Advanced Detection Methods of Genomic Barcodes for Genotyping Escherichia coli Libraries
Nicole Eger
Degree project in biology, Master of science (2 years), 2021 Examensarbete I biologi 30 hp till masterexamen, 2021
Biology Education Center and Institure of Cell and Molecular Biology, Uppsala University Supervisors: Daniel Camsund, Jimmy Larsson and Johan Elf
I hereby declare that this submission is my own work and that, to the best of my knowledge and belief, it contains no material previously published or written by another person, except where due acknowledgment has been made in the text.
Nicole Eger
Uppsala, 8th January 2020
ii
Abstract
Pooled cell strain libraries are a powerful tool allowing to investigate the influence of genetic modifications on phenotypes in high throughput single-cell assays. To link the genotype to phe- notype in each cell of the library, unique 20 base pairs (bp) long barcodes are used to allow in situ genotyping after phenotyping via fluorescence microscopy. In previous studies, these barcode sequences were expressed from high copy number plasmids resulting in a high number of targets for detection via fluorescence in situ hybridisation (FISH) and thus, a strong readout signal. How- ever, constant selection pressure must be applied on the cells to maintain the foreign plasmid DNA which may influence the phenotype. Inserting unique barcodes on the chromosome ensures sta- bility of the construct which is required for some genomic library applications. However, the low copy number of the barcode sequence often requires an additional step of DNA amplification for efficient detection. In this study, two methods for barcode amplification were investigated. First, amplification from the double stranded DNA upon binding of peptide nucleic acids and subsequent amplification via rolling circle amplification (AmPPR). Second, amplification from genomic DNA or cDNA via loop-mediated isothermal amplification (LAMP). Whereas the AmPPR approach re- mained unsuccessful, chromosomal barcode sequences were successfully amplified in situ via LAMP and subsequently detected using FISH. I show that LAMP can potentially be a quick, specific, and elegant amplification technique for in situ genotyping in microfluidic devices. However, nonspecific amplification and partly nonspecific readout signals when using LAMP remain a problem and need to be further investigated before implementing this method on pooled libraries.
iii
Declaration ii
Abstract iii
List of abbreviations vi
1 Introduction 1
1.1 In Situ Genotyping on the Basis of Barcode Sequences . . . . 1
1.2 DNA Amplification via PNA Binding and thus Opening of the dsDNA Helix, Pad- lock Probe Hybridisation, and Rolling Circle Amplification . . . . 3
1.3 DNA Amplification via Loop-Mediated Isothermal Amplification (LAMP) . . . . 6
1.4 Project Aim . . . . 9
2 Materials and Methods 10 2.1 Cell Culture . . . . 10
2.2 Cloning Methods . . . . 10
2.3 Microfluidics . . . . 11
2.4 Genomic Barcode Amplification via AmPPR . . . . 12
2.4.1 Design . . . . 12
2.4.2 Testing AmPPR In Situ . . . . 13
2.4.3 Testing AmPPR In Vitro . . . . 14
2.5 Genomic Barcode Amplification via LAMP . . . . 15
2.5.1 Design . . . . 15
2.5.2 Experimental Setup and Sample Preparation . . . . 17
2.5.3 Testing the LAMP Construct in Bulk . . . . 18
2.5.4 Testing the LAMP Construct in a Closed Microfluidic System . . . . 18
2.5.5 Testing the LAMP construct in an Open Microfluidic System . . . . 19
2.5.6 Testing DNA Amplification Arising from Primer Dimers . . . . 20
2.6 Microscopy settings . . . . 20
3 Results 21 3.1 Genomic Barcode Amplification via AmPPR . . . . 21
3.1.1 DNA Amplification via AmPPR in situ remains nonsuccessful . . . . 21
3.1.2 PNAs bind Specifically and Unspecifically to Plasmids in Vitro . . . . 23
3.1.3 Presence of PNAs Inhibits Padlock Probe Binding and Ligation in Vitro . . 24
iv
Master Thesis Uppsala University
3.2 Genomic Barcode Amplification via LAMP . . . . 28 3.2.1 LAMP Combined with FISH in Bulk Experiments Results in a Fluorescent
Signal in Situ in High Producer Cells . . . . 28 3.2.2 LAMP Amplifies Barcode Sequences in a Closed Microfluidic System . . . . 29 3.2.3 LAMP Amplifies DNA Nonspecifically Still Leading to a Specific Output Signal 30 3.2.4 LAMP Amplifies DNA from Primer Dimers . . . . 32 3.2.5 Variations in Primer Sequences lead to Nonspecific Fluorescent Signals . . . 33
4 Discussion 36
4.1 Genomic Barcode Amplification via AmPPR . . . . 36 4.2 Genomic Barcode Amplification via LAMP . . . . 37
5 Conclusion 39
Acknowledgements 40
Bibliography 41
Nicole Eger v
Ap ampicillin
BIP backward inner primer bp base pairs
CRISPR clustered regularly interspaced short palindromic repeats csp cold shock protein
DNA deoxyribonucleic acid
DAPI 4,6-diamidino-2-phenylindole dsDNA double stranded DNA E. coli Escherichia coli
EtOH ethanol
FIP forward inner primer
FISH fluorescent in situ hybridisation GOI gene of interest
Km kanamycin LB luria broth
LAMP loop-mediated isothermal amplification NEB New England Biolabs
PBS phosphate buffered saline
PBS-T phosphate buffered saline 0.05 % tween
vi
Master Thesis Uppsala University
PCR polymerase chain reaction PNA peptide nucleic acids PDMS polydimethylsiloxane RCA rolling circle amplification RFU relative fluorescent units RNA ribonucleic acid
ssDNA single stranded DNA SPB sodium phosphate buffer SSC saline sodium citrate WT wild type
Nicole Eger vii
1.1 In Situ Genotyping on the Basis of Barcode Sequences
Natural expression patterns of living cells can be altered by inserting specific genetic modifica- tions. Foreign deoxyribonucleic acid (DNA) fragments can be inserted into a genome or genes can be mutated or knocked out which can lead to variations in protein synthesis and thus, in a cells phenotype (Song et al., 2015). To study the influence of different modifications on the phenotype of a cell, large cell strain libraries can be created (A. Garst et al., 2017). Genotyping of individual cells within a pooled cell strain library is essential to identify their specific modification after phe- notyping. To enable in situ genotyping, each specific modification can be linked to a unique 20 base pairs (bp) long barcode sequence which is either inserted into the chromosome of the cell or located on plasmid DNA (Camsund et al., 2020; Chen et al., 2018; Lawson et al., 2017). The detection of the barcodes in situ can be performed either on the DNA sequence directly or on their transcribed ribonucleic acid (RNA) molecules (Chen et al., 2018). The number of barcodes affects the strength of the readout signal. Using high copy number plasmids provides a high number of target DNA and thus can lead to high expression levels of target RNA. However, constant selection pressure in form of e.g. antibiotics must be applied in order to keep the foreign plasmid DNA. This is not desirable as certain antibiotics can affect phenotypes or provoke mutations (Athamneh et al., 2014;
Cira et al., 2018). Inserting the barcode sequence into the chromosome results in a more stable construct that does not normally need constant selection (Blundell & Levy, 2014). However, this reduces the barcode copy number to one single copy per chromosome compared to over hundreds when using high copy number plasmids. Thus, when using genome integrated barcodes, amplifying them on either the DNA duplex directly or from transcribed RNA molecules often is required to obtain a detectable output signal.
Different methods for in situ DNA or RNA target amplification were developed using for example padlock probe approaches (Ke et al., 2013; Lee et al., 2014, Smolina et al., 2007). A popular method for detecting amplified barcode sequences for genotyping is fluorescent in situ hybridisation (FISH) (Camsund et al., 2020; Lawson et al., 2017). In FISH, single stranded oligonucleotide sequences are labeled with a fluorophore. Specific binding to their target combined with fluorescence microscopy, allows for detecting the presence and localisation of molecules in situ. FISH is an important re- search tool in cell biology and genomics, and furthermore applied in clinical diagnostics (Huber et al., 2018). FISH applications in personalized medicine is a growing field. In molecular cancer diagnostics for example, FISH can be used to detect genetic aberrations for individual patients (Hu et al., 2014). The method is sensitive enough to detect single nucleotide polymorphism on the
1
Master Thesis Uppsala University
genome (Beliveau et al., 2015). The specificity and sensitivity of FISH is determined by the accu- rate design of the fluorescent probes whereby sequence length is a key player (Huber et al., 2018).
With increasing probe length, the penetration speed into a cell decreases as probe uptake is based on diffusion (Cassidy & Jones, 2014). For low copy number targets, long incubation times of up to 16 h are typical to obtain a sufficiently strong FISH signal (Huber et al., 2018). This incubation time can be reduced, when applying a constant flow rate of FISH probe solution over the cells. This can be achieved using microfluidic devices (Huber et al., 2018), also referred to as a Lab-on-a-Chip.
Microfluidics is an attractive approach for genotyping because it enables quick switching of media and reagents, precise control of media flow, real-time monitoring with a microscope and allows for small reaction volumes. Furthermore, microfluidics provides the possibility for high-throughput genotyping of cells in combination with detection of individual phenotypes such as dynamics and distribution of proteins inside of an a single cell. A recent study investigated genotypes on the basis of barcodes in high-thoughout genetic libraries in microfluidic devices via FISH (Camsund et al.
2020). However, in situ genotyping was done with RNA barcodes expressed from high copy number plasmids. The next step is to incorporate these barcode sequences into the chromosome and to be able to detect them despite the anticipated lower expression levels expected from the chromosome.
Previously, the signal given by fluorescent probes which bind RNA molecules expressed from the chromosome had been too weak for detection within a practical timeframe (unpublished results, Elf, Uppsala University). Therefore, to enable genotyping using these single barcode copies either on the chromosome directly or on transcribed RNA molecules, their sequence must be amplified.
Nicole Eger 2
1.2 DNA Amplification via PNA Binding and thus Opening of the dsDNA Helix, Padlock Probe Hybridisation, and Rolling Circle Amplification
In situ genotyping on the basis of chromosomal barcodes using FISH and analysis by microscopy requires amplification of the readout signal. A study by Smolina et al. (2007) presents a three step DNA amplification method from double stranded DNA (dsDNA) targets. The three steps are (i) the opening of dsDNA via the binding of peptide nucleic acids (PNA) (Figure 1.1 A), (ii) the invasion and ligation of a padlock probe (Figure 1.1 B), and (iii) the amplification of the padlock sequence via rolling circle amplification (RCA) (Figure 1.1 C). In this previous study, a 21 nt long sequence within the major cold shock protein gene cold shock protein (csp) in the Escherichia coli (E. coli) chromosome was targeted. To use this method for genotyping, a genomic barocde sequence serves as the dsDNA target. In my thesis, I refer to this DNA amplification method generally as AmPPR (Amplification via PNA, Padlock probe, and RCA).
In the first key step, PNAs hybridize to the target sequence opening the DNA duplex. PNAs are single stranded DNA mimics which can form duplex or triplex structures with RNA or DNA.
A PNA molecule consist of a pseudo peptide backbone of N-(2-aminoethyl)-glycine units to which pyrimidine and purine bases are attached via methyl carbonyl linkers. The hybridisation of PNAs to DNA or RNA molecules is of high stability as the backbone of a PNA molecule is neutral in charge and consequently, there is no electrostatic repulsion (Jensen et al., 1997). Hence, a short sequence of only seven bases is enough for a highly specific and stable hybridisation to the tar- get. Due to their unnatural backbone, PNAs are resistant to nucleases and proteases (Demidov et al., 1994). These features make PNAs an attractive tool in molecular medicine. Due to their high stability of binding, PNAs can efficiently inhibit the synthesis of proteins on a transcriptional or translational level (Nielsen et al., 1999). However, the uptake of PNA molecules into living cells remains a challenge. To improve cellular uptake, PNAs can be attached to cell-penetrating molecules or be conjugated with cationic lipids (Shiraishi et al., 2020).
In the present work, two PNA oligos were used to open up the dsDNA target to allow the hy- bridisation of a padlock probe on the complementary strand (Smolina et al., 2007). A padlock probe is a single stranded oligonucleotide which hybridizes to the target sequence with its 3’ end as well as with its 5’ end but not its center, thereby forming a circle (Larsson et al., 2004; Figure 1.1).
A DNA ligase such as T4 Ligase closes the circle and adds a further specificity level as non-perfect base-pairing of the padlock’s DNA ends to the target inhibits or even prevents its ligation. After the ligation step, the circularized padlock probe sequence can be amplified via RCA by the DNA polymerase Phi29. The RCA product is a micron-sized ball of DNA repeats with multiple copies
Nicole Eger 3
Master Thesis Uppsala University
of the padlock probe sequence. For the detection of the RCA product, one or several FISH probe target sequences can be integrated into the padlock probe sequence. Multiple fluorescent probes can thus bind to one RCA product making it visible as a fluorescent dot under a fluorescence mi- croscope. The combination of padlock probe hybridisation to a target sequence and amplification via RCA has been used for genotyping in several studies (Ke et al., 2013; Larsson et al., 2004;
Larsson et al., 2010).
Two main approaches in using padlock probes are distinguishable; the no-gap padlock approach, and the gap-filling padlock approach. The no-gap approach has a higher sensitivity in RNA detec- tion whereas the gap-filling approach detects the actual sequence of the target (Chen et al., 2017;
Ke et al., 2013; ). In the no-gap approach, the padlock probe binds to the target sequence (the barcode sequence), is ligated and only the padlock probe itself is amplified by RCA. However, for each barcode, a new padlock probe must be designed (Larsson et al., 2010) making this approach challenging to use for genotyping of large barcoded libraries. In the gap-filling approach, the pad- lock probe binds to its target sequence with a gap of several nucleotides between its 3’ and 5’ ends.
This gap is filled by a polymerase, synthesizing the complementary strand of the target sequence.
After ligation of the padlock probe to the newly synthesized DNA, the filled in gap, which is the actual barcode sequence, is amplified via RCA (Ke et al., 2013). As the barcode sequence of the gap is copied, the same padlock probe can be used for all targets as far as they are all flanked by the same padlock probe binding sequences. After RCA reaction, the amplified barcode can be detected via FISH and fluorescence microscopy.
Nicole Eger 4
Figure 1.1: Illustration of barcode amplification and detection from dsDNA via AmPPR. (A) PNA molecules bind to their target sequence which flank the genomic barcode, thereby opening up the dsDNA helix. (B) A circularizable padlock probe invades into the double strand and hybridizes to the target sequence with its 3’ and 5’ ends. A DNA ligase seals the nick between the two ends. (C) Annealing of RCA primers initiate the amplification of the padlock probe sequence and thereby the bar- code by Phi29 polymerase. Hybridisation of fluorescent probes to the RCA product enables visualization via microscopy.
Nicole Eger 5
Master Thesis Uppsala University
1.3 DNA Amplification via Loop-Mediated Isothermal Amplification (LAMP)
Another DNA amplification method is loop-mediated isothermal amplification (LAMP) which was developed by Natomi at al. (2000). Its simplicity, high specificity, and rapidness makes LAMP to a popular detection method. It is widely used in diagnostics and its reliability has been proven as results agree with the standard detection method polymerase chain reaction (PCR). In pre- vious studies, pathogenic microorganisms including bacteria, fungi, viruses, and parasites were detected via LAMP (Fu et al., 2011). LAMP provides a cost-effective point-of-care diagnosis and is hence suitable for the high demand of detection of the virus COVID-19 in the ongoing pandemic.
COVID-19 is an RNA virus responsible for the current Corona pandemic which started in 2019, in Wuhan China. R. Soares et al. (2020) developed a LAMP protocol for the the detection of the viral RNA directly from nasopharyngal swab samples using an integrated smartphone-based cen- trifugal microfluidic platform. Besides disease diagnostics, LAMP is applicable in various fields for example for identification of genetically modified organisms (Guan et al., 2010; D. Lee et al., 2009), for embryonic sex determination (Hirayama et al., 2013; Zoheir et al., 2011), or for genotyping of single nucleotide polymorphisms and mutations in nucleic acid sequences (Gill et al., 2020).
Compared to DNA amplification by PCR, LAMP has several advantages, being (i) LAMP has a simpler experimental setup as it amplifies DNA in an isothermal reaction and does not need a thermocycler instrument like PCR does, (ii) LAMP has a shorter reaction time of less than one hour, and (iii) LAMP is more suitable to be used for direct detection from biological samples as it is more robust to the presence of inhibitors than PCR enzymes are (Rekha, 2014). These advantages make LAMP an attractive method for field diagnosis. Furthermore, a positive reaction of LAMP is detectable by eye by its turbidity. This is because the amplification can result in up to 10 µg of DNA leading to a high amount of pyrophosphate ions which together with magnesium form a white precipitate (Mori et al.; 2001, Parida et al.; 2008). Other detection methods of LAMP products are gel electrophoresis, the use of fluorescent probes, or SYBR green dye (Gill et al., 2020).
LAMP amplifies its target DNA by an autocyclic strand displacement reaction. The isother- mal reaction is carried out by a DNA polymerase with a high strand displacement activity.
A DNA polymerase optimized for LAMP is Bst 2.0 which originates from Bacillus stearother- mophilus. It amplifies DNA at temperatures between 60-65°C (New England Biolabs (NEB), https://international.neb.com/products/m0538-bst-20-warmstart-dna-polymeraseProduct Informa- tion). Besides a DNA polymerase, four to six primers, dNTPs, magnesium sulphate, betaine, en- zyme buffer, and template DNA are needed for DNA amplification. The target DNA sequence contains six distinct regions: F1c, F2c, F3c, and B1, B2, B3 (Figure 1.2). Two outer primers (F3
Nicole Eger 6
and B3) and two inner primers (forward inner primer (FIP) and a backward inner primer (BIP)) bind to this sequence. FIP contains the F1c sequence at its 5’ end and initiates the amplification reaction by binding to F2c. The newly synthesized strand is displaced by the binding and elonga- tion of the F3 primer. The removed strand forms a loop structure as F1 and F1c hybridize. This one-loop structured DNA strand is amplified by the binding, elongation and strand displacement of the primers BIP and B3. The resulting product is a dumbbell-like structure which is further amplified by the binding and elongation of FIP and BIP (Notomi et al., 2000). The addition of two loop primers to the reaction which bind to the loop structures were shown to increase the specificity of the reaction and reduce the amplification time (Nagamine et al., 2002). As newly synthesized strands can serve as templates for the next amplification round, LAMP products are different long concatamers that resemble cauliflower-like structures (Notomi et al., 2000).
One challenge of LAMP lies in the design of the primers and the selection of the target. The target sequence can not be longer than 300 bp and the primers should not form hairpin structures or primer dimers. As LAMP is a very sensitive amplification method, primer dimers might trigger amplification and lead to nonspecific LAMP products. Despite the non-specific amplification from primer dimers, LAMP can lead to unspecific amplification of DNA in negative controls. In a former study, unspecific LAMP products were identified by adding fluorescent FIP and BIP primers in the ongoing amplification reaction (Hardinge & Murray, 2019). After amplification, samples with the LAMP target showed no fluorescence signal whereas negative controls did. It was speculated that the high amount of DNA amplification and the dense cauliflower-like structure of LAMP products quench the fluorescent signal. In another study it was shown that the addition of the polysaccharide pullulan to a LAMP reaction inhibited non-specific DNA amplification (Gao et al., 2019). Besides unspecific amplification, further drawbacks of LAMP are that LAMP products are not suitable for further downstream DNA analysis, and that LAMP reactions are not ideal for multiplexing in diagnostics (Rekha, 2014). However, its several advantages made LAMP to a popular DNA amplification method in diagnostics.
Nicole Eger 7
Master Thesis Uppsala University
Figure 1.2: Illustration of the LAMP mechanism and barcode amplification.
(A) The target sequence contains the segments F1c, F2c, F3c, B1, B2, and B3. Barcode sequences (BC) can be integrated between the segments F1c and F2c, and B1 and B2;
the later stem-loop structures. The amplification reaction is initiated by the annealing and elongation of the forward inner primer (FIP) to F2c. The primer includes the complementary sequences to F2c to which it hybridizes, and to F1c on its 5’ end.
Hybridisation and elongation of the F3 primer to the target sequence leads to strand displacement of the before newly synthesized strand. (B) Then, the backward inner primer (BIP), which includes the segments B2 and B2c, hybridizes to sequence B2c, and is elongated. This strand is then replaced by a new strand elongated from primer B3.
(C) The now displaced strand includes the segments F1c and F1, and B1 and B1c. The complementary segments hybridize thereby forming two loop structures, which result in a dumbbell-like structure of the strand. (D) This construct is amplified in an autocyclic strand displacement reaction leading to the production of multiple concatemers. The addition of two loop primers LB and LF that hybridize to the loop structures promote further amplification. Barcode sequences which are now located in loop structures can be detected using fluorescent probes.
Nicole Eger 8
1.4 Project Aim
The overall goal of this thesis project was to develop and optimize a method that can be used to identify different genotypes in pooled E. coli strain libraries in situ using microscopy. The method aims to detect chromosomal barcode sequences within a practical timeframe and to be performed in microfluidic devices. However, as only one barcode per chromosome is present in a cell, their sequence must be amplified before detection via FISH is possible.
In this thesis, two different DNA amplification methods were investigated in microfluidic devices with the aim to amplify genomic chromosomal barcodes and to detect them with FISH and mi- croscopy.
Nicole Eger 9
2 Materials and Methods
2.1 Cell Culture
All E. coli cell strains used in this study were grown in luria broth (LB) media at 37°C unless otherwise stated. For selection of plasmid carrying strains in liquid media, the respective antibiotic was added; 25-50µg/ml kanamycin (Km) for the AmPPR construct, and 100 µg/ml ampicillin (Ap) for the LAMP construct. LB agar plates contained Ap 100 or 50 µg/ml, or Km 50 µg/ml. For experiments requiring cells in the exponential growth state, cells were cultured overnight, diluted 1:1000 next morning, and grown for approximately 3 hours until an OD600 of 0.3. For microfluidic experiments, cells were diluted in filtered LB media containing 0.2 % pluronics.
2.2 Cloning Methods
For cloning the desired gene of interest (GOI) into a target plasmid backbone, the Golden Gate method was used. Golden Gate was developed by Engler et al. in 2013. One advantage of this cloning method is its use of class II restriction enzymes. These enzymes cut a definite number of bp away from their specific recognition sequences. After enzymatic digestion, the recognition sequences are removed leaving sticky or blunt ends. The advantage of removing the recognition sequences is that restriction enzymes can not repeatedly cut as is the case for standard cloning methods with class I restriction enzymes. For the ligation of two overhangs, these must contain a complementary four nucleotide long sequence. Usually four nucleotides of either site of the con- struct are chosen to enable scar-less cloning.
To conduct the Golden Gate method, first type II recognition sites were added to the 3’ and 5’ ends of the GOI and the plasmid via PCR (Table 2.1). For my thesis study, the enzyme BpiI recognizing the sequence 5’- GAAGAC -3’ was used. To ensure a high fidelity amplification during PCR, Q5 High-Fidelity DNA Polymerase (NEB) was used. After the PCR reactions of both GOI and plasmid, both constructs were digested with DpnI (ThermoFisher) for 1 h. DpnI is a restriction enzyme cutting methylated DNA and thus all DNA templates which do not contain the recognition sites for the Golden Gate cloning. After PCR purifications (ThermoFisher Qiagen PCR Purification Kit), the construct consisting of GOI and plasmid was assembled using 5 - 10 U of the restriction type II enzyme BpiI, and 3 U of T4 ligase in corresponding ligase buffer which was incubated at 37°C for 1 h. The Golden Gate products were transformed via electroporation into E. coli Top10 cells for plasmid enrichment. After plasmid isolation (Thermo Fisher Qiagen Quick Plasmid Isolation Kit), and GOI being sequence confirmed (Eurofins Genomics Europe),
10
about 50 ng of plasmid DNA was transformed into 50 µl electrocompetent E. coli wild type (WT) cells for AmPPR constructs or MG1655 araB::T7 rnap-tetA DELaraB for LAMP constructs. The last-mentioned strain expresses T7 polymerase from its chromosome in the presence of arabinose as the gene is under the control of a ParaBAD promotor. Additionally, this strain is tetracycline resistant.
After recovery, cells were plated on respective antibiotic LB agar plates. For long time storage, 20 % glycerol cryo stocks were made from liquid selective overnight cell cultures and kept at -80°C.
Table 2.1: Primer sequences for Golden Gate cloning Primer Name Primer (5’- 3’)
AmPPR target forward
CTAGAAGACTAAGACTCAAAAGAAGGTCACCCTATAGT- GAGTCGTATTAATCTCC
AmPPR target reverse
ATTGAAGACTGGTCTCTCTCCGCTGGGGGAGGACTCC- CACAGTC
AmPPR sequencing CCTTCCTTTCTCGCCACGTT
2.3 Microfluidics
The design for the microfluidic devices used in this study is based on the work of Baltekin et al.
(2017). The devices had 1500 nm wide cell traps to line up bacteria. The constrictions in the end of each trap keep aligned bacteria in their position when flowing liquids through the device and over the bacteria in the traps. The devices were manufactured from liquid polydimethylsiloxane (PDMS) (Silicon Elastomer Kit, Sylgard 184) which consists of different long monomeric siloxane chains. To initiate solidification of the PDMS, it was mixed with the kit’s cross-linking agent in a 10:1 ratio. After mixing for 30 min, it was poured onto a microfluidic wafer. Air bubbles were removed by applying vacuum and the PDMS was hardened at 80°C for 2 h. The PDMS was then carefully removed from the mold, leaving behind an imprinted replica of the micro-channels on the PDMS surface. Holes were punched into the device to later allow for the connection of tubings to the micro-channels. These tubings were used to inject fluids and control flow rates. Next, the chips were bound onto glass cover slips after first treating both with plasma (Henniker plasma).
Plasma transforms CH3-groups on the PDMS surface to OH-groups which enables binding to the glass surface. As this reaction creates water molecules, the chips were heated again to 80°C for 1 h to promote evaporation. Microfluidic devices were kept in room temperature in a dust free environment and were utilized within two weeks.
Nicole Eger 11
Master Thesis Uppsala University
2.4 Genomic Barcode Amplification via AmPPR
2.4.1 Design
Using AmPPR for genotyping of cell strains on the basis of barcodes requires the design of bar- code sequences which are specific for every strain, PNA molecules, padlock probe sequences, and FISH probes. The strategy was that two different 7 nt long PNA molecules bind to the sequences that flank the antisense barcode sequence, thereby opening the dsDNA helix. Once the DNA is accessible, a specific padlock probe can bind to the sense barcode sequence. After padlock probe ligation, a Phi29 polymerase amplifies the entire padlock probe sequence in multiple copies. One padlock probe encodes several barcode sequences to which specific FISH probes can bind.
Unfortunately, in this thesis project, this design was not tested as it was not possible to reproduce the previous results (Smolina et al., 2007), which formed the basis of the new method. Hence, all experiments in this section are reproduction attempts of the Smolina et al. study. Here, the target sequence was a 21 bp long region within the major csp gene of WT E. coli. Sequences for PNAs, primers, and padlock probes were taken from the paper and are listed in Table 2.2. However, the AmPPR target sequence was elongated by flanking nucleotides of the 21 nt long csp target sequence to ensure padlock probe binding.
Table 2.2: Sequences for the signature site, the PNAs, the padlock probe, the fluores- cent probe, and the RCA primer used in this work. Sequences were taken from Smolina et al., 2007. The AmPPR target sequence was elongated by 4 nt on the 3’ and by 5 nt on the 5’ end.
Molecule Sequence (5’-3’)
AmPPR target CCAGCGGAGAGAGACTCAAAAGAAGGTCAC
PNA 1 H-Lys2-CTCTCTCC-(eg1)3-JJTJTJTJLys-NH2
PNA 2 H-Lys2-TTTJTTJJ-(eg1)3-CCTTCTTT-Lys-NH2
Padlock Probe p-TCAAAAGAAGGTCACGGAATGGTTACTTGCCAGCCAG
CAGCCTCACGGAATGGTACTTGCCAGCGGAGAGAGAC
RCA Primer GTGAGGCTGCTGGCTG
Fluorescent Probe Cy3-TCACGGAATGGTTACTTGCACAGC-biotin-3
As stated before, the in situ method of Smolina et al. (2007) was not reproducible. Hence, in addition to the in situ experiments, in vitro experiments were performed mainly to control sequence compatibility of PNAs, target, padlock probe and to test reagents and reaction conditions. For the in vitro experiments, the AmPPR target sequence from the in situ experiments was cloned into the 2100 bp long pGuide7 plasmid backbone (Lawson et al., 2017, ”pGuide”), termed AmPPR target plasmid. pGuide7 is a high copy number plasmid with the marker gene neo, which confers resistance to Km. To confirm correct sense and antisense sequences and a working protocol, a single stranded DNA (ssDNA) oligonucleotide encoding the padlock probe target sequence (equal
Nicole Eger 12
to the AmPPR target sequence, compare Table 2.2) named ’ssDNA padlock target’ was used as a positive control target. For a negative control, an empty pGuide7 plasmid was used.
2.4.2 Testing AmPPR In Situ
The experimental key steps of in situ AmPPR include (i) cell fixation on glass slides, (ii) PNA invasion and dsDNA opening, (iii) padlock probe binding to target DNA and padlock probe lig- ation, (iv) RCA reaction, and (v) analysis via microscopy. Before the experiment, the padlock probes were phosphorylated on their 5’ ends in order to enable circularization. Thus, 2 µM pad- lock probe was incubated with 1 mM ATP, 0.1 U/µl T4 Kinase, and 1x PNK buffer A (NEB) at 37°C for 30 min. Subsequently, the enzyme was inactivated by heating the reaction mix up to 60°C for 20 min. Until further usage, phosphorylated padlock probes were stored at -20°C. In the in situ experiments, E. coli WT cells and E. coli WT cells termed ’high producers’ which were transformed with pGuide7-AmPPR and therefore have higher numbers of the target DNA, were tested.
Cells were grown until the culture reached the exponential phase and were subsequently placed on a glass slide previously coated with polylysine (Sigma) solution. Then, glass slides were put in freshly prepared ice-cold fixative (3:1 methanol-glacial acid) solution for fixation for 15 min at room temperature. After one washing step with phosphate buffered saline (PBS) (pH 7.4), cells were permeabilized through an ethanol (EtOH) dilution series (70%, 90%, and 96%). After gluing 50 µl hybridisation and incubation chambers onto the glass slides with fixed cells, slides were kept at -20°C until continued use.
For PNA invasion, fixed cells were washed twice with 10 mM sodium phosphate buffer (SPB) (pH 7.0), air dried, and incubated with 40 nM PNAs in 10 mM SPB overnight at 37°C. To prevent evaporation, reaction chambers were sealed.
Next, the samples were washed twice with SPB, air dried, and then 50 µl of the padlock probe reaction mix including 0.04µM padlock probe, 5 U T4 DNA ligase (ThermoFisher), 1x ligase buffer (ThermoFisher), was applied. As a negative control, ligase was not added to the mix preventing padlock probe circularization. The reaction was incubated for 2 h at room temperature after chambers were sealed.
After washing the cells twice with 2x saline sodium citrate (SSC) buffer and once with buffer A (100 mM TrisHCl, 150 mM NaCl and 0.05 % Tween20), RCA was performed. The RCA mixture containing 10 U Phi29 DNA polymerase and 1x Phi29 buffer (NEB), 200 µM dNTPs, 2 µM RCA primer, and 0.2 µM fluorescent detection probe was added onto the sample. The reaction chamber was sealed and the sample was incubated for 4 h at room temperature and subsequently at 37°C overnight.
Next, the samples were washed twice in buffer A, once in buffer 4XT (4x SSC, 0.05 % Tween20),
Nicole Eger 13
Master Thesis Uppsala University
and incubated with the DNA dye 4,6-diamidino-2-phenylindole (DAPI) (ThermoFisher) for 10 min.
Before imaging, the samples were washed in 2XT (2x SSC, 0.05 % Tween20), the buffer reaction chambers were removed, mounting media (ThermoFisher) was applied on the sample, and a cover slip was placed on top. Then, cells were imaged under a fluorescence microscope.
2.4.3 Testing AmPPR In Vitro
In the in vitro experiments, briefly, PNA were added to bind to the AmPPR target sequence on a plasmid, then, the PNA-DNA construct was transferred into a reaction mix for padlock probe binding and ligation, next, a portion of this mix was transferred into the RCA reaction mix for DNA amplification.
To test specific PNA binding to the target sequence, AmPPR target plasmid, the ssDNA padlock target, as well as the negative control (empty pGuide7) were separately incubated with both PNA molecules. For each of these three samples, different concentrations of DNA and PNA molecules were tested; 5 nM DNA in combination with 10 µM of each PNA, and 20 nM in combination with 27.5 µM of each PNA. As a negative control for each sample, no PNAs were added. The samples were incubated in 8 mM tris (pH 7.9) and 0.8 mM EDTA in a total reaction volume of 20 µl at 37°C for 2.5 h. Then, the samples were analyzed using gel electrophoresis and the DNA stain SYBR Safe (ThermoFisher) as a DNA binding dye, and were subsequently imaged in a UV transilluminator.
Next, target DNA amplification via RCA subsequent to PNA and padlock probe binding was investigated. AmPPR target plasmid sample and ssDNA padolck target sample were tested in duplicates. For DNA-PNA binding, 20 nM DNA was incubated with 27.5 µM of each PNA. For padlock probe binding, 10 nM DNA of each duplicate were transferred to a 20 µl reaction mix containing 0.02µM padlock probe, 0.02 U/µl T4 ligase, 0.2 µg/µl BSA, and 1x T4 ligase buffer. In addition to these samples, 20 nM of ssDNA padlock target oligo DNA was directly incubated with padlock probe and ligase in the same conditions as stated before. Furthermore, padlock probe binding and ligase reaction mixes were prepared without ligases for each sample and served as negative controls. After incubating all samples for 20 min at 37°C, 5 µl of each ligation mix was used in a RCA reaction. RCA mixes contained 0.1 U/µl Phi29 DNA polymerase, 1x Phi29 buffer, 0.25 mM dNTPs, and 0.2µg/µl BSA in a total volume of 50 µl. The reaction mix was incubated at 37°C for 30-60 min. Next, the DNA content was verified via gel electrophoresis using 0.8 % agarose gels and SYBR Safe. For gel electrophoresis, pure sample DNA, DNA-PNA binding mix, ligation mix, and RCA reaction mix were loaded onto the gel. Images were taken in a UV transilluminator.
For the specific detection of RCA products, fluorescent probes were added to the different RCA samples in a 2x SSC and 15 % formamide hybridisation buffer. After 5 h of incubation and DNA
Nicole Eger 14
fragment separation via gel electrophoresis, the samples were analyzed with a ChemiDoc (BioRad).
2.5 Genomic Barcode Amplification via LAMP
2.5.1 Design
For testing in situ genotyping on the basis of barcodes with LAMP, a 232 nt long DNA target sequence was designed by my supervisor Daniel Camsund specifically for my project (Figure 2.1).
The construct had four barcode sequences for later detection via FISH, and included the regions B1-B3 and F1-F3 for primer binding and amplification. This LAMP construct was designed as a proof of concept construct, to test LAMP in the microfluidic device used in this thesis.
Figure 2.1: Illustration of the LAMP target construct with four integrated barcode sequences and annealing primers.
The genomic barcodes are 20 bp long sequences that were taken from a previous study (Camsund et al., 2020). In this study, 40 barcode sequences were selected after being randomly generated, and screened for different properties like sequence heterogeneity, the presence of certain restriction enzyme recognition sequences or the likelihood of forming hairpin structures. For the detection of these barcodes, fluorescent probes with varying excitation and emission wavelengths were designed.
Each FISH probe is marked with one of four different fluorophores, hence the fluorescent signal can be detected in four different channels. The reason is that different combinations of barcodes, binding probes of different colors, can encode a larger number of unique genotypes than the dif- ferent barcodes by themselves. In LAMP, the barcodes are amplified in both sense and antisense directions, hence the FISH probes are either sense or antisense sequences of the barcode. All bar- code sequences and their corresponding FISH probe sequences used in the present study are listed in table 2.3. The barcode sequence 4 was part of the LAMP construct, however if was never part of the analysis.
The LAMP target sequence was either expressed from the chromosome leading to ’low producer’
cells, or from high copy number plasmids leading to ’high producer’ cells as here more LAMP tar- gets were present in one cell. To create the low producers, the target sequence was cloned into the chromosome of E. coli Eco MG1655 araB::T7 rnap-tetA DELaraB SS9::PT7 by Daniel Camsund
Nicole Eger 15
Master Thesis Uppsala University
Table 2.3: Barcode sequences within the LAMP construct and the corresponding FISH probes. Sequences were taken from Camsund et al., 2020
BC Sequence (5’- 3’) Fluorescent Probe (5’- 3’) Fluorescent Channel 1 CACGACGGATTCTATGTTTC CACGACGGATTCTATGTTTC Cy3
2 TTGTGACAGGTTATCGCAGT TTGTGACAGGTTATCGCAGT Cy5 3 ATCGCTCCTTATTGTCGCCG CGGCGACAATAAGGAGCGAT Alexa 488 4 ATAAGGTTGATCGTGCCGTT AACGGCACGATCAACCTTAT Texas Red
via clustered regularly interspaced short palindromic repeats (CRISPR) facilitated recombineering.
For the high producers, the LAMP target was cloned into the high copy number pGuide8 plasmid backbone (Lawson et al., 2017) termed LAMP target plasmid and transformed into E. coli Eco MG1655 araB::T7 rnap-tetA DELaraB SS9::PT7.
The expression of the LAMP target sequence in vivo was controlled by T7 polymerase which was expressed from the chromosome of the used E. coli strain Eco MG1655 araB::T7 rnap-tetA DE- LaraB SS9::PT7. In the absence of T7 polymerase, the LAMP target sequence was not transcribed.
The expression of T7 polymerase was under the control of the inducible ParaBAD promoter on the chromosome. The ParaBAD promoter is turned on by the presence of arabinose and silenced by glucose.
The LAMP primers were designed by using the online software NEB LAMP Primer Design Tool (NEB, https://lamp.neb.com/#!/), and LAMP primer designing software Primer Explorer V5 (Primer Explorer V5, (http://primerexplorer.jp/lampv5e/index.html). Primers for two sets were designed (Table 2.4); a third set consisted of primers B3, BIP, and LB from set 1 and F3, FIP, and LF, from primer set 2. For preventing the formation of primer dimers, the deltaG limit for free energy of dimer interaction was set to -2.1 kcal/mol in both software. DeltaG values of -9 kcal/mol and more negative have a higher likelihood for the formation of hetero- or homo-dimers (Inte- grated DNA Technologies, IDT, https://eu.idtdna.com/pages/support/faqs/how-do-i-use- the-oligoanalyzer-tool-to-analyze-possible-hairpins-and-dimers-formed-by-my-oligo).
Nicole Eger 16
Table 2.4: LAMP primer sequences.
Primer
Name Primer Set 1 (5’- 3’) Primer Set 2 (5’- 3’)
F3 ACACTCGTAGTTAACAACGG CCCGGCCCTATCAGAACA
FIP GCGCTCCCTGGTAATGGACT-
GAACACGAGTGCCTTGAATG
GCGCTCCCTGGTAATGGAC- TACAGTGCCTTGAATGCACGAC
B3 GGGATTACGACTGACCGT CGTAAGACCGCTATGTACCG
BIP
CTCTCGCCAGCACTG-
TAATAGGGACCGCTATGTACC- GACG
TTCATTCTCTCGCCAGCACTGT- GAACGGCACGATCAACCTTA
LF GCGATAACCTGTCACAAAA-
GAAACA TGCGATAACCTGTCACAAAAG
LB CGTTATAAGGTTGATCGTGCCG GGATCGCTCCTTATTGTCG
During amplification, the regions between B2 and B1, and between F2 and F1 form the two loops within the LAMP-specific dumbbell structures. Different barcode sequences were incorpo- rated into these loop regions for posterior detection via FISH probes. The inspiration of this design originates in a previous study which demonstrated that fluorescent probes can bind to the loops of LAMP products (Mori et al., 2006). The loop primers bind to these regions thereby promoting further DNA amplification. The amplification of the target sequence in LAMP is an isothermal process often driven by Bst 2.0 DNA polymerase. This polymerase requires a DNA template for amplification, hence a reverse transcriptase is contained in the NEB 2x Warm- Start LAMP mastermix (https://international.neb.com/products/ e1700-warmstart-lamp-kit-dna- rnaProduct Information) transcribing RNA into cDNA to enable LAMP assays also on RNA tar- gets.
The LAMP reaction was performed on fixed cells at 65°C. This high temperature leads to desta- bilization of the hydrogen bonds in dsDNA allowing primers to bind to dsDNA targets whereby Bst 2.0 polymerase can amplify the LAMP target sequence directly from the plasmid as well as from cDNA molecules.
2.5.2 Experimental Setup and Sample Preparation
The target DNA amplification via LAMP was tested in three different systems: in cell bulk ex- periments, in microfluidic experiments within a closed system controlled by an air pressure pump (Elveflow), and in microfluidic experiments within an open system controlled by a syringe pump
Nicole Eger 17
Master Thesis Uppsala University
(New Era Pump Systems Inc.). All experiments share the same LAMP reaction mix with 1x LAMP WarmStart mastermix (NEB), and 1 µM primer mix. For the expression of the LAMP target se- quence in bulk and in open microfluidic system experiments, cells were cultured overnight, diluted 1:1000 in the morning, and after 1 h growth the cells were induced for 1 hour at 37°C with ei- ther 0.2 % arabinose, or 0.5 % arabinose, when inducing the cells in the microfluidic device using the closed system. After induction, cells were collected by centrifugation at 4500 g for 4 min at 4°C. For the in bulk experiments and the open system microfluidic experiments, cells were fixed in 4 % formaldehyde in PBS for 10 min followed by two PBS washing steps. Then, cells were permeabilized in 70 % EtOH and kept at -20°C until further use.
2.5.3 Testing the LAMP Construct in Bulk
For the bulk experiments, E. coli cells expressing the LAMP target sequence from a plasmid ’high producer’ cells and cells expressing the LAMP target from the chromosome ’low producer’ cells were prepared as mentioned above. For the LAMP reaction, cells were incubated with the LAMP mix containing primer set 1 at 65°C for 1 h in a thermocycler (BioRad). To stabilize the LAMP products inside the cells, cells were postfixed with 4 % formaldehyde in PBS for 10 min. Then, after two washing steps with PBS, FISH probes 2 and 3 were added to the samples in a 30 % formamide, and 0.4 % SSC hybridisation buffer. To allow high binding levels, samples were labelled overnight.
For microscopy imaging, 0.5 µl of cell sample was applied onto 2.5 % agarose pads made with PBS.
2.5.4 Testing the LAMP Construct in a Closed Microfluidic System
In the closed microfluidic system, all outlets of the PDMS device were directly connected to an air pressure pump via tubings. With this system, the media flow was precisely controlled.
For the experiments in this system, E. coli high producer cells, and cells carrying an empty pGuide plasmid ’negative control’ which were grown to the exponential growth phase, were loaded into the same PDMS device, grown for 1 h at 37°C, and induced with 0.5 % arabinose for 1 h. Sub- sequently, cells were fixed (4 % formaldehyde in PBS), washed with PBS, and both dehydrated and permeabilized with 70 % EtOH for 30 min. After rehydration of the cells by a 10 min PBS washing step, the microchannels of the PDMS device were coated with 1 % casein. The phosphoprotein casein, a component of milk, is a hydrophobic protein which interacts with the hydrophobic surface of the PDMS chip. Thereby, the PDMS surface gets coated preventing other proteins relevant for the experiment to be adsorbed by the PDMS surface. Next, the LAMP reaction mix containing set 1 primers was applied for 1 h at 65°C with a high flow rate to prevent evaporation through the PDMS at this high temperature. To further keep the humidity inside of the device during the reaction, it was placed into a 65°C water bath. After the LAMP reaction, cells were labelled with
Nicole Eger 18
fluorescent probes 2, and 3 in hybridisation buffer consisting of 30 % formamide and 2x SSC. Then, cells were washed with PBS to remove excess probes, and then imaged via fluorescence microscopy.
2.5.5 Testing the LAMP construct in an Open Microfluidic System
For the open microfluidic system, liquid injection and pressure control was supplied by a syringe pump which was connected by only one tube to the inlet of the PDMS device. The outlets re- mained unplugged. Using a syringe pump enables a simpler experimental setup compared to the closed system. Hence, this system was used for optimization of basic LAMP parameters inside of a PDMS device. Different conditions such as reaction temperature, amplification time, and the use of different primer sets were tested. The flow rate of the fluids was controlled in µl/min.
In these experiments, LAMP reactions on (i) E. coli cells expressing the LAMP target from a plasmid ’high producers’, (ii) cells expressing the LAMP target from the chromosome ’low produc- ers’, (iii) cells carrying an empty pGuide plasmid ’negative control’, and LAMP reactions on (iv) latex beads (ThermoFisher) were tested.
Cells were cultured, induced, and fixed as described in the Experimental Setup and Sample Prepa- ration section and loaded into the PDMS device in phosphate buffered saline 0.05 % tween (PBS-T).
For each cell strain a different device was used. After cell loading, 1 % casein (ThermoFisher) was perfused for 10 min at 1 µl/min. Next, the LAMP reaction was applied at 3 µl/min for 30-60 min in a 65-69°C water bath. First experiments were performed at 65°C, later the temperature was increased to 68°C. However, the temperature of the water bath had a variation of about 1°C. The high flow rate as well as the water bath prevented high evaporation rates of the reaction mix through the PDMS. Water from the water bath did not enter the PDMS device despite its open ports as the pressure inside the device was much higher than in the surroundings.
Following the LAMP reaction, fluorescent probes were perfused in a high stringency buffer con- taining 30 % formamide and 0.4 % SSC. The device was kept at 65°C for the first 5 min of the hybridisation step. This might enable easier accessibility for hybridisation of the FISH probes to the LAMP products as these products are large tightly packed bundles of single and double stranded DNA and high temperature destabilizes dsDNA (Mori et al., 2006). Then, the hybridi- sation continued for 10 min at room temperature. Thereafter, cells were stained with DAPI in a high stringency buffer containing 30 % formamide and 0.4 % SSC and imaged.
The open system experiments were performed in cooperation with Ruben Soares, Stockholm University, Mats Nilsson Laboratory.
Nicole Eger 19
Master Thesis Uppsala University
2.5.6 Testing DNA Amplification Arising from Primer Dimers
For testing the formation of primer dimers, in vitro LAMP reactions were prepared in duplicates and run in a real-time PCR machine (BioRad). Samples included (i) LAMP reaction mix without DNA ’no template’, (ii) LAMP reaction mix with the LAMP target plasmids, and (iii) negative control plasmids not encoding the LAMP target sequence. In a total reaction mix of 10 µl, DNA was added to 1x LAMP WarmStart MasterMix (NEB), 1 µM primer mix, and 1x of the DNA dye SYBR Green (ThermoFisher). The reactions were incubated for 1 h at 65 or 68°C in a real-time PCR machine.
LAMP reactions were tested with different plasmid numbers, 102 and 109. To reduce amplifi- cation rates, the use of only four primers (excluding loop primers) instead of all six primers was tested. All samples were run in duplicates. The fluorescence signal given by the intercalation of SYBR Green into double stranded DNA represents the DNA amount in relative fluorescent units (RFU).
2.6 Microscopy settings
Images were taken with a Ti2-E inverted microscope (Nikon) with a 20x or 100x Plan Apo Lambda objective (Nikon). All LAMP open microfluidic system experiments cells were imaged with a Zeiss Axio Imager Z2 microscope. For fluorescence images DAPI, Cy3, Cy5, and Alexa 488 filter cubes were used. For image analysis, the ImageJ software (NIH, USA) was used.
Nicole Eger 20
3.1 Genomic Barcode Amplification via AmPPR
Low copy number sequences can by amplified from double stranded DNA directly by using specific PNAs to open dsDNA, padlock probes which then hybridize and circularize on the target sequence, and DNA polymerases that can thereafter perform RCA of the padlock probe sequence (Smolina et al., 2007). In this master thesis, this three-step method is referred to as AmPPR. Before testing the amplification of chromosomal barcodes with AmPPR, an in situ experiment from a previous study was reproduced (Smolina et al., 2007). In that study, a 21 nt long region of the gene csp, located on the chromosome of E. coli was amplified. This sequence is referred to as ’AmPPR target sequence’ in this thesis. Additionally to in situ experiments, the method was performed in vitro on plasmid DNA encoding the AmPPR target sequence. Unfortunately, both attempts to amplify the target sequence on dsDNA remained nonsuccessful.
3.1.1 DNA Amplification via AmPPR in situ remains nonsuccessful
In the attempt to reproduce the results of Smolina et al. (2007) in situ, the AmPPR target se- quence located on the chromosome was targeted in WT E. coli cells, and in E. coli cells referred to as ’high producers’, where the AmPPR target sequence was additionally present on high copy number plasmids. To evaluate the output signal, negative controls were added to each experiment for which the AmPPR target sequence amplification was prevented. For these controls, no ligase was added to the samples in the second reaction step of AmPPR which prevented the closing of the hybridized padlock probe and thereby its amplification.
For the in situ experiments, first, cells were fixed, incubated with PNAs, padlock probes, and Phi29 DNA polymerase for the RCA reaction. Thereafter, Cy3 labeled fluorescent probes hybridiz- ing to the amplified target sequence were added. For analysis, cells were imaged on an agarose pad under a fluorescence microscope in brightfield and in the fluorescent channel Cy3 (Figure 3.1).
Fluorescent signals overlap with cell structures in all samples. Further, the intensity of the fluorescent signal varies within one sample. In the high producer sample, strong fluorescent signals as well as weaker signals which are similar in strength to the high producer ’no ligase’ control and both WT samples, are seen. In the WT ’no ligase’ control sample accumulations of fluorescent probes were observed around the cell clusters or were not detectable at all. In general, the negative controls have similar strong fluorescent signals as the E. coli WT and high producer samples.
This implies that no RCA products are present in either E. coli WT nor high producer samples.
21
Master Thesis Uppsala University
The experiment had been repeated several times with varying incubation times in all experimental steps. As these experiments all lead to the same result, only one representative example is shown (Figure 3.1).
Figure 3.1: In situ DNA amplification via AmPPR. E. coli cells which produce high amounts of AmPPR target ’high producers’ and WT E. coli cells which have one copy of the AmPPR target sequence on each chromosome, were incubated with specific PNAs, and padlock probes to enable amplification of the target DNA in a RCA reaction. For negative controls of each sample, no ligase was added to the padlock probe reaction.
After adding FISH probes, high producer cells (A, B, C), high producers ’no ligase’
(D), WT E. coli cells (E, F), and WT ’no ligase’ cells (G, H) were imaged in phase contrast (grey shades) and fluorescence Cy3 channel (green).
Nicole Eger 22
3.1.2 PNAs bind Specifically and Unspecifically to Plasmids in Vitro
As the target amplification in situ remained non successful, first PNA binding to the dsDNA tar- get sequence was tested in vitro to verify the sequence comparability. The PNA hybridization to DNA was tested on (i) plasmids with inserted AmPPR target sequence ’AmPPR target plasmid’, (ii) plasmids without target sequence ’negative control plasmid’, and ssDNA encoding the padlock probe target sequence which is complementary to the PNA target sequence ’ssDNA padlock target’.
No PNA-DNA interaction was expected for the negative control plasmid nor for the ssDNA padlock target as for the latter, both, the PNA molecules and ssDNA padlock target oligonucleotide have the sense target sequence and therefore are not expected to hybridize.
Samples were incubated with PNAs and additionally, a duplicate of each DNA sample was incubated without PNAs, serving as negative controls for each of the samples. With these negative controls a direct comparison of samples with PNAs added and without was possible to test the influence of PNAs on the reaction and the samples. After 4 h of incubation, the samples were analyzed by agarose gel electrophoresis and DNA staining (Figure 3.2).
The DNA bands close to the well of the gel that are visible for all PNA incubated samples, indicate that PNAs bind to plasmid molecules, leading to DNA-PNA complexes that travel slower through the gel. The slowed movement can be explained by positively charged lysine residues included in the PNA sequence. Originally, these lysine residues were added to facilitate PNA-DNA interaction (Ishizuka et al., 2008). Plasmid samples where no PNAs were added, show three bands of plasmid DNA representing the three topological forms: supercoiled, linear, and nicked-circular.
These bands appear stronger in the AmPPR target plasmid samples than in the negative control plasmid samples. For all plasmid samples incubated with PNAs, DNA bands can be seen close to the gel wells. In AmPPR target plasmid samples incubated with PNAs, only one DNA band is visible located close to the gel well, whereas negative control plasmid samples show additionally DNA bands at a length of 2000 bp indicating unbound plasmids. For the ssDNA padlock target samples no bands can be observed. This was expected as PNAs can not bind to the the padlock probe target oligo as they share same sequences.
Nicole Eger 23
Master Thesis Uppsala University
Figure 3.2: Agarose gel electrophoresis image of plasmid DNA incubated with PNA molecules. Different concentrations of AmPPR target plasmid, plasmids without target sequence ’negative control plasmid’, and ssDNA padlock probe target oligo were incubated with PNA molecules. As a negative control for PNA-DNA incubation for each of these samples, the DNA was not incubated with PNAs. Components of reactions are specified on top of the gel image.
3.1.3 Presence of PNAs Inhibits Padlock Probe Binding and Ligation in Vitro
The previous results showed that PNAs interact with plasmid DNA. To investigate if upon PNA binding padlock probes can invade the double strand and be amplified by RCA, AmPPR target plasmids PNA complexes were incubated with padlock probes and Phi29 polymerase. Addition- ally, ssDNA encoding the padlock probe target sequence which served as a positive control for padlock probe binding and amplification was tested. For negative controls, no ligase was added to the samples during the padlock probe incubation step to prevent complete circularization of the padlock probes and thereby the formation of RCA products.
Samples were taken after each of the three AmPPR reaction steps, (i) PNA binding, (ii) padlock probe hybridisation and ligation, and (iii) padlock probe amplification via RCA, and were analysed by gel electrophoresis and DNA staining with SYBR Safe (Figure 3.3).
Nicole Eger 24
