No. Not all the time. There is an unfortunate habit of biochemists to get too wrapped up in the genetics to think about epigenetics (control of genetic expression by proteins), especially with 'simpler' organisms like bacteria. Persistence is the property of a totally identical population of bacteria to respond differently to antibiotics, despite having identical genomes they are phenotypically (phenotype = set of observable characteristics) different.
Persistent bacteria aren't only seen in response to antibiotic's either...they can be found for many types of stress, such as heat or starvation. They've been roughly split into two different types:
- Type I persisters: form in response to stress, usually at stationary phase (i.e after the initial burst of bacterial growth)
- Type II persisters: are seen forming throughout the bacterial lifecycle
Although these sound dangerous at first glance, the medical implications of persisters are relatively limited (probably the reason why the mechanism has not been more thoroughly studied). As they have a slower growth rate, and as there are so few of them within an overall susceptible population, they can usually be cleared relatively easily by the immune system. There are, therefore, only three main areas where they are clinically relevant: immunosuppressed patients, pathogens that have adapted to the immune system, and in niches in the body that are less available to the immune system (such as within a biofilm).
The main ecological and evolutionary implications of this are that a colony of identical bacteria can 'survive' antibiotic attack by having a few of its members able to withstand that attack. Although they will grow slower under normal conditions, their ability to withstand stressful conditions means that they can essentially recover the whole colony, without the need for a major genetic change. This is true especially of diseases such as tuberculosis, where very few bacteria are needed to re-start an infection.
One thing that would be really interesting would be to study the behaviour of persisters within biofilms. Along with swarming and quorum sensing, biofilms are an example of pseudo-multicellular bacterial behaviour, circumstances under which having a small population of cells that can regenerate the whole colony would be very useful.
Orit Gefen & Nathalie Q. Balaban (2009). The importance of being persistent: heterogeneity of bacterial populations under antibiotic stress FEMS Microbiology Reviews, 33 (4) : 10.1111/j.1574-6976.2008.00156.x