Interestingly, the senior author on this paper is Martin Brasier and he was one of the scientists who challenged the earlier result of William Schopf. Read all about it on Jerry Coyne's
The fossils are associated with a sulfur-rich mineral called pyrite. This mineral is produced by modern sulfate-reducing bacteria and it's reasonable to assume that the primitive bacteria detected in these ancient rocks also carried out sulfate reduction. That's not surprising since there wasn't much oxygen in the deep ocean 3.5 billion years ago.
Jerry Coyne didn't explain the biochemistry so let me give it a try. We need to start with a brief lesson on how modern living bacteria make ATP. The key to this understanding is a knowledge of Oxidation-Redution Reactions—one of the important basic concepts in biochemistry [see also Pushing Electrons].
Oxidation reactions are reactions where one or more electrons are given up. They are always coupled to reduction reactions that take up electrons. The direction of electron flow will depend on something called the "reduction potential" of the two half reactions. For our purposes we can think of the oxidation reaction as giving up "high energy" electrons that lose some of their energy before being taken up by the reduction reaction. It's this loss of energy that drives ATP synthesis.
Here's a simplified (!?) version of the electron transport pathway in the cell membranes of bacteria or mitochondria. The initial oxidation reaction is on the left below the membrane. NADH is a molecule produced during cell metabolism; notably glycolysis (degradation of glucose) and the citric acid cycle. NADH is the electron donor.
The electrons given up by NADH travel through various enzyme complexes that are embedded in the membrane (follow the red line). As the electrons give up some of their energy, the complexes pump protons from the inside of the cell to the outside. This sets up a proton gradient across the membrane and that gradient is used to drive ATP synthesis by ATP Synthase as the protons move back into the cell. This is the essence of chemiosmotic theory for which Peter Mitchell won the Nobel Prize in 1978. It's the most important pathway in all of biochemistry so every biochemist should know it intimately (they don't).
The electrons have to go somewhere eventually. The terminal electron acceptor in most modern species is molecular oxygen (O2) as shown in the figure. This is why we need oxygen to survive.
There are many modern species of bacteria that use different electron donors and electron acceptors to create the proton gradient. We're interested in those that use inorganic electron donors and something other than oxygen as an electron acceptor since those kind of bacteria are going to give us clues about the metabolism of primitive bacterial cells such as those preserved in the 3.4 billion year old rocks.
Here's a possible scheme based on what we know about modern sulfur-reducing bacteria.
The electron donor in this case is molecular hydrogen (H2). This avoids the need to use organic compounds such as NADH and glucose as a source of electrons. Hydrogen is just one of many possible electron donors but it happens to be one of the donors used by sulfur-reducing bacteria.
The electrons pass through various cytochrome complexes that are simplified versions of the ones in the first figure but they achieve the same purpose [Ubiquinone and the Proton Pump].
The electron acceptor is sulfate (SO42-). It is reduced (gaining electrons) to hydrogen sulfide (HS) which reacts with iron to produce the mineral pyrite. These bacteria do not need oxygen and they do not need an external source of energy in the form of a complex organic molecule.
We understand a lot about basic metabolism and bioenergetics and we can deduce the basic evolution of the modern complex pathways through studying unusual bacterial species that have adapted to unique environments, such as those that lack oxygen.
Wacey, D., Kilburn, M.R., Saudners, M., Cliff, J., and Brasier M.D. (2011) Microfossils of sulphur-metabolizing cells in 3.4-billion-year-old rocks of Western Australia. Nature Geoscience Published online Aug. 21, 20110 [doi:10.1038/ngeo1238]