68 4 Rational Efforts to Streamline the Escherichia coli Genome
4.7.4 Additional Genome Reduction Studies
In an early proof-of-concept work, random genomic deletions were constructed
in MG1655 by developing a method involving repeated integration/deletion of a
Tn5 transposon derivative. Deleted regions could be rescued on a conditionally
replicating plasmid, allowing identification of essential genes. The extent of the
genome reduction in the most deleted strain was about 5.6%, as estimated by
pulsed-field electrophoresis [43] (Figure 4.4).
Utilizing a pre-mapped transposon insertion library, a semi-random method
was applied to reduce the genome of MG1655 by 6.7% (Figure 4.4) [73]. A pair of
selected transposon insertions could be combined in a single cell by P1 transduc-
tion, and the genomic region between them could be excised by the Cre/lox sys-
tem. Combining of deletions in a single genome was also achieved by P1
transduction. In some multiple deleted strains, synthetic lethality was observed:
some deletions were individually viable but were lethal when combined. This
genome engineering strategy, producing large sets of mapped transposon inser-
tions ready for pairwise combination, followed by Cre/lox-mediated in between
deletion, is most useful when deletion of a particular region of the genome is
desired.4.8 Selected Research Applications of Streamlined-Genome E. coli
Genome E. coli
4.8.1 Testing Genome Streamlining Hypotheses
The MDS series with increasing number of genomic deletions provides a con-
venient model for studying the impact of stepwise genome streamlining on cel-
lular traits, addressing unsettled questions of reductive genome evolution [100].
A comprehensive study showed that deletions caused a gradual fitness loss,
decreased nutrient utilization, and induced a general stress response. Growth
yield and maintenance energy were measured in chemostat cultures of MG1655,
MDS42, and MDS69 under nutrient limitation. Both carbon and nitrogen utili-
zation efficiencies decreased in the multideletion strains without significantly
affecting the maintenance energy requirement of the cell. These results argue
against the adaptive genome streamlining hypothesis [102, 103]. Results sup-
ported the notion that selection for reduced DNA synthesis per se is unlikely to
reduce genome size in the course of evolution of small genomes. No general
trend was found between growth rate and genome size, neither between cell size
and genome size. Genome reduction was also shown to cause transcriptome
reprogramming. Many targets of the general stress sigma factor RpoS were
upregulated in MDS42 and MDS69. rprA, a small regulatory RNA that facilitates
RpoS translation was strongly induced, and, as expected, the MDS42 and MDS69
had elevated acid resistance. These studies revealed an unexpectedly significant
role of horizontally transferred genes not only in stressful environments but also
under routine growth conditions.