Grasping at Shadows: Revealing the Elusive Nature of Essential Genes

The conundrum of studying a gene that is essential for bacterial viability poses a nontrivial genetic challenge that has been met with mixed success using multiple strategies. Most of the time, cursory analyses merely report the inability to inactivate a gene of interest, hence suggesting it is vital for the cell. The identification of essential genes has long been a “holy grail” of bacterial genetics (16) and is receiving increasing attention in this age of “omics,” as these genes are often considered to be potential targets for novel antimicrobial compounds in pathogenic bacteria (14, 24). Several genome-wide strategies have thus been used to identify essential genes, ranging from directed approaches using comparative genomics with minimal-sized genomes (2) or massive undertakings such as systematic genome-wide gene inactivation (3, 22) to global methods including genetic footprinting through high-density transposon mutagenesis (13, 18) and antisense RNA expression libraries (20). Yet these reports rarely extend beyond the primary observation that certain sets of genes are essential, and none of these methods is entirely without pitfalls. As the search for relevant phenotypes for these genes is de facto hampered by their essential nature, more-detailed studies often involve the use of inducible promoters allowing programmed depletion of the corresponding protein, plasmid-based overexpression of the gene, or the generation of conditional mutants such as thermosensitive alleles. In this issue, an elegantly designed in-depth study by Handford et al. combines several of these methods and many more in a multidisciplinary approach which has allowed them to convincingly characterize not one but three essential genes in Escherichia coli, yjeE, yeaZ, and ygjD, which they show as belonging to the same pathway, apparently playing a role in linking DNA metabolism with cell division (15).

When is a gene essential?

Microbiologists, through the pressures of the ever-increasing pace of research and the timetable imposed by their circadian social lifestyle, have in turn systematically subjected bacteria to life in the fast lane. Indeed, most bacterial studies involve growth in idealized laboratory conditions far removed from those of their natural habitat, with correspondingly artificially reduced generation times. The impact of such an accelerated life cycle is currently drawing increased attention, with more and more research devoted to comparisons of laboratory strain responses with those of clinical or natural isolates, emphasizing the genetic variations now observed in so-called “domesticated” laboratory strains (1, 4, 12, 29). Clearly, the conditions under which the essential nature of a gene is to be examined are critical, since biosynthetic pathway genes, for example, will often prove essential during growth in minimal media or within the host. Nevertheless, most studies to date have focused instead on genes that are required for survival in rich media.

In a pioneering systematic gene inactivation approach carried out in Bacillus subtilis, a minimal set of 271 genes were characterized as essential during optimal growth conditions (22), whereas 303 were later reported as such in Escherichia coli (3). The B. subtilis study was completed and extended by Jeff Errington and colleagues, first by showing that of the 271, 4 in fact were not essential, and then, through protein depletion and localization, by assigning potential functions to several of these, in RNA processing and lipid and cell wall synthesis (17). In the case of coupled gene pairs encoding toxin/antitoxin gene modules, inactivation of the antitoxin gene alone is often lethal, whereas inactivating both genes rarely leads to a given phenotype (28). Conversely, functional redundancy or gene duplication can seriously impede efforts to identify essential pathways. This aspect was also addressed by Jeff Errington and colleagues, in a study targeted at paralogous gene pairs, allowing them to identify six such essential pairs in B. subtilis (27). In the rare examples of essential regulatory genes, identifying the operator binding site for the DNA-binding regulator can allow its function to be deduced through analysis of regulon members identified by genome scanning, as shown for the Staphylococcus aureus WalK/WalR (née YycG/YycF) two-component system, which controls cell wall homeostasis and is highly conserved in low-G+C-content gram-positive bacteria (8-11).

When serendipity strikes.

When Tracy Palmer's team began their study by embarking upon this road less traveled, little did they know it would soon prove to be long and winding, rather than paved with yellow bricks. In fact, it would take several years before their ambitious undertaking hit pay dirt, with a kitchen-sink molecular approach that remains an impressive testimony to their will and determination. Following up on an initial observation that mutations inactivating the E. coli Tat secretion pathway led to defects in cell division (26), they began their arduous journey in earnest when they isolated transposon insertions reversing this cell filamentation phenotype (19) that were found to lie directly upstream from the yjeE gene. They later realized that the cell filamentation reversal was in fact not due to yjeE but to upregulation of the neighboring amiB gene encoding an amidase, yet once their interest had been piqued by yjeE, there was to be no turning back.

Conservation of gene order in a genomic context often suggests that the corresponding proteins interact (7). In B. subtilis, the yjeE counterpart (ydiB) is located within an operon that also includes orthologs of the E. coli yeaZ (ydiC) and ygjD (ydiE) genes (23), although these genes are not colocalized in E. coli. In this issue, Handford et al. confirm earlier reports that all three of these genes are indeed essential for E. coli's viability, which they accomplished by placing each of them under the control of an arabinose-inducible promoter and then going on to examine in greater detail phenotypes linked to depletion of the corresponding proteins (15). YjeE, a small, highly conserved eubacterial protein, is essential in some bacteria but not in others and is thought to bind ATP and act as an ATP hydrolysis-driven molecular switch. As shown in the work of Handford et al. (15) and in Fig. 1, YjeE depletion leads to dramatic changes in cell morphology, with increases in size, length, and width, as well as distortions and a curved aspect (Fig. 1B). They also observed an unusual peripheral localization of the nucleoid within the cell, suggesting a defect in DNA synthesis (15).

FIG. 1.

Fluorescence microscopy of Escherichia coli MC4100-A derivative JH17 that allows arabinose-dependent expression of yjeE (15), showing the effects of YjeE depletion on cell morphology (courtesy of J. I. Handford, A. Hunt, J. Errington, and T. Palmer). Cell membranes were stained with FM4-64. (A) Strain JH17 grown in LB containing 0.2% arabinose, expressing yjeE. (B) Strain JH17 grown in LB containing 0.4% glucose-6-phosphate, depleted of YjeE.

YgjD was identified by comparative genomics more than 10 years ago as being required for survival during growth in complex media in both E. coli and B. subtilis (2), and as it shares similarities with O-sialoglycoprotein endopeptidases, it is often referred to as Gcp. YeaZ shares some similarities with YgjD, particularly within the amino-terminal domain, and belongs to the HSP70 actin-like-fold protein family. Using the same protein depletion approach as for YjeE, Handford et al. show that cells lacking YeaZ were enlarged and misshapen but, most of all, had highly condensed nucleoids, whereas those without YgjD had a more-heterogeneous appearance, with about two-thirds displaying phenotypes similar to those of strains lacking YjeE (15).

Guilt by association.

An earlier report indicated that YgjD is copurified with YeaZ, indicating that the two proteins interact in vivo (5). One of the most successful and powerful techniques recently developed for studying protein-protein interactions is the bacterial two-hybrid system (21). Using this approach, Handford et al. showed that YjeE also interacts with YeaZ but not with YgjD and confirmed these interactions by two-hybrid screening of several random genomic libraries (15). Their results suggest that YeaZ switches partners between YjeE and YgjD, with a preference for the latter. In an effort to identify genes that could compensate for the lethal phenotype caused by starving cells for any of these three proteins, two complementary approaches were used. One, unsuccessful, involved a search for transposon insertions that restored growth to depleted cells (T. Palmer, personal communication). The other, described in this issue, involved screening plasmid-based genomic libraries to identify genes that would allow cells to survive when depleted of YeaZ, YjeE, or YgjD. They were thus able not only to confirm an earlier report that overexpression of the rstA gene, encoding the response regulator of the RstB/RstA two-component system, restored growth to cells depleted of YjeE (6) but to show that the same held true for cells lacking either YeaZ or YgjD (15). In an in vitro approach, they showed that the three purified proteins formed dimers in solution, that YgjD could be copurified with YeaZ, and that YeaZ mediates proteolysis of YgjD (15). This latter activity was discovered quite serendipitously by J. I. Handford and B. Ize as they tested whether YeaZ or YgjD or a mixture of the two, given their similarities to sialoglycoprotein endopeptidases, could cleave glycophorin A (which they cannot) and noticed that YgjD disappeared whenever YeaZ was present.

Significance of this work.

The yjeE, yeaZ, and ygjD genes are highly conserved in eubacteria, and although the genes and their essential nature have been known for some time in E. coli, this is the first thorough study to link these three essential genes together in the same cellular pathway. Furthermore, as shown in Fig. 2 and in this issue, Handford and colleagues' model suggests that within this complex essential network, both YjeE and YeaZ interact to regulate the activity of YgjD, likely via proteolysis, and thus, the model provides an important framework for investigating the function of these genes in other bacteria in an effort to fully understand their role. Indeed, they propose that although preferentially associated with YgjD, YeaZ engages in partner switching in response to an undefined signal, binding with YjeE instead, allowing free YgjD to accumulate within the cell (Fig. 2). ATP binding and hydrolysis by YjeE is thought to reverse the partner switch, allowing YeaZ to form a complex with YgjD which is then subjected to proteolysis (Fig. 2). As it remains to be determined whether YgjD is in fact activated or inactivated through YeaZ-dependent proteolysis, both possibilities are presented in Fig. 2. Likewise, YgjD may be involved either in the signaling pathway which leads to accumulation of the active phosphorylated form of RstA or in controlling some common cellular function (Fig. 2), although the RstB/RstA two-component system is not essential. Interestingly, the expression of the rstAB operon is itself controlled by another two-component system, PhoP/PhoQ, known to respond to magnesium ion concentrations (25), and it will be important to determine the hierarchical relationships of these two signal transduction pathways with the essential YjeE/YeaZ/YgjD protein network.

FIG. 2.

Model of the YjeE/YeaZ/YgjD essential signal transduction network. Although preferentially associated with YgjD, YeaZ engages in partner switching in response to an undefined signal, binding with YjeE instead, and free YgjD can then accumulate within the cell. ATP binding or hydrolysis by YjeE is thought to reverse the partner switch, allowing YeaZ to form a complex with YgjD, which is then subjected to proteolysis. The RstB histidine kinase phosphorylates its cognate response regulator RstA, thus activating the expression of a specific set of genes whose identity is still unknown. It remains to be determined whether YgjD is in fact activated through proteolysis or simply degraded, where it acts in the cell or within the RstB/RstA pathway to control essential functions, and how RstA overexpression is able to compensate for the absence of YjeE, YeaZ, or YgjD.

Future questions.

Somewhat frustratingly for all, despite the considerable amount of work and the beginnings of an inkling as to the pathway in which YjeE, YeaZ, and YgjD are involved, their true cellular function still remains elusive. Morphological changes linked to depletion of YjeE and YgjD suggest a role in cell envelope metabolism, and the unusual nucleoid distribution implies an effect on DNA replication. Many questions still remain, such as determining the role of the RstA response regulator and why its overproduction can compensate for the lack of these three essential proteins. These questions will likely be addressed both by defining the RstA regulon and determining what effects on global gene expression are caused by depleting cells of YjeE, YeaZ, and YgjD. Additionally, protein subcellular colocalization experiments using translational fusions of RstB, RstA, YjeE, YeaZ, and YgjD with derivatives of green fluorescent protein are also likely to provide important information on how and where the components of this essential signal transduction network interact within the cell and may help in determining the signals they respond to, as well as their true function.