Tuberculosis (2008) 88 Suppl. 1, S19-S24 The many roads to essential genes Jun-Rong Wei, Eric J. Rubin* Department of Immunology and Infectious Diseases, Harvard School of Public Health, 200 Longwood Ave, Boston, MA 02115, USA KEYWORDS Summary Mycobacterium Antibiotics target functions that are required for bacterial growth and survival. As tuberculosis; genetic tools for studying Mycobacterium tuberculosis continue to improve we are Essential genes increasingly able to identify genes that encode these important effectors. Here we review the strategies that have been used to identify and validate essential genes in mycobacteria and look forward to possible future advances. © 2008 Elsevier Ltd. All rights reserved Introduction conditions. For example, in the case of Mycobacterium tuberculosis, several genes must be intact so that While there are many approaches to developing new bacteria can survive during infection in a mouse but not antibacterial drugs, all share a common property. Drugs for growth in defined medium.1 Others are required for must target processes that are critical for bacterial growth under all assayed conditions, a group of genes growth or survival during infection. Thus, the genes that that is far more technically difficult to study. Here we encode these functions, so-called essential genes, will review the strategies that have been employed to represent the complement of possible antibiotic targets identify and validate such essential genes in myco- and identifying these genes and their encoded pathways bacteria and indicate their advantages and disadvan- can help lead drug discovery efforts. In the case of tages. These methods produce tradeoffs. In general, tuberculosis (TB), a disease for which current treatment methods that analyze single genes have very low is suboptimal, starting with target identification could throughput but have far greater specificity. Screening lead to drugs with different mechanisms of action and, methods can identify large numbers of candidates but possibly, improved efficacy. are far more likely to make incorrect predictions. What is meant by “essential genes?” This term suggests genes that are absolutely required for bacterial survival. In fact, most of the studies described rely on Choosing candidate essential genes bacterial growth to assay for essentiality. Thus, they actually define genes required for growth, not survival. Most of the approaches described below are designed to Many of these approaches cannot distinguish between study targeted genes. Several strategies for selecting bacteria that grow very slowly and those that either die which genes to target have been employed. Principal or cannot grow at all. Therefore, the function of among these is an understanding of bacterial physiology identified genes can vary considerably depending on the and structure. For example, several genes are known to method used for their identification. Of course, gene play important parts in cell division while others are products are often only required under specific critical for obtaining and metabolizing nutrients. On the other hand, gene products might be attractive drug targets because of what is known about them in other *Corresponding author: Tel: +1 617 432 3335; systems. For example, protein kinases are particularly Fax: +1 617 432 3354 druggable targets so it is valuable to understand their E-mail address: [email protected] (E.J. Rubin) role in bacterial growth. Finally, screens, described below, 1472-9792/$ – see front matter © 2008 Elsevier Ltd. All rights reserved. S20 J.-R. Wei, E.J. Rubin have identified a number of potential drug targets. Hasan Using site-specific recombination system et al.2 have collated these results and assigned weighted scores for the potential of gene products to serve as One of the major limitations of using homologous targets for antibiotic development. recombination is the low rate at which it occurs. One way to overcome this is to use a higher frequency, site- specific recombination system. To do this, Pashley and Attempted knockouts Parish took advantage of the mycobacteriophage L5 excisionase16 (Fig. 1B). This enzyme efficiently removes The first evidence for gene essentiality generally arises DNA integrated into the ΦL5 attB site. These from negative data — the inability to knockout a researchers constructed a merodiploid strain in which targeted gene. Generating knockout mutants is usually one copy of the gene was integrated at the L5 attB site the first step in determining the function of a gene. then deleted the native copy of the gene. Introduction Several methods have been employed to produce such of another vector carrying the L5 excisionase gene will knockouts, including the use of non-replicating vectors,3 remove the integrated copy. For non-essential genes, long linear DNA fragments4 and incompatible plasmids.5 transformation efficiency of excisionase carrying Gene replacement can be performed using plasmids that plasmid should be the same for all 3 strains. On the require two recombination steps6 or as a single step other hand, if the target gene is essential, no transfor- using mycobacteriophage-mediated delivery.7 Unfortu- mants will be obtained in strain B. Another modification nately, the rate of illegitimate recombination is often uses the L5 integrase10,17,18 (Fig. 1C). This method employs high, particularly when attempting to introduce an un- the same recombination constructs; however, instead of favorable mutation. The introduction of mycobacterio- the excisionase, another vector carrying L5 integrase is phage proteins Che9c gp60 and gp61 — the homologs of used to promote high frequency “switching” of intro- RecE and RecT, respectively — can both increase the duced genes at the attB site. The rate of switching can rate of true homologous recombination allowing the be easily monitored using a different antibiotic marker. insertion of single nucleotide changes.8,9 A failure to If the targeted gene is essential, no switching will occur obtain appropriate recombinants is often initially attri- in strain B. buted to technique and later to biology. But how can a These approaches have two advantages. The rela- failed experiment be used to determine the probability tively high recombination rates make ambiguous results of a gene actually being required for optimal growth? less likely. In addition, it is quite simple to integrate The simplest method is to use homologous recom- variants of the complementing gene (e.g., alleles from bination rates to estimate the probability of gene other species and mutant alleles) and test for comple- essentiality.6,10–13 This is performed as shown in figure mentation. Once again, however, results can be influ- 1A. In this strategy, homologous recombination is used enced by local effects at the site of the deletion of the to insert a vector sequence into the appropriate native chromosomal gene. In addition, this method chromosomal location by a single crossover event. A cannot be used to study the behavior of these genes second crossover event (resolution) can either restore during infection. the original structure of the chromosome or result in a deletion of the targeted gene. In the absence of selection, either second crossover event should occur Antisense oligonucleotides at equal frequency. However, if the targeted gene is essential, the only allowable crossover is the one that Inhibiting gene expression using antisense has been a very preserves the gene. If there are a sufficiently large successful strategy for identifying important genes in number of resolution reactions that fail to remove the other pathogens such as Staphylococcus aureus.19 Since gene, it is very likely that the gene is essential. This antisense RNA can be supplied in trans, this approach is can be further controlled by adding a second copy of very attractive as it does not require that site-directed the targeted gene, either on an episome or integrated mutations be introduced into the chromosome. at a remote site, a process that can be facilitated with In fact, the Horwitz lab has taken this one step the use of vectors that are temperature sensitive for further.20–22 As shown in a series of publications, they are replication.14 If the resolution reactions occur only in able to partially silence genes using exogenous oligo- the presence of the functional second copy, this nucleotides. These phosphorothioate-modified nucleic strongly suggests gene essentiality and crossover acids are stable and can be taken up into target cells frequency can be used to estimate the probability that resulting in decreased synthesis of several targeted a gene is truly essential.15 proteins. Although individual oligonucleotides had only While simple, this approach is often not persuasive. limited effects, mixtures of different targeting anti- The readout can be biased by the specific attributes of sense molecules could result in significant inhibition. the constructs that are employed. For example, inserts They have used this method to show the potential of might create toxicity independent of the effect on the various proteins as targets for antibiotic development. targeted gene (e.g., by producing transcriptional Despite the ease of this approach, it has not seen effects on adjacent regions of the chromosome). Of widespread application. Perhaps this is because it is not course, failure to disrupt a gene might be evidence of yet clear what rules govern successful design of these essentiality but does not provide any evidence for