in cell lysis [31].
23
Table 1. Molecular and physiological changes of mammalian cells caused by poliovirus
infection.
Hours post-infection
0-1
Microscopic observations
Intact virus particles in the
cytoplasm directly below the
plasma membrane and in
micropinocytotic vesicles
Biochemical events
Transcient increase in amino acid
uptake, elevated Na+/K+ pump
activity, decrease in membrane
fluidity, parental capsid proteins
and RNA in polysomes and
lysosomes
1-2
Decrease in cell size, distortion
of nucleus, wrinkling of
nuclear membrane,
chromosome condensation.
Inhibition of host protein and
RNA synthesis, release of host
mRNA from cytoskeleton,
beginning of viral protein
synthesis, peak activity of Na+/K+
pump, exponential phase of RNA
synthesis.
2-2.5
Membrane proliferation,
appearance of large clusters of
membrane bound polysomes in
cytoplasmic periphery,
structural arrangement of the
cell cytoskeleton.
Increase in chlorine
incorporation, host cell mRNA
inactive, viral protein synthesis
and all newly formed viral
proteins membrane associated,
inactivation of Na+/K+ pump,
increase in intracellular Na+,
decrease in intracellular K+.
2.5-3
Several foci of vesicle
formation
Switch to linear phase of viral
RNA synthesis: predominantly
membrane associated vRNA
synthesis, peak of RNA synthesis
2.5-3
Several foci of vesicle
formation
Switch to linear phase of viral
RNA synthesis: predominantly
membrane associated vRNA
synthesis, peak of RNA synthesis
3-4
Perinuclear conglomeration of
membrane enclosed vesicles
into one large mass, nuclear
extrusions, appearance of
progeny virions
Peak of virion assembly, increase
in membrane permeability for
monovalent cations, declining
rates of protein and RNA
synthesis.
4-6
Autophagic vacuoles,
redistribution of lysosomal
enzymes over mass of
membrane bound vesicles
Accumulation of ds RNA,
depletion of metabolic precursors,
release of lysosomal enzymes
6-8
Viral crystals in cell periphery
lysis of host cell
Modified from The Molecular Biology of Poliovirus [32].
24
2.3 Globalization, Outbreaks and Drinking Water Problems
Waterborne infectious diseases have continued to emerge due in part to (1) an increase in
sensitive populations; (2) recognition of the importance of additional health effects; (3)
globalization of commerce and travel including importation of food from countries with
poor water quality; (4) natural evolution of microbes with increased virulence; (5)
changes in drinking water treatment technologies; and (6) development of molecular
methods for detection to improve identification of outbreaks and their sources. Changes
in drinking water treatment technology or food supply production may also contribute,
for example, an increase in the use of ultraviolet light for water treatment followed efforts
to address concerns over protozoan contamination of water and chlorine resistance.
However, increased reliance on UV light treatment has subsequently raised concerns over
virus resistance [33, 34] .
Most viruses are more environmentally stable than enteric bacteria and not cultivable. For
this reason, it is completely unsafe to rely on bacteriological standards to assess the
virological content of water. Nearly half of all documented waterborne outbreaks since
1971 remain uncategorized but many have characteristics consisten with virus infection
patterns [35].
Waterborne disease outbreaks are considered to be primarily the result of technological
failures or failure to treat the water [36]. Figure 2 shows the percentage of illness related
to drinking water and sanitation in the world. Water treatment for human consumption
25
requires, at a minimum, reducing infectious viruses by 99.99% and protozoan parasites
by 99.9% [24]. The US-EPA states an acceptable risk of one infection per 10,000
consumers annually.
Attribution of disease burden from water, sanitation and
hygiene to areas
22%
62%
16%
Related behavior and
other
Ecosystem
management
Drinking-water and
sanitation
Figure 2. Attribution of disease burden from water, sanitation and hygiene to
areas/sectors. Modified from Safer water, better health: costs, benefits and sustainability
of interventions to protect and promote health, adapted from W.H.O. 2008 [29].
These studies emphasize the need to improve drinking water monitoring systems to better
detect viral contamination, and the setting up of more efficient water treatments for virus
inactivation once viruses are detected in the water supply. Future advances in biosensor
technologies to detect the presence of human enteric viruses in a water source will be a
valuable instrument in the prevention of waterborne diseases [37].
26
2.4 Water monitoring
Water monitoring is performed to protect human health against contamination of the
water supply with pathogens. In the United States, the Environmental Protection Agency
(EPA) and the Center for Disease Control and Prevention (CDC) regulate water quality.
The EPA establishes drinking water standards including maximum contaminant levels
and treatment techniques that states must meet through legislation. The EPA also issues
advisories about drinking water contaminants. From 1971 to 2002 there have been 764
documented waterborne outbreaks associated with drinking water, with 8% caused by
viruses. Reynolds et al. [24] estimated that 19.5 million people are estimated to be at risk
from drinking water in the U.S. annually, indicating the possible health care impacts of a
deficient drinking water monitoring.
A recent study made by Craun et al. [38] showed that the annual number of reported
drinking water outbreaks analyzed by water system type, in public water systems
(community and noncommunity) was found to decrease during 1971 to 2006 . In contrast,
the annual proportion of drinking water outbreaks associated with individual water
distribution systems increased [38]. The reduction in total waterborne outbreaks is largely
attributed to new regulations by the US-EPA, including the surface water treatment rule,
and improving water treatment. In 2006, EPA signed the Ground Water Rule (GWR)
with the purpose to reduce disease incidence associated with pathogens in drinking water.
27
One of the GWR requirements is compliance monitoring to ensure that treatment
technology installed to treat drinking water reliably achieves at least 99.99% inactivation
or removal of viruses. Although the current regulatory requirements do not appear to
reduce the proportion of outbreaks associated with distribution systems [39]. Water
treatment technologies need to increasingly address changes in water quality and quantity
requirements along with changes in microbial populations caused by these used treatment
technology, which could potentially produce a problem of detection [40].
28
3. Current Enterovirus Detection Methods
Current methods for enterovirus detection can be divided in techniques that determine
only the total virus particle number (do not differentiate between the presence of inactive
virus particles and viable virus) for example, immunological, nucleic acid-based, and
techniques that are based in cell culture (infectivity based). The standard method for the
detection of infective viruses uses mammalian cell culture. The principal drawback of
this method is the requirement of complex analyses (visible monolayer cytopathic
effects) that demand several days of laboratory time [41]. Polymerase chain reaction
(PCR) methods for the detection of viruses have been developed, offering specificity,
speed and cost advantages over cell culture methods [42, 43]. PCR methods alone do not,
however, differentiate between the presence of physical (inactive) virus particles and
viable (active) virus particles [44, 45]. Nonetheless, the development of molecular
methods for pathogen detection and source tracking has aided in the monitoring of water
supplies and identifying causative agents in outbreak situations.
In-Situ Hybridization (ISH) can be used to detect nucleic acids inside living cells [46].
Yeh et al. 2009 [47] reviews some advantages of fluorescent probes in conjunction with
fluorescence microscopy. This techniqu, allows the study of dynamic interactions of the
viruses with different cellular structures in living cells that cannot be detected by other
experiments. For example, Hwang et al. [48] used Fluorescence Resonance Energy
Transfer (FRET) for rapid detection of enteroviral infection in vivo. General limitations
of using ISH for pathogen detection include non-specific binding, high limits of detection
29
due to degradation of the probes, and complex methods [46].
3.1 Problems With Current Detection Methods
The current method in use for enterovirus detection use mammalian cell culture and
require complex analyses (such as visible monolayer cytopathic effects) that require
several days of laboratory time [41]. In contrast, PCR based methods offer specificity and
speed but their major disadvantage is their inability to distinguish active from inactive
virus particles leading to false positives on sample infectivity [42]. Cultural methods are
often necessary for the determination of pathogen viability in water.
Differentiation between a viral and bacterial etiology based on clinical symptoms is
currently very difficult. Treating viral infections with antibiotics is ineffective and
contributes to the development of antibiotic resistance, allergic reactions, toxicity and
greater healthcare costs. Limiting the use of unnecessary antibiotics may help prevent the
development of antibiotic resistance, reduce the number of patients with adverse effects
of antibiotics, and substantially decrease healthcare costs [49]. Therefore, an alternative
rapid, sensitive, and commercially available diagnostic tool for the detection of active
viral particles and whether the infection is viral or bacterial in medical diagnostics and
drinking water distribution systems is needed [50].
30
4. Research Objectives
The overall goal of this study was to evaluate the potential of Fourier transform infrared
(FTIR) spectroscopy coupled with cell culture to detect the response of mammalian cell
cultures to viral infection, understand early dynamics of infection, characterize the
limitations as well as the possibility to discriminate between viral infections, and to
develop a practical application of the method in drinking water quality monitoring and
clinical analysis for a proper selection of early medical treatment.
4.1 Specific Aims
1. To develop techniques and evaluate the limits of performance (Appendix A).
2. To assess the ability to identify different types of viruses (Appendix B).
3. To design a continuous monitoring of infective viral particles (Appendix C).
31
PRESENT STUDY
Appendix A describes the development of this novel method for virus detection using the
well-studied poliovirus. It details the first-pass at virus detection using a standard
protocol and then provides a revised, more effective protocol. This method can detect
poliovirus at a titer as low as 17 plaque-forming units (PFU); is most effective at lower
titers (101-103 PFU) and at an interval of 8 hours post infection. Changes in the
carbohydrates, phosphates and lipids were observed relative to uninfected control cells
and likely represent changes in the cell metabolic precursors viral RNA synthesis as the
virus enters and begins to take over cellular functions and processes.
Appendix B summarizes a successful discrimination between three different
enteroviruses, poliovirus PV1, echovirus 30, and Coxsackievirus B5, based on the
changes that they induce in mammalian cell components. The primary cellular
components that provide distinct infrared spectral absorbance features include proteins
(represented by the amide features at 1655 and 1544 cm-1) and the phosphodiester region
between 900 – 1300 cm-1. The results of this study encourage the further pursuit of the
development of FTIR microscopy for virus infection diagnosis.
Appendix C summarizes the work done in the development of a continuous monitoring
systems for active viral particles using attenuated total reflectance (ATR)-FTIR to
continuously monitor the changes in the cells cause by a viral infection. Poliovirus was
used as an model obtaining changes in the peak 915 cm-1, located between the
32
phosphodiester feature region (900 – 1300 cm-1), the peak height increased about 10% in
duplicate experiments after 8 hours post infection, using a poliovirus titer of 106 plaque
forming units per ml (PFU/ml). In response to poliovirus titer with a 104 PFU/ml, the
feature at 967 assigned to DNA deoxyribose, increased by an average of 58.57%. These
two biochemical changes could potentially be related to different stages of poliovirus
infection.
The combination of Appendixes A-C illustrate the progression of the engineered cellbased biosensor in the following manner: Appendix A provides insights of the detection
limits when using BGMK cells to quantify enterovirus and also demonstrates that the
assay perform best when used to detect lower viral titers. Appendix B shows the potential
of this method to be used in the discrimination of enteroviruses and unexploited potential
to be applied in the discrimination between bacterial and viral infection for clinical
applications. Appendix C reviews the steps for the design of an ATR-FTIR cell-based
biosensor for the continuous measurement of biochemical changes in vivo.
Overall Conclusions and Recommendations
Given the complexity required to effectively quantify infective viral particles, FTIR cell
culture in combination with sensitive mammalian cells appears to be a successful
approach toward the design of a faster and sensitive system that could be potentially
applied in monitoring against a bioterrorist attack or for better clinical decisions that lead
to improved patient treatments against a variety of pathogens.
33
REFERENCES
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
Pancrazio JJ, Whelan JP, Borkholder DA, Ma W, Stenger DA: Development and
Application of Cell-Based Biosensors. Annals of Biomedical Engineering 1999,
27:697-711.
El-Ali J, Sorger PK, Jensen KF: Cells on chips. Nature 2006, 442:403-411.
Rodriguez-Mozaz S, Lopez de Alda M, Barceló D: Biosensors as useful tools for
environmental analysis and monitoring. Analytical and Bioanalytical
Chemistry 2006, 386:1025-1041.
Veiseh M, Veiseh O, Martin MC, Bertozzi C, Zhang M: Single-cell-based
sensors and synchrotron FTIR spectroscopy: A hybrid system towards
bacterial detection. Biosensors and Bioelectronics 2007, 23:253-260.
Curtis TM, Widder MW, Brennan LM, Schwager SJ, van der Schalie WH, Fey J,
Salazar N: A portable cell-based impedance sensor for toxicity testing of
drinking water. Lab on a Chip 2009, 9:2176-2183.
Gilchrist KH, Giovangrandi L, Whittington RH, Kovacs GTA: Sensitivity of cellbased biosensors to environmental variables. Biosensors and Bioelectronics
2005, 20:1397-1406.
DeBusschere BD, Kovacs GTA: Portable cell-based biosensor system using
integrated CMOS cell-cartridges. Biosensors and Bioelectronics 2001, 16:543556.
Daniels JS, Pourmand N: Label-Free Impedance Biosensors: Opportunities
and Challenges. Electroanalysis 2007, 19:1239-1257.
Yu N, Atienza JM, Bernard J, Blanc S, Zhu J, Wang X, Xu X, Abassi YA: RealTime Monitoring of Morphological Changes in Living Cells by Electronic
Cell Sensor Arrays:   An Approach To Study G Protein-Coupled Receptors.
Analytical Chemistry 2005, 78:35-43.
Aguilera-Herrador E, Cruz-Vera M, Valcarcel M: Analytical connotations of
point-of-care testing. Analyst 2010, 135:2220-2232.
Maquelin K, Kirschner C, Choo-Smith LP, van den Braak N, Endtz HP, Naumann
D, Puppels GJ: Identification of medically relevant microorganisms by
vibrational spectroscopy. Journal of Microbiological Methods 2002, 51:255271.
Salman A, Erukhimovitch V, Talyshinsky M, Huleihil M, Huleihel M: FTIR
spectroscopic method for detection of cells infected with herpes viruses.
Biopolymers 2002, 67:406-412.
Quaroni L, Zlateva T: Infrared spectromicroscopy of biochemistry in
functional single cells. Analyst 2011. 136:3219-3232
Haes AJ, Van Duyne RP: A Nanoscale Optical Biosensor:   Sensitivity and
Selectivity of an Approach Based on the Localized Surface Plasmon
Resonance Spectroscopy of Triangular Silver Nanoparticles. Journal of the
American Chemical Society 2002, 124:10596-10604.
Petrich W: Mid-infrared and raman spectroscopy for medical diagnostics.
Applied Spectroscopy Reviews 2001, 36:181-237.
34
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
27.
28.
Alam MK, Timlin JA, Martin LE, Williams D, Lyons CR, Garrison K, Hjelle B:
Spectroscopic evaluation of living murine macrophage cells before and after
activation using attenuated total reflectance infrared spectroscopy.
Vibrational Spectroscopy 2004, 34:3-11.
Jackson M, Sowa MG, Mantsch HH: Infrared spectroscopy: a new frontier in
medicine. Biophysical Chemistry 1997, 68:109-125.
Perromat A, Melin A-M, Déléris G: Pharmacologic Application of Fourier
Transform Infrared Spectroscopy: The in Vivo Toxic Effect of Carrageenan.
Applied Spectroscopy 2001, 55:1166-1172.
Liu K-Z, Jia L, Kelsey SM, Newland AC, Mantsch HH: Quantitative
determination of apoptosis on leukemia cells by infrared spectroscopy.
Apoptosis 2001, 6:269-278.
Riley MR, Lucas P, Le Coq D, Juncker C, Boesewetter DE, Collier JL, DeRosa
DM, Katterman ME, Boussard-Plédel C, Bureau B: Lung cell fiber evanescent
wave spectroscopic biosensing of inhalation health hazards. Biotechnology
and Bioengineering 2006, 95:599-612.
Erukhimovitch V, Mukmanov I, Talyshinsky M, Souprun Y, Huleihel M: The use
of FTIR microscopy for evaluation of herpes viruses infection development
kinetics. Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy
2004, 60:2355-2361.
Hastings G, Krug P, Wang R, Guo J, Lamichhane HP, Tang T, Hsu Y-s, Ward J,
Katz D, Hilliard J: Viral infection of cells in culture detected using infrared
microscopy. Analyst 2009, 134:1462-1471.
Vargas CA, Wilhelm AA, Williams J, Lucas P, Reynolds KA, Riley MR:
Integrated Capture and Spectroscopic Detection of Viruses. Appl Environ
Microbiol 2009, 75:6431-6440.
Reynolds KA, Mena KD, Gerba CP: Risk of Waterborne Illness Via Drinking
Water in the United States. Reviews of Environmental Contamination and
Toxicology 2008, 192:117-158.
Pallansch MA, RP R: Enteroviruses: Polioviruses, Coxsackieviruses,
echoviruses, and newer enteroviruses. In. Edited by In: KnipeDM, HowleyPM.
Fields' virology, 4th edn. Philadelphia: Lippincott, Williams & Wilkins.;
2001:723–775.
Khetsuriani N, Lamonte-Fowlkes A, ObersteS, MA P: Centers for Disease
Control and Prevention. Enterovirus surveillance—United States, 1970–
2005. MMWR Surveillance Summ 2006; 55:1–20.
Strikas RA, Anderson LJ, Parker RA: Temporal and Geographic Patterns of
Isolates of Nonpolio Enterovirus in the United States, 1970–1983. Journal of
Infectious Diseases 1986, 153:346-351.
Federal Register: Drinkingwater: regulatory determinations regarding
contaminants on the second drinking water contaminant candidate list—
preliminary determinations; proposed rule. 2007, 72:24016–24058.
35
29.
30.
31.
32.
33.
34.
35.
36.
37.
38.
39.
40.
41.
42.
43.
Pruss-Ustun A, Bos R, Gore F, Bartram J: Safe water, better health: costs,
benefits and sustainability of interventions to protect and promote health.
World Health Organization, Geneva, 2008.
Vuorinen T, Vainionp R, Heino J, Hyypi T: Enterovirus receptors and virus
replication in human leukocytes. Journal of General Virology 1999, 80:921927.
Rotbart HA: Human Enterovirus Infections. Washington, DC: ASM press;
1995.
Koch F, Koch G: The molecular biology of poliovirus. Wien: Springer-Verlag;
1985.
Eischeid AC, Meyer JN, Linden KG: UV Disinfection of Adenoviruses:
Molecular Indications of DNA Damage Efficiency. Applied Environmental
Microbiology 2009, 75:23-28.
Nwachuku N, Gerba CP, Oswald A, Mashadi FD: Comparative Inactivation of
Adenovirus Serotypes by UV Light Disinfection. Applied Environmental
Microbiology 2005, 71:5633-5636.
Straub Tymothy, K HzB, PO O-C, A D, BK M, RA B: In vitro cell culture
infectivity assay for human noroviruses. . Emerging Infectious Diseases [serial
on the Internet] 2007, 13.
Craun, F. M, F. G, Calderon, L. R, Beach, J. M: Waterborne Outbreaks
Reported In The United States. London, Royaume-Uni: Iwa Publishing; 2006.
Bosch A: Human enteric viruses in the water environment: a minireview.
International Microbiology 1998, 1:191-196.
Craun GF, Brunkard JM, Yoder JS, Roberts VA, Carpenter J, Wade T, Calderon
RL, Roberts JM, Beach MJ, Roy SL: Causes of Outbreaks Associated with
Drinking Water in the United States from 1971 to 2006. Clinical Microbiology
Reviews 2010, 23:507-528.
Blackburn BG, Craun GF, Yoder JS, Hill V, Calderon RL, Chen N, Lee SH, Levy
DA, MJ. B: Surveillance for waterborne-disease outbreaks associated with
drinking water--United States, 2001-2002. Division of Parasitic Diseases,
National Center for Infectious Diseases, CDC, USA 2004.
Riley M, Gerba C, Elimelech M: Biological approaches for addressing the
grand challenge of providing access to clean drinking water. Journal of
Biological Engineering 2011, 5:2.
Mantsch HH, Chapman D: Infrared spectroscopy of biomolecules. New York:
Wiley-Liss; 1996.
Reynolds KA, Gerba CP, M. A, I.L P: ICC/PCR detection of enteroviruses and
hepatitis A virus in environmental samples. Canadian Journal of Microbiology
2001, 47:153-157.
Nijhuis M, van Maarseveen N, Schuurman R, Verkuijlen S, de Vos M,
Hendriksen K, van Loon AM: Rapid and Sensitive Routine Detection of All
Members of the Genus Enterovirus in Different Clinical Specimens by RealTime PCR. Journal of Clinical Microbiology. 2002, 40:3666-3670.
36
44.
45.
46.
47.
48.
49.
50.
Metcalf, G. T, Melnick, L. J, Estes, K. M: Environmental Virology : From
Detection Of Virus In Sewage And Water By Isolation To Identification By
Molecular Biology : A Trip Of Over 50 Years. Annual Reviews Microbiology;
1995, pp. 461-487
Koopmans M, Duizer E: Foodborne viruses: an emerging problem.
International Journal of Food Microbiology 2004, 90:23-41.
Procop GW: In situ hybridization for the detection of infectious agents.
Clinical Microbiology Newsletter 2002, 24:121-125.
Yeh H-Y, Yates MV, Chen W, Mulchandani A: Real-time molecular methods
to detect infectious viruses. Seminars in Cell & Developmental Biology 2009,
20:49-54.
Hwang Y-C, Chen W, Yates MV: Use of Fluorescence Resonance Energy
Transfer for Rapid Detection of Enteroviral Infection In Vivo. Appl Environ
Microbiol 2006, 72:3710-3715.
Moreno L, Krishnan JA, Duran P, Ferrero F: Development and validation of a
clinical prediction rule to distinguish bacterial from viral pneumonia in
children. Pediatric Pulmonology 2006, 41:331-337.
Prilutsky D, Shneider E, Shefer A, Rogachev B, Lobel L, Last M, Marks RS:
Differentiation between Viral and Bacterial Acute Infections Using
Chemiluminescent Signatures of Circulating Phagocytes. Analytical Chemistry
2011, 83:4258-4265.
37
APPENDIX A
A NOVEL DETECTION SCHEME FOR POLIOVIRUS INFECTION USING FTIR
SPECTROSCOPY AND CELL CULTURE
Felipe T Lee-Montiel1, Kelly A Reynolds2, Mark R Riley1§
1
Agricultural and Biosystems Engineering, University of Arizona, Tucson, Arizona
85721
2
Mel and Enid Zuckerman College of Public Health, University of Arizona, Tucson,
Arizona 85724
§
Corresponding author
Email addresses:
FTLM: [email protected]
KAR: [email protected]
MRR: [email protected]
To be submitted to the Journal of Biological Engineering.
38
ABSTRACT
Background
In a globalized word, prevention of infectious diseases is a major challenge. Rapid
detection of viable virus particles in water and other environmental samples is essential to
public health risk assessment, homeland security and environmental protection. Current
virus detection methods, especially assessing viral infectivity, are complex and timeconsuming, making point-of-care detection a challenge. Faster, more sensitive, more
specific methods are needed to quantify potentially hazardous viral pathogens and to
determine if suspected materials contain viable viral particles. Fourier transform infrared
(FTIR) spectroscopy combined with cellular–based sensing, may offer a rapid way to
detect specific viruses. This approach utilizes infrared light to monitor changes in
molecular components of mammalian cells by measuring absorbance patterns produced
following virus infection. In this work poliovirus (PV1) was used to evaluate the utility of
FTIR spectroscopy with cell culture for rapid detection of infective virus particles.
Results
Buffalo green monkey kidney (BGMK) cells infected with different virus titers were
studied at 1 – 12 hours post-infection (h.p.i.). A partial least squares (PLS) regression
method was used to analyze and model cellular responses to different infection titers and
times post-infection. Our model performs best at 8 h.p.i., resulting in an estimated root
mean square error of cross validation (RMSECV) of 17 plaque forming units (PFU)/ml
39
when using low titers of infection of 10 and 100 PFU/ml. Higher titers were also tested
up to 106 PFU/ml.
Conclusions
This new approach to poliovirus detection and quantification using FTIR spectroscopy
and cell culture also illustrates the biochemistry behind the stages of viral infection and
could potentially be extended to compare different biochemical cell responses to
infection with different viruses. This virus detection method could be adapted to an
automated plate reader scheme for use in areas such as water safety monitoring and
medical diagnostics.
Keywords: enterovirus, Fourier Transform Infrared (FTIR), zinc selenide (ZnSe), midinfrared, partial least squares, cell culture, buffalo green monkey kidney (BGMK) cells,
virus detection, poliovirus (PV1).
40
BACKGROUND
Increased population density and movement of people around the globe have generated a
rise in the number of outbreaks of infectious diseases and the emergence of new
infectious diseases [1]. Worldwide, 3.575 million people die each year from water-related
diseases [2]. The water and sanitation crises claim more lives through disease than any
warfare [2]. A key step in the prevention of outbreaks of communicable diseases is the
early detection of virulent particles [3]. Rapid detection of active viral pathogens is of
central importance for public health risk assessment and environmental protection.
Waterborne viruses are particularly important for public safety monitoring due to their
environmental stability and low infectious dose; a single virion is sufficient to initiate
illness in previously unexposed, healthy adults [4].
Enteroviruses (family Picornaviridae) are a genus of waterborne viruses that infect
humans and other mammals. They are a health problem worldwide, for example causing
10 to 15 million cases of symptomatic infection in humans annually in the United States
[5]. Enteroviruses are single, positive-strand RNA viruses that include polioviruses,
Coxsackieviruses and echoviruses, among others. Some enteric virus groups have
emerged as waterborne pathogens because of their high levels of resistance to current
water treatment processes, which include ultraviolet light inactivation and heat
inactivation [6, 7]. Poliovirus was used here as a model virus because a large body of
research data exists on the physical, chemical and biological properties of the virus,
vaccination is available and its ease of cell culturing [8-10]. In addition, poliovirus
41
remains endemic in four countries. During 2002 the rejection of polio immunization led
to a worrying resurgence of polio in some areas of Nigeria, followed by re-infection in 21
other countries; resurgence of the disease was also observed in India. Auxiliary
vaccination actions were restarted and by 2007 most re-infected countries had become
polio-free again. The goal of global polio eradication was re-set to 2010, but concerns
continue to be expressed about the progress of this eradication program [11].
Current methods for enterovirus detection use mammalian cell culture and require
complex analyses (visible monolayer cytopathic effects) that require several days of
laboratory time [12]. Polymerase chain reaction (PCR) methods for the detection of
viruses have been developed, offering specificity, speed and cost advantages over cell
culture methods [13, 14]. PCR methods alone do not, however, differentiate between the
presence of physical (inactive) virus particles and viable (active) virus particles [6, 15].
The major disadvantage of most current methods of virus detection is the inability to
provide information about whether a viral particle can start an infection or not. Faster
methods with increased sensitivity and specificity are needed to quantify active viral
pathogens from medical and environmental samples.
FTIR spectroscopy is a noninvasive measurement method that has previously been
applied for identifying various biological components of cells by detecting changes in
vibrations of molecules leading to changes in spectral patterns [12, 16]. It can be used as
part of a sensitive method for the detection of specific cellular molecular changes [12, 17-
42
21]. Quantitative infrared absorption methods such as FTIR spectroscopy differ from
ultraviolet/visible molecular spectroscopic methods because of the greater complexity of
the spectra. FTIR spectroscopy has been applied in medicine, particularly to study the
process of herpes virus infection [5] and in the diagnosis of cancers and other disorders
[22-25]. In addition, FTIR spectroscopy has been used for the quantification of blood
serum components such as glucose, protein, cholesterol and urea [26].
Cell based sensors detect changes in the physiological state of cells following exposure to
an environmental stimulus. Changes in cell state can provide information about the
stimulus; for example, cells have been used to sense toxins in water samples [27, 28].
Cell based sensors have recently been combine with spectroscopy and applied to viral
detection [29, 30]. Cantera et al. [31] used an optical system that involved the use of
molecular beacons as a way to detect infective virus particles, resulting in a detection
limit of 1 PFU. Using live cells to assist in identifying and quantifying viruses in samples
helps to bring the detection closer to an in vivo setting, allowing the natural and complex
interactions between cell and virus to be part of the experimental setup.
Here we present the development of a novel strategy for virus detection using a
combination of FTIR spectroscopy and live BGMK cells as biosensors (Figure 1). The
approach utilizes mammalian cell culture and infrared light absorbance to monitor
specific absorbance patterns produced following changes in cell components (such as
lipids, proteins, nucleic acids and sugars) subsequent to the virus infection, effectively
43
using the cells as biosensors [32, 33]. We also provide insights into the small-scale
changes that host cells undergo upon interacting with a virus and discuss the remaining
challenges to applying this method for virus detection in water samples.
44
RESULTS
Optimal time for virus detection
BGMK cells were infected with PV1 at different multiplicities of infection (m.o.i.) of 10
PFU (0 – 106 PFU/ml) and studied at 1, 1.5, 2, 4, 5, 6, 8 and 12 h.p.i. Virus infection
regression models were developed to correlate changes in spectral features with time of
infection. Changes in the spectra varied depending on the progress of the viral infection,
with biochemical alterations appearing in poliovirus infected cells within 2 h.p.i.
Example regression models for 1.5, 4, 6 and 8 h.p.i. are shown in Figure 2 and a
summary of the regression model parameters are given in Table 1. We selected an
infection time of 8 h for subsequent experiments based on the results here and the viral
replication time [34].
Viral infection at 8 h.p.i.
Spectra of cells at 8 h.p.i infected with different m.o.i. of PV1 are shown in Figure 3. The
highlighted areas are the optimal regions of the spectra used in the virus detection model,
as determined using interval Partial Least Square (iPLS) with an interval size of 10 cm-1
and a maximum of 8 latent variables. These 9 areas of the spectra correspond to the
following wavenumbers: 660.90 – 667.26, 806.11 – 823.46, 844.68 – 862.03, 979.67 –
997.03, 1095.38 – 1112.74, 1191.80 – 1209.16, 1230.37 – 1267.02 and 1326.80 –
1344.16 cm-1.
Changes in absorbance can be correlated with the development of poliovirus infection.
45
The region between 600 – 900 cm-1 corresponds to C2′ endo/anti (B-form helix)
conformation, DNA and RNA molecules. 1000 to 1300 cm-1 relates to symmetric
stretching mode of dianionic phosphate monoester in phosphorylated proteins and left
handed helix DNA (Z form) [35].
Figure 4 shows a regression model of poliovirus at 8 h.p.i. that correlates changes in
absorbance spectra with virus infection titer. This model has a RMSECV of 0.57 log. We
achieved similar results using the whole spectral region or the nine regions selected by
iPLS (data not shown). This FTIR spectroscopy with cell culture method can detect a
viral titer of 101 – 102 PFU/ml with a RMSECV of 17 PFU/ml; at higher titers, 102 –104
PFU/ml, a RMSECV of 2009 PFU/ml for 8 h.p.i. was achieved.
Effect of virus titer on characteristic spectra peak height at 8 h.p.i.
The characteristic peaks of the BGMK cell line showed changes in relative absorbance
following infection with different viral titers. These peaks correspond to different
biomolecules. A graphical representation of the absorbance values of the eighteen
characteristic peaks of each data set for the different infection titers (106, 105, 104, 103,
102 and 101 PFU/ml) for the 8 h.p.i. time point is given in Figure 5. A summary of the
change in peak height upon virus infection and the corresponding biomolecules
represented at the peak wavenumbers are given in Table 2.
46
Some trends are evident for the change in peak heights at different titers. For example,
lower virus concentrations (102 and 103 PFU/ml) showed negative values relative to
controls at 1399 – 2956 cm-1, indicating a decrease in the amount a specific biochemical
component. In contrast, the mean absorbance of the cells with higher virus titers (104,
105 and 106 PFU/ml) only showed average increases in absorbance across all peaks
relative to uninfected controls.
The most significant differences were found for the peak at wavenumber 3293 cm-1 (Oneway ANOVA, d.f. 6, F = 7.63, p < 0.0001). The absorbance of the 101 PFU/ml sample
was significantly higher than the absorbance of the 106 PFU/ml sample, whereas the 106
PFU/ml sample and uninfected control had significantly lower absorbance values than the
101 – 105 PFU/ml samples with an alpha value of 0.05 at this wavenumber (TukeyKramer HSD).
Another significant difference was found for the peak at wavenumber 1043 cm-1 (Oneway ANOVA, d.f. 6, F = 2.47, p = 0.032). Pair-wise comparisons performed for each
absorbance value for each titer (Tukey-Kramer HSD) showed that the absorbance of the
101 PFU/ml sample was significantly higher than the absorbance of the uninfected control
using an alpha value of 0.05.
The mean absorbance values for the different titers at the 700 cm-1 (ANOVA, d.f. 6, F =
2.37, p = 0.0383), 835 cm-1 (ANOVA, d.f. 6, F = 2.36, p = 0.0387) peaks also formed
47
significantly separate groups and the absorbance for the 1079 cm-1 (ANOVA, d.f. 6, F =
2.08, p = 0.0659), 1313 cm-1 (ANOVA, d.f. 6, F = 1.99, p = 0.0781), 2852cm-1
(ANOVA, d.f. 6, F = 1.89, p = 0.0942), 2924 cm-1 (ANOVA, d.f. 6, F = 2.16, p = 0.0575)
titers showed trends.
Microscopy of cell structure on crystal
The cell architecture of the BGMK cells attached to a titanium oxide crystal can be seen
in Figure 6. Actin, CD155 and cell nuclei are shown. Actin mediates a variety of cellular
functions including internal and external cell movement and cell support. The
organization of the actin cytoskeleton is highly regulated; the special distribution of actin
is used as a measure of cell health. The average height of the cells on a titanium oxide
crystal was approximately 6.87 µm. Cells were distributed evenly on the crystal with
many focal adhesion points (data not shown). CD155 poliovirus receptors were observed
to be close to the cell nuclei.
CPE analysis
The error associated with a standard CPE analysis was compared to the error of the new
FTIR spectroscopy with cell culture method. The results of this assay are shown in Table
3. Standard error was calculated from duplicate experiments except where there were too
many PFUs to count accurately. The average standard error was 12.0% of the mean
estimated number of PFU/ml. For comparison, we saw an error of cross validation of
48
~17% of the mean across titers using the FTIR with cell culture method. Note that these
calculations assume a perfect serial dilution to obtain the starting known viral titers.
49
DISCUSSION
Cells infected with PV1 showed consistent alterations in their infrared spectral features.
The broad, undulating features observed relatively weakly in the spectra of entire cells
were previously attributed to Mie scattering of the cellular nuclei [39, 40]. The overall
variations in intensity in the initial method are most likely due to variations in the
thickness of the cell, as well as the nucleus/cytoplasm (N/C) ratio.
Multiple models were developed using a variety of pre-processing approaches, different
spectral regions and different latent variables. For example, this FTIR spectroscopy with
cell culture method could detect a viral titer of 101 –102 PFU/ml with a RMSECV of 17
PFU/ml in cells at 8 h.p.i. Appendix A illustrates the relationship between RMSECV and
latent variables. At higher titers the accuracy decreased slightly, for example at 102-104
PFU/ml, a RMSECV of 2009 PFU/ml for 8 h.p.i. was achieved.
Changes in cell absorbance spectra were correlated with specific cell components to
better understand the progress of the poliovirus infection. The most significant changes
were associated with –OH stretching, particularly at low virus titers. Another significant
change associated with infection was a symmetric stretching vibration of PO2 due to
RNA and DNA, which was most evident in the 101 PFU/ml virus titer. Broad cis-C-H out
of plane bend and left handed helix of DNA also showed significant changes associated
with PV1 infection.
50
In the first method cells were grown in a culture flask, trypsin was used to detach the
cells, they were centrifuged and transferred to the ZnSe crystals. This method, however,
resulted in models with a high level of error. The improved method of growing cells
directly on the ZnSe crystals used a greatly simplified protocol, removing many steps
where error and sample handling variation could occur. The direct cultivation method
produced cleaner, more stable spectral patterns and had a higher reproducibility
compared to the initial method. The modified method required less time and improved
the quality of the spectra. The cleaner spectra results were likely due to reduced light
scattering from more uniform thickness of the cell monolayer.
This virus detection method was more accurate for lower virus titers. This could be
because the lower titers cause less variation in the viral infection process of adsorption
and penetration inside of each individual cell. For example, at low titers the viruses
present may be invading the cells in synchrony, causing more equal changes in the cell
components among cells. With many virus particles present there might be more than one
virus infecting each cell, meaning that the stage of infection is not synchronized among
cells. Cantera et al. [31, 36] have reported greater success for detection of lower virus
titers of poliovirus after 12 h.p.i.. It may be that for lower virus infection it is necessary to
wait longer until the cells reach some steady infection state so that the changes in cell
components are saturated causing the overall signal pattern to stabilize.
51
Poliovirus replicates and lyses the cell in approximately 8h [34] and a previous study
using molecular beacons also found this time period to be the best for viral quantification
[31]. An infection period of eight hours was the most successful of the times tested here.
Prior works support this capability, for example, eight hours of infection was found to be
the best for detection of poliovirus with a method based on engineered BGMK cells
expressing fluorescent proteins undergoing fluorescence resonance energy transfer
(FRET) [36] and [37]. Perhaps this is the time point when the most dramatic changes in
cell components are occurring. The poliovirus is known to have a replication time of
eight hours under standard laboratory conditions in BGMK cells [36], meaning that the
eight hour time point in our protocol may be just prior to virus release from infected cells.
A CPE assay is the standard method for enterovirus detection and requires three to
fourteen days to perform depending on the virus type. The poliovirus assay takes three to
ten days in which the presence of a virus is marked by the death of animal cells in culture
[3]. The cells are grown in a monolayer in a semi-solid media such that new virus
particles infect surrounding cells. Cells killed by the virus form a plaque and the
remaining living cells are stained with a dye. Another method, integrated cell culturePCR (ICC-PCR) is a fast molecular detection method that can identify low levels of
viable virus [3]. A sample is applied to cell culture and then molecular methods are used
to detect replicating virus in the cell culture. This method takes 1 to 3 days from infecting
the cells, depending on the replication time of the virus [13].
52
We present a relatively fast protocol for the detection of enterovirus. Cells need 24h to
attach to the infrared transparent crystal, but with cells ready the sample can be added to
the cells on the crystal and then left undisturbed for 8h for the virus to infect. The sample
measuring process is then completed in a number of minutes. This method could
conceivably be developed into an automated system. For example, a plate reader of ZnSe
with wells for cell growth could be integrated with FTIR microspectroscopy. Software
could be built to automatically compare known changes in cell components with a
database to give an estimate of number of virus particles or to detect the type of virus
present.
53
CONCLUSIONS
Here we present a fast new method for the detection of low titers of viable poliovirus
using mammalian cells as a biosensor combine with FTIR spectroscopy to detect changes
in cell components. Virus titers from 101 to 106 PFU/ml were assayed over a range of
infection times from 1h.p.i. to 12h.p.i., with detection most accurate at low titers from 101
to 104 PFU/ml. Prediction models of infected cells were best detected as early as 6 h.p.i.,
and having the most accurate prediction of virus titer at 8 h.p.i. The model gave a
RMSECV of 17 viral particles. The major changes in cell components following virus
infection were an increase in the peak 3293 cm-1 and negative trend in the absorbance for
lipids region (3200 – 2800 cm-1) in the poliovirus infected cells with 102 – 103 PFU/ml.
The cell culture method is still considered the standard for viral diagnosis as it has the
advantages of detecting infectious viral particles and the ability to achieve low detection
limits [6]. Combining the cell culture method with FTIR spectroscopy, we can enhance
the cell culture method with the increased speed of FTIR spectroscopy. This new method
has the potential to be extended for the detection of other viruses and adapted into a
portable, automated system for detection of viruses from environmental samples.
54
METHODS
Protocol development
An overview of the method is presented in Figure 1. Our first experiments involved
growing cells on the plastic surface of 24-well plates, removing the cells with an
enzymatic treatment, pelleting the cells by centrifugation and then transferring them to
ZnSe crystals to scan using FTIR spectroscopy. We then made several changes to the
initial method to optimize both the laboratory steps and the spectra results by reducing
the room for variation between experimental replicates. We present both methods for
comparison.
Cell culture
BGMK cells were selected due to their high level of sensitivity to enteroviruses and their
use in existing studies [38]. Cells were grown in T25 flasks to a confluent monolayer in
Dulbecco’s Modified Eagle Medium (DMEM) with 584 mg/L L-glutamine (Cellgro)
containing 0.1% NaHCO3, 10,000 units/ml penicillin, 10,000 µg/ml streptomycin
(HyClone) and 10% (vol/vol) fetal bovine serum (FBS; HyClone) and buffered with 12
mM HEPES. All reagents were from Sigma Chemical Co. unless otherwise noted. Cells
were maintained in culture at 37°C in a humidifying incubator with 5% carbon dioxide
and passaged when confluent, approximately every two days. Cells were detached from
the growth flask with 2.5% EDTA-trypsin, centrifuged at 18 g for 8 minutes, resuspended
in media and transferred into a flask of fresh media.
55
Initial method
For all viral exposure experiments, T25 flasks with 95% confluent cell monolayers were
trypsinized, resuspended in media and seeded in 24-well plates. The 24-well plates were
individually seeded with 125,000 cells per well. Cells were allowed to grow to 100%
confluence by incubating for 48 h at 37°C and 5% CO2 before viral exposures. Different
viral exposure times were administered during experiments for cells. Cells were exposed
for different periods of time to monitor viral infection progression. Once cells were
confluent different viral titers were made using DMEM with 10% NCS to expose the cell
layers. The different viral titers were added in the eight central wells of a 24-well plate in
the amount of 0.5 ml and 0.5 ml of media to give a total of 1 ml/well. A minimum of two
control samples per experiment was inoculated with virus-free media.
When viral exposure times were completed, cell layers were exposed to 70 µL of trypsin
per well to remove cells from the wells for FTIR analysis. Once cells were removed, cells
were pelleted by centrifugation at 18 x g for 8 minutes. Cells were washed by adding 1
ml of a 0.9% NaCl saline solution to remove any residual cell media and resuspended in
20 µL of the saline solution. Experiments were performed in a biosafety cabinet under
laminar flow to maintain aseptic conditions.
Revised method
After the quantification of the cell pellet, individual ZnSe crystals were seeded with
150,000 cells with a final volume of 1 ml DMEM media per well in a 24-well plate. Cells
56
were incubated for 24 h to allow attachment to the crystal surface and infected with
different virus titers (0.5 ml of media was removed and replaced with virus in
suspension). ZnSe crystals were washed with soap, pure acetone and sterilized in 70%
EtOH between uses. Experiments with different infection times were performed with the
initial method. The 8 h.p.i. and 12 h.p.i. experiments were subsequently performed with
the revised method.
Virus titers and infection times
Multiple serial dilutions were made from the initial stock concentration of 107 PFU/ml of
purified vaccine strain poliovirus type 1 (PV1, LSc-2ab) using different time of infection
experiments. Previous studies observed molecular and morphological changes in
poliovirus infected cells within 2 h.p.i. [34]. This infection time was therefore used as a
starting point for identification of the optimum incubation time for virus detection using
FTIR spectroscopy. BGMK cells were infected with PV1 at different multiplicities of
infection (m.o.i.) of 10 PFU (0 – 106 PFU/ml) and studied at 1, 1.5, 2, 4, 5, 6, 8 and 12
h.p.i.
FTIR spectroscopy
Infrared spectra of healthy cells and cells post-infection were collected in transmission
mode on a ThermoNicolet Magna 560 FTIR spectrometer equipped with a liquidnitrogen cooled MCT-A detector, KBr beamsplitter and infrared light source. The
57
spectral collection parameters used were 128 co-added scans (requiring about 128
seconds) and a spectral resolution of 4 cm-1 with a 20% aperture opening.
Preprocessing of spectra
The advantage of preprocessing the data is the reduction of noise and other interferences
in the data analysis. This preprocessing included computation of the first derivative of
spectral intensity with respect to wavenumber, vector normalization and mean center.
The derivative spectra exhibited nearly completely flat baselines and collapsed
bandwidth. Spectral expansion emphasized the most diagnostic “mid-IR” region between
650 and 1700 cm−1 and eliminates a large spectral range (1800 – 2800 cm−1), which
contains no vibrational spectroscopic information. Subsequently, spectra were vector
normalized using amide I as the maximum absorbance equal to unity. This last step
removed the variation in spectral intensity discussed above.
Data analysis
Collection of the spectra was performed using OMNIC 7.3 (Thermo Electron
Corporation). The analysis of spectra from individual cells was carried out using the PLS
toolbox. PLS is a well-established multivariate method ideally suited to analyze data with
strongly collinear (correlated), noisy and numerous X-variables distinguishing small
spectral variations in large data sets [41, 42]. Interval Partial Least Square (iPLS) was
used to assist in the selection of narrower regions of the spectra that could be used to
generate a more accurate virus prediction model.
58
OMNIC 7.3 was used to find the absorbance values of the eighteen characteristic peaks
of each data set for the different infection titers (106, 105, 104, 103, 102 and 101 PFU/ml)
for the 8 h.p.i. time point. Differences in peak height were analyzed using the statistical
software JMP version 8.0. Significant differences were assessed among peak heights at
each titer using a one-way ANOVA. Pair-wise comparisons were performed for each
absorbance value for each titer using a Tukey-Kramer HSD test that corrects for multiple
comparisons.
Microscopy of cell structure on crystal
An Actin Cytoskeleton and Focal Adhesion Staining Kit (Millipore) was used to visualize
the cell architecture of the BGMK cells attached to the ZnSe crystal. Preliminary
experiments were conducted growing cells on titanium oxide crystals. Actin was stained
using TRITC conjugated Phalloidin. Nuclei were stained with DAPI. The average height
of the cells on a titanium oxide crystal was determined using a z-axis stack of cell
images. We also used the same kit to perform immunohistochemistry to observe the
CD155 poliovirus receptor in the cell using FITC (Santa Cruz Biotechnology).
Cytopathic effect assay
A CPE assay was conducted as described previously [3]. Briefly, an initial sample with a
known poliovirus titer of 107 PFU/ml was diluted in series. The CPE assay was
performed on each dilution to estimate the virus concentration. 1 ml of 1% (wt/vol)
carboxymethylcellulose (CMC; Sigma-Aldrich) in maintenance medium (with 2% FBS)
59
was overlaid onto the infected-cell monolayer. After 48 h of incubation at 37°C, plaques
were visualized by staining with 0.8% (wt/vol) crystal violet (Sigma-Aldrich) in 3.7%
(vol/vol) formaldehyde solution.
60
List of abbreviations
(CPE): cytopathic effect; (FTIR): Fourier transform infrared; (BGMK): buffalo green
monkey kidney; (PFU): plaque forming units; (DMEM) Dulbecco’s Modified Eagle
Medium; (NCS): new calf serum; (PCR): polymerase chain reaction; (RNA): ribonucleic
acid; (DNA): deoxyribonucleic acid; (RMSECV): root mean square error of cross
validation; (PV1): poliovirus 1; (FBS): fetal bovine serum
Competing interests
The authors declare that they have no competing interests.
Authors' contributions
FLM conducted the experiments, performed the data analysis, participated in the design
and coordination of experiments and drafted the manuscript. MR and KR conceived of
the study and participated in its design and coordination and helped to draft the
manuscript. All authors read and approved the final manuscript.
Acknowledgements
FLM was funded by a CONACYT fellowship. Southwest Environmental Health Sciences
Center (SWEHSC) supported this research. Dr. Charles Gerba and Dr. Kelly Reynolds
generously provided purified vaccine strain poliovirus type 1 (PV1, LSc-2ab), at
University of Arizona. Jonathan Sexton assisted with CPE analysis of the virus samples
used in the development of this method.
61
REFERENCES
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
Jones KE, Patel NG, Levy MA, Storeygard A, Balk D, Gittleman JL, Daszak P:
Global trends in emerging infectious diseases. Nature 2008, 451:990-993.
Pruss-Ustun A, Bos R, Gore F, Bartram J: Safe water, better health: costs,
benefits and sustainability of interventions to protect and promote health.
World Health Organization, Geneva, 2008.
Reynolds K, Gerba C, Pepper I: Detection of infectious enteroviruses by an
integrated cell culture-PCR procedure. Applied Environmental Microbioly
1996, 62:1424-1427.
Schiff GM, Stefanovic GM, Young B, Pennekamp JK: Minimum human
infectious dose of enteric viruses (echovirus-12) in drinking water.
Monographs in Virology 1984, vol. 15:222-228.
Non-polio enterovirus infections
Koopmans M, Duizer E: Foodborne viruses: an emerging problem.
International Journal of Food Microbiology 2004, 90:23-41.
Gerba CP, Gramos DM, Nwachuku N: Comparative Inactivation of
Enteroviruses and Adenovirus 2 by UV Light. Applied Environmental
Microbiology 2002, 68:5167-5169.
Karttunen Å, Pöyry T, Vaarala O, Ilonen J, Hovi T, Roivainen M, Hyypiä T:
Variation in enterovirus receptor genes. Journal of Medical Virology 2003,
70:99-108.
Wimmer, E., Hellen, T. Cu, Cao X: Genetics Of Poliovirus. Palo Alto, Ca, EtatsUnis: Annual Reviews of Genetics; 1993:p. 353-436
Gerba CP: Applied and Theoretical Aspects of Virus Adsorption to Surfaces.
In Advances in Applied Microbiology. Volume 30. Edited by Allen IL: Academic
Press; 1984, 133-168
Rey M, Girard MP: The global eradication of poliomyelitis: Progress and
problems. Comparative Immunology, Microbiology and Infectious Diseases
2008, 31:317-325.
Mantsch HH, Chapman D: Infrared spectroscopy of biomolecules. New York:
Wiley-Liss; 1996.
Reynolds KA, Gerba CP, M. A, I.L P: ICC/PCR detection of enteroviruses and
hepatitis A virus in environmental samples. Canadian Journal of Microbiology
2001, 47:153-157.
Nijhuis M, van Maarseveen N, Schuurman R, Verkuijlen S, de Vos M,
Hendriksen K, van Loon AM: Rapid and Sensitive Routine Detection of All
Members of the Genus Enterovirus in Different Clinical Specimens by RealTime PCR. Journal of Clinical Microbiolgy 2002, 40:3666-3670.
Metcalf, G. T, Melnick, L. J, Estes, K. M: Environmental Virology : From
Detection Of Virus In Sewage And Water By Isolation To Identification By
Molecular Biology : A Trip Of Over 50 Years. Palo Alto, Ca, Etats-Unis:
Annual Reviews of Microbiology; 1995: 461-487
62
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
27.
28.
29.
Erukhimovitch V, Talyshinsky M, Souprun Y, Huleihel M: FTIR Microscopy
Detection of Cells Infected With Viruses. DNA Viruses 2005, 292:161-172.
Jackson M, Kim K, Tetteh J, Mansfield JR, Dolenko B, Somorjai RL, Orr FW,
Watson PH, Mantsch HH: Cancer diagnosis by infrared spectroscopy:
methodological aspects. Edited by Mantsch HH, Jackson M. SPIE 3257; 1998:
24-34.
Vargas CA, Wilhelm AA, Williams J, Lucas P, Reynolds KA, Riley MR:
Integrated Capture and Spectroscopic Detection of Viruses. Applied
Environmental Microbiology 2009, 75:6431-6440.
Riley MR, Lucas P, Le Coq D, Juncker C, Boesewetter DE, Collier JL, DeRosa
DM, Katterman ME, Boussard-Plédel C, Bureau B: Lung cell fiber evanescent
wave spectroscopic biosensing of inhalation health hazards. Biotechnology
and Bioengineering 2006, 95:599-612.
Lucas P, Le Coq D, Juncker C, Collier J, Boesewetter DE, Boussard-Plédel C,
Bureau B, Riley MR: Evaluation of Toxic Agent Effects on Lung Cells by
Fiber Evanescent Wave Spectroscopy. Applied Spectroscopy 2005, 59:1-9.
Wilhelm AA, Lucas P, Reynolds K, Riley MR: Integrated capture and
spectroscopic detection of viruses in an aqueous environment. Edited by
Gannot I. Proc. SPIE; 2008: 68520K-68528.
Argov S, Ramesh J, Salman A, Sinelnikov I, Goldstein J, Guterman H, Mordechai
S: Diagnostic potential of Fourier-transform infrared microspectroscopy and
advanced computational methods in colon cancer patients. Journal of
Biomedical Optics 2002, 7:248-254.
Dukor, K. R, Liebman, N. M, Johnson, L. B: A New, Non-Destructive Method
For Analysis Of Clinical Samples With FT-IR Microspectroscopy. Breast
Cancer Tissue As An Example. Cellular And Molecular Biology 44; 1998:p.
211-217.
Wang HP, Wang HC, Huang YJ: Microscopic FTIR studies of lung cancer cells
in pleural fluid. Science of The Total Environment 1997, 204:283-287.
Choo LP, Mansfield JR, Pizzi N, Somorjai RL, Jackson M, Halliday WC,
Mantsch HH: Infrared spectra of human central nervous system tissue:
Diagnosis of alzheimer's disease by multivariate analyses. Biospectroscopy
1995, 1:141-148.
Shaw, A. R, Kotowich, S., Mantsch, H. H, Leroux, M.: Quantitation Of Protein,
Creatinine, And Urea In Urine By Near-Infrared Spectroscopy. Amsterdam,
Pays-Bas: Elsevier; 1996.
Stenger DA, Gross GW, Keefer EW, Shaffer KM, Andreadis JD, Ma W,
Pancrazio JJ: Detection of physiologically active compounds using cell-based
biosensors. Trends in Biotechnology 2001, 19:304-309.
Banerjee P, Bhunia AK: Mammalian cell-based biosensors for pathogens and
toxins. Trends in Biotechnology 2009, 27:179-188.
Sakudo A, Suganuma Y, Kobayashi T, Onodera T, Ikuta K: Near-infrared
spectroscopy: Promising diagnostic tool for viral infections. Biochemical and
Biophysical Research Communications 2006, 341:279-284.
63
30.
31.
32.
33.
34.
35.
36.
37.
38.
39.
40.
41.
42.
43.
Hastings G, Krug P, Wang R, Guo J, Lamichhane HP, Tang T, Hsu Y-s, Ward J,
Katz D, Hilliard J: Viral infection of cells in culture detected using infrared
microscopy. Analyst 2009, 134:1462-1471.
Cantera JL, Chen W, Yates MV: A simple, rapid and sensitive flow cytometry
method to detect infective poliovirus based on fluorescence resonance energy
transfer (FRET) technology. Applied Environmental Microbiology 2009:
01851-01809.
Pancrazio JJ, Whelan JP, Borkholder DA, Ma W, Stenger DA: Development and
Application of Cell-Based Biosensors. Annals of Biomedical Engineering 1999,
27:697-711.
Haes AJ, Van Duyne RP: A Nanoscale Optical Biosensor: Sensitivity and
Selectivity of an Approach Based on the Localized Surface Plasmon
Resonance Spectroscopy of Triangular Silver Nanoparticles. Journal of the
American Chemical Society 2002, 124:10596-10604.
Koch F, Koch G: The molecular biology of poliovirus. Wien: Springer-Verlag;
1985.
Movasaghi Z, Rehman S, ur Rehman DI: Fourier Transform Infrared (FTIR)
Spectroscopy of Biological Tissues. Applied Spectroscopy Reviews 2008,
43:134-179.
Cantera JL, Chen W, Yates MV: Detection of Infective Poliovirus by a Simple,
Rapid, and Sensitive Flow Cytometry Method Based on Fluorescence
Resonance Energy Transfer Technology. Applied Environmental Microbiology
2010, 76:584-588.
Hwang Y-C, Chen W, Yates MV: Use of Fluorescence Resonance Energy
Transfer for Rapid Detection of Enteroviral Infection In Vivo. Applied
Environmental Microbiology 2006, 72:3710-3715.
Huang YT, Yam P, Yan H, Sun Y: Engineered BGMK Cells for Sensitive and
Rapid Detection of Enteroviruses. Journal of Clinical Microbiology 2002,
40:366-371.
Romeo M, Mohlenhoff B, Diem M: Infrared micro-spectroscopy of human
cells: Causes for the spectral variance of oral mucosa (buccal) cells.
Vibrational Spectroscopy 2006, 42:9-14.
Mohlenhoff B, Romeo M, Diem M, Wood BR: Mie-Type Scattering and NonBeer-Lambert Absorption Behavior of Human Cells in Infrared
Microspectroscopy. Biophysical Journal 2005, 88:3635-3640.
Liebmann B, Filzmoser P, Varmuza K: Robust and classical PLS regression
compared. Journal of Chemometrics 2010, 24:111-120.
Abdi H: Partial least squares regression and projection on latent structure
regression (PLS Regression). Wiley Interdisciplinary Reviews: Computational
Statistics 2010, 2:97-106.
Rodriguez RA, Gundy PM, Gerba CP: Comparison of BGM and PLC/PRC/5
Cell Lines for Total Culturable Viral Assay of Treated Sewage. Applied
Environmental Microbiology 2008, 74:2583-2587.
64
44.
Cervera AE, Petersen N, Lantz AE, Larsen A, Gernaey KV: Application of nearinfrared spectroscopy for monitoring and control of cell culture and
fermentation. Biotechnology Progress 2009, 25:1561-1581.
65
FIGURES
1. BGMK cells grown
in cell culture
2. Cells seeded onto crystals
ZnSe crystal
ZnSe crystal
4. Cells infected with
virus particles
3. Cells incubated
for 24h to allow for
cell adhesion
Incubation time
(h.p.i.)
5. Cells act as a biosensor
6. FTIR spectroscopy to
detect changes in cells
Detector
Detector
Absorbance
Absorbance
7. Raw
spectral data
Wavenumber cm-1
Amines
Lipids
Wavenumber cm-1
9. Prediction of virus titer
Predicted titer
Absorbance
8. Changes in cell components
Wavenumber cm-1
Actual titer
Figure 1 - Schematic representation of viral detection method using cell culture and
FTIR spectroscopy (not to scale)
66
(")$%"&"'"$
*"#$%"&"'"$
!"#$%"&"'"$
+"#$%"&"'"$
Figure 2. Poliovirus prediction models comparing the estimated PFU and predicted
PFU at 1.5 – 8 h.p.i.
The green lines indicate a 1:1 regression model.
67
Figure 3. Example FTIR spectra showing changes in absorbance when cells are infected with poliovirus
Spectra in the wavelength region of 650 – 1750 cm-1 show the absorbance of BGMK cells infected with different PV1 titers
101 – 106 PFU/ml at 8 h.p.i. Uninfected cells served as a control. The nine highlighted regions illuminate the areas chosen by
the PLS model as the most informative for detecting changes in cell components following virus infection. The different colors
represent individual samples.
68
Figure 4 – Example regression analysis for cells infected with PV1 at 8 h.p.i.
This model uses 7 latent variables. The regression uses a log scale and 0–103 PFU/ml in the 650 – 1600 cm-1 wavenumber
region.
69
Figure 5 - Average peak absorbance values compared to uninfected control for different virus titers at 8 h.p.i.
Error bars show standard error. Note that 1654 cm-1 was used to normalize the data and therefore shows no change.
70
Figure 6 - Confocal (Actin filaments and CD155 receptor) and bright field image of BGMK cells
The left panel shows an immunofluorescence image of BGMK cell on a titanium dioxide crystal. Green color indicates CD155
receptor required for the poliovirus to infect the cells. The right panel shows a bright field image of a confluent monolayer of
BGMK cells on a tissue culture flask.
71
TABLES
Table 1 - Comparison of poliovirus prediction models using different infrared
regions and virus titers
Table shows root mean square error of calibration (RMSEC), root mean square error of
cross validation (RMSECV) and correlation coefficient for each model.
Method
IR region used
h.p.i
Virus titers
(PFU)
650 – 1650 cm-1
1.5
0, 102 – 104
4
0.910 0.4225
0.7093
650 – 1650 cm-1
4
0, 101, 103,
104
3
0.891 0.5225
0.8900
-1
650 – 1650 cm
6
0, 101, 102,
104
6
0.868 0.4914
0.8800
650 – 1650 cm-1
8
0 – 104
7
0.911 0.4290
0.8471
650 – 3600 cm-1
8
0 – 104
7
0.819 0.6480
0.9187
650 – 1650 cm-1
8
0 – 103
7
0.917 0.3298
0.5726
iPLS 9 regions
8
0 – 103
7
0.903 0.3577
0.5581
iPLS 9 regions
8
0,102 – 104
7
0.944 0.4019
0.6628
iPLS 9 regions
8
0 – 102
7
0.964 0.1640
0.3163
Initial
Revised
Latent
variables
R2
RMSEC RMSECV
72
Table 2 - Observed changes in relative peak absorbance values with corresponding biomolecules
Peak wavenumber (cm-1)
700
Corresponding biomolecules
Observed changes in cell absorbance following PV1
infection relative to uninfected control cells
(Broad) cis-C-H out-of-plane bend
General increase
835
– 840
C2′ endo/anti (B-form helix) conformation and left handed helix DNA (Z form)
Gradual increase corresponding to increased PV1 titer
900
– 1350
Phosphodiester stretching bands (from absorbance due to collagen and glycogen)
Decrease; trend is less clear at higher PV1 titers
– 100
Symmetric stretching vibration PO2 due to RNA and DNA
General increase
1040
1079/80
Sharp increase in BGMK cells infected with 101 PFU/ml titer
2
Symmetric phosphate stretching modes or ν(PO2 −) symmetric. Phosphate stretching modes followed by a decrease in the 10 PFU/ml infected cells then
a gradual height incremental with the higher titers.
originate from the phosphodiester groups in nucleic acids.
1153
Stretching vibrations of hydrogen bonding
C-OH groups
1236/7
Amide III, stretching PO2−asymmetric (phosphate I)
Decrease in the peak height for the cells infected with 102
and 103 PFU/ml; higher titers shown an increase
1312-1317
Collagen related, amide III band components of proteins
General increase
1455/6
Asymmetric CH3 bending modes of the methyl groups of proteins
General decrease
2852, 2873, 2924/5,
2956
Lipids region (CH2 symmetric, symmetric stretching vibration of CH3 of acyl chains,
stretching C-H and asymmetric stretching vibration of CH3 of acyl chains respectively)
3084
Stretching N-H symmetric
3293
OH stretching (associated)
Virus titer 102 PFU/ml shows a decrease
Decrease for infection titers 102
– 103 PFU/ml
General increase
General increase; increase gradually less from 101
PFU/ml titers
– 106
73
Table 3 - CPE Analysis for Poliovirus at a virus titer of 107 PFU/ml.
Sample of Known Titer (PFU/ml)
Dilution
Factor
Estimated PFU/ml
S.E.
1000000
-6
7000000
2000000
100000
-5
550000
50000
100000
-4
380000
40000
10000
-4
35000
5000
10000
-3
39500
2500
1000
-3
4000
0
1000
-2
3300
100
10
-1
5
5
A standard CPE analysis was performed to compare the error of this method with the
error of the new FTIR spectroscopy with cell culture method. An initial sample with a
known poliovirus titer of 107 PFU/ml was diluted in series. A CPE assay was performed
on each dilution to estimate the virus concentration. Standard error (S.E.) was calculated
from duplicates. The dilution factor indicates which of the subsequent dilutions in the
CPE protocol had a number of plaque forming units in the correct range to be counted.
74
SUPPLEMENTARY DATA
RMSEC and RMSECV
Losing
information
R2
RMSECV
RMSEC
!"#$%&''
Latent Variables
Supplementary Figure 1 - Optimization of the number of latent variables by minimizing
the RMSECV. Image adapted from [44].
75
APPENDIX B
FTIR SPECTROSCOPY AND PCA TO DISCRIMINATE CELLS WITH DIFFERENT
ENTEROVIRUS INFECTIONS
Felipe T Lee-Montiel1, Kelly A Reynolds2, Mark R Riley1§
1
Agricultural and Biosystems Engineering, University of Arizona, Tucson, Arizona
85721
2
Mel and Enid Zuckerman College of Public Health, University of Arizona, Tucson,
Arizona 85724
§
Corresponding author
To be submitted to: Journal of Biotechnology and Bioengineering
Email addresses:
FTLM: [email protected]
KAR: [email protected]
MRR: [email protected]
76
ABSTRACT
Background
Fourier transform infrared (FTIR) spectroscopy was used to analyze buffalo green
monkey kidney (BGMK) cells infected with different enteroviruses. FTIR spectra were
compared among healthy cells and cells infected with one of poliovirus, Coxsackie virus
or echovirus. Advanced statistical methods were applied to analyze the spectra for
discriminating viral infections in BGMK cells. Partial Least Squares (PLS) toolbox
(Eigenvector Research) was used to perform a principal component analysis (PCA),
which allowed us to identify when spectral differences were correlated with different
viral infection types.
Results
It was possible to discriminate and classify the absorbance spectra of BGMK cells with
different viral infections using multivariate chemometric analysis. PCA plots are
presented for the complete spectral range from 3600 to 650 cm-1 and for viruses at three
different titers, 101, 102, and 103 PFU/ml. The partial least squares discriminant analysis
(PLS-DA) using 3 latent variables (LV) for the complete data set had sensitivity and
specificity values of 100% and 100% for poliovirus and echovirus, however Coxsackie
had 95.9% and 94.3% respectively. Using 4 LV for the complete data set, the three
viruses had sensitivity and specificity values of 100% and 100% respectively. A selected
subset of the spectral range from 1614 – 650 cm-1 was found to have the most power for
discriminating cells by virus infection type. A PLS-DA using this subset of the
absorbance data and 4 LV gave rise to sensitivity and specificity values of 100% and
100% respectively.
77
Conclusions
These results demonstrate the utility of FTIR spectroscopy combined with chemometrics
to detect and discriminate virus species based on changes in the biochemical components
of live cells with an active virus infection. This method has the potential to be applied to
distinguish cells with other environmental stimuli. The relative speed of this method,
clearly discriminating virus infections within 8 hours post infection (h.p.i.), indicates that
it could be scaled to use in an automated plate reader to adapted for water safety
monitoring or medical diagnostics.
Keywords: enterovirus, Fourier transform infrared (FTIR) spectroscopy, zinc selenide
(ZnSe), partial least square regression, virus detection, poliovirus (PV1), Coxsackievirus
B5, echovirus 30, principal components analysis (PCA).
78
BACKGROUND
Worldwide, waterborne viruses are of central importance in public health due to their
stability in the environment and their easy transmission throughout water sources. Early
stage detection and discrimination of infectious diseases leads to more effective medical
treatments. Recent outbreaks of poliovirus and other enteroviruses [1, 2] further
emphasize the importance of rapid virus diagnostics to aid in clinical treatment decisionmaking, monitoring of drinking water supplies and assessment of other environmental
samples [3]. Molecular detection methods (nucleic acid-based) for infectious viruses
determine only the total virus particle number and do not differentiate between the
presence of physical virus particles and viable virus (infective particles). Unfortunately,
current methods that asses virus infectivity require several days before a diagnosis can be
made [4, 5]. As a result, the development of faster methods for the detection and
identification of different viruses before the onset of symptoms is needed [6, 7].
Many current detection methods for rapid diagnosis of pathogens are based on molecular
biology. Most of these tests are targeted at specific DNA sequences using PCR, enabling
specific identification or broad spectrum pathogen detection [4]. The major advantages of
molecular diagnostics are the high sensitivity, specificity, and potential speed. For
example, PCR can detect bacteria directly from a sample in a short time [8] and can
theoretically detect pathogen DNA from just a single copy. However, a significant
disadvantage with PCR based methods is their inability to distinguish active from
inactive virus particles leading to false positives on sample infectivity [9]. PCR methods
79
can also reach a limit of detection for rapidly evolving viruses and can be inhibited by
environmental contaminants [10].
Here we present the use of FTIR to discriminate between enteroviruses of the
Picornaviridae family. The Picornaviridae virus family contains small, positivestrand RNA viruses members including polioviruses, Coxsackie A viruses (CA),
Coxsackie B viruses (CB), and echoviruses. Enteroviruses are a worldwide health
problem causing 10 to 15 million cases of symptomatic infection in humans and produce
significant costs associated with the 25,000 to 50,000 hospitalizations for "aseptic"
meningitis annually in the United States [11]. Faster diagnosis of viral infection could
consequently aid patients with suspected infections to choose appropriate treatment
depending on the pathogen and virulence. Discrimination between bacterial or viral
infection would reduce the incidence of antibiotic resistance, allowing more accurate
pharmacological involvement with an overall reduction in health care costs [12].
Spectroscopic methods are being developed for use in the rapid detection of
microorganisms. These methods include those that destroy the sample such as pyrolysis
mass spectrometry and matrix-assisted laser desorption ionization time of flight mass
spectrometry [13, 14]. Fourier transform infrared (FTIR) spectroscopy is a nondestructive
technique based on vibration spectroscopy that has also shown much potential as a tool
for use in microorganism identification [15].
80
The potential use of FTIR spectroscopy to detect and identify changes in cells was
proposed in the early 1990s [16] and increases in computing power have aided its
development and application in biology and medicine [17-19]. It has been applied to
prostate cancer detection [20, 21], to explore the biochemical changes caused by human
papillomavirus [22] and used with partial least squares (PLS) analysis to investigate
problems from Alzheimer disease to waste water management, among others [23].
Salman et al. [24] applied cluster analysis on FTIR microscopic signatures to differentiate
normal cells from herpes-infected cells [25]. Similarly, Alam et al. [26] distinguished
activated from inactivated murine macrophages using FTIR and principal components
analysis (PCA). These prior studies provide a foundation for discriminating healthy from
infected cells but were not able to differentiate types of infection [17].
FTIR spectroscopy uses infrared light to monitor small changes in the molecular structure
of a sample, such as a collection of cells. It can detect changes in absorbance at different
wavelengths that correspond to specific cell components such as lipids, amino acids and
carbohydrates which in turn can be used for the identification, classification and
diagnosis of virus infections [27].
Infrared spectral data provide many predictor variables (wavelengths) relative to the
number of observations, leading to multicollineality. The data therefore requires
chemometric analysis to extract useful information. PCA is a multivariate statistical
method that collapses the many predictor variables into principle components, extracting
81
meaning from the data by combining variables into those that best describe the major data
trends [28]. These combinations of variables can be used to describe and predict trends in
the data and are often more robust than single variables.
Here we investigated the effectiveness of FTIR spectroscopy with PCA for the
discrimination of viruses in the family Picornaviridae. We used the partial least squares
toolbox (Eigenvector Research) to perform PCA analysis on the infrared absorption
spectra of BGMK cells infected with PV1, Coxsackievirus B5 and echovirus 30 and were
able to demonstrate a capability for differentiating the type of viral infection from the
information contained in the cell spectra alone. The potential of this method for virus
identification is discussed.
82
RESULTS
Studies were performed to evaluate the ability to differentiate between three types of viral
infection and uninfected control within BGMK cell cultures. An individual virus type was applied
at any one time. Analysis was performed on the entire data set after collection.
Analysis of raw spectra of enterovirus infected cells
Differences were observed in the raw spectral data of BGMK cells infected with
poliovirus PV1, Coxsackievirus B5 and echovirus 30 for data combined from three
different virus titers: 101, 102 and 103 PFU/ml. Figure 1a shows the complete data set for
the vector-normalized spectra with approximately eight independent replicates per virus
and per titer. Clear differences in absorbance spectra were observed, correlating with
virus infection type, with the most distinct features present in the region from 1614 – 650
cm-1, but especially between 1399 – 1237 cm-1. The region of 3600 – 1750 cm-1 was less
useful for distinguishing cells infected with the different viruses. Figure 1b shows the
vector normalized mean absorbance spectra for cells infected with the three different
viruses, demonstrating clear differences in the region between 1399 – 1237 cm-1.
PCA of complete spectral range
Several methods of data pre-processing were applied.
The first derivative vector
normalization showed better results than the second derivative vector normalization for
the PCA. Figure 2 shows the results of this PCA analysis, which distinguished the
absorbance spectra based on virus infection type for cells infected with three different
83
virus titers: 101, 102 and 103 PFU/ml for each virus type. The total percentage of variance
captured by the PCA model was 93.49% using 3 principal components (PC) with the
complete spectral range of 3600 – 650 cm-1.
PCA using spectral subset 1614 – 650 cm-1
The PCA analysis resulted in a total percentage of variance captured by the PCA model
of 84.20% using 3 PC and 91.01% using 4 PC for the vector-normalized spectral region
between 1614 – 650 cm-1 for the three virus types at the three titers. Figure 3 shows the
vector-normalized spectra for the enterovirus-infected cells in this wavenumber region.
Figure 4 shows the results of the PCA scores plot for this region. A more specific subset
within this region, between 1399 – 1237 cm-1, highlighted in Figure 3, resulted in a total
percentage of variance captured by the PCA model of 98.44% using 3 PC and 99.23%
using 4 PC. The vector normalized spectra and PCA scores plot for this more specific
region are shown in the Appendix.
PLS-DA
The PLS-DA using 4 latent variables (LV) for the complete data set had sensitivity and
specificity values of 100% and 100%, respectively. PLS-DAs using 4 LV for the selected
subsets of the spectral range from 1614 – 650 cm-1 and from 1399 – 1237 cm-1 also had
sensitivity and specificity values of 100% and 100%, respectively. The sensitivity and
specificity values for both the complete spectral range and the large subset did not
decrease when 3 LV were used. The sensitivity and specificity values for the smaller
84
subset dropped to 95.5% and 85.7% respectively when 3 LV were used. Table 1 shows
the specificity and sensitivity values for the different models. Figure 5 shows the PLSDA plots for the full spectral region, the 1614 – 650 cm-1 and 1399 – 1237 cm-1 regions;
the calculated threshold is shown as the horizontal red dashed line. The threshold is
estimated using Bayer’s theorem and the available data in order to minimize total errors.
The Bayesian threshold calculation assumes that the predicted y values follow a
distribution similar to what will be observed for future samples.
PCA using spectral subset 1614 – 650 cm-1 divided by virus titer
Figure 6 shows the PCA score plots for the vector-normalized spectra in the 1614 – 650
cm-1 region divided by virus titer (101, 102, and 103 PFU/ml for Figure 5a, 5b and 5c
respectively). This PCA also distinguished the absorbance spectra based on virus
infection type.
85
DISCUSSION
We have used advanced statistical methods to demonstrate clear discrimination of
absorbance spectra obtained from enterovirus-infected cellular monolayers. This extends
previous work to discriminate uninfected cells from cells infected with poliovirus using
similar methods [29]. This method separated cells by virus infection type with strong
statistical support using the entire FTIR spectral region or a subset. The results of this
study attest to the efficacy of FTIR spectroscopy to discriminate three kinds of virusinfected cells at early stages of infection. It encourages the further pursuit of the
development of FTIR microscopy for virus infection diagnosis.
The primary cellular components that provide distinct infrared spectral absorbance
features include proteins (represented by the amide features at 1655 and 1544 cm-1) and
the phosphodiester region between 900 – 1300 cm-1. A viral infection would be expected
to lead to changes in the cellular composition both due to replication of viral particles and
to damage to the host cell. These would likely lead to distinct changes in the magnitude,
position, or shape of infrared spectral features. It is possible that different viruses, each
having their own mechanisms and rate of replication, may provide unique changes to
these spectral features that would permit a diagnostic method to differentiate between
types of viral infection. Being able to quantify active viral particles is the first step in this
analysis and has been demonstrated by our group recently, which was able to identify
polio infection over a range of titers from 101 to 106 PFU/ml [40]. An increasingly useful
application would be to permit identification of the type of viral infection in addition to
86
the presence of an active infection so as to be able to implement appropriate treatment
methodologies. We report here a clear and quantitative result.
Differentiation between types of viral particles infecting a cell using the infrared spectra
can be performed by visual inspection only when differences among spectra are large.
Figure 1b demonstrates some of these differences, but using this alone is difficult to
quantify the spectral differences. Detecting more subtle changes necessitates the use of
chemometric methods such as PLS, PCA, and PLS-DA. These methods are described in
more detail elsewhere [35]. In brief, PCA results are commonly displayed as a two
dimensional plot of the values of different loading vectors. Samples with similar
underlying features will tend to cluster into unique regions of this phase space.
Differences in loadings are at least in part due to sample-to-sample variation that is
typical of cellular studies. Figure 2 shows a clear demonstration of this clustering
indicating that the spectral information alone can be used to classify a viral infection (at
least from within this group of three evaluated here). This capability is somewhat
surprising, especially with the degree to which clusters do not overlap across virus types.
Providing less information into the analysis (by using a smaller spectral range) or
reducing the number of LV’s applied greatly reduces classification success thus
indicating an approximate level of information required for successful prediction. PLSDA, on the other hand, provides a model that predicts the class type for each sample. The
results shown in Figure 6 provide a more simplified quantitative assessment of
classification performance.
87
We found the spectral region from 1614 – 650 cm-1 to be the most useful for
discriminating virus types at 8 h.p.i. This region incorporates the primary cellular
components that interact with the different viruses including phosphates in DNA and
RNA, amide II, amide III, collagen and glycogen. Further inspection reveals that the
region between 1399 – 1237 cm-1 has the most useful spectral information.
Peak
assignments have been made for this region and indicate the most likely source is due to
collagen and amide III.
This type of data could be used in the future to develop specific vaccines for rapidly
evolving viruses. Alternatively, these cell components could interact similarly to virus
infections but at different h.p.i., such that we observe differences in this snapshot of
infection taken at 8 h.p.i. A continuous spectral monitoring system of live cells infected
with viruses would help to discriminate these two possibilities. The other regions of the
spectra that were not as useful for virus infection type discrimination may still involve
cell components that have substantial interactions with the viral particles, but not in a
way specific to different picornaviruses. Testing of more virus types would help to
explore the general interactions of viruses with the cell components monitored with FTIR
spectroscopy.
The cell components found in the spectral region with the most differences among virus
infections are shown in Appendix A. Interestingly, the region with the clearest
88
differences between virus types corresponds to a collagen receptor. Collagen is a key
component of the extracellular matrix. The alpha-2-beta-1 integrins mediate collagen
formation in the extracellular matrix. Picornaviruses are known to interact with collagen
receptors [30]. The function of the receptor could possibly be modified by the interaction
with the virus. Changes in the extracellular matrix can also lead to changes inside the cell
with gene expression and signal transduction, which could also be perceived by FTIR
spectroscopy. The changes in collagen would then detected by the FTIR spectroscopy
and could be different depending on the virus [31].
Other components in this spectral region include Stretching C-O, deformation C-H,
deformation N-H. A similar spectral region was previously used to correlate FTIR spectra
in the discrimination of different cancer cell lines [32]. Various biological markers, such
as the phosphate level, the RNA/DNA, amide III and collagen, also displayed significant
differences between the control and infected cells based on the analysis of the FTIR
spectra.
The data set provided here is comparatively small and so the high degree of success in
classification may be due in part to low variability across the data set. As the calibration
becomes larger, more variation is incorporated and performance classification typically
improves. Even with this comparatively small data set, sufficient sample variation is
present to calibrate such that 100% classification success is achieved. Reducing the
information content reduces performance as anticipated.
89
The results shown here open new possibilities for classification of viral infections, which
could be performed in an automated manner especially for assessment of safety of
commercial products or municipal drinking water. This method using spectroscopic data
does present some difficulties, including challenges in obtaining high spectral
reproducibility. This could be due to biological or experimental variation causing sample
variation and shifts in the baseline of the spectra. Data processing using corrections such
as taking the first or second derivative of the data can reduce the baseline shift but loses
spectral details. Another problem can be to determine which variables and therefore
which cell components make the most significant contributions to the spectral changes;
here mathematical tools aided this process. Here we focused on the region of the spectra
with the most promising changes correlating with virus infection type. Note that this does
not exclude the utility of other spectral regions and suggest that further analysis may
pinpoint other regions to use in the discrimination of the above enteroviruses used.
90
CONCLUSIONS
We demonstrated that FTIR cellular spectroscopy can discriminate between three types
of enterovirus infection based on small alterations in the overall cellular molecular
composition. This method could be very practical for use in a clinical diagnostic setting
once automated, owing to the minimal need for sample processing and reagents. This
method also provided specific, basic information about the different progress of viral
infections.
ACKNOWLEDGMENTS
FLM was funded by a CONACYT fellowship. Dr. Kelly Reynolds generously provided
purified vaccine strain poliovirus type 1 (PV1, LSc-2ab), Coxsackievirus B5 and
Echovirus 30, at University of Arizona. Jonathan Sexton assisted with CPE analysis of
the virus samples used in the development of this method. Brooke Beam help in the
confocal microscopy.
91
MATERIALS AND METHODS
Cell culture
Cell culture was performed as described in [29]. Briefly, cells were grown in T25 flasks
to a confluent monolayer in Dulbecco’s Modified Eagle Medium (DMEM) with 584
mg/L L-glutamine (Cellgro) containing 0.1% NaHCO3, 10,000 units/ml penicillin, 10,000
µg/ml streptomycin (HyClone) and 10% (vol/vol) fetal bovine serum (FBS; HyClone)
and buffered with 12 mM HEPES at 37°C in a humidifying incubator with 5% carbon
dioxide and passaged when confluent, approximately every two days.
Sample preparation
Zinc selenide crystals, which are highly transparent to IR radiation, were seeded with
150,000 BGMK cells with a final volume of 1 ml DMEM media per well in a 24-well
plate. Cells were incubated for 24 h to allow attachment to the crystal surface. Cells were
then infected with one of three different enteroviruses (or left as uninfected control) with
at estimated concentration of 101, 102 or 103 PFU/ml by removing a volume of 0.5 ml of
media and replacing it with virus in suspension, followed by an incubation period for 8
h.p.i. Uninfected control samples were included for each experiment day, but the data for
these was not included in the PCA because our goal was to discriminate different virus
infection types rather than uninfected from infected cells, as this has been achieved
previously [29]. Samples were air-dried for 10 minutes under laminar flow then dried
further by incubating overnight at room temperature. The dried cells were then examined
then by FTIR spectroscopy. ZnSe crystals were washed with soap, pure acetone and
92
sterilized in 70% EtOH between uses. Virus type and titer replicates were conducted
across many different experiment days to separate treatment from cell passage number.
FTIR spectroscopy
Infrared spectra of healthy cells and cells post-infection were collected in transmission
mode on a Thermo-Nicolet Magna 560 FTIR spectrometer equipped with a liquidnitrogen cooled MCT-A detector, KBr beamsplitter and infrared light source. Spectral
collection parameters used were 128 co-added scans (requiring about 128 seconds) and a
spectral resolution of 4 cm-1 with a 20% aperture opening. A minimum of eight spectra
samples of the cell infected with the different enteroviruses were collected from three
different infection titer and the uninfected control to serve as multivariate biological
fingerprints.
Data processing
Data processing was carried out immediately after spectral acquisition using OMNIC
v.6.3 software. Classification of the enterovirus infected cells was achieved using partial
least squares discriminant analysis that was performed using Partial Lease Squares (PLS)
Toolbox version 6.2.1 (Eigen Vector Research Inc., Wenatchee, WA), operating in a
MATLAB environment (v7.1, The Mathworks Inc., Natick, MA). PLS was used to
preprocess the spectra by taking the first derivative and mean centering of the data set.
Two spectral regions were used: 650 – 3600 cm-1 and 650 – 1614 cm-1. This resulted in
501 spectral data points for analysis in the smaller region. The spectra were then vector
93
normalized in Excel or Matlab, to correct for baseline shifts (scattering) [33]. The spectra
were normalized using an arbitrarily chosen band, amide I (1654 cm-1). Each spectrum
was assigned to one of three defined classes, poliovirus, echovirus 30 and Coxsackie B5.
Two different pre-processing procedures were used on raw spectra and assessed for
optimum discrimination; these steps were (1) vector normalization combined with first
derivatization and (2) vector normalization combined with second derivatization (data not
shown).
Principal components analysis (PCA)
Principal components analysis (PCA) of the baseline corrected and vector normalized
mid-IR spectra data was first performed using data covering the range from 650 – 3600
cm-1 to preliminarily assess the separation of samples. Partial least square (PLS) toolbox
(Eigenvector Research) was used to perform the analysis using two or three principal
components based on the loadings and the percentage of variation captured by the model.
Error ellipses in the PCA plots show 95 % confidence intervals.
PLS-DA
Partial least squares discriminant analysis (PLS-DA) was used to establish statistically
significant differences between mid-infrared spectra for BGMK cells infected with
poliovirus, echovirus 30 and Coxsackie B5. PLS-DA is a multivariate, full-spectrum
calibration method. We specified the descriptor matrix as the spectral data of the cells
infected with the three virus types, and used the PLS-DA to determine the best-fit
94
relationship that minimized the influence of spectral features that vary among replicate
samples. This created a training dataset that could later be used as a classification model
to test unknown samples [37, 38]. The PLS-DA model was built using cross validation
(leave one out). This method used 90% of the data to make the model and tests the
remaining 10% to assess the model, with a total of 10 iterations covering all of the
samples as the test data.
95
FIGURES
(a)!
(b)!
Figure 1 – Complete data set of absorbance of BGMK cells infected with three
different enteroviruses
a. The absorbance spectra of BGMK cells infected with either poliovirus PV1 (red),
96
Coxsackievirus B5 (green) or echovirus 30 (blue) at 8 h.p.i. Each spectra line represents a
different sample. The spectral region ranges from 649.9 – 3600 cm-1. Data is shown for
cells infected with three different virus titers: 101, 102 and 103 PFU/ml.
b. Mean absorbance spectra of BGMK cells infected with either poliovirus PV1 (red),
cocksaki virus B5 (green) or echovirus 30 (blue) at 8 h.p.i. The spectral region ranges
from 649.9 – 3600 cm-1. Spectra represent the mean absorbance for cells infected with
three different virus titers: 101, 102 and 103 PFU/ml.
97
Figure 2 – Principal components analysis for complete data set
PCA score plot for the complete data set using three components, preprocessed with
mean center and first derivative of the spectra. Each point represents one infrared
spectrum. This PCA discriminates the absorbance of BGMK cells infected with either
poliovirus PV1 (red), Coxsackievirus B5 (green) or echovirus 30 (blue) at 8 h.p.i. using
the spectral region from 649.9 – 3600 cm-1. Data is shown for cells infected with three
different virus titers: 101, 102 and 103 PFU/ml.
98
Figure 3 – Absorbance of BGMK cells infected with three different enteroviruses from 1614
– 650 cm-1
The absorbance spectra of BGMK cells infected with either poliovirus PV1 (red),
coxsakievirus B5 (green) or echovirus 30 (blue) at 8 h.p.i. in a region selected to
highlight where the most substantial differences between cells infected with the different
viruses. Each spectra line represents a different sample. The spectral region ranges from
1614 – 650 cm-1. Data is shown for cells infected with three different virus titers: 101, 102
and 103 PFU/ml.
99
Figure 4 - Principal components analysis for BGMK cells infected with three different
enteroviruses at 1614 – 650 cm-1
PCA score plot for the region where the most clear differences between cells infected
with the three viruses were seen in the raw spectral data. Data was preprocessed with
mean center and first derivative of the spectra, using three components. Each point
represents one spectrum. This PCA discriminates the absorbance of BGMK cells infected
with either poliovirus PV1 (red), Coxsackievirus B5 (green) or echovirus 30 (blue) at 8
h.p.i. using the spectral region from 937 – 1614 cm-1. Data is shown for cells infected
with three different virus titers: 101, 102 and 103 PFU/ml.
100
Figure 4b- Principal components analysis for BGMK cells infected with three
different enteroviruses at 1614 – 937 cm-1.
PCA score plot for the region where the most clear differences between cells infected
with the three viruses were seen in the raw spectral data. Data was preprocessed with
mean center and first derivative of the spectra, using three components. Each point
represents one spectrum. This PCA discriminates the absorbance of BGMK cells infected
with either poliovirus PV1 (red), Coxsackievirus B5 (green) or echovirus 30 (blue) at 8
h.p.i. using the spectral region from 937 – 1614 cm-1. Data is shown for cells infected
with three different virus titers: 101, 102 and 103 PFU/ml.
101
(a)
(b)
(c)
Figure 5. PLS-DA class predictions for three different classes of enterovirus
102
In Figure 5, each data point represents a single spectrum. PLS Toolbox software
generates an optimum threshold for sample classification that is plotted as the dashed
line. Spectra that result in Y prediction values greater than the threshold value are
classified as a different type of enterovirus. Figure 5a shows the prediction model for
poliovirus, Figure 5b shows the prediction model for Coxsackievirus and Figure 5c
shows the prediction model for echovirus. Red triangles represent poliovirus, green stars
Coxsackievirus and blue squares echovirus.
103
Figure 6 - Principal component analyses for BGMK cells infected with three
different enteroviruses at 1614 – 650 cm-1 at three virus titers
a. PCA score plot for the region where the clearest differences between cells infected
with the three viruses were seen in the raw spectral data at 101 PFU/ml. This PCA
discriminates the absorbance of BGMK cells infected with either poliovirus PV1 (red),
coxsakievirus B5 (green) or echovirus 30 (blue) at 8 h.p.i. using the spectral region from
1614 – 650 cm-1.
104
b. PCA score plot for the region where the clearest differences between cells infected
with the three viruses were seen in the raw spectral data at 102 PFU/ml. This PCA
discriminates the absorbance of BGMK cells infected with either poliovirus PV1 (red),
coxsakievirus B5 (green) or echovirus 30 (blue) at 8 h.p.i. using the spectral region from
1614-650 cm-1.
c. PCA score plot for the region where the clearest differences between cells infected
with the three viruses were seen in the raw spectral data at 103 PFU/ml. This PCA
discriminates the absorbance of BGMK cells infected with either poliovirus PV1 (red),
coxsakievirus B5 (green) or echovirus 30 (blue) at 8 h.p.i. using the spectral region from
1614-650 cm-1.
105
TABLES
Region
Latent
Variables
Poliovirus
Sensitivity
Specificiy
Coxsackievirus
Echovirus
Sensitivity
Sensitivity
Specificiy
Specificiy
650 – 1650
2
1.000
1.000
0.909
0.943
1.000
0.952
650 – 1650
3
1.000
1.000
0.955
0.943
1.000
1.000
650 – 1650
4
1.000
1.000
1.000
1.000
1.000
1.000
1284 – 1400
2
1.000
1.000
0.818
0.514
0.933
0.881
1284 – 1400
3
1.000
1.000
0.955
0.829
1.000
1.000
1284 – 1400
4
1.000
1.000
1.000
1.000
1.000
1.000
Table 1. Shows the specificity and sensitivity values for the different PLS-DA models
using two different regions of the spectra and 2 – 4 latent variables.
106
APPENDIX
Appendix A - Absorbance of BGMK cells infected with three different enteroviruses from
1400 – 1284 cm-1
All spectra have been normalized at the amide I feature.
107
Appendix B – Principal components analysis using 3PC for the region 1400 – 1284
cm-1
108
1282 cm−1
1284 cm−1
1287 cm−1
13061312 cm−1
1317 cm−1
1327/8 cm−1
1328 cm−1
1330 cm−1
1337/8 cm−1
1339 cm−1
1340 cm−1
1358 /67cm−1
1368 cm−1
1369/70 cm−1
1370/1 cm−1
1370/1/3 cm−1
1373 cm−1
1380 cm−1
1390 cm−1
1395 cm−1
1396 cm−1
1398 cm−1
1399 cm−1
1400 cm−1
Amide III band components of proteins
Collagen
Amide III band components of proteins
Collagen
Deformation N-H
Amide III band components of proteins
Amide III band components of proteins
Collagen
Stretching C-N thymine, adenine
Benzene ring mixed with the CH in-plane bending from the phenyl ring and the ethylene
bridge
CH2 wagging
CH2 wagging
Collagen
In‐plane C-O stretching vibration combined with the ring stretch of phenyl
CH2 wagging
Collagen
Stretching C-O, deformation C-H, deformation N-H
δ(CH2), ν(CC) (polysaccharides, pectin)
Stretching C-N cytosine, guanine
Stretching C-O, deformation C-H, deformation N-H
Deformation N-H, C-H
Stretching C‐N cytosine, guanine
δCH3
Stretching C-O, deformation C-H, deformation N-H
Carbon particle
Less characteristic, due to aliphatic side groups of the amino acid residues
Symmetric CH3 bending of the methyl groups of proteins
CH3 symmetric deformation
Extremely weak peaks of DNA & RNA-arises mainly from the vibrational modes of
methyl and methylene groups of proteins and lipids and amide groups
Symmetric CH3 bending modes of the methyl groups of proteins
δ[(CH3)] sym.
δ[C(CH3)2] symmetric
Symmetric stretching vibration of COO− group of fatty acids and amino acids
δsCH3 of proteins
Symmetric bending modes of methyl groups in skeletal proteins
Specific absorption of proteins
Symmetric stretch of methyl groups in proteins
Appendix C - Biomolecules found in spectral region with more clear differences
(1400 – 1284 cm-1)
Table modified from [34]
109
REFERENCES
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
Arita I, Nakane M, Fenner F: Is Polio Eradication Realistic? Science 2006,
312:852-854.
Rey M, Girard MP: The global eradication of poliomyelitis: Progress and
problems. Comparative Immunology, Microbiology and Infectious Diseases
2008, 31:317-325.
Sinclair RG, Jones EL, Gerba CP: Viruses in recreational water-borne disease
outbreaks: a review. Journal of Applied Microbiology 2009, 107:1769-1780.
Yang S, Rothman RE: PCR-based diagnostics for infectious diseases: uses,
limitations, and future applications in acute-care settings. The Lancet
Infectious Diseases 2004, 4:337-348.
Jungkind D: Molecular Testing for Infectious Disease. Science 2001, 294:15531555.
Lazcka O, Campo FJD, Muñoz FX: Pathogen detection: A perspective of
traditional methods and biosensors. Biosensors and Bioelectronics 2007,
22:1205-1217.
Yeh H-Y, Yates MV, Chen W, Mulchandani A: Real-time molecular methods
to detect infectious viruses. Seminars in Cell & Developmental Biology 2009,
20:49-54.
Belgrader P, Benett W, Hadley D, Richards J, Stratton P, Mariella R, Milanovich
F: PCR Detection of Bacteria in Seven Minutes. Science 1999, 284:449-450.
Reynolds KA, Gerba CP, M. A, I.L P: ICC/PCR detection of enteroviruses and
hepatitis A virus in environmental samples. Canadian Journal of Microbiology
2001, 47:153-157.
Dubois E, Le Guyader F, Haugarreau L, Kopecka H, Cormier M, Pommepuy M:
Molecular epidemiological survey of rotaviruses in sewage by reverse
transcriptase seminested PCR and restriction fragment length polymorphism
assay. Applied Environmental Microbiology 1997, 63:1794-1800.
Non-polio enterovirus infections
[http://www.cdc.gov/ncidod/dvrd/revb/enterovirus/non-polio_entero.htm]
Prilutsky D, Shneider E, Shefer A, Rogachev B, Lobel L, Last M, Marks RS:
Differentiation between Viral and Bacterial Acute Infections Using
Chemiluminescent Signatures of Circulating Phagocytes. Analytical Chemistry
2011, 83:4258-4265.
Haag AM, Taylor SN, Johnston KH, Cole RB: Rapid identification and
speciation of Haemophilus bacteria by matrix-assisted laser
desorption/ionization time-of-flight mass spectrometry. Journal of Mass
Spectrometry 1998, 33:750-756.
Claydon MA, Davey SN, Edwards-Jones V, Gordon DB: The rapid
identification of intact microorganisms using mass spectrometry. Nature
Biotechnology 1996, 14:1584-1586.
110
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
27.
28.
Maquelin K, Kirschner C, Choo-Smith LP, van den Braak N, Endtz HP, Naumann
D, Puppels GJ: Identification of medically relevant microorganisms by
vibrational spectroscopy. Journal of Microbiological Methods 2002, 51:255271.
Helm D, Labischinski H, Schallehn G, Naumann D: Classification and
identification of bacteria by Fourier-transform infrared spectroscopy.
Journal of General Microbiology 1991, 137:69-79.
Cohenford MA, Godwin TA, Cahn F, Bhandare P, Caputo TA, Rigas B: Infrared
Spectroscopy of Normal and Abnormal Cervical Smears: Evaluation by
Principal Component Analysis. Gynecologic Oncology 1997, 66:59-65.
Jackson M, Kim K, Tetteh J, Mansfield JR, Dolenko B, Somorjai RL, Orr FW,
Watson PH, Mantsch HH: Cancer diagnosis by infrared spectroscopy:
methodological aspects. Proc. SPIE; 1998: 24-34.
Jackson M, Sowa MG, Mantsch HH: Infrared spectroscopy: a new frontier in
medicine. Biophysical Chemistry 1997, 68:109-125.
Baker MJ, Gazi E, Brown MD, Shanks JH, Gardner P, Clarke NW: FTIR-based
spectroscopic analysis in the identification of clinically aggressive prostate
cancer. British Journal of Cancer 2008, 99:1859-1866.
Harvey TJ, Gazi E, Henderson A, Snook RD, Clarke NW, Brown M, Gardner P:
Factors influencing the discrimination and classification of prostate cancer
cell lines by FTIR microspectroscopy. Analyst 2009, 134:1083-1091.
Ostrowska KM, Malkin A, Meade A, O'Leary J, Martin C, Spillane C, Byrne HJ,
Lyng FM: Investigation of the influence of high-risk human papillomavirus
on the biochemical composition of cervical cancer cells using vibrational
spectroscopy. Analyst 2010, 135:3087-3093.
Barker M, Rayens W: Partial least squares for discrimination. Journal of
Chemometrics 2003, 17:166-173.
Salman A, Erukhimovitch V, Talyshinsky M, Huleihil M, Huleihel M: FTIR
spectroscopic method for detection of cells infected with herpes viruses.
Biopolymers 2002, 67:406-412.
Mariey L, Signolle JP, Amiel C, Travert J: Discrimination, classification,
identification of microorganisms using FTIR spectroscopy and
chemometrics. Vibrational Spectroscopy 2001, 26:151-159.
Alam MK, Timlin JA, Martin LE, Williams D, Lyons CR, Garrison K, Hjelle B:
Spectroscopic evaluation of living murine macrophage cells before and after
activation using attenuated total reflectance infrared spectroscopy.
Vibrational Spectroscopy 2004, 34:3-11.
Diem M, Romeo M, Boydston-White S, Miljkovic M, Matthaus C: A decade of
vibrational micro-spectroscopy of human cells and tissue (1994-2004). Analyst
2004, 129:880-885.
Jackson J: Principal Components and Factor Analysis. Part 1: Principal
Analysis. Journal of Qualitative Technology 1980, 12, :201.
111
29.
30.
31.
32.
33.
34.
35.
Lee-Montiel Felipe, Reynolds Kelly A, Riley Mark R: A novel detection scheme
for poliovirus infection using FTIR spectroscopy and cell culture. (in
progress).
Marjomaki V, Pietiainen V, Matilainen H, Upla P, Ivaska J, Nissinen L,
Reunanen H, Huttunen P, Hyypia T, Heino J: Internalization of Echovirus 1 in
Caveolae. Journal of Virololy 2002, 76:1856-1865.
Schwimmer R, Ojakian GK: The alpha 2 beta 1 integrin regulates collagenmediated MDCK epithelial membrane remodeling and tubule formation.
Journal of Cell Science 1995, 108:2487-2498.
Gazi E, Baker M, Dwyer J, Lockyer NP, Gardner P, Shanks JH, Reeve RS, Hart
CA, Clarke NW, Brown MD: A Correlation of FTIR Spectra Derived from
Prostate Cancer Biopsies with Gleason Grade and Tumour Stage. European
Urology 2006, 50:750-761.
Bassan P, Byrne HJ, Bonnier F, Lee J, Dumas P, Gardner P: Resonant Mie
scattering in infrared spectroscopy of biological materials - understanding
the 'dispersion artefact'. Analyst 2009, 134:1586-1593.
Movasaghi Z, Rehman S, ur Rehman DI: Fourier Transform Infrared (FTIR)
Spectroscopy of Biological Tissues. Applied Spectroscopy Reviews 2008,
43:134-179.
Eriksson L, Johansson E., Kettanah-Wold, N & Wold S: Introduction to multi
and megavariate data analysis using projection methods (PCA and PLS-DA).
Umetrics AB, Sweden 1999.
112
APPENDIX C
REAL-TIME, CONTINUOUS MEASUREMENT OF CHANGES IN CELLULAR
COMPONENTS FOLLOWING VIRAL INFECTION
Felipe T Lee-Montiel1, Kelly A Reynolds2, Mark R Riley1§
1
Agricultural and Biosystems Engineering, University of Arizona, Tucson, Arizona
85721
2
Mel and Enid Zuckerman College of Public Health, University of Arizona, Tucson,
Arizona 85724
§
Corresponding author
Email addresses:
FTLM: [email protected]
KAR: [email protected]
MRR: [email protected]
To be submitted to: Applied Spectroscopy
113
ABSTRACT
Living cells are microscopic biosensors, constantly perceiving and responding to external
stimuli. Cells can undergo dramatic changes in cellular components following an
encounter with a toxin or pathogen. The precise details of these changes have been
difficult to elucidate using existing methods. A continuous monitoring method for live
cells is needed to understand some of the cellular changes in response to different stimuli.
Here we present a technique for real-time, continuous measurement of alterations in
cellular components following infection with poliovirus. A mammalian cell culture of
Buffalo Green Monkey Kidney (BGMK) cells was grown on a Zinc Selenide trapezoid
crystal and continuously monitored in situ via attenuated total reflectance Fourier
transform infrared (ATR-FTIR) spectroscopy for approximately 12 h. The cells were
grown in a specially designed batch-cell bioreactor that controlled physiological
conditions. Infrared spectra were taken every 3 minutes after 24h from seeding the cells
on the ZnSe crystal to allow cell adhesion. A feature at 915 cm-1 in between the
phosphodiester region 900 – 1300 cm-1 increased about 10% in duplicate experiments
after 8 hours post infection, using a poliovirus titer of 106 plaque forming units (PFU)/ml.
In response to poliovirus titer 104 PFU/ml, the feature at 967 cm-1 assigned to DNA
deoxyribose, increased by an average of 58.57%. These two biochemical changes could
potentially be related to the stages of infection. Analogous optical biosensors are being
applied in many research areas due to the advantages offered by their small size and ease
of implementation. This reagent-free biosensor provides detailed information about cell
114
components and improves on previous methods with potential for use in medicine,
toxicology and environmental sciences.
115
INTRODUCTION
The ability to track changes in living cells under normal physiological conditions is a
vital part of understanding cellular function and the effects of environmental stimuli, such
as drugs, toxins and biologically active molecules. Early detection of infectious diseases
within an individual often leads to an improved outcome by allowing more effective
treatment. Unfortunately, most of the current disease detection methods involve the use
of pathogen-specific macromolecules or host antibody production, each requiring days, if
not weeks, of pathogen replication in the host before a diagnosis can be made.
The mechanisms by which animal viruses, especially non-enveloped viruses, deliver their
genomes are poorly understood [1]. This is due in part to technical difficulties involved in
direct visualization of viral gene delivery and to uncertainties in distinguishing
productive and non-productive pathways caused by the high particle-to-plaque forming
unit ratio of most animal viruses. Furthermore, living organisms are extremely sensitive
to changes in the environment; they can act as biosensors, detecting otherwise
imperceptibly small-scale changes such as the presence of virus particles or trace
amounts of toxins.
Cell-based biosensors are being applied in medicine, drug discovery, toxicology and
environmental monitoring [2]. They are generally constructed by interfacing cells to a
transducer that converts cellular responses into signals detectable by electronic or optical
devices [3]. These biosensors bridge the gap between in vitro, biochemical assays and in
116
vivo, whole organism experiments. They have the advantages of providing more complex
information about biochemical pathways and they can place information gained in a
biochemical assay into the context of the living cell. Cell-based sensors can detect
phenotypic responses and explore functional information in the native cell environment,
leading to a better understanding of the response of all the cellular components as a
whole. The major advantages of these sensing arrays over conventional biosensors
include rapid and inexpensive analyses, much smaller sample size requirement, high
throughput and sensitivity [4]. A good example of a cell-based biosensor is label-free
impedance electrical biosensors; these biosensors can continuously track biochemical and
morphological changes of adherent cells providing quantitative data from in vivo
experiments [5, 6]
The ATR-FTIR system employs evanescence waves to measure changes in the surface of
transparent infrared material. Optical biosensors that employ evanescent waves have seen
widespread utility in both basic and applied research [7-11]. Interest in this field is
rapidly growing as this technology helps to probe the activities of living cells, such as cell
adhesion and spreading, toxicity, and proliferation [12-15].
Many cellular components emit signals in the mid-infrared region, 4000 – 650 cm-1. Midinfrared spectroscopy can be used to observe changes in these molecular machineries. It
can be used to track many aspects of the cell, with broad applications in biomedicine,
cellular biochemistry, disease monitoring, and observation of single-celled organisms
117
[16-18]. It has previously been used to detect changes in cellular components following
exposure to toxins [19] and to detect cell apoptosis [20]. FTIR has been used to detect
herpes infection [21, 22] and to monitor cell changes on infection with poliovirus [43]. Previous studies present a similar approach to detect changes in cell components, but do
not apply the methods to virus detection [23, 24]. Miyamoto et al. [23] performed in situ
observation of cell adhesion using infrared with cells grown on a crystal, however they
used different ATR prism crystal materials, which have different physical and optical
properties. Curtis et. al. [25] designed a portable automated bench-top mammalian cellbased toxicity sensor for drinking water applications that incorporates enclosed fluidic
biochips containing endothelial cells monitored by Electric Cell-substrate Impedance
Sensing technology. A desirable feature for any cell-based biosensor implementation is
the capacity for continuous monitoring or the ability to be reused [26].
The goal of this study was to characterize the early infection steps that occur in cells
caused by poliovirus infection. In the United States, waterborne disease outbreaks have
been associated with treatment deficiencies in the water supply and distribution system
contamination [27]. Close to 50% of all waterborne disease outbreaks are due to acute
gastrointestinal illness caused by unknown agents [28]. The size and life cycle of viruses
can make them difficult to sample and their disease symptoms may be unspecific,
meaning viruses may be the most prevalent undiagnosed disease agent [29].
118
We used an ATR ZnSe crystal covered with a monolayer of BGMK cells to achieve a
non-invasive continuous monitoring of the virus infection process using ATR-FTIR. This
system allowed us to access near-continuous spectral information (measurements were
recorded every 2.6 minutes following infection) and to instantaneously track changes in
cellular activity following a stimulus. We present the relative changes in lipids,
carbohydrates, proteins, and other cell components caused by the viral infection process
in live BGMK cells that characterize the early events in the poliovirus (PV1) infection.
119
RESULTS
Cell adhesion to crystal
The way in which cells attach in monolayer culture depends on the cell type and the
characteristics of the surface. BGMK cells were confirmed to be biocompatible with the
ATR ZnSe prism. The cell architecture of the BGMK cells attached to the ZnSe crystal
can be seen in Figure 1a. Actin, vinculin and cell nuclei were stained to study cell
adhesion of BGMK cells on the ZnSe crystal. The actin filaments are well distributed
throughout the cell with many focal adhesion points, indicating healthy cells. The average
height of the cells on the ZnSe crystal was approximately 4.42 µm. Figure 1b shows the
cells under bright field microscopy, where they can be seen elongated and spread out
over the crystal surface.
Bioreactor design
Figure 2 provides a schematic representation of the experimental setup. The first panel
(Figure 2a) shows how the ATR-FTIR functions, indicating how the infrared light passes
through the crystal creating an evanescent layer that encompasses the cells. The depth of
penetration of the evanescent field is determined by the refractive index of the ATR
crystal, the refractive index of the cells, the angle of the prism and the wavelength of the
incident light [30]. Using a 45° angle in a ZnSe crystal achieves an evanescence field
depth of penetration between 0.3 to 1.6 µm for the midinfrared region, sufficient to
penetrate into the cell membrane.
120
A prototype metal plate biochamber for the cell-based biosensor was built entirely out of
aluminum. Figure 2b shows the prototype drawn using Solidworks 2010 software. The
water jacket surrounds the aluminum base, which is used to control the temperature of the
cells on the crystal. The temperature was monitored and found to be stable at 35.6°C at
the crystal surface using water at 40°C from the water bath driven through jacket. The
chamber is coated in a layer of insulation tape to assist with temperature stability. The
acrylic cover with a polydimethylsiloxane gasket is not shown in Figure 2b. Figure 2c
shows the complete set up of the experiment with the FTIR machine and water bath.
Tracking real-time changes in cells
The ability of this system to track an immediate change in the cells was tested by adding
trypsin to the cell layer to cause the cells to detach from the crystal surface. The spectra
pre- and post-trypsinization is shown in Figure 3. A sharp decrease in the number of
peaks and peak heights was observed upon addition of trypsin, corresponding to a
decrease in cell components being detected by the FTIR spectroscopy. This also confirms
that the cells were attached to the crystal prior to adding the trypsin.
Monitoring changes in cell absorbance following virus infection
The experimental system successfully recorded complete absorbance patterns from 3600
– 650 cm-1 at 2.6-minute intervals for 8 h.p.i. Figure 4 shows spectra collection of DNA
feature (PO2 vibrations) spectrum from a single experiment run for 8 h.p.i. with a
poliovirus titer of 106 PFU/ml. There were no baseline shifts in this experiment, allowing
121
tracking of feature heights without normalization. Spectra taken only at 30-minute
intervals are shown for ease of viewing.
Some interesting regions are shown in more detail with the 2.6-minute readings. The
spectral region between 965 – 996 cm-1 which is related to DNA was one of the regions
with most interesting changes. In experiments using 104 PFU/ml, changes in the feature
monitored every 2.6 min has a total average increase of 58.57% after 8 h.p.i.
Changes in absorbance features over time with 106 PFU/ml poliovirus infection
Figure 5 shows the results of a duplicate experiment. The upper half of Figure 5 shows
the change in the height of the detected absorbance features in the spectra of cells
infected with 106 PFU/ml of poliovirus monitored every 2.6 minutes for 8 h.p.i. Changes
include a jump in the feature absorbance height at 1084, 1239 and 1400 cm-1 for a 120minute window from approximately 160 minutes post infection. These features
correspond to collagen, phosphate bands and methyl groups in skeletal protein. The peak
related to amide I located at 1646 cm-1 decreased over time. The peaks corresponding to
lipids, 2853, 2873, 2924 and 2955 cm-1, displayed a fluctuating absorbance pattern. In
this lipid region are CH2 and CH3 symmetric and asymmetric vibrations, which increase
with chain length.
Changes in absorbance peaks over time with 104 PFU/ml poliovirus infection
The lower half of Figure 5 shows the change in the height of the detected absorbance
122
features in the spectra of cells infected with 106 PFU/ml and 104 PFU/ml of poliovirus
monitored every 2.6 minutes for 8 h.p.i.
Figure 6 shows continuous monitoring of four biochemical features present in BGMK
cells. DNA PO2 vibrations located at 1086 cm-1 increased over time, Amide I and Amide
II protein bands located at 1645 cm-1 and 1545 cm-1 respectively decreased after 4 h.p.i,
lipids (vibrations of CH2 in acyl chains) 2853 cm-1 showed a minimal decrease over 8
h.p.i. Figure 7 shows the changes in DNA PO2 vibrations feature located at 1086 cm-1;
only 30-minute intervals are shown for ease of visualization.
123
DISCUSSION
In the present study, ATR-FTIR was used to study the kinetics of development of
infection with poliovirus PV1 virus. ATR-FTIR spectroscopy uses a multi-bounce
infrared pattern that creates evanescence in the trapezoidal ZnSe crystal. This allowed
scanning of the complete monolayer of cells simultaneously, obtaining more uniform
information than by examining a single cell or a pinpoint sample of cells. Using the ATR
method provides an overall view of the cell states within each experimental batch,
overcoming the fluctuating processes in individual living cells [31].
A new ATR biochamber was designed and built for these experiments. This biochamber
enabled control of the physiological conditions of the system to permit cell growth and
create a realistic cellular environment. The standard, horizontal ATR plate made of
stainless steel reacted with the cell culture media, causing corrosion. This has been
observed previously and occurs because buffers with amino acids react with the stainless
steel [32-34]. The new biochamber was constructed using 100% aluminum, which
prevented further problems with corrosion. The chamber was held in place with four
large screws and sealed with a silicone gasket. The silicone caused negligible interference
with the spectra.
One of the challenges for the design of this cell-based biosensor is the proper
maintenance of the environment suitable for cell or tissue survival, which requires a
humid environment. The presence of high water concentrations in the cell culture media
124
causes absorbance of mid-IR region. This is a major experimental challenge for the
accurate IR absorption measurements from cell culture [35]. Interference from water is
caused by its strong absorption bands in the region between 3600 – 3200 cm-1[31] .
The commercially available ATR plate with a volume of 1.5 ml did not hold a sufficient
volume of media to complete the experiment, thereby limiting nutrient availability for the
cells. The increase in chamber volume to a capacity of >5 mL allowed the cells to survive
longer; 2 ml of media was sufficient for an 8 h experiment. The acrylic lid also prevented
evaporation of the media and thereby reduced the noise in the spectra. The cells adhered
to the surface of the crystal with or without a functionalization layer. The only factor in
this protocol that appears to affect cell adhesion to the crystal surface is hydrophobic
residuals, which can be cleaned from the surface.
Confocal microscopy was used to take 3D images of the cytoskeleton of the cells on the
crystal surface to characterize the height and degree of cell adhesion across the surface.
Cell adhesion plays an important role in many fundamental cellular processes such as cell
morphology [43]. One of the distinctive features of adhesion to a rigid surface is the
ability of forces to develop at the contact site. Such forces can be external, such as shear
flow in blood vessels, or can be generated by the cells own contractile system.
The main sites of cell adhesion to the substrate are focal adhesions (vinculin). These
complex multimolecular assemblies link the extracellular matrix, via membrane-bound
125
receptors, to the cell’s cytoskeleton [44]. Focal adhesions are therefore also the sites at
which forces are transmitted to the substrate [41]. The cells grown on a ZnSe surface had
a height average of 5.43 µm; this height indicates that the cells spread evenly across the
crystal surface. The BGMK cells also had many well-distributed focal adhesion points,
showing a good attachment to the surface. Cells attached only to an underlying stiff
surface differ in their spreading and cytoskeletal organization [43], for example, in cell
types that grow preferentially on hard matrices, the tension will stimulate such a cell to
extend about its periphery [45].
Previous experiments exhibited a cell height average of approximately 6.87 µm on
titanium oxide crystals and 8.25 µm on glass. On very stiff substrates, stress fibers and
strong focal adhesions predominate, suggesting equal contraction [42]. These
observations by confocal microscopy show that the cells are biocompatible with the
ZnSe. The cell height of 5.43 µm mean that only a maximum of 36% is probed by the
penetration of the evanescence that reaches a range between 0.3 – 1.6 µm in the IR region
between 650 – 4000 cm-1.
The DNA feature at 1084 cm-1 corresponding to the phosphate vibrations increased in
peak absorbance height over time. This could be related to the activity of the virus in the
cell. The amide I and amide II peaks (1645 and 1545 cm-1) are associated with the
amount of biomass. When the virus causes cell lysis, the cells will no longer be attached
to the surface. The decrease in the absorbance of the amide peaks may be related to a
126
reduction in protein due to cells starting to lyse or change shape such that they are no
longer in the evanescence field.
The ability to study in vivo dynamic processes has numerous potential applications. This
new method of continuous measurement of a live cell using the ATR biosensor is not
constrained to viral detection. The most immediate application in this respect is to boost
molecular and morphological data by measuring changes of biomolecules in single cells
as a function of time following a stimulus (for example, a growth factor, nanoparticles,
other viruses or toxins) [41, 42] or to compare different cell types (for example,
cancerous and non-cancerous) following the same stimulus. It could be developed further
for more complex samples (clinical blood samples or drinking water), although using the
existing system the samples would first need to be filtered or centrifuged.
This method has given an insight into the real-time changes that occur in cellular
components upon virus infection. It will be interesting to perform future comparisons
between different viruses and to track the steps of infection of each virus, as there is
limited available information about infection kinetics with this timing and level of detail.
For example, it is know that after 30 minutes the virus will release RNA, but the detailed
biochemistry in between these steps could be elucidated with an ATR-FTIR biosensor.
We see this experiment with poliovirus as a starting point to investigate other viruses
with this method.
127
CONCLUSIONS
Cell-based biosensors have promising potential in numerous applications ranging from
pharmaceutical screening to environmental monitoring. Cellular responses of a
monolayer of mammalian cells caused by the poliovirus were studied using ATR-FTIR
spectroscopy. Changes in the IR spectra were correlated with some changes in the
molecular cell components at different stages of the infection process. A custom designed
biochamber allowed near-continuous acquisition of real-time spectral data for many
hours following infection. The results of this study attest to the unexploited potential of
ATR-FTIR spectroscopy for tracking molecular changes in cells following an encounter
with an environmental stimulus.
This type of biosensor could be used for continuous monitoring of viral agents or toxic
chemicals in drinking water distribution systems for a rapid hazard assessment in
potential public health threats. Potentially, the first use of this biosensor could be used as
a presumptive test and not as a substitute for standard analytical tests.
128
FIGURES
a
b
FIGURE 1. Microscopy of cells attached to ZnSe crystal
a) Confocal image of BGMK cells adhered to a ZnSe ATR prism. Actin (red), Vinculin
(green) and cell nuclei (blue) are shown.
b) Bright microscopy image of BGMK cells adhered to a ZnSe ATR prism.
Confocal image of cells on the ZnSe crystal showing the focal addition points, the actin
filaments and the height of the cells. The height of the cells is linked to the evanescence
wave of penetration of the ATR, which has an estimated depth of penetration between 0.3
129
– 1.6 µm allowing us continuous monitoring of the cellular processes. This parameter
depends on the angle of crystal (45°) and the refractive index of the prism material.
FIGURE 2 – Experimental setup
a) Schematic of the ATR-FTIR biosensor.
b) Schematic of ATR-FTIR biochamber design.
c) Complete experimental setup.
Absorbance
130
Wavenumbers cm-1
FIGURE 3. Continuous spectral collection of BGMK cells attached to a ZnSe crystal
Cells were scanned every 2.6 minutes to track any change in their biochemistry. Scans 14
– 20 show the fast response of the biosensor to changes in the cell adhesion produced by
trypsin, which detached the cells from the surface causing a reduction in absorbance.
131
DNA (band due to PO2 ! vibrations)
20min
8 hr
FIGURE 4. Spectra collection of BGMKS cells infected with a poliovirus titer of 104
PFU/ml. DNA feature is shown between the region 1250 – 945 cm-1
Each series represents continuous collection of the spectra with 30 minutes intervals to
track cellular changes cause by the virus infection.
132
PV1 106 PFU
!#*"
8%%"
8''"
!#&$"
(&&"
!#&"
('*"
7%$"
!#%$"
PV1 106 PFU
!#&$"
7%$"
78%"
!#&"
%!)'"
%!()"
!#%$"
%&*7"
78%"
%!()"
!#%"
%)!!"
!#%"
%)$'"
%&*7"
%&'("
!#!$"
%$)$#'"
!#!$"
%')$#7"
%)!!"
!"
!"
'!"
%&!"
%(!"
&)!"
*!!"
*'!"
)&!"
)(!"
+,-./01"231/",-40563-"
%)$'"
&($*"
!"
!"
'!"
%$)$#'"
&(8*"
%&!" %(!" &)!" *!!" *'!" )&!" )(!"
&7&)#$'"
+,-./01"231/",-40563-"
PV1 104 PFU
PV1 104 PFU
!#*"
!#)$"
87)#$"
!#&$"
(''#("
78$#("
!#&"
%!)8#%"
%!()#8"
!#%$"
%&)%#7"
!#%"
!#!$"
!"
!"
'!"
%&!" %(!" &)!" *!!" *'!" )&!" )(!"
+,-./01"231/",-40563-"
(*(#("
!#*$"
7%)#%"
!#*"
7'8#%"
!#&$"
%!()#8"
!#&"
%)!!"
!#%$"
%)$'"
!#%"
%$)$#'"
!#!$"
%'$&#8"
!"
&($!#*"
8&%"
!#)"
%%%(#$"
%&*!#*"
%)!!"
%)$'#7"
%$)'#'"
!"
'!"
%&!"
%(!"
&)!"
*!!"
*'!"
)&!"
+,-./01"231/",-40563-"
Figure 5. BGMK peaks detected using OMINIC 7.3 software after an infection
process
Peaks heights were monitored every 2.6 min for 8 h.p.i. to study the early steps of
poliovirus infection.
%')$#7"
133
BGMK cells infected with PV1 10^6 peak height average 0.25 Absorbance 0.2 0.15 1084 1545.6 0.1 1645.9 2853 0.05 0 0 60 120 180 240 300 360 420 480 minutes post infection Figure 6. BGMK peak average height monitoring every 2.6 minutes for cells
infected with Poliovirus 106 PFU/ml for 8 h.p.i. to study the early steps of poliovirus
infection.
134
Peak assigned to DNA dexoyribose, increased in an average of 58.57%.
975.8 cm-1
967.1 cm-1
!#("
!#!)"
!#!$7"
!#!'"
*+,-.+/012"
"
*+,-.+/012"
!#!$"
!#!&"
!#!%"
!#!(7"
!#!("
!#!$"
!#!!7"
!"
!"
!"
&!"
($!" ('!" $%!" )!!" )&!" %$!" %'!"
340562,"
!"
&!"
($!" ('!" $%!" )!!" )&!" %$!" %'!"
340562,"
Figure 7. BGMK peak height monitoring every 2.6 minutes for 8 h.p.i. to study the
early steps of poliovirus infection
135
MATERIALS AND METHODS
Cell Culture
Cell culture was performed as previously described [40] with the following
modifications: 20 mM HEPES buffer was used instead of sodium bicarbonate and
DMEM 1X media was used without phosphates (MP Biomedical) to avoid any
phosphates interference.
Virus Titers
Multiple serial dilutions using DMEM 1X without phosphates (MP biomedical) were
made from the initial stock concentration of 107 PFU/ml of purified vaccine strain
poliovirus type 1 (PV1, LSc-2ab). These were used in experiments with varying infection
times.
FTIR Spectroscopy
Infrared spectra of healthy cells and cells post-infection were collected in transmission
mode on a ThermoNicolet Magna 560 FT-IR equipped with a liquid-nitrogen cooled
MCT-A detector, KBr beamsplitter, and infrared light source. The spectral collection
parameters used were 128 co-added scans (requiring 2.5 minutes) and a spectral
resolution of 2cm-1 with a 68% aperture opening. A trapezoid ZnSe 45° angle, 80mm x
40mm x 10mm (Pike Technologies) was used for all experiments.
136
Biochamber Design
To analyze the living culture in near real-time, we constructed a horizontal ATR
biochamber of 100% aluminum around the ATR crystal. We increased the height of the
channel to hold a total of 5ml of media. We also included an acrylic cover to prevent
evaporation. The chamber was held together with a PDMS gasket between the acrylic top
and the aluminum body with 4 screws to compress the gasket and seal the top. The sides
of the chamber were further insulated by polyethylene tape foam. A schematic view of
the environment control system is shown in Fig. 2c. The system consisted of a water bath
(VWR) with a temperature controller and a pump to circulate hot water (41°C) in the
biochamber (around the outside of the gasket). The temperature of the water bath allowed
us to maintain the crystal at a constant 36°C.
Experimental Setup
After the cells in culture were confluent in the T25 flasks, cells were trypsinized and
resuspended. Cells were quantified using a hemocytometer and approximately 106 cells in
a total volume of 2ml of DMEM media with phosphates seeded into the biochamber
directly onto the crystal. The biochamber was held for 24h in an incubator at 36°C to
allow the cells to adhere to the crystal, forming a confluent monolayer. Cells were
infected with 0.5ml of poliovirus samples at different titers and an additional 1.5ml of
DMEM media without phosphates was added to the biochamber. The biochamber was
used inside the FTIR sample compartment, connected to the water bath as described
above.
137
The surface of the crystal was treated occasionally with 5% hydrogen peroxide for 1
minute to remove any hydrophobic residuals and make the surface more hydrophilic and
favorable for cell adhesion. Between experiments, the crystal was rinsed with soapy
water and distilled water, then sterilized using 70% ethanol.
Continuous Data Measurement
Collection of the spectra was performed using OMNIC 7.3 from Thermo Electron
Corporation. We recorded spectra from the cells every two minutes following virus
infection for several hours. Collection was automated using Macros/Basic OMINC
software and converted to a single Excel file using a Matlab code.
Data Analysis
Absorbance was computed using the DMEM without phosphates spectrum as a
background. Automatic water vapor and CO2 correction was implemented using this
software. The analysis of spectra from the cell monolayer was carried out in Excel 2007
where we were looking to distinguish small variations in data that is strongly collinear
and noisy and has many X-variables [37, 38].
Confocal Microscopy
An Actin Cytoskeleton and Focal Adhesion Staining Kit (Millipore) was used to visualize
the cell architecture of the BGMK cells attached to the ATR ZnSe prism. Actin filaments
were stained using TRITC conjugated Phalloidin. Cell nuclei were stained with DAPI
138
and viculin adhesion points were stained with FITC. The average height of the cells on a
ZnSe prism was determined using a z-axis stack of cell images.
ACKNOWLEDGEMENTS
FLM was funded by a CONACYT fellowship. Southwest Environmental Health Sciences
Center (SWEHSC) supported this research. Virus stock samples were generously
provided by Dr. Charles Gerba and Dr. Kelly Reynolds at The University of Arizona. Dr.
Brooke Beam helped in the confocal imaging of the cells. We would also like to thank
Charlie DeFer for his help and advice with the milling machine.
139
REFERENCES
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
Brandenburg B, Lee LY, Lakadamyali M, Rust MJ, Zhuang X, Hogle JM:
Imaging Poliovirus Entry in Live Cells. PLoS Biology 2007, 5:e183.
Yamaguchi R-t, Hirano-Iwata A, Kimura Y, Niwano M, Miyamoto K-i, Isoda H,
Miyazaki H: Real-time monitoring of cell death by surface infrared
spectroscopy. Applied Physics Letters 2007, 91:203902-203903.
Veiseh M, Veiseh O, Martin MC, Bertozzi C, Zhang M: Single-cell-based
sensors and synchrotron FTIR spectroscopy: A hybrid system towards
bacterial detection. Biosensors and Bioelectronics 2007, 23:253-260.
Pancrazio JJ, Whelan JP, Borkholder DA, Ma W, Stenger DA: Development and
Application of Cell-Based Biosensors. Annals of Biomedical Engineering 1999,
27:697-711.
Giaever I, Keese CR: A morphological biosensor for mammalian cells. Nature
1993, 366:591-592.
Daniels JS, Pourmand N: Label-Free Impedance Biosensors: Opportunities
and Challenges. Electroanalysis 2007, 19:1239-1257.
Cooper MA: Optical biosensors in drug discovery. Nature Reviews Drug
Discovery 2002, 1:515-528.
Rich RL, Hoth LR, Geoghegan KF, Brown TA, LeMotte PK, Simons SP, Hensley
P, Myszka DG: Kinetic analysis of estrogen receptor/ligand interactions.
Proceedings of the National Academy of Sciences 2002, 99:8562-8567.
Bureau B, Zhang XH, Smektala F, Adam J-L, Troles J, Ma H-l, Boussard-Plèdel
C, Lucas J, Lucas P, Le Coq D, et al: Recent advances in chalcogenide glasses.
Journal of Non-Crystalline Solids 2004, 345-346:276-283.
Lucas P, Riley MR, Boussard-Plédel C, Bureau B: Advances in chalcogenide
fiber evanescent wave biochemical sensing. Analytical Biochemistry 2006,
351:1-10.
Lucas P, Le Coq D, Juncker C, Collier J, Boesewetter DE, Boussard-Plédel C,
Bureau B, Riley MR: Evaluation of Toxic Agent Effects on Lung Cells by
Fiber Evanescent Wave Spectroscopy. Applied Spectroscopy 2005, 59:1-9.
Ramsden JJ, Li SY, Prenosil JE, Heinzle E: Kinetics of adhesion and spreading
of animal cells. Biotechnology and Bioengineering 1994, 43:939-945.
Horvath R, Pedersen HC, Skivesen N, Selmeczi D, Larsen NB: Monitoring of
living cell attachment and spreading using reverse symmetry waveguide
sensing. Applied Physics Letters 2005, 86:071101-071103.
Vörös J, Graf R, Kenausis GL, Bruinink A, Mayer J, Textor M, Wintermantel E,
Spencer ND: Feasibility study of an online toxicological sensor based on the
optical waveguide technique. Biosensors and Bioelectronics 2000, 15:423-429.
Quinn JG, O'Neill S, Doyle A, McAtamney C, Diamond D, MacCraith BD,
O'Kennedy R: Development and Application of Surface Plasmon ResonanceBased Biosensors for the Detection of Cell-Ligand Interactions. Analytical
Biochemistry 2000, 281:135-143.
140
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
27.
28.
29.
30.
31.
Petrich W: Mid-infrared and raman spectroscopy for medical diagnostics.
Applied Spectroscopy Reviews 2001, 36:181-237.
Alam MK, Timlin JA, Martin LE, Williams D, Lyons CR, Garrison K, Hjelle B:
Spectroscopic evaluation of living murine macrophage cells before and after
activation using attenuated total reflectance infrared spectroscopy.
Vibrational Spectroscopy 2004, 34:3-11.
Jackson M, Sowa MG, Mantsch HH: Infrared spectroscopy: a new frontier in
medicine. Biophysical Chemistry 1997, 68:109-125.
Perromat A, Melin A-M, Déléris G: Pharmacologic Application of Fourier
Transform Infrared Spectroscopy: The in Vivo Toxic Effect of Carrageenan.
Applied Spectroscopy 2001, 55:1166-1172.
Liu K-Z, Jia L, Kelsey SM, Newland AC, Mantsch HH: Quantitative
determination of apoptosis on leukemia cells by infrared spectroscopy.
Apoptosis 2001, 6:269-278.
Erukhimovitch V, Mukmanov I, Talyshinsky M, Souprun Y, Huleihel M: The use
of FTIR microscopy for evaluation of herpes viruses infection development
kinetics. Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy
2004, 60:2355-2361.
Hastings G, Krug P, Wang R, Guo J, Lamichhane HP, Tang T, Hsu Y-s, Ward J,
Katz D, Hilliard J: Viral infection of cells in culture detected using infrared
microscopy. Analyst 2009, 134:1462-1471.
Miyamoto K-i, Yamada P, Yamaguchi R-t, Muto T, Hirano A, Kimura Y, Niwano
M, Isoda H: In situ observation of a cell adhesion and metabolism using
surface infrared spectroscopy. Cytotechnology 2007, 55:143-149.
Hutson TB, Mitchell ML, Keller JT, Long DJ, Chang MJW: A technique for
monitoring mammalian cell growth and inhibition in situ via Fourier
transform infrared spectroscopy. Analytical Biochemistry 1988, 174:415-422.
Curtis TM, Widder MW, Brennan LM, Schwager SJ, van der Schalie WH, Fey J,
Salazar N: A portable cell-based impedance sensor for toxicity testing of
drinking water. Lab on a Chip 2009, 9:2176-2183.
Paddle BM: Biosensors for chemical and biological agents of defence interest.
Biosensors and Bioelectronics 1996, 11:1079-1113.
Craun GF, Calderon RL, Wade TJ: Assessing waterborne risks: an
introduction. Journal of Water Health 2006, Suppl. 2:3-18.
Reynolds KA, Mena KD, Gerba CP: Risk of Waterborne Illness Via Drinking
Water in the United States. Reviews of Environmental Contamination and
Toxicology 2008, 192:117-158.
Yeh H-Y, Yates MV, Chen W, Mulchandani A: Real-time molecular methods
to detect infectious viruses. Seminars in Cell & Developmental Biology 2009,
20:49-54.
Chittur KK: FTIR/ATR for protein adsorption to biomaterial surfaces.
Biomaterials 1998, 19:357-369.
Quaroni L, Zlateva T: Infrared spectromicroscopy of biochemistry in
functional single cells. Analyst 2011, 136:3219-3232.
141
32.
33.
34.
35.
36.
37.
38.
39.
40.
41.
42.
43.
44.
45.
Vinnichenko MV, Pham MT, Chevolleau T, Poperenko LV, Maitz MF: In situ
ellipsometric investigation of stainless steel corrosion behavior in buffered
solutions with amino acids. Applied Surface Science 2003, 207:176-182.
In vitro degradation performance and biological response of a Mg–Zn–Zr
alloy. Materials Science and Engineering B 2011, (in press).
Gu XN, et al.: Influence of artificial biological fluid composition on the
biocorrosion of potential orthopedic Mg–Ca, AZ31, AZ91 alloys. Biomedical
Materials 2009, 4:065011.
Judith RM, et al.: Methods for measuring the infrared spectra of biological
cells. Physics in Medicine and Biology 2003, 48:243-257.
Heumann D, Roger T: Initial responses to endotoxins and Gram-negative
bacteria. Clinica Chimica Acta 2002, 323:59-72.
Gribbon P, Sewing A: Fluorescence readouts in HTS: no gain without pain?
Drug Discovery Today 2003, 8:1035-1043.
Heraud P, Wood BR, Tobin MJ, Beardall J, McNaughton D: Mapping of
nutrient-induced biochemical changes in living algal cells using synchrotron
infrared microspectroscopy. FEMS Microbiology Letters 2005, 249:219-225.
Hoi-Ying NH, Michael CM, Wayne RM: Tracking chemical changes in a live
cell: Biomedical applications of SR-FTIR spectromicroscopy. Spectroscopy
2003, 17:139-159.
Lee-Montiel Felipe, Reynolds Kelly A, Riley Mark R: A novel detection scheme
for poliovirus infection using FTIR spectroscopy and cell culture. (in
progress).
Balaban NQ, Schwarz US, Riveline D, Goichberg P, Tzur G, Sabanay I, Mahalu
D, Safran S, Bershadsky A, Addadi L, Geiger B: Force and focal adhesion
assembly: a close relationship studied using elastic micropatterned
substrates. Nature Cell Biology 2001, 3:466-472.
Discher DE, Janmey P, Wang Y-l: Tissue Cells Feel and Respond to the
Stiffness of Their Substrate. Science 2005, 310:1139-1143.
Yeung T, Georges PC, Flanagan LA, Marg B, Ortiz M, Funaki M, Zahir N, Ming
W, Weaver V, Janmey PA: Effects of substrate stiffness on cell morphology,
cytoskeletal structure, and adhesion. Cell Motility and the Cytoskeleton 2005,
60:24-34.
Yamada KM, Geiger B: Molecular interactions in cell adhesion complexes.
Current Opinion in Cell Biology 1997, 9:76-85.
Solon J, Levental I, Sengupta K, Georges PC, Janmey PA: Fibroblast
Adaptation and Stiffness Matching to Soft Elastic Substrates. Biophysical
Journal 2007, 93:4453-4461.