Much of the facultative gene expression system, especially including the virulon, is temporally organized so that the component genes must contain regulatory sequences that are activated combinatorially in a time-dependent manner by different incoming signals acting through intracellular response elements. As suggested by the model in Fig. 6A, the entire accessory gene regulatory network must also be coupled to the overall energy metabolism of the cell, and it has been suggested that there must be a key coupling parameter, such as the levels of nucleotide polyphosphates or of other energy-transducing cofactors such as NADH2 (R. Proctor, personal communication). Key enzymes of intermediary metabolism, such as aconitase (Somerville et al., 2002a), could be involved in this coupling, possibly acting through one or more TCS, such as srr (srh). On the basis of results obtained with in vitro cultures, it appears that the surface proteins are probably required earlier in the course of an infection than the secreted enzymes, immunotoxins and cytotoxins and the above-mentioned intracellular metabolic enzymes. This sequential activation seems to be, at least in part, a function of population density. Starting with stationary phase, there would seem to be three key transition points in the in vitro growth cycle, possibly occurring in response to intracellular signals, such as GTP levels. First is the transition to exponential phase, which involves not only the revival of biosynthetic and other metabolic pathways required for growth and cell division, but also the synthesis of some surface proteins, coagulase and possibly other accessory proteins. The synthesis of these is probably initiated during the transition from stationary phase to exponential phase and may come under the general metabolic programme governing this transition. The nature of the signals (?) acting at this stage represents a key area for study. Other surface protein genes are switched on shortly after the onset of exponential growth and, as typified by spa, are switched off shortly thereafter, concomitantly with the appearance of agr-RNA III (Vandenesch et al., 1991). This clear reciprocity, however, is not seen with all strains and under all conditions (Tegmark et al., 2000), and may be related to σB activity (S. Herbert and R. P. Novick, unpublished data) or to growth conditions and media. The agr AIP reaches its threshold around mid-exponential phase, activating agr expression. In 8325 derivatives, however, certain exoprotein genes, such as coa, are sharply downregulated well before the appearance of RNA III, suggesting that some other inhibitory signal is responsible. The second transition, between the exponential and post-exponential phases (possibly a consequence of decreasing availability of oxygen owing to increasing population density), is, in most strains, accompanied by upregulation of the genes encoding secreted proteins. Agr, which, in 8325, is activated two or more hours earlier, sets the level of expression of most of these proteins, but not the timing (Vandenesch et al., 1991); in fact, upregulation of these genes occurs at the onset of the post-exponential phase, regardless of when, or even whether, RNA III transcription is activated (unpublished data). This is consistent with the results of temporal activation studies, in which activation of hla transcription may occur as much as 6 h after RNA III (cloned to the β-lactamase promoter and induced) reaches its maximum level (Vandenesch et al., 1991). In a sarA mutant, RNA III transcription is delayed by an hour and is closely co-ordinated with the onset of hla transcription. This effect can be attributed to SarS (also known as SarH1) because, in the double SarA/SarS mutant, RNA III is not delayed and there is again a timing differential (Tegmark et al., 2000). In some strains, such as those of agr group IV, hla and other exoprotein genes are upregulated earlier, concomitantly with RNA III synthesis (Jarraud et al., 2000), suggesting that the exponential to post-exponential phase transition may not be a critical regulatory point for hla and other exoprotein genes in these strains. A further complication is the apparent post-exponential upregulation of DNA gyrase by agr (≈ six-fold; Dunman et al., 2001), raising the possibility that agr regulation could involve changes in superhelix density, which are well known to occur during post-exponential growth and are well known to affect a variety of promoters (although there are very little data on this in staphylococci).