Circadian Clocks in Microorganisms

Many papers in chronobiology state that circadian clocks are ubiqutous. That has been a mantra since at least 1960. This suggests that most or all organisms on Earth possess biological clocks.

In the pioneering days of chronobiology, it was a common practice to go out in the woods and collect as many species as possible and document the existence of circadian rhythms. Technical limitations certainly influenced what kinds of organisms were usually tested.

Rhythms of locomotor activity are the easiest to measure. Rodents, as well as large walking insects like cockroaches, will turn running wheels, each revolution triggering a switch that sends a signal to the computer. Songbirds will jump from one perch to another, each perch flipping a switch connected to a computer. Lizards, while walking around the cage will tilt the cage from left to right around an axis - a metal bar on the bottom - which will turn a switch. Plants that exhibit leaf movements (closing at night, opening during the day) were the prime experimental models for a while (e.g., Kalanchoe, mimosa, tobacco).

Monitoring rhythms in other organisms is much harder: it is mighty difficult to make a fish run in a running wheel, or build hopping perches sensitive enough to be triggered by the landing of a butterfly. That was even harder back in the late 1940s and early 1950s when most of this work was done.

It is no suprise that nobody looked at microorganisms back then - it was just technically too hard. The fact is that most of the pioneers in the field came in from vertebrate physiology, ethology or ecology. It is easy for us, large mammals, to forget that we are not among the dominant life-forms on the planet - that title goes to bacteria, in terms of numbers of individuals, in terms of biodiversity, and in terms of total biomass. See if you can find mammals, or even all animals on the Tree of Life (click to enlarge):
Some old papers, mostly parts of Conference Proceedings of various kinds, mention as fact that Bacteria do not have clocks but do not provide any citations. It took me years to dig out three papers (Rogers and Greenbank, 1930, Halberg and Connor, 1961; and Sturtevant, 1973) with relevance to this question and all three are ambiguous about the final verdict. Why is nobody revisiting this with modern molecular techniques?

Being unicellular does not preclude one from having a clock, though, as single-cell Protista and Fungi all have circadian rhythms, which have been studied quite extensively since the 1970s or so (I intend to delve some more in that literature and write some posts on them in the future).

One group of bacteria does have a clock - the unicellular Cyanobacteria (if you are above a certain age, you may remember them under their old name: blue-green algae), in particular those species that do not form chains, e.g., Synechococcus and Nostoc. This was discovered very recently - only ten years ago (Mori et al. 1996). I was two years into my Masters when that paper appeared and I remember the excitement. I will certainly write a post or two on those soon:
There has not yet been a single study of any kind of rhythmicity in Archaea. Most of those microorganisms live in strange places - miles deep under the surface of the earth, in rocks, in ice, on the ocean floors and in the hydrothermal vents. They mostly do not inhabit rhythmic environments, so perhaps they do not need to have clocks - but it would be really nice to know if that is really the case:
Old Faithful, the famous geyser in Yellowstone park contains Archea. As the geyser erupts every 45 minutes or so, the microbes are suddenly exposed to very different environment: light, turbulence, lower temperature. Should we expect them to evolve a 45-minute clock that will help them predict the eruption so they can limit some sensitive biochemical reactions to the quiet periods and switch on the defenses agains light and cold every 45 minutes?

In The Geometry of Biological Time, Arthur T. Winfree suggested an experiment (on Page 580) that it

"... should be possible to demonstrate the effect by bacterial selection experiments in a chemostat. By alternating the nutrient influx from glucose without oxygen, to oxygen without glucose, to alanine and oxygen, cells would be forced into a three-point metabolic cycle." and "... reversing the order of the driving cycle, it should be possible also to select cells whose clocks run backward."

In a later edition (after we learned that cyanobacteria have clocks) he suggested, instead, to use

"one of the species of cyanobacteria that revealed no circadian rhythms in surveys before Mori et al. (1996), and use light as the alternative nutrient".

As of today, nobody has performed such an experiment, although Elowitz and Leibler (2000) came pretty close with a study in which they produced oscillations in Escherichia coli with periods of 3-4 hours, which are slower than the cell-division cycle:
So, if most of Life on Earth is Prokaryotic (Eubacteria and Archaea), and those groups do not have clocks, then clocks are not ubiqutous, are they? In my papers and in my Dissertation I try to hedge a bit by stating that they are found in "organisms that live on or close to the surface of the Earth", thus at least avoiding the deep-oceanic, deep-soil, and parasitic microorganisms (as well as burrowing and cave organisms that may have secondarily lost their clock).