INTEGRATION OF MICRO FLUIDICS AND FLUORESCENCE IN SITU

INTEGRATION OF MICROFLUIDICS AND FLUORESCENCE IN SITU HYBRIDIZATION (FISH) FOR THE RAPID IDENTIFICATION OF MICROORGANISMS

CÁTIA DANIELA CRUZ MOREIRA DISSERTAÇÃO DE MESTRADO APRESENTADA À FACULDADE DE ENGENHARIA DA UNIVERSIDADE DO PORTO EM ENGENHARIA BIOMÉDICA

M 2014

i

This thesis was supervised by:

Professor Nuno Filipe Azevedo

Faculdade de Engenharia, Universidade do Porto

Professor João Mário Miranda

Faculdade de Engenharia, Universidade do Porto

The host institution of this thesis was:

FEUP – Faculdade de Engenharia da Universidade do Porto

LEPABE – Laboratório de Engenharia de Processos, Ambiente, Biotecnologia e Energia

The research described in this thesis was financially supported by:

FEDER funds through the Operational Programme for

Competitiveness Factors – COMPETE, ON.2 - O Novo Norte - North

Portugal Regional Operational Programme and National Funds

through FCT - Foundation for Science and Technology under the

projects: PEst-C/EQB/UI0511, NORTE-07-0124-FEDER-000025 -

RL2_ Environment&Health and DNA mimics Research Project

PIC/IC/82815/2007.

ii

“The two most powerful warriors are patience and time”

Lev Tolstoy, War and peace

iii

Acknowledgements

iv

Acknowledgments

This work was made possible by the valuable contributions of an important number of

people. Firstly, I would like to express my gratitude to my supervisor, Prof. Nuno

Azevedo, and co-supervisor Prof. João Mário Miranda, for their help and support,

excellent advice and encouragement during the development of this work. I also wish

to express my sincere gratitude to Laura Cerqueira, for all her help and support.

I would like to acknowledge the collaboration of several persons from the following

institutions:

Laboratório de Engenharia de Processos, Ambiente, Biotecnologia e Energia

(LEPABE) – Andreia Azevedo, who provided me with the best support I could have

had, for being a friend more than a co-worker, and for always being there for me –

thank you for your patience and understanding. To my colleagues at LEPABE that

have in anyway contributed to this work and provided a good working environment:

Bruno Pereira, Rui Rocha, André Ferreira, Sílvia Fontenete, Nuno Guimarães, Luciana

Gomes, Rita Santos, and Carolina Lima.

Centro de Estudos de Fenómenos de Transporte (CEFT) – Maha Ponmozhi, Lúcia

Sousa, Romeu Matos and João Carneiro for their assistance and technical support.

Mariline Pinto, from who I learnt my first steps on microfluidics, for the suggestions and

assistance, for sharing his scientific knowledge, for all the helpful discussions and for

providing me the friendship I needed.

Finally, I wish to thank my close friends and family, namely to:

Mom: I can’t thank you enough.

My father: thanks for being always there!

BabyBro: I’m sorry for the lack of patience and attention over these last months!

And you can’t imagine how important was your joy every time we jumped on the bed

singing «I’m on top of the world»

A last and very special thanks to my friends José Pedro Gonçalves, Marlene

Morgado, Tatiana Teixeira, and Claudia Tonelo who offer me great moments, and for

having the amazing ability to laugh on my busiest and annoying days!

Abstract

v

Abstract The frequency of invasive fungal infections caused by yeasts has increased in

intensive care units, being Candida spp. reported among the ten leading causes of

bloodstream infections in Europe and in the USA. Candida sepsis has proven to be a

life-threatening infection with high prevalence and crude morbidity and mortality rates.

A rapid identification of the microorganism leads to an efficient intervention and a

significant reduction of the intake of broad spectrum antibiotics, with results in overall

cost savings. Peptide nucleic acid fluorescence in situ hybridization (PNA-FISH) is

already used as a highly precise and sensitive molecular method for microbial

identification. However, the entire procedure can take 1-2 days, since an enrichment

step of blood culture is necessary before applying the technique. Microfluidic devices

appear as a solution to reduce the time of the diagnosis by just consuming the time of

the PNA-FISH methodology, without wasting it on sample enrichment.

Therefore, the aim of this work is to develop a microfluidic system capable of retain

yeasts from a sample in a specific place within, increasing this way the sample

concentration, so the PNA-FISH procedure can be performed and the detection of the

pathogen can be made directly. Saccharomyces cerevisiae, a generally recognized as

safe (GRAS) microorganism, was chosen as a surrogate microorganism, in order to

avoid the contact with pathogens during this study. The construction of the microfluidics

system comprised micro-channels produced in polydimethylsiloxane (PDMS) using soft

lithography and a fluid injection system (including tubing, needles and a syringe pump).

To monitor the yeast retention, an optical microscope was used for the experiments of

flow visualization and to assess the hybridization an epifluorescence microscope

equipped with a red filter, sensitive to the fluorochrome attached to the probe, was

used.

PNA-FISH protocol was optimized in terms of fixation, temperature of hybridization

and the composition of hybridization solution. Paraformaldehyde/ethanol fixation, 54ºC

and a simplified hybridization solution without formamide was set as the optimal

procedure. Several microchannel geometries were conceived and a CFD simulation for

each one was performed in order to evaluate the best option. In some cases, the

exclusion of possibilities was made by geometrical characterization of the channels or

experimental microfluidics. Finally, the PNA-FISH was performed within the

microchannel with limited success.

In the future, deeper optimizations in the hybridization process within the microfluidic

device and the flow rate and yeast concentration need to be also optimized in order to

Abstract

vi

increase, as much as possible, the efficiency of the retention process. Additionally, the

time of hybridization and washing must be reduced without compromising the output.

Nonetheless, a microchannel geometry that makes feasible the rapid identification of S.

cerevisiae inside a microchip was achieved.

Resumo

vii

Resumo

A frequência de infeções fúngicas invasivas causadas por leveduras tem

aumentado nas unidades de cuidados intensivos, estando as espécies de Candida

reportadas entre as dez principais causas de infeções sistémicas na Europa e nos

EUA. A sepses tem proado ser uma infeção que ameaça a vida com elevada

prevalência e com índices de morbidade e mortalidade brutos. Uma identificação

rápida dos microrganismos leva a uma intervenção eficiente e a uma redução

significativa da toma de antibióticos de largo espectro, com resultados numa poupança

global. A técnica de hibridação in situ fluorescente com utilização de ácidos péptido-

nucleicos (PNA-FISH) é já usada como um método molecular para identificação de

microrganismos. No entanto, o processo completo pode levar 1-2 dias, uma vez que é

necessário um passo de enriquecimento da cultura sanguínea, antes de aplicar a

técnica. Os dispositivos de microfluídica aparecem como a solução para reduzir o

tempo do diagnóstico para unicamente o consumo do tempo da metodologia do PNA-

FISH, sem o gastar no enriquecimento da amostra.

Portanto, o objetivo deste trabalho é desenvolver um sistema de microfluídica

capaz de reter leveduras de uma amostra num local específico no seu interior,

aumentando, desta forma, a concentração da amostra, para que o procedimento de

PNA-FISH possa ser realizado e a deteção dos patogénios possa ser feita

diretamente. Saccharomyces cerevisiae, um microrganismo geralmente reconhecido

como seguro (GRAS), foi escolhida como o microrganismo substituto, para evitar o

contacto com patogénios em estudos preliminares. A construção destes sistemas de

microfluídica compreendem a produção de microcanais em polidimetilsiloxano

(PDMS), usando a litografia suave e um sistema de injeção de fluidos (incluindo

tubagem, agulhas e uma bomba de seringa). Para monitorizar a retenção das

leveduras, um microscópio ótico foi usado para os ensaios de visualização de

escoamento e para avaliar a hibridação foi utilizado um microscópio epifluorescente

equipado com um filtro vermelho, sensível à sonda usada.

O protocolo de PNA-FISH foi otimizado em termos de fixação, da temperatura

de hibridação e da composição da solução de hibridação. A fixação de

paraformaldeído/etanol, 54ºC e uma solução de hibridação simplificada sem

formamida foi estabelecida como ótima. Várias geometrias de microcanais foram

concebidas e foi realizada uma simulação CFD para cada, com o objetivo de avaliar a

melhor opção. Em alguns casos, foi feita a exclusão de possibilidades pela

caracterização geométrica do canal ou pela microfluídica experimental. Finalmente,

Resumo

viii

procedeu-se ao PNA-FISH dentro do microcanal.

Os resultados são promissores para a obtenção de bons resultados, utilizando

o método proposto, mas precisam de otimizações profundas no processo de

hibridação no interior do dispositivo de microfluídica assim como a concentração de

leveduras precisam, também, de ser otimizados, com o objetivo de aumentar, tanto

quanto possível, a eficiência do processo de retenção. Adicionalmente, o tempo de

hibridação e lavagem deve ser reduzido, sem comprometer o resultado. Conclui-se

que foi encontrada uma geometria de microcanal que torna fazível a identificação

rápida de S. cerevisiae dentro do microchip.

Contents

ix

Contents

Acknowledgments .......................................................................................................... iv

Abstract ............................................................................................................................ v

Resumo ........................................................................................................................... vii

Contents .......................................................................................................................... ix

List of figures .................................................................................................................. xi

List of tables ................................................................................................................... xii

Acronyms ...................................................................................................................... xiii

Chapter I – Introduction .................................................................................................. 1 1. Framework ...................................................................................................................... 2 2. Aim ................................................................................................................................... 3 3. Thesis Outline ................................................................................................................ 3

Chapter II - Literature review .......................................................................................... 5 1. Candida spp. .................................................................................................................. 6

1.1 Microbiology and pathogenesis ............................................................................... 6 1.2 Epidemiology and treatment ..................................................................................... 9

2. Diagnosis of Candida spp. ......................................................................................... 12 2.1 Phenotypic methods ................................................................................................ 12 2.2 Molecular Techniques ............................................................................................. 13

3. Fluorescence in situ hybridization ............................................................................. 14 3.1 Brief historical overview .......................................................................................... 17 3.2 Nucleic acid mimics ................................................................................................. 18 3.2.1 Peptide nucleic acid (PNA).................................................................................. 19

4. Microfluidics for miniaturizing biological assays ..................................................... 20 4.1 Entrapment strategies ............................................................................................. 20 4.2 Microfabrication ........................................................................................................ 22 4.2.1 Photolithography (SU-8) ...................................................................................... 22 4.2.2 Xurography ............................................................................................................ 23 4.2.3 Soft Lithography .................................................................................................... 24 4.2.3.1 PDMS .................................................................................................................. 25 4.2.4 Other techniques ................................................................................................... 26 4.3 Concepts in fluid mechanics ................................................................................... 27 4.4 Liquid handling systems .......................................................................................... 29

5. Computational fluid dynamics .................................................................................... 31 5.1 Finite volume method .............................................................................................. 33

Chapter III – Materials and methods ........................................................................... 35 1. Yeast characterization ................................................................................................ 36

1.1. Surrogate microorganism selection, yeast strain and culture maintenance . 36 1.2 S. cerevisiae growth ................................................................................................. 36

2. PNA-FISH protocol optimization ............................................................................... 37 2.1 Probe sequence ....................................................................................................... 37 2.2 Hybridization conditions optimization .................................................................... 37 2.3 Fixation conditions optimization ............................................................................. 38 2.4 Assessment of dextran sulfate influence on hybridization .................................. 39

Contents

x

3. Material preparation .................................................................................................... 40 3.1 Microchip design ...................................................................................................... 40 3.1.1 Production of molds – channels with a dam-like structure .............................. 40 3.1.2 Production of molds – channels with pillar-based structure ............................ 41 3.2 Microfabrication of microchannels .......................................................................... 41

4. CFD simulation ............................................................................................................ 42 5. Characterization of the microchannels ..................................................................... 43

5.1 Geometrical characterization of the microchannels ............................................ 43 5.2 Volumetric measurement of channels capacity ................................................... 44 5.3 Material testing ......................................................................................................... 44

6. Microfluidic channels retention test........................................................................... 44 7. PNA-FISH methodology within the microfluidic channel ....................................... 44 8. Microscopic visualization ............................................................................................ 46

Chapter IV – Results and Discussion ......................................................................... 48 1. S. cerevisiae growth .................................................................................................... 49 2. PNA-FISH protocol optimization ............................................................................... 51 3. Microchannels geometry design ............................................................................... 58

3.1 Channels with a dam-like structure ....................................................................... 59 3.2 Channels with a pillar-based filter .......................................................................... 59 3.2.1 Single channel devices ......................................................................................... 60 3.2.2 Two inlet devices ................................................................................................... 62

4. Computational fluid dynamics analysis .................................................................... 64 5. Retention tests in channels with a dam-like structure ........................................... 67 6. Geometrical characterization of channels with pillar-based filter ......................... 69 7. Retention tests in single channels with pillar-based filter ...................................... 72 8. PNA-FISH methodology within the microfluidic channel ....................................... 74

Chapter V – Conclusions and Future Work ................................................................ 77 Conclusions ........................................................................................................................... 78 Future work ........................................................................................................................... 79

References ..................................................................................................................... 80

List of Figures

xi

List of figures Figure 1| Morphologies assumed by Candida albicans ____________________________________________________ 6 Figure 2| Candida albicans tissue invasion ________________________________________________________________ 7 Figure 3| Schematic of cell biology of yeast, pseudohyphal and hyphal development ___________________ 8 Figure 4| Fluconazole drug resistance by Candida species and country ________________________________ 11 Figure 5| CHROMagar Candida ___________________________________________________________________________ 13 Figure 6| API Candida System strips ______________________________________________________________________ 13 Figure 7| Basics steps of FISH ______________________________________________________________________________ 15 Figure 8| Chemical structure of DNA, RNA LNA and PNA ________________________________________________ 18 Figure 9| Examples of microfilters _________________________________________________________________________ 21 Figure 10| Schematic of photolithography technique ____________________________________________________ 23 Figure 11| Xurography technique__________________________________________________________________________ 24 Figure 12| Schematic of soft lithography technique ______________________________________________________ 24 Figure 13| Interconnected channels _______________________________________________________________________ 25 Figure 14| Structure of PDMS polymer chain _____________________________________________________________ 26 Figure 15| Motion of cylindrical fluid element within a pipe _____________________________________________ 28 Figure 16| Schematic of equipment network for pressure driven laminar flow _________________________ 30 Figure 17| Flow focusing structure ________________________________________________________________________ 30 Figure 18| Steps in CFD simulation ________________________________________________________________________ 32 Figure 19| Bi-dimensional finite volume P and its neighbors. ____________________________________________ 33 Figure 20| Construction of the microfluidic device with a dam-like structure with 8-µm gap __________ 41 Figure 21| Cross section of the microchannel _____________________________________________________________ 43 Figure 22| Schematic illustration of the concept to perform FISH assay in the microfluidic device ____ 46 Figure 23| Saccharomyces cerevisiae growth curve ______________________________________________________ 50 Figure 24| Saccharomyces cerevisiae calibration curve __________________________________________________ 51 FIGURE 25| Fluorescence microscope results for in suspension PNA-FISH with different hybridization solutions at different temperatures _________________________________________________ 53 Figure 26| Fluorescence microscope results for in suspension PNA-FISH with simplified hybridization

solution (0% formamide) at 54˚C __________________________________________________________________________ 54 Figure 27| Fluorescence microscope results for in suspension PNA-FISH simplified hybridization

solution (0% formamide) at 54˚C after different fixation protocols _____________________________________ 56 Figure 28| Fluorescence microscope results for in suspension PNA-FISH simplified hybridization solution (0% formamide) at 54˚C with presence or absence of dextran sulfate. _________________________ 58 Figure 29| Mask structure __________________________________________________________________________________ 59 Figure 30| Geometries of cell-filtering single microfluidic devices _______________________________________ 60 Figure 31| Geometries of enlarged cell-filtering microfluidic devices ____________________________________ 61 Figure 32| Geometries of two streams cell-filtering microfluidic devices ________________________________ 63 Figure 33| Geometries of three streams cell-filtering microfluidic devices ______________________________ 63 FIGURE 34| Velocity contours and streamlines in the conceived geometries for microfluidic devices along the horizontal plane of symmetry as predicted by the CFD simulation

performed __________________________________________________________________________________________________ 66 FIGURE 35| Contour plot in a vertical plane in the middle of the horseshoe structure from Simulation X _______________________________________________________________________________________________ 67 Figure 36| Retention test in channels with a dam-like structure _________________________________________ 68 Figure 37| Usage of PS particles as filter for retention of S. cerevisiae in channels with a dam-like structure ____________________________________________________________________________________________________ 69 Figure 38| Geometrical characterization of a channel with a nominal height of 50µm ________________ 70 Figure 39| Geometrical characterization of a channel with a nominal height of 30µm ________________ 70 Figure 40| Channel with diamond-like pillars exhibiting barrier imperfections ________________________ 71 Figure 41| Channel with 15x5 µm rectangular pillars exhibiting barrier imperfections _______________ 71 Figure 42| Channel with pyramidal structures ____________________________________________________________ 72 Figure 43| Microspheres retention test for channels with pyramidal structures ________________________ 73 Figure 44| S. cerevisiae retention test for channels with pyramidal structures _________________________ 74 Figure 45| Channel with pyramidal structures ____________________________________________________________ 75 Figure 46| Positive for PNA-FISH methodology within the microfluidic channel. _______________________ 75

List of Tables

xii

List of tables

Table 1| Risk factors for development of invasive Candidiasis ____________________________________________ 9 Table 2| Probe thermodynamic parameters177 ____________________________________________________________ 37 Table 3| Comparison of different fixation protocols ______________________________________________________ 39 Table 4| Characteristics of each 2D mesh ____________________________________________________________ 43

Acronyms

xiii

Acronyms

BSI – Bloodstream infection

CFD – Computational fluid dynamics

DAPI – 4',6-diamidino-2-phenylindole

FISH – Fluorescence in situ hybridization

ICU – Intensive care unit

LNA – Locked nucleic acid

MALDI-TOF – matrix-assisted laser desorption ionization – time of flight

MEMS – Microelectromechanical system

MRI – Magnetic resonance imaging

MS – Mass spectrometry

NAC – Non-albicans Candida

OD - Optical Density

PC – Polycarbonate

PCR – Polymerase chain reaction

PDMS – Polydimethylsiloxane

PE – Polyethylene

PMMA – Polymethylmethacrylate

PNA – Peptide Nucleic Acid

PVC – Polyvinylchloride

Re –Reynolds number

2’-OMe – 2’-O-methyl nucleic acid

3D – Three dimensional

Chapter I – Introduction

1. Framework

2. Aim

3. Thesis Outline

Integration of microfluidics and FISH for the rapid identification of microorganisms

2 Daniela Cruz Moreira

1. Framework

The World Health Organization (WHO) has recently warned against what can be the

entry into the ominous “post-antibiotic” era, in its first global report on antibiotic

resistance in microoganisms1. This first-of-its-kind report has reached prominence in

the press and rang the alarm bell regarding the need to start a global effort to fight

against emerging resistance. According to the information included in the report, this

public health threat can be tackled essentially by avoiding the overuse of antibiotics,

preventing the quick proliferation of drug-resistant pathogens. In spite of not being a

new problem, modern medicine is now facing an apocalyptic scenario in which

common infections can kill, and healthcare professionals should act in order to use

antibiotics effectively. The consciousness of antimicrobial-resistant emergence in every

part of the world provides the kick-start needed to the urgent and consolidated

development of rapid and effective diagnostic tools1.

Identification of microorganisms causing sepsis is a determinant outcome because it

can lead to an appropriate antimicrobial therapy, instead of an empiric broad-spectrum

antibiotic treatment. Prior antibiotic use in patients with fungal infection is a risk factor

since it will not lead to fungus eradication but to its opportunistic spread. Rapid

diagnostic methods can revolutionize modern medicine by avoiding those current

misdirected practices that can worsen the patient’s condition and also by softening the

selective pressure of intense antibiotic use that promotes antibiotic resistance.

Molecular techniques require prior enrichment the biosample to reach detectable levels

of target cells. These assays are generally traditional cultural techniques, that can take

several hours. Miniaturization technology provides facilities for creating tools with

feature sizes matching the dimensions of cells. A particular cell is fixed and labeled by

its position, enabling high spatial resolution. Bypassing the enrichment step, the time

consumed during the entire procedure can be reduced to the time of the technique

alone.

Diagnosis delay also implies a heavy and unnecessary economic burden for the

hospital, due to treatments inefficiency, increased patient’s hospitalization time and

resources2. Despite the progress and improvement of hygiene and sanitary conditions,

hospital-acquired infections, also called nosocomial infections, are a major cause of

morbidity and mortality3. This fact is due to the hospital environment, that facilitates the

transmission of microorganisms among patients. Bloodstream infections (BSI) are

frequent in intensive care units (ICU), essentially because of immunocompromised

patient’s vulnerability and their prolonged hospital stay4. Although Gram-positive

CHAPTER I - INTRODUCTION

Daniela Cruz Moreira 3

bacterial pathogens persist as the most common cause of sepsis, fungal organisms

infections are increasing rapidly5,6.

Development of rapid diagnostic technology can be the golden key to positively impact

the economic and social issues described above, considering that can provide faster,

cheaper and more precise tools than time-consuming procedures available nowadays7.

Raising new advances in the diagnosis of infectious diseases is the first step to

generate significant progresses in healthcare.

2. Aim

This work is part of a project, whose ultimate goal is the conception of a fully integrated

and automated lab-on-a-chip sensor platform capable of being clinically implemented

for point-of-care pathogen detection where the sensing relies on nucleic acid

hybridization.

The aim of this work is the development of a microfluidic filter-based device to achieve

rapid detection of yeast cells in which peptide nucleic acid fluorescence in situ

hybridization (PNA-FISH) assays can be performed. The low number of cells present in

clinical samples implies an enrichment step before the PNA-FISH procedure can be

carried out. This system should allow the effective entrapment and concentration of

non-adherent yeast cells in defined locations, in order to bypass the prior enrichment

step required by standard FISH. All geometries developed should be conceived to trap

Candida spp., the most common cause of fungal infection worldwide5. Although

Candida spp. is the targeted microorganism, a morphologically similar non-pathogenic

surrogate microorganism must be carefully chosen to facilitate the testing phase of the

product.

This represents one of the first attempts to integrate PNA-FISH procedures applied to

microorganisms into miniaturized fluidic devices.

3. Thesis Outline

This thesis is composed by 6 chapters. On this chapter (Chapter I), a brief introduction

to the work in question is presented, and the framework and description of the thesis

structure is mentioned, to allow a better understanding of each chapter’s content.

Chapter II is dedicated to the literature review, where concepts of microbiology and

epidemiology of Candida spp., fluorescence in situ hybridization, microfluidics,



computational fluids dynamics (CFD) and current available methods in the market for

microorganism infections are discussed

Chapter III describes the experimental methodology, providing the necessary

information to replicate the procedures. . An introduction about the microorganism that

was selected as a Candida surrogate, S. cerevisiae, is also presented.

Chapter IV presents the experimental results obtained.

Chapter V discusses the results in detail.

Finally, chapter VI summarizes the main conclusions of this thesis and presents an

outlook for future work.

Chapter II - Literature review

1. Candida spp.

2. Diagnosis of Candida spp.

3. Fluorescence in situ hybridization

4. Microfluidics for miniaturizing

biological assays

5. Computational fluid dynamics



1. Candida spp.

1.1 Microbiology and pathogenesis

Mammalian mucosal surfaces and skin harbor millions of microorganisms that live as

benign commensals. Most of these microorganisms are beneficial to their host but a

few of them are considered opportunistic pathogens since they are capable of causing

illness when they cross the host’s protective barriers and colonize internal organs8.

Candida spp. are naturally present in the skin and the mucosal surfaces of the oral

cavity, gastrointestinal and urogenital tracts and vagina of the human host8. The genus

Candida belongs to the Fungi kingdom, class of Deuteromycetes, and includes several

species such as yeasts with ascomycetous and basidiomycetous affinities9. Some of

them can assume a variety of cell morphologies, as Figure 1 shows, ranging from

yeast-like cells (blastospore) to a variety of elongated growth forms, including germ

tubes (transition between yeast and hyphae), hyphae and pseudohyphae10. Only

Candida albicans and Candida dubliniensis are able to form true hyphae and these two

species are considered to be polymorphic11. This phenotypic plasticity is crucial to C.

albicans pathogenesis12,13, since hyphal cells may promote tissue invasion (Fig. 2),

whereas yeast cells facilitate dissemination of the pathogen (see below)14,15.

The yeast form is the one most commonly found in the laboratory because the

transition to elongated growth has to be stimulated by a wide range of environmental

conditions that mimics the microenvironment that it encounters in the host10. Hyphal

growth can be experimentally induced, at 37˚C (physiological temperature), by the

presence of serum16; a neutral pH17; 5% CO2 partial pressure, experienced in the

bloodstream18 and N-acetyl-D-glucosamine19. There are two synthetic growth media

that often induce hyphal growth due to the presence of amino acids20 or mannitol as a

carbon source21.

FIGURE 1| Morphologies assumed by Candida albicans

Candida albicans can assume a variety of morphologies: Yeast (Y), pseudohyphal (P) or hyphal (H) forms10 .

CHAPTER II – LITERATURE REVIEW


FIGURE 2| Candida albicans tissue invasion

The several steps in tissue invasion by Candida

albicans, for a stylized epithelial cell surface:

adhesion to the epithelium; epithelial penetration

and invasion by hyphae; vascular dissemination,

which involves hyphal penetration of blood vessels

and seeding of yeast cells into the bloodstream; and,

finally, endothelial colonization and penetration

during disseminated disease22

.

The growth behavior of yeast is also similar to bacteria. Yeast cells display a lag

phase prior to an exponential period of division. As some nutrients become depleted,

the increase in cell number slows and stops. If refrigerated in this stationary phase,

cells can remain alive for months23.

Figure 3 shows fundamental differences in the cell cycle progression of a yeast,

pseudohyphal and hyphal cells. It is observed that what is common in all types is the

evagination of a single germ tube from the mother cell, that exhibit highly polarized

growth. First, the fixation site of the septin ring24,25 and the nucleus migration26,27 of

yeast (Figure 3a) and pseudohyphal (Fig. 3b) cells are very similar to each other but

different from hyphal cells28 (Fig. 3c). The main difference among these cells types are

related to cytokinesis. In budded yeast cells results on a constriction between adjacent

cells until they separate and both reenter the cycle29. In pseudohyphae, daughter cell

remained associated with the mother cell in both types but the attachment is easily

disrupted by mechanical agitation30. In hyphae, cytokinesis does not result in any

constriction, but in the formation of a primary and secondary septum, and the two

daughter compartments remain firmly attached to each other, being hyphae shape a



sparsely branched tube-like structure29. After hyphal cytokinesis, the subapical

compartment of hypha becomes highly vacuolated and remains in G1 and the apical

compartment is continuing dividing.

FIGURE 3| Schematic of cell biology of yeast, pseudohyphal and hyphal development

(a) Yeast cell. As the mother cell germinates, septin bars (orange) are visible at the germ tube base, as well as

a septin cap at the tip (red). The nucleus (blue) moves to the neck, assisted by astral microtubules (green), followed

by division across the neck. (b) Pseudohyphal cell. Similar features to yeast cells are observed in the beginning of

the cycle. The main exception is that cells do not separate following by cytokinesis, and the polarisome persists for

longer. (c) Hyphal cell. As the germ tube extends, the septin bars disappear and a septin ring appears in the hypha.

The nucleus migrates out of the mother cell to undergo mitosis in the germ tube. Septum (yellow) is formed

between the two rings of septin. The apical compartment remains in the cell cycle and the subapical remains in G1

phase29

.

The alteration of cell wall proteins composition during the yeast-to-hypha transition

that triggers a potent defense mechanism. The human host mucosa reveals

immunological tolerance to colonizing microorganisms but when C. albicans infects the

host, hypha specific cell wall proteins are major adhesins and invasins that function as

antigens capable of modulate immune response31.

The disease can be caused by the toxins that pathogen releases in the host

organism, or due to the direct destruction of living tissues23. Disseminated candidiasis

is fatal in immunocompromised individuals. Kidney and brain are the primary target

organs of this yeast during infection. Within the kidney, massive fungal invasion and

growth can occur, resulting in inflammatory reactions that lead to tissue necrosis32.



1.2 Epidemiology and treatment

Candida spp. are considered the most frequent fungi in the hospital. The frequency of

invasive fungal infections caused by yeasts has increased in intensive care units (ICU),

being Candida reported among the ten leading causes of bloodstream infections (BSI)

in Europe and USA33,34. Fungal infections in immunocompromised patients are

increasing and are associated with immunosuppression, illness, prematurity, exposure

to broad-spectrum antibiotics and patients of older age35 (Table 1). Fungal infections

may be superficial, such as in skin, hair and nails; subcutaneous and systemic,

affecting several internal organs, blood, and internal epithelia. Candida sepsis has

proven to be a life-threatening infection with high prevalence and high morbidity and

crude mortality rates36,37.

Table 1| Risk factors for development of invasive Candidiasis38

Risk factors for development of invasive Candidiasis

1. Length of stay in the ICU

2. Use of central venous catheters

3. Use of broad-spectrum antimicrobial agents

4. Parenteral nutrition

5. Development of acute renal insufficiency,

hemodialysis

6. Severe pancreatitis

7. Diabetes mellitus

8. Neutropenia

9. Gastrointestinal tract surgery

10. Solid organ or stem cell transplantation

11. Candida colonization

There are about 150 species of Candida, but only a small number are human

pathogens. C. albicans colonizes the mucus membranes of 30-60% on humans39 and

is the cause of vaginal infections, diaper rash in infants, and thrush in the mouth and

throat. The infections caused by C. albicans affect mostly immunocompromised

patients that suffer from acquired immunodeficiency syndrome, for instance8. C.

albicans also invades the brain during acute infections and causes



meningoencephalitis40. Navarathna et al.41 recently confirmed in vivo, using magnetic

resonance imaging (MRI), that the integrity of the blood-brain barrier is lost during C.

albicans invasion. Although C. albicans is the most prevalent species involved in

invasive mycoses, the incidence of infections due to non-albicans Candida (NAC)

species is increasing42. The most common NAC species are C. tropicalis, C. glabrata,

C. krusei and C. parapsilosis, which as a group, represents about one-half of all

Candida spp. isolates from blood cultures43.

C. tropicalis infections are usually preceded by a band of progressive necrotic tissue,

which is not observed with C. albicans44. Moreover, infections due to C. tropicalis have

increased dramatically on a global scale because of its resistance to fluconazole. C.

tropicalis candidaemia has been associated with cancer, especially in patients with

leukemia or neutropenia45.

Candidaemia due to C. glabrata has been reported as related to the use of

fluconazole46. C. rugosa has been described in the oral cavity of diabetic patients47.

C. parapsilosis has emerged as a significant nosocomial pathogen with manifestations

in endophthalmitis, endocarditis, septic arthritis, peritonitis and fungaemia, usually

associated with invasive procedures48.

In Europe, in 2006, more than half of the cases of candidaemia were caused by C.

albicans, and the incidence rates for NAC infections were 14% each for C. glabrata and

C. parapsilosis, 7% for C. tropicalis and 2% for C. krusei37. In North America a

predominance of NAC species was observed, although C. albicans was the most

frequently isolated species, followed by C. glabrata38. Latin American Countries

showed changes in prevalence of C. albicans, and a progressive increase of NAC

infections. C. parapsilosis was the most frequent species, followed by C. tropicalis and

C. glabrata49. According to the Brazilian Network Candidaemia Study, C. albicans

accounted for 40.9% of cases in Brazil, followed by C. tropicalis (20.9%), C.

parapsilosis (20.5%) and C. glabrata (4.9%)47,49. In a four years retrospective study in a

Portuguese hospital it was observed that C. albicans was the most frequently yeast,

with an overall incidence of 69.9% followed by C. glabarata, C. tropicalis, C.

parapsilosis and C. krusei50.

Response to antifungal therapy differs for each Candida spp.. Currently, there are

only three classes of antifungal agents available to treat Candida infections: azoles,

echinocandins and the polyenes51. The main reason to discriminate between different

Candida species is to proceed with an effective treatment, especially because there are

some species resistant to some antifungals like azoles, such as fluconazole, the

standard antifungal drug of choice in many countries52,48. Fig. 4 shows resistance rates



against fluconazole for Candida albicans, non-C. albicans, and all Candida isolates

combined in selected countries from which data are available1. Significant cost, toxicity

and absence of an oral formulation can present barriers to the use of drugs to treat

azole resistant Candida infections.

FIGURE 4| Fluconazole drug resistance by Candida species and country

A rapid microorganism identification leads to an efficient intervention, a significant

reduction of the intake of broad spectrum antibiotics with an overall cost savings. The

early diagnosis is determinant in critically ill patients because the indiscriminate intake

of antibiotics usually kills bacteria that are essential for normal digestion and favors the

opportunistic spread of Candida spp. on the digestive tract walls, which can be

worsened when associated with a diet rich in sugars and carbohydrates53,54. Once

colonizing the intestinal walls, the fungus is very difficult to eradicate and infection

treatment is usually followed by recurrence23. According to a review from intensive care

units in USA and Canadian cities, the risk of death from sepsis increases by 6 to 10%

with every hour that passes from the onset of septic shock until the start of effective

antimicrobial therapy55.



2. Diagnosis of Candida spp.

Invasive candidiasis is difficult to diagnose. Tissue sampling has a low diagnostic

yield in patients who have received empirical therapy. Cultures other than blood (e.g.

bronchoalveolar lavage samples) sometimes reflect colonization instead of invasion

and blood cultures have a great incidence of false negatives56.

Several methods have been developed for the identification of Candida spp.. These

methods can be based on phenotypic aspects using culturing methods or can be

molecular-based methods. General insights regarding each of these methods will be

hereafter discussed in detail.

2.1 Phenotypic methods

Phenotypic methods allow the identification of the microorganism based on

observable characteristics such as morphology, development, and biochemical or

physiological properties resulting from gene expression57. Phenotypic methods use

culturing methods such as dextrose agar medium, germ-tube test, chlamydospore-

inducing media-based test, fermentation test, and CHROMagar test.

A dextrose agar medium can be used and a direct observation of yeast colonies on

slants or under the microscope is performed by an observer which leads to a subjective

reliance on fungal morphological aspects58. The germ-tube test is considered a simple,

economical and efficient procedure to differentiate C. albicans from other Candida spp.

and it is based on the observation of germ tubes formed by C. albicans59. This

technique, nevertheless, can lead to a misidentification since some NAC species such

as C. dubliniensis, C. tropicalis, and C. parapsilosis can also generate germ tubes or

pseudohyphae60. Cornmeal and rice extract agar are the best-known chlamydospore-

inducing media. Chlamydospore formation is a peculiarity of C. albicans, C.

dubliniensis and more rarely of C. tropicalis. The chlamydospore is a highly refractile,

thick walled, asexual fungal spore that is derived from of hyphal cell and can function

as a resting spore.61 The difference between C. dubliniensis chlamydospore and those

formed by C. albicans is that in C. dublinienses they are attached in pairs or larger

clusters. However, the method may not be consistent and may require additional

tests.62 Another method to differentiate C. albicans from C. dublienensis is the

fermentation test, based on the demonstration of acid or carbon dioxide production that

occurs in liquid media63. CHROMagar Candida (Fig. 5) is a plate-based test for the

simultaneous isolation and identification of various Candida species. This medium is



useful to differentiate the species C. dubliniensis, C. albicans, C. krusei, C. tropicalis

and C. glabrata based on colour development64. API Candida system (bioMérieux,

Marcy l’Etoile, France) (Fig. 6) is based on the detection of enzyme reactions and

acidification of sugars. The reactions are visually interpreted by spontaneous color

changes.65

FIGURE 5| CHROMagar Candida

Different isolations of Candida spp. appear in different colors. C. albicans, C.

tropicalis and C. krusei appea as green, blue and pink, respectivelya.

FIGURE 6| API Candida System strips

The system consists on a strip of tubes, which allows

identification after 18-24h of incubation at 35ºCb

2.2 Molecular Techniques

Conventional culture-based methods are mostly employed because of their user-

friendliness and low cost. However these methods are time-consuming and do not give

promptly useful results59. In addition, in critically-ill patients, to whom pre-emptive

therapy is administered, the presence of fastidious microorganisms, that sometimes

are in low densities, can lead to false negative cultures66. Molecular identification

methods have been used as suitable alternatives due to their high accuracy, sensitivity

and specificity for the identification of microorganisms 67. From these, the most

frequently used are the polymerase chain reaction (PCR)-based methods.

PCR-based methods are based on the amplification and detection of microbial nucleic

acid directly from clinical samples. This method comprises repeated DNA denaturation

a Source:http://www.frilabo.pt/fcms//index.php?option=content&task=view&id=172 b Source: http://www.biomerieux- diagnostics.com/servlet/srt/bio/clinical-diagnostics/dynPage?open=CNL_CLN_PRD&doc=CNL_PRD_CPL_G_PRD_CLN_11&pubparams.sform=5&lang=en



cycles, annealing of two oligonucleotide primers to the target DNA, flanking a region of

interest, and extension of a new strand by the DNA polymerase enzyme. The

synthesized sequence in one cycle can serve as template in the next and the number

of target DNA copies doubles approximately at every cycle68.

Multiplex-PCR combines different species-specific primers in a single PCR tube.

Carvalho et al. 69, described a method based on the amplification of two fragments from

ITS1 and ITS2 regions that made possible the identification of eight Candida spp. with

high specificity. Nested PCR consists of two primers sets for specific amplification of

Candida spp. DNA that are used in two successive PCR runs to improve specificity of

Candida identification70. The TaqMan assay (Perkin-Elmer Corp. Applied Biosystems,

Foster City, CA, USA) is a real-time procedure that can also be applied to identify

Candida spp.. Guiver et al. 71, used this method for the rapid identification of clinically

relevant Candida spp. within 4-5h.

Non PCR-based methods identify specific genotypes, without needing DNA

amplification. For instance, matrix-assisted laser desorption ionization – time of flight

mass spectrometry (MALDI-TOF MS) is based on protein fingerprints. Using Burker

Daltonic’s MALDI BioTyper system (Burker Daltonics, Bremen, Germany), fungi are

submitted to a thermal degradation, and the resulting small molecules are cleaved at

their frailest points, producing volatile fragments. Using mass spectrometry, a pyrolysis

mass spectrum is produced, which can be analyzed as a chemical fingerprint of the

yeast72. It is considered a rapid, reliable and cost-effective alternative for the

identification of Candida spp.73.

Another non-PCR-based method is fluorescence in situ hybridization (FISH) that will

be discussed in detail in the next chapter.

3. Fluorescence in situ hybridization

Fluorescence in situ hybridization (FISH) is a molecular method used to identify and

quantify microbial populations, attaching a fluorescent label to nucleic acids74,75. Three

essential steps illustrated in Fig. 7, fixation and permeabilization, hybridization and

washing, define FISH procedure76.



FIGURE 7| Basics steps of FISH

FISH follows three basic steps: 1| Cells are fixed and the cell membrane permeabilized. 2| The labelled probe

hybridize to the target rRNA sequence. 3| The unbound probe is washed away. Then the sample is ready for

visualization or quantification76

.

The fixation and permeabilization step is one of the most crucial steps assuring the

accuracy of detection protocols and is therefore decisive in determining the subsequent

success of a given experiment. An effective sample fixation preserves cellular

structures by rapidly ending all enzymatic and other metabolic activities maintaining the

cellular morphology77. An ideal fixative should reflect the in vivo situation by preventing

autolysis and stabilizing cellular structures. However, a compromise must be attained

since over fixation leads to fixation artifacts, loss of signal and increased nonspecific

background78.

Glutaraldehyde, a five-carbon dialdehyde, and formaldehyde, a one-carbon

monoaldehyde, are the most commonly used fixative agents, and act by cross-linking

proteins78. Glutaraldehyde is a harsh fixative due to its ability to polymerize and form

extended cross-linkings. However, formaldehyde is a small molecule that penetrates

quickly in cellular structures, stabilizing them and has proven to be the most effective

fixative for nucleic acids78. Paraformaldehyde is a polymerized form of formaldehyde,

and is preferred for most immunohistochemical procedures because it provides a

fixative free of extraneous additives79.



In addition to fixation, it is also necessary to permeabilize the cell wall, so probes can

diffuse and attach to intracellular target molecules. Some of the more commonly used

permeabilization agents are used after fixation, but some protocols suggest fixing and

permeabilizing cells at the same time80. Cell membrane permeabilizing agents can be

divided in two major groups: organic solvents and detergents. Alcohols, such as

methanol and ethanol, and acetone are harsh organic solvents that dissolve lipids from

the cell wall and also precipitate proteins and carbohydrates79. These reagents have

the advantage of performing rapid one-step fixation at low temperatures and they are

also considered permeabilizing agents. However, shrinkage of the sample can occur

due to simultaneous dehydration and fixation81. At low temperatures (0 to -20ºC)

ethanol precipitate proteins without denaturing them81.

Detergents are used especially after fixation with cross-linking agents. Triton-X is a

non-ionic detergent with uncharged, hydrophilic head groups of polyoxyethylene

moiety. It solubilizes phospholipids from the plasma membrane but, as it is non-

selective, may extract proteins along with lipids82. The importance of fixation step relies

on its osmotic and ionic effects on cellular structures, particularly organelles83.

Hybridization, the following step, is when the annealing between a fluorescently

labeled probe with the complementary target sequence occurs. In order to identify

microbial species, the preferred targets used are sequences in 16 or 23S rRNA that

comprises highly conserved regions, specific for each microorganism, being accepted

as discrimination criteria of microorganisms84. There are nearly 200 000 ribosomes in a

single Saccharomyces cerevisiae cell 85, and due to its abundance in fixed cells,

targeting rRNA specific sequences facilitate the visualization and quantification of

individual cells86. Nevertheless, one possible limitation of the method is associated with

probe inaccessibility to the target region within the secondary structure of ribosome,

which may hinder probe hybridization87. It is necessary to adequately design a probe

that targets sites located in the rRNA regions already known to be accessible. Other

limitations of this technique can be explained by decreased probe permeabilization and

low cellular ribosome content88. To guarantee that the probe accesses and hybridizes

with the target sequence, parameters like temperature, pH, ionic strength and

formamide concentration should be optimized74. When added to the hybridization

buffer, formamide act as a denaturant solvent that weakens the secondary structures of

rRNA89.

To avoid false positives, the washing step, as the name suggests, washes away

unbound or loosely bound probes. The sample is then ready for identification and



quantification by fluorescence microscopy or flow cytometry depending on the aim of

the analysis.

3.1 Brief historical overview

Watson and Crick’s discovery of the three-dimensional double-helice model for the

structure of DNA was a landmark on the history of genetic research 90. This discovery

marked a milestone in science and gave rise to modern molecular biology, yielding the

understanding of genes control, chemical processes within cells and the development

of powerful molecular techniques.

In 1961, Schildkraut et al.91 introduces the technique of DNA-DNA hybridization, used

to study genetic relationships among several species and boosted the first genome

studies92,93. Later in the same decade, Pardue and Gall94 show that hybridization can

be used to identify the position of a DNA sequence in situ, hybridizing a radioactive

labelled DNA probe to DNA within a cytological preparation. In 1977, Rudkin and

Stollar95 use fluorescent labeled probes instead of radioactive labeled probes

demonstrating the significant advantages over autoradiographic procedures, like ease

of use, safety, specificity, stability, diminished hybridization and microscopic analysis

times, sensitivity and resolution. The first application of fluorescence in situ

hybridization came in 1980, when a specific DNA sequence was hybridized with RNA

labeled with a fluorophore96. The following year, to overcome the difficulty in detecting

short segments of DNA, Harper and Saunders97 demonstrated the suggestion of Wahl

et al.98 that the use of dextran sulfate could accelerate the rate of hybridization. Langer

et al.99 incorporated enzymatically, amino-allyl modified bases throughout the length of

the RNA, which in turn conjugates to haptens or fluorophores allowing the production of

low-noise probes. Biotin covalently attaches to the nitrogenous base ring and then

avidin is conjugated. The interaction between biotin and avidin has a very high binding

constant and when avidin is coupled to indicator molecules, a small quantity of biotin

can be detected. It is considered an indirect detection since there are secondary

reporters that bind to the hybridization probe. Secondary detection by fluorescent

streptavidin conjugate assays were also performed for the detection of DNA and

mRNA100,101. In order to achieve a rapid phylogenetic identification of single microbial

cells, DeLong102 and co-workers described fluorescence in situ hybridization (FISH)

technique with rRNA-target probes.



Although traditionally used in FISH procedures, DNA probes suffers from limitations

related to cell permeability due to its negatively charged sugar-phosphate backbone

and the fact that the hydrophobic core of the lipid bilayer of the membrane is

impermeable to charged species. Other disadvantages of DNA-FISH are associated

with hybridization affinity and target site accessibility, leading to poor signal-to-noise

ratio and lack of target site specificity and sensitivity75. In the early 1990s, synthetic

single-stranded PNA probes were designed and synthesized, allowing direct detection

and circumventing DNA-FISH shortcomings103. In the next section, synthetic single-

stranded DNA mimic probes will be discussed in detail, with particular focus on peptide

nucleic acid (PNA) probes.

3.2 Nucleic acid mimics

FISH probes target specific sequences of nucleic acids by complementarity. Nucleic

acid analogues appear as a breakthrough to improve sensitivity of FISH. Different

types of nucleic acid mimics, which were reviewed by Cerqueira et al.74, have been

reported to potentially have application on FISH procedures, being the most common

peptide nucleic acid (PNA), locked nucleic acid (LNA) and 2’-O-methyl (2’-OMe) RNA

(Fig. 8).

FIGURE 8| Chemical structure of DNA, RNA LNA and PNA

A, B| DNA and RNA have a negatively charged sugar-phosphate backbone C| The ribose ring of LNA is locked to a

C3’ endo-conformation. D| PNA have a neutral polyamide backbone composed of N-(2-aminoethyl) glycine units 74

.



LNA resembles RNA, being the ribose ring locked to a C3’ endo-conformation. The

ribose is linked to a methylene bridge between 2’-Oxigen and 4’-carbon (Fig.8C)104.

LNA forms duplexes with complementary DNA and RNA. The most important

characteristics of this mimic for FISH implementation are the increased thermal stability

and improved selectivity74,105. 2’O-Me RNA is another modified oligonucleotide that

shows a high affinity towards an RNA106.

3.2.1 Peptide nucleic acid (PNA)

PNA probes contain the same nucleotide bases as DNA. However, the negatively

charged sugar-phosphate backbone of DNA is replaced by a non-charged polyamide

backbone composed of N-(2-aminoethyl) glycine units, as shown in Figure 8D.

An important aspect is the physical configuration of nucleobases in PNA, that are

positioned at the same distance from each other, similarly to what occurs in the DNA

molecular structure, enabling the hybridization by complementarity, still obeying the

Watson-Crick hydrogen bonding rules107.

PNA/DNA complementary bounds are stronger than DNA/DNA. Due to the lack of

charge of PNA molecules, electrostatic repulsions are not verified and the hybridization

is rapid, more stable and more specific108. However, it is important to bear in mind the

self-complementarity of the probe design since complementary PNA sequences

hybridize strongly109. Moreover, the neutrally charged nature of the molecule implies

relatively low aqueous solubility which means that PNA tends to aggregate.

Fortunately, solubility-enhancing modifications to the molecule are possible to make110.

The relatively smaller molecular size, about 15 base pairs, and the hydrophobic

nature of the peptide backbone facilitates the diffusion through the lipidic cell wall86.

Also, shorter sequences mean a higher destabilization by a single-base mismatch, and

an effective discrimination of single-base differences109.

PNA probes thermodynamic properties were studied by Ylmaz and Noguera111,

providing an essential tool to study the mismatches effect. Due to its synthetic

backbone, PNA revealed an extended lifespan because of its increased resistance to

nucleases and proteases, which is a great advantage in contrast to DNA probes112.

As described below, PNA probes own such characteristics that make it so favorable to

many molecular biology applications. However, the required enrichment steps that

some clinical samples need, to assure considerable number of cells in the sample, can

hinder PNA-FISH use for microorganism clinical diagnosis since it can be more time-



consuming. To overcome this difficulty, a promising solution appears as a lab-on-a-chip

device form, a miniaturized and automated system, where a significant time and cost

reduction is achieved. PNA-FISH adapted to fit microfluidic channels is believed to

become a new rapid and extremely effective tool in clinical diagnosis. For better

understanding the advantages of these two techniques (PNA-FISH and microfluidics)

integration, microfluidic systems are introduced in the next section.

4. Microfluidics for miniaturizing biological assays

Science has been driven to develop new technology and miniaturization has been a

central component. When applied to the analytical and diagnostic assays field, the

main goals of miniaturized devices are to lower the cost per test (because the amount

of materials and reagent consumption decreases by scaling down the assay volume),

to shorten the time to obtain a result, provide higher sensitivity, portability and to

decrease the amount of laboratory space needed113. In vitro diagnostics (e.g. point-of-

care and central laboratory-based testing) offer huge potential for microfluidics since is

possible to fabricate a hand-held device with low energy consumption and high

precision.

Microfluidics is the handling and analyzing of fluids in structures of micrometer scale,

namely microfluidic devices, which are a tool with an integrated channel where a fluid

circulates 114. This concept can be applied to obtain ideal conditions for well-defined

microenvironments and offers spatial control and manipulation of cells in microfluidic

systems115. The microdomain provides conditions for performing biological experiments

on cells and offers precise definition of compound concentrations in a confined space

with high resolution115. By implementing these systems to concentrate cell

suspensions, the time consuming enrichment step of standard procedures can be

bypassed.

4.1 Entrapment strategies

Moving cell-based studies into a microfluidic network means achieving new strategies

to retain cells at defined locations over time. Culturing cells in microfluidic environments

requires fundamental knowledge of multiple topics, including cell biology, cell culture

techniques, fluid dynamics and the physics of the microscale.

The strategy adopted to concentrate cell suspensions depends on the cell adherence

to the material116. Most of the experiments with bacteria inside microchannels do not



need specific structures for entrapment. Bacteria adhere to the substratum surface due

to hydrophobic interactions, stronger than repulsive forces, and include irreversible

interaction as hydrogen, ionic and covalent bonding117. In the case of non-adherent

cells a trapping mechanism is mandatory. The most commonly used technologies for

cell trapping in microfluidic systems are based on surface bound cell assays118,

localized surface modifications119 or chemical immobilization120. Current trends and

technologies make use of hydrodynamic121, electrical122, optical123, magnetic124 or

acoustic125 fields. However, incorporation of these principles in complex biomedical

microfluidic platforms requires a set of components that need to be combined together.

These components are expensive and are not widely available; therefore the general

trend in commercial devices has been to fabricate simple, disposable devices that are

designed to interface with external equipment that houses the required control

electronics, reagent supply, detectors and programming126.

Various techniques have been adopted to incorporate within the microdevice a

simple-to-obtain 3D architecture to confine micrometer-sized particles to specific

regions in a microfluidic network. These microfluidic filters (Fig. 9) can be microporous

membranes incorporated as one of the walls of a microfluidic channel127, arrays of

pillars arranged in a regular geometric pattern128 and porous beds distributed by

sizes129.

FIGURE 9| Examples of microfilters

a| Porous membranes Schematic of microfluidic blood filtration device, in which top and bottom pieces are made

up of PDMS and the membrane is a commercially-available filter129

. b| Arrays of pillars SEM image of pillar-type

microfilter on silicon. These pillars serve as filters with a well-defined size130

. c| Packed beds Schematic of

microfluidic filter of 20 and 10 µm beads, which inhibit the passage of cells out of the culture chamber preventing

blockage of the flow131

.

Micro and nanoporous membranes (Fig. 9a) may either be manufactured using

microfabrication techniques130 or commercial routinely used membranes may be

bought and incorporated into microfluidic systems. The integration is made by having

two separated open microchannels that are glued together sandwiching the

membrane127. Although the design has a high degree of flexibility relatively to pore size



distribution, the main drawbacks of the method are expensiveness and limitation to few

materials.

Arrays of pillars (Fig. 9b) are designed and fabricated along with the channels. Pillars

have well-defined gaps size that can be easily manipulated, considering the size

resolution of the fabrication process128. The ratio between channel depth and the size

of the space among pillars is a critical concern, because pillars are liable to buckle

when this ratio is increased too much. The main disadvantage of the method is that the

failure of even a few pillars may compromise the filtration process131.

Porous beads (Fig. 9c) consist of a large number of solid particles packed by sizes.

They are typically held in a dam-like structure, and beads are packed by size at high

pressure, maintaining a network of interconnected pores through which fluid is

driven129. The main advantage of this type of filters is that porous characteristics can be

rapidly modified with minimum changes in fabrication procedure.

4.2 Microfabrication

Microfabrication is how three dimensional (3D) microstructures are produced132. The

fabrication technique and material chosen are dependent on the application and are

fundamental to achieve the goal of the microfluidic device at a low cost and high

throughput. Microelectromechanical systems (MEMS) technologies, like molding and

embossing techniques, are the most cost-effective for mass production, since the fixed

cost associated with the expensive tools can be spread over a large number of

devices.

Replica molding process is used to replicate with high reproducibility topographical

features from a rigid or elastomeric mold into another material by solidifying a liquid in

contact with the original patterned master133. Several procedures are used to obtain a

master and their characteristics are determinant to choose which is the most

appropriate. The most popular techniques to produce microfluidic devices are

photolithography and soft lithography. Other techniques are xurography, hot imprinting

or embossing, injection molding or laser ablation134.

4.2.1 Photolithography (SU-8)

Photolithography is the most common technique to produce patterned masters with

high accuracy. The development of complex 3D geometries in microfluidic devices are



made possible through the implementation of photoactivated polymers, such as SU-

8135.

As illustrated by Figure 10, the SU-8 photolithographic process consists on spin-

coating a SU-8 photoresist on a silicon wafer and then soft-bake on a hot plate to

evaporate the solvent and harden the photoresist. The UV exposure onto a

photopatternable polymer, initiate a cross-linking reaction across exposed regions and

the pattern on the photomask is replicated into the photoresist. Excess polymer and

reactants need to be washed away leaving the microstructure136.

SU-8 was made a popular photoresist due to its harsh environmental durability and

lithographic contrast, that enables the production of high resolution structures for direct

molding137. An important feature of SU-8 prototyping is the ability to control layers

thickness and quality varying the solvent, which will be reflected on channel depth.

Also, tight control of structures width and height has elevated SU-8 to the photoresist of

choice for creating microchannels molds138. However, several baking sessions to

condition the photoresist and clean room facilities to reduce dust and particles that may

compromise film quality are required.

FIGURE 10| Schematic of photolithography technique

A silicon wafer is spin-coated with SU-8 photoresist and soft-baked on a hot plate to evaporate the solvent and

harden the photoresist; a photomask is aligned above the photoresist on silicon wafer and exposed to UV to

replicate the pattern into the photoresist; following agitation in SU-8 developer the unexposed photoresist is

removed; finally, the mold is rinsed and dried with clean air136

.

4.2.2 Xurography

Xurography is a photolithography replacement technique for rapid prototyping of

microstructures using a cutting plotter (Fig. 11). The mold is designed using a software



tool and the limits of the mold are cut on film with a cutting plotter. The surrounding film

is removed and the pattern is adherently attached to a glass substrate.

The control settings of the cutting plotter need to be considered, since cutting speed

and force influence the features precision. Cutting force has to be enough to

completely cut the film, but it only must be slightly scored. High cutting speed and a dull

blade are closely related to material tearing, resulting in rougher edges139,140.

Xurography is cheaper because it uses low-cost equipment and materials and can

rapidly prototype microfluidic devices for higher resolution devices141. In spite of being a

fast, effective and an easy method, dimensions obtainable with this technique can be

limiting because patterns with small dimensions not only are difficult to handle but also

depends on the resolution of the equipment142. Xurography proved to have poor

accuracy. Due to small dimensions of the structures, irregularities are critical and a

more precise technique has to be adopted to create the master.

FIGURE 11| Xurography technique

First the film is cut with a plotter blade and then unnecessary film around pattern is weed

145.

4.2.3 Soft Lithography

After mold fabrication, soft lithography (Fig.12) represents a non-photolithographic

strategy based on replica molding. An elastomeric stamp with patterned relief

structures is used to generate microstructures. In order to avoid damage of molded

structures, the molded part should have elastic properties so that demolding can be

performed under reversible deformation143. Usually a two-part polymer in a liquid form

is poured over the prefabricated mold at room temperature and cured at a slightly

elevated temperature. Once cured, the polymer is separated from the mold and access

points are made with a needle (Fig. 13) and finally sealed to a flat surface143.

FIGURE 12| Schematic of soft lithography

technique

Elastomer is poured on master, polymerized and removed

from the mold. Ports are pouched and the block is bond to a

glass slide previously spin-coated with a thin layer of

elastomer136

.



Polymer molding is a powerful tool for microfabrication, allowing the manufacture of

many transparent polymer chips from a single mold. The practicability of polymers for

fabrication with both rapid prototyping and mass production as well as lower cost

relative to silicon and glass make them particularly attractive144. Elastic properties of

elastomers have proven to reduce breakage of structures during demolding, due to

reversible deformation. Taking advantage of soft elastomers properties, whose Young’s

modulus makes it moderately stiff (~1MPa), micropatterns formed on the walls of

microfluidic channels have proven to enhance cell capturing and cell cultivation. A

variety of polymers are candidates for low cost, mass production of microfluidic

devices, such as polycarbonate (PC), polymethylmethacrylate (PMMA),

polyvinylchloride (PVC), polyethylene (PE), and polydimethylsiloxane (PDMS)145.

FIGURE 13| Interconnected channels

1| Using a twist motion, the needle punch the block from the outside of the PDMS to the buried channel 2| A hole is cut with a diameter identical to the inner diameter of the needle. 3| When the needle is removed, the hole diameter readjusts 4| When a second needle is inserted into the hole, a compression seals is formed around the needle, preventing leakage

146.

4.2.3.1 PDMS

Poly(dimethylsiloxane) or PDMS is an optically transparent, non-toxic, soft silicone

elastomer, widely used to fabricate highly reproducible microfluidic devices by replica

molding147. PDMS base monomers are in the liquid phase at room temperature and

cure at 80ºC, after being combined with a curing agent. The PDMS base monomer is

vinyl terminated, while the crosslinking monomers are methyl terminated and contain

silicone hydride units. PDMS cure process is due to the crosslinking of PDMS

monomer vinyl group and the crosslinker silicon hydride groups148. Structure of PDMS

polymer chain is illustrated on Figure 14, and is composed by repeating –O-Si(CH3)2-

units.

Hydrophobicity of this material is due to the CH3, which may become a disadvantage

since it makes cell adhesion harder and also microchannels are susceptible to air

bubbles trapping149. The surface may become hydrophilic if it is exposed to oxygen

plasma147.



FIGURE 14| Structure of PDMS polymer chain150

A large number of PDMS devices can be made from a single master, reducing the time

and cost of fabrication149. PDMS can be molded easily to form simple, complex, or

multilayered channel systems. It is a soft elastomer with a low Young’s modulus, with

great demolding behavior and reduced breakage of structures. However, depending on

the ratio of the designed microstructures, they can collapse because of the low

modulus of elasticity of PDMS and it can be a problem when dealing with structures of

a few micrometers147. Thin PDMS structures become crumpled easily, making it very

difficult to handle.

Experimental feasibility depends on the intrinsic properties of the device, like optical

transparency and imaging compatibility, low water permeability to prevent leakage,

thermal stability and is chemically inert to provide conditions to a controlled

environment151. An important characteristic as a diagnostic tool is that PDMS-based

microfluidic devices are easily sterilized through conventional means and are

biocompatible151.

PDMS offers versatility of the substrate, since it can be reversibly or irreversibly sealed

to glass, polystyrene or silicone. Substrate properties can be chosen taking in account

hydrophobicity, stiffness, topography, and the ability to be functionalized with bound

biomolecules or substrate patterning.

4.2.4 Other techniques

Hot embossing, injection molding and laser ablation are other low cost methods for

manufacturing of MEMS.

Hot embossing uses the replication of a micromachined embossing master to

generate microstructures on a polymer substrate. The plastic substrate is heated, then

a mold is insert to above plastic and forces are applied to emboss plastic substrate.

The substrate is cooled to solidify and de-molding152. It is a very quickly and simple



fabrication for small structures with high aspect ratios. However, because of the high

residual stress, created by the use of temperature in conjunction with pressure, can

cause cracks and broken edges152.

Injection molding, as the name suggests, utilizes a plunger to force the molten

polymer into a mold cavity, which forms the shape of the desired component. Polymer

solidifies, acquiring the contour of the mold153. This method offers, above all, the

opportunity of a fully automated production. Limitations for injection molding are high

tooling costs, skrinkage and partial mold filling. For microscale components, tighter

tolerances are required and mold inserts are often fabricated by costly high-precision

micromachining techniques, such as laser ablation. Laser ablation uses CO2 lasers to

cut through a polymer sheet and quickly generate patterns on it154.

4.3 Concepts in fluid mechanics

Flow properties and laws that govern the microscale are different of those

experienced in everyday life. Some applications can only be performed due to the

physics of the scale. Learning to think on an entirely different scale is a new and

exciting challenge in microfluidics. Forces that work on the microscale result in other

dominant effects such as laminar flow, diffusion, fluidic resistance, surface area to

volume ratio, and surface tension155. Projecting complex microfluidic systems requires

a study of flow inside microchannels.

In a microchannel, the flow rate, Q (m3/s), is given by:

� = � × � × � (1)

being �(m/s) the mean fluid velocity, � (m) the thickness of the channel and �(m) the

width of channel155.

The shear stress, � (Pa), represents the hydrodynamic force associated to a surface

which can cause detachment of particles. On a channel with rectangular cross section

it is given by:

� = 3�2�(�/2) × �

(2)

where � (kg/(m.s)) is the viscosity of the fluid156.

Due to shear stress at a stationary surface, e.g. the wall of a pipe, the fluid touching

the surface is brought to rest (no-slip condition). The region where there is a velocity

profile in the flow, due to the shear stress at the wall, is called the boundary layer. The



velocity increases from the wall to a maximum in the mainstream of the flow and,

between two fixed walls, the flow profile is parabolic (Fig. 15).

FIGURE 15| Motion of cylindrical fluid element within a pipe157

The velocity profile in fully developed laminar flow in pipe is parabolic with a maximum at the centerline and

minimum near to zero at the pipe wall. These flow leads to the deformation of any fluid element in the pipe.

Another important aspect is the Reynolds number, which give us information about

fluid behavior and is calculated by:

�� = ��

(3)

where � (kg/m3) is the density and � (kg/(m.s)) the dynamic viscosity of the fluid, and l

(m) a characteristic dimension of the flow. It is known that for cylindrical tubes the flow

is laminar for �� < 2300, else is transient or turbulent, and this threshold is used as

approximate reference for other geometries158. Inside microchannels a laminar flow is

observed, without turbulence, due to drastic decrease of Reynolds on microliter volume

scale159. One consequence of laminar flow is that fluid particles move in distinct and

separated layers and the streams flowing in contact without mixing except by diffusion.

Diffusion is a transport phenomenon by which a concentrated solute spreads out in a

volume, over time, , through a random movements, until becomes uniformly distributed.

In the case of particles or cells the random motion is due to Brownian motion.

Therefore, laminar flow conditions and controlled diffusion enable temporally and

spatially highly resolved reactions with little reagent consumption.

The distance d a particle moves in a time t, can be modeled in one dimension by:

� = √2�� (4)

where D is the diffusion coefficient of the particle. The hydrodynamic properties of

spherical particles can be express in the translational self-diffusion coefficient form, Dt

(cm2/s), by Einstein’s equation :

�� = � × ��

(5)



where � (K) is the absolute temperature, � (1,38 x 10-23 m2.kg.s-2.K-1) is the

Boltzmann’s constant and �� (m2/s)) the translational frictional coefficient.

The frictional coefficient for rod-like molecules was developed by Tirado and Garcia

de la Torre160, employing a bead-model procedure that accounted rigorously for the

hydrodynamic interactions and expressing the end-effect term as an interpolating

equation.

�� = 3��ln ! + 0,312 + 0,565!'( − 0,100!' (6)

Another factor that becomes important at the microscale is surface area to volume

(SAV) ratio. Surface phenomena become increasingly dominant over volume

phenomena. A very large SAV ratio and low thermal mass makes heat transfer

between the fluid and the environment more efficient161. That fact enables precise

temperature control due to quick temperature changes. However, large SAV can also

be a disadvantage in processes involving transporting fluids, because it allows

macromolecules to quickly diffuse and adsorb to channel surfaces162.

4.4 Liquid handling systems

Microfluidic platforms can be defined by its liquid handling system. Different

microfluidic platforms were summarized in a critical review by Mark et al.163, focusing

on the main requirements and characteristics of lateral flow tests, linear actuated

devices, pressure driven laminar flow, microfluidic large scale integration, segmented

flow microfluidics, centrifugal microfluidics, electrokinetics, electrowetting, surface

acoustic waves and dedicated systems for massively parallel analysis. The principal

parameter that differentiates these devices is the fluid propulsion system within

microchannels with sub-millimeter cross section. For instance, in pregnancy tests,

liquids are driven by capillary forces, and controlled by wettability and feature size of

the microstructured substrate, and it is known as a lateral flow test (e.g. test strips).

A special attention here will be given to pressure driven laminar flow, whose fluid

moves due to pressure gradients. The samples and reagents are injected into the chip

inlets in a continuous mode using syringe pumps and tubes (Fig. 16)164. Pressure

driven laminar flow offers predictable velocity profiles and it is a key requirement for

several cell-based assays.



FIGURE 16| Schematic of equipment network for pressure driven laminar flow

Reagents are stored in a syringe and are connected to the microfluidic device by a tube. Syringe pump injects the

fluid at a constant flow rate. All the process can be observed, using a microscope165

.

The strictly laminar flow effect can be utilized to implement hydrodynamic focusing

technology, illustrated by Fig. 17, whereby particles or cells suspended in the liquid

flowing through the central channel are focused and aligned to a well-defined

streamline position. The network consists of a junction of more than one inlet channels,

connected at a junction to form a common outlet channel. By varying the ratio of flow

rates, the width of streamlines within the common outlet channel can be adjusted very

accurately.

FIGURE 17| Flow focusing structure

Three different liquid streams are

symmetrically contacted at an intersection

point. Cells suspended in the liquid flowing

through the central channel are focused

and aligned to streamline position

Another technique allowed by pressure driven laminar flow is the implementation of

active and passive valves166. A disadvantage of pressure driven platforms is the

necessity of a pressure source, which requires manual steps and decreases the

portability of the device.

The requirements on microfluidic devices differ between market segments. Several

aspects need to be considered while designing a microfluidic device, such as selection

of materials, dimensions of the microfluidics devices and fluidic control devices, as

described above. The optimization of the developed device requires several

experimental tests to achieve the best conditions for an effective procedure.



5. Computational fluid dynamics

As shown above, the development of microfluidic devices is a process that requires

time and resources to build and test successive prototypes of a selected chip design.

Since the development of lab-on-a-chip devices becomes a highly competitive field,

computational simulation has become the standard method to reduce the time from

concept to chip167. Virtual prototyping can reduce the number of experimental iterations

allowing a rapid determination of how design changes will affect chip performance. A

new trend is the use of computational fluid mechanics (CFD) tools to simulate the flow

combined with mass and heat transfer in the design of a new equipment and

process171.

There are many commercial general-purpose CFD programs available. Fig. 18

illustrates the steps that must be defined in solving a problem using CFD.

A CFD solution displays the result of the specific chosen model with the given mesh,

as color maps in which is possible to obtain detailed local information (pressure,

velocity) on the simulated system that will help building and understanding of the

process.

CFD software uses an iterative solution method where the equation sets for variables

such as pressure, velocity and temperature are solved sequentially and repeatedly until

a converged solution is obtained169.

It is possible to obtain very accurate flow simulations for single-phase laminar flow

systems, the kind of flow behavior present in microfluidic devices. Simulation of heat

and mass transfer is also often very accurate and a good prediction can be easily

obtained168.



FIGURE 18| Steps in CFD simulation169

A CFD simulation starts with geometry drawing in a CAD program. The computational volume must be properly

discretized, then computational domain is meshed. A refined mesh leads to an accurate numerical solution of the

equation. A selection of the most appropriate model should be done. Simple, single-phase laminar flow, can be

simulated very accurately, because the Navier-Stokes equations can be solved directly, and for the most single-

phase laminar flows the simulations are reliable. All physical properties of the fluids must be defined. All inlet and

outlet conditions must be defined, as well as conditions on the wall and other boundaries. The solver and the

quality of an acceptable solution in terms of the convergence criteria must also be defined. Analysis of the final

simulation gives local information about fluid in a colorful display.

Geometry modelling

•Define geometry

•Define boundaries

Mesh Generation

•Divide the geometry into small computational cells

Defining models

•Add models for turbulence, laminar, etc.

Set properties

•Density, viscosity, etc.

Set boundary conditions

•The initial conditions, inlet and outlet conditions and conditions at the walls are set

Solve

•Choose solver, iteration methods, transient or steady state, convergence requirement

Post-Processing

•Analyse results



5.1 Finite volume method

There are three distinct streams of numerical solution techniques: finite difference,

finite element and spectral methods. Solely finite volume method (FVM) will be

described because is central to the most well-established CFD codes, like Fluent ® 170.

FVM is a numerical technique that divides the entire computational domain into a

number of cells known as control volumes (CV), each corresponding to a single grid

point (node)171, as shown in Fig. 19.

FIGURE 19| Bi-dimensional finite volume P and its neighbors172.

The balance of a given property in a volume P should take into account the influence of its neighbors at east (E),

west (W), north (N) and south (S), when bi-dimensional problems are considered. Capital letters refer to the center

of the elementary volume and small letters to the frontier between P and its respective neighbor.

The general transport equation for an arbitrary variable ϕ in conservative form, written

using Einstein notation, is:

� *+*� + � *(,-+)*.- = **.- /Γ

*+*.-1 + 23

(7)

where � is the specific mass, ,- is the velocity vector, Γ the diffusion coefficient, and 2 is

the source term.

Since this equation is nonlinear and often contain both spatial and temporal

derivatives, generally is not possible to solve it analytically, and requires a numerical

method. In the finite volume approach, the governing equations are made discrete and

finite, and then numerically integrated over each CV169:

4 �*+*�56�7 + 4 �*(,-+)*.-56

�7 = 4 **.- /Γ

*+*.-1�756

+4 23�756

(8)



where 9 � :(;<3):=<56 �7 represents the convective term, 9 :

:=< >Γ :3:=<?�756 is the diffusion

term that takes into account the transport of ϕ by diffusion, and 9 23�756 , the source

term, takes into account any generation or dissipation of ϕ.

According to Gauss’ theorem it is possible to rewrite eq. 8 as

4 �*+*� �756+4 �,-@+5A

�B = 4 Γ *+*.- @�B5A+4 23�756

(9)

where CS denotes the control-volume surfaces (the faces that surround the cell), dA is

the area that surrounds the volume dV, and n is a normal vector pointing outwards from

dA.

The original transport equation is now transformed into an equation that can be solved

algebraically in an iterative manner and a numerical error is introduced into the

solution. The smaller cell size is, the lowest is magnitude of the error. However, it is a

question of cell size trade-off, because reducing the cell size too much will create an

unnecessarily large number of cells and the consequences are an increased

computational effort, slowing down the process to achieve the solution169.

The FVM is perhaps the simplest to understand and to program, because it does not

require a structured grid and the values of the neighboring nodes are unknown173.

Although unstructured grids make it possible to mesh complex geometries, CFD

programs using unstructured grids are slower and require more memory than those

using structured grids169.

Chapter III – Materials and methods

1. Yeast characterization

2. PNA-FISH protocol optimization

3. Material preparation

4. CFD simulation

5. Characterization of the microchannels

6. Microfluidic channels retention tests

7. PNA-FISH methodology within the

microfluidic channel

8. Microscopic visualization



1. Yeast characterization

1.1. Surrogate microorganism selection, yeast strain and culture maintenance Experimental research involving pathogens requires precautions to prevent infection

and contamination during microbial procedures. Candida spp. are categorized as

biohazard level 2 and LEPABE laboratory available for this project is classified as

Biosafety level 1. In the scope of the project, a surrogate microorganism for Candida

spp. had to be selected.

Saccharomyces cerevisiae, a non-pathogenic yeast well-known from the alimentary

industry174, met all criteria.

Saccharomyces cerevisiae PYCC4072 was kindly provided by Dr. Manuela Rodrigues

from University of Minho - Biology Department and maintained YEPD medium with

25% glycerol at -80ºC. The strain was then streaked onto a YEPD-agar and incubated

at 30ºC, as described by Larsson et al175.

To prepare liquid cultures, a loopful of biomass was transferred into 100mL of YEPD

medium and incubated overnight (~16h) at 30ºC (VELP Scientifica Incubator, FOC

225E model, Usmate, Italy) and 150 rpm (IKA KS 130 basic shaker, Vidrolab,

Portugal), under aerobic conditions, yielding a stationary phase culture176.

YEPD-agar and YEPD media recipes may be found in Appendix I.

1.2 S. cerevisiae growth

The growth curve for S. cerevisiae PYCC4072 strain was carried out by transferring a

small portion of the overnight culture to 100mL of fresh YEPD medium until it reaches

an OD600 of 0.1, and incubated at 30ºC and 150 rpm. Cellular growth was determined

by measuring optical density at 600nm (OD600) every 30 min using a

spectrophotometer (VWR V-1200), until the culture reach the stationary phase. This

experiment determined the OD600 of the culture that corresponds to the mid-log phase.

In order to assess the number of cells present in a given suspension, a calibration

curve was also built where an overnight culture of S. cerevisiae was prepared being

serially diluted in YEPD medium (1:2, 1:4, 1:8, 1:16, 1:32). OD600 of each sample was

measured in triplicates. Cells suspensions (500 µL) were filtered onto black nucleopore

polycarbonate membrane (Ø 25 mm) with a pore size of 0.2µm (Whatman, Japan) and

3 drops of 4’,6-diamidino-2-phenylindole (DAPI) (Sigma-Aldrich, USA) (0,2 mg/mL)

were added. Samples were placed in the dark, incubated at room temperature for 10

CHAPTER III – MATERIALS AND METHODS


min. After that, the samples were filtered by vacuum and mounted with 1 drop of

immersion oil (Merk, Germany) on a slide. Cells were analyzed using an

epifluorescence microscope (Leica DM LB2, Leica Microsystems, Germany) (see used

features in microscopic visualization description below) . For each sample, 10 fields

with an area of 0,0063 mm2 were counted. For every field, the cells were manually

counted and averaged. A calibration curve was prepared by plotting the mean values of

OD600 against the corresponding concentration. All experiments were repeated at least

three times.


2.1 Probe sequence

A PNA-probe targeting S. cerevisiae 26S rRNA, already designed by Meireles177 and

ordered to Panagene Inc., was used.

The final probe sequence was Alexa 594-OO-AGGCTATAATACTTACC (5’-3’), being

HPLC purified > 90%. The theoretical specificity is 91.15% and sensitivity of 89.47%.

Probe thermodynamic parameters calculated by Meireles are summarized in Table 2.

Table 2| Probe thermodynamic parameters177

Probe ∆H

(Kcal/mol)

∆S

(kcal/K)

∆G

(Kcal/mol)

Tm

(ºC)

Tm PNA

(ºC)

AGGCTATAATACTTACC -115 -321.7 -15.26 62.4 66.2

2.2 Hybridization conditions optimization

Although hybridization conditions for this particular PNA probe were already studied, it

was decided to test several different parameters on the hybridization assays

(description bellow) to simplify the methodology. Hence, different hybridization

temperatures, formamide concentrations and the use of simplified hybridization

solutions were tested to increase the signal to noise ratio (See Chapter II, section 3).

For the hybridization conditions optimization, assays were based on the PNA-FISH

suspension protocol described by Perry-O’Keefe et al178. One mL of 1x107 cells /ml of

S. cerevisiae was pelleted by centrifugation at 10 000G for 5 min (Centrifuge 5418,

Eppendorf, USA), and the pellet was re-suspended in 400µL of 4% paraformaldehyde



(wt/vol) (Acros Organics, UK) and left for 1 hour at room temperature. Then,

suspension was pelleted by centrifugation again at 10 000G for 5 min and resuspended

in 500 µl of 50%(vol/vol) ethanol and placed for at least 30 min at -20ºC. Subsequently,

100 µl of the fixed cells aliquot was pelleted by centrifugation and resuspended in 100

µl of hybridization solution containing the PNA probe at 200nM and incubated for 60

min (FD 23, Binder, Germany). Besides the standardly used hybridization solution179,180

two simplified hybridization solutions with different formamide concentration (30% and

0%) were tested and compared. The detailed composition of these solutions may be

found in Appendix II. Also, different temperatures, ranged between 50 and 60ºC, for

each hybridization solution were tested.

After hybridization, cells were rinsed by centrifugation at 10.000 G for 5 min, and 500

µl of wash solution was added (composition in Appendix III). Cells were incubated at

the same temperature as in hybridization for 30 min. Washed suspension was pelleted

by centrifugation, and resuspended in 500 µl of sterile water. Finally, 20 µl of the cell

suspension were placed on a microscope slide. Samples were allowed to air dry. The

sample is then ready to be observed in the epifluorescent microscope.

2.3 Fixation conditions optimization

The choice of the agents to be used in the fixation step can compromise the

performance of subsequent FISH procedure (See Chapter II, section 3). Different

fixation protocols were performed and the quality of fluorescent signal was compared.

One mL of S. cerevisiae inoculum was pelleted by centrifugation at 10 000G for 5 min

and re-suspended in 1mL of sterile water. 20µL of the suspension was dispensed in 14

mm well slides (Thermo Scientific) and allowed to dry by flame. Subsequently, smears

were chemically fixed using ten different protocols, described in Table 3, based on the

previously described by Guimarães et al. 181, Reller et al.182, Rigby et al.183, Harris and

Hata184, Shepard et al.185, protocols described by the manufacturer abcam®186

(methods 2, 5 and 7) and protocols suggested by the manufacturer Li-Cor ®187

(methods 9 and 10). In protocols with most volatile reagents, such as alcohols, the

procedure was performed at low temperatures (-20ºC).

The subsequent hybridization step was performed according to Guimarães et al.181,

where samples were covered with 20µL of hybridization solution containing 200nM of

PNA probe, covered with cover slips and incubated for 60 min at 54ºC, which was the

temperature with better signal to noise ratio (see chapter IV for these results).



Subsequently coverslips were removed and the slides were immersed in the wash

solution and incubated for 30min at 54ºC. Then, samples were allowed to air dry and

were ready to be observed by epifluorescence microscopy.

2.4 Assessment of dextran sulfate influence on hybridization

In microfluidic devices, excessive solution viscosity should be avoided mostly because

of the difficulty in flowing high viscosity fluids on the microscale (See equations (2) and

(3) of Chapter II-Section 5.3). Also, when cells contact with the hybridization solution

containing the probe, the assessment of the probe to the target is made by diffusion,

and according to equations (4) and (5) (see Chapter II, Section 5.3) the diffusion is

faster as less viscous the solution is. Dextran sulfate is known for accelerating the rate

of hybridization but also to increase the viscosity of the hybridization solution188. This

experiment was performed to evaluate the influence of dextran sulfate on the quality of

the hybridization. For this purpose the protocol in suspension was used (see

description above in point 2.2).

Two simplified hybridization solutions with 10% (w/v) and 0%(w/v) of dextran sulfate

were compared. A negative control was performed simultaneously, where no probe

was added during the hybridization procedure.

Table 3| Comparison of different fixation protocols

Protocol Fixation step in slide

Reagent Time Temperature Reference

Method 1 4% (w/v) Paraformaldehyde 10min Room temp. Guimarães et

al.181 50% (v/v) Ethanol 10min -20˚C

Method 2

4% (w/v) Paraformaldehyde 10min Room temp. Anon. 2014a186 0,2% (v/v) Triton X-100 in

PBS 10min -20˚C

Method 3 4% (w/v) Paraformaldehyde 10min Room temp.

Reller et al.182 100% (v/v) Methanol 10min -20˚C

Method 4

1% (v/v) Triton X-100 in PBS

10min Room temp. Rigby et al.183

96% (v/v) Ethanol 10min -20˚C

Method 5 100% (v/v) Methanol 10min -20˚C Anon. 2014a186

Method 6 100% (v/v) Methanol 10min -20˚C Harris and

Hata184 80% (v/v) Ethanol 10min -20˚C

Method 7 100% (v/v) Acetone 10min -20˚C Anon. 2014a186



Method 8 50% (v/v) Ethanol 10min -20˚C Shepard et al.

(adapted)185

Method 9 100% (v/v) Acetone 10min -20˚C

Anon. 2014b187

100% (v/v) Methanol 10min -20˚C

Method 10 50% (v/v) Acetone in

Methanol 10min -20˚C Anon. 2014b187

3. Material preparation

3.1 Microchip design

Microfluidic systems were conceived to operate by pressure driven laminar flow.

Therefore, solutions were injected into inlets using syringe pumps and flow to the

microchannel due to pressure gradients, coming out in the outlet. To concentrate the

cell suspension, entrapment strategies focused on dam-like structures and filtering

pillars were employed. It is necessary to take special care in the design of geometries

in order to avoid clogging, because exposure of cells to different solutions during PNA-

FISH procedure is performed by passing each solution from the inlet to the outlet of the

cell-trapping microchannels.

3.1.1 Production of molds – channels with a dam-like structure

A microfluidic device possessing an 8 µm gap was planned for cell entrapment. A

cross-sectional view of the microfluidic device is shown in Fig. 20. It is meant that cells

get trapped in the region of the dam, in front of the 8 µm gap, which is called the

“detection zone”.

As can be seen in Fig. 20, the microfluidic device was constructed by bonding of two

(upper and lower) parts. The mold production of the upper part was performed by

xurography technique. Two geometries composed by two macro-to-micro interfaces

(inlet and outlet) and a microchannel connecting them with 25 000µm long and 300µm

wide were drawn by CorelDRAW ®. Geometries were cut with a cutting plotter on a

100 µm thickness self-adhesive paper using GreatCut® software. The desired mold

was placed on a petri dish and the surrounding paper was removed.

A small rectangle was cut and adherently attached to a glass slide substrate. Substrate

was covered with PDMS mixed in 1:20 (or 1:10) ratio, using a spin coater (ω=3000 rpm

for 50s) to obtain a thin layer and was cured in an oven for 20 minutes. The rectangle



was removed, forming a 8µm well, and PDMS block was bonded to glass slide

substrate and connected to an injection system.

FIGURE 20| Construction of the microfluidic device with a dam-like structure with 8-µm gap

To construct the channel with a dam-like structure the upper part with a 100 µm high barrier is bonded to a lower

part with a 8 µm depth gap

3.1.2 Production of molds – channels with pillar-based structure

The geometry of the channels was drawn using AutoCAD 2013 ® software (Autodesk

Inc, USA). Eight different geometries composed of microchannels with filtering

structures were evaluated. Five of these eight geometries are composed by one inlet

and one outlet and remaining three have two inlets and one outlet, to apply the

hydrodynamic focusing principle. The main differences between these approaches are

the arrangement and the dimensions of the gaps between the filtering pillars, and also

the influence of an enlarged section in which the velocity of the fluid is amended.

SU-8 molds on a silicon wafer were produced by photolithography using chrome masks

by a specialized manufacturer (microLIQUID, Spain). Molds were ordered in two

different depths, 50µm and 30µm, to ascertain what is the most appropriate.

3.2 Microfabrication of microchannels

Microchannels were manufactured by soft lithography. A two-part PDMS kit (Sylgard,

USA) was used to create the liquid polymer. The PDMS used to create a negative of

the molds was hand mixed, using a glass stirring rod, with curing agent at a ratio of 1:5.

To degas, remove air bubbles formed during the mixing, the liquid polymer was placed

in a desiccator connected to a vaccum pump until no air bubbles were observed. The

liquid PDMS was poured over the prefabricated mold at room temperature and pre-

cured for 20 minutes at 80ºC. Then, a PDMS block with the channels imprinted on it

was removed from the mold and access points were created using a needle. The

PDMS used to create a coverslip in the substrate to seal the channel, was hand mixed



with curing agent at a ratio of 1:20 and the air bubbles were removed used the same

conditions as above. A small portion of the liquid polymer was applied on the glass

cover slip and placed on a spin-coater at a speed of 3000 rpm for 50s to form a thin

layer of elastomer, and then was pre-cured for 5 min at 80ºC. Finally the PDMS block

is bound to the spin coated glass slide and cured at 80ºC overnight.

4. CFD simulation

Computational fluid dynamics simulations were made in Ansys Fluent CFD package

(version 14.5, Ansys, Inc., Canonsburg, PA, USA). Channel geometries designed in

AutoCAD were imported to Design Modeller 14.5 (Ansys, Inc., Canonsburg, PA, USA)

and discretized into a grid of quadrilateral cells by Meshing 14.5 (Ansys, Inc.,

Canonsburg, PA, USA).

Table 4 contains the characteristics of each mesh created. A mesh independence

study was performed to make sure that the solution is also independent of the mesh

density. The mesh was refined and the number of cells across the gap was increased

until a certain point in which the solution do not change because of mesh density. To

reduce the simulation run time, the smallest mesh that gives the mesh independent

solution was used.

For each simulation, the initial velocity of the inflow from the inlets was assumed to be

0.003 m/s, corresponding to an inlet flow rate of 1µL/min in a cross section of 100µm x

50 µm, and the fluid was assumed to be water for the sake of simplicity. A pressure

outlet boundary condition was implemented at the downstream and was assumed to be

zero. The boundary condition for the inner walls of the microchannels was defined as

the no slip condition.

In Ansys Fluent, the convergence was defined by reducing Residual Error to an

acceptable value (10-6) to ensure a valid solution is achieved. The number of iterations

was set to 100, ensuring that the main outputs from the simulation have converged to a

steady solution.

Ansys Fluent was used to carry out the two dimensional (2D) simulation of velocity

magnitude and stream function of each channel geometry conceived.



Table 4| Characteristics of each 2D mesh

Cells Across Gap Nodes Elements

Simulation I 20 5275 5003

Simulation II 10 329844 322479

Simulation III 20 994975 985397

Simulation IV 20 313105 303838

Simulation V 10 40358 37507

Simulation VI 10 519249 506888

Simulation VII 10 83092 79682

Simulation VIII 10 240640 235096

Simulation IX 10 40890 37438

5. Characterization of the microchannels

5.1 Geometrical characterization of the microchannels

Several measurements were performed using Leica Application Suite tool, to confirm

if the results from microfabrication were in agreement to the designed.

To confirm if the channel height correspond to the depth ordered to the manufacturer,

a cross section of the channel was cut as displayed in Fig. 21, the cross section was

placed in a glass slide, observed using a microscope (DMI 5000M, Leica Microsystems

GmbH) and its dimensions were measured using Leica Application Suite tool.

FIGURE 21| Cross section of the microchannel189



5.2 Volumetric measurement of channels capacity

Due to the irregular geometry of the microchannels, determining the capacity of each

microchannel geometry seems to be more accurate by experimental methods instead

calculations. Concentrated violet crystal, was pumped into the device at a constant flow

rate and the filling time was clocked. The volume was achieved using a simple

calculation, with the formula 7 = � × �, being V (µL) the volume, Q (µL/s) the flow rate

and t (s) the time.

5.3 Material testing

In order to evaluate the suitability of the PDMS for fabricating the microchannels and

complete the PNA-FISH methodology within its interior without compromising the

characteristics of the channel, a simple test was performed. A 50 µm wide

microchannel with a narrowed section of 20 µm was filled with each reagent of the

PNA-FISH procedure during the specific time of each one, at room temperature. The

channel was visually analyzed using a microscope (DMI 5000M, Leica Microsystems

GmbH) and its dimensions were measured at two points using Leica Application Suite

tool, to assess the resistance of the PDMS under those conditions.

6. Microfluidic channels retention test

In order to test channels ability to entrap yeasts, retention was first evaluated using 10

µm polystyrene microspheres (Sigma-Aldrich, USA). Theoretically, microparticles will

have a behavior similar to yeasts and are integrated in the initial tests because are

considerably easier to handle and require less limitation laboratorial conditions than

microorganisms. The same test was later made using S. cervisiae.

7. PNA-FISH methodology within the microfluidic channel

The concept of trapping and labeling yeast cells within the microfluidic device is

illustrated in Fig 22. For a FISH assay, a new microfluidic chip was washed with sterile

water for 30 s at a constant flow rate of 100 µL/min using a syringe pump (Cetoni,



neMESYS syringe pump, Germany). The flow rate in this first washing step was so

high to drag any dust and remove air bubbles that sometimes remain in the detection

zone. Then, 5 µL of solution containing approximately 1x107 cells/ml of previously fixed

S. cerevisiae were delivered into the device using a micro-syringe. Yeast size is larger

than the gap between any two given pillars, and for this reason, S. cerevisiae were

trapped in front of the pillars and retained in the detection zone by the flow pressure

toward the pillars. The hybridization solution chosen as optimum was pumped at a

constant flow rate of 1µL/min into the device at a constant flow rate until fill the channel,

allowing the probe to access the target. The device was disconnected from the pump

system and placed in the incubator at 54ºC for 1 h. After hybridization, the chip was

reconnected to the pump system and washing solution was injected at a constant flow

rate of 1µL/min until fill the channel, to remove exceed probe. The device was again

disconnected from the pump system and placed in the incubator at 54ºC for 30 min.

Then, the chip was reconnected to the pump system and sterile water was injected to

wash away artifacts that can influence the result and DAPI (0.2 mg/mL) was injected

until fulfill the channel, to counterstain the cells. After 10 min, sterile water was injected

to wash any residuals. Microfluidic device containing labeled yeasts is ready to be

observed by epifluorescence microscopy.

Integration of FISH methodology with these new microfluidic devices requires some

adjustments, performed as a trial and error procedure. The volume of fixed cells

injected was adjusted, as well as the flow rate of the hybridization and washing

solution.



FIGURE 22| Schematic illustration of the concept to perform FISH assay in the microfluidic

device

I| Yeast cells are trapped and retained. II| PNA probe solution is pumped into the device at a constant flow rate . III| After hybridization, cells are washed to remove loosely-bound probes and examined using an epifluorescence microscopy.

8. Microscopic visualization

Fluorescently labeled samples were observed using Leica DM LB2 (Leica

Microsystems, Germany) epifluorescence microscope connected to a Leica DFC300

FX camera (Leica Microsystems, Germany) and equipped with a Live/Dead Filter

(Excitation 530 to 550 nm; Barrier 570 nm; Emission LP 591 nm) sensitive to the

fluorophore Alexa Fluor 594 attached to PNA probe. To confirm the absence of cells

autofluorescence a N2.1 Filter (Excitation 515 to 560 nm; Barrier 580 nm; Emission LP

590 nm) was used. The microscope software parameters (exposure, gain and

saturation) were carefully maintained constant in all the experiments involving FISH.



For DAPI stained cells a Chroma 61000-V2 Filter, consisted of a 545/30 nm excitation

filter combined with a dichromatic mirror at 565 nm and suppression filter 610/75, was

used.

During experimental microfluidics, a Leica DMI 5000 microscope was used.

Chapter IV – Results and Discussion

1. S. cerevisiae growth


3. Microchannels geometry design

4. Computational fluid dynamics analysis

5. Retention tests in channels with a dam-

like structure

6. Geometrical characterization of

channels with pillar-based filter

7. Retention tests in single channels with

pillar-based filter

8. PNA-FISH methodology within the

microfluidic channel

CHAPTER IV – RESULTS AND DISCUSSION


1. S. cerevisiae growth

. Being the goal of the project the development of a microfluidic filter-based device in

which yeasts can be identified by FISH, the surrogate microorganism has to mimic

especially the morphology of Candida spp.. To correctly evaluate the practicability of

any conceived trapping structures, surrogate must have similar shape, size and life

cycle form.

Phylogenetically, Saccharomyces spp. is closely related to Candida spp., sharing a

common ancestor190. Due to its long history of safe use and consumption, and lack of

production of toxins, most strains of S. cerevisiae have been considered as generally

recognized as safe (GRAS)191. Also, it is easy to obtain and to produce a large number

of samples at low cost192. The morphology of S. cerevisiae is similar to Candida spp.,

nonetheless, small differences in terms of size are verified. S. cerevisiae cells are ovoid

with 5-8 µm whilst Candida spp. diameter is about 2-3µm193. Since microfluidic

channels fabrication may be easily scaled up or down according to the used technique,

small differences in term of size are insignificant.

Under appropriate nutritional conditions that trigger cellular morphogenic programs, S.

cervisiae can alternate between yeast-like growth form and pseudohyphae194,195. Also,

environmental conditions highly influence the thickness of its cell wall that can vary in a

range of 10 to 25% of its total biomass, depending on the nature of the carbon

source196.

As already referred, probe penetration into the cell wall is crucial to an effective

hybridization during FISH procedure (see Chapter I, Section 3). Understanding the

properties of the cell wall can be helpful to optimize the FISH protocol. S. cerevisiae

cell wall resembles C. albicans cell wall197 and is essentially composed by β1,3- linked

glucans and mannoproteins198. β1,6- linked glucan is a minor component of the wall

but has a central role in cross-linking wall components199. Another important minor

component is chitin, which can be covalently joined to β1,6- glycosylated

mannoproteins to form high order complexes and contributes to the insolubility of the

fibers200. Although chitin levels in the cell wall are normally low, cell wall stress activate

the cell integrity pathway, which result in a strong increase in the deposition of chitin in

the lateral wall, increasing the resistance of the cells201. Another interesting fact is that

proteins incorporated in the cell wall depend not only on environmental conditions but it

is observed differences at different phases of the cell cycle202. Cell wall of

logarithmically growing cells has relatively high porosity, that could reflect a lower



degree of cross-linking, whereas walls of stationary-phase cells are less porous203,204. It

is easy to understand that PNA probes would penetrate easier through the cell wall

pores to access the target rRNA in the cytosol in logarithmically growing cells than in

stationary-phase cells with insoluble and resistant cell walls.

To predict an efficient model for the S. cerevisiae growth, OD600 measures were

performed with time (Fig. 23). Obtained results match typical yeast growth curve,

where three essential phases are observed: a lag phase, when newly pitched yeast

acclimate to the environment; a log phase, when cells grow rapidly due to relative

excess of nutrients and insignificant waste in the environment; and a stationary phase,

when cell growth decelerate due to substrate consumption and high waste

concentration205.

In order to obtain the best fluorescence intensity in fluorescent-labeled yeast cells by

PNA-FISH, reducing the possibility of false negatives, some cell cycle-dependent

characteristics of yeasts need to be considered. Ribosomal content are closely linked

to growth rate and, according to Warner’s review85, rapidly growing yeast cells contain

a higher number of ribosomes per cell. Being the targeted sequence present in the 26S

ribosome, higher number of ribosomes per cell leads to more targets and,

consequently, enhanced fluorescence intensity. According to Smits et al.202 study in

cell cycle-regulated incorporation of cell wall proteins, mid-log phase yeast cell wall are

highly porous, what improve the accessibility of the probe to the target by ease the

penetration of probe into the cell through the cell wall pores. Thereupon, it is thought

that intensity of probe-conferred fluorescence reach a higher value in the mid-log

phase.

FIGURE 23| Saccharomyces cerevisiae growth curve



In addition, to determine S. cerevisae concentration used in the experiments, a

calibration curve was obtained (Fig. 24)

FIGURE 24| Saccharomyces cerevisiae calibration curve

To the FISH experiments, S. cerevisiae culture was grown until the mid-log phase with

an OD600 of approximately 0.90 what, according to the obtained calibration curve,

corresponds to a cell density of 8,7x106 cells/mL.


PNA-FISH protocol optimization main goal is to achieve the best output, in order to

enhance signal-to-noise ratio, to a better discrimination of the results when performed

inside microfluidic devices. Several assays were performed to define the best fixation

and hybridization conditions, such as fixation protocols, hybridization temperature, and

hybridization buffer composition. Results were analyzed comparing the fluorescent

signal-to-noise resulting of each assay.

According to SantaLucia206, thermal stability, also referred as melting temperature, is

related to the probe affinity to the target. It can be estimated by Gibbs free energy

change during hybridization reaction89. However, this theoretic calculation does not

consider, among other parameters, the presence of formamide, a denaturant agent

known to reduce the melting temperature of double-stranded nucleic acids207. Aware

that hybridization stringency is affected by the interplay of several parameters and to

properly adjust hybridization conditions during experimental PNA-FISH methodologies,

a trial and error procedure was performed. Temperatures ranging from 51-59ºC were

y = 1,50E+07x - 4,79E+06R² = 9,18E-01

-5,E+06

0,E+00

5,E+06

1,E+07

2,E+07

2,E+07

3,E+07

3,E+07

0,0 0,5 1,0 1,5 2,0

Ce

ll c

on

cen

tra

tio

n (

cell

s/m

L)

OD600



tested with different hybridization solutions to assess PNA-FISH performance, and the

results are displayed in Fig. 26. Initial temperature optimizations test lower

temperatures than the theoretically melting temperature, because higher temperatures

are not favorable in biological terms208. The chosen temperature is according to

previous results obtained by Meireles177, were the best temperature for hybridization in

suspension was 53ºC and for hybridization in slide was 55ºC. Lima209 also predicted,

with a response surface methodology (RSM) approach, the maximum hybridization

efficiency at a temperature of 53.9ºC. As such, and according to the results observed in

here it seems appropriate to use the medium temperature of 54ºC in the subsequent

tests. Regarding the comparison between hybridization solutions, the difference from

the standard hybridization solution developed by Azevedo179 and simplified one used

by Santos et al.210 is the lack of Denhardt’s solution components (sodium

pyrophosphate, polyvinylpyrrolidone, and FICOLL), NaCl and disodium EDTA.

According to Azevedo179, sodium pyrophosphate, polyvinylpyrrolidone and FICOLL act

as blocking reagents for preventing unspecific binding to nitrocellulose membranes;

NaCl is used to increase salt concentration, increasing reaction rate and stabilize

secondary structures of ribosomes; and disodium EDTA act as a chelator. Also the

presence of formamide, a denaturing agent, in the simplified solutions is evaluated.

Formamide is known to increase the accessibility to rRNA target by weakening

hydrogen bounds responsible for secondary structures of ribosome.

Analyzing the results presented in Fig. 25, it is clear that the best results occur for

54ºC with a simplified solution without formamide. Contrarily to what was thought,

hybridization without a denaturant agent is stronger compared to hybridization with

formamide. Actually, it is possible and clearly distinguished compared with a negative

control (data not shown). Earlier accessibility studies in Bacteria and Eukarya211

conclude that the absence of formamide in the hybridization buffer result in kinetic

limitations, in agreement with Yilmaz et al.207. However, recently, Santos et al. 210,

proved that the use of formamide in FISH is beneficial for certain species of bacteria,

but could also have a harmful effect on integrity of cells with a relatively thick wall. In

the same study, S. cerevisiae was used to demystify the influence of formamide in the

hybridization buffer, leading to the conclusion that do not affect the hybridization

specificity. Accordingly, the absence of fomamide in PNA-FISH may not affect the

specificity of the method.



Temp.

(ºC)

Standard hybridization

solution

Simplified hybridization solution

(30% formamide)


(0% formamide)

51

53

54

55

57

59

FIGURE 25| Fluorescence microscope results for in suspension PNA-FISH with different hybridization solutions at different temperatures Images were obtained with equal exposure times. Original magnification x 1000. Scale bar 20µm.



What is also observable in Fig. 25 is that the hybridization is not perfect and the

distribution of the signal is not uniform. In Fig. 26b is observable the heterogeneity of

the fluorescence signal in different cells of the same samples.

a. b.

Figure 26| Fluorescence microscope results for in suspension PNA-FISH with simplified hybridization solution (0% formamide) at 54˚C a| Smears without probe as negative control b| Positive. Images were obtained with equal exposure times. Original magnification x 1000. Scale bar 20µm.

As already explained, rRNA content depends on physiologic state of the cell which is

directly correlated with growth rate102. Even though the culture was used when reach a

mid-log phase, it is possible that cells appear in different states of growth, since cell

division is not a synchronized event. Depending on the cell cycle phase of a single cell,

low rRNA content or increased wall resistance can thereby result in low signal intensity

or false negatives as well76.

Despite all significant differences in intensity of probe-conferred fluorescent in this set

of temperature optimization tests, even in the optimum temperature, the signal is less

intense than expected. Another consequence of low signal intensity may be insufficient

probe penetration into cell wall. Optimal fixation should result in good probe penetration

and retention of the maximal level of target rRNA, as well as maintenance of cell

integrity and morphologic detail76.

Various fixation protocols were tested to assess the best fixation conditions to S.

cerevisiae in order to obtain a successful hybridization. All the fixation procedures were

performed in slide for the simplicity of the method. Due to the results obtained in

hybridization conditions optimization, simplified hybridization solution with 0% formamide

and hybridization temperature of 54ºC were used in all the subsequent experiments.

The results were displayed in Fig. 27. The negative control was obtained with N2.1

filter that was not able to detect the probe.



Protocol Positive

(Live/Dead Filter) Negative control

(N2.1 Filter)

Method 1

Method 2

Method 3

Method 4

Method 5

Method 6

Method 7



Protocol Positive

(Live/Dead Filter) Negative control

(N2.1 Filter)

Method 8

Method 9

Method 10

Figure 27| Fluorescence microscope results for in suspension PNA-FISH simplified hybridization solution (0% formamide) at 54˚C after different fixation protocols Images were obtained with equal exposure times. Original magnification x 1000. Scale bar 20µm.

Examining all positives, it might be possible to think that Methods 9 (acetone and

methanol) and 10 (50% (v/v) acetone in methanol) offered the best results. However,

looking closely at their negative controls, it is observable that there is an unexpectedly

high signal when looking at any other negative. Autofluorescence is the natural

emission of light by biological structures that masks specific fluorescent signal. It can

be concluded that fixation with acetone and methanol trigger autofluorescence in S.

cerevisiae. Some moulds and yeasts have been reported by Graham212, to be

themselves autofluorescent, therefore the interpretation of fungal fluorescent signals

has to be done with caution213,214. Graf et al.213 state that autofluorescence in C.

albicans depends on the fixation and excitation wavelength. Nevertheless, S.

cerevisiae background do not appear to increase in paraformaldehyde-fixed cells, as

can be seen by previous results (Method 1 by Guimarães et al181 ). When Methods 6

(Methanol and 80%(v/v) ethanol), 7 (acetone) and 8 (50% (v/v) ethanol) were used, a

weaker signal was observed, probably due to alcohols evaporation, before completing

the 10 minutes of fixation. These alcohols were applied at -20˚C to prevent this

situation, but it seems to be inefficient. In addition hybridization results after Method 2

(4% paraformaldehyde and 0.2% (v/v) Triton X-100 in PBS) and 4 (1% Triton X-100 in



PBS and 96% ethanol) fixation protocols were very similar, presenting a low signal-to-

noise ratio, which may be due to Triton X-100 reagent, common in both protocols.

Triton X-100 is a nonionic surfactant widely use to permeabilize cells215. It solubilizes

phospholipids, forming transmembrane pores that serve as channels to periplasm216.

However, if large amounts are added or cells are subject to prolonged exposure to

Triton X-100, disruption of polar head group on the hydrogen bonding present within

the cell’s bilayer could lead to destruction of compactness and integrity of the lipid

membrane217. Miozzari and colleagues215 report that S. cerevisiae are completely

permeabilized by Triton X-100 concentration of 0.05% (vol/vol), reiterated by Vasileva-

Tonkova et al218. Therefore, it leads to the conclusion that maybe the concentration of

Triton X-100 used was high enough to compromise cell wall integrity, that is intrinsically

linked to cell viability76.

In the end, Method 1 and 5 appears to be the most effective among the tested

protocols. Therefore, the fixation method chosen was Method 1 by Guimarães et al.181

since it is the more standard one219,220. The conditions achieved in slides hybridizations

were also used in suspension in all subsequent experiments.

As any of these fixation optimization protocols result in an increased of fluorescence

intensity, it was supposed that the limitation of the method could be related to the

accessibility of the target region of the ribosome. Conformational structure of

ribosomes can be favorable to hinder the targeted sequence and difficult the probe

access and consequent hybridization76. Inácio et al.87 evaluated the accessibility of D1

and D2 domains in the 26S rRNA of S. cerevisiae. To identify D1-D2 position of the

sequence targeted by the probe used in this work [AGGCTATAATACTTACC (5’-3’)], a

multiple sequence alignment software, ClustalW [European Bioinformatics

Institute(http://www.ebi.ac.uk/Tools/msa/clustalw2/)] was used. The entire sequence of

the 26S rRNA D1 and D2 domain of S. cerevisiae (sequence retrieved from GenBank

under accession number U44806) used by Inácio et al87f was aligned with the

probe221. According to results of the referred paper, the probe targets a lower

accessibility target sequence (brightness class V with 6-20 % of relative probe

fluorescence). However, low salt concentration is used, which destabilizes native

nucleic acid structures and results in an improved access to target sequnces.

The last attempt of PNA-FISH protocol optimization was made to evaluate if a less

viscous hybridization solution, more advantageous to flow into microchannels, could

perform an efficient hybridization222. Dextran sulfate is a volume exclusion polymer

incorporated in the hybridization solution that causes increasing viscosities and surface

tensions. Brigatti223 supports that increased viscosity inhibits both probe diffusing in and



out from the target region and during washing steps to wash away probe excess,

making such hybridization solutions non-ideal for capillary gap technology, which can

be a problem when transporting FISH procedure to microfluidic devices.

As hypothesized, to assess dextran sulfate influence on quality of hybridization, cells

were fixed using Guimarães et al.181 protocol, and hybridization was performed at 54ºC

with simplified hybridization solution (0% formamide) and simplified hybridization

solution (0% formamide and 0% dextran sulfate), as shown in Fig 28.

In a brief analysis of the results, is clearly observable that sulfate dextran is crucial for

an efficient hybridization under these conditions and the chemical effect overlaps the

increased viscosity. Dextran sulfate functions as an accelerant by excluding the nucleic

acid from volume of the solution occupied by the polymer, resulting in local

concentration of the probe. This principle had been shown to be successful previously

by Zaytseva et al224 and Ku et al225.

For these reasons, the dextran sulfate concentration of 10% (vol/vol) was used in all

subsequent experiments.

Positive

Negative control (No probe)


(0% formamide)


(0% formamide and 0%

dextran sulfate)

Figure 28| Fluorescence microscope results for in suspension PNA-FISH simplified hybridization solution (0%

formamide) at 54˚C with presence or absence of dextran sulfate.

Images were obtained with equal exposure times. Original magnification x 1000. Scale bar 20µm.

3. Microchannels geometry design

Most of the forces relevant for microfluidics are negligible at the macro scale, so the

development of new geometries has to be thought having in mind the scale effect.

Several geometries were conceived to obtain various microfluidic systems and choose



the best approach for rapid and efficient microorganism identification by PNA-FISH.

Two types of structures were evaluated: dam-like structures and pillar-like structures

3.1 Channels with a dam-like structure

Geometries designed with CorelDRAW ® software are represented by Fig. 28. The

results are two microchannels composed by one inlet and one outlet carrying out the

macro-to-micro interface. The microchannel connecting them is 25 000 µm long, 300

µm wide. A 250 µm thick stripe was purposefully removed in the middle of the channel

to be filled with PDMS during soft lithography and form the 250µm wide central barrier

of the dam structure in the upper part. Channel represented by Fig. 28a. is 10 times

enlarged to increase the cross-sectional area at dam-like structure and consequently

decrease Re, comparatively to channel represented by Fig. 28b. It was designed in an

attempt for decrease velocity in this region for the fluid do not exert pressure enough to

push cells beyond the dam-like structure through the gap. The gap will be formed

below the barrier, during the microfabrication of the device, and it will have the shape of

a rectangle.

(a)

(b)

FIGURE 29| Mask structure

a| Geometry to produce a dam-like structure; b| Geometry with the same entrapment strategy but enlarged at the

detection zone.

3.2 Channels with a pillar-based filter

Geometries for the pillar filtering strategy guarantee that the gaps between pillars are

smaller than the cell size, so that, the gaps between each two pillars of every designed

structure have 5µm width. In some cases, it was decided to incorporate three (or more)

lines of pillars so if the cells can escape from one line can be entrapped at the next.



3.2.1 Single channel devices

Fig. 30 and 31 represent microfluidic systems with one inlet and one outlet connected

by a 10 000 µm long and 100 µm wide channel, being the channel represented by Fig.

31 enlarged at the detection zone, exactly for the same reasons of the channel

represented by Fig. 29b, to reduce the velocity at this level. The detection zone is

where cells are retained over time. It located in the middle of the channel to be easy to

find when observing the results with the microscope. Total structure was maintained in

Fig. 8 but the geometry of the pillars was altered. Channel represented by Fig. 30a is

100 µm wide, but when the fluid passes through the gaps, the section is reduced to a

quarter, 5 gaps of 5µm wide. The expected consequence is an increased velocity in the

detection zone.

(a)

(b) Mask structure

FIGURE 30| Geometries of cell-filtering single microfluidic devices

a| 15x45µm rectangular pillars ; b| 5x15µm rectangular pillars



Trying to circumvent the abrupt increase of fluid velocity, geometry represented by

Fig. 30b was designed. As can be seen, the number of gaps is increased from 5 to 10.

However, pillars are thinner, which may have as a consequence more unstable and

less effective structures, compared with the robust pillars of Fig. 30a.

Thinner pillars are more susceptible to bend than robust ones, and this possibility is

higher when they suffer from effects of increased velocity. In an attempt to minimize the

effects of increased velocity and reduce not only damages on the structures but also

the loss of cells that are pushed through the gaps by the flow, channels in Fig. 31 were

designed. Channel represented by Fig. 31 is 100 µm wide, but the enlargement

reaches 500 µm wide. The sum of the gaps width (total width available for the flow

when passing a pillar row) 150 µm, 1.5 times the channel width. This conformation will

reduce the velocity of the fluid at the gaps.

(a)

(b)

(c)

Mask Structure

FIGURE 31| Geometries of enlarged cell-filtering microfluidic devices

a| 5x10µm rectangular pillars; b| diamond-shape pillars; b| pyramidal structures



Yeast cells have a cell wall that confers some rigidity to keep its form, but in the inside

they are full of liquid cytoplasm. When pushed through a gap they have the capability

to deform. Wherefore, it is possible, and probable, that cells, when subject to high

pressure gradients near the gap, deform and may be trapped, clogging it. Fig. 31b

represents a geometry designed keeping that in mind. The rectangular pillars were

substituted by diamond shape pillars, to reduce the length of the structure in which the

cell may get stuck. The main advantage is the reduced probability of clogging with the

chance of cells to get trapped on further structures.

Fig. 31c represents a different alternative to filter the cell suspension. The first

geometries developed had the pillars aligned, and the cells were trapped in a plane

perpendicular to the flow. This new geometry has intermediate gaps where the velocity

rate is supposed to decrease. The smallest cells should deposit between the pillars and

the larger ones in the taper gaps.

3.2.2 Two inlet devices

Two inlets-channels (Fig. 32) were also designed, trying to develop alternatives if the

one inlet geometries were prone to clogging.

The aim of the insertion of another inlet, is to use the properties of laminar flow, and

use one auxiliary stream to confine and direct the cell suspension.

For instance, as represented by Fig. 17 (hydrodynamic focus) from Section 5.4 of

Chapter II, it is possible to direct the cell suspension. The strategy employed is

injecting the cell suspension through the one inlet and sterile water through the other

inlet, focusing the cell flow to side of the microchannel, in both cases represented as

Fig. 32a or 32b. Cells would be entrapped in the lower region and, in case of clogging,

the cell suspension, and later on the remaining reactant streams, can be deviated from

the traps switching off the sterile water stream. The channel would not be totally

useless in case of clogging and the reagent would reach the cells by diffusion.

Figure 33 is a typical structure to perform hydrodynamic focusing. Sterile water is

injected by outer inlet and the cells suspension by the inner inlet. The streams contact

symmetrically at an intersection point and cells are focused in the flow central position,

until bump a horseshoe-like structure and be retained there.



Mask Structure

(a)

(b)

FIGURE 32| Geometries of two streams cell-filtering microfluidic devices a| arc aligned rectangular pillars ; b| rectangular pillars

Mask structure

(a)

FIGURE 33| Geometries of three streams cell-filtering microfluidic devices

a| horseshoe structure



4. Computational fluid dynamics analysis

Numerical simulations are able to determine the fluid behavior inside a microchannel

even before the production of the microchip. This is advantageous because save

resources and time. 2D simulation are assumed as a modelling of the central horizontal

section of the fluid. Figure 34 is based entirely on numerical result which shows the

velocity contours (velocity magnitude) and the stream function contours (streamlines)

of all geometries conceived. It is observable in all simulations that the velocity

increases from the wall to a maximum in the center of the channel, in the zone free of

pillars. In almost all simulations, the fluid flows through the traps, except in Simulation

X. Fig. 35 represents a contour plot in a vertical plane in the middle of the horseshoe

structure. As we can interpret, in the middle, the horseshoe structure, the velocity of the

fluid is near to zero, which means that in this zone the fluid is almost stagnant.

Analyzing the streamlines in the stream function contour is perfectly notorious that the

fluids detour from the structure and flow at the sides, not resulting the openings in the

horseshoe structures in a bottleneck effect, as expected. For that reason, the geometry

of three streams cell filtering device is not the best approach for cell entrapment.

Simulation III shows an interesting pattern of velocity magnitude between the pillars.

This pattern was investigated, to try to understand why this happens. Since a grid

independence analysis was made and the gaps have all the same size.

This set of simulations can be divided into subsets, according to the order of magnitude

and position of their points with maximum velocity. Straight channels have maximums

of velocity magnitude between pillars and the order of magnitude is 10-2. It is due to the

narrowing of cross sectional area between the pillars. Maximums of enlarged channels

(IV-VI) not always are in the detection zone, and its magnitude is ten times smaller.

This is an advantage of the channels with an enlarged region compared to the straight

ones because cells will be less prone to be dragged. For that reason, single straight

channels do not seem as appealing as the enlarged and were excluded from the set of

channels to test experimentally.



Simulation Contours of velocity magnitude (m/s) Contours of stream function

(kg/s)

I

Magnitude Range: [0; 5.60x10-2]

II


III


IV


V




VI


VII


VIII


X


FIGURE 34| Velocity contours and streamlines in the conceived geometries for microfluidic

devices along the horizontal plane of symmetry as predicted by the CFD simulation

performed



FIGURE 35| Contour plot in a vertical plane in the middle of the horseshoe structure from

Simulation X

Two inlet channels were conceived as backup options, in the case clogging was

observed in single channel devices. Although their maximum velocity magnitudes are

high, they were not immediately excluded. Simulation I does not represent an

appealing velocity profile because the velocity increases drastically in the gap. This can

result in the bent of the barrier, or the fluid can drag the cells through the gap.

However, this geometry is still being submitted to experimental tests because of its low

manufacturing costs.

5. Retention tests in channels with a dam-like structure

The first set of retention tests was performed in microfluidic devices that integrates

channels with a dam-like structure. The criteria for this choice was the relative low cost

producing the masks, since masks produced by xurography technique use adhesive

paper that are so much cheaper than SU-8 molds.

The elevated height of the barrier in opposition to the slight depth of the well, most of

the times, caused the collapse of the structure, closing the channel. The percentage of

successful fabricated channels was quite low.

Fig. 36a shows the results for the retention tests in channels with dam-like structures

using 10µm polystyrene microspheres. It is possible to see that spheres were retained

in the detection zone, but not where was expected. Fig. 36b supports the idea of similar

behavior between yeasts and PS particles. The channel was designed so that the cells

get trapped upstream the region of the dam, and what is observed is that they get stuck

in the well, the gap below the barrier. The fact can have two explanations. First, due to

the technique low accuracy, it is impossible to guarantee that the gap is 8µm high all



over the well, essentially because the ratio between the height of the barrier and the

height of the gap is very high, making possible a collapse of the structure. Second,

analyzing the image, it is perfectly visible that the well zone is not clean, it has some

residuals that are thought to be adhesive residue, from the adhesive paper. It is

impossible to remove the mold that forms the well without leaving some residuals of

adhesive, essentially because the adhesive paper glue is molten after the step of cure

of the PDMS in the oven for 20min at 80ºC. These residues are randomly and where

there are no residuals, the fluid flows. This molten adhesive problem also makes the

fabrication reproducibility impossible.

a. b.

FIGURE 36| Retention test in channels with a dam-like structure

a| PS particle retention b| S. cerevisiae retention.

Original magnification x 100.

2 µm microspheres were injected and entrapped into the microfluidic device, after the

injection of 10 µm microparticles to form a filter. The gaps tend to be smaller, and it is

expected that S. cerevisiae, injected after the microparticles, will be entrapped in an

easier and more efficient way.

When PS particles were used as a filter (Fig. 37), although yeast retention was

significantly higher and flow were effectively more difficult, because the gaps between

microparticles were so small, after some time, the microchannel clogged and it was

impossible to continue the experiments.



FIGURE 37| Usage of PS particles as filter for retention of S. cerevisiae in channels with a

dam-like structure

.Original magnification x 100.

It was concluded that microparticles as a filter were not a viable method, at least using

this method of microfabrication. As such, the microchannels with dam-like structures

were discarded.

6. Geometrical characterization of channels with pillar-based

filter

Molds of channels with pillar-based filter were ordered in two different depths, 50µm

and 30µm. To confirm the channel height, a cross section of the channels fabricated by

soft lithography technique was measured using Leica Application Suite Tools.

The analysis was made for 50 µm high channels, represented in Fig.38 . The geometry

chosen was the enlarged channel with 5x10µm pillars. As can be observed, the

channel has the expected height, but, pillars are bent. Square pillars are very high and

thin, and can collapse due to lack of support. For that reason, 50 µm channels are

excluded from the range of channels to test.



FIGURE 38| Geometrical characterization of a channel with a nominal height of 50µm


Analyzing 30 µm channels with other geometries, they seem more stable. Fig. 39

shows two cross sections of different geometries and the differences to the 50µm

channels, in terms of stability, are notorious.

a.

b.

FIGURE 39| Geometrical characterization of a channel with a nominal height of 30µm

a| pyramidal structures; b| diamond-shape pillars


The first criterium to choose a channel to test was related to the microfabrication

outcome. Channels with frailer pillars are more prone to flaws than channels with

robust ones. When it comes to fragile pillars, it was very frequent to miss a pillar

because it was plucked during demolding. Additionally, the pillars arrangement were

distorted due to the bending of the pillars. For instance, Fig. 40 shows a channel with

diamond-shaped pillars and where it is notorious that two pillars are missing, the fifth of

the first barrier and the fifteenth of the third barrier. The volume of the missing pillars is

a preferential escape to yeasts and has strong repercussions on their entrapment.



FIGURE 40| Channel with diamond-like pillars exhibiting barrier imperfections

Original magnification x100.

Fig. 41 shows a channel with 5x15µm rectangular pillars and, even being 30µm high,

the channels are bent. 5x15µm rectangular pillars proved to be very frail, maybe due to

the towering height in opposition to the reduced dimensions of the base.

FIGURE 41| Channel with 15x5 µm rectangular pillars exhibiting barrier imperfections


From the above, diamond shape and 5x15µm rectangular pillars were excluded from

the set of channels to test.

Fig. 42 is a microphotography of a microchannel with pyramidal structures. Some

pillars can lean but since the gap of the barrier is maintained, no relevant imperfections



are observed. The gaps obtained between the biggest pillars in the last line are

considerably uniform.

FIGURE 42| Channel with pyramidal structures


With the resources available, the volumetric measure of the channel has poor

precision and accuracy, but it was relevant to understand that with an inflow rate of

1µL/min during almost 1 min is enough to fill the channel.

The material testing did not show any relevant PDMS expansion, when in contact with

any FISH reagents.

7. Retention tests in single channels with pillar-based filter

Retention tests were performed in enlarged channels with pyramidal structures using

10µm PS spheres. Fig. 43 shows better results. The microspheres were perfectly

entrapped in the narrowing space between pyramidal structures.



FIGURE 43| Microspheres retention test for channels with pyramidal structures


In Fig. 43, an air bubble is observable in the right side. During retention tests, it was

very frequent to have bubbles dragging all the cells and fluid inside the channel. It is a

common issue in microfluidics procedures, and has to be minimized226.

Given the results for this geometry, the work advanced to similar tests using S.

cerevisiae cells. Fig 44 shows the entrapment of yeast cells. These results were not so

satisfactory as those for the microspheres retention test. S. cerevisiae diameter is

about 5-8µm (see Chapter III, section 1), smaller than 10 µm microspheres, meaning

that S. cerevisiae are more prone to pass through the gaps than the microspheres.

Also, yeasts are not bulky as microspheres. We can compare them with microcapsules,

since the cytoplasm are encapsulated by the membrane, which means that they are

deformed by pressure gradients in the gaps. In this retention test, S. cerevisiae prove

to be able to contract and pass throw the gaps, behavior that is not similar to the

experienced by microspheres. However, this characteristics also are advantageous for

the yeast to entry in the gaps perpendicular to the flow, being safeguarded, since it is

an objective of this work to depend only on hydrodynamic traps to retain the cells in the

detection zone.



FIGURE 44| S. cerevisiae retention test for channels with pyramidal structures


Two inlets devices were conceived as a backup plan, in case of clogging in all single

channel devices tested. As clogging did not appear as a problem, and after the

entrapment, the fluid continues to flow without visible limitations, two inlet devices were

excluded from the set of channels tested. This does not mean they can be used in

different assays in future work, but, for this work, a reasonable solution was achieved.

8. PNA-FISH methodology within the microfluidic channel

The implementation of the full PNA-FISH methodology offers great difficulties,

particularly due to the formations of air bubbles during all the process. Several attempts

failed as a result of bubble formation that dragged all the cells or completely blocked

the channel. Even by the end of this work, bubble formation was impossible to control

at a full extent, which compromises reproducibility. Fig. 45 illustrates the bubble

formation during the retention step. These frequent events can result in the rupture of

the channel and in the abortion of the assays. These problems may be due to the

insertion of bubbles already present in the tube connected to the microchip or, when

the pressure inside the channel is lower than the atmospheric pressure, the air can

enter by the connection between the needle of the tube and the chip.



FIGURE 45| Channel with pyramidal structures


Some changes were adopted, trying to solve this problem. To minimize the number of

times that the chip was disconnected and reconnected to the pumping system, 3µL of

each reagent were inserted in the tube sequentially, and each time that the sample had

to incubate, the syringe was placed inside the incubator together with the microchip.

This procedure was effective during the retention step, but, when the system

experienced a temperature increase during the hybridization step in the incubator,

bubbles were created inside the channel and the cells were dragged. This may occur

due to fluid evaporation in low pressure points. The solution was to immerge the chip in

water previously heated inside the incubator. This minimizes bubble formation but do

not solve it completely.

After controlling to some extent the problem related to air-bubbles, the entire PNA-

FISH methodology was performed within the microfluidic channel. Figure 46 shows the

results for this assay.

a. b.

FIGURE 46| Positive for PNA-FISH methodology within the microfluidic channel.

a| DAPI staining (N2.1 Filter) b| Positive (Dead/Live Filter). Images were obtained with equal exposure times. Original magnification x 100. Scale bar 20µm.



The DAPI staining (Fig. 46a) confirms the presence of cells after all the FISH

procedure. However, when observing the sample with the filter sensitive to the probe

(Fig. 46b), it is notorious that the hybridization only occurs in a few cells in the central

section of the channel. Due to the height of the upper block of the device, the higher

magnification that can be used is 10x (lenses magnifications). It is possible that the

hybridization solution is not passing through the traps very well or the flow is increased

in the central part and this zone is more exposed to the probe. This would explain the

higher fluorescence intensity of the cells trapped in the central part of the

microchannel.

Successful implementation of FISH procedures are reported by Sieben227,228 and

Vedarethinam229, and the work of Zhang126 that uses pillar-based filters to entrap cells

and successfully hybridizes RNA with quantum-dots shows that this procedure is

possible. Despite some limitations, the fact that a cluster of cells fluorescing red is

identified within the microchip works as a proof of concept. A trustable geometry was

developed but the PNA-FISH protocol within the microchip has however to be strongly

optimized.

Chapter V – Conclusions and Future Work



Conclusions

This work describes the development of a novel microfluidic device for the retention

of S. cerevisiae and subsequent identification by PNA-FISH. Unlike other devices, this

system is much less susceptible to clogging, as observed in Meireles177, and is much

simpler to prepare than the other systems128, while retaining a significant amount of S.

cerevisiae cells. Furthermore, it was one of the first times that PNA has been used

successfully in microchannels using simple hydrodynamic methods.

Strategies for successful retention of yeast cells are reported by Li et al230, using U-

shape barriers, which, according to Meireles177, is a strategy more difficult to handle,

since it is susceptible to fluid recess. Zhang et al231also proposes the use of dam-like

structures to retain yeast cells in a study of yeast aging. Some pillar-based geometries

for cell retention was also described232 but there are no evidences of geometries similar

to the ones described in this work. The chosen geometry allows cells to be entrapped

not only in the narrowing space between pyramidal structures but also in the gaps

perpendicular to the flow, as suggested by numerical simulations.

The PNA-FISH proved to be successful for S. cerevisiae identification, in spite of a

better signal being expected. Some new probe design can be performed in an attempt

to achieve a better signal. Regarding the microfabrication, no significant leakages were

observed.

Forrest et al.233, studied the impact of a diagnosis using PNA-FISH on the expenses

related to antifungal treatment. This study proves that although laboratory costs for

doing PNA-FISH test exceed those for blood culture, savings through a decrease in

antifungal drug costs, particularly caspofungin were obtained and can result in

substantial savings for hospitals. This supports the significance of rapid identification of

the microorganism. An appropriated and effective treatment drastically reduces

mortality. Unfortunately, standard microbiology laboratory methods are too slow to

support rapid interventions, requiring approximately 24 hours to detect the presence of

hematopathogens and at least 3 days to confirm the selection of appropriate

antimicrobial therapy due to low density of microbes and the need of an enrichment

step. Empiric treatment is common in these situations and, if it is not effective, may

contribute to the evolution of drug-resistant microbes and increase treatment costs234.

As proposed initially, this work validates the concept proposed. Reproducible

retention of S. cerevisiae was ensured and its detection within the microchannel by

fluorescence microscopy was successful and unequivocally achieved.

CHAPTER V – CONCLUSIONS AND FUTURE WORK

Future work

The formation of air bubble was the main issue during the development of the device.

The process of cells retention could be optimized, if it was possible to perform PNA-

FISH under continuous flow, avoiding reflux. It would be necessary a heated plate with

high accuracy, since the hybridization process is extremely sensitive. Also, in an ideal

situation of continuous flow, the outlet of the channel could be elevated, increasing the

pressure inside the channels and preventing the formation of air bubbles. The problem

of air bubbles is compounded by the fact that the FISH procedure needs to hybridize at

high temperatures, which leads to the evaporation of organic reagents in low pressure

points226. Also, it is thought that if the time of the procedure could be reduced, and

Zhang126 proved it possible, organic reagents would be less prone to evaporate. This

definitely demands a PNA-FISH optimization to the new conditions. The design of a

bubble entrapment device, suggested by Zheng235, is very appealing to solve the

bubble problem. The surface may become hydrophilic if it is exposed to oxygen

plasma, solving the bubble problem147.

As a final remark, to determinate the device’s limit of detection, a flow rate and yeast

concentration optimization should also be performed.


References 1 World Health Organization (2014) Antimicrobial resistance: Global Report on Surveillance. WHO Library Cataloguing-in-Publication Data. ISBN 978 92 4 156474 8

2 BURGMANN, H et al. (2010) Impact of nosocomial infections on clinical outcome and

resource consumption in critically ill patients. Int Care Med 36: 1597-601

3 STRUELENS MJ (1998) The epidemiology of antimicrobial resistance in hospital-acquired

infections: problems and possible solutions. BMJ 317: 652-4

4 ANGUS, DC; WAX RS (2001) Epidemiology of sepsis: an update. Crit Care Med 29: S109-16

5 MAYR, FB; YEND, S; ANGUS, DC (2014) Epidemiology of severe sepsis. Virulence 5:4-11

6MARTIN, GS (2012) Sepsis, severe sepsis and septic shock: changes in incidence, pathogens

and outcomes. Expert Ver Anti Infect Ther 10:701-6

7 WOLK DM; FIORELLO AB (2010) Code Sepsis: Rapid Methods To Diagnose Sepsis and Detect

Hematopathogens Part II: Challenges to the Laboratory Diagnosis of Sepsis Clinical Microbiology Newsletter 32: 41-49.

8 PARHAM, P. (2009), The Immune System, 3rd Edition, ISBN 978-0-8153-4146-8 (Garland Science)

9 McCULLOUGH MJ, ROSS BC, READE PC (1996), Candida albicans: a review of its history, taxonomy,

epidemiology, virulence attributes and methods of strain differentiation Int J Oral Maxillofac Surg, 25:136-144

10 CORNER BE, MAGEE PT (1997), Candida pathogenesis: unravelling the threads of infection

Curr Biol 7: R691-R694

11 CALDERONE RA (2002), Candida and candidiasis ISBN 1-55581-212-0 (ASM PRESS)

12 LEBERER E. et al (1997), Virulence and hyphal formation of Candida albicans require the Ste20p-like protein kinase CaCla4p Curr Biol, 7:539-546

13 LO HJ et al (1997), Nonfilamentous mutants of C. albicans are avirulent Cell, 90:939-949.

14 FILLER SG, SHEPPARD DC (2006), Fungal invasion of normally nonphagocytic host cells PLoS Pathog 12: 1099-1105

15 THOMPSON DS, CARLISLE PL, KADOSH D (2011) Coevolution of morphology and virulence in

Candida species Eukaryot Cell 10: 1173-1182

16 TASCHDJIAN CL, BURCHILL JJ, KOZINN PJ (1960) Rapid identification of Candida albicans by

filamentation on serum and serum substitutes. AMA J Dis Child 99:212

17 BUFFO, J et al. (1985) A characterization of pH regulated dimorphism in Candida albicans. Mycopathologia 85, 21-30

18 MARDON D, BALISH E, PHILIPS AW (1969) Control of dimorphism in a biochemical variant of

Candida albicans.J Bacteriol 100:701-707

19 SIMONNETI N, STRIPOLLI V, CASSONE EA (1974) Yeast-mycelial conversion induced by N-

acetyl-D-glucosmine in Candida albicans. Nature 250:344-346


20 LEE KL, BUCKLEY HR, CAMPBELL CC (1975) An amino acid liquid synthetic medium for the

development of mycelial and yeast forms of Candida albicans. Sabouraudia 13: 148-53

21 LIU HP, KOHLER JR, FINK GR (1994) Suppression of hyphal formation in Candida albicans by

mutation of a STE12 homolog. Science 266: 1723-6

22 GOW, NAR et al. (2012) Candida albicans morphogenesis and host defence: discriminating

invasion from colonization Nat Rev Microb 10:112-122

23 LERNER KL, LERNER BW (2002) World of Microbiology and Immunology, 1st Edition, ISBN 0787665401, (Gale)

24 LEBERER E, THOMAS DY, WHITEWAY M (1997) Pheromone signaling and polarized

morphogenesis in yeast. Curr Opin Genet Dev 7:59-66

25 SUDBERY PE (2001) The germ tubes of Candida albicans hyphae and pseudohyphae show

different patterns of septin ring localization. Mol Microbiol 41: 19-31

26 VIRAG A, HARRIS SD (2006) The spitzenkorper: a molecular perspective. Mycol Res 110:4-13

27 SUDBERY P, GOW N, BERMAN J (2004) The distinct morphogenic state of Candida albicans.

Trends Microbiol 12:317-324

28 FINLEY KR, BERMAN J (2005) Microtubules in Candida albicans hyphae drive

nuclear dynamics and connect cell cycle progression to morphogenesis. Eukaryot Cell 4:1697-711

29 BERMAN J (2006) Morphogenesis and cell cycle progression in Candida albicans. Curr Opin Microbiol 9:595-601

30 SUDBERY PE (2011) Growth of Candida albians hyphae. Nat Ver Microbiol 9:737-48

31 TSAI, PW et al (2013) Study of Candida albicans and its interaction with the host: A mini

review Biomedicine 3: 51-64

32 NAVARATHNA DH et al. (2007) Effect of farnesol on mouse modelo f systemic candidiasis , determined by use of a DPP3 knockout mutante of Candida albicans Infect Immun 75: 1609-1618

33 BOUZA E; MUÑOZ P(2008) Epidemiology of candida in intensive care units. Int J Antimicrob Agents.

32:Suppl 2 S87–S91

34 TORTORANO AM, et al. (2006) Candidaemia in Europe: epidemiology and resistance. Int J Antimicrob

Agents. 27:359–366

35 HORN, DL et al. (2009). Epidemiology and outcomes of candidemia in 2019 patients: data from the

prospective antifungal therapy aliance registry. Clin Infect Dis. 48:1695–1703

36 CHAHOUD J: KANAFANI ZA; KANJ SS (2013) Management of candidaemia and invasive candidiasis in critically ill patients. Int J Antimicrob Agents. Pii: S0924-8579

37 MORGAN J, et al.(2005) Excess mortality, hospital stay, and cost due to candidemia: a case–controlstudy using data from population-based candidemia surveillance. Infect Con-trol Hosp Epidemiol. 26:540–547

38 OBEROI JK (2010) Invasive candidiasis JIMSA 23: 25-28

39 ODDS FC (1998), Candida and candidosis, 2nd Edition, ISBN 0-7020-1265-3 (Balliere Tindall)


40 WITEK-JANUSEK L, SHAREEF MJ, MATHEWS HL (2002) Reduced lymphocyte mediated

antifungal capacity in high-risk infants J Infect Dis 186: 129-133

41 NAVARATHNA D et al. (2013) MRI confirms loss of blood-brain barrier integrity in a mouse

modelo f disseminated candidiasis NMR Biomed

42 SUDBERY P, GOW N, BERMAN J (2004) The distinct morphogenic states of Candida albicans Trends Micobiol 12: 317-324

43 KRCMERY V, BARNES AJ (2002) Non-albicans Candida spp. Causing fungaemia: pathogenicity and antifungal resistance J Hosp Infect 50:243-260

44 KOTHAVADE RJ et al. (2010) Candida tropicalis: its prevalence, pathogenicity and increasing

resistance to fluconazole J Med Microbiol 59: 873-880

45 COLOMBO AL et al. (2007) Prospective observational study of candidemia in São Paulo, Brazil: incidence rate, epidemiology, and predictors of mortality Infect Control Hosp Epidemiol 28: 570-576

46 NUCCI M et al. (2010) Epidemiology of opportunistic fungal infections in Latin America Clin Infect Dis 51: 561-570

47 PIRES-GONÇALVES RH et al (2007) Genetic relatedness of comensal strains of Candida

albicans carried in the oral cavity of patients’ dental prosthesis users in Brazil

Mycopathologia 164:255-263

48 CANTÓN E et al. (2011) Prospective multicentre study of the epidemiology, molecular identification, and antifungal susceptibility of Candida parapsilosis, Candida orthopsilosis,

and Candida metapsilosis isolated from patients with candidemia. Antimicrob Agents Chemother 55: 5590-5596

49 AJENJO H et al (2011) Epidemiological profile of invasive candidiasis in intensive care units

at a university hospital Ver Chilena Infectol 28: 118-122

50 PAULO C et al. (2009) Retrospective analysis of clinical yeast isolates in a hospital in the centre of Portugal: spectrum and revision of the identification procedures Med Mycol 19:1-10

51 MOUDGAL V, SOBEL J (2010) Antifungals to treat Candida albicans. Expert Opin Pharmacother 11:2037-48

52 PFALLER MA et al (2006) Candida rugosa, na emerging fungal pathogen with resistance to azoles: geographic and temporal trends from the ARTEMIS DISK antifungal surveillance program J Clin Microbiol 44: 3578-3582

53 VARGAS SL et al. (1993) Modulating effect of dietary carbohydrate supplementation on Candida albicans colonization and invasion in neutropenic mouse model. Infect Immun 61:619-26

54 RODAKI A et al. (2009) Glucose promotes stess resistance in the fungal pathogen Candida

albicans. Mol Biol Cell 20: 4845-55

55 KUMAR A et al (2006) Duration of hypotension before initiation of effective antimicrobial therapy is the critical determinant of survival in human septic shock Crit Care Med 34: 1589-1596

56 AZOULAY E et al (2004) Practice in non-neutropenic ICU patients with Candida-positive

airway specimens Intensive Care Med 30: 1384-1389


57 NEPPELENBROEK KH et al. (2014) Identification of Candida species in the clinical laboratory: a review of conventional, comercial, and molecular techniques. Oral Dis 20:329-44

58 SINGH G et al (2013) Candidal infection: epidemiology, pathogenesis and recente advances

for diagnosis. BOPAMS 1: 1-8

59 HOPPE JE, FREY P (1999) Evaluation of six comercial tests and the germ-tube test for presumptive identification of Candida albicans Eur J Clin Microbiol Infect Dis 18:188-191

60 KIM TH et al. (2010) Candida dubliniensis screening using the germ tube test in clinical

yeast isolates and prevalence of C. dubliniensis in Korea. J Clin Lab Anal 24:145-148

61 STAIB P, MORSCHHAUSER J (2007) Chlamydospore formation in Candida albicans and

Candida dubliniensis – an enigmatic developmental program Mycoses 50: 1-12

62 KIM D et al (2002) Rapid differentiation of Candida albicans from other Candida species

using its unique germ tube formation at 39 degrees C Yeast 19: 957-962

63 ROBERTS GD et al (1976) Evaluation of API 20 c microtube system for the identification of clinically important yeasts J Clin Microbiol 3:302-395

64 HORVATH LL et al (2003) Direct isolation of Candida spp. From blood cultures on the

chromogenic medium CHROMagar Candida J Clin Microbiol 41:2629-2632

65 FRICKER-HIDALGO et al. (1996) Comparison of the new API Candida system to the ID 32C

system for identification of clinically important yeast species J Clin Microbiol 34:1846-1848

66WOLK DM; FIORELLO AB (2010) Code Sepsis: Rapid Methods To Diagnose Sepsis and Detect

Hematopathogens Part II: Challenges to the Laboratory Diagnosis of Sepsis Clinical Microbiology Newsletter 32: 41-49.

67NEPPELENBROEK KH et al. (2006) Molecular fingerprinting methods for the discrimination between

C. albicans and C. dubliniensis. Oral Dis 12:242-53

68 ERLICH HA (1989) Polymerase Chain Reaction. J Clin Immunol 9: 437-47

69 CARVALHO A, et al(2007) Multiplex PCR identification of eight clinically relevant Candida

species Med Mycol 45: 619-627

70 BOUGNOUX M, et al (1999) Serum is more suitable than whole blood for diagnosis of systemic candidiasis by nested PCR J Clin Microbiol 37: 925-930

71 GUIVER M, LEVI K, OPPENHEIM BA (2001) Rapid identification of Candida species by TaqMan PCR J Clin Pathol 54: 362-366

72 SPANU T et al (2012) Direct MALDI-TOF mass spectrometry assay of blood culture broths for rapid identification of Candida species causing bloodstream infections: an observational

study in two large microbiology laboratories J Clin Microbiol 50: 176-179

73DHIMAN N et al (2011) Performance and cost analysis of matrix-assisted laser desorption ionization-time of flight mass spectrometry for routine identification of yeast J Clin Microbiol 49: 1641-1616

74CERQUEIRA L: et al (2008) DNA Mimics for the Rapid Identification of Microorganisms by

Fluorescence in situ Hybridization (FISH) Int J Mol Sci 1944-1960


75

ALMEIDA C; et al. (2010) Fluorescence In Situ Hybridization Method Using a Peptide Nucleic Acid

Probe for Identification of Salmonella spp. in a Broad Spectrum of Samples Appl Environ Microbio 4476-4485

76 AMANN R, FUCHS BM (2008) Single-cell identification in microbial communities by

improved fluorescence in situ hybridization. Nat Rev Microbiol 6:339-348

77 WAGNER M (2003) Fluorescence in situ hybridisation for the identification and characterisation of prokaryotes. Curr Opin Microbiol 6:302-9

78 SPECTOR DL, GOLDMAN RD (2006) Basic methods in microscopy: protocols and concepts

from cells. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY

79 JAVOIS LC (1999) Immunocytochemical methods and protocols. 2nd Ed. Humana Press.

80 OLIVEIRA K et al. (2001) Differentiation of Candida albicans and Candida dubliniensis by fluorescente in situ hybridization with peptide nucleic acid probes. J Clin Microbiol 39:4138-41

81 KIERNAN JÁ (1981) Histological and histochemical methods. Theory and practice. Pergamon, Oxford, UK

82 BHAIRI SM (2001) Detergents: a guide to the properties and uses of detergents in biological

systems. Calbiochem – Novabiochem Corporation, San Diego, CA.

83 MELAN MA, SLUDER G (1992) Redistribution and differential extraction of soluble proteins in permeabilized cultured cells: implications for immunofluorescence microscopy. J. Cell Sci 101:731-43

84 AMANN RI et al (1992) Identification of individual prokaryotic cells by using enzyme-labeled, rRNA targeted oligonucleotide probes. Appl Environ Microbiol 58: 3007-11

85 WARNER JR (1999) The economics of ribosome byosynthesis in yeast. Trends Biochem Sci 24:437-40

86 STENDER H et al (2002) PNA for rapid microbiology. J Microb Method 48:1-17

87 INÁCIO J et al (2003) In situ accessibility of Saccharomyces cerevisiae 26S rRNA to Cy3-labeled oligonucleotide probes comprising the D1 and D2 domains. Appl Environ Mibrobiol 69, 2899-905

88 AMANN R, LUDWIG W (2000) Ribossomal RNA-targeted nucleic acid probes for studies in

microbial ecology. FEMS Microbiology Reviews 24:555-65

89 YILMAZ LS, NOGUERA DR (2004) Mechanistic approach to the problema of hybridization

efficiency in fluorescent in situ hybridization. Appl Environ Microbiol 70:7126-39.

90 WATSON, JD; CRICK, FH (1953) Molecular structure of nucleic acids; a structure for

deoxyribose nucleic acid. Nature 171: 737-8

91 SCHILDKRAUT CL, MARMUR J, DOTY P (1961) The formation of hybrid DNA molecules and their use in studies of DNA homologies. J Mol Biol 3:595-617

92 HOYER BH; McCARTHY BJ; BOLTON ET (1964) A molecular approach in the systematics of higher organism. DNA interactions provide a basis for detecting common polynucleotide

sequences among diverse organisms. Science 22: 959-67.

93 McCARTHY BJ; HOYER BH (1964) Identify of DNA and diversity of messenger RNA molecules in normal mouse tissues. Proc Natl Acad Sci USA 52:915-22


94 PARDUE ML; GALL JG (1969) Molecular hybridization of radioactive DNA to the DNA of

cytological preparations. Proc Natl Acad Sci USA 64:600-4

95 RUDKIN GT, STOLLAR BD (1977) High resolution detection of DNA-RNA hybrids in situ by

indirect immunofluorescence. Nature 265:472-3

96 BAUMAN, JG et al. (1980) A new method for fluorescence microscopical localization of specific DNA sequences by in situ hybridization of fluorochrome labelled RNA. Exp Cell Res 128: 485-490

97 HARPER ME, SAUNDERS GF (1981) Localization of single copy DNA sequences of G-banded

human chromosomes by in situ hybridization. Chromosoma 83: 431-9

98 WAHL GM; STERN M; STARK GR (1979) Efficient transfer of large DNA fragments from agarose gels to diazobenzyloxymethyl-paper and rapid hybridization by using dextran sulfate. PNAS USA 79: 3683-7

99 LANGER PR; WALDROP AA; WARD DC (1981) Enzymatic synthesis of biotin-labeled polynucleotides: novel acid affinity probes. Proc Natl Acad Sci USA 78:6633-7

100 MANUELIDIS L; LANGER-SAFER PR; WARD DC (1982) High-resolution mapping of satellite DNA using

biotin-labeled DNA probes. J Cell Biol 95:612-25

101 SINGER RH; WARD DC (1982) Actin gene expression visualized in chiken muscle tissue culture by

using in situ hybridization with a biotinated nucleotide analog. Proc Natl Acad Sci USA 79: 7331-5

102 DELONG EF; WICKHAM GS; PACE NR (1989) Phylogenetic stains – ribossomal RNA-based

probes for the identification of single cells. Science 243:1360-3

103NIELSEN, PE et al (1991) Sequence-selective recognition of DNA by strand

displacement with a thymine-substituted polyamide. Science 254:1497-500

104 FONTENETE S. et al. (2013) Hybridization-based detection of Helicobacter pylori at human

body temperature using advanced locked nucleic acid (LNA) probes. PLOS ONE 8: e81230

105 PETERSEN M, WENGEL L (2003) LNA: A versatile tool for the therapeutics and genomics.

Trends Biotechnol 21: 74-81

106 SØE MJ et al. (2011) A sensitive alternative for microRNA in situ hybridizations using

probes of 2’-O-methyl RNA + LNA. J Histochem Cytochem 59:661-72

107 EGHOLM M et al (1993) PNA hybidizes to complementry oligonucleotides obeying the Watson-Crick hydrogen bonding rules. Nature 365: 556-8

108 NIELSEN PE (2001) Peptide nucleic acid: a versatile tool in genetic diagnostics and

molecular biology. Curr OPin Biotechnol 12:16-20

109 NIELSEN PE, HAAIMA G (1997) Peptide nucleic acid (PNA): a DNA mimic with a

pseudopeptide backbone. Chem Soc Rev 26:73-8

110 GILDEA BD et al (1998) PNA solubility enhancers. Tetrahedron Lett39:7255-8

111 YLMAZ LS, NOGUERA DR (2007) Development of thermodynamic models for simulating probe dissociation profiles in fluorescence in situ hybridization. Biotechnol Bioeng 96:349-63

112 DEMINOV VV et al. (1994) Stability of peptide nucleic acids in human serum and cellular

extracts. Biochem Pharmacol 48:1310-3.


113 HEO J, HUA SZ (2009) An overview of recent strategies in pathogen sensing. Sensors 9: 4483-501

114 LI, D (2008) Encyclopedia of microfluidics and nanofluidics. Springer Science+Business Media, LLC, NY USA

115 YEON JH, PARK JK (2007) Microfluidic cell culture systems for celular analysis. Biochip J 1:17-27

116 YOUNG, EWK; ZERINGUE HC (2010) Fundamentals of microfluidic cell culture in controlled microenvironments. Chemical Society Reviews 39: 1036-48

117 KATSIKOGIANNI M; MISSIRLIS YF (2004) Concise review of mechanisms of bacterial adhesion to biomaterials and of the techniques used in estimating bacteria-material

interactions. European cells & materials 8:37-57

118 KIM, L et al. (2007) A practical guide to microfluidic perfusion culture of adherent

mammalian cells. Lab on a chip 9:2132-9

119 KANE, RS et al. (1999) Patterning proteins and cells using soft lithography Biomaterials 20:2363-76

120 DAVIDSON R et al. (2004) Microfluidic biosensing systems – Part II. Monitoring the dynamic production of glucose and ethanol from microchip-immobilised yeast cells using enzymatic

chemiluminescent µ-biosensors. Lab Chip 4:488

121 SKELLEY AM et al. (2009) Microfluidic control of cell pairing and fusion. Nature Methods 6:147-52

122 LAPIZCO-ENCINAS BH et al (2004) Insulator-based dielectrophoresis for the selective

concentration and separtion of live bacteria in water. Electrophoresis 25:1695-704

123 MOFFITT JR et al. (2008) Recent advances in optical tweezers. Annu Rev Biochem 77:205-28

124 PAMME N (2006) Magnetism and microfluidics. Lab chip 6:24-38

125 BAZOU D; KUZNETSOVA LA; COAKLEY T (2005) Physical environment of 2D animal cell aggregates formed in a short pathlength ultrasound standing wave trap. Ultrasound Med Biol 31:423-30

126 FAIR RB (2007) Digital microfluidics: is a true lab-on-a-chip possible? Microfluid Nanofluid 3:245-81

127 THORSLUND S et al. (2006) A hybrid poly(dimethylsiloxane) microsystem for on-chip

whole blood filtration optimized for steroid screening. Biomed Microdev 8:73-9

128 ZHANG Q et al. (2006) Microbial detection in microfluidic devices through dual staining of

quantum dots-labeled immunoassay and RNA hybridization. Analytica Chimica Acta 556:171-7

129 YAMAZOE H et al. (2014) Facile immunostaining and labeling of nonadherent cells using a

microfluidic device to entrap the cells. J Biosci and Bioeng 117: 375-8

130 YANG X et al. (1993) Micromachined membrane particle filters. Sens Actuators A Phys 73:184-91

131 ANDERSSON H et al (2000) Micromachined flow-through filter chamber for chemical

reactions on beads. Sens Actuators B Chem 67: 203-8

132 WHITESIDES, GM (2006). The origins and the future of microfluidics. Nature. 442: 368-373


133 GATES BD et al (2004) Unconventional Nanofabrication Annu Rev Mater Res 34: 339-72

134 BECKER H; LOCASCIO LE (2002) Polymer microfluidic devices. Talanta 56:267-87

135 SHAW JM et al (1997) Negative photoresist for optical lithography. IBM J Res Dev 41:81-94

136 MAZUTIS L et al (2013) Single-cell analysis and sorting using droplet-based microfluidics. Nat Protoc 8: 870-91

137 DUFFY DC et al. (1998) Rapid prototyping of microfluidic systems in polydimethylsiloxane

(PDMS) Anal Chem 70: 4974-84

138 ANDERSON JR et al. (2000) Fabrication of topologically complex three-dimensional

microfluidic systems in PDMS by rapid prototyping. Anal Chem 72: 3158-64

139 GREER J et al (2007) Comparision of glass etching to xurography prototyping of microfluidic channels for DNA melting analysis. J Micromech Microeng 17: 2407-13

140 TREISE I et al (2005) Efficient energy based modeling and experimental validation of liquid

filling in planar microfluidic componentes and networks. Lab on a chip 5:285-97

141BARTHOLOMEUSZ, DA (2005). Xurography: rapid prototyping of microstructures using a cutting

plotter Journal of Microelectromechanical systems, 14: 1364-1374

142 GALE , BK et al (2008). Low-Cost MEMS Technologies. Comprehensive Microsystems (pp. 1–

36).

143 XIA, Y; WHITESIDES GM (1998) Soft lithography. Angewandte Chemie International Edition 37: 550-

575

144 QUAKE SR; SCHERER A (2000) From micro- to nanofabrication with soft materials. Science 290:1536-40

145 KITSARA M; DUCRÉE (2013) Integration of functional materials and surface modification for polymeric microfluidic systems. J Micromech Microeng 23: 033001

146 CRISTENSEN AM; CHANG-YEN DA; GALE BK (2005) Characterization of interconnects used in

PDMS mirofluidic systems J Micromech Microeng 15:928

147 SIA, SK; WHITESIDES GM (2003) Microfluidics devices fabricated in poly(dimethylsiloxane)

for biological studies. Electrophoresis 24: 3563-76

148 BHAGAT AA, JOTHIMUTHU P, PAPAUTSKY I (2007) Photodefinable polydimethylsiloxane

(PDMS) for a rapid lab-on-a-chip prototyping. Lab Chip 7: 1192-7

149 McDONALD JC et al (2000) Review General Fabrication of microfluidic systems in poly(dimethylsiloxane) Electrophoresis 21:27-40

150 BOWEN J, CHENELER D, ROBINSON APG (2012) Direct e-beam lithography of PDMS.

Microelectronic Engineering 97: 34-7

151 MATA A, FLEISCHMAN AJ, ROY S (2005) Characterization of polydimethylsiloxane (PDMS)

properties for biomedical micro/nanosystems. Biomed Microdev 7:281-93

152 BECKER H, HEIM U (2000) Hot embossing as a method for the fabrication of polymer high aspect ratio structures. Sensors and Actuators A: Physical 83: 130-5

153 ATTIA UM, MARSON S, ALCOCK JR (2009) Micro-injection moulding of polymer microfluidic

devices. Microfluid Nanofluid 7: 1-28


154 LOCASCIO LE et al. (2006) Fabrication of polymer microfluidic systems by hot embossing

and laser ablation. Methods Mol Biol 339: 37-46

155 BEEBE DJ et al. (2002) Physics and Applications of Microfluidics in Biology Annual Review of

Biomedical Engineering 4: 261-286

156 BUSSCHER HJ, VAN DER MEI hc. (2006) Microbial Adhesion in Flow Displacement Systems. Clin Microbiol Rev, 19(1), 127–141.

157 KUDELA H Viscous flow in pipe (http://fluid.itcmp.pwr.wroc.pl/~znmp/dydaktyka/fundam_

FM/Lecture13.pdf) (Adapted)

158 WHITE FM (2003) Fluid Mechanics, 5th edition ISBN 0072402172 (McGraw-Hill Professional)

159 SQUIRES TM; QUAKE SR (2005) Microfluidics: Fluid physics at the nanoliter scale. Ver Mod Phys. 77:

977-1026

160 TIRADO MM, MARTINEZ CL, GARCIA DE LA TORRE J (1990) Comparison of theories for the translational and rotational diffusion coefficients of rod like macromolecules. Application to

short DNA fragments. J Chem Phys 81: 2047-52

161 DAMEAN N, REGTIEN PPL, ELWENSPOEK M (2003) Heat transfer in a MEMS for

microfluidics. Sensors and Actuators A 105: 137-49

162 LOCASCIO L, PERSO C, LEE C (1999) Measurement of electroosmotic flow in plastic imprinted microfluid devices and the effect of protein adsorption on flow rate. J Chromatogr A 857: 275-84

163 MARK D et al (2010) Microfluidic lab-on-a-chip platforms: requirements, characteristics

and applications. Chem Soc Rev 39: 1153-82

164 GUDIPATY T et al (2010) Cluster formation and growth in microchannel flow of dilute

particle suspensions. Microfluidics and Nanofluidics 10: 661-9

165 EIMER W et al (1990) Characterization of the overall and internal dynamics of short

oligonucleotides by depolarized dynamic light scattering and NMR relaxation measurements.

Biochemistry 29:799-811

166 OH KW, AHN CH (2006) A review of microvalves. J Micromech Microeng 16: R13-R39

167 ERICKSON D (2005) Towards numerical prototyping of lab-on-chip: modeling for

integrated microfluidic devices. Microfluid Nanofluid 1: 301-18

168 NORTON T, TIWARI B, SUN DW (2013) Computational fluid dynamics in design and analysis of termal processes: a review of recente advances. Critical Reviews in food science and nutrition 53:251-75

169 ANDERSSON B et al. (2011) Computational fluid dynamics for engineers. Cambridge University Press, Cambridge

170 VERSTEEG HK, MALALASEKERA W (2007) An introdution to computational fluid dynamics:

The finite volume method, 2nd Ed., Pearson Education Limited, Edinburgh

171 JUAREZ LH (2012) Fluid dynamics, computational modeling and applications. InTech, Rijeka, Croatia

172 LIMA ER, TAVARES FW, BISCAIA Jr EC (2007) Finite volume solution of the modified Poisson-Boltzmann equation for two coloidal particles. Phys Chem Chem Phys 9:3174-80


173 FERZIGER J, PERIC M (2002) Computational methods for fluid dynamics. Springer-Verlag. Berlin

174 FELDMANN, H (2010) Yeast: Molecular and Cell Biology. Wiley-Blackwell, EUA

175 LARSSON C et al. (1993) Growth and metabolism of Saccharomyces cerevisiae in chemostat cultures under carbon-, nitrogen-, or carbon- and nitrogen-limiting conditions. J Bacteriol 175: 4809-16

176 ANDORRA I et al. (2011) Analysis and direct quantification of Saccharomyces cerevisiae and Hanseniaspora guilliermondii populations during alcoholic fermentation by fluorescence in

situ hybridization, flow cytometry and quantitative PCR. Food microbiol 28: 1483-91

177 MEIRELES F (2012) PNA-FISH with microfluidics. Biomedical Engineering, Porto, Faculty of Engineering.

178 PERRY-O’KEEFE, H et al. (2001) Identification of indicator microorganisms using a

standardized PNA FISH method. J Microbiol Methods 47: 281-92

179 AZEVEDO, NF (2005) Survival of Helicobacter pylori in drinking water and associated biofilms. Chemical and Biological Engineering, University of Minho and University of Southampton

180 CERQUEIRA, L et al. (2011) PNA-FISH as a new diagnostic method for the determination of

clarithromycin resistance of Helicobacter pylori. BMC Microbiol 11:101

181 GUIMARÃES N et al. (2007) Deve Development and Application of a Novel Peptide Nucleic Acid Probe for the Specific Detection of Helicobacter pylori in Gastric Biopsy Specimens. J Clin Microbiol 45:3089-94.

182 RELLER ME et al. (2007) Use of peptide nucleic acid-fluorescence in situ hybridization for

definitive, rapid identification of five common Candida species. J clin Microbiol 45: 3802-3

183 RIGBY S et al (2002) Fluorescence in situ hybridization with peptide nucleic acid probes for rapid identification of Candida albicans directly from blood culture bottles. J clin Microbiol 40:2182-6

184 HARRIS DM, HATA DJ (2013) Rapid identification of bacteria and candida using pna-fish from blood and peritoneal fluid cultures: a retrospective clinical study. Ann Clin Microbiol and Antimicrob 12:2

185 SHEPARD JR et al. (2007) Multicenter Evaluation of the Candida albicans/ Candida glabrata

peptide nucleic acid fluorescent in situ hybridization method for simultaneous dual-color identification of C. albicans and C. glabrata directly from blood culture bottles. J Clin Microbiol 46: 50-55

186 Anon. Fixation and permeabilization in IHC/ICC. abcam 2014a [cited 27/5/2014]. Available from http://docs.abcam.com/pdf/protocols/Permeabilization-and-fixation.pdf

187 Anon. In-cell western™ assay – Cell fixation/permeabilization. Li-Cor 2014b [cited 27/5/2014]. Available from http://biosupport.licor.com/docs/ICW_fix_and_perm.pdf

188 FARREL RE (2009) RNA methodologies: Laboratory guide for isolation and characterization. 4th Ed. Academic Press, USA

189 PINTO M (2013) Study under dynamic environment of bacterial adhesion to a model

substrate in the presence of proteins Biomedical Engineering, Porto, Faculty of Engineering


190 MARCET-HOUBEN M, GABALDÓN T (2010) Aquisition of prokaryotic genes by fungal

genomes. Trends Genet 26:5-8

191 OSTERGAARD S, OLSSON L, NIELSEN J (2000) Metabolic Engineering of Saccharomyces

cerevisiae. Microbiol and Mol Biol Rev 63: 34-50

192 FELDMANN, H (2010) Yeast: Molecular and Cell Biology. Wiley-Blackwell, EUA

193 KURTZMAN CP, FELL JW, BOEKHOUT T (2011) The yeast: a taxonomic study. Elsevier, Amsterdam

194 GAGIANO M et al (2002) The sensing of nutritional status and the relationship to filamentous growth in Saccharomyces cerevisiae. FEMS Yeast Res 2: 433-70

195 GIMENO CJ et al (1992) Unipolar cell divisions in yeast S. cerevisiae lead to filamentous

growth: regulation by starvation and RAS. Cell 68:1077-90

196 AGUILAR-USCANGA B, JM FRANÇOIS (2003) A study of the yeast cell wall composition and structure in response to growth conditions and mode of cultivation. Lett Appl Microbiol 37: 268-74

197 RUIZ-HERRERA J et al. (2006) Molecular organization of the cell wall of Candida albicans and its relationship to pathogenicity. FEMS Yeast Res 6: 14-29

198 LIPKE PN, OVALLE R (1998) Cell Wall architecture in yeast: new structure and new

challenges. J Bacteriol 180: 3735-40

199 KÓLLAR R et al. (1995) Architecture of the yeast cell wall. The linkage between

chitin and beta (1��3) glucan. J Biol Chem 270:1170-8

200 KAPTEYN JC et al (1997) Altered extend of cross-linking of β1,6-glucosylated mannoproteins to chitin in Saccharomyces cerevisiae mutants with reduced cell wall β1,3-

glucan content. J Bacteriol 179: 6279-84

201 VALDIVIESO MH et al. (2000) Chitin synthesis in a gas1 mutant of Saccharomyces

cerevisiae. J Bacteriol 182: 4752-7

202 SMITS GJ et al. (2006) Role of cell cycle-regulated expression in the localized incorporation

of cell wall proteins in yeast. Mol Biol Cell 17: 3267-80

203 DE NOBEL JG et al. (1990) The glucanase-soluble mannoproteins limit cell wall porosity in

Saccharomyces cerevisiae. Yeast 6: 491-9

204 KURANDA MJ, ROBBINS PW (1991) Chitinase is required for cell separation during during

growth of Saccharomyces cerevisiae. J Biol Chem 266: 19758-67

205 BROACH JR (2012) Nutritional control of growth and development in yeast. Genetics 192: 73-105

206 SANTALUCIA J, ALLAWI HT, SENEVIRATNE PA (1996) Improved nearest-neighbor parameter

for predicting DNA duplex stability. Biochemistry 35: 3555-62

207 YILMAZ L, OKTEN H, NOGUERA D (2006) Making all parts of the 16S rRNA of Escherichia coli

ccessible in situ to single DNA oligonucleotides. Appl Environ Microbiol 72: 733-44. 208SILVERMAN AP, KOOL ET (2007) Oligonucleotide probes for RNA-targeted fluorescence in

situ hybridization. Adv Clin Chem 43: 79-115.


209 LIMA C (2013) Application of response surface methodology (RSM) to optimize the hybridization efficiency of a PNA probe targeting Sacchoromyces cerevisiae. Biological Engineering, Porto, Faculty of Engineering

210 SANTOS R et al. (2014) Optimization of a peptide nucleic acid fluorescence in situ hybridization (PNA-FISH) method for the detection of bacteria and disclosure of a

formamide effect. J Biotechnol pii: S0168-1656(14)00314-9

211 OKTEN HE, YILMAZ LS, NOGUERA DR (2011) Exploring the in situ accessibility of small subunit ribosomal RNA of members of the domains Bacteria and Eukarya to oligonucleotide

probes. System Appl Microbiol 35: 485-95

212 GRAHAM AR (1983) Fungal autofluorescence with ultravioleta illumination. Am J Clin Pathol 79: 231-4

213 GRAF B, GOBEL UB, ADAM T (1998) Qualitative and quantitative analysis of autofluorescence in fungi. Mycoses 41: 39-46

214 SORENSEN AH et al. (1997) Whole-cell hybridization of Methanosarcina cells with two new

oligonucleotide probes. Appl Environ Microbiol 63: 3043-50

215 MIOZZARI GF, NIEDERBERGER P, HÜTTER R (1978) Permeabilization of microorganisms by

Triton X-100. Anal Biochem 90:220-33

216 KING AT et al. (1991) Surfactant effects on yeast cells. Enzyme and Microbial Technology 13: 148-53

217 KOLEY D, BARD AJ (2010) Triton X-100 effects on membrane permeability of a single HeLa cell by scanning electrochemical microscopy (SECM). Proc Natl Acad Sci USA 107: 16783-7

218 VASILEVA-TONKOVA E et. al. (2001) Biosurfactant-rhamnolipid effects on yeast cells. Lett Appl Microbiol 33: 280-4

219 ALMEIDA C et al. (2013) Detection of Salmonella enterica serovar Enteritidis using real time PCR, immunocapture assay, PNA FISH and standard culture methods in different types of food samples. Int J Food Microbiol 161: 16-22

220 MALIC S (2009) Detection and identification of specific bacteria in wound biofilm using

peptide nucleic acid fluorescent in situ hybridization (PNA). Microbiology 155:2603-11

221 THOMPSON JD, HIGGINS DG, GIBSON TJ (1994) CLUSTAL W: improving the sensitivity of progressive multiple sequence alignment through sequence weighting, position-specific gap

penalties and weight matrix choice. Nucleic Acids Res 22:4673-80

222 YUEN PK et al (2003) Microfluidic devices for fluidic circulation and mixing improve hybridization signal intensity on DNA arrays. Lab Chip 3: 46-50

223 BRIGATI DJ, inventor; Iniziative Marittime 1991, SRL, assignee (1992) And an unsulfated

monosaccharide, reduced viscosity and surface tension. United States patent US 5116727 A

224 ZAYTSEVA NV et al (2005) Microfluidic biossensor for the serotype-specific detection of

dengue vírus RNA. Anal Chem 77:7520-7

225KU WC et al. (2004) Dextran sulfate provides a quantitative and quick microarry hybridization

reaction. Biochem Biophys Res Commun 27: 30-7

226 SUNG, J; SHULER M (2009) Prevention o fair bubble formation in a microfluidic perfusion cell culture system using microscale bubble trap. Biomedical Microdevices 11:731-8.


227 SIEBEN V et al. (2007) FISH and chips: chromossomal analysis on microfluidic platforms.

Nanobiotechnology 1: 27-35

228 SIEBEN VJ et al (2008) Na integrated microfluidic chip for chromosome

enumeration using fluorescence in situ hybridization. Lab Chip 8: 2151-6

229 VEDARETHINAM I et al. (2010) FISH on a Chip: Miniaturized microfluidic device for

fluorescence in situ hybridization. Sensors 10: 9831-46

230 LI PCH et al (2004) Transport retention and fluorescence measurement of single biological

cells studied i microfluidic chips. Lab Chip 4:174-80

231 ZHANG Y et al (2012) Single Cell analysis of yeast replicative aging using a new generation

of microfluidic device. PLoS ONE 7:e48275

232 ALVANKARIAN J et al (2013) A pillar-based microfilter for isolation of white blood cells on elastomeric substrate. Biomicrofluidics 7: 014102

233 FORREST GN et al (2006) Peptide nucleic acid fluorescence in situ hybridization-based identification of Candida albicans and its impact on mortality and antifungal therapy costs J Clin Microbiol 44: 3381- 3383

234FERNANDEZ J; et al (2009) Time to positive culture and identification for Candida bloodstream

infections. Diagn Microbiol Infect Dis. 64:402-407 235 ZHENG W et al (2010) A simple PDMS-based microfluidic channel design that removes

bubbles for long-term on-chip culture of mammalian cells. Lab Chip 10:2906-10


Appendix

I – YEPD-agar and YEPD medium recipe

a) YEPD-agar medium recipe

Reagent Concentration Manufacturer

Yeast Extract 1% (w/v) Merck, Germany Peptone 2% (w/v) Liofilchem, Italy Dextrose 2% (w/v) Merck, Germany Agar 2% (w/v) Merck, Germany Water - - Sterilize by autoclaving.

b) YEPD medium recipe


Yeast Extract 1% (w/v) Merck, Germany Peptone 2% (w/v) Liofilchem, Italy Dextrose 2% (w/v) Merck, Germany Water - - Sterilize by autoclaving.

II – Composition of FISH Solutions

a) Standard hybridization solution (pH 7.5)


Sodium Chloride 10 mM Sigma-Aldrich, USA Dextran Sulfate 10% (w/v) Fisher Scientific, UK Formamide 30% (v/v) Acros Organics, UK Sodium pyrophosphate 0.1% (w/v) Acros Organics, UK Polyvinylpyrrolidone 0.2% (w/v) Sigma-Aldrich, USA Ficol 0.2% (w/v) Fisher Bioreagents, UK Disodium EDTA 5 mM Panreac Quimica, Spain Triton X-100 0.1% (v/v) Panreac Quimica, Spain Tris-HCL 50mM Fisher Scientific, UK


b) Simplified hybridization solution (30% formamide) (pH 7.5)


Tris-HCL 50mM Fisher Scientific, UK Dextran Sulfate 10% (w/v) Fisher Scientific, UK Formamide 30% (v/v) Acros Organics, UK Triton X-100 0.1% (v/v) Panreac Quimica, Spain

c) Simplified hybridization solution (0% formamide) (pH 7.5)


Tris-HCL 50mM Fisher Scientific, UK Dextran Sulfate 10% (w/v) Fisher Scientific, UK Triton X-100 0.1% (v/v) Panreac Quimica, Spain

II – Composition of washing solutions

a) Standard hybridization solution (pH 10)


Tris Base 5 mM Fisher Scientific, UK Sodium Chloride 15 mM Sigma-Aldrich, USA Triton X-100 1% (v/v) Panreac Quimica, Spain

Documents

INTEGRATION OF MICRO FLUIDICS AND FLUORESCENCE IN SITU