Transformation and Dyeing
GFP and RFP expressing amoebae were obtained by transforming cells with pTX-‐GFP (Levi et al. 2000) or pTX-‐RFP plasmids (constructed in our lab) (Figure 2-‐1). The GFP (or RFP) gene was put under ubiquitously expressing promoter actin15. Standard electroporation procedure was optimized in the lab by my supervisor Clement Nizak and co-‐workers prior to my arrival in the lab. Cells were grown in 75cm2 flasks until dense but not confluent (usually 1 day before confluency). The medium was changed 4-‐ 6h before transformation. For transformation cells were re-‐suspended in 10mL of ice-‐ cold HL5 and kept on ice for 30min. Cells were centrifuged for 5min, 500g at 4°C. Supernatant was re-‐suspended in 800μl of electroporation buffer and transferred into ice cold 4mm electroporation cuvettes containing 30μg of plasmid DNA. Cells were electroporated at 0.85 kV and 25 mF twice, waiting for 5 s between pulses. Cells were transferred from the cuvette to 75cm2 flask with HL5. The next day, transformants were selected with 5μg/ml G418 (SIGMA). The concentration of G418 was gradually increased to 20μg/ml G418 over 1-‐2 weeks. Transformed strains were maintained at this concentration of G418, yielding GFP and RFP-‐expressing cell lines that were analyzed by flow cytometry to confirm their unimodal cellular fluorescence distribution (>99% of fluorescent cells upon analysis of 106 cells).
Transforming D.discoideum natural isolates
No standardized cell transformation protocol for natural isolates of D.discoideum exists. We have tried to develop our own protocol. Here are details of the protocol and obtained results.
Protocol: Natural isolate strain was grown on SM/5 plates with K. aerogenes or in flasks with SorC and heat killed bacteria as food source (10ml of SorC + 300μl of heat killed bacteria). In case of liquid cultures in SorC medium was changed several hours before transformation (10ml SorC + 250μl of heat killed bacteria). In both cases exponentially growing cells were re-‐suspended is ice cold SorC and let sit for 30min on ice. Suspension was centrifuged 3 times on 500g, 5min, 4°C after which standard transformation protocol was performed. At the end cells were transferred to HL5 medium with streptomycin. Next day the medium was changed to a) SorC + 200μl heat killed bacteria + streptomycin + G418 (7.5μg/ml) or b) SorC agar plates with 7.5μg/ml G18 and 200μl heat killed bacteria. SorCan heat killed bacteria were used to prevent bacterial division that causes the consumption of selective G418 antibiotic.
The transformation was performed with self replicating plasmids pTX-‐GFP and pTX-‐RFP and integrating plasmid V18 GFP.
Results: transformation showed no success rate. In some cases isolated colonies that were resistant to G418 were observed. In every case they would not multiply and die within several days.
Cell staining with chemical dyes
Invitrogen CellTracker Probes Red CMTPX and Green CMFDA are commonly used for staining D.discoideum cells. We worked according to previously established protocols (Nizak et al. 2007; Buttery et al. 2009). Working solution was 10mM dye re-‐suspended in DMSO. Cells were washed and re-‐suspended in ice colds phosphate buffer. If natural isolates were used they were re-‐suspended in phosphate buffer and centrifuged 3 times at 300g, 5min, 4°C. Cells were counted and re-‐suspended to 1x104cells/μl. CellTracker dye was thawed quickly and a dye was added to a final concentration of: Red dye-‐ 25μM for 30min, or 10μM for 1h, and Green dye -‐ 50μM for 30min. Cells were incubated in the dark with gentle shaking for time indicated for each concentration. After incubation excess dye was removed by centrifuging the cells 2-‐3 times in ice-‐cold phosphate buffer. Cells were observed under the microscope for their fluorescence.
When only part of the population was dyed, the “non-‐dyed” population was treated the same way as dyeing one. The only exception being that the cells were not incubated with the dye but with the same concentration of DMSO.
Results: The dyeing was successful for both D.discoideum and P. pallidum natural isolates and D. discoideum AX3 strain. In all cases dyeing was not sufficiently strong for tracking single cell fluorescent, but population fluorescence was strong enough. In addition dye would bleach fast making it difficult to track non-‐aggregated cells over time. When natural isolates were used bacteria that were still left in suspension and that were dyed together with cells were a big source of noise. This made it difficult to distinguish the background from single fluorescent cells.
Aggregation D. discoideum starvation experiment
Cells were subjected to two different starvation conditions: sudden and gradual starvation. Sudden starvation is a standard starvation protocol used in all D.discoideum and P. pallidum studies (Fey et al. 2007). We established gradual starvation protocols in liquid and on bacteria during the period of this PhD.
Sudden starvation: If not mentioned otherwise sudden starvation was used as a standard plating protocol: When cell culture was near confluency, medium with antibiotics was replaced with antibiotic free medium. No difference was found when cells from exponential growth phase (before confluence) were used. After 4-‐6h cells were washed out of nutrient medium and centrifuged in KK2 buffer at 500g for 5 min. The pellet was re-‐suspended in KK2 buffer to the concentration of 1x105cells/μL. For density dependent aggregation experiment cells were re-‐suspended to the concentration of 1×103, 1×104, 5×104,1×105 or 5-‐7.5×105 cells/μL. Green and red fluorescent cells were mixed in ratios indicated in Image analysis section. 30μl of suspension was plated on 6cm plates filled with 2mLof 2% Phytagel (SIGMA). In case of pairwise mixtures, strains grown in different medium or genetically different strains, the ratio of two strains was 1:1.
Gradual starvation in liquid: the cells were collected 1-‐2 days after reaching confluency in HL5. Cell washing and plating was done as in sudden starvation experiment described above.
Gradual starvation on bacteria: another way of slowly starving the cells is to plate them with bacteria and to let them deplete the food source as in natural conditions. Two types of plating were done: homogenous and heterogeneous plating. In both cases RFP-‐ expressing AX3 and GFP-‐expressing AX3 cells were grown in HL5 medium with 20μg/mL G418. When confluent cells were re-‐suspended in KK2 buffer and centrifuged at 500g for 5min. The cell pellet was re-‐suspended in KK2 to the concentration of 1x105cells/μL. Green and red fluorescent cells were mixed in ratios indicated in Image analysis section. For heterogeneous plating 200μL of heat-‐killed bacteria was mixed with 100μl of cell suspension. The mixture was spread on a 6cm plate with 2mL of 2% Phytagel. This gave rise to heterogeneous distribution of cells and bacteria (Sup. Figure S2). For homogenous plating 100μl of heat-‐killed bacteria were mixed with 100μl of cell suspension. A 100μl drop was plated on a 6cm plate with 2ml of 2% Phytagel and let to dry under the hood. This gave a very homogeneous cell distribution (Sup. Figure S3-‐2). In both cases, cells fed for ~8h on heat-‐killed bacteria before the beginning of starvation, and thus divided at most twice after plating. The density of cells at the onset of starvation (measured via a similar method as the one for measuring the non-‐aggregating cell fraction, see below) was comparable to that of cells processed according to the sudden starvation protocol.
P. pallidum aggregation
Starvation induced aggregation was observed by: i) plating cells on nutrient free agar or ii) plating cells on bacterial plates and letting them deplete food gradually.
Plating on nutrient free agar: cells were grown on SM/5 plates with K. aerogenes as food source. Cells were washed of in ice cold KK2 and centrifuged 3 times on 500g, 4°C to remove access bacteria. Bacteria free cells were plated on phytagel plates at density around 1×105 cells/μl. Under these conditions cells start aggregating within several hours.
Plating cells on bacterial plates: cells were prepared as above. Cells were plated on 5ml SM/5 plates with 100μl overnight K. aerogenes culture.
Time-‐laps microscopy **
The 6 cm diameter Petri dish was imaged on an automated inverted microscope setup duplicated from (Houchmandzadeh 2008). The set up was constructed by Clement Nizak prior to this PhD. The setup was made of: OlympusIX70 inverted microscope, Photometrics CoolSNAP HQ2 CCD camera, Zeiss NHBO 100 microscope illuminating system, Thorlabs SH05 shutter, Thorlabs TSC001 shutter controller, and 2.5x-‐5x-‐10x-‐ 20x objectives (5x was used for all experiments shown here). Images were acquired in WinView/32 and the whole setup was controlled by custom-‐made visual basic software. The setup allows Petri dish scanning at regular time intervals, with phase contrast and fluorescence image acquisition for each image at all time points. An area of around 1cm x 1cm was scanned every 15-‐30 min for P. pallidum aggregation and 1h-‐2h for non-‐ aggregated cells experiments, with 5x or 2.5x objective. A mosaic image is reconstructed by combining all the images of contiguous areas of the Petri dish at a given time point by a custom-‐made macro using Image J software.
All images were analyzed in Image J. Image J is an open source, public domain, Java-‐ based image processing program. Program and various plugins can be found at official website: http://rsbweb.nih.gov/ij/
Non-‐aggregated cells **
Mixing a small percentage of red fluorescent cells in a population of green fluorescent cells allowed us to get the image single cells as single red fluorescent dots (Fig 2-‐2). This a experimental approach was inspired from studies of cell motion within aggregates (Dormann et al. 1997). The approach was developed by my supervisor Clement Nizak and co-‐workers prior to beginning of my PhD and optimized during the course of my PhD. We optimized the red to green cell ratios depending on plated cell density: for experiments with 1x105and 5-‐7.5×105 cells/μL0.25-‐0.5% of RFP cells was mixed with 99.5-‐99.75% GFP cells, for 1×104 cells/μL 1% of RFP cells were used and for 1×103 cells/μL 2% RFP cells were used. For pairwise mixtures the ratio was made as following: 50% of strain A in GFP was mixed with 49.75% of strain B in GFP and 0.25% of strain B in RFP. Reciprocal mixing was done as well, for example 0.25% of GFP cells were mixed with 99.75% of RFP cells. The choice of fluorescent signal had no effect on cell behavior.
Images were acquired by time-‐laps fluorescence microscopy. All the images were analyzed in Image J by custom-‐made macro made by us. The analysis consisted in counting fluorescent dots before and after aggregation. For each experiment 1000 – 10 000 dots/cells were monitored. Dead cells were excluded from counting by looking at cell displacement as an indicator of cell viability. Two fluorescent images taken 1-‐2h apart were overlapped and cells that showed no displacement were counted and subtracted from the overall non-‐aggregating population.
The density of red dots (RFP-‐expressing cells) was used to estimate cell density at the onset of starvation in all experiments. Cell density was comparable at the onset of starvation for all starvation protocols used.
Figure 2 -‐ 2 Image analysis of D. discoideum non-‐aggregating cells. Fluorescent images (A and C) are treated in Image J to get a binary image with fluorescent cells as black dots (B and D). Number of dots before and after aggregation is counted using Analyze Particles command.
P. pallidum aggregation **
All images were analyzed in Image J. Phase contrast images were converted to a binary image of white background and black aggregates. This was done by manually setting the image threshold. Image was cleared of background noise by using Despeckle and Remove Outliers commands, and manually. The process demanded a great deal of manual work for most mid-‐aggregation steps when a real aggregate boundary was difficult to estimate based on phase contrast alone. Final working image is show in Figure 2-‐3.
Figure 2 -‐ 3 Image analysis of aggregation in P. pallidum. A) phase contrast image and B) binary image ready for analysis.
To estimate spatial structure of aggregates we used two point correlation function. We used an Image J plugin built by Johannes Schindelin, free to download at http://wbgn013.biozentrum.uni-‐wuerzburg.de/ImageJ/two-‐point-‐correlation.html. The plugin is based on the paper: J. G. Berryman and S. C. Blair, Use of digital image analysis to estimate fluid permeability of porous materials I. Application of two-‐point correlation functions (Berryman & Blair 1986). The function calculates the correlation of finding two points (pixels of radius r) at distance d. The program uses FFT transformation to accelerate the calculation. Original images were too big for analysis, therefore before analysis images were scaled 0.1 times its original scale.
Aggregate size was calculated in Image J using Analyze particles command. Aggregate size is represented as aggregate area in pixels.
Strain Specific Life Cycle Properties
Growth rate on bacterial plates **
Standard growth rate measurements for D.discoideum are done in liquid culture. To measure growth rate on bacterial plates we developed our own experimental procedure. Experiment was done with 6 natural isolates; 34.1, 28.1, 105.1, 63.2, 85.2 and 98.1. Spores were plated on SM/5 plates with 200μL of overnight K. aerogenes culture. Spores germinated into cells and cells started dividing. 15-‐20h after plating spores cells were removed from the plates by washing the plates in ice-‐cold KK2 buffer. Suspension was centrifuged 3 times in ice cold KK2 for 5min, 300g to remove bacteria so we could count the cell concentration. 1×105 cells were re-‐suspended in 500μL of overnight K. aerogenes and plated on 15cm petri dish with SM/5 agar. 16-‐20h after plating we started to measure cell growth. Growth was measured for 3 independent plates/time point every 2h during 8h. For each measurement cells were removed by scraping the cells from the plate in ice-‐cold KK2 buffer to prevent cell division. Suspension was centrifuged 3 times for 5min, 300g on 4°C to remove bacteria. Cells were counted using haemocytometer. Growth curve was represented as log2 of the cell number over time. Growth rate was calculated as slope of the linear regression of the growth curve.
Sporulation efficiency tests the “efficiency” of a cell to become a spore. It is calculated as a ratio of Nspores/Nscells. It is a common procedure used in many studies (Buttery et al. 2009; Fortunato, Queller, et al. 2003). We used the same procedure as in their studies with minor adaptations (centrifugation force and timing before cell plating)
Spores were plated on SM/5 pates with K. aerogenes. Spores germinated and a fresh lawn of cells was collected. Cells were scraped from plate in ice-‐cold KK2 buffer. Suspension was centrifuged 3 times on 300g, 4°C to remove bacteria. Cells were re-‐ suspended in KK2 buffer to a final concentration of 1×105 cell/μl. Cells were plated on 2% phytagel plates or filter papers. For plating on phytagel plates 1mL (1×108 cells) was plated and spread on 6 cm plate. The plate was left to dry under the hood. For plating on filters, cells were left for 2-‐3h in phosphate buffer in order to consume any bacteria left and to finish division. Cells we counted and re-‐suspended to final concentration 1×105 cell/μl. 30μl (3×106 cells) was plated as a drop on filter. Plates were kept in the incubator for 2 days for fruiting bodies to form. Plates were washed in SORC with 0.1% TWEEN. Spore concentration was counted with the hemocytometer.
On bacterial plates
Germination efficiency protocol was adopted from previous studies (Castillo et al. 2011; Jack et al. 2008).
Experiment was done with 6 natural isolates; 34.1, 28.1, 105.1, 63.2, 85.2 and 98.1. Spores were plated on 9cm SM/5 plates with 200μl of overnight K. aerogenes culture. Spores germinated, cells divided, consumed bacterial food and when starved formed new spores. Fresh spores were re-‐suspended in ice-‐cold HL5 with 0.1% TWEEN and vortexed. HL5 was used instead of phosphate buffer cause spores would attach to surface of tube when in suspension with phosphate buffer. 100 spores were plated with 500μl of overnight K. aerogenes culture and plated on 14cm petri dish with SM/5 agar. After 3 days we counted number of formed plaques. Germination efficiency was counted as Nspores/Nplaques. For each strain experiment was repeated 7-‐9 times with 3 replicas per measurement.
Table of contents :
CHAPTER 1 GENERAL INTRODUCTION
POPULATION-‐LEVEL BHAVIOURS AND EFFECTS
Population level perspective – making sense of individual behaviors
Individual and population level interactions
Population level perspective in our work
Life cycles -‐ Survival strategies
Social life cycle -‐ Morphogenesis and differentiation
Evolution and phylogeny
CHAPTER 2 MATERIALS AND METHODS
Strains and Culturing
Transformation and Dyeing
Strain Specific Life Cycle Properties
MEDIA AND BUFFERS
Non-‐aggregating cells (Chapter 3)
Strain competition (Chapter 4)
CHAPTER 3 POPULATION PARTITIONING BETWEEN UNICELLULAR AND MULTICELLULAR STRATEGIES IN SOCIAL AMOEBAE D. DISCOIDEUM
Not all cells aggregate
New microscopy technique for quantifying non-‐aggregating cells
Phenotypic plasticity affects population partitioning
Genetics of population partitioning
Individual-‐level costs and benefits of the non-‐aggregating cell fraction
Cell history and cell fate
Model: evolutionary framework
Population partitioning into aggregating and non-‐aggregating cells
Consequences of population partitioning on cooperation
SUPPLEMENTARY FIGURES AND MOVIES
APPENDIX: FOR ECONOMISTS: ADAPTATION TO UNCERTAINTY IN BIOLOGY AND ECONOMY
CHAPTER 4 HOW LIFE CYCLE COMPLEXITY AFFECTS GENETIC DIVERSITY AND COOPERATION IN SOCIAL AMOEBAE D. DISCOIDEUM
Genetic diversity in cooperative systems
Social amoebae: competition and cooperation
Environmental sources of phenotypic variability
Life cycle complexity, competition and cooperation
Germination efficiency –questions and speculations
APPENDIX: DIVERSITY AND INDIVIDUALITY
Costs and benefits of being diverse
Individuality and levels of selection
CHAPTER 5 DYNAMICS OF AGGREGATION IN P. PALLIDUM
Aggregation in social amoebae
Aggregation in D. discoideum – the best studied example
Aggregation in P. pallidum
Our focus: dynamical aggregation in P. pallidum
CONCLUSIONS AND PERSPECTIVES