PARTICLE INTEGRITY, SAMPLING, AND APPLICATION OF A DNA-TAGGED TRACER
FOR AEROSOL TRANSPORT STUDIES
By
Cynthia Jeanne Kaeser

A DISSERTATION
Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of
Chemistry – Doctor of Philosophy
2017

ABSTRACT
PARTICLE INTEGRITY, SAMPLING, AND APPLICATION OF A DNA-TAGGED TRACER
FOR AEROSOL TRANSPORT STUDIES
By
Cynthia Jeanne Kaeser
Aerosols are an ever-present part of our daily environment and have extensive effects on
both human and environmental health. Particles in the inhalable range (1-10 µm diameter) are of
particular concern because their deposition in the lung can lead to a variety of illnesses including
allergic reactions, viral or bacterial infections, and cancer. Understanding the transport of
inhalable aerosols across both short and long distances is necessary to predict human exposures
to aerosols. To assess the transport of hazardous aerosols, surrogate tracer particles are required
to measure their transport through occupied spaces. These tracer particles must not only possess
similar transport characteristics to those of interest but also be easily distinguished from the
background at low levels and survive the environmental conditions of the testing environment. A
previously-developed DNA-tagged particle (DNATrax), composed of food-grade sugar and a
DNA oligonucleotide as a “barcode” label, shows promise as a new aerosol tracer.
Herein, the use of DNATrax material is validated for use in both indoor and outdoor
environments. Utilizing passive samplers made of materials commonly found in indoor
environments followed by quantitative polymerase chain reaction (qPCR) assay for endpoint
particle detection, particles detection was achieved up to 90 m from the aerosolization location
and across shorter distances with high spatial resolution. The unique DNA label and PCR assay
specificity were leveraged to perform multiple simultaneous experiments. This allowed the
assessment of experimental reproducibility, a rare occurrence among aerosol field tests. To
transition to outdoor testing, the solid material provides some protection of the DNA label when

exposed to ultraviolet (UV) radiation, with 60% of the DNA remaining intact after 60 minutes
under a germicidal lamp and the rate of degradation declining with irradiation time. Additionally,
exposure of the DNATrax material using formulations of two different food-grade sugars
(maltodextrin and erythritol) to humidity as high as 66% had no significant effect on the DNA
label’s degradation or the particle’s aerodynamic diameter, confirming particle stability under
such conditions. In summary, confirmation of the DNATrax particles’ size and label integrity
under variable conditions combined with experiment multiplexing and high resolution sampling
provides a powerful experimental design for modeling aerosol transport through occupied indoor
and outdoor locations.

ACKNOWLEDGEMENTS

I would like to first thank my research advisor at MSU, Dan Jones, for his mentorship
and encouragement throughout the last 7 years. You once told me that successful scientists know
when to take risks and I am thankful for your encouragement to take those risks, especially in
moving to Lawrence Livermore National Laboratory. Thank you to Elizabeth Wheeler, who has
served as my LLNL advisor for the last 2+ years, and to George Farquar for initially recruiting
me to the DNATrax project. I must also thank my current committee members, Dana Spence and
Gary Blanchard, and my former committee members, Gavin Reid and David Kramer.
Beyond those who have advised me in an official capacity, I must also thank the current
and former Jones group members, especially Prabodha, Kristen, and Fanny, and the current and
former DNATrax project members for passing along their knowledge and perspective.
I am grateful for the unyielding support of my family and the friends that have become
family, especially my California family, Katie, the members of St. Charles, and those that have
walked with me in various capacities along the way. I am especially grateful to Thuan for the
constant pep talks, perspective, and prayer. I look forward to more life adventures with you.
I would like to acknowledge the funding sources that have supported me and my work
throughout graduate school: the MSU Chemistry Department, the LLNL Lawrence Scholar
Program, the Defense Threat Reduction Agency, and the Department of Homeland Security.
This work was performed under the auspices of the U.S. Department of Energy by Lawrence
Livermore National Laboratory under Contract DE-AC52-07NA27344. LLNL-TH-735173.
Thank you to my Lord and God for leading me through this path that is wilder than I
could have ever imagined. This has been quite the journey – now, on to the next one!

iv

TABLE OF CONTENTS

LIST OF TABLES .................................................................................................................vi
LIST OF FIGURES ...............................................................................................................ix
CHAPTER ONE: Introduction to Aerosol Transport and Tracers ........................................ix
1.1 Motivation and Scope of Work ............................................................................1
1.2 Transport Study Considerations ...........................................................................2
1.3 Development of the Aerosol Tracer DNATrax ...................................................4
1.4 Specific Aims .......................................................................................................8
REFERENCES ......................................................................................................................9
CHAPTER TWO: Multiplexed Indoor Aerosol Transport Determination Using DNABarcoded Aerosols and Passive Sampling .............................................................................13
2.1 Introduction ..........................................................................................................13
2.1.1 Indoor Air Hazards ...............................................................................13
2.1.2 Aerosol Tracer Selection.......................................................................14
2.1.3 Aerosol Sampling Techniques ..............................................................16
2.1.4 Aerosol Field Tests ...............................................................................17
2.1.5 Specific Aims ........................................................................................17
2.2 Materials and Methods .........................................................................................18
2.2.1 DNATrax Material Preparation ............................................................18
2.2.2 Passive Sampling Material Validation ..................................................19
2.2.3 Field Test Set-ups .................................................................................20
2.3 Results and Discussion ........................................................................................22
2.3.1 Passive Sampler Validation ..................................................................22
2.3.2 Single Barcode Testing .........................................................................24
2.3.3 Multiple Barcode Testing .....................................................................27
2.4 Conclusions ..........................................................................................................30
APPENDIX ............................................................................................................................31
REFERENCES ......................................................................................................................36
CHAPTER THREE: Stability of DNATrax Material as a Function of Exposure to UV
Radiation ................................................................................................................................39
3.1 Introduction ..........................................................................................................39
3.1.1 Photodegradation Mechanisms .............................................................39
3.1.2 Specific Aims ........................................................................................43
3.2 Materials and Methods .........................................................................................43
3.2.1 Material Production and DNA Analysis ...............................................43
3.2.2 UV Degradation Under Germicidal Lamp ............................................45
3.2.3 Gel Electrophoresis Analysis ................................................................46
3.3 Results and Discussion ........................................................................................47
3.3.1 DNA Degradation Determination using qPCR .....................................47
v

3.3.2 Degradation Products Analysis by Gel Electrophoresis (GE) and
qPCR ..............................................................................................................53
3.3.3 Degradation Mitigation .........................................................................58
3.4 Conclusions ..........................................................................................................59
APPENDIX ............................................................................................................................61
REFERENCES ......................................................................................................................64
CHAPTER FOUR: Stability of DNATrax Material as a Function of Exposure to
Humidity ................................................................................................................................67
4.1 Introduction ..........................................................................................................67
4.1.1 Hygroscopicity of Sweeteners ..............................................................67
4.1.2 Particle Size and Aerosol Transport .....................................................69
4.1.3 DNA Degradation Mechanisms ............................................................72
4.1.4 Specific Aims ........................................................................................73
4.2 Materials and Methods .........................................................................................73
4.2.1 Solid Material Production Conditions ..................................................73
4.2.2 Particle Size Determination ..................................................................74
4.2.3 Humidity Chamber Set-up ....................................................................75
4.2.4 Humidity Chamber Incubation .............................................................76
4.3 Results and Discussion ........................................................................................77
4.3.1 Particle Production using Alternative Sweeteners ................................77
4.3.2 Effect of Humidity Exposure on Bulk Weight .....................................78
4.3.3 Effect of Humidity Exposure on Particle Size ......................................82
4.3.4 Effect of Humidity Exposure on DNA Degradation.............................86
4.4 Conclusions ..........................................................................................................89
APPENDIX ............................................................................................................................91
REFERENCES ......................................................................................................................96
CHAPTER FIVE: Conclusions .............................................................................................99

vi

LIST OF TABLES

Table 2.1 Spray drying parameters for the production of DNATrax material. .....................19
Table 2.2 Protocols for the a) master mix and b) thermal profile used with the PCR analysis.
................................................................................................................................................19
Table 2A.1 Average cycle threshold for DNA templates used during the multiple barcode testing
with each template’s master mix (1 fg DNA per reaction; n=3 reactions). Combinations in which
the fluorescence did not exceed the threshold after 40 cycles are indicated by “N/A.” .......32
Table 2A.2 Amount of DNATrax material collected on Post-It® notes and cotton t-shirt samplers
(n=1; copies DNA/cm2) along a 90-m stretch of hallway. ....................................................34
Table 2A.3 Average amount and relative standard deviation (RSD) of DNATrax material
collected on Post-It® notes (n=3; ng/cm2) along at 1-m intervals during the second hallway
testing. ....................................................................................................................................34
Table 2A.4 Average amount of DNATrax material collected on glass slides (n=3; ng/cm2) along
the cross-sections of the center line during the multiple barcode testing. .............................35
Table 3.1 Composition of each solid material used for degradation experiments. ................44
Table 3.2 Composition of each aqueous solution used for degradation experiments. ...........44
Table 3A.1 Base composition and length of the double-stranded DNA sequences used for
DNATrax labeling and length of the primer and probe sequences. .......................................62
Table 3A.2 Average relative concentrations of Barcode A DNA in solid DNATrax particles and
aqueous solutions both with and without Fluorescent Brightener 220 as a function of irradiation
time, assessed using qPCR. Standard deviations are the three replicate samples irradiated
simultaneously. ......................................................................................................................62
Table 3A.3 Average relative concentrations of Barcode B DNA in solid DNATrax particles and
aqueous solutions as a function of irradiation time, assessed using qPCR. Standard deviations are
the three replicate samples irradiated simultaneously. ..........................................................63
Table 4.1 Physical properties and chemical structure of commercially available sweeteners with
low hygroscopicity. ................................................................................................................68
Table 4.2 Incubation location for the rotation, high, low, and ambient aliquots during the
hygroscopicity study. .............................................................................................................77

vii

Table 4A.1 Changes in bulk weight relative to the starting weight for maltodextrin material.
Averages and standard deviations are for n=3 aliquots. ........................................................93
Table 4A.2 Changes in bulk weight relative to the starting weight for erythritol material.
Averages and standard deviations are for n=3 aliquots. ........................................................93
Table 4A.3 Average MMAD values for maltodextrin material. Averages and standard deviations
are for n=3 aliquots. ...............................................................................................................93
Table 4A.4 Average d90 values for maltodextrin material. Averages and standard deviations are
for n=3 aliquots. .....................................................................................................................94
Table 4A.5 Average MMAD values for erythritol material. Averages and standard deviations are
for n=3 aliquots. .....................................................................................................................94
Table 4A.6 Average d90 values for erythritol material. Averages and standard deviations are for
n=3 aliquots............................................................................................................................94
Table 4A.7 Copies of DNA per 2-µm particle of maltodextrin-based DNATrax as determined
using qPCR. Averages and standard deviations are for n=3 aliquots. ...................................95
Table 4A.8 Copies of DNA per 2-µm particle of erythritol-based DNATrax as determined using
qPCR. Averages and standard deviations are for n=3 aliquots. .............................................95

viii

LIST OF FIGURES

Figure 1.1 Schematic of the spray dryer indicating the flows of both the gas and liquid as well as
the temperature measurement points. The air flow heating and cooling is indicated by the color
of the arrows. .........................................................................................................................5
Figure 1.2 Summary of DNA replication in a polymerase chain reaction (PCR). ...............7
Figure 1.3 a) Determination of cycle threshold in qPCR for known standards (9.89*107 copies
DNA per reaction to 98.9 copies DNA per reaction) and conversion to b) a calibration curve for
the quantification of DNA in a sample (red trace). ................................................................8
Figure 2.1 PCR curves to compare assay inhibition (green circles) and DNA recovery (orange
triangles) to control curves (blue diamonds) using a) glass microscope slides and b) Post-It®
notes. Inhibition of undiluted samples (solid lines) in the Post-it® notes was mitigated using a
10x dilution of each solution (dashed lines). The fluorescence threshold is denoted by the black
line..........................................................................................................................................23
Figure 2.2 Amount of DNATrax material deposited at each sampling point along a 90-m stretch
of hallway. Collection of material used Post-It® notes and cotton t-shirt squares (n=1 each).
Numerical values for each data point can be found in Table 2A.2 ........................................25
Figure 2.3 a) Amount of DNATrax material deposited at each sampling point along a hallway.
Collection of material used Post-It® notes (n=3). The error bars indicate the standard deviation
for the average amount collected. b) Schematic of release hallway on the same scaling as the data
in a). Air flow varied from 0.2 to 0.5 m-s-1 along the length of the hallway at approximately 1.5
m above the floor. Numerical values for each data point can be found in Table 2A.3. ........26
Figure 2.4 Average amount of DNATrax material collected on glass slides (n=3) along the crosssections of the center line in a) the Mix 1 (A/B) direction and b) the Mix 2 (C/D) direction. Error
bars indicate the standard deviation for the average amount collected on the glass slides. c)
Schematic of testing location. Numerical values for each data point can be found in Table 2A.4.
................................................................................................................................................29
Figure 2A.1 Aerodynamic size distributions for DNATrax material of each label used during the
multiple barcode testing. ........................................................................................................32
Figure 2A.2 PCR fluorescence curves for control (blue), inhibition (green), and recovery
(orange) DNA solutions for the passive sampling materials a) polyester filter, b) glass slide, c) tshirt cotton, d) filter paper, e) Post-It® note and f) Post-It® note with a 10x dilution. The
threshold value is indicated by the black line. .......................................................................33
Figure 3.1 The pyrimidine base a) thymine and two products of photoinduced crosslinking of
adjacent thymine monomers: b) 6,4-photoproduct and c) cyclobutane dimer.......................40
ix

Figure 3.2 Summary of photooxidation reaction mechanisms. .............................................41
Figure 3.3 Formation of 8-oxo-7,8-dihydroguanine a) through ionization and reaction with water
or oxygen and b) through reaction with singlet oxygen. .......................................................42
Figure 3.4 Intensity factors for light emitted across UV and visible wavelengths for the
germicidal lamp used, as provided by the manufacturer. ......................................................45
Figure 3.5 Average relative concentrations of DNA Barcode A sequence in solid DNATrax
material and aqueous solutions both with and without Fluorescent Brightener 220 as a function
of irradiation time, assessed using qPCR. Error bars indicate the standard deviation of three
replicate samples irradiated simultaneously. Numerical values for this data is found in Table
3A.2........................................................................................................................................48
Figure 3.6 Ratio of the average a) natural log and b) inverse of the amplifiable DNA Barcode A
sequence in aqueous solutions by qPCR after UV exposure to the amount of DNA present with
no exposure. The error bars indicate the standard deviations of triplicate samples. The best fit
lines for the data, forced through an intercept of 1, are also shown. .....................................51
Figure 3.7 Ratio of the average a) natural log and b) inverse of the amplifiable DNA Barcode A
sequence in solid material by qPCR after UV exposure to the amount of DNA present with no
exposure. The error bars indicate the standard deviations of triplicate samples. The best fit lines
for the data, forced through an intercept of 1, are also shown. ..............................................52
Figure 3.8 GE results of aqueous DNA (Aq) and solid material with DNA labeling (High and
Low), after varying lengths of irradiation with no PCR amplification prior to GE analysis. Serial
dilutions of a standard aqueous DNA solution (Std Aq) as well as a DNA ladder (DNA Lad) are
included for reference. The image is shown both a) as acquired at the time of analysis and b)
after adjustment of the brightness scale to improve visualization of low intensity bands,
highlighted using the black box, using Adobe PhotoShop. ...................................................54
Figure 3.9 Average pixel brightness of horizontal pixels across a GE lane as a function of
migration distance for UV-degraded samples of the Barcode A-labelled solid with high DNA
labeling. All data have been background subtracted. ............................................................55
Figure 3.10 Fraction of DNA Barcode A sequence remaining quantified through qPCR as well as
through GE with image brightness analysis for a) aqueous DNA, b) solid material with a high
level of DNA labeling, and c) solid material with a low level of DNA labeling...................56
Figure 3.11 Average relative concentrations of either the Barcode A sequence (blue traces) or
Barcode B sequence (orange traces) of DNA in a) solid DNATrax particles and b) aqueous
solutions as a function of irradiation time, assessed using qPCR. Standard deviations are the
three replicate samples irradiated simultaneously. Numerical data for these graphs can be found
in Table 3A.3. ........................................................................................................................59

x

Figure 4.1 Example of a lognormal distribution of particle size plotted a) by the fraction of the
total number of sample counts at each aerodynamic diameter and b) by the fraction of the total
mass of particles represented by particles of each aerodynamic diameter. The mass median
aerodynamic diameter (MMAD) and 90th mass percentile (d90) are marked in panel b. ....72
Figure 4.2 Diagram of the aerosol generation and sizing method used. ................................75
Figure 4.3 Change in bulk weight of a) maltodextrin-based or b) erythritol-based DNATrax
material in response to c) the relative humidity in each chamber. Error bars indicate the standard
deviations of triplicate samples. Numerical values for bulk weights can be found in Tables 4A.1
and 4A.2. ................................................................................................................................80
Figure 4.4 Average aerodynamic diameter of maltodextrin-based particles assessed using a) the
MMAD and b) the d90 after incubation at high humidity, low humidity, or both on a rotating
basis according to the c) humidity profiles. Error bars indicate the standard deviation of triplicate
samples. Numerical values for bulk weights can be found in Tables 4A.3 and 4A.4. .........84
Figure 4.5 Average aerodynamic diameter of erythritol-based particles assessed using a) the
MMAD and b) the d90 after incubation at high humidity, low humidity, or both on a rotating
basis according to the c) humidity profiles. Error bars indicate the standard deviation of triplicate
samples. Numerical values for bulk weights can be found in Tables 4A.5 and 4A.6. .........85
Figure 4.6 Copies of replicable DNA per particle with a 2-µm aerodynamic diameter in
maltodextrin-based and erythritol-based DNATrax material as determined by qPCR after storage
at low humidity, high humidity, rotation between low and high humidity, or ambient conditions.
Error bars are the standard deviation of 3 samples. Significant differences between DNA levels
are indicated by the brackets (95% confidence). ...................................................................87
Figure 4.7 Relative humidity experienced by aliquots stored in high- or low-humidity chambers,
rotated between them, or stored on the benchtop under ambient condition. .........................89
Figure C.1 Particle size distributions for maltodextrin and erythritol-based DNATrax plotted as
the a) physical diameter scaled by counts, b) aerodynamic diameter scaled by counts, and c)
aerodynamic diameter scaled by mass. ..................................................................................92

xi

CHAPTER ONE: Introduction to Aerosol Transport and Tracers
Chapter 1
1.1 Motivation and Scope of Work
Aerosols are solid particulates or liquid droplets dispersed in the air and can be either
naturally-occurring (e.g. fog, pollen, salt from ocean spray) or anthropogenic (e.g. combustion
products, smoke). While some aerosols have environmental effects such as aiding cloud
nucleation or forming a “haze” that inhibits cloud formation and suppresses precipitation1,
aerosols are also associated with a range of health effects, including allergic reaction to pollens,
infection from viral or bacterial pathogens, and cancer due to combustion and smoking
byproducts2. Spending time indoors does not inherently minimize aerosol exposure, with molds
and microbial growth affecting the health of building inhabitants and leading to instances of
“sick building syndrome”3. However, the ability of aerosols to penetrate the human lung is not
inherently negative as a range of pharmaceuticals treatments from asthma medications to
sedatives are administered by aerosolization and inhalation4.
While aerosols are generated from a broad range of sources and provide a variety of
effects, the scope of this work will focus on the transport properties of aerosols in the range of
human inhalation. A particle’s transport through air and subsequent deposition in the lung
depends on a variety of factors including the particle’s size, shape, charge, density and
hygroscopicity5. However, in general, particles less than 10 µm in aerodynamic diameter can
penetrate the alveolar region of the lung with 1-5 µm particles having the highest deposition rate
in that region6. Common aerosols in this size range include some pollens and animal dander,
bacteria, cooking aerosols, and dust7.

1

1.2 Transport Study Considerations
While personal aerosol sampling can monitor aerosol exposure as it occurs, it is also valuable
to predict the particle transport that leads to exposure when a release of aerosols is expected, as
in cases of occupational exposures. Computational models provide the basis for such predictions
and have been used to predict aerosol transport in both small air volumes, such as within an
aerosol sampler8 or human airways9, to larger volumes such as cross-country distances10-12.
Models have also been used to assess particle transport in indoor rooms with varying layouts and
levels of occupation13-17. While they are extensively used, the validation of transport models is
plagued by a lack of clear definitions as to how models should be validated and the minimal
requirements for refining such models through experimental testing18.
Experimental investigation of particle transport has been performed through monitoring of
either indigenous aerosols or material aerosolized for the sole purpose of the experiment. While
investigating transport of aerosols that affect human health, it is advantageous to use a nonhazardous surrogate for the aerosol of interest, such as an aerosol tracer. Aerosol tracers can be
used to model the transport of hazardous aerosols in locations where it is not feasible to use the
hazardous aerosols. These surrogates must have transport characteristics similar to the aerosol of
interest, be easily distinguished from aerosols already present, be safe for human exposure with
little or no decontamination necessary before the space can be reoccupied, and be stable under
the environmental conditions of the experiment19. Aerosol tracers in the inhalable range have
utilized both biological-based particles (e.g. bacteria spores20,21) and man-made particles (e.g.
polymer or silica spheres22,23) with either fluorescent or DNA-based detection schemes providing
sensitivity needed to measure individual aerosol particles. While each tracer offers its own
advantages, DNA-based detection schemes, including real-time quantitative polymerase chain

2

reaction (qPCR), offer low detection limits and quantification of individual components24.
Detection limits as low as single-digit particle counts for an aerosol tracer, even in locations with
a high level of background aerosols present, is a primary advantage of using an aerosol tracer
with a unique fluorescent or DNA identifier over using indigenous aerosols.
For experimental validation of transport in indoor environments, using a test chamber is often
preferred over an occupied space, as transport studies performed outside of an enclosed test
chamber provide have less control over the characteristics of the environment14. Some of these
characteristics, such as light and humidity exposure, have the potential to alter the detectability
or transport properties of an aerosol. Light exposure can lead to photobleaching of fluorophores
or degradation of DNA in aerosol tracers, increasing the lower limits of detection. For
hygroscopic materials, fluctuations in the humidity can alter the amount of adsorbed water,
leading to alterations in particle mass and potential issues with agglomeration. Thus, to
accurately assess aerosol transport using an aerosol tracer, the stability of the aerosol tracer with
respect to these variables must be understood.
An additional, often overlooked, concern in transport studies is the reproducibility of results.
With constant variation in the environment, guaranteeing that replicate studies are under
identical conditions from the moment of aerosolization to collection requires that the studies be
performed simultaneously. Based on the ability of the qPCR assay to quantify specific DNA
sequences from a mixture, a DNA-tagged tracer material may be a suitable platform for
performing aerosol transport reproducibility studies as well.

3

1.3 Development of the Aerosol Tracer DNATrax
Leveraging the power of DNA detection schemes with the safety of food-grade materials,
an aerosol tracer was developed. This tracer, denoted DNATrax (DNA-Tagged Reagent for
Aerosol eXperiments), paired unique double-stranded DNA sequences (dsDNA) with organic,
food-safe sweeteners as the bulk material. Original production utilized the sugar glucono-deltalactone (GDL) with inkjet printing for particle production25, but an alternate production scheme
using maltodextrin, a sugar with a higher glass transition temperature than GDL, and a
commercially available spray dryer resulted in the gram-scale production necessary for aerosol
releases in larger spaces26. The spray drying process, summarized in Figure 1.125, begins by
spraying an aqueous solution as fine droplets into a heated airstream. The droplets rapidly dry in
the airstream, condensing the materials dissolved in the aqueous solution into a solid particle.
The solid particles are collected by impaction on the walls of the collection cyclone and drop into
the collection chamber.

4

Figure 1.1 Schematic of the spray dryer indicating the flows of both the gas and liquid as well as
the temperature measurement points. The air flow heating and cooling is indicated by the color
of the arrows.
After production, detection and quantification of the DNA label in the powder is
performed using qPCR, a process summarized in Figure 1.2. In qPCR, the double-stranded DNA
oligonucleotides are denatured and complementary strands synthesized. This theoretically
doubles the copies of dsDNA with each thermal cycle, with a single parent dsDNA strand
yielding 1.1 billion copies of DNA after 30 cycles through the process. In addition to the use of
sequence-specific primers that allow amplification of a single dsDNA sequence even when
multiple sequences are present, the addition of a sequence-specific fluorescent probe provides
the basis for DNA quantification. The probe is a short (~20 bp) oligonucleotide sequence that is
complementary to one of the dsDNA strands with a fluorophore (6-FAM; fluorescein) at the 5′5

end and a quencher (Black Hole Quencher®-1) at the 3′-end. During the strand synthesis step,
the fluorescent probe is cleaved from the oligonucleotide. With increased distance between the
fluorophore and the quencher, the fluorophore’s fluorescence can be detected. The intensity of
fluorescence within the solution is directly proportional to the number of dsDNA strands
produced, doubling with each cycle. For each fluorescence curve, a threshold cycle (Ct) is
defined as the PCR cycle at which the fluorescence intensity passes a set threshold, for known
DNA concentrations (Figure 1.3a). There is a linear relationship between the cycle threshold and
number of DNA copies at the start of the PCR cycling (Figure 1.3b). This allows quantification
of DNA in a solution of unknown concentration using a calibration curve of solutions of known
DNA concentrations. Thus, for quantification of particle transport in a field test, the ratio of
DNA content per mass of DNATrax material assessed prior to an experiment is compared with
the DNA content collected at various locations across the test space.
The DNATrax material posits several advantages as an aerosol tracer. First, its synthetic
DNA tag and food-grade bulk material provide a low burden of approval for testing in occupied
spaces. Also, by using unique DNA sequences, each batch of material produced can have a
different signature that is easily separated from both background aerosols and DNATrax particles
containing other sequences. This provides a platform for multiplexed experiments, allowing
multiple variables to be investigated simultaneously as well as assessing the reproducibility of an
experiment’s results. However, to serve as a tracer for outdoor environments, the bulk material
and DNA oligonucleotide must be stable under the variable humidity and light exposures for the
duration of the experiment.

6

Figure 1.2 Summary of DNA replication in a polymerase chain reaction (PCR).

7

Figure 1.3 a) Determination of cycle threshold in qPCR for known standards (9.89*107 copies
DNA per reaction to 98.9 copies DNA per reaction) and conversion to b) a calibration curve for
the quantification of DNA in a sample (red trace).

1.4 Specific Aims
This work challenges the proposed advantages of using DNATrax material as an aerosol
tracer by examining its transport and label stability under a range of indoor and outdoor
environments. Chapter 2 details its use indoors, showing the low limits of detection enable
detection with passive samplers as far as 90 meters from the point of aerosolization. Using
multiple materials each with a unique barcode, the reproducibility of indoor experiments is also
assessed. Chapters 3 and 4 discuss the effects of exposure to UV radiation and variable humidity,
respectively, on both the DNA tag and the particle size, as these characteristics dictate the
particle detection and transport properties.

8

REFERENCES

9

REFERENCES

1. Tao W-K, Chen J-P, Li Z, Wang C, Zhang C. Impact of aerosols on convective clouds and
precipitation. Rev Geophys. 2012. 50: RG2001.
2. Douwes J, Thorne P, Pearce N, Heederik D. Bioaerosol health effects and exposure
assessment: progress and prospects. Ann Occup Hyg. 2003. 47: 187-200.
3. Walinder R, Norback D, Wessen B, Venge P. Nasal lavage biomarkers: effects of water
damage and microbial growth in an office building. Arch Environ Health. 2001. 56: 30–
36.
4. Stein SW, Thiel CG. The history of therapeutic aerosols: a chronological review. J Aerosol
Med Pulm Drug Deliv. 2017. 30: 20-41.
5. Schlesinger RB. Comparative deposition of inhaled aerosols in experimental animals and
humans: a review. J Tox and Environ Health. 1985. 15: 197-214.
6. Byron PR. Prediction of drug residence times in regions of the human respiratory tract
following aerosol inhalation. J Pharm Sci. 1986. 75: 433–438.
7. Brain JD, Valberg PA. Deposition of aerosol in the respiratory tract. Am Rev Respir Dis. 1979.
120: 1325-1373.
8. Sajjadi H, Tavakoli B, Ahmadi G, Dhaniyala S, Harner T, Holsen TM. Computational fluid
dynamics (CFD) simulation of a newly designed passive particle sampler. Environ Pollut.
2016. 214: 410-418.
9. Oldham MJ. Computational fluid dynamic predictions and experimental results for particle
deposition in an airway model. Aerosol Sci Technol. 2000. 32: 61-71.
10. Owega S, Khan BUZ, Evans GJ, Jervis RE, Fila M. Identification of long-range aerosol
transport patterns to Toronto via classification of back trajectories by cluster analysis and
neural network techniques. Chemometr Intell Lab. 2006. 83: 26-33.
11. Chin M, Diehl T, Ginoux P, Malm W. Intercontinental transport of pollution and dust
aerosols: implications for regional air quality. Atmos Chem Phys. 2007. 7: 5501-5517.
12. Wong MS, Nichol JE, Lee KH. Estimation of aerosol sources and aerosol transport pathways
using AERONET clustering and backward trajectories: a case study of Hong Kong. Int J
Remote Sens. 2013. 34: 938-955.

10

13. Tao Y, Inthavong K, Petersen P, Mohanarangam K, Yang W, Tu J. Experimental and CFD
modelling of indoor air and wake flow from a moving manikin. 20th Australasian Fluid
Mechanics Conference. 2016.
14. Sajjadi H, Salmanzadeh M, Ahmadi G, Jafari S. Simulations of indoor airflow and particle
dispersion and deposition by the lattice Boltzmann method using LES and RANS
approaches. Build Environ. 2016. 102: 1-12.
15. Lu W, Howarth AT, Adams NM, Riffat SB. CFD modeling and measurement of aerosol
particle distributions in ventilated multizone rooms. ASHRAE Trans. 1999. 105, part 2:
116-127.
16. Sohn MD, Apte MG, Sextro RG, Lai ACK. Predicting size-resolved particle behavior in
multizone buildings. Atmos Environ. 2007. 41: 1473-1482.
17. Richmond-Bryant J, Eisner AD, Bixey LA, Wiener RW. Transport of airborne particles
within a room. Indoor Air. 2006. 16: 48-55.
18. Oldham MJ. Challenges in validating CFD-derived inhaled aerosol deposition predictions.
Inhal Toxicol. 2006. 18: 781-786.
19. Sinclair RG, Rose JB, Hashsham SA, Gerba CP, Haas CN. Criteria for selection of
surrogates used to study the fate and control of pathogens in the environment. Appl
Environ Microbiol. 2012. 78: 1969-1977.
20. Emanuel PA, Buckley PE, Sutton TA, Edmonds JM, Bailey AM, Rivers BA, Kim MH,
Ginley WJ, Keiser CC, Doherty RW, Kragl FJ, Narayanan FE, Katoski SE, Paikoff S,
Leppert SP, Strawbridge JB, VanReenen DR, Biberos SS, Moore D, Phillips DW,
Mingioni LR, Melles O, Ondercin DG, Hirsh B, Bieschke KM, Harris CL, Omberg KM,
Rastogi VK, Cuyk SV, Gibbons HS. Detection and tracking of a novel genetically tagged
biological simulant in the environment. Appl Environ Microb. 2012. 78: 8281-8288.
21. King MF, Noakes CJ, Sleigh PA, Camargo-Valero MA. Bioaerosol deposition in single and
two-bed hospital rooms: a numerical and experimental study. Build Environ. 2013. 59:
436-447.
22. Liljegren JC, Brown DF, Lunden MM, Silcott D. Particle deposition onto people in a transit
venue. Health Security. 2016. 14: 237-249.
23. Lai ACK, Byrne MA, Goddard AJH. Experimental studies of the effect of rough surfaces and
air speed on aerosol deposition in a test chamber. Aerosol Sci Technol. 2002. 36: 973982.
24. Lopez-Andreo M, Lugo L, Garrido-Petierra A, Preito MI, Puyet A. Identification and
quantitation of species in complex DNA mixtures by real-time polymerase chain reaction.
Anal Biochem. 2005. 339: 73-82.
11

25. Udey RN, Jones AD, Farquar G R. Aerosol and microparticle generation using a commercial
inkjet printer. Aerosol Sci Technol. 2013. 47: 361-372.
26. Harding RN, Hara CA, Hall SB, Vitalis EA, Thomas CB, Jones AD, Day JA, Tur-Rojas VR,
Jorgensen T, Herchert E, Yoder R, Wheeler EK, Farquar GR. Unique DNA-barcoded
aerosol test particles for studying aerosol transport. Aerosol Sci Technol. 2016. 50: 429435.

12

CHAPTER TWO: Multiplexed Indoor Aerosol Transport Determination Using DNA-Barcoded
Aerosols and Passive Sampling

2.1 Introduction
Accurate assessment and prediction of indoor health hazards posed by exposure to
aerosols in the respirable range (1-10 μm) requires an understanding of particle transport in
occupied spaces. Occupied spaces present several constraints when conducting aerosol releases
including, 1) concerns for the safety of the occupants, 2) maintaining the integrity of the test
space’s HVAC air flow, and 3) the necessity for experimental replicates that can be
distinguished. DNA Tagged Reagents for Aerosol eXperiments (DNATrax), a safe simulant with
tunable DNA barcodes, provides a platform to overcome these constraints. Coupled with passive
sampling, multiple DNATrax materials were used simultaneously to assess both the
reproducibility of aerosol transport and the effect of aerosol release direction on the dispersion
pattern in an occupied space.

2.1.1 Indoor Air Hazards
Indoor air pollutants such as mold spores, allergens, and bacteria can cause negative
immediate and long-term health effects1. Of most concern are those pollutants with sizes in the
respirable range (1-10 μm aerodynamic diameter), as these are capable of entering the lungs and
are impacted onto the airway surfaces without being exhaled2. Heating, ventilation and air
conditioning (HVAC) systems filter out such airborne hazards while also providing air
circulation and indoor/outdoor air mixing to maintain indoor air quality. To improve indoor
ventilation systems, the manner in which these systems affect aerosol transport through indoor

13

spaces must be better understood. The observation and prediction of aerosol particle dispersion
and deposition patterns in occupied areas is limited by complex and dynamic airflows that
provide a changing set of underlying variables which influence particle motion3. The presence of
furniture and people add additional complexities to the air movement. While test chambers offer
control of these variables, such experiments are not always feasible. Passive sampling of an
aerosol simulant released in a test space bridges the gap between the laboratory and the field as it
allows testing in occupied locations while introducing as few extra variables as possible. Under
uncontrolled airflow conditions, the challenge of assessing the reliability and reproducibility of
dispersion is overcome by using multiple samplers at each test site and simultaneous
performance of multiple experiments, e.g. with different particle sizes or types or releases from
multiple locations.

2.1.2 Aerosol Tracer Selection
Aerosol tracers are used in chamber and field aerosol dispersion studies to monitor and
predict the transport of aerosols. An ideal tracer would have similar size and transport properties
to the aerosol of interest, be easily distinguished from aerosols already present, and be safe for
human exposure with little or no decontamination necessary before the space can be reoccupied4.
Additionally, for multiple simultaneous experiments to be successful, the tracers released must
also be distinguishable from each other without having altered transport properties.
Several aerosol tracers have been used in previous studies. The tracer gas sulfur
hexafluoride5 is nontoxic and can be accurately measured at low concentrations. However, these
molecules are smaller in size than the inhalable range particles and will be more affected by
Brownian motion. Thus, it is useful for modelling gas transport, but does not provide an accurate

14

assessment of particle transport. Bacterial spores6-7 are of a more appropriate size, falling within
the 1-10 µm range, but fluorescence detection may be unable to distinguish these spores from
biological aerosols already present and lack the ability to tune the fluorescence spectrum so
multiple experiments can be simultaneously performed. DNA-based detection schemes such as
real-time quantitative polymerase chain reaction (qPCR) offer an alternative to fluorescence that
allows quantification of trace levels of different DNA markers from a mixture of collected
particles8. Bacteria with tunable DNA barcodes have been reported6 but still suffer from a
negative public perception of their safety in occupied spaces. Leveraging the advantages of
qPCR detection without the connotations of bacteria, the aerosol simulant material DNATrax
(DNA Tagged Reagents for Aerosol eXperiments)9 provides an alternative aerosol tracer.
DNATrax are sugar-based particles tagged with unique and tunable DNA barcodes. These are
spherical particles with aerodynamic diameters within the inhalable range so their transport is
representative for the indoor health hazards listed above. Each batch of DNATrax is labelled
with a batch-specific DNA sequence as a barcode for detection by a quantitative real-time
polymerase chain reaction (qPCR) assay. By using multiple DNATrax batches, each with a
differing DNA barcode, DNATrax material released during multiple simultaneous experiments
can be collected on the same samplers. The amount of each DNA label collected per sampler is
determined at the end of the experiment by qPCR and is not influenced by other DNA sequences
present. Utilizing several DNATrax materials each with a differing barcode allows simultaneous
releases to be performed and the initial release conditions of each detected particle to be known.

15

2.1.3 Aerosol Sampling Techniques
In addition to selection of the aerosol tracer, an appropriate sampling technique must also
be determined. Aerosol sampling techniques can be classified as either active samplers, which
employ vacuum systems to draw air into the sampler, or passive samplers, which rely on
deposition through gravitational settling or impaction. With their ability to condense aerosols
from large volumes of air into a single measurement, active samplers are most useful when the
tracer is at a low concentration10. However, because they draw in large volumes of air, these
samplers alter the airflow of the test space and are impractical when high spatial resolution is
desired. Passive sampling techniques have much lower sampling rates, relying on the direct
deposition of particles onto the sampler for detection, but preserve the natural air movements in a
test space. These samplers are usually small (<100 cm2) and can provide high spatial resolution
in when multiple samplers are placed across the test space.
Several aspects should be considered when selecting a passive sampling material. To
preserve the authenticity of the sample space, the material selected should be similar to those
already found in the testing area, inspiring the use of clean floor tiles6 and polyethylene films11
as passive samplers. However, compatibility with the planned detection scheme is vital for an
experiment, leading to the use of materials such as glass microscope slides12, SEM stubs13-14, and
Petri dishes with growth media7. Other desirable qualities in a sampler include durability,
portability, and low cost per sampler. Driven by these desirable traits, several potential passive
sampling materials were examined to expand the suite of suitable materials to include small,
easily manipulated samplers made from materials commonly found in indoor spaces including
various cloth and paper textures.

16

2.1.4 Aerosol Field Tests
Computational models of aerosol transport allow prediction of particle dispersion under a
variety of circumstances, but requires validation using experimental data15. Several previous
studies have sought to provide such datasets using a variety of experimental designs including
empty rooms with moving mannequins16, empty rooms with subdivided areas17-19, and furnished
rooms20. However, these transport experiments are typically performed as single experiments of
the aerosolization of one tracer under one set of circumstances without any assessment of the
experiment reproducibility. For model validation, it would be advantageous to have an
assessment of the reproducibility of the experimental set-up from tracer release to collection and
quantification to include a level of confidence in the experimental data used. In order to assess
the reproducibility of the experimental design, all of the other variables must be held constant, a
near impossibility for sequential experiments performed outside of a controlled test chamber.
Thus, simultaneous experiments must be performed. With the qPCR assay able to quantitate
individual DNA sequences even in mixed samples, multiple DNATrax tracers could be used
simultaneously to assess experiment reproducibility. The use of multiple tracers could also be
leveraged for multiplexed experiments, distinguishing the effect of multiple release variables
(e.g. location and direction of release, aerosolization duration, particle size, etc.) under identical
environmental conditions.

2.1.5 Specific Aims
The main goal of this work is to validate that the DNATrax aerosol tracer is a suitable
tracer for use across long distances (>50 m), over short distances (10 m) with strong airflow, and
in cases where simultaneous testing of multiple variables or high spatial resolution sampling is

17

desired. This validation includes the production and use of DNATrax with differing unique DNA
labels, confirmation of the compatibility of prospective passive sampling materials with the
qPCR detection scheme, and 3 separate field studies.

2.2 Materials and Methods
2.2.1 DNATrax Material Preparation
Solid DNATrax material was produced via spray-drying using a Mini Spray Dryer B-290
(Büchi; Switzerland), using the parameters listed in Table 2.19,21. The carrier material was
organic food-grade tapioca maltodextrin (DE 10; Ciranda; Thailand) with short synthetic
oligonucleotides serving as the DNA label, each approximately 100bp in length. While the
standard solutions of each DNA sequence used for calibration curves were purchased (Biosearch
Technologies; Petaluma, CA), the stock solutions of DNA used for solid production were from
standard DNA solutions previously amplified using a polymerase chain reaction (PCR) assay
(Invitrogen; Carlsbad, CA) with sequence-specific primers (Integrated DNA Technologies;
Coralville, IA). The components in the qPCR master mix and the thermal profile used for
amplification are listed in Table 2.2a and 2.2b, respectively. These qPCR parameters were also
used for quantification of DNA concentrations in the solid material and to determine the
specificity of the PCR assay with respect to each DNA sequence using standard PCR crossreactivity analyses (Table 2A.1).

18

Aspect
Value
Aspirator
475 L/min (70%)
Liquid Flow Rate
9 mL/min (30%)
Nitrogen Flow Rate
13.85 L/min (45mm)
Inlet Temperature
190 °C
Initial Outlet Temperature
110 °C
Solids Concentration (Aq)
3.0% (w/v%)
Table 2.1 Spray drying parameters for the production of DNATrax material.

a) Master Mix
b) Thermal Profile
Reagent
µL
Step Temperature (˚C)
Time
PCR water
14.4
1
95
2 minutes
10x Reaction buffer
2.5
2
95
30 sec
50 mM MgCl2
1.5
3*
55
30 sec
dNTP @ 10 µM
0.5
4
72
30 sec
F/R Primers @ 10 µM
0.5
5
go to step 2 39x (40 cycles total)
Probe @ 10 µM
0.5
*Optic collection on step 3
Platinum Taq
0.1
Sample
5.0
Table 2.2 Protocols for the a) master mix and b) thermal profile used with the PCR analysis.

The size distribution of the resulting particles was assessed by aerosolizing material into
a small (1L) test chamber and sampling using an aerodynamic particle sizer (APS) (TSI;
Shoreview, MN). The size distributions of all batches were observed to be log normal with
aerodynamic diameters ranging from 0.5 to 5 μm with maxima centered between 1 and 2 μm
(Figure 2A.1).

2.2.2 Passive Sampling Material Validation
The passive sampling materials investigated in this study represent a variety of materials
and textures commonly found in indoor environments, including

19

•

Polyester filters used in dry filter units (Lockheed Martin; Bethesda, MD; round with
4.75 cm diameter),

•

Glass microscope slides (Gold Seal; Portsmouth, NH; 2.5 cm by 7.5 cm),

•

White cotton t-shirts (Hanes; Winston Salem, NC; cut to 5.1 cm by 5.1 cm),

•

Yellow Post-it® notes (3M; Santa Clara, CA; 5.1 cm by 5.1 cm), and

•

Grade 1 filter paper (Whatman; Buckinghamshire, UK; round with 11.5 cm diameter).
The extent of matrix inhibition of the PCR assay was determined by extraction of the

sampler into PBS (Amresco; Solon, OH) containing 0.1% Triton-X (Acros Organics; Geel,
Belgium) and 0.2 pg DNA/μL solution. For barcode recovery efficiency, 10 µL of a 0.2
pg/µL solution of aqueous DNA was added directly to each material and allowed to dry
before extraction into the buffer and subsequent quantification. All samples and controls
were measured in triplicate using the qPCR assay. Inhibition or incomplete recovery was
defined as sampler results having a statistically significant shift in the average response curve
to more positive cycle threshold (Ct) values relative to the positive control.

2.2.3 Field Test Set-ups
Three field test locations were selected; all indoor occupied spaces controlled by HVAC
systems. The HVAC systems were not adjusted before or during the tests to provide an accurate
assessment of particle transport under normal operating procedures. All areas were cleared of
people apart from those performing the tests, with those present remaining outside of the
sampling area from the time of aerosol release until the time of sampler collection.
The first location was a long (>100 m), straight interior hallway. All doors adjacent to the
hallway were closed and there were several air vents spaced along the length of the hallway.

20

DNATrax was aerosolized using an eductor from one end of the hallway in several aliquots
totaling 6.85 g of material. The material was passively sampled at 7 locations along the hallway,
using three Post-it® notes and three cotton t-shirt squares at each location, set out for the
duration of the experiment.
The second location was a shorter hallway measuring 16.0 m long by 1.8 m wide with an
overhead air vent at one end and a door opening into an adjacent part of the building at the other.
All other doors into the hallway were closed. The DNATrax simulant was aerosolized using an
eductor at 2.5 m into the hallway from the open door and approximately 1.5 m above the floor.
Aerosolization took place over a period of 10 seconds, dispersing approximately 300 milligrams
of material in the direction of the air vent. The hallway air was passively sampled at 1-meter
intervals along the hallway for 45 minutes using 3 Post-It® notes at each sampling location. The
samplers were then collected and the amount of DNATrax on each sampler was determined
using the same procedure for extraction and quantitation used during sampler validation. Note
that while the HVAC system was running for the duration of the experiment and provided a
definite directional air flow from the vent toward the point of release, the air flow was not
specifically controlled during the experiment.
The third location was a larger room (5.9 m by 9.5 m) with one overhead intake air vent
and three doors, all remaining closed for the duration of the test. Two batches of barcoded
DNATrax material (barcodes A and B) were mixed together and then 110 milligrams of the
mixed material was aerosolized from approximately 1.5 m off the ground in the center of the
room facing one direction. Fifteen minutes later, 110 milligrams of a second mix of barcodes (C
and D) was aerosolized from the same location facing the opposing direction for a total of 220
milligrams of barcoded material aerosolized over the course of the experiment. Each of the

21

mixed batches was fully aerosolized in under 10 seconds using in-house built eductors with
portable carbon dioxide cartridges as a compressed gas source. Sampling was performed at 1meter intervals along lines perpendicular to the aerosolizing direction with the nearest location
being 2.5 m away from the release point in each direction. Three glass slides were used at each
location and all samplers were collected 50 minutes after the second release. Similar to the
hallway tests, the HVAC system for this room was left on and not specifically controlled.

2.3 Results and Discussion
2.3.1 Passive Sampler Validation
The compatibility of the samplers with PCR was determined by assessing assay inhibition
and DNA recovery. As the inhibition, recovery and control samples all have the same
concentration of DNA in theory, any significant shift of the cycle threshold for an inhibition or
recovery test sample to higher values than the control would indicate inhibition of the assay or
incomplete recovery. Figure 2.1a shows the PCR curves for glass slide testing as an example of a
sampler that did not inhibit the PCR assay and allowed for complete recovery of DNA (i.e.
complete dissolution of the DNATrax material). PCR curves for additional passive sampling
materials that passed inhibition and recovery tests (filter paper, polyester filters, and cotton tshirt fabric) can be found in Figure 2A.2.

22

b) 1000

800

800

Fluorescence
Intensity (a.u.)

1000

Fluorescence
Intensity (a.u.)

a)

600
400

600
400

200

200

0

0

20

25
30
Cycle number

20

35

25
30
Cycle number

35

Figure 2.1 PCR curves to compare assay inhibition (green circles) and DNA recovery (orange
triangles) to control curves (blue diamonds) using a) glass microscope slides and b) Post-It®
notes. Inhibition of undiluted samples (solid lines) in the Post-it® notes was mitigated using a
10x dilution of each solution (dashed lines). The fluorescence threshold is denoted by the black
line.

Initial measurements of DNA deposited on Post-It® notes showed no inhibition and
complete recovery, but testing with additional packages of the notes yielded evidence of
complete inhibition of the PCR assay (Figure 2.1b). However, using 1:10 dilutions of the sample
extracts mitigated this effect by diluting the inhibiting compounds and the resulting DNA
quantification fell within the assay’s margin of error, defined as the standard deviation of three
technical replicates. Using this knowledge, the first two field tests were successfully performed
using Post-It® notes as passive samplers using the 1:10 dilution before qPCR analysis without
sample inhibition. However, while the portability, durability and knowledge that they will not
shift position during an experiment are major advantages to using Post-It® notes, the pack-to-

23

pack variation in inhibition described above is troubling for sampler reproducibility in the longterm. Thus, glass slides were used for the third field test.

2.3.2 Single Barcode Testing
In the first hallway test, particles were successfully detected along the length of the
hallway, as far as 90 m from the release location (Figure 2.2) using both Post-It® notes and
cotton t-shirt squares. The two sampler types showed similar results at each location along the
hallway, with the ratio of copies of DNA between the two sampler types (Post-It® note/T-shirt)
ranging from 0.82-1.02. The consistency in amount collected means the two samplers indicates
that the sampler composition does not affect its sampling efficiency and these measurements can
be considered replicates in sampling for each location. While the sharpest decrease in amount of
material collected was from 3 m to 9 m (120 million copies DNA per cm2 to 4.7 million copies
DNA per cm2) from the release point, there was a gradual decrease of more than two orders of
magnitude in DNA levels across the 90 meters. As a point of reference, at 10 copies of DNA per
2-µm particle, the 120 million copies of DNA per cm2 at the 3-m location would equate to
approximately 50 µg of solid material collected per cm2, while the 4.7 million copies of DNA
per cm2 is approximately 2.0 µg solid material per cm2. The successful collection and detection
of DNATrax material as far as 90 m from the release point demonstrates DNATrax’s usefulness
for passively monitoring aerosol transport across large areas. Given that the samplers set out 90
m from the release point still collected a substantial amount (360,000 copies DNA per cm2; 0.15
µg solid material per cm2) of DNATrax material, small particles with longer residence time
travelled further, which indicates potential for studies involving even greater distances.

24

DNA Collected on Sampler
(Copies DNA/cm2)

1.E+09
1.E+08
1.E+07
1.E+06
1.E+05
0
Post-It Note
T-Shirt Square

20

40
60
Distance From Release (m)

80

100

Figure 2.2 Amount of DNATrax material deposited at each sampling point along a 90-m stretch
of hallway. Collection of material used Post-It® notes and cotton t-shirt squares (n=3 each; error
bars represent standard deviation). Numerical values for each data point can be found in Table
2A.2.

25

Figure 2.3 a) Amount of DNATrax material deposited at each sampling point along a hallway.
Collection of material used Post-It® notes (n=3). The error bars indicate the standard deviation
for the average amount collected. b) Schematic of release hallway on the same scaling as the data
in a). Air flow varied from 0.2 to 0.5 m-s-1 along the length of the hallway at approximately 1.5
m above the floor. Numerical values for each data point can be found in Table 2A.3.

While the first field test demonstrated the distance across which studies can be
performed, the second hallway field test aimed to demonstrate sampling with improved spatial
resolution. Thus, this test used a much shorter hallway, less released material, and sampling with
1-meter spatial resolution. The amount of solid DNATrax material deposited on samplers along
the hallway (in nanograms) spanned nearly 2 orders of magnitude per square centimeter sampled
(Figure 2.3a), a decrease similar to that seen across the first 10 meters of the longer hallway
study. The steep decrease in amount collected between 2 and 4 m from the release location and
26

subsequent leveling off to a constant level is consistent with visual observations during the
experiment. After aerosolization, a visible cloud of material was observed moving in the
direction of release (against the airflow), indicating that the initial aerosol velocity was higher
than the HVAC air movement in the hallway. This allowed the particles to travel a short distance
before particles decelerated to the ambient air flow conditions defined by the HVAC system and
stopped moving down the hallway. This hypothesis is supported by the sharp particle decrease at
~4 m in Figure 2.3a.
Having acquired multiple samples at each location, statistical analysis was used to divide
the hallway into 3 zones of distinct exposure amounts. Using a student’s t-test with a 95%
confidence level, the first two locations away from the release point were found to be statistically
similar to each other, but different from all other locations. The location 3 m from the release
point was also statistically different from locations 4-10, which were indistinguishable from one
another. Thus, the hallway might be divided into areas of high exposure (up to 2 meters from
release, ~500-600 ng-cm-2 DNATrax), moderate exposure (2 to 4 meters from release, ~70 ngcm-2 DNATrax) and low exposure (4 or more meters from release, <15 ng-cm-2 DNATrax).
Thus, when considering the exposure of a person to an indoor health hazard, the person’s
proximity to the point of aerosolization is of great importance.

2.3.3 Multiple Barcode Testing
Passive sampling with high spatial resolution is a powerful tool for observing differences
in particle deposition across a test space and the use of replicate samplers at each location adds
confidence to the results. To further leverage these advantages when air flow fluctuations within
the test space are likely, multiple experiments were performed simultaneously. Identical release

27

conditions were accomplished by mixing particles of two different DNATrax barcodes and
dispersing the mixed material. In doing this, particles of both barcodes in a “mix” are exposed to
the same conditions through the duration of the test. Two sets of mixed barcodes (Mix 1 and Mix
2) were released so particles of 4 distinct barcode sequences were present (barcodes A, B, C, and
D). In both mixed barcode releases (Figure 2.4), the ratio between levels of the two barcodes
remained constant across all locations before and after the release with a Mix 1 (A/B) ratio of
1.49 ± 0.10 and a Mix 2 (C/D) ratio of 1.45 ± 0.16, by weight. This consistency between
barcodes indicates the particle deposition observed at different locations is independent of
barcode and highlights the experimental reproducibility inherent with DNATrax.

28

Figure 2.4 Average amount of DNATrax material collected on glass slides (n=3) along the crosssections of the center line in a) the Mix 1 (A/B) direction and b) the Mix 2 (C/D) direction. Error
bars indicate the standard deviation for the average amount collected on the glass slides. c)
Schematic of testing location. Numerical values for each data point can be found in Table 2A.4.

The shape and relative relationships between of the deposition curves shown in Figure
2.4a represent the particle distribution expected for a room with no natural airflow. Barcodes A
and B were released in the direction of these samplers, so there were more of Barcodes A and B
collected than Barcodes C and D, which were released in the opposite direction. All four
barcoded materials have nearly symmetrical distributions across the center line, also expected if
there is no influence from the air flow. The Mix 2 direction (Figure 2.4b) has an asymmetry in

29

particle collection that was observed for all barcodes. While the exact source of the asymmetry is
unclear, its presence across all barcodes adds confidence to the conclusion that this result
accurately reflects the particle deposition in the room under normal, occupied conditions. The
ability to simultaneously perform experiments in an occupied space with reproducible results is
an advantageous characteristic of DNATrax that is clearly demonstrated within this experiment.

2.4 Conclusions
Accurate determination of the distribution of aerosol particles in an occupied space
without disruption of the natural airflow in the room was accomplished using passive samplers in
conjunction with DNATrax particles. Several passive sampling materials were compatible with
the DNATrax simulant and subsequent PCR detection scheme. Using passive sampling,
detection of DNATrax was achieved at both long distances up to 90 m from the release point and
at shorter distances with 1-meter sampling resolution. Sampling reproducibility at the high
resolution gave additional confidence in measurements of particle deposition at each location.
Mixing together DNATrax powders with differing DNA barcodes before release provided two
datasets experiencing identical experiment conditions from release through collection and
detection. Using an additional pair of DNA barcodes, the effect of changing the initial direction
of aerosol release was investigated. The combination of simultaneous DNATrax releases with
passive sampling eliminates many of the underlying variables in aerosol transport studies
including aerosolization efficiency, air flow fluctuations, sampling reproducibility and the impact
of occupants in the test facility. By using simultaneous safe barcoded particles for aerosol test
particles the environmental variability from experiment to experiment is eliminated.

30

APPENDIX

31

APPENDIX

Master Mix
Barcode A (359) Barcode B (771) Barcode C (431) Barcode D (610)
Barcode A (359)
27.34±0.23
N/A
N/A
N/A
Barcode B (771)
N/A
28.97±0.07
N/A
N/A
Barcode C (431)
N/A
N/A
28.47±0.12
N/A
Barcode D (610)
N/A
N/A
N/A
29.01±0.14
Table 2A.1 Average cycle threshold for DNA templates used during the multiple barcode testing
DNA Template

with each template’s master mix (1 fg DNA per reaction; n=3 reactions). Combinations in which

Number of Particles in Sample by
Counts, Normalized to Maximum

the fluorescence did not exceed the threshold after 40 cycles are indicated by “N/A.”

1.20
1.00

Barcode A
0.80
Barcode B
0.60
Barcode C
0.40
Barcode D
0.20
0.00
0

5

10
Aerodynamic Diameter (µm)

15

20

Figure 2A.1 Aerodynamic size distributions for DNATrax material of each label used during the
multiple barcode testing.

32

1400

Inhibition
Recovery
Control
Threshold

Fluorescence
Intensity (a.u.)

1200
1000
600
400

600
400

0

0
30
35
Cycle number

40
d)

25

30
35
Cycle number

40

25

30
35
Cycle number

40

1400
1200

Fluorescence
Intensity (a.u.)

Fluorescence
Intensity (a.u.)

800

200

1200
1000
800
600
400

1000
800
600
400

200

200

0

0
25

30
35
Cycle number

40

1400

f)

1400
1200

Fluorescence
Intensity (a.u.)

1200

Fluorescence
Intensity (a.u.)

1000

200

1400

e)

1400

1200

800

25
c)

b)

Fluorescence
Intensity (a.u.)

a)

1000
800
600
400

1000
800
600
400

200

200

0

0
25

30
35
40
25
30
35
40
Cycle number
Cycle number
Figure 2A.2 PCR fluorescence curves for control (blue), inhibition (green), and recovery
(orange) DNA solutions for the passive sampling materials a) polyester filter, b) glass slide, c) tshirt cotton, d) filter paper, e) Post-It® note and f) Post-It® note with a 10x dilution. The
threshold value is indicated by the black line.
33

Distance
Amount Collected
Amount Collected
Ratio of Amount
from Release
on Post-It® Note
on Cotton T-shirt
Collected, (Post(m)
(copies DNA/cm2)
(copies DNA/cm2)
It®/T-shirt)
3
1.19 ± 0.16 E+08
1.32 ± 0.41 E+08
0.90
9
4.68 ± 1.62 E+06
5.63 ± 0.68 E+06
0.83
27
3.10 ± 0.32 E+06
3.74 ± 0.77 E+06
0.83
0.82
60
1.61 ± 0.47 E+06
1.97 ± 0.11 E+06
63
2.71 ± 0.21 E+06
2.87 ± 0.31 E+06
0.94
72
2.28 ± 0.13 E+06
2.23 ± 0.14 E+06
1.02
90
3.66 ± 0.27 E+05
3.60 ± 0.21 E+05
1.02
Table 2A.2 Amount of DNATrax material collected on Post-It® notes and cotton t-shirt samplers
(n=3; copies DNA/cm2) along a 90-m stretch of hallway.

Distance from
Average amount
Release (m)
collected (ng DNA/cm2)
RSD
1
604.48
44.6%
2
496.52
12.4%
3
72.81
44.6%
4
13.63
26.9%
5
11.02
27.2%
6
11.20
36.7%
7
10.31
22.7%
8
9.98
33.7%
9
8.11
16.9%
10
14.26
40.8%
Table 2A.3 Average amount and relative standard deviation (RSD) of DNATrax material
collected on Post-It® notes (n=3; ng/cm2) along at 1-m intervals during the second hallway
testing.

34

Barcode A
Barcode B
Barcode C
Barcode D
Distance
(Mix 1)
(Mix 1)
(Mix 2)
(Mix 2)
from
Release
St.
St.
St.
St.
Avg.
Avg.
Avg.
Avg.
Line (m)
Dev.
Dev.
Dev.
Dev.
-2
90.6
11.1
66.6
9.7
77.0 18.1 53.0
8.8
-1
96.3
10.5
65.4
5.8
71.7
5.6
54.8
1.4
Mix 1
0
49.7
3.0
34.5
2.3
58.3
4.9
51.8
11.1
1
33.6
5.3
21.8
2.7
35.2
2.5
23.2
2.0
2
30.9
1.0
20.6
1.0
29.9
4.2
21.7
1.3
-2
107.9 12.9
71.8
6.5
42.5
9.7
26.2
1.4
-1
364.4 37.2 244.6 25.0
95.8 11.0 67.2
5.6
0
448.5 75.9 264.6 46.4
90.5 10.3 64.6
2.6
Mix 2
1
494.7 31.1 312.8
8.6
113.9 4.9
69.3
5.9
2
58.8
5.5
43.5
6.2
67.8
5.4
41.9
6.2
Table 2A.4 Average amount of DNATrax material collected on glass slides (n=3; ng/cm2) along
Direction
of
Release

the cross-sections of the center line during the multiple barcode testing.

35

REFERENCES

36

REFERENCES

1. An Introduction to Indoor Air Quality. [EPA website]. March 16, 2016.
http://www2.epa.gov/indoor-air-quality-iaq/introduction-indoor-air-quality. Accessed
July 13, 2016.
2. Byron PR. Prediction of drug residence times in regions of the human respiratory tract
following aerosol inhalation. J Pharm Sci. 1986. 75: 433–438.
3. Tsuda A, Henry FS, Butler JP. Particle transport and deposition: basic physics of particle
kinetics. Compr Physiol. 2013. 3: 1437-1471.
4. Sinclair RG, Rose JB, Hashsham SA, Gerba CP, Haas CN. Criteria for selection of surrogates
used to study the fate and control of pathogens in the environment. Appl Environ
Microbiol. 2012. 78: 1969-1977.
5. Underwood DM, Herron DL, Croisant WJ. Whole-building dispersion of tracer gas after
internal release in an administrative/classroom building. ASHRAE Tran. 2007. 113: 457465.
6. Emanuel PA, Buckley PE, Sutton TA, Edmonds JM, Bailey AM, Rivers BA, Kim MH, Ginley
WJ, Keiser CC, Doherty RW, Kragl FJ, Narayanan FE, Katoski SE, Paikoff S, Leppert
SP, Strawbridge JB, VanReenen DR, Biberos SS, Moore D, Phillips DW, Mingioni LR,
Melles O, Ondercin DG, Hirsh B, Bieschke KM, Harris CL, Omberg KM, Rastogi VK,
Cuyk SV, Gibbons HS. Detection and tracking of a novel genetically tagged biological
simulant in the environment. Appl Environ Microb. 2012. 78: 8281-8288.
7. King MF, Noakes CJ, Sleigh PA, Camargo-Valero MA. Bioaerosol deposition in single and
two-bed hospital rooms: A numerical and experimental study. Build Environ. 2013. 59:
436-447.
8. Lopez-Andreo M, Lugo L, Garrido-Petierra A, Preito MI, Puyet A. Identification and
quantitation of species in complex DNA mixtures by real-time polymerase chain reaction.
Anal Biochem. 2005. 339: 73-82.
9. Harding RN, Hara CA, Hall SB, Vitalis EA, Thomas CB, Jones AD, Day JA, Tur-Rojas VR,
Jorgensen T, Herchert E, Yoder R, Wheeler EK, Farquar GR. Unique DNA-barcoded
aerosol test particles for studying aerosol transport. Aerosol Sci Technol. 2016. 50: 429435.
10. Cable-Dunlap P, Trowbridge L, Bostick D, Lee D, Anderson B, Harter A, Kapsimalis R,
Sexton L, De Gange J, Radford D. Comparison of active and passive environmental
sampling for safeguards applications. J Radioanal Nucl Chem. 2013. 296: 943-949.

37

11. Sze To GN, Wan MP, Chao CYH, Fang L, Melikov A. Experimental study of dispersion and
deposition of expiratory aerosols in aircraft cabins and impact on infectious disease
Transmission. Aerosol Sci Technol. 2009. 43: 466-485.
12. Leith D, Sommerlatt D, Boundy MG. Passive sampler for PM10–2.5 aerosol. J Air Waste
Manage Assoc. 2007. 57: 332-336.
13. Wagner J, Macher JM. Comparison of a passive aerosol sampler to size-selective pump
samplers in indoor environments. AIHA Journal. 2003. 64: 630-639.
14. Ott DK, Kumar N, Peters TM. Passive sampling to capture spatial variability in PM10–2.5.
Atmos Environ. 2008. 42: 746-756.
15. Oldham MJ. Challenges in validating CFD-derived inhaled aerosol deposition predictions.
Inhal Toxicol. 2006. 18: 781-786.
16. Tao Y, Inthavong K, Petersen P, Mohanarangam K, Yang W, Tu J. Experimental and CFD
modelling of indoor air and wake flow from a moving manikin. 20th Australasian Fluid
Mechanics Conference. 2016.
17. Sajjadi H, Salmanzadeh M, Ahmadi G, Jafari S. Simulations of indoor airflow and particle
dispersion and deposition by the lattice Boltzmann method using LES and RANS
approaches. Build Environ. 2016. 102: 1-12.
18. Lu W, Howarth AT, Adams NM, Riffat SB. CFD modeling and measurement of aerosol
particle distributions in ventilated multizone rooms. ASHRAE Trans. 1999. 105, part 2:
116-127.
19. Sohn MD, Apte MG, Sextro RG, Lai ACK. Predicting size-resolved particle behavior in
multizone buildings. Atmos Environ. 2007. 41: 1473-1482.
20. Richmond-Bryant J, Eisner AD, Bixey LA, Wiener RW. Transport of airborne particles
within a room. Indoor Air. 2006. 16: 48-55.
21. Udey, RN. Statistical data analyses of trace chemical, biochemical, and physical analytical
signatures. Doctoral dissertation. 2013. Retrieved from MSU Libraries.

38

CHAPTER THREE: Stability of DNATrax Material as a Function of Exposure to UV Radiation

3.1 Introduction
Assessments of aerosol transport are facilitated by releases of tracer aerosols and
quantitative measurements of particle numbers at various locations and times. The suitability of
DNA-barcoded tracer aerosols (DNATrax) depends on stability of the DNA barcodes during the
time between release and tracer measurement. Outdoor field studies that employ DNATrax
present additional challenges not encountered with indoor particle releases including the
potential for photodegradation of the DNA barcodes.

3.1.1 Photodegradation Mechanisms
Although DNA primarily absorbs UVC (λ<280 nm) radiation, absorption by stratospheric
ozone results in minimal solar energy below 290 nm reaching the earth’s surface1. With few
solar UVC rays, the effects of UVB (λ=280-320 nm) and UVA (λ=320-400 nm) wavelengths
must be considered in the photodegradation of DNA in outdoor environments. Direct absorption
of a photon is more likely to occur with UVC or UVB radiation as the absorption of these
wavelengths by DNA is more efficient than UVA2. However, the presence of photosensitizers, or
chromophores that become excited by the absorption of UVA photons, can lead to degradation
reactions of the DNA3-4. Photosensitization is a common occurrence in vivo where chromophores
such as riboflavin5 and benzophenone6 are present, but photosensitization can also play a role in
isolated DNA degradation when other chromophores are present.
There are several mechanisms through which DNA photodegradation can occur including
pyrimidine dimerization7-8 and purine oxidation5-6. Dimerization of adjacent pyrimidine bases

39

(thymine or cytosine) results in intra-strand crosslinking, while dimerization between pyrimidine
bases on separate strands is termed inter-strand crosslinking. Intra-strand crosslinking of adjacent
pyrimidines forms a range of photoproducts including 6,4-linked bases and cyclobutane dimers
(Figure 3.1). Formation of cyclobutane dimers is favored over the 6,4-photoproduct because the
6,4-photoprocts require more extensive DNA unwinding and base rotations than the cyclobutane
dimers9. Cyclobutane dimerization occurs with highest frequency between adjacent thymines10.

Figure 3.1 The pyrimidine base a) thymine and two products of photoinduced crosslinking of
adjacent thymine monomers: b) 6,4-photoproduct and c) cyclobutane dimer.

Oxidation of guanine and the formation of 8-oxo-7,8-dihydroguanine (8-oxo-dGuo)
occurs through two main types of reactions: Type I and Type II, summarized in Figure 3.2. In
their simplest definitions, Type I reactions involve radical formation through electron transfer
while Type II reactions involve energy transfer to oxygen11. In Type I reactions, the electron
transfer steps result in formation of a radical guanine, as guanine is the DNA base with the
lowest ionization energy of DNA bases12, followed by reaction of the guanine radical cation with
oxygen to give the oxidized product (8-oxo-dGuo) (Figure 3.3a). While the final oxidation steps
are the same, the radical guanine can be formed through either direct adsorption of UVC
radiation which provides sufficient energy to ionize guanine (~7.8 eV; single photon of ~160
40

nm)12-13 or through reactions with radical reactive oxygen species (ROS), the product of electron
transfer reactions between an activated photosensitizer and water. With little solar UVC radiation
reaching the Earth’s surface, Type I reactions that involve photosensitizers are more probable.

Figure 3.2 Summary of photooxidation reaction mechanisms.

41

Figure 3.3 Formation of 8-oxo-7,8-dihydroguanine a) through ionization and reaction with water
or oxygen and b) through reaction with singlet oxygen13.

Rather than transferring electrons, the activated photosensitizers in Type II reactions
transfer energy to molecular oxygen to form singlet oxygen (1O2), another ROS. Upon energy
transfer from photosensitizer excited state to molecular oxygen ground state, the spin of one of
the unpaired electrons flips and the first singlet state, 1∑g+, is formed5-6,14. This singlet state
relaxes to the second, lower energy, singlet state (1∆g) in which the electrons are spin paired15. In
contrast to triplet oxygen (3O2) which has two half-filled π* orbitals, the singlet state (1O2) has an
empty orbital and acts as an electrophile, adding to the conjugated diene in guanine’s imidazole
ring (Figure 3.3b)16. Further reduction of this intermediate yields the 8-oxo-dGuo product17.
DNA lesions caused by both oxidation and dimerization interrupt DNA replication and
inhibit the quantitative real-time polymerase chain reaction (qPCR) amplification used for
DNATrax detection and quantitation18. qPCR analyses of UV-exposed biological material
recommend amplifying DNA with short (100 bp or less) sequences to minimize the likelihood
that amplified degradation products from nontarget DNA will interfere with measurement of the
target sequence19. In this investigation, however, the entire length of the oligonucleotides used as
42

the DNA label in DNATrax is only ~100 bp in length, so any sites of degradation along the DNA
sequence will result in a loss of amplification using the PCR assay. It is, therefore, important to
understand the extent of this degradation with UV exposure to predict the decrease of detectable
DNA label expected over the course of a release experiment performed outdoors.

3.1.2 Specific Aims
Outdoor releases of DNATrax aerosol tracer expose DNA to solar UV radiation. With
quantification of the DNA vital to the detection of the tracer, degradation of the DNA that
inhibits qPCR detection may compromise use of DNATrax in outdoor environments. Thus, the
aim of this work is to determine the extent of DNA degradation that occurs with UV irradiation
and assess whether steps might be taken to minimize photodegradation.

3.2 Materials and Methods
3.2.1 Material Production and DNA Analysis
Solid DNATrax material was produced via spray-drying using a Mini Spray Dryer B-290
(Büchi; Switzerland), using the parameters discussed in Chapter 2, Table 2.120-21. Several batches
of solid material were produced with different DNA sequences, DNA concentrations, and
concentrations of the UV-absorbing compound fluorescent brightener 220 (AK Scientific; Union
City, CA), according to the concentrations in Table 3.2, preparations 1-5. Additionally, aqueous
solutions containing DNA, organic food-grade tapioca maltodextrin (DE 10; Ciranda; Thailand),
and fluorescent brightener 220 were prepared in PCR-grade water (Teknova; Hollister, CA)
according to the concentrations in Table 3.2, preparations 6-8. Two different sequences of
double-stranded DNA were used during these studies, denoted as Barcodes A and B, respectively

43

(see Table 3A.1. for sequence details). While the standard solutions of each DNA sequence used
for calibration curves were purchased (Biosearch Technologies; Petaluma, CA), the stock
solutions of DNA used for solid production and in the aqueous solutions listed in Table 3.1 were
from standard DNA solutions previously amplified using a Taq-man quantitative real-time
polymerase chain reaction (qPCR) assay (Invitrogen; Carlsbad, CA) with sequence-specific
primers and probes (Integrated DNA Technologies; Coralville, IA). The components in the
qPCR master mix and the thermal profile used for amplification are the same as those listed in
Chapter 2, Tables 2.2a and 2.2b, respectively. These qPCR parameters were also used for
quantification of DNA concentrations in the solid material and aqueous solutions before and
after irradiation.

DNA
DNA Concentration
Maltodextrin
Fluorescent Brightener 220
Barcode
(ng DNA/mg solids)
Concentration (w/w)
Concentration (w/w)
S1
A
12.5
99%
1%
S2
A
8.7
100%
N/A
S3
A
127.5
100%
N/A
S4
A
12.7
100%
N/A
S5
B
51.7
100%
N/A
Table 3.1 Composition of each solid material used for degradation experiments.
Prep

Prep

DNA
Barcode

DNA Concentration
(pg DNA/µL solution)

Maltodextrin Concentration
in Solution (w/v)

A1

A

20.2

0.2%

A2

A

20.4

0.2%

Fluorescent Brightener
220 Concentration
0.002% (w/v);
1% (w/w) total solids
N/A

A3

B

21.4

3%

N/A

Table 3.2 Composition of each aqueous solution used for degradation experiments.

44

3.2.2 UV Degradation Under Germicidal Lamp
To assess the photodegradation of DNATrax, solid material and aqueous solutions
containing DNA were irradiated using a germicidal lamp for up to 1 hour. The germicidal lamp
used was a 15-watt lamp, delivering 136.8 µW-cm-2 at 18 inches (46 cm) from the lamp
(AirClean; Creedmoor, NC). This lamp generates significant emission in the UVC, UVA and
visible ranges of wavelengths, and substantially less in the UVB wavelength range (Figure 3.4)22.
While an imperfect simulator of solar irradiation, this lamp allows a consistent, known dose of
each wavelength to be administered to the samples.

Figure 3.4 Intensity factors for light emitted across UV and visible wavelengths for the
germicidal lamp used, as provided by the manufacturer.

For photoirradiation experiments, 10-20 mg of each solid material listed in Table 3.1 was
weighed into uncapped glass scintillation vials that were placed 18 inches below the germicidal
lamp and held at ambient temperature. Three vials of each material were set aside at the
beginning of the experiment and 3 additional vials were removed from under the light at 5, 10,
30 and 60 minute time points. The solids were dissolved in Tris-EDTA buffer (Teknova;
Hollister, CA) and further diluted for DNA quantification. For the aqueous solutions (Table 3.2),
45

50-µL aliquots were prepared, placed in opened 2 mL polypropylene Eppendorf tubes, placed
35.5 cm directly below the lamp, and removed at appropriate time points (3 aliquots removed at
each time point). The aqueous solutions were also diluted before DNA quantification so final
concentrations would fall within the calibration curve concentrations. The amount of DNA in
each sample was quantified by comparing the cycle threshold of each sample to a calibration
curve of known DNA concentrations. Quantification of the extent of degradation was achieved
by comparing the DNA concentration after each length of irradiation time with the control
sample exposed to no irradiation.

3.2.3 Gel Electrophoresis Analysis
Formation of DNA degradation products was also assessed using gel electrophoresis
(GE). GE analysis was performed using 20-µL aliquots of dissolved solid material (Preps S3 and
S4, Table 3.1) or aqueous solutions (Prep A2, Table 3.2). These aliquots were removed after
irradiation and dissolution (in the case of the solids) but prior to dilution and PCR amplification.
Aliquots were loaded onto a 4% agarose gel, pre-loaded with ethidium bromide stain, that
provides 3.2 cm of run length for samples (Invitrogen; Carlsbad, CA). The electrophoretic
separation was conducted for 15 minutes using the recommended program for the gel E-Base
(Invitrogen; Carlsbad, CA). Images were visualized and recorded using a GeneFlash Bio
Imaging UV transilluminator with an amber filter (Syngene; Frederick, MD). DNA length was
determined by comparison to a 25-base pair (bp) DNA ladder (Invitrogen; Carlsbad, CA).
Standard concentrations of the DNA sequence were also analyzed to ensure the solutions fell
within the linear range of the GE analysis. Quantitation of band brightness was performed using
ImageJ23 and Microsoft Excel for Mac 2011 (Microsoft Corporation, Redmond, WA).

46

3.3 Results and Discussion
3.3.1 DNA Degradation Determination using qPCR
Quantification of DNA photodegradation by qPCR following irradiation showed
exponential decay for both solid an aqueous samples (Figure 3.5). In all cases, the concentration
of amplifiable DNA significantly decreased over the 60 minutes of light exposure (t-test, 95%
confidence). While all irradiated samples showed evidence degradation, the decrease in
amplifiable DNA was most pronounced for the aqueous solutions, exhibiting a half-life of less
than 10 minutes and yielding less than 3% of the DNA remaining intact after 60 minutes of
irradiation. The addition of a UV absorber, fluorescent brightener 220, only provided minimal
but statistically significant (t-test, 95% confidence) difference from solutions without the
brightener at the 60-minute measurement, at which point the solutions containing the brightener
underwent 0.62% more degradation than the solutions without it (2.37±0.08% of DNA
remaining with brightener; 2.99±0.43% of DNA remaining without brightener). Thus, the
addition of the brightener did not protect the DNA in aqueous samples.

47

Amplified DNA Label Relative to
Starting Value by qPCR

1.4

Solid Maltodextrin
with DNA and
Brightener

1.2
1.0

Solid Maltodextrin
with DNA

0.8
0.6

Aqueous DNA with
Maltodextrin and
Brightener

0.4
0.2
0.0
0

10

20
30
40
50
Irradiation Time (min)

60

70

Aqueous DNA with
Maltodextrin

Figure 3.5 Average relative concentrations of DNA Barcode A sequence in solid DNATrax
material and aqueous solutions both with and without Fluorescent Brightener 220 as a function
of irradiation time, assessed using qPCR. Error bars indicate the standard deviation of three
replicate samples irradiated simultaneously. Numerical values for this data is found in Table
3A.2.

When incorporated into the solid material, the DNA experienced less degradation than in
aqueous solutions. After 10 minutes of irradiation, there was a significant difference between the
degradation in the solid material (93±6% and 79±17% DNA remaining with and without
brightener, respectively) and that of the aqueous DNA (34±2% and 36±7% DNA remaining with
and without brightener, respectively). The brightener provided additional protection of the DNA
in the solid samples, with a significant difference between the solid samples with and without
brightener after 30 minutes of exposure (95% confidence), preserving at least 20% more DNA
from degradation by its presence. Degradation in the solid samples occurred mainly during the
first 10 minutes of exposure, with little further change in the amount of undegraded DNA
48

remaining after 30 and 60 minutes of exposure. This suggests that degradation mechanisms
responsible for the DNA degradation over the first 10 minutes are inhibited at the later time
points through either depletion of a reactant or generation of a photoprotective product that
protects the DNA from photodegradation.
The shapes of the aqueous and solid DNA degradation curves provide information about
the kinetics of the degradation reactions dominating in each material type. Direct absorption of
UV radiation and its degradation via intramolecular thymine crosslinking is expected to proceed
via first order kinetics, with reaction rates proportional to DNA concentration. The reaction rate
of oxidative degradation depends on the formation of ROS more than the DNA concentration
and can exhibit either zero-order kinetics when ROS concentration is in excess relative to the
DNA, or higher order kinetics when the formation of ROS is slower than its reaction with DNA.
To determine if the degradation proceeds via a first- or second-order reaction, the natural log of
the concentration of amplifiable DNA and the inverse of the concentration were also graphed as
a function of time for both the DNA in aqueous solution (Figure 3.6) and solid material (Figure
3.7). Both the first- and second-order plots give good linear correlation (R2>0.9) for the aqueous
solutions, although visual comparison reveals the data at irradiation times of 30 minutes or less
are better fit by the first-order plot. This suggests that the rate-limiting step in the degradation
process is the reaction between a ROS and the DNA. Each strand of DNA has multiple sites with
the potential for oxidation or thymine cross-linkage, so multiple ROSs can interact with each
strand. However, after only one site has been altered, the DNA will no longer replicate by qPCR
and will be considered degraded. Thus, as the concentration of undegraded DNA decreases in the
aqueous solution, the probability that the ROS will interact with an undegraded DNA strand also
decreases, slowing the rate of reaction over time. The same first-order kinetics in effect in the

49

aqueous solution are not observed for degradation in the solid material, suggesting a more
complex degradation mechanism. In the solid material, the extent of DNA degradation is not
only limited by ROS formation and its reaction with undegraded DNA, but also by the mobility
of reactive species within the solids to facilitate that reaction, and these processes may be slow.
The degradation may also be inhibited by formation of photoproducts that either consume ROS
or absorb UV light without leading to DNA degradation.

50

Natural Log of the Amplified DNA
Label Relative to Starting DNA Value
by qPCR

a)

1.2
1.0
R² = 0.9494

0.8

0.6
R² = 0.9318

Linear (Aqueous DNA
with Maltodextrin and
Brightener)
Linear (Aqueous DNA
with Maltodextrin)

0.4
0.2
0.0
0

Inverse of the Amplified DNA
Label Relative to Starting DNA Value
by qPCR

b)

Aqueous DNA with
Maltodextrin and
Brightener
Aqueous DNA with
Maltodextrin

10

20
30
40
50
Irradiation Time (min)

60

70

50

45
40

Aqueous DNA with
Maltodextrin and
Brightener
Aqueous DNA with
Maltodextrin

R² = 0.9414

35
30
25
20

R² = 0.9572

Linear (Aqueous DNA
with Maltodextrin and
Brightener)
Linear (Aqueous DNA
with Maltodextrin)

15
10
5
0
0

10

20
30
40
50
Irradiation Time (min)

60

70

Figure 3.6 Ratio of the average a) natural log and b) inverse of the amplifiable DNA Barcode A
sequence in aqueous solutions by qPCR after UV exposure to the amount of DNA present with
no exposure. The error bars indicate the standard deviations of triplicate samples. The best fit
lines for the data, forced through an intercept of 1, are also shown.

51

Natural Log of the Amplified DNA
Label Relative to Starting DNA Value
by qPCR

a)

1.02
1.01
R² = 0.533

1.00
0.99

0.98

Solid Maltodextrin with
DNA

0.97

Linear (Solid Maltodextrin
with DNA and Brightener)

0.96

Linear (Solid Maltodextrin
with DNA)

R² = 0.2557

0.95
0

Inverse of the Amplified DNA
Label Relative to Starting DNA Value
by qPCR

b)

Solid Maltodextrin with
DNA and Brightener

10

20
30
40
50
Irradiation Time (min)

60

70

2.5
R² = 0.3702

2.0

Solid Maltodextrin with
DNA and Brightener

1.5

Solid Maltodextrin with
DNA

1.0

R² = 0.5805

Linear (Solid Maltodextrin
with DNA and Brightener)

0.5

Linear (Solid Maltodextrin
with DNA)

0.0
0

10

20
30
40
50
Irradiation Time (min)

60

70

Figure 3.7 Ratio of the average a) natural log and b) inverse of the amplifiable DNA Barcode A
sequence in solid material by qPCR after UV exposure to the amount of DNA present with no
exposure. The error bars indicate the standard deviations of triplicate samples. The best fit lines
for the data, forced through an intercept of 1, are also shown.

52

3.3.2 Degradation Products Analysis by Gel Electrophoresis (GE) and qPCR
Additional analyses, including gel electrophoresis and qPCR of a sequence with different
guanine and thymine content, were performed to further evaluate degradation mechanisms
dominating in each material type. Despite qPCR evidence for substantial photodegradation of
DNA in aqueous solutions and somewhat less in solids, gel electrophoresis of products after
irradiation showed minimal change in band migration or band brightness, with all of the major
DNA bands present at approximately 100 bp (Figure 3.8a). However, visual assessment revealed
a very weak band present in the “high” solid samples. After enhancement (Figure 3.8b), these
low-intensity bands were shown to be at ~200 bp. Since DNA migrates as a function of size in
GE, any changes in the migration of DNA or the presence of new bands would indicate that
degradation has resulted in a change in the molecular size, either through chain fragmentation
(shift to smaller size DNA) or inter-strand crosslinking that results in dimer formation (shift to
larger size DNA). Given that the DNA of the low intensity band was roughly double the size of
the DNA strand, this band suggests the formation of dimers within the “high” solid material.
These dimer bands were not present after irradiation of either the aqueous or “low” solid
material, indicating that either inter-strand crosslinking did not occur in these samples or the
concentration of the dimers was below the limit of detection for GE analysis.

53

Figure 3.8 GE results of aqueous DNA (Aq) and solid material with DNA labeling (High and
Low), after varying lengths of irradiation with no PCR amplification prior to GE analysis. Serial
dilutions of a standard aqueous DNA solution (Std Aq) as well as a DNA ladder (DNA Lad) are
included for reference. The image is shown both a) as acquired at the time of analysis and b)
after adjustment of the brightness scale to improve visualization of low intensity bands,
highlighted using the black box, using Adobe PhotoShop.

Direct comparison between the GE results and qPCR was achieved by quantifying the
pixel brightness of the 100-bp band across each lane of the gel using ImageJ. These values were
background subtracted using the pixel brightness for PCR water blanks and then plotted as a
54

function of migration distance. The results from each time point of the “high” solid material are
shown as an example in Figure 3.9. Calculating the area under each curve and then dividing by
the value of the “no UV” sample converts the pixel brightness into numerical values that can be
compared to the qPCR values. This process was repeated for the aqueous DNA solutions and

Average Pixel Brightness Across GE
Lane

“low” solid material with the results of each shown in Figure 3.10a-c, respectively.

140
High, No UV
High, 10 minutes
High, 30 minutes
High, 60 minutes
PCR water

120
100
80
60
40
20
0
0.0

0.5

1.0

1.5

2.0

2.5

3.0

3.5

4.0

4.5

Migration Distance (cm)

Figure 3.9 Average pixel brightness of horizontal pixels across a GE lane as a function of
migration distance for UV-degraded samples of the Barcode A-labelled solid with high DNA
labeling. All data have been background subtracted.

55

Fraction DNA Remaining
Compared to Time 0

a)

Fraction DNA Remaining
Compared to Time 0

b)

1.2
1.0
0.8
0.6

Aqueous,
GE

0.4
0.2

Aqueous,
PCR

0.0
0

Fraction DNA Remaining
Compared to Time 0

20
30
40
50
Irradiation Time (min)

60

70

1.2
1.0
0.8
0.6
High, GE

0.4
High, PCR

0.2
0.0
0

c)

10

10

20
30
40
50
Irradiation Time (min)

60

70

1.2
1.0
0.8
0.6
Low, GE

0.4

Low, PCR

0.2

0.0
0

10

20
30
40
50
60
70
Irradiation Time (min)
Figure 3.10 Fraction of DNA Barcode A sequence remaining quantified through qPCR as well as
through GE with image brightness analysis for a) aqueous DNA, b) solid material with a high
level of DNA labeling, and c) solid material with a low level of DNA labeling.

56

For all three sample types, the qPCR measurements reflect more extensive degradation
than the changes in GE band intensity (Figure 3.10). This difference was most pronounced for
the aqueous DNA solutions (Figure 3.10a). The aqueous solutions experienced no discernable
change in GE band brightness, even after 60 minutes of irradiation while qPCR indicated 80% of
the DNA was converted to unamplifiable forms. Thus the degradation mechanisms occurring in
the aqueous samples result in no change in DNA length and occur through oxidation or intrastrand crosslinking. GE analyses did not yield evidence that lower molecular mass forms of
DNA were produced during irradiation.
For the solid DNATrax particles, the results were more complex (Figure 3.10b and c).
The GE results revealed that, at its steady state (measured by qPCR) at 30 minutes or more
irradiation time, the 100-bp DNA concentration was reduced by 10-20% from the initial
concentration. However, the qPCR results showed that the concentrations of amplifiable DNA
had decreased by 30-40% from the initial concentration over the same time frame. Thus, only
about half of the degradation that inhibited PCR also resulted in a change in the observed DNA
size, so multiple degradation mechanisms must be present and yield more than one product.
Based on the weak band seen with image enhancement indicating dimer formation (Figure 3.8b),
the decrease the 100-bp band intensity corresponded to dimer formation due to inter-strand
crosslinking. The remaining degradation assessed by qPCR that did not result in a size change is
attributed to intra-strand crosslinking or oxidation similar to what was observed for the aqueous
samples.

57

3.3.3 Degradation Mitigation
To determine if the extent of photodegradation is sequence-specific and thus favors
mechanisms involving thymine or guanine, materials containing a second DNA sequence
(Barcode B) were also irradiated under identical conditions. The original double-stranded DNA
sequence, Barcode A, was 103-bp long with 58 thymine residues and 45 guanine residues
between the two strands. Forty (69%) of these thymine residues were adjacent to another
thymine residue and would provide a location for intra-strand crosslinking. The alternate doublestranded DNA sequence, Barcode B, was 100-bp long with a lower thymine content (48 thymine
resides) and a higher guanine content (52 guanine residues). Twenty-six (54%) of the thymine
residues in Barcode B were adjacent to another thymine.
Regardless of sequence, the aqueous and solid degradation curves resulted in similar
behavior for the two sequences, with Barcode B DNA in the solid material again undergoing less
extensive degradation than the aqueous DNA (Figure 3.11). However, using the Barcode B DNA
sequence resulted in significantly less DNA degradation at each time point than when using the
Barcode A DNA sequence for both aqueous DNA (77.7% vs. 3.0% DNA remaining after 60
minutes) and the solid material (93.2% vs. 63.4% DNA remaining after 60 minutes) (Figure
3.11; 95% confidence). While the similar curve shape suggests the same mechanisms dominate
for both sequences, the lower thymine content and decreased extent of degradation in the 359
sequence is consistent with intra-strand crosslinking of adjacent thymines being a primary
degradation mechanism. However, additional investigations using other sequences with varied
thymine and guanine content are needed to establish principles useful for predicting rates of
photodegradation.

58

b)

Amplified DNA Label
Relative to Starting Value by
qPCR
Amplified DNA Label
Relative to Starting Value by
qPCR

a)

1.4
1.2
1.0
0.8
0.6
0.4
0.2
0.0
0

10

20

30
40
Irradiation Time (min)

50

60

70

0

10

20

30
40
Irradiation Time (min)

50

60

70

1.4
1.2
1.0
0.8
0.6
0.4
0.2
0.0

Figure 3.11 Average relative concentrations of either the Barcode A sequence (blue traces) or
Barcode B sequence (orange traces) of DNA in a) solid DNATrax particles and b) aqueous
solutions as a function of irradiation time, assessed using qPCR. Standard deviations are the
three replicate samples irradiated simultaneously. Numerical data for these graphs can be found
in Table 3A.3.

3.4 Conclusions
Irradiation of the original formulation of DNATrax (solid sample, maltodextrin, 610
sequence) yielded less than 40% loss in quantifiable DNA, even after 60 minutes of irradiation
using a germicidal lamp that included wavelengths across the UVC, UVB, and UVA ranges. The
59

rate of degradation of the DNA barcode was more pronounced for aqueous samples of DNA than
for the solid material, suggesting the solid matrix affords protection against UV-induced damage.
The addition of a UV-absorber, Fluorescent Brightener 220, provided added protection for solid
samples irradiated for 30 minutes or longer.
With a variety of wavelengths used for photodegradation, evidence of multiple
degradation mechanisms was unsurprisingly observed. Comparison of qPCR and GE results
revealed that the majority of degradation does not cause substantial change in DNA strand
length, likely occurring though mechanisms such as intra-strand crosslinking or oxidation. While
not present in aqueous samples, minor amounts of DNA dimers formed in the solid material
through inter-strand crosslinking. The rate of degradation was qualitatively linked to the DNA
sequence, although additional studies must be done to quantitatively correlate thymine or
guanine content to the rate of reaction. While the use of a germicidal lamp provided control over
some parameters associated with UV degradation, additional irradiation experiments that expose
the DNA directly to sunlight should follow this work.

60

APPENDIX

61

APPENDIX

DNA
Sequence
Name

Length
(bp)

AT
Content

%T
GC
Adjacent
Content
to T

Reverse
Primer
Length
(bp)

Forward
Primer
Length (bp)

Probe
Length
(bp)

56%
69%
44%
20
20
31
(116/206) (40/58)
48%
54%
Barcode B
100
52%
21
21
31
(96/200) (26/48)
Table 3A.1 Base composition and length of the double-stranded DNA sequences used for
Barcode A

103

DNATrax labeling and length of the primer and probe sequences.

Solid

Aqueous
Average
Average
Concentration,
Standard
Concentration,
Standard
Composition
Time
Relative to
Deviation
Relative to
Deviation
Time 0
Time 0
0
1.000
0.083
1.000
0.037
DNA,
5
1.119
0.068
0.766
0.057
Maltodextrin,
10
0.930
0.061
0.343
0.020
and
30
0.957
0.131
0.071
0.005
Brightener
60
0.852
0.065
0.024
0.001
0
1.000
0.106
1.000
0.141
5
0.760
0.083
0.822
0.089
DNA and
10
0.814
0.105
0.361
0.065
Maltodextrin
30
0.602
0.093
0.082
0.023
60
0.634
0.102
0.030
0.004
Table 3A.2 Average relative concentrations of Barcode A DNA in solid DNATrax particles and
aqueous solutions both with and without Fluorescent Brightener 220 as a function of irradiation
time, assessed using qPCR. Standard deviations are the three replicate samples irradiated
simultaneously.

62

Solid

Aqueous
Average
Average
Concentration,
Standard
Concentration,
Standard
Composition
Time
Relative to
Deviation
Relative to
Deviation
Time 0
Time 0
0
1.000
0.027
1.000
0.009
5
1.064
0.128
0.996
0.025
DNA and
10
0.930
0.039
0.918
0.055
Maltodextrin
30
0.947
0.047
0.787
0.082
60
0.932
0.031
0.777
0.115
Table 3A.3 Average relative concentrations of Barcode B DNA in solid DNATrax particles and
aqueous solutions as a function of irradiation time, assessed using qPCR. Standard deviations are
the three replicate samples irradiated simultaneously.

63

REFERENCES

64

REFERENCES

1. Gates DM. Spectral distribution of solar radiation at the Earth’s surface. Science. 1966. 151:
523-529.
2. Mouret S, Philippe C, Gracia-Chantegrel J, Banyasz A, Karpati S, Markovitsi D, Douki T. UVAinduced cyclobutane pyrimidine dimers in DNA: a direct photochemical mechanism?
Org Biomol Chem. 2010. 8: 1706–1711.
3. Foote CS. Type I and Type II mechanisms of photodynamic action. ACS Symposium. 1987.
339: 22–38.
4. Cadet J, Berger M, Decarroz C, Wagner JR, Van Lier JE, Ginot YM, Vigny P. Photosensitized
reactions of nucleic acids. Biochimie. 1986. 68: 813-834.
5. Joshi PC, Gray TA, Keane TC. Protection of riboflavin and UVB sensitized degradation of
DNA and RNA bases by natural antioxidants. Ecotoxicol Environ Saf. 2012. 78: 86–90.
6. Cuquerella MC, Lhiaubet-Vallet V, Cadet J, Miranda MA. Benzophenone photosensitized
DNA damage. Accounts Chem Res. 2012. 45: 1558-1570.
7. Ballari RV, Martin A. Assessment of DNA degradation induced by thermal and UV radiation
processing: Implications for quantification of genetically modified organisms. Food
Chem. 2013. 141: 2130–2136.
8. Jenkins GJS, Parry JM. PCR-based methodology for the determination of the photoprotection
afforded by sunscreen application. Biotechniques. 2000. 29: 1318-1320, 1323-1326.
9. Pfeifer GP. Formation and processing of UV photoproducts: effects of DNA sequence and
chromatin environment. Photochem Photobiol. 1997. 65: 270-283.
10. Ravanat JL, Douki T, Cadet J. Direct and indirect effects of UV radiation on DNA and its
components. J Photoch Photobio B. 2001. 63: 88-102.
11. Baptista MS, Cadet J, Mascio PD, Ghogare AA, Greer A, Hamblin MR, Lorente C, Nunez
SC, Ribeiro MS, Thomas AH, Vignoni M, Yoshimura TM. Type I and Type II
photosensitized oxidation reactions: guidelines and mechanistic pathways. Photochem
Photobiol. 2017. DOI: 10.1111/php.12716.
12. Orlov VM, Smirnov AN, Varshavsky YM. Ionization potentials and electron-donor ability of
nucleic acid bases and their analogues. Tetrahedron Lett. 1976. 48: 4377-4378.

65

13. Gomez-Mendoza M, Banyasz A, Douki T, Markovitsi D, Ravanat J-L. Direct oxidative
damage of naked DNA generated upon absorption of UV radiation by nucleobases. J
Phys Chem Lett. 2016. 7: 3945−3948.
14. Brash, DE. Sunlight and the onset of skin cancer. Trends Genet. 1997. 13: 410-414.
15. DeRosa MC, Crutchley RJ. Photosensitized singlet oxygen and its applications. Coordin
Chem Rev. 2002. 233-234: 351-371.
16. Ravanat, JL, Cadet J. Reaction of singlet oxygen with 2’-deoxyguanosine DNA. Isolation
and characterization the main and oxidation products. Chem Res Toxicol. 1995. 8: 379388.
of

17. Wei H, Cai Q, Rahn R, Zhang X. Singlet oxygen involvement in ultraviolet (254 nm)
radiation-induced formation of 8-hydroxy-deoxyguanosine in DNA. Free Radic Biol
Med. 1997. 23: 148-154.
18. Moore PD, Bose KK, Rabkin SD, Strauss BS. Sites of termination of in vitro DNA synthesis
on ultraviolet- and N-acetylaminofluorene-treated ϕX714 templates by prokaryotic and
eukaryotic DNA polymerases. Proc Natl Acad Sci. 1981. 78: 110-114.
19. Ballari RV, Martin A. Assessment of DNA degradation induced by thermal and UV radiation
processing: implications for quantification of genetically modified organisms. Food
Chem. 2013. 141: 2130-2136.
20. Harding RN, Hara CA, Hall SB, Vitalis EA, Thomas CB, Jones AD, Day JA, Tur-Rojas VR,
Jorgensen T, Herchert E, Yoder R, Wheeler EK, Farquar GR. Unique DNA-barcoded
aerosol test particles for studying aerosol transport. Aerosol Sci Technol. 2016. 50: 429435.
21. Udey, RN. Statistical data analyses of trace chemical, biochemical, and physical analytical
signatures. Doctoral dissertation. 2013. Retrieved from MSU Libraries.
22. American Air & Water, Inc. Determining ultraviolet intensities using germicidal UV lamps.
Received from AirClean Systems Service Technician on 4/9/17.
23. Rasband WS. ImageJ, U. S. National Institutes of Health, Bethesda, Maryland, USA,
https://imagej.nih.gov/ij/, 1997-2016.

66

CHAPTER FOUR: Stability of DNATrax Material as a Function of Exposure to Humidity

4.1 Introduction
The suitability of DNA-barcoded tracer aerosols (DNATrax) as tracer aerosols for particle
transport studies depends on the stability of both the particle’s size and DNA barcode in the test
environment. Outdoor field studies introduce conditions that can affect this stability that are
otherwise controlled in indoor environments, including variable humidity levels. For example, in
2016 the relative humidity in East Lansing, MI, varied by an average of 37% within a single day
with a difference between the highest and lowest humidities as high as 70% on June 29, 20161.
With such a large change in relative humidity possible over the course of a single day, a better
understanding of how this environmental change affects the particle size and DNA label is
necessary. Both of these aspects are assessed within this Chapter.

4.1.1 Hygroscopicity of Sweeteners
Stability of a material under variable humidity depends in part on the material’s
hygroscopicity, or its affinity to absorb water. Hygroscopic materials will readily absorb the
moisture present at high humidity environments and desorb water at low humidity. For a material
made up of free-flowing particles, the absorption of water not only increases the particle mass
but can also cause particle aggregation and a decrease in powder flowability2,3. This is often a
concern for the powdered food industries (e.g. spices, powdered milk) and is mitigated by the
addition of bulking agents with low hygroscopicity such as maltodextrin, erythritol, xylitol and
isomalt to the formulations4-6. The original DNATrax formulation utilizes maltodextrin as its

67

base material7, and other sweeteners may also serve as suitable materials. The chemical
structures and select physical properties of these sweeteners are shown in Table 4.1.

Compound
Melting
Point (°C)
Glass
Transition
Temperature
(°C)
Material
Density
(g/cm3)
Critical
Relative
Humidity
(%)

Erythritol8

Isomalt8

Xylitol8

Maltodextrin9

121

137

94

240

-42

34

-22

138

1.45

1.50

1.52

1.53

90%

85%

85%

>60%

Structure

Table 4.1 Physical properties and chemical structure of commercially available sweeteners with
low hygroscopicity.

While often expressed as a qualitative comparison (e.g. high or low), hygroscopicity can
be more quantitatively compared using each material’s sorption isotherms, which indicate the
amount of water each absorbs as a function of water activity or relative humidity at a given
temperature. From these isotherms, each sweetener’s critical relative humidity, or relative
humidity above which the moisture content rapidly increases with increasing humidity (a sharp
increase in slope of the absorption), can be determined. While all four sweeteners are considered
to be of low hygroscopicity, the value of this critical relative humidity and the extent of the
material’s moisture absorbance before the critical relative humidity varies. Erythritol and xylitol
68

are both sugar alcohols, with xylitol being more hygroscopic than erythritol. Erythritol absorbs
less than 2% of its weight in water until the relative humidity exceeds 90%10-11, while at 85%
relative humidity xylitol absorbs nearly enough moisture to liquefy its crystals12. Like erythritol,
isomalt is stable across a range of humidities, but particularly absorbs moisture at relative
humidities above 85%13.
Unlike the other sweeteners, which have well-defined physical properties, maltodextrin is
a hydrolysis product of starch, so the oligosaccharide chain lengths and the corresponding
physical properties in maltodextrin vary. Maltodextrins are described by their dextrose
equivalent (DE) value, a measure of the extent of hydrolysis of the starch based on the presence
of reducing sugar end groups, normalized to total mass9. The DE scale defines glucose as a value
of 100, representing complete hydrolysis, and all other sweeteners relative to that value.
Maltodextrins have a DE<20 and their hygroscopicity increases as the DE increases, with the
critical relative humidity at or above 60% RH3,14. Below the critical relative humidity,
maltodextrins can still absorb nearly 20% of their weight in water, with the exact value again
depending on their DE15. With the absorption of water leading to particle aggregation, using a
material with a higher critical relative humidity than maltodextrin that exhibits less water
absorption prior to that point would reduce the probability of particle agglomeration and increase
DNATrax’s usefulness in humid environments.

4.1.2 Particle Size and Aerosol Transport
A main concern with moisture absorption by DNATrax is its potential to exert
subsequent change on particle transport. The persistence of a particle as an aerosol is defined by
its terminal settling velocity (VTS), which in turn depends not only on the viscosity () of the gas

69

in which the particle is suspended and the gravitational constant (g), but also on the particle’s
density (ρp) and physical diameter (dp) (Equation 4.1)16. If the density of the particle material
increases due to water absorption or the effective diameter increases due to particle swelling or
agglomeration, the settling velocity of the particle will also increase and the particle will settle at
a faster rate than expected, altering the transport of the particle.

Equation 4.1.
Since particle transport is linked to the physical diameter of an aerosol particle and
defining that diameter is difficult for irregularly-shaped particles, an alternate metric of particle
size is preferred: the aerodynamic diameter. The aerodynamic diameter (da) is defined as the
diameter of a spherical particle of unit density (ρ0, 1 g/cm3) with the same settling velocity as the
particle being measured (Equation 4.2). This allows for convenient comparison and prediction of
the transport of particles of differing densities or shapes. The aerodynamic diameter can be
expressed as a function of the physical diameter and density with a shape correction factor (X)
that accounts for differences in resistive force as a function of shape (Equation 4.3). Since
aerodynamic diameter is defined using spherical particles, the shape correction factor for
spherical particles is 1.

Equation 4.2

70

Equation 4.3

The settling velocity and the corresponding aerodynamic diameter of a particle can be
measured using the particle’s time of flight in a controlled airflow17. In this work, aerodynamic
diameter was measured using an Aerodynamic Particle Sizer Model 3321 (TSI; Shoreview, MN),
which accelerates particles through a nozzle and into 2 partially overlapping laser beams18. The
light scattered as the particle passes through the beams is focused onto a photodetector and
recorded as a function of time. The overlapping beams result in a double-crested signal with the
time between the peaks correlating to the particle’s velocity and the aerodynamic diameter, with
larger particles requiring more time to travel between the beams than smaller particles. The
aerodynamic diameters recorded are binned into 23 bins spanning ~20 µm in diameter.
For particles of a broad size distribution binned into discrete values, the aerodynamic
diameter is reported using the mass median aerodynamic diameter (MMAD) and the 90th mass
percentile (d90). The calculation of these values is summarized in Figure 4.1. The size
distribution is converted from the number of particles of each size initially reported to the
fraction of the mass of material represented by the counts of each size. The particle diameter at
which half of the mass is distributed across smaller diameters and half across larger diameters is
the mass median aerodynamic diameter. The 90th percentile of this distribution is the d90. The
combination of MMAD and d90 indicate the center and width of the mass distribution.

71

Figure 4.1 Example of a lognormal distribution of particle size plotted a) by the fraction of the
total number of sample counts at each aerodynamic diameter and b) by the fraction of the total
mass of particles represented by particles of each aerodynamic diameter. The mass median
aerodynamic diameter (MMAD) and 90th mass percentile (d90) are marked in panel b.

4.1.3 DNA Degradation Mechanisms
In addition to size stability, the DNA label also needs to be stable with exposure to
variable humidities to be useful in outdoor environments. While the presence of moisture in the
particle is not detrimental to DNA by itself19, it does facilitate the formation of reactive oxygen
species and subsequent oxidation of guanine residues in the presence of UV radiation20. A more
72

detailed discussion of oxidative degradation can be found in Section 3.1.1 of this dissertation. In
addition to the oxidative mechanisms, degradation could also occur through contamination by
deoxyribonuclease (DNase) enzymes, which hydrolytically cleave the phosphodiester linkages of
the DNA backbone. In the case of both degradation mechanisms, the resulting DNA might no
longer be amplified and detected using quantitative real-time polymerase chain reaction (qPCR),
offering potential loss of particle detection sensitivity under these conditions. Therefore, it is
important to determine the extent to which the DNA is affected by humid environments.

4.1.4 Specific Aims
With potential for water absorption or desorption by DNATrax particles under variable
humidities, the stability of the particle size and DNA label is uncertain. To assess the stability of
both the aerodynamic diameter and the DNA label, the size and DNA label of DNATrax material
was monitored over time with exposure to high, low, and cycling humidities. Particle production
and hygroscopic stability of additional bulk materials were also performed to investigate the
DNATrax formulation for environments with variable humidities.

4.2 Materials and Methods
4.2.1 Solid Material Production Conditions
Four food-grade sweeteners were selected as potential bulk material for DNATrax:
xylitol (NOW Foods; Bloomingdale, IL), isomalt (Confectionery Arts International; New
Britain, CT), erythritol (NOW Foods; Bloomingdale, IL), and maltodextrin (DE 10; Ciranda;
Thailand). Aqueous solutions of each were prepared in PCR-grade water (Teknova; Hollister,
CA) and spray-dried using a Mini Spray Dryer B-290 (Büchi; Switzerland), according to the

73

parameters previously listed in Table 2.17,21. Of these sweeteners, only erythritol and
maltodextrin produced the free-flowing powder necessary for the particles to be used as an
aerosol simulant (see Section 4.3.1 for discussion) and particles were made using each of these
sweeteners both with and without DNA labeling. For the DNA label, standard solutions of a
custom double-stranded DNA sequence were purchased (Biosearch Technologies; Petaluma,
CA) and used for calibration curves. The stock solutions of DNA used for solids production were
standard DNA solutions amplified using a Taq-man quantitative real-time polymerase chain
reaction (qPCR) assay (Invitrogen; Carlsbad, CA) with sequence-specific primers and probes
(Integrated DNA Technologies; Coralville, IA). The components of the qPCR master mix and
the thermal profile used for amplification and quantitation are identical to those listed in Tables
2.2a and b, respectively. These qPCR parameters were also used for quantitation of the DNA
labeling in the material, using approximately 20 mg (exact mass known) of solid material
dissolved in Tris-EDTA buffer (Teknova; Hollister, CA) and diluted before qPCR analysis. Each
sample was analyzed with triplicate instrumental replicates at two different dilution factors
(1:104 and 1:105). The level of DNA in each solution was determined using a calibration curve of
standard concentrations and the labeling levels in solid materials are reported as the copies of
DNA per particle of 2-µm aerodynamic diameter.

4.2.2 Particle Size Determination
The aerodynamic diameter of the material was determined by first aerosolizing material
into a closed 5-gallon (18.9 L) chamber using an eductor (MIT Lincoln Laboratory; Lexington,
MA) with a compressed gas source (Falcon Safety Products; Branchburg, NJ) (Figure 4.2).
Aerosolization took place over a period of 2-3 seconds. Air was drawn out of the chamber by an

74

Aerodynamic Particle Sizer (APS) Model 3321 (TSI; Shoreview, MN) at a rate of 1 L/min and
was sampled for 30 seconds immediately after aerosolization. These data were exported from the
APS software (TSI; Shoreview, MN) and further calculation of the distribution’s MMAD and
d90 were performed in Microsoft Excel for Mac 2011 (Microsoft Corporation, Redmond, WA).

Figure 4.2 Diagram of the aerosol generation and sizing method used.

4.2.3 Humidity Chamber Set-up
Humidity chambers were used to maintain a constant relative humidity for sample
incubation. The chambers consisted of airtight containers (Rubbermaid; Atlanta, GA) measuring
22.9 cm x 22.9 cm x 7.6 cm and were stored in an oven set to 25°C. Each chamber contained a
30-mL beaker holding a saturated salt solution of either sodium chloride (high humidity
chamber) (Sigma Aldrich; St. Louis, MO) or lithium chloride (low humidity chamber) (Sigma
Aldrich; St. Louis, MO). Saturated salt solutions have defined vapor pressures which control the
humidity within a closed chamber, with predicted relative humidities of 75.8% and 12.0% at
25°C for the high and low humidity chambers, respectively22. The conditions within the chamber
were monitored using an RHT20 Humidity and Temperature Datalogger (Extech; Waltham,
MA), which recorded both the temperature and relative humidity in the chamber every 1 minute
for the duration of the experiment. An additional datalogger was placed on the benchtop to

75

monitor the temperature and humidity of the ambient lab air. After the experiment was
completed, the recorded data were exported from the manufacturer’s software and plotted using
Microsoft Excel for Mac 2011.

4.2.4 Humidity Chamber Incubation
To study the effects of humidity on particle size, ~100 mg of unlabeled (no DNA)
maltodextrin or erythritol material were weighed into scintillation vials. Three aliquots of each
material were placed, uncapped, in the high humidity chamber; 3 were placed in the low
humidity chamber; and 3 on the benchtop (ambient). An additional 3 aliquots of each material
were rotated between the high and low humidity chambers according to the schedule in Table 4.4
to mimic the fluctuations in humidity encountered in outdoor environments. At each time point
(Day 0, 1, 2, 5, and 8), each sample was removed from the chamber, capped, weighed, sized
according the method described in Section 4.2.2, reweighed, and placed into the appropriate
chamber. While Radosta3 et al. reported maltodextrin (DE4) reaching equilibrium in 3-6 days at
98% relative humidity, the humidity of an outdoor environment fluctuates throughout a single
day and these incubation times represent a middle ground between the two time scales.

76

Time
Into
Chamber

Time Out
of
Chamber

Incubation
Time

Rotation

High

Low

Ambient

12% RH
76% RH
12% RH
Benchtop
(Theor.)
(Theor.)
(Theor.)
76% RH
76% RH
12% RH
Day 1
Day 2
1 Day
Benchtop
(Theor.)
(Theor.)
(Theor.)
12% RH
76% RH
12% RH
Day 2
Day 5
3 Days
Benchtop
(Theor.)
(Theor.)
(Theor.)
76% RH
76% RH
12% RH
Day 5
Day 8
3 Days
Benchtop
(Theor.)
(Theor.)
(Theor.)
Table 4.2 Incubation location for the rotation, high, low, and ambient aliquots during the
Day 0

Day 1

1 Day

hygroscopicity study.

To determine the effect of humidity on the DNA label, ~20 mg aliquots of labeled
maltodextrin or erythritol material were weighed into scintillation vials. Similar to the nonlabeled material study described above, triplicate aliquots were prepared as high, low, ambient,
and rotation aliquots and incubated according to the schedule in Table 4.2. Three additional
aliquots of each labeled material were prepared and the DNA content was determined using the
qPCR method described in Section 4.2.1. At Day 8 of the experiment, the DNA in all incubated
aliquots was also quantified using qPCR.

4.3 Results and Discussion
4.3.1 Particle Production using Alternative Sweeteners
Of the four sweeteners – maltodextrin, erythritol, xylitol, and isomalt – only maltodextrin
and erythritol produced particles suitable for use as an aerosol tracer. Production with xylitol did
not yield solid material, but rather produced a viscous sticky liquid. The melting point of xylitol
(94°C) is below the spray dryer’s outlet temperature (110°C) and the solid material produced
melted as it was collected. Reducing the outlet temperature to below 94°C required reducing the
77

overall temperature of the drying gas, which resulted in incomplete drying of the solids and still
no particle formation. Given this, no further testing was performed using xylitol. While the
melting point of isomalt (137°C) is above the outlet temperature and a solid material formed, the
solid was a rubbery texture rather than a free-flowing, crystalline solid. This is due to the
production temperature being above the glass transition temperature (34°C), or the temperature at
which an amorphous solid changes from a hard and brittle “glassy” state at temperatures below
this point to a viscous or rubbery state with at temperatures above this point23. Once cooled to
below the glass transition temperature, the rubbery state will relax into a crystalline solid.
However, the solid formed is not a free-flowing powder, a necessary characteristic for use as an
aerosol tracer. Erythritol and maltodextrin both produced solids as free-flowing powders.
Production with erythritol gave a much lower percent yield than maltodextrin (~5% vs. ~50%) as
much of the erythritol material was accumulated as a sticky coating on the walls of the spray
dryer due to its lower melting point and glass transition temperature. However, free-flowing
material made of each of these sweeteners was used in further studies.

4.3.2 Effect of Humidity Exposure on Bulk Weight
The changes in weight of the maltodextrin-based material with time in each chamber
reflect maltodextrin’s hygroscopic nature (Figure 4.3a). When stored continuously in the high
humidity chamber (62.8 ±2.2% relative humidity), the weight of the maltodextrin material
increased by 3.0% after 1 day in the chamber and did not significantly change for 7 days
afterward (t-test; 95% confidence), having reached equilibrium within the first day. Radosta3 et
al. reported maltodextrin (DE 4) increased in weight by ~10% per day until it reached
equilibrium under 98% humidity. While the humidity in this chamber is lower than in their

78

study, the 3% increase is well below this upper limit for the maximum amount of water absorbed
within a day and it is reasonable that the material can reach equilibrium in this time scale. When
stored in the low humidity chamber (19.4±3.7% relative humidity), the weight of the
maltodextrin-based material continually decreased by at least 0.5% of the bulk weight at each
time point over the 8 days of the experiment (95% confidence), consistent with moisture being
present in the material at the beginning of the experiment. A total of 5.5% of the bulk weight was
lost by the low-humidity aliquots over the 8 days of the experiment with no weight equilibrium
reached in that time. The lack of weight equilibrium with the low-humidity incubation is likely
an effect of the gradual decrease in RH in the low humidity chamber over the course of the
experiment. The humidity level in the low humidity chamber gradually decreased from 23.7%
RH just before the samples were removed from the chamber on Day 1 to 14.4% RH just before
sample removal on Day 8. As the saturated lithium chloride solution absorbs water over time to
reduce the humidity level, the concentration of the dissolved salt at the surface of the slurry is
decreased compared to the bulk solution. With the humidity controlled by the vapor pressure of
the solution (related to the salt concentration), this dilution over time would limit the
effectiveness of the lithium chloride solution at reducing the humidity in the chamber and
prevent the humidity chamber from reaching its literature value for humidity. The solution was
stirred at each time point, increasing the salt concentration at the surface and allowing more
water to be absorbed, effectively lowering the relative humidity in the chamber with cycle of
stirring and humidity equilibration. In future experiments, the salt slurry should be continuously
stirred or have a higher ratio of slurry surface area to chamber volume to bring the chamber
humidity to match the literature values. The change in sample weight has a strong positive
correlation with this humidity change (R-squared value of 0.8700).

79

Total Percent Weight Change,
Relative to Time 0 (%)

a)

4.0%
2.0%
0.0%
0

1

2

3

6

7

8

7

8

7

8

-6.0%

Time in Chamber (Days)
High

Total Percent Weight Change,
Relative to Time 0 (%)

5

-4.0%

-8.0%

b)

4

-2.0%

Low

Rotation

4.0%
2.0%
0.0%
0

1

2

3

4

5

6

-2.0%
-4.0%
-6.0%
-8.0%

Time in Chamber (Days)
High

Measured Relative Humidity in
Chamber (%RH)

c)

Low

Rotation

80.0
70.0
60.0
50.0
40.0
30.0
20.0
10.0
0.0
0

1

2

3

4

5

6

Time in Chamber (Days)
High

Low

Rotation

Figure 4.3 Change in bulk weight of a) maltodextrin-based or b) erythritol-based DNATrax
material in response to c) the relative humidity in each chamber. Error bars indicate the standard
deviations of triplicate samples. Numerical values for bulk weights can be found in Tables 4A.1
and 4A.2.
80

The bulk weight of maltodextrin material rotated between the low- and high-humidity
chambers decreased and increased according to the chamber in which it was stored (Figure 4.3a).
Moisture was removed from the material after each incubation at low humidity, with decreases in
weight from the starting value of 1.7±0.2% and 3.2±0.1% on days 1 and 5, respectively, relative
to the starting weight. There was no significant difference between the weight change in the
rotated material and those stored consistently in the low humidity chamber, even on day 5 when
the rotated samples lost 2.1% of their weight more in water over 2 fewer days of incubation. The
consistency between the low humidity samples and the rotated samples on day 5 supports the
conclusion that the gradual decrease in weight of the low humidity material is due to the gradual
decrease in chamber humidity. Similarly, when rotated samples were placed in the high humidity
chamber, they reached the same equilibrium weight change at days 2 and 8 (2.1±0.4% and
0.6±0.3%, respectively) as those consistently stored in the high humidity chamber (2.1±.9% and
1.3±1.2%, respectively). This consistency of the rotated maltodextrin material with its singlechamber counterparts indicates the reversibility of the water absorption/desorption process in
maltodextrin.
For particle transport purposes, it is not the bulk weight of the material, but the effect of
changes in the material density and the corresponding aerodynamic diameter that is of concern.
Changes in weights of maltodextrin particles ranged from -5.5% (low humidity incubation for 8
days) to +3.0% (high humidity incubation for 1 day) An 8.5% increase in weight with no change
in physical diameter, as might be expected if moisture was absorbed into pores, would increase
the density by 8.5%. Since the aerodynamic diameter of a particle is proportional to the square
root of the particle’s density (Equation 4.3), this would result in a 4.16% change in aerodynamic
diameter. Thus, a maltodextrin particle with a 2.00-µm aerodynamic diameter at low humidity

81

would have an aerodynamic diameter of 2.08 µm at high humidity. This change is small
compared to the variability between replicate aerodynamic diameter measurements of the same
aliquot, which can be as high as 10%. Thus, the change in aerodynamic diameter due to the
absorption of moisture by the maltodextrin material is negligible.
While maltodextrin’s hygroscopic nature was revealed in the high and low humidity
chambers, incubation of erythritol material confirmed its nonhygroscopic nature at these
humidities (Figure 4.3b). The weight of the erythritol material decreased between 0.2% and 1.7%
at each time point, regardless of which humidity condition was used for storage. Additionally,
there was no significant difference in weight between the two storage conditions occurring until
after 8 days of storage. Given that the high humidity chamber reached a maximum relative
humidity of 66.7%, this result is consistent with previous reports that erythritol does not readily
absorb moisture at relative humidities below 90%10. However, the gradual decrease in weight
over time suggests that the material had residual moisture present at the start that was desorbed
over time. Since the material was stored in a sealed vial in a desiccator between production and
the start of the experiment, the desorbed moisture is likely residual moisture from the spraydrying production process. In future hygroscopicity studies, it would be beneficial to dry material
to a constant weight under defined relative humidity before making the initial weight
measurements.

4.3.3 Effect of Humidity Exposure on Particle Size
While neither material exhibited enough water absorption to significantly change the
material’s density and corresponding aerodynamic diameter, the aerodynamic diameter can also
be altered through particle agglomeration. Particle agglomeration will form larger, irregularly

82

shaped clusters with higher aerodynamic diameters. To determine if this agglomeration
significantly affected either material, the aerodynamic diameters of particles in each material
were also assessed at each time point.
As the bulk weight changes suggested, the MMAD and d90 aerodynamic diameters for
both maltodextrin (Figure 4.4) and erythritol (Figure 4.5) showed no significant difference
between material stored at either high and low humidity until after 8 days of incubation (95%
confidence). Additionally, the MMADs and d90s of the rotated aliquots were indistinguishable
from the high- and low-humidity aliquots with few exceptions. The d90 of the rotated
maltodextrin was significantly lower than that of both the high and low humidity samples on day
1, and both the MMAD and d90 were significantly higher than those for the low humidity
samples on day 8 (95% confidence). For erythritol, the d90 of the rotated material was
significantly lower than that of the high humidity samples on day 1 and significantly higher than
that of the low humidity samples on day 2 (95% confidence).

83

Average Mass Median Diameter,
n=3 (µm)

a)

4.0
3.5
3.0
2.5
2.0
1.5
0

1

2

3

High RH

Average Diameter at 90% Mass
Accounted for, n=3 (µm)

b)

Measured Relative Humidity in
Chamber (%RH)

5

6

Low RH

7

8

7

8

7

8

Rotation

8.0
7.5
7.0
6.5
6.0
5.5
5.0
4.5
4.0
3.5
3.0
0

1

2

3

4

5

6

Time in Chamber (days)
High RH

c)

4

Time in Chamber (days)

Low RH

Rotation

80.0
70.0
60.0
50.0
40.0
30.0
20.0
10.0
0.0
0

1

2

3

4

5

6

Time in Chamber (Days)
High

Low

Rotation

Figure 4.4 Average aerodynamic diameter of maltodextrin-based particles assessed using a) the
MMAD and b) the d90 after incubation at high humidity, low humidity, or both on a rotating
basis according to the c) humidity profiles. Error bars indicate the standard deviation of triplicate
samples. Numerical values for bulk weights can be found in Tables 4A.3 and 4A.4.
84

Average Mass Median
Aerodynamic Diameter, n=3 (µm)

a)
6.0
5.5
5.0
4.5
4.0
3.5
3.0
0

1

2

3

4

5

6

7

8

7

8

7

8

Time in Chamber (days)
High RH

Low RH

Rotation

Average Aerodynamic Diameter
at 90% Mass Accounted for, n=3
(µm)

b)
8.5
8.0
7.5
7.0
6.5
6.0
5.5
5.0
0

1

2

3

High RH

Measured Relative Humidity in
Chamber (%RH)

c)

4

5

6

Time in Chamber (days)
Low RH

Rotation

80.0
70.0
60.0
50.0
40.0
30.0
20.0
10.0
0.0
0

1

2

3

4

5

6

Time in Chamber (Days)
High

Low

Rotation

Figure 4.5 Average aerodynamic diameter of erythritol-based particles assessed using a) the
MMAD and b) the d90 after incubation at high humidity, low humidity, or both on a rotating
basis according to the c) humidity profiles. Error bars indicate the standard deviation of triplicate
samples. Numerical values for bulk weights can be found in Tables 4A.5 and 4A.6.
85

The lack of size changes indicates that significant, irreversible agglomeration of particles
did not occur during sample storage and the aerodynamic diameters of maltodextrin and
erythritol particles were stable at both humidities tested. Any particles agglomerates formed were
weakly held together and broken apart by the eductor during aerosolization, resulting in no net
change in the aerodynamic diameter. The lack of a significant change in particle size across this
range of humidities (14.4% to 64.9%) is a positive characteristic for material to be used in field
tests, where humidity can be monitored but not controlled. However, while there was minimal
size change across this range, the humidity levels were kept below the critical relative humidity.
At humidities above the critical value, the material will rapidly increase in moisture content with
a slight increase in humidity. Since the extent of moisture absorbance affects aerodynamic
diameter not only by increasing the density of the material but also through facilitating particle
agglomeration, the results here cannot be extended to humidities above the critical relative
humidity. Additional tests are needed to draw conclusions about particle size stability under
those conditions.

4.3.4 Effect of Humidity Exposure on DNA Degradation
Unlike the particle size, which experienced no significant change during incubation in the
humidity chambers, a significant decrease in the amount of DNA quantified using qPCR was
measured for both maltodextrin and erythritol DNATrax material after 8 days of chamber
incubation as well when stored at ambient conditions (Figure 4.6). In all cases, light exposure
was restricted to ambient light in the lab during weight measurements and processing for PCR.
Both the DNA-labeled maltodextrin and erythritol materials started with similar levels of

86

labeling (670 copies of DNA per particle of 2-µm diameter) and DNA labeling decreased
significantly with the 8-day incubation in all storage conditions (95% confidence).

Figure 4.6 Copies of replicable DNA per particle with a 2-µm aerodynamic diameter in
maltodextrin-based and erythritol-based DNATrax material as determined by qPCR after storage
at low humidity, high humidity, rotation between low and high humidity, or ambient conditions.
Error bars are the standard deviation of 3 samples. Significant differences between DNA levels
are indicated by the brackets (95% confidence).

In the erythritol material, the three sets of aliquots stored within closed chambers (low,
high, and rotation aliquots) degraded in a similar manner with 584±14, 587±6, and 587±26
copies/particle (respectively) remaining after storage (Table 4A.8). Given that erythritol is
nonhygroscopic at the humidity levels in either chamber and did not absorb moisture during the
time of incubation, it can be concluded that the primary mechanism in DNA degradation in the
erythritol material is independent of humidity level or the presence of water in the material. One
87

possible mechanism of this degradation is through hydrolytic cleavage of the DNA backbone by
DNases present in the material. Some precautions were taken to limit the contamination of
DNases in the DNATrax material during material production and storage including bleaching
glassware before use, using PCR-grade (DNase- and RNase-free) water, and autoclaving storage
vials. However, the drying gas used for production was heated ambient air that was not filtered
prior to use. Thus, airborne contaminants could be incorporated into the material during particle
generation. The likelihood of DNase activity contributing to the degradation of DNA is also
supported by the erythritol ambient samples undergoing significantly more degradation than
those in closed humidity chambers with only 543 copies/particle remaining. While the low, high,
and rotation aliquots were placed in sealed chambers, the ambient samples were placed uncapped
on the benchtop and would be more susceptible to contamination by microbial DNases. Although
DNase contamination and the subsequent DNA degradation is independent of humidity and
would be more likely to occur for material open to room air rather than in a closed chamber,
further experiments are required to test this hypothesis.
The results for the DNA stability in the maltodextrin material are not as clear. While the
ambient samples did experience the greatest decrease in DNA after incubation (543
copies/particle; 81.1% remaining), this decrease was not significantly more than storage in the
humidity chambers. Additionally, the samples stored in the low humidity chamber underwent
less degradation than those in the high chamber or rotated between chambers. Maltodextrin did
absorb moisture at high humidity and, while this did not significantly alter the particle size, it
does appear to play a role in the extent of DNA degradation. At all times, the ambient samples
were at relative humidities between those of the high and low humidity chambers (Figure 4.7), so
if the primary mechanism of degradation depended solely on the presence of humidity, then the

88

ambient aliquots would have DNA labeling between those of the high and low chambers. Since
this is not the case, there are likely multiple degradation mechanisms occurring and the DNase
activity suspected in the erythritol material also plays a role in DNA degradation in the

Measured Relative Humidity
(%RH)

maltodextrin material.

80.0
70.0
60.0
50.0
40.0
30.0
20.0
10.0
0.0
0
High
Rotation

1

2
Low
Ambient

3

4

5

6

7

8

Time in Chamber (Days)

Figure 4.7 Relative humidity experienced by aliquots stored in high- or low-humidity chambers,
rotated between them, or stored on the benchtop under ambient condition.

4.4 Conclusions
While fluctuations in humidity are likely during aerosol transport tests in outdoor
environments, the stability of the aerodynamic diameter and persistence of the DNA label in
DNATrax material were not significantly affected by this variable. Two sweeteners,
maltodextrin and erythritol, both produced free-flowing powders and were both determined to be
suitable for bulk material production. While maltodextrin exhibited some (<10%) absorption of
moisture, this did not result in a significant increase in the aerodynamic diameter through either
particle swelling or agglomeration. Production yields of erythritol were an order of magnitude

89

lower than maltodextrin, but no moisture was absorbed during incubation at high relative
humidity (63% RH). While the amount of amplifiable DNA decreased between 7 and 18% after
8 days of storage, there was no correlation between the extent of degradation and the humidity
used, indicating the DNA label is not significantly affected by the presence or lack of humidity at
these values. Further work is required to confirm the stability of the material at humidities above
the critical relative humidity of each material. Overall, DNATrax material made from either
maltodextrin or erythritol is stable over a range of humidities (up to 65% RH) and can be used in
environments where the humidity is variable.

90

APPENDIX

91

Average Amount in
Sample by Counts,
Normalized to
Maximum

Amount in Sample by
Counts, Normalized to
Maximum Value

APPENDIX

1.2
1.0
0.8
0.6
0.4
0.2
0.0

Maltodextrin
Erythritol
0

10
Physical Diameter (um)

15

1.2
1.0
0.8
0.6
0.4
0.2
0.0

20

Maltodextrin
Erythritol

0

Average Amount in
Sample by Mass,
Normalized to
Maximum

5

5

10
15
Aerodynamic Diameter (um)

1.2
1.0
0.8
0.6
0.4
0.2
0.0

20

Maltodextrin
Erythritol

0

5

10
15
Aerodynamic Diameter (um)

20

Figure 4A.1 Particle size distributions for maltodextrin and erythritol-based DNATrax plotted as
the a) physical diameter scaled by counts, b) aerodynamic diameter scaled by counts, and c)
aerodynamic diameter scaled by mass.

92

Maltodextrin Metric
Day 0
Day 1
Day 2
Day 5
Day 8
Average
0.00%
-1.64%
-2.71%
-3.34%
-5.49%
Low
St. Dev.
0.00%
0.08%
0.30%
0.39%
0.85%
Average
0.00%
3.01%
2.15%
2.53%
1.27%
High
St. Dev.
0.00%
0.60%
0.88%
1.04%
1.21%
Average
0.00%
-1.75%
2.12%
-3.24%
0.64%
Rotation
St. Dev.
0.00%
0.19%
0.36%
0.09%
0.34%
Table 4A.1 Changes in bulk weight relative to the starting weight for maltodextrin material.
Averages and standard deviations are for n=3 aliquots.

Erythritol

Metric
Day 0
Day 1
Day 2
Day 5
Day 8
Average
0.00%
-0.51%
-1.66%
-1.83%
-3.52%
Low
St. Dev.
0.00%
0.28%
0.67%
0.85%
0.94%
Average
0.00%
-0.73%
-1.40%
-1.25%
-2.39%
High
St. Dev.
0.00%
0.12%
0.51%
0.30%
0.27%
Average
0.00%
-1.16%
-2.36%
-2.99%
-3.93%
Rotation
St. Dev.
0.00%
0.03%
0.04%
0.24%
0.26%
Table 4A.2 Changes in bulk weight relative to the starting weight for erythritol material.
Averages and standard deviations are for n=3 aliquots.

Maltodextrin,
MMAD
Metric
Day 0
Day 1
Day 2
Day 5
Day 8
Average
2.38
2.56
2.17
2.51
2.01
Low
St. Dev.
0.14
0.27
0.20
0.13
0.09
Average
2.38
2.78
2.65
2.90
2.81
High
St. Dev.
0.14
0.76
0.60
0.59
0.16
Average
2.38
2.13
2.45
2.65
2.83
Rotation
St. Dev.
0.14
0.30
0.19
0.27
0.45
Table 4A.3 Average MMAD values for maltodextrin material. Averages and standard deviations
are for n=3 aliquots.

93

Maltodextrin,
d90
Metric
Day 0
Day 1
Day 2
Day 5
Day 8
Average
5.25
5.46
4.79
5.56
3.74
Low
St. Dev.
0.40
0.45
0.43
0.20
0.12
Average
5.25
5.71
5.48
5.42
5.68
High
St. Dev.
5.25
2.02
0.69
0.67
0.24
Average
0.40
4.61
5.04
5.38
5.90
Rotation
St. Dev.
5.25
0.34
1.22
0.66
1.04
Table 4A.4 Average d90 values for maltodextrin material. Averages and standard deviations are
for n=3 aliquots.

Erythritol,
MMAD

Metric
Day 0
Day 1
Day 2
Day 5
Day 8
Average
4.89
3.94
3.97
4.48
4.17
Low
St. Dev.
0.17
0.33
0.49
0.21
0.29
Average
4.89
4.26
4.37
4.81
4.61
High
St. Dev.
0.17
0.51
0.62
0.61
0.38
Average
4.89
4.27
4.50
4.03
4.55
Rotation
St. Dev.
0.17
0.26
0.29
0.48
0.49
Table 4A.5 Average MMAD values for erythritol material. Averages and standard deviations are
for n=3 aliquots.

Erythritol,
d90

Metric
Day 0
Day 1
Day 2
Day 5
Day 8
Average
7.47
6.31
6.74
7.13
7.18
Low
St. Dev.
0.45
0.68
0.43
0.27
0.26
Average
7.47
7.22
7.19
7.59
7.70
High
St. Dev.
0.45
0.55
1.09
0.36
0.24
Average
7.47
6.00
7.75
7.09
7.44
Rotation
St. Dev.
0.45
0.53
0.24
0.88
0.17
Table 4A.6 Average d90 values for erythritol material. Averages and standard deviations are for
n=3 aliquots.

94

Maltodextrin
Initial QC
Low
High
Rotation
Ambient
Average
670
620
553
574
543
Standard
16
28
13
20
64
Deviation
% Intact
100.0%
92.5%
82.6%
85.6%
81.1%
DNA
Table 4A.7 Copies of DNA per 2-µm particle of maltodextrin-based DNATrax as determined
using qPCR. Averages and standard deviations are for n=3 aliquots.

Erythritol
Initial QC
Low
High
Rotation
Ambient
Average
669
584
587
587
517
Standard
8
14
6
26
37
Deviation
% Intact
100.0%
87.3%
87.8%
87.7%
77.3%
DNA
Table 4A.8 Copies of DNA per 2-µm particle of erythritol-based DNATrax as determined using
qPCR. Averages and standard deviations are for n=3 aliquots.

95

REFERENCES

96

REFERENCES

1. Weather Underground. www.wunderground.com. The Weather Channel Interactive, Inc.
Atlanta, GA. Last accessed June 12, 2017.
2. Juarez-Enriquez E, Olivas GI, Zamudio-Flores PB, Ortega-Rivas E, Perez-Vega S, Sepulveda
DR. Effect of water content on the flowability of hygroscopic powders. J Food Eng.
2017. 205: 12-17.
3. Radosta S, Schierbaum F, Reuther F, Ager H. Polymer-water interaction of maltodextrins. Part
I. Water vapour sorption and desorption of maltodextrin powders. Starch/Stärke. 1989.
41: 395-401.
4. Ávila EL, Rodríguez MC, Velásquez HJC. Influence of maltodextrin and spray drying process
conditions on sugarcane juice powder quality. Rev Fac Nac Agron. 2015. 68: 7509-7520.
5. Goula AM, Adamopoulos KG. Effect of maltodextrin addition during spray drying of tomato
pulp in dehumidified air: II. powder properties. Dry Technol. 2008. 26: 726–737.
6. Caparino OA, Tang J, Nindo CI, Sablani SS, Powers JR, Fellman JK. Effect of drying
methods on the physical properties and microstructures of mango (Philippine ‘Carabao’
var.) powder. J Food Eng. 2012. 111: 135-148.
7. Harding RN, Hara CA, Hall SB, Vitalis EA, Thomas CB, Jones AD, Day JA, Tur-Rojas VR,
Jorgensen T, Herchert E, Yoder R, Wheeler EK, Farquar GR. Unique DNA-barcoded
aerosol test particles for studying aerosol transport. Aerosol Sci Technol. 2016. 50: 429435.
8. Perko R and DeCock P. Erythritol. In: Sweeteners and Sugar Alternatives in Food
Technology. H Mitchell, ed. Blackwell Publishing: Oxford, UK. 2006. 151-176
9. Avaltroni F, Bouquerand PE, Normand V. Maltodextrin molecular weight distribution
influence on the glass transition temperature and viscosity in aqueous solutions.
Carbohyd Polym. 2004. 58: 323-334.
10. DeCock P. Erythritol. In: Alternative Sweeteners, 4th Ed. LOB Nabors, ed. CRC Press,
Taylor & Francis Group: Boca Raton, FL. 2012. 249-264.
11. Bhandari BR, Adhikari BP. Water activity in food processing and preservation. In: Drying
Technologies in Food Processing, 1st Ed. XD Chen, AS Mujumdar, eds. Blackwell
Publishing: Oxford, UK. 2008. 55-89.
12. Hyvönen L, Koivistoninen P. Food technological evaluation of xylitol. Advs Food Res. 1983.
373-403.
97

13. Sträter PJ. Palatinit - Technological and processing characteristics. In: Alternative
Sweeteners. LOB Nabors, RC Gelardi, eds. Marcel Dekker: New York. 1986. 217-244.
14. Morris CE. New applications for maltodextrin. Food Eng. 1984. 56: 48-50.
15. Castro N, Durrieu V, Raynaud C, Rouilly A. Influence of DE-value on the physicochemical
properties of maltodextrin for melt extrusion processes. Carbohyd Polym. 2016. 144:
464–473.
16. Hinds WC. Aerosol Technology: Properties, Behavior, and Measurement of Airborne
Particles, 2nd Ed. John Wiley and Sons, Inc.: New York, NY. 1999.
17. Mitchell JP, Nagel MW. Time-of-flight aerodynamic particle size analyzers: their use and
limitations for the evaluation of medical aerosols. J Aerosol Med. 1999. 12: 217-240.
18. TSI. Aerodynamic Particle Sizer Model 3321: Theory of Operation. 2012. www.tsi.com.
19. Lund S, Dissing J. Surprising stability of DNA in stains at extreme humidity and
temperature. Int Congr Ser. 2004. 1261: 616–618.
20. Mouret S, Philippe C, Gracia-Chantegrel J, Banyasz A, Karpati S, Markovitsi D, Douki T. UVAinduced cyclobutane pyrimidine dimers in DNA: a direct photochemical mechanism?
Org Biomol Chem. 2010. 8: 1706–1711.
21. Udey, RN. Statistical data analyses of trace chemical, biochemical, and physical analytical
signatures. Doctoral dissertation. 2013. Retrieved from MSU Libraries.
22. Wexler A, Hasegawa S. Relative humidity-temperature relationships of some saturated salt
solutions in the temperature range 0 to 50 C. J Res Nat Bur Stand. 1954. 53: 19-26.
23. Roos YH. Water activity and physical state effects on amorphous food stability. J Food
Process Pres. 1993. 16: 433–447.

98

CHAPTER FIVE: Conclusions

From individual barcode quantification during multiplexed experiments to environmental
stability for outdoor studies, the advantages posited at the inception of DNATrax have been
largely proved true. When coupled with passive sampling, this aerosol tracer can provide highresolution spatial distributions of particle deposition across a defined space. Unlike the detection
scheme in many other tracers, the DNA signature used here is tunable, allowing for experiment
multiplexing and reproducibility assessments during field tests, a rarity among aerosol transport
studies. The particle size and DNA label are stable at relative humidities up to 66%, as well as in
varying humidities in that range, and the solid particle affords the DNA label some protection
when exposed to UV radiation. However, there remain additional aspects of field studies that
should be considered before its use.
For indoor studies, the combination of passive sampling with the DNA tagged particles is
a powerful combination for high-resolution sampling. However, the passive sampling material
selected for an experiment should be checked for compatibility with the qPCR assay before each
use, as some commercial products such as Post-It® notes vary in qPCR compatibility between
packs.
Outdoor studies present additional variables to consider. The DNA tag was found to
degrade with exposure to UV radiation, although the solid material offered some protection.
However, it should be noted that the UV exposure here is an imperfect model of the wavelengths
and relative intensities within sunlight and additional work should be done using actual sunlight.
While shown to be stable under variable humidities up to 66%, the DNA tag did experience
some degradation over a week of storage in uncapped vials. When not kept under sterile

99

conditions, contamination by DNases in the environment will lead to DNA degradation.
Contamination by DNases becomes even more prevalent in outdoor studies and could cause
significant degradation in material during experiments lasting 1 week or longer. This would have
to be taken into account when evaluating results from such studies. An additional consideration
for longer studies is the variation in humidity that can occur. While neither maltodextrin or
erythritol bulk materials were negatively affected by humidities as high as 66%, increased
moisture absorption at humidity levels above their critical relative humidity values could lead to
altered particle transport and agglomeration. Additionally, dew formation or precipitation
occurring during the course of a longer study would be detrimental to the DNATrax material, as
the particles are water-soluble. While the DNA tag could still be collected from the aqueous
samples after dissolution, the DNA tag is more susceptible to UV photodegradation in its
aqueous form. For long-term outdoor studies where precipitation is likely to occur, alternative
materials should be investigated for production of particles that are not water soluble and would
maintain the protection of the DNA tag for the duration of such experiments.
The ability to perform simultaneous aerosol transport experiments is a novel aspect of
this work and the power of this experimental design is fully realized when it is coupled with
computational models. The validation of models using experimental data is vital to their accurate
prediction of particle transport in complex spaces. An experimental design that allows particle
deposition pattern to be determined from a variety of release and sampling locations, each under
identical environmental conditions, provides a simple platform for acquiring the necessary data
for model validation, with DNATrax already being used for this purpose by other research
scientists and government agencies.

100