10. Ancestor
Evo 1
Evo 2
Evo 3
Evo 4
MIC of
Clone
Generation50,000clones
Experiment – Assess the antibiotic susceptibility of
the ancestor and generation 50,000 clones
12. Is antibiotic resistance maintained in a
population that evolved hypermutability?
Wielgoss S., et al. PNAS. (2013)
Hypermutable
phenotype evolves
13. Experiment – Assess the antibiotic susceptibility
of a population that evolved hypermutability
0K 0.5K 2K 5K 10K 20K 30K 40K 50K
Measure each clone’s susceptibility
to ampicillin using selection assays
15. Conclusions
• On balance, antibiotic resistance
traits decay during bacterial
evolution in the absence of
drugs
• Antibiotic resistance was
maintained in one population
• The removal of antibiotics is
likely a promising strategy to
reverse resistance over time
16. Acknowledgments
• Rich Lenski
• Guidance committee:
• Chris Waters
• Chris Adami
• Frances Downes
• Neerja Hajela
• Jasper Gomez
• Thomas LaBar
18. Inoculate culture medium
with LTEE strain
Two-fold increasing concentration of antibiotic
Positive control
(no antibiotic)
Streak for isolated
clones
Pick one isolated clone
and inoculate culture
Resuspend
cells from
culture in
saline
Spread
suspension
onto
selection
plates
~2x106
cells/plate
Incubate plates for 2
days
MIC of
Clone
Antibiotics used:
• Tetracycline
• Ampicillin
• Ciprofloxacin
• Ceftriaxone
Selection
assay
MIC of
Mutant
19. Experimental
blocking scheme
Ancestral
streak plate
Replicate cultures
each founded with
a single ancestral
clone
Evo 2 streak
plate
Evo 1 streak
plate
Evo 4 streak plate
Evo 3 streak
plateExperimental
Block
Each replicate
culture founded with
single evolved clone
per strain
20. Statistical methodology
• Four independent tests of hypotheses
• Given AP and MA, evolved clones are more
susceptible than their ancestral counterparts
• Given global epistasis, evolved clones have a
greater evolutionary potential in proportion to
their ancestral counterparts
• Determine p-value for each test and combine
values using Fisher’s method for combining
probabilities
• Individual independent tests may not be significant,
but the p-values are overall lower than would be
expected by chance
χ2𝑘
2
~ − 2
𝑖=1
𝑘
ln(𝑝𝑖)
p = 0.00459
Ancestor
Evo 3
Evo 4
Evo 2
Evo 1
-2.0
-1.5
-1.0
-0.5
0.0
0.5
1.0
1.5
2.0
Susceptible
Genotype
Resistant
Genotype
MedianLog2MICof
ampicillin(μg/ml)
21. Statistical methodology
• Idiosyncratic epistasis predicts that there is
significant heterogeneity in fitness increases
among lines that is independent of each line’s
initial fitness
• Performed a Kruskal – Wallis test
• Followed by testing all pairwise
comparisons post hoc using Dunn’s test
with a conservative critical P-value of
0.005
p = 0.0376
Ancestor
Evo 3
Evo 4
Evo 2
Evo 1
-0.5
0.0
0.5
1.0
1.5
2.0
2.5
Susceptible
Genotype
Resistant
Genotype
MedianLog2MICof
tetracycline(μg/ml)