Madhadron:PloidyMeasurement

Two parts: flow cytometry to determine distribution, and DNA concentration measurement to calibrate it.

DNA Concentration

1. Grow M. smegmatis
2. Take 100μL of culture, and do dilution plating to get population count.
3. Centrifuge for 20 minutes at 3,000rpm, resuspend to concentration ~10^9 cells/mL in STE buffer (0.1M NaCl, 50mM Tris HCl, 1mM Na₂EDTA) to ~10^7 cells/mL.
4. Add sodium dodecyl sulfate to 0.1%, incubate for 10 minutes at 60C
5. Add proteinase K (100μg/mL). Incubate at 37C for 30 minutes.
6. Add KCl to final concentration 40mM, incubate on ice for 30 min.
7. Centrifuge at 10,000g for 20 min at 5C, and discard pellet.
8. Stain supernatant with Hoechst 33258 at concentration 0.05 μg/mL.
9. Measure fluorescence at ex 350nm, em 450nm.

(Adapted from "Determination of DNA Content of Aquatic Bacteria by Flow Cytometry by Button and Robertson, Applied and Environmental Microbiology, Apr 2001.)

Controls: same procedure with no cells, vary the number of cells and see fluorescence variation

Compare number that comes out with Sigma's standard curve for calf thymus. Correct for mycobacterial GC/AT ratio as compared to calf.

Flow Cytometry

1. Grow M. smegmatis
2. Stain with orange cell cycle stain from Invitrogen

Get a pile of events. The mean of this distribution should be the value measured above.

Analysis

The DNA concentration gives a mean intensity $$\langle I_c \rangle$$ = a\langle n \rangle + b = f(s\langle n \rangle)$$, where $$f(\langle n \rangle)$$ we can find from Sigma's curves for DNA concentration vs. intensity, compensating for G/C content in mycobacteria, and $$s$$ is the weight of one chromosome. Let $$f_0 (\langle n \rangle)$$ be the curve from Sigma, $\gamma_{calf}$ be the GC/AT fraction in calf DNA and $\gamma_{myco}$ be that in mycobacteria. Then $f(\langle n \rangle) = $$ (FIXME! What does Hoechst bind to? Same thing as DAPI?).

The flow cytometry produces a set of intensity values. We let $$N(I)$$ be the number of chromosomes as a function of intensity, and assume it is linear, $$N(I) = cI+d$$. We find $$c$$ and $$d$$ by minimizing the three functions $$\langle N(I) \rangle - \langle n \rangle$$, $$\textrm{round}(N(I_m^i)) - N(I_m^i)$$, where I_m^i is the $i$th maximum of the density, which we assume must correspond to an integer.