Why both slow and fast dividing subpopulations of bacteria are missed during routine culture?

During routine culture, the number of bacteria on solid agar medium that can form distinct colonies, separated from each other are usually around 100. Among 100 colonies, we may not notice slow or fast dividing subpopulations.

It is easy to explain why we miss slow dividing bacteria during routine culture- their number is very small and they divide slowly. By the time these subpopulation of bacteria start to grow and form colonies, the normal bacterial population would have already grown and covered the agar. Hence, to isolate those subpopulations of small colony variats (SCV), the normally dividing population needs to be eliminated, which can be done with the help of aminoglycosides (see the blogpost on Oct.5).

On the other hand, one would expect to notice fast dividing subpopulation more frequently. Even if their initial number is low, one would expect them to gradually dominate the population since their growth rate is higher than the normal. However, this may not happen if this fast dividing subpopulation undergoes senescence. As they divide, the growth rate of mother cell gradually reduces and becomes comparable to the normal population whereas the daughter cell may be a rejuvenated offspring. It may be due to the mother cell undergoing senescence that the fast dividing subpopulation does not dominate the culture. However, the fast dividing subpopulation can be selected by removing the normal population by repeatedly growing the culture in early exponential phase (as described in the previous blog on Sep.19).

Thus, bacterial senescence can explain why the fast dividing, hypervirulent subpopulation of bacteria that can be isolated by repeatedly growing the culture in early exponential phase does not dominate the whole population even if they have the growth advantage over the normal population.

Next- Shifting bacterial population distribution to the right or left

Experiment that could indicate that SCVs are senescent bacteria

One of the markers of senescence is the accumulation of oxidative carbonylated proteins. Cells of an E. coli population show asymmetry not only with respect to growth rate, but also with respect to protein oxidation levels (Desnues et al. 2003; Aguilaniu et al. 2003). An E. coli population consists of relatively low damaged daughter cells (low protein oxidation) that are reproductively competent and damaged mother cells with reduced reproductive ability (Desnues et al. 2003). In exponentially growing E. coli, the amount of protein aggregates increases over time and were found to be more prevalent in dead cells than in culturable cells (Maisonneuve et al. 2008a). Similarly, aggregated proteins accumulate in cells with older poles, which are associated with a decrease in reproductive ability (Lindner et al. 2008).

In the earlier blog, I had hypothesized that small colony variants of E. coli isolated using subinhibitory concentrations of aminoglycosides are senescent bacteria which are hypovirulent, slow dividing and form small colonies on solid medium. Measuring the levels of protein carbonylation can give an indication whether they are senescent bacteria. If they are senescent bacteria, one can expect the protein carbonylation levels to be high. As far as I know, level of protein carbonylation in small colony variants has never been measured (a search of “small colony variants” and protein carbonylation returned only two results in Google scholar).

However, increased protein carbonylation may not be conclusive evidence that SCVs are replicative senescent bacteria. Increased carbonylation can be a feature of both conditional and replicative senescence. Hence further research may be required to differentiate between these two. In fact, there is possibility that mutants that form SCVs (like hemin, menadione or thiamine mutants) may also show increased carbonylation which may be due to conditional senescence and not replicative senescence. Hence, I assume that both mutant SCVs and non-mutant SCVs may show increased protein carbonylation, the former due to conditional senescence and the latter due to replicative senescence.

If the yeast cells isolated by repetitively growing in early exponential phase have low number of bud scars (described earlier) and SCVs isolated using aminoglycosides have increased protein carbonylation levels, I can say with increased confidence that my model of bacterial aging is different from that proposed by Stewart et al. (2005) and that SCVs are senescent bacteria

Next- Why both slow and fast dividing subpopulation of bacteria are missed during routine culture?

Stewart et al. (2005). Aging and death in an organism that reproduces by morphologically symmetric division. PLoS Biol 3(2), e45.

Aguilaniu et al. (2003). Asymmetric inheritance of oxidatively damaged proteins during cytokinesis. Science 299(5613), 1751-3.

Desnues et al. (2003). Differential oxidative damage and expression of stress defence regulons in culturable and non-culturable Escherichia coli cells. EMBO Rep 4(4), 400-4.

Maisonneuve et al. (2008). Protein aggregates: an aging factor involved in cell death. J Bacteriol 190(18), 6070-5.

Lindner et al. (2008). Asymmetric segregation of protein aggregates is associated with cellular aging and rejuvenation. Proc Natl Acad Sci U S A 105(8), 3076-81.

Isolation of slow dividing, small colony forming, hypovirulent, senescent bacteria

                   

                   In a normally dividing population, a small subpopulation of slow dividing bacteria is present which can be isolated using aminoglycoside antibiotics. They are termed as small colony variants (SCV). SCVs had already been discussed before (please check the posts in the month of August). They constitute a naturally occurring, slow-growing subpopulation of bacteria that form small colonies (less than one-tenth of the size of parent colonies) on solid media (Proctor et al. 2006). Much has been published on the biochemical aspects and the significance of SCVs. However, there are two areas where I have difference of opinion from those in published articles.

1. SCVs are mutants that revert to normal growth in the presence of auxotrophic agents

2. SCVs are responsible for chronic infections

Whereas a number of mutants form SCVs and can be reverted to normal growth after adding hemin, menadione, thiamine or thymidine, all SCVs isolated in vitro after adding aminoglycosides may not be specific mutants. In fact all SCVs are not similar and may exhibit different protein profiles (Kriegeskorte et al. 2011). Similarly, the role of SCVs in chronic infections is questionable (please read the previous posts).

A pure culture of SCVs of E. coli DH-5alpha cells can be isolated after treating cells with subinhibitory concentration of aminoglycosides like kanamycin as explained in Jacob (2007). In short, 50 ul of stationary phase culture is added to 3 ml of fresh LB medium containing kanamycin at different concentrations and incubated for 2 days. Three factors are important to get a pure culture of SCVs- initial inoculum size, concentration of antibiotic and the total time of incubation. If the inoculum size is very low, SCVs may be missed, but if high, some normally dividing bacteria that have escaped killing may overgrow and mask SCVs. Since they are slow dividing bacteria, SCVs may take longer time to grow. With different concentrations of kanamycin, colonies of different sizes can be obtained.

The slow dividing SCVs have been shown to be hypovirulent also (Sifri et al. 2006). But, how can it be proved that they are senescent bacteria?

Next- Experiment that could indicate that SCVs are senescent bacteria

 

Proctor et al. (2006). Small colony variants: a pathogenic form of bacteria that facilitates persistent and recurrent infections. Nat Rev Microbiol 4(4), 295-305.

Kriegeskorte et al. (2011). Small colony variants of Staphylococcus aureus reveal distinct protein profiles. PROTEOMICS, 11: 2476–2490.

Jacob, J (2007). Persisters show heritable phenotype and generate bacterial heterogeneity and noise in protein expression . Available from Nature Precedings http://hdl.handle.net/10101/npre.2007.1411.1

Sifri et al. (2006). Virulence of Staphylococcus aureus small colony variants in the Caenorhabditidis elegans infection model. Infection and Immunity, 74(2);1091-1096.

Experimental proof that the fast dividing, hypervirulent bacteria are young bacterial population

In the last blog, I assumed that the fast dividing, hypervirulent bacteria isolated after repeated growth in early exponential phase are young bacterial population. This can be confirmed by doing the same experiment in Saccharomyces cerevisiae.

When yeast cells undergo replication by budding, a bud scar is left behind on mother cell’s surface. Bud scars remain permanently deposited on the surface and get accumulated as mother cells undergo more divisions. Thus, an old mother cell will have more number of bud scars whereas the number will be less in relatively young mother cells. Thus, the bud scars are used as a marker for the number of divisions a cell had undergone or the budding index. The budding index can be calculated by analyzing the cell wall for bud scars using confocal microscopy.

For the experiment, when a normally dividing yeast culture reaches an O.D. of 0.3-0.5, 100 ul of the culture should be withdrawn and added to 3 ml of fresh culture medium and incubated for growth. This process should be repeated 3-4 times. At the end of the fourth cycle, when yeast cells remain in early exponential phase itself, the budding index should be calculated. This can be compared with a yeast culture in one-time exponential phase and also with a culture in stationary phase. If the cells obtained after repeated culturing in early exponential phase are indeed young yeast cells, the number of bud scars will be lower than that in other two.

If they are found to be young yeast cells, the same can be true for bacteria also.

Next- Isolation of slow dividing, small colony forming, hypovirulent, senescent bacteria

Isolation of fast dividing, hypervirulent, young bacteria- a simple experiment that can reveal many interesting properties

In one of my experiments aimed at eliminating persister bacteria (a small subset of slow growing bacterial population) from a normally dividing population, a culture of bacteria was repeatedly grown in early exponential phase (Jacob 2007). This experiment is simple and inexpensive (only an incubator and a spectrophotometer are needed) and quick (can be finished in a few days), but give some interesting results regarding bacterial growth kinetics.

In this experiment, when a culture of bacteria reached an O.D. of 0.3- 0.5 ( i.e. at a stage of light turbidity of the medium), 100 ul of the culture was transferred to 3 ml of fresh medium and incubated again. The same procedure was repeated 4 times. Repeated selection of bacteria at early exponential phase eliminated all slow growing bacteria. At the end of 4^th cycle, when the bacterial culture was allowed to continue its growth and reach a stationary phase, some interesting properties were noticed.

1. The growth rate of the selected bacteria was found to be much higher than a normally dividing culture

2. At stationary phase, the number of bacteria per ml of medium was higher

3. Persisters that were absent initially after repeatedly grown in early exponential phase reappeared at stationary phase

4. GFP expression of selected bacteria were higher

5. Activity of ampicillin was much reduced against the selected bacteria especially at higher initial inoculum size (in fact, bacteria grew as if antibiotic was not added; however, this lack of activity was not due to antibiotic resistance)

In the above article, only the third property was reported since the article was focused on the phenotypic shift of persisters. Based on the increased GFP expression and lack of activity of ampicillin (which I attribute to fast growth and increased production of enzymes that destroy the antibiotic, especially at higher inoculum size), I assumed that those bacteria were hypervirulent. Indeed, the hypervirulence of selected bacteria was later reported by Chapuis et al. (2011) after injection of Xenorhabdus nematophila into insects. My interpretation to the above result is that these fast dividing, hypervirulent bacteria are very young bacteria that are found in low in numbers in a normally dividing population (see the earlier blogpost).

Just because the selected bacteria are fast dividing and hypervirulent does not prove that they are young bacteria. However, it is possible to prove it with the help of another experiment.

Next- The experiment that can prove that the fast dividing, hypervirulent bacteria are young bacteria

Jacob, J (2007). Persisters show heritable phenotype and generate bacterial heterogeneity and noise in protein expression . Available from Nature Precedings <http://hdl.handle.net/10101/npre.2007.1411.1>

Chapuis et al. (2011). Virulence and pathogen multiplication: A serial passage experiment in the hypervirulent bacterial insect-pathogen Xenorhabdus nematophila. PLoS ONE 6(1): e15872. doi:10.1371/journal.pone.0015872

Population distribution of an E. coli colony- comparison with Stewart et al. (2005) model

As per Stewart et al. (2005) (discussed in the last blog), the daughter cell formed from a mother cell is a rejuvenated offspring with full reproductive potential. So, if we designate x for virgin cells, x+1 for cells that have undergone 1 division, x+2 for those undergone 2 divisions and so on, then approximately 50% of a normally dividing  population in a colony of bacteria will be virgin cells (x), 25% will be x+1, 12.5% will be x+2 and so on. Hence if we plot the percentage of different populations on a graph, a half-bell curve distribution will be obtained. However, it may not be as perfect as one below as the growth rate of all cells are not the same (mother cells have reduced growth rate as it undergoes senescence).

However, my model of E. coli senescence is closer to that of S. pombe senescence (discussed on September 6 blogpost). In this model, the old cells do not give rise to rejuvenated offspring, but generate old cells itself. However, majority of the cells are relatively young cells with limited damaged or carbonylated proteins and this pool of young cells are large enough to prevent the extinction of the population. Another important feature in my model is the presence of a small population of virgin cells with no damaged or carbonylated proteins. Thus a major difference between Stewart et al. (2005) model and my model is that in the former, majority of the population are virgin cells (rejuvenated offspring) whereas in the latter, majority of the population are young bacteria with limited carbonylated proteins, along with a small population of virgin cells.

Thus, the population distribution in my model follows a normal distribution curve. However, it may not be a perfect bell shaped curve but can be skewed in favor of young cells. A bell-shaped curve indicates that there are small populations of fast-dividing young bacterial cells as well as slow-dividing senescent cells in any bacterial colony.

Is it possible to isolate any of these small populations? The answer is yes – both young cells and old cells can be isolated. In the next few sections, I will discus how to isolate these two populations separately.

Next- Isolation of fast dividing, hypervirulent, young bacteria 

Stewart et al. (2005) model of replicative senescence in E.coli

Perhaps, the earliest evidence of replicative senescence in bacteria was provided by Liu (1999). By tracking the bacterial growth in liquid media with high viscosity, Liu (1999) observed the unidirectional growth and reproduction of E. coli. He proposed that the bacterium has an intrinsic cell polarity with one end behaving as a mother compartment and the other end as the daughter compartment resulting in the formation of two bacteria of succeeding generations. His model defined bacterial age by its experienced chronological time. Based on this model, he predicted that, on bacterial division, the old strand of DNA remain with the mother bacterium whereas the new strand goes to the daughter bacterium and that this distribution of old and new strands of DNA between the mother and daughter cells is responsible for the intrinsic differences between the two.

Later, Stewart et al. (2005) studied the senescence in E. coli using automated time lapse microscopy by following repeated cycles of reproduction. They followed individual exponentially growing cells up to nine generations of growth and reproduction. Their findings were comparable to those reported previously by Liu (1999). The bacterium exhibits cell polarities which give rise to an old pole with a reduced growth rate and a new pole with higher growth rate. They found that the average growth rate of old pole cells was 2.2% slower than that of new pole cells and that the new pole cells were larger and divided sooner than the old pole cells. In addition, the old pole cells were also more likely to die than the new pole cells. They concluded that the two apparently identical cells are functionally asymmetrical, with the old pole cell behaving as the aging mother cell and the new pole cell as the rejuvenated offspring.

Thus, according to these models, bacteria undergo aging and that the growth rate of mother cell decreases with age whereas the daughter cell produced from the mother cell is a rejuvenated offspring with high growth rate which helps to maintain the bacterial lineage.

Next- Population distribution of an E. coli colony- a comparison with Stewart et al. (2005) model

Liu, S. V. (1999). Tracking bacterial growth in liquid media and a new bacterial life model. Science in China 42, 644-654.

Stewart et al. (2005). Aging and death in an organism that reproduces by morphologically symmetric division. PLoS Biol 3(2), e45.