Universal Stress Proteins in Escherichia coli

Multiple members of the UspA family of proteins are found in the genomes of bacteria, archaea, fungi, protozoa, and plants, but the biological and biochemical functions of these proteins are not known. In this issue of Journal of Bacteriology, Nachin et al. (6) describe a functional genomics approach to determining the cellular function(s) of the UspA family of proteins. A complex array of phenotypes, affecting motility and adhesion, extend the biological roles of these proteins beyond coping with stress. The work of Nachin et al. illustrates both the promise and challenges of functional genomics, even for an organism as well studied as Escherichia coli.

The UspA domain (Pfam accession number PF0582) is found in more than 1,000 different proteins (1). The domain occurs either in isolation or fused to other domains, as is seen with KdpD, the sensor kinase regulating high-affinity K⁺ transport (10). KdpD contains an UspA domain in addition to the conserved KdpD sensor domain (Pfam accession number PF02702) and the ATPase and phosphoacceptor domains characteristic of sensor kinases. Proteins that contain only one UspA domain or have two in tandem represent ca. 80% of the UspA proteins in Bacteria and Archaea and about 50% of the family members in Eucarya. The atomic structures of two UspA domains, MJ0577 from Methanococcus jannaschii (11) and UspA from Haemophilus influenzae (9), reveal asymmetric dimers with similar α/β tertiary folds. However, there are some important differences. MJ0577 crystallizes with a bound ATP, while H. influenzae UspA lacks both ATP-binding activity and conserved ATP-binding residues in the P loop. Interestingly, these structural differences correlate with groupings defined by amino acid sequence comparisons. UspA, UspC, UspD, and the N-terminal domain of UspE fall into one group, and UspF, UspG, and the C-terminal domain of UspE fall into a second group (5). UspA from H. influenzae is most similar to E. coli UspA, while MJ0577 is most similar to E. coli UspG. However, none of the E. coli UspA paralogs have the G-2X-G-9X-G-(S/T) motif that is found in MJ0577 and shared by many other members of the UspA family (9).

Previous work on regulation and analysis of mutant phenotypes implicated the UspA proteins in stress responses. UspA in E. coli K-12 was named a universal stress protein in 1992 by Nyström and Neidhardt, who showed that its synthesis is induced in response to a large number of stresses (7). The only stress that doesn't induce synthesis of UspA is cold shock (8). Synthesis of UspA and at least four of the five UspA paralogs is induced by overlapping but nonidentical sets of stresses (2, 3, 4). Synthesis of UspA (uspA), UspC (yecG), UspD (yiiT), UspE (ydaA), and UspG (ybdQ) is induced by starvation for glucose or phosphate and upon entry to stationary phase in rich medium. All five also are induced by treatment with the uncoupler dinitrophenol and by heat shock. However, the UspA proteins differ in their responses to oxidative stress and DNA damaging agents: UspA, UspC, UspD, and UspE are induced by exposure to mitomycin C, cadmium, and hydrogen peroxide, but UspG is not. Currently there is no information on the regulation of UspF (ynaF). These similarities and differences in expression pattern correlate with biological roles defined by analysis of usp mutants.

Mutants missing one of the UspA proteins often have similar phenotypes. UspA, UspC, UspD, and UspE are each needed to protect cells from DNA damage (3, 4). At least five of the UspA proteins are important for recovery from starvation in E. coli. Inactivation of uspA, uspC, uspD, uspE, or uspG causes an extended lag when stationary-phase cells are transferred to fresh medium (2, 3, 4).

These previous studies support the idea that Usp family members have partially overlapping but distinct biological functions. To study these, organisms missing combinations of Usp proteins are needed. The use of engineered and targeted knockouts in functional genomics allows one to make combinations of mutations that are difficult to obtain by traditional genetic methods. However, even with targeted knockouts, there are 63 possible combinations of single and multiple mutations for the six usp genes. Rather than constructing and testing all possible mutants, Nachin et al. were guided by groupings based on similarities in regulation and sequence analysis.

Nachin et al. demonstrate that the E. coli UspA proteins also have a variety of specialized roles in the cell. UspA and UspD, but not UspC, UspE, UspF, or UspG, protect cells against superoxide stress during exponential growth. Strikingly, after addition of phenazine methosulfate, growth of the ΔuspA mutant is indistinguishable from that of the wild-type strain for several generations, until the transition to stationary phase (see Fig. 1 in reference 6).

Several of the E. coli UspA proteins also function during steady-state growth in the absence of external stress. This was known previously for UspA, whose absence alters carbon utilization (8). The work of Nachin et al. shows that intracellular iron levels appear to be higher in the ΔuspD mutant than in wild-type cells, which could account for the superoxide sensitivity of this mutant. Nachin et al. noticed serendipitously that UspA proteins affect cell surface properties, noting that ΔuspE mutants did not settle to the bottom of the tube when cultures were left standing on the bench. Deletion of uspC, uspE, or uspG reduces autoaggregation, while mutants lacking UspF form larger clumps than wild-type cells. This observation led to the investigation of adhesion and motility phenotypes. Two pairs of UspA proteins (UspC-UspE and UspF-UspG) affect adhesion and motility, but in opposing directions. UspC and UspE each promote flagellum synthesis and motility and decrease adhesion by type 1 fimbriae, while UspF and UspG each increase adhesion and reduce motility. Surprisingly, deleting both uspF and uspG restores the wild-type phenotypes.

Despite the progress that has been made in studying the UspA family of proteins, their biochemical activities and the mechanisms by which they function remain elusive. This family of proteins clearly warrants further study. The work described by Nachin et al. points to new avenues of investigation that will help to solve this puzzle.