The GASP phenotype in baylyi

A Senior Thesis

submitted to the

Department of Biology,

The Colorado College

by

Leland Krych

Date ______

Approved by:

______

Primary Thesis Advisor

______

Secondary Thesis Advisor

Introduction

Microbiologist Steven Finkel once referenced Thomas Hobbes’s Leviathan to describe the realities of a bacterium’s life; it is: “solitary, poor, nasty, brutish and short (Finkel 2006).” All theatrics aside, Dr. Finkel makes the point that have it rough, and that this fact is often overlooked in modern bacterial experimentation. Therefore, common laboratory models for bacterial growth might not reveal how bacteria survive in nature, where resources are scarce and competition is fierce. Dr. Finkel goes on to suggest that a more accurate way of studying bacteria in the lab would be to focus on an overlooked aspect of the bacterial life cycle: long-term stationary phase. In order to address why this concept is significant we need to review the five stages in the typical life cycle of a bacterial population raised in the laboratory.

These phases are known as lag phase, exponential phase, stationary phase, death phase, and long- term stationary phase. After inoculation of bacteria into liquid medium there is an initial phase in population growth in the media known as lag phase which is then followed by a phase of exponential growth of the culture. This burgeoning population growth is known as the exponential growth phase. Eventually, the population growth levels off into stationary phase as the amount of space and nutrients start to dwindle. After this period approximately 99% of the cells perish. Survivors from death phase undergo high rates of mutation and begin great rates of cell division in the next stage of the bacterial life cycle: long term stationary phase (LTSP). Long term stationary phase is a highly dynamic phase in which the birth rates and death rates of the bacteria are approximately equal (figure 1). This small population of cells can survive long periods without addition of nutrients. A batch of E. coli in LTSP survived for five years with only the addition of sterile distilled water to maintain the volume and osmolarity (Finkel et al., 2000).

Finkel (2006) suggests that the study of LTSP is important because in a natural environment most bacteria probably live under threat of starvation where the activation of stress-response genes and the use of alternative metabolic pathways is essential for life. One phenotype that is associated with cells found in LTSP is the Growth Advantage in Stationary Phase (GASP) phenotype (Zambrano 1996, Finkel 2001, Farrell 2003, Finkel 2006,). The GASP phenotype was first described in 1993 by Zambrano who noticed that 10 day old E coli could out-compete 1 day old E coli when co-incubated under starvation conditions. To test for the GASP phenotype, cells are grown until they reach immediate stationary phase, and then these cells are competed with cells that have already reached LTSP. The older cells are introduced into the competition experiment in a ratio of 1:1000 old:young. Eventually the aged cells increase in population and drive the younger cells to extinction. This phenomenon is known as a class 1 GASP phenotype. The GASP phenotype is determined by observing the population densities of the two separately labeled strains. Three other classes of GASP phenotypes have been reviewed by Dr. Finkel (figure 2). There is the class 2 GASP phenotype whereby the aged cell population increases but this increase does not lead to the demise of the younger cells. The class 3 phenotype where aged cells initially exhibit the GASP phenotype but then die out; leaving the younger cells at stable population levels. And finally, there is the class 4 phenotype in which aged cells die without any serious population growth. The competitive advantage of the older cells is a result of mutations formed in LTSP (Zambrano 1996). Currently, all known gene mutations that cause GASP phenotypes increase the ability of the bacteria to catabolize various substrates as a source of energy (Finkel 2006). An example is DNA, which E. coli cells use as a source of food during LTSP (Finkel 2001). Drake (1998) found that in E. coli, the rate of mutation in LTSP was one point mutation for every 300-400 cells. In contrast, cells in exponential growth phase show a mutation frequency of one in every 10,000 cells.

While high mutation frequency is typically not a favorable trait for a cell to possess, it is thought that when environmental pressures are extreme enough, a high mutation frequency might create new alleles that provide a competitive advantage. A possible mechanism for the formation of mutations in the GASP phenotype in E. coli is mutations in the SOS-induced DNA polymerases. SOS DNA polymerases are enzymes that repair DNA lesions that DNA polymerase III cannot copy through. DNA polymerase III is the polymerase that synthesizes the leading and lagging strand at the replication fork of DNA. When DNA polymerase III encounter abasic sites, photodimers, or other damaged bases the SOS DNA polymerases are called in to repair the damage (Finkel 2006). Yeiser et al. (2002) demonstrated that in competition with wild type E. coli, mutants lacking one or more SOS-induced DNA polymerases show a dramatic reduction in fitness. Additionally, Yeiser et al (2002) performed GASP tests on wild type cells and the polymerase deficient mutants and found that 10 day old wild type cells experienced a class 1 GASP phenotype when competed against 1 day old wild type cells. In contrast, when 10 day old mutants were competed with 1 day old mutants, the mutants displayed variable GASP phenotypes (Yeiser 2002) indicating that DNA polymerases play some role in generating a class 1 GASP phenotype. It is also prescient to note that the mutants displayed type 1 GASP phenotypes when in competition with other polymerase deficient mutants. It is important to note that A. baylyi does not possess the lexA gene, a regulator of the SOS DNA response (Hare et al. 2006). Recent findings suggest, however, that this process might be mediated by the UmuDAb gene that possesses similar characteristics (Hare et al. 2012).

The model in which we chose to study the GASP phenotype is the gram-negative soil bacterium Acinetobacter baylyi. While Finkel has speculated that bacteria in addition to E. coli should exhibit a GASP response, GASP has been demonstrated in only a few species. A. baylyi and E. coli are distant relatives, found in the same Class, , but different Orders ( and Enterobacteriales, respectively). One striking physiological trait that is very different between the two bacteria is that A. baylyi is well known for high natural transformation efficiency, while E. coli have never been shown to be naturally competent. Competence for natural transformation is a physiological state that permits the uptake of exogenous DNA from the environment. It requires synthesis of a complex competence machine, encoded by proteins in the Type IV Pilus family. E. coli does have genes homologous to this system, but appears to use them for catabolism of DNA under starvation conditions (Finkel 2001).

The evolutionary advantage caused by natural competence is highly debated and as of now there are many competing hypotheses as to why it evolved. Here are two examples:

1) Bacteria exploit natural transformation for nutrient acquisition. Palmen et al. (1994) found that the addition of a five-fold saturation amount of DNA in comparison to normal amounts of DNA present in solution did not result in growth of Acinetobacter calcoaceticus cells under carbon and nitrogen limitation. However, under conditions of limited phosphate, the addition of DNA facilitated further growth. Under these conditions, none of the cells were transformed (as evidenced by lack of radio-labeled DNA in the bacteria). Finkel et al. (2001) found that wild-type strains of E. coli showed 50 times greater growth yield in minimal media with added DNA than mutants lacking specific homologs of competence genes.

2) Bacteria use exogenous DNA as fodder for DNA repair.

Redfield et al (1988) found using a computer simulated model that natural transformation could reduce mutational load but that this benefit hinged upon the source of DNA. The reduction of mutation load was reduced and or negligible when the DNA came from cells that were killed by selection against deleterious mutations; which is what one would expect from unfit cells under high stress conditions. Redfield (1997) also found that according to her numerical model that the benefit of mutational reduction deriving from natural transformation is only very slight. Redfield’s models do not consider the possibility that healthy cells might release DNA in response to environmental changes.

Discussion of natural transformation would not be complete without an understanding of the competence machinery. The mechanism for DNA uptake through natural transformation is mediated by the competence genes; however the overall structure of the multi-protein competence machine has never been visualized. Efforts are being taken by Kristine Lang and Phoebe Lostroh to realize this goal.

Sequence analysis has revealed that competence machinery is related to the type-IV secretion system (Chen 2004). Type IV secretion systems have been adapted to several distinctive physiological purposes. The best understood is use for building type IV pili, which are surface-exposed, retractable fibers that mediate the formation of biofilms, fruiting bodies and motility (Averhoff 2003). Other type IV secretion systems, however, have been adapted for secretion of other protein substrates or for uptake of DNA during natural competence. In all cases, the machines are comprised of homologous protein subunits. Averhoff et al. (2003) proposed a model for DNA translocation through the Acinetobacter baylyi competence system based on homologues to Type IV secretion systems (figure 3). This system is composed of three subunits: a portal in the outer membrane composed of the secretin ComQ, a pilus- like filament that protrudes through the portal and binds DNA, and finally a DNA transport machine that provides passage for the bound DNA into the periplasm (Averhoff 2008).

Secretins form a portal through the outer membrane; in type IV pili, the pilus passes through this pore. This multimeric ring structure is likely composed of multiple copies of the secretin ComQ in A. baylyi. This structure is homologous to other secretins observed in Type IV secretion systems and serves as a portal for pilus assembly and penetration through the portal (Averhoff 2003). The current model suggests that the pilin acts like a piston that is thrust out into space to grasp objects and is then retracted back through the portal. Actual entry into the periplasm is mediated by pilin-like proteins. ComP, ComB, ComE, and ComF are similar to prepilins, the structural subunits of type IV pilins (Averhoff 2008). Prepilins are proteolytically digested to form the type IV pilins. DNA binding and uptake assays have shown that ComP is essential for these functions (Porstendorfer 1997). The DNA that makes it into the periplasm is ferried into the cytoplasm through the highly conserved ComEC (ComA in figure 3) transmembrane protein.

It is already known that E. coli homologs of competence genes are needed for them to survive in LTSP and to exhibit a normal GASP response. But the evolutionary benefits possibly conferred upon naturally competent cells are not clear. Here, we set out to determine whether A. baylyi cells, which unlike E.coli are naturally competent, also exhibit a GASP response. After determining that they do exhibit a GASP response, we next tested whether natural competence is required for them to exhibit a normal GASP response, using non-competent mutants lacking either the ComP or ComQ proteins. We determined that A. baylyi do require competence for development of a normal GASP response when competed against young wild type cells, which has evolutionary implications.

Methods

Bacterial strains used. All experiments were performed with Acinetobacter baylyi strain ADP1. The wild-type strains used were ACN1202 (gentamycin resistant at 5 micrograms/mL) and ACN1203 (Chloramphenicol resistant at 5 micrograms/mL). Antibiotic resistance was used as a marker and did not have an effect on growth advantage in flask competitions. An unpaired t-test between the ratio of old to young in normal GASP controls and switched GASP controls revealed that there was not a significant difference between these two strains (p=0.562) Mutant strains 3335::kan-tdk and 3338:kan-tdk were used. For 3335::kan-tdk, the competence gene comQ had been knocked out and replaced with kanamycin resistance (10 micrograms/mL). Mutant strain 3338::kan-tdk had competence gene comP knocked out and replaced with kanamycin resistance (10 micrograms/mL) (Berardinis 2008).

Creation of mutants used for population dynamics and GASP tests. 3335::kan-tdk and 3338::kan-tdk were transformed onto the ACN1202 strain by puddle transformation. Single colonies from the mutant streak plates were inoculated into 50 mL of deionized water. The colony was heated at 95 degrees Celsius for 30 minutes to induce lysing. Donor DNA was tested for sterility by plating on plain LB agar. After sterility had been verified, two microliters of donor DNA was placed into a 50 microliter puddle of the ACN1202 liquid culture on an LB agar plate. This puddle was grown overnight, and then 1 colony was inoculated into liquid LB media containing 10 micrograms of kanamycin and 5 micrograms of gentamycin per mL. After growth overnight, the liquid culture was streak plated.

Preparation of overnight cultures for Population Dynamics. All strains that were going to compete in GASP tests had to be grown in isolation to ensure that they did not harbor any growth defects. Henceforth, all liquid cultures were grown overnight at 37 degrees Celsius with high aeration. Strains were directly inoculated into 25mL of LB broth from a streak plate and grown for 168 hours. Samples were harvested at 0 (time of mixing), 24, 72, 120, and 168 hours. Cell counts were determined by ten-fold serial dilutions, followed by enumeration on LB agar with antibiotics: gentamycin (ACN1202), chloramphenicol (ACN1203), or gentamycin and kanamycin (mutant strains). At 24 hours and 168 hours the strains were frozen with 5 mL of LB-50% glycerol at negative 70 degrees Celsius for further use in GASP testing. Preparation of overnight cultures for GASP experiments. Four different competitions were established: two control experiments; one with young (24 hours growth) ACN1202 and old ACN1203, the other with young ACN1203 and old ACN1202. The third and fourth competitions were against old (168 hours growth) ACN1202/3338::kan-tdk and young ACN1203, and old ACN1202/3335::kan-tdk and young ACN1203. Four independent trials were conducted on each. Flasks of 25mL of LB broth were inoculated with frozen stocks of young cells, and were incubated at 37 degrees Celsius overnight. Test tubes of 2.5mL of LB broth were inoculated with frozen stocks of 10 day old strains, and were incubated at 37 degrees Celsius overnight.

Competitive GASP experiments in flask cultures. Time 0 was designated as the mixing of 25 microliters of aged cells from the test tube into the flask with 25mL of young cells. Samples of each culture were taken at 0, 24, 48, 120, and 168 hours. Cell counts were determined by serial dilutions of these samples, followed by plating on LB agar: chloramphenicol (ACN1203), gentamycin (ACN1202), or double antibiotic plates for the mutants. Plates were incubated at 37 degrees Celsius overnight and recorded 24 hours later.

Results

Population Dynamics. Figure 4 represents the growth rates of ACN1202, ACN1203 and our two ACN1202 mutants when grown separately. All strains show similar growth.

GASP tests. Irrespective of which antibiotic marker was associated with the older cells (in eight independent trials), the concentration of older cells increased over time in the control GASP tests. Antibiotic resistance was used as a marker and had no effect on growth in same-age flask competitions. An unpaired t-test between the ratio of old to young in normal GASP controls and switched GASP controls revealed that there was not a significant difference between these two strains (p=0.562). Five out of eight of the trials experienced class 1 GASP phenotypes, 2 of the trials experienced class 2 phenotypes and the last trial was compromised due to early cell death (figures 5-8).

The five experiments exhibiting the class 1 phenotype are trials 1 and 2 of figure 5 and trials 2, 3 and 4 of figures 7 and 8 where the antibiotic markers are switched. The two trials that experienced the class 2 phenotype are trials 3 and 4 of figure 6. Trial one of old cells marked with switched antibiotic markers showed no growth at day 5 (Figure 7). 7 out of the 7 trials that showed growth until day 7 experienced a GASP phenotype. This is the first evidence of A. baylyi experiencing the GASP phenotype.

GASP tests were performed with the hypothesis that if competence genes were necessary for the expression of the GASP phenotype then we would observe a different growth rate in the mutant cells. We performed four GASP trials with comP::kan-tdk knock outs where the old cells were comP::kan- tdk knock outs and the young cells were ACN1202. Double antibiotic plates were used to distinguish the mutants. In all four independent trials we found that the cells expressed a class 3 GASP phenotype (figures 9 and 10). This indicates that the older cells could not compete with the younger cells, the opposite effect of what we witnessed in our control GASP experiments. We performed the same independent GASP trials against comQ::kan-tdk mutants and the results show three trials of a class 3 GASP phenotype, although to a less dramatic extent as the comP::kan-tdk trials (figures 11 and 12). The first trial in the comQ::kan-tdk GASP experiments showed no cell growth after day 1 (figure 11).

Figure 13 is a compilation of GASP results. It was constructed by creating a ratio of the average values of growth for old cells over the average value of growth for young cells. It demonstrates that there is a significant difference (p=0.007) between the ratio of population density of old cells to younger cells between the comP::kan-tdk mutants and the controls, and a non-significant difference (p=0.088) between the ratio of population density of old cells to younger cells between the comQ::kan-tdk mutants and the control.

Discussion

We are the first to observe that the GASP phenotype occurs in the genus Acinetobacter. The wild type GASP results are similar to results seen in G.sulfurreducens (Helmus et al. 2011), E. cloacae (Martinez-Garcia et al. 2003), S. typhimurium (Martinez-Garcia et al. 2003), Sh. dysenteriae (Martinez- Garcia et al. 2003), Vibrio fischeri (Petrun & Losroh 2012) and E. coli (Zambrano et al. 1993). The congruence of our results with the previous studies suggests that A. baylyi should be added to the above list.

Additionally, we found that mutations inactivating two of the competence genes of A. baylyi caused alterations in the GASP phenotype during competition with wild type cells. Figure 13 demonstrates that there is a difference in the ratio between population density of old cells and young cells between wild type GASP competition and experiments done with mutants. The failure of A baylyi with knocked out competence genes to express the GASP phenotype suggests that these genes play a role in the expression of the phenotype. Our results do not definitively suggest what this ability might be. When we applied an un-paired t-test to assess these results we found that there is a significant difference between the comP knockout and the control (p=0.007). Notably, there is no significant difference between comQ knock outs and the control growth rates (p=0.088). It is difficult to gauge the significance of this data considering no previous work has implicated competence genes in the GASP phenotype. Additionally, the comQ::kan-tdk mutant behaved very differently across trials, so more trials are needed to determine its most frequent phenotype. To further elucidate the roles of these genes in the GASP response of A. baylyi many paths could be taken. Competence genes have been theorized to be used for nutrient acquisition, DNA repair and genetic recombination. One avenue for study would involve the testing of these mutants on minimal media with only DNA for sustenance, this of course addressing whether the competence genes play a role in nutrient acquisition. This theory was purported originally by Finkel et al in 2001 and further evidence was provided in 2006. These experiments were however performed on E. coli, and A. baylyi could use its homologous competence structure in a completely different way. Another study by Muratov (2010 unpublished) found that both ComP and ComQ are required for DNA repair. The quantity of information about the proposed role of competence genes is truly daunting to sift through, and more research is going to need to be done to establish any definitive results for A. baylyi. Our experiment does not take into account other environmental factors such as temperature extremes and chemical composition of the environment. However, it does provide a simple way of studying bacterial behavior under nutrient depletion, and this aspect is beneficial because it singles out one of the many variables that might encourage transcriptional change in a bacterium’s genome. We think it’s important to note this shortcoming in our experiments because Bacher et al (2006) found that non-competent lineages performed similar to competent lineages when serially passaged at 40 degrees celsius with 300mM NaCl in the media. Note that this experiment still provides ample nutrients, but tests the bacteria’s fitness under other environmental extremes. Nonetheless, their results suggest that under these specific “harsh” conditions that non-competent and competent mutants are essentially similar in their ability to adapt. Interestingly, Bacher et al. (2006) also found that under direct competition during exponential growth phase that, non-competent lineages also performed better than competent lineages. This is exactly opposite of what we found in our experiments. Again, with a discriminating eye it is easy to detect that our experiments were fundamentally different from each other’s. In Bacher et al.’s (2006) experiments, they serially passaged the bacteria in a 1:10,000 ratio into fresh broth every twelve hours. This procedure is obviously different from ours in which the original media was used for 168 hours, and no genetic variation in the population was removed or omitted by the experimenter. This blatant difference provides an explanation for the drastically different results that were observed in our experiments.

A more appropriate comparison to our experiment would be against Finkel et al. (2006) where competition experiments were initiated with knock outs of E. coli lacking one of several competence gene homologs, which can use DNA as a growth substrate but are not naturally transformable. These experiments showed that the E. coli homologs of competence genes provide an advantage during long term stationary phase. These results along with our own provide some weight to the argument that competence machinery is utilized during LTSP and the GASP phenotype.

The above results demonstrate that Acinetobacter baylyi is capable of expressing the growth advantage in stationary phenotype. The mechanisms by which this phenotype asserts itself are still shrouded in uncertainty, but we have determined that aged cells lacking the ComP or ComQ proteins are not able to mount a normal GASP response when competed against wild type cells. In the best- studied model for LTSP, E. coli, mutations in rpoS, SOS DNA polymerases, the methyl-mismatch repair system and the mut genes have contributed to the GASP phenotype (Finkel 2006). Our results show that in a naturally competent bacterium, A. baylyi, competence genes are also important for development of a normal GASP response.

Acknowledgements

Thesis advisor: Phoebe Lostroh

Secondary Advisor: Nancy Huang

Figge-Bourquin Grant Colorado College Venture Grant

Colorado College Biology Department

References

1) Averhoff, B., and I. Graf. 2008. The Natural Transformation System of Acinetobacter baylyi ADP1: A Unique DNA Transport Machinery, p.119-149. In U. Gerischer (ed.), Acinetobacter Molecular Microbiology. Caister Academic Press, Norfolk, UK. 2) Averhoff, B., Friedrich, A. 2003. Type IV pili-related natural transformation systems: DNA transport in mesophilic and thermophilic bacteria. Archives of Microbiology. 180:6. 385-393. 3) Bacher, JM., Metzgar, D., de Crecy-Lagard, V. 2006. Rapid evolution of diminished transformability in Acinetobacter baylyi. Journal of Bacteriology 188:24-8534-8542. 4) De Berardinis, V., D. Vallenet, V. Castelli, M. Besnard, A. Pinet, C. Cruaud, S. Samair, C. Lechaplais, G. Gyapay, C. Richez, M. Durot, A. Kreimeyer, F. Le Fevre, V. Schachter, V. Pezo, V. Doring, C. Scarpelli, C. Medigue, G.N. Cohen, P. Marliere, M. Salanoubat, and J. Weissenbach. 2008. A complete collection of single-gene deletion mutants of Acinetobacter baylyi ADP1. Molecular Systems Biology. 4:174. 5) Drake, JW., Charlesworth, B., Charlesworth, D., Crow, JF. 1998. Rates of spontaneous mutation. Genetics. 148, 1667- 1686 6) Farrell, MJ., Finkel, SE. 2003. The growth advantage in stationary-phase phenotype conferred by rpoS mutations is dependent on the pH and nutrient environment. Journal of Bacteriology. 185:24. 7044-7052. 7) Finkel, SE., 2006. Long-term survival during stationary phase: evolution and the GASP phenotype. Nature Reviews Microbiology. 4:2. 113-120. 8) Finkel, SE., Kolter, R. 2001. DNA as a nutrient: novel role for bacterial competence gene homologs. Journal of Bacteriology. 183:21. 6288-6293. 9) Finkel, SE., Zinser, E., Kolter, R. 2000. Bacterial Stress Responses (EDS Storz, G., Hengge-Aronis, R.) 231-238 (ASM Press, Washington DC). 10) Hare, JM., Adhikari, S., Lambert, KV., Hare, AE., Grice, AN. 2012. The Acinetobacter regulatory UmuDAb protein cleaves in response to DNA damage with chimeric LexA/UmuD characteristics. FEMS Microbiology letters. 334, 57-65. 11) Hare, JM., Perkins, SN., Gregg-Jolly, LA. 2006. A constitutively expressed, truncated umuDC operon regulates the recA-dependent DNA damage induction of a gene in Acinetobacter baylyi strain ADP1. Applied and Environmental Microbiology. 72, 4036-4043. 12) Helmus, RA., Liermann, LJ., Brantley SL., Tien, M. 2011. Growth advantage in stationary phase (GASP) phenotype in long-term survival strains of Geobacter sulfurreducens. FEMS microbiology ecology. 79:1. 218-228. 13) Ines, C., Dubnau, D. 2004. DNA uptake during bacterial transformation. Nature Reviews Microbiology. 2. 241-249. 14) Martinez-Garcia, E., Tormo, A., Navarro-Llorens, JM. 2003. GASP phenotype: presence in enterobacteria and independence of sigma(S) in its acquisition. FEMS microbiology letters. 225:2. 201-206. 15) Palchevskiy, V., Finkel SE. 2006. Eschrichia coli competence gene homologs are essential for competitive fitness and the use of DNA as a nutrient. Journal of Bacteriology. 188:11. 3902-3910. 16) Palmen, R., P.Buisjsman, and K.J. Hellingwerf. 1994. Physiological Regulation of Competence Induction for Natural Transformation in Acinetobacter calcoaceticus. Archives of Microbiology. 162. 344-351. 17) Porstendorfer, D., U. Drotschmann, and B. Averhoff. 1997. A novel competence gene, comP, is essential for natural transformation of Acinetobacter sp. Strain BD413. Applied Environmental Microbiology. 63, 4150-4157. 18) Redfield, RJ., Schrag, MR., Dean, AM. 1997. The evolution of bacterial transformation: sex with poor relations. Genetics. 146:1. 27-38. 19) Redfield, RJ. 1988. Evolution of Bacterial Transformation- is sex with dead cells ever better than no sex at all. Genetics. 119:1. 213-221. 20) Yeiser, B., Pepper, E.D., Goodman, M. F., Finkel, SE. 2002. SOS-induced DNA polymerases enhance long-term survival and evolutionary fitness. Proceedings from the National Academy of Science. USA. 88. 7160-7164. 21) Zambrano, M. M., Kolter R. 1996. GASPing for life in stationary phase. Cell. 86, 181-184. 22) Zambrano, M. M., Siegele, D. A., Almirôn, M., Tormo, A., Kolter, R. 1993. Microbial competition: E. coli mutants that take over stationary phase cultures. Science. 259, 1757-1760.

Figure 1: Model of normal bacterial life cycle. From hours 0 to 6 you see lag phase, 6-12 exponential phase, 12-48 stationary phase, 48-36 death phase, and onward from 36 hours we see the GASP phenotype during long term stationary phase (Finkel 2006).

Figure 2: Graphical representation of the four classes of the GASP phenotype. The red line denotes older cells, the green line the young cells (Finkel 2006).

Figure 3: Proposed competence machinery. (Averhoff 2003)

Figure 4- Population Dynamics: This graph shows changes in population density between the individual strains used in this experiment. The control strains are labeled 1203 and 1202 and correspond to strains ACN1203 and ACN1202 respectively. The mutants are labeled comQ and comP and correspond to strains 3335::kan-tdk and 3338::kan-tdk respectively. T1 and T2 correspond to independent trials.

9

8

7

6

5 Young Trial 1

4 Old Trial 1 CFU/mL(log) 3 Young Trial 2 Old Trial 2 2

1

0 0 1 2 3 4 5 6 7 8 Days after inoculation

Figure 5: Control trials 1 and 2. Young demarcates ACN1202 and old demarcates ACN1203.

9 8 7

6 5 Young Trial 3 4 Old Trial 3

CFU/mL(log) 3 Young Trial 4 2 Old Trial 4 1 0 0 2 4 6 8 Days after inoculation

Figure 6: Control trials 3 and 4. Young demarcates ACN1202 and old demarcates ACN1203

Figure 7: Control trials 1 and 2 with switched antibiotic markers where young demarcates ACN1203 and old demarcates ACN1202.

10 9 8

7

6 Young Trial 3 5 Old Trial 3 4 CFU/mL(log) Young Trial 4 3 2 Old Trial 4 1 0 0 2 4 6 8 Days after inoculation

Figure 8: control trials 3 and 4 with switched antibiotic markers where young demarcates ACN1203 and old demarcates ACN1202.

9 8 7

6 5 Young Trial 1 4 Old Trial 1

CFU/mL(log) 3 Young Trial 2 2 Old Cells 2 1 0 0 2 4 6 8 Days after inoculation

Figure 9: Com P trials 1 and 2 where old demarcates ACN1202/3338::kan-tdk and young demarcates young 1203.

Figure 10: Com P trials 3 and 4 where old demarcates ACN1202/3338::kan-tdk and young demarcates young 1203.

9 8 7

6 5 Young Trial 1 4 Old Trial 1

CFU/mL(log) 3 Young Trial 2 2 Old Trial 2 1 0 0 2 4 6 8 Days after inoculation

Figure 11: Com Q trials 1 and 2 where old demarcates ACN1202/3335::kan-tdk and young demarcates ACN1203.

Figure 12: Com Q trials 3 and 4 where old demarcates ACN1202/3335::kan-tdk and young demarcates ACN1203.

Figure 13: GASP compilation