How can (or did) Deinococcus radiodurans continue to evolve after developing resistance to mutation?

How can (or did) Deinococcus radiodurans continue to evolve after developing resistance to mutation?

We are searching data for your request:

Forums and discussions:
Manuals and reference books:
Data from registers:
Wait the end of the search in all databases.
Upon completion, a link will appear to access the found materials.

Deinococcus radiodurans has a remarkable ability to resist damage to its DNA due to radiation, dehydration or (to my knowledge) any other source. It keeps multiple copies of its genome and has a repair mechanism as well.

My question is: how could this thing evolve to be so sophisticated at preventing the processes behind evolution? It seems that a mechanism to prevent errors would also prevent changes, so that as soon as the organism got OK at preventing changes it wouldn't be able to evolve to get better at it. Were genes for these processes borrowed from other extremophiles? Is the assumption that it was in such extreme environments that the damaging influences were keeping up with the repair mechanisms? That's my guess, but I'd love to hear from others with more knowledge of evolutionary processes.

According to the "Functional adaptation" section of this webpage, Deinococcus radiodurans uses repair proteins to stitch up DNA after radiation damage. Meaning, after a burst of radiation, the DNA will be in pieces and the repair proteins put it back together. It seems like this mechanism would not come into play during other forms of DNA mutation that happen during replication, such as point mutations or inversions.

Deinococcus radiodurans did not "develop resistance to mutations".

It is able to repair its chromosome when scatered in pieces by radiations or desiccation, while other bacteria would die in such conditions.

So this is adaptive in extreme environments, such as deserts (where it has evolved) or canned corned beef (where it was discovered).

Radiation-resistant Organism Reveals Its Defense Strategies The Secret To Its Strength Is A Ring, Weizmann Institute Researchers Report In Science

Rehovot, Israel &mdash January 9, 2003 &mdashWeizmann Institute scientists have found what makes the bacterium Deinococcus radiodurans the most radiation-resistant organism in the world: The microbe's DNA is packed tightly into a ring. The findings, published in the January 10 issue of Science, solve a mystery that has long engaged the scientific community.

The red bacterium can withstand 1.5 million rads &ndash a thousand times more than any other life form on Earth and three thousand that of humans. Its healthy appetite has made it a reliable worker at nuclear waste sites, where it eats up nuclear waste and transforms it into more disposable derivatives. The ability to withstand other extreme stresses, such as dehydration and low temperatures, makes the microbe one of the few life forms found on the North Pole. It is not surprising, then, that it has been the source of much curiosity worldwide, recently leading to a debate between NASA and Russian scientists &ndash the latter saying that it originated on Mars, where radiation levels are higher.

Since DNA is the first part of a cell to be damaged by radiation and the most lethal damage is the breakage of both DNA strands, scientists have focused on DNA repair mechanisms to find the answer to the microbe's resilience. Cells, including human cells, can mend only very few such breaks in their DNA. Microbes, for example, can repair only three to five. Yet D. radiodurans can fix more than 200. Thus scientists believed that the microbe must possess uniquely effective enzymes that repair DNA. However, a series of experiments showed that the microbe's repair enzymes were very similar to those existing in ordinary bacteria.

Using an assortment of optical and electron microscopy methods, Prof. Avi Minsky of the Weizmann Institute of Science&rsquos Organic Chemistry Department found that the microbe's DNA is organized in a unique ring that prevents pieces of DNA broken by radiation from floating off into the cell's liquids. Unlike other organisms, in which DNA fragments are lost due to radiation, this microbe does not lose genetic information because it keeps the severed DNA fragments tightly locked in the ring &ndash by the hundreds, if necessary. The fragments, held close, eventually come back together in the correct, original order, reconstructing the DNA strands.

As exciting as these findings may be, they are not expected to lead to the protection of humans from radiation. &ldquoOur DNA is structured in a fundamentally different manner,&rdquo says Minsky. The results may, however, lead to a better understanding of DNA protection in sperm cells, where a ring-like DNA structure has also been observed.

Minsky's team also found that the microbe undergoes two phases of DNA repair. During the first phase the DNA fixes itself within the ring as described. It then performs an even more unusual stunt.

The bacterium is composed of four compartments, each containing one copy of DNA. Minsky's group found two small passages between the compartments. After about an hour and a half of repair within the ring, the DNA unfolds and migrates to an adjacent compartment &ndash where it mingles with the copy of DNA residing there. Then the &ldquoregular&rdquo repair machinery, common in humans and bacteria alike, comes into play &ndash repair enzymes compare between the two copies of DNA, using each as a template to fix the other. Since the DNA has already been through one phase of repair in which many of the breaks are fixed, this phase can be completed relatively easily.

The finding of a tightly packed ring made the team wonder how the bacterium could live and function under normal conditions. DNA strands must unfurl to perform their job of protein production. How can they do that if they can barely budge? This question led to the uncovering of another of the microbe's survival strategies: out of the four copies of DNA, there are always two or three tightly packed in a ring while the other copies are free to move about. Thus at any given moment there are copies of DNA that drive the production of proteins and others that are inactive but continuously protected.

Minsky, along with other scientists, believes that the bacterium&rsquos answer to acute stresses evolved on Earth as a response to the harsh environments from which it might have emerged. It is one of the few life forms found in extremely dry areas. The unique defense mechanism that evolved to help it combat dehydration proves useful in protecting it from radiation.

Deinococcus radiodurans was discovered decades ago in canned food that was sterilized using radiation. Red patches appeared in the cans &ndash colonies of the bacterium &ndash setting off questions as to how it could have survived. Though these questions have now been answered, the tide of speculation as to how these defense mechanisms evolved &ndash and where &ndash is likely to continue.

Prof. Abraham Minsky's research is supported by Verband der Chemischen Industrie, Teva Pharmaceuticals, Israel, and the Helen & Milton A. Kimmelman Center for Biomolecular Structure & Assembly.

Prof. Minsky is the incumbent of the Professor T. Reichstein Professorial Chair.

The Weizmann Institute of Science in Rehovot, Israel, is one of the world's top-ranking multidisciplinary research institutions. Noted for its wide-ranging exploration of the natural and exact sciences, the Institute is home to 2,500 scientists, students, technicians and supporting staff. Institute research efforts include the search for new ways of fighting disease and hunger, examining leading questions in mathematics and computer science, probing the physics of matter and the universe, creating novel materials and developing new strategies for protecting the environment.

Story Source:

Materials provided by Weizmann Institute. Note: Content may be edited for style and length.

Mismatch repair ensures fidelity of replication and recombination in the radioresistant organism Deinococcus radiodurans

We have characterized the mismatch repair system (MMR) of the highly radiation-resistant type strain of Deinococcus radiodurans, ATCC 13939. We show that the MMR system is functional in this organism, where it participates in ensuring the fidelity of DNA replication and recombination. The system relies on the activity of two key proteins, MutS1 and MutL, which constitute a conserved core involved in mismatch recognition. Inactivation of MutS1 or MutL resulted in a seven-fold increase in the frequency of spontaneous Rif R mutagenesis and a ten-fold increase in the efficiency of integration of a donor point-mutation marker during bacterial transformation. Inactivation of the mismatch repair-associated UvrD helicase increased the level of spontaneous mutagenesis, but had no effect on marker integration—suggesting that binding of MutS1 and MutL proteins to a mismatched heteroduplex suffices to inhibit recombination between non identical (homeologous) DNAs. In contrast, inactivation of MutS2, encoded by the second mutS -related gene present in D. radiodurans, had no effect on mutagenesis or recombination. Cells devoid of MutS1 or MutL proteins were as resistant to γ-rays, mitomycin C and UV-irradiation as wild-type bacteria, suggesting that the mismatch repair system is not essential for the reconstitution of a functional genome after DNA damage.

This is a preview of subscription content, access via your institution.

World's toughest bacterium holds promise for rapid vaccine development against deadly diseases

Scientists from the Uniformed Services University of the Health Sciences (USU) have developed a new preparation method that renders a virus or bacterium non-infectious while preserving its immune-boosting ability after exposure to gamma radiation. A vaccine exposed to megadoses of gamma radiation was successfully tested in mice against drug-resistant Staphylococcus aureus bacteria by colleagues at the National Institutes of Health (NIH), and holds promise for other such deadly diseases.

The results of the breakthrough study were published in the July edition of Cell Host and Microbe.

High doses of radiation typically destroy a pathogen's genome, rendering it unable to cause infection when used in a vaccine. However, radiation also damages a microbe's protein epitopes, which the immune system must recognize for a vaccine to be protective. Organisms inactivated, or killed, by radiation trigger better immune responses than those inactivated by traditional heat or chemical methods. Although live vaccines may provide better immune protection than irradiated vaccines, live vaccines are frequently not an option as they can carry an unacceptable risk of infection with an otherwise untreatable disease (e.g., HIV). Lethally irradiated vaccines could also help the developing world, where the need for cold storage limits the availability of live vaccines.

To separate genome destruction from epitope survival, the researchers borrowed some complex chemistry from the world's toughest bacterium Deinococcus radiodurans, nicknamed "Conan the Bacterium," which can withstand 3,000 times the radiation levels that would kill a human being. In 2000, Deinococcus was engineered for cleanup of highly radioactive wastes left over from the production of atomic bombs. Now, unusual Mn(II)-antioxidants discovered in this extremophile have been successfully applied to preparing irradiated vaccines.

Deinococcus accumulates high concentrations of manganese and peptides, which the scientists combined in the laboratory -- forming a potent antioxidant complex which specifically protects proteins from radiation. They found that the complex preserves immune-related epitopes when applied to viruses and bacteria during exposure to gamma radiation, but did not protect their genomes.

Michael J. Daly, Ph.D., professor of Pathology at USU, and his research team, collaborated on the work with Sandip K. Datta, M.D., and colleagues at NIH's National Institute of Allergy and Infectious Diseases (NIAID). Daly devoted 20 years to studying Deinococcus radiodurans, which has led to three patents for his work.

The scientists used the Mn-peptide complex in a laboratory setting to successfully protect from radiation damage the protein epitopes of Venezuelan equine encephalitis virus, a microbe that causes a mosquito- borne disease of the nervous system. They also used the preparation method to develop an effective vaccine against methicillin-resistant S. aureus (MRSA) infections in mice.

Learning to Care for Those in Harm's Way

The researchers believe the whole-microbe vaccine approach could extend to any infectious organism that can be cultivated, whether fungi, parasites, protozoa, viruses or bacteria -- including agents that mutate rapidly, such as pandemic influenza and HIV. The groups aim to demonstrate this method of irradiation as a rapid, cost-effective approach to vaccine development.

The project was funded by the Air Force Office of Scientific Research (AFOSR) and the intramural research program of the NIAID.

Community Ecology of Deinococcus in Irradiated Soil

Deinococcus is a genus of soil bacteria known for radiation resistance. However, the effects of radiation exposure on its community structure are unknown. We exposed soil to three levels of gamma radiation, 0.1 kGy/h (low), 1 kGy/h (medium), and 3 kGy/h (high), once a week for 6 weeks and then extracted soil DNA for 16S rRNA amplicon sequencing. We found the following: (1) Increasing radiation dose produced a major increase in relative abundance of Deinococcus, reaching

80% of reads at the highest doses. Differing abundances of the various Deinococcus species in relation to exposure levels indicate distinct “radiation niches.” At 3 kGy/h, a single OTU identified as D. ficus overwhelmingly dominated the mesocosms. (2) Corresponding published genome data show that the dominant species at 3 kGy/h, D. ficus, has a larger and more complex genome than other Deinococcus species with a greater proportion of genes related to DNA and nucleotide metabolism, cell wall, membrane, and envelope biogenesis as well as more cell cycle control, cell division, and chromosome partitioning-related genes. Deinococcus ficus also has a higher guanine–cytosine ratio than most other Deinococcus. These features may be linked to genome stability and may explain its greater abundance in this apparently competitive system, under high-radiation exposures. (3) Genomic analysis suggests that Deinococcus, including D. ficus, are capable of utilizing diverse carbon sources derived from both microbial cells killed by the radiation (including C5–C12-containing compounds, like arabinose, lactose, N-acetyl- d -glucosamine) and plant-derived organic matter in the soil (e.g., cellulose and hemicellulose). (4) Overall, based on its metagenome, even the most highly irradiated (3 kGy/h) soil possesses a wide range of the activities necessary for a functional soil system. Future studies may consider the resilience and sustainability of such soils in a high-radiation environment.

This is a preview of subscription content, access via your institution.

Materials and Methods

Ethical Statement

The research complied with requirements for research on birds in Ukraine and Denmark and permission was given by the administration of the Chernobyl Exclusion Zone. All sampling was approved in an ethical review by the University of South Carolina Institutional Animal Care and Use Committee the methods were carried out in accordance with the approved guidelines. All birds were handled briefly and none died or showed signs of suffering during the short sampling period. All individuals flew upon release. The field studies did not involve endangered or protected species.

Isolation of bacterial strains

In 2007 and 2008 wearing sterile gloves we collected eight feathers from the body of barn swallows randomly captured as part of an ongoing long-term project on birds breeding in Ukraine and Belarus 42 . Collections were made at collective farms in areas with high levels of radiation outside the southern exclusion zone near Chernobyl and following the procedure described by Czirják et al. 30 . All feathers were placed in a sealed plastic bag and stored at -20 °C in the dark until microbiological analysis. Host sex ratios were balanced (N2007 = 31 Females, 29 Males N2008 = 58 Females, 56 Males). All farms presented similar climate and habitat types, surrounded by plantations, scattered trees and open farmland, but were exposed to different environmental radiation levels 44 . The α, β and γ radiation at ground level have previously been measured at each farm and cross-validated with measurements by the Ukrainian Ministry of Emergencies 45 . The two series of measurements were strongly positively correlated 30,44 . We chose bird samples from places with four different background radiation intensities: high, 2.9 μGy/h (Vesniane Farm, Ukraine) intermediate, 0.45 μGy/h (Farm 49, Vetka, Belarus) low, 0.1 μGy/h (Farm 43, Vetka, Belaryus) and, control, 0.03 - 0.05 μGy/h (Kraghede, Denmark).

In order to obtain both free-living and attached microorganisms, five feathers were sonicated at high frequency for 15 minutes in 3 repeats (5 minutes each with a 5-minute pause between repeats) in 0.8 ml of sterile physiological (0.90% w/v) saline solution for bacterial declumping. This process does not affect bacterial viability 46 . After sonication samples were vortexed for 20 seconds and bacterial suspensions were transferred to a sterile 1.5 ml eppendorf tube. Then, feathers were re-suspended in 0.5 ml sterile physiological saline and vortexed again for 20 seconds. The supernatant was transferred to a second sterile eppendorf, obtaining an overall volume of

1.3 ml solution. The method is fully described by Czirják et al. 30 . All samples were treated in the same way and thus, any potential bias would have affected bacterial communities equally.

In duplicate, we spread 100 μl each of the microbial solutions on Tryptic Soy Agar (TSA, #22091, Fluka), a rich growth medium and the plates were incubated at 25 °C, for 3 days. To inhibit fungal growth we added 100 mg mL −1 of cycloheximide (#01810, Fluka) to the medium. Then, we selected all bacterial colonies showing morphological differences based on colour, shape and size, for all the sixteen plates per site and we inoculated tubes containing 15 mL of Tryptic Soy Broth (TSB, #22092, Fluka) medium with single colonies until cultures reached an optical density of 0.6 at a wavelength of 600 nm. The tubes were centrifuged and the pellets re-suspended in 10 mL of TSB.


Bacterial morphotypes isolated as above were randomly allocated across the wells of a 96-well microtiter flat plate (84 isolates and three replicates of an Escherichia coli K-12 W3110 laboratory strain as control). We used 15 μl of each culture to inoculate 135 μl of TSB per well. The same inoculation order was kept to culture 16 microtiter plates that were allocated to radiation and no radiation treatments with a replicate each. Thus, each bacterial isolate had two replicates and the control E. coli six replicates for each treatment. Three samples did not grow well in TSB medium during the experiment and they were removed from the experiments (two from high and one from low background radiation respectively).

Samples were exposed to four different doses of radiation with a BIOBEAM 8000 γ irradiator (IFR 31, Service de Médicine Nucléaire de Rangueil, Hôpital de Rangueil, Toulouse, France) equipped with a 137 Cs source to deliver radiation at a dose rate of 3.82–4.4 Gy/min: 2-hours exposure, 0.458–0.528 kGy 4 hours, 0.917–1.056 kGy 8 hours, 1.833–2.112 kGy and, 15 hours, 3.438–3.960 kGy. Control plates (No radiation) were not exposed, but placed close to the irradiator in the same room for the same period of time as the radiation treatment plates to ensure similar environmental conditions.

Immediately after the treatment the corresponding plates were kept at 4 °C. Five μL of each individual culture were serially diluted into a new microtiter plate containing 195 μL of TSB. For each sample five μL of each dilution were spotted on LB agar plates to calculate the number of C.F.U.s per mL after 24 h of incubation at 37 °C. Overall bacterial populations were calculated by multiplying the number of CFUs for the corresponding dilution factor. Five samples were removed from all the analysis and three more were excluded from the non-parametric pairwise comparisons (Supplementary Table S3).

Microbial species identification

The 16 S rDNA was amplified with the universal primers fD1 and rP2 47 . The PCR products were sequenced in 2012 to identify the microbial species when possible and sequences are deposited in GenBank (Supplementary Table S1).

Statistical analyses

We calculated both the geometric and the arithmetic means of the two counts of the number of CFUs for each bacterial morphotype, the former to lessen the impact of dissimilar values on meaningful statistical analyses and the latter because under radiation one of the CFUs counts was zero in few cases. All analyses shown in the text were done using the arithmetic mean except mentioned otherwise. The means were transformed using log (y + 1) to approximate a normal distribution. We further recorded the mortality of strains and we used a χ 2 -test to examine differences. We used a three-way ANOVA to examine overall differences in the number of CFUs produced by each bacterial morphotype with experimental exposure to radiation (irradiated vs. no irradiated), time of exposure (2, 4, 8 and 15 hours) and background radiation (high, intermediate, low and control) as factors. ANOVAs were performed with the R software 48 .

Due to the nature of the observations, however, our transformed data did not meet the assumption of homoscedasticity in many of the analysis. Therefore, we conducted Kruskall-Wallis tests to test for overall differences in bacteria survival for experimental exposure to radiation, time of exposure and background radiation intensity and post hoc Mann-Whitney pairwise comparisons within each factor, Bonferroni corrected. Then, we calculated 1-way ANOSIM 49 , using Bray-Curtis distance, to test for the interactions between background radiation intensity and treatment, treatment and time of exposure and background radiation intensity and time of exposure. We estimated both sequential Bonferroni significance and corrected Bonferroni P-values. We used the free software PAST 50 .

To examine differences in the performance of each community of bacteria isolated from sites with different background radiation intensity within the experimental exposure to radiation treatment, we used a two-way ANOVA with time of exposure (2, 4, 8 and 15 hours) as and background radiation (high, intermediate, low and control) as factors.

Finally, we explored whether colony colour had an effect on bacterial mortality under irradiation conditions. To do so, first, we visually allocated bacterial morphotypes into eight different categories, established beforehand by the authors, that are based on colony colour or morphology by sight: Creamy (all those beige, pale yellow or slightly pink and creamy colonies), Orange (all orange, peach or pale orange or peach colonies), Paenibacillus sp1 (big pale chiral pattern-forming bacteria), Paenibacillus sp2 (translucent swarming bacteria), Pink (pink colonies), Shiny (white and very lustrous colonies), White (white colonies) and Yellow (yellow colonies). Then, we quantified the overall mortality across treatments and places for each bacterial morphotype. We used a χ 2 -test to examine differences across background radiation intensity, experimental exposure to radiation and single bacteria morphotypes across treatments when the morphotype was represented by at least 10 colonies.


We acknowledge helpful discussions with John R. Battista that contributed to the interpretation of the results in this paper. We also acknowledge with gratitude the assistance of D. B. Knowles and M. Thomas Record in the analysis and interpretation of the pH-rate profile data in Fig. 6 A.

* This work was supported, in whole or in part, by National Institutes of Health Grant GM32335 from the NIGMS (to M. M. C.). This work was also supported by an Advanced Opportunity Fellowship from the University of Wisconsin College of Agricultural and Life Sciences and a Genentech predoctoral fellowship (both to K. V. N.). Michael M. Cox is a board member and shareholder of Recombitech, Inc., which is applying recombinational DNA repair to problems in medical diagnostics. His relationship to the company is managed by the University of Wisconsin-Madison in accordance with its conflict of interest policies.

History of Bioinformatics

The term Bioinformatics was coined by Paulien Hogeweg and Ben Hesper in the year 1970. However, the emergence of Bioinformatics can be traced back to the 1960s. The reason this field emerged is due to the development of protein sequencing methods.

Fredrick Sanger calculated the sequence of Insulin in the early 1950s. So as the protein sequences developed there was a need for a tool that would analyze and compare a huge number of protein sequences with each other. It was manually impossible to read these sequences and analyze. To give you an idea, a human genome consists of 3 Billion pairs of DNA strands. And figuring out the exact order of these DNA strands is next to impossible without computing powers.

So, as the analysis and comparison of the protein sequences looked impossible, the researchers worked on developing computer methods that would help them. That’s when the first “Protein Information Resources” (PIR) was developed by Margaret Oakley Dayhoff and her collaborators at the National Biomedical Research Foundation.

The PIR (Protein Information Resources) was like an atlas (a map) of protein sequences. These protein sequences were classified into different groups and subgroups according to their sequence similarity and percent accepted mutation (PAM) matrices. This atlas has been used widely in the field of Bioinformatics since then.

In the 1970s, Elvin A. Kabat further led to the development of the field by doing extended protein sequence analysis of antibodies. He collaborated with Tai Te Wu on this project and released the antibodies sequence b/w 1980 and 1991.

Further, the field was enriched by the following events-

· Collection of DNA sequences into GenBank* during 1982-1992. GenBank was prepared by Walter Goad’s group.

*GenBank is a comprehensive database that contains publicly available nucleotide sequences for more than 300,000 organisms.

· The DNA sequencing database became more useful when the researchers developed web-based searching algorithms. This helped researchers to find and compare the data they needed at that moment. And saved them to go through the entire sequence when they needed just a part of it only.

· Subsequently, a software developed called GENEINFO, by which researchers could rapidly search the sequences and match them with the other sequences.

· After that other software developed by the National Center of Biotechnology Information (NCBI), helped in the analysis, comparison and visualization of Molecular sequence.

· Development of FASTA and BLAST* greatly improved the biological data analysis

* FASTA and BLAST are two similarity searching programs that identify homologous DNA sequences and proteins. They provide facilities for comparing DNA and proteins sequences with the existing DNA and protein databases .

· Other than this, tools were developed for predicting the putative protein sequences, their structures and the functions of proteins based on DNA sequences. Full genome sequences were completed with these tools and a base genome database of various organisms was created.

· Finally, the ability for identification, data storing, mining and querying for large volumes of biological datasets has led to the unprecedented popularity and applications of Bioinformatics.

To further clarify the history of bioinformatics , following the timeline of events would be helpful.

Alfred Day Hershey and Martha Chase proved that the DNA carries genetic information

Sidney Brenner, Francois Jacob, Matthew Meselson identified RNA

Pauling gave the theory of molecular evolution

Margaret Dayhoff developed an Atlas of Protein Sequences

Needleman-Wunsch algorithm developed

Software to analyse DNA sequencing developed

Smith-Waterman algorithm developed

Sequence database searching algorithm

Creation of National Center for Biotechnology Information (NCBI)

EMBnet network developed for the distribution of the database

BLAST, software for fast searching of sequence

First bacterial genomes sequenced completely

Human genome gets published

2 Protein Synthesis Inhibitors

The bacterial protein synthesis machinery is a major target for antibiotics and it has been used for efficient structure-based interventions against antibiotic resistance. 17, 18 Protein synthesis, which is conducted by ribosomes, converts mRNA into the corresponding polypeptide chain. 19 This process can be divided into four steps: initiation, elongation, termination, and recycling. The first two are depicted in Figure 2 and will be described here briefly because they provide several targets for antibiotic intervention.

The mechanism of bacterial translation. Antibacterial targets highlighted.

Bacteria have 70S ribosomes that are made up of two subunits. 19-22 The large 50S subunit, which includes the 23S and 5S rRNAs, binds aminoacyl tRNA (aa-tRNA), catalyzes peptidyl transfer, and controls the elongation process, while the smaller 30S subunit, which includes the 16S rRNA, binds mRNA and initiates protein synthesis. The initiation step involves assembling these two subunits around the mRNA and the initiator fMet-tRNA. This process is catalyzed by three prokaryotic initiation factors: IF1, IF2 and IF3. The resulting 70S initiation complex has three main tRNA binding sites, the A, P, and E sites. The A site is the site at which the charged aminoacyl-tRNA matching the mRNA codon enters the ribosome. The P site carries the peptidyl-tRNA, the tRNA carrying the growing peptide chain. The E site harbors the deacylated or uncharged tRNA before it exits the ribosome. The elongation step is the cycle that adds amino acids to the growing peptide chain in a stepwise manner and can be considered the heart of protein synthesis. In the first step, a ternary complex composed of aa-tRNA, the elongation factor Tu, and guanosine triphosphate (aa-tRNAEF-TuGTP) binds at the A site. After successful decoding, the complex is hydrolyzed, thereby resulting in the departure of EF-TuGDP (GDP: guanosine diphosphate) and an inorganic phosphate (Pi), which allows the aa-tRNA to enter the A site and bind. Peptide bond formation then occurs through transfer of the entire peptide chain from the peptidyl-tRNA in the P site to the aa-tRNA in the A site. This pre-translocation ribosomal state (PRE state) is often referred to as a hybrid state since the tRNAs are moving back and forth between the A/A, P/P, A/P and P/E sites. In the next step, elongation factor G (EF-G) catalyzes translocation of the tRNA2mRNA complex by a distance of one codon (POST state). This results in the deacylated tRNA being moved to the E site, the peptidyl-tRNA being moved to the P site, and the A site becoming free to bind the next aa-tRNA. Until a stop codon enters the A site, this cycle continues, gradually building the full peptide chain. As seen in Figure 2, protein synthesis is targeted by a number of different classes of antibiotics at essentially every step of the process. While some classes of antibiotics such as macrolides, oxazolidinones, and pleuromutilins bind the large 50S subunit, others such as aminoglycosides and tetracyclines interfere with the smaller 30S subunit.

2.1 Oxazolidinones

Structures of the first two oxazolidinones synthesized by Upjohn.

These two molecules were the results of a wide structure–activity relationship (SAR) study that revealed the required substitution on the central oxazolidinone core. 27 Essential factors for antibacterial activity were the N-aryl substituent, the 5S configuration, and the C5 acylaminomethyl group. The meta-fluoro substitution of the phenyl ring was not essential, but usually helped to increase activity, and the para-substitution could be varied to expand the antibacterial spectrum.

While oxazolidinones are protein synthesis inhibitors that bind to the ribosome, it has taken a number of years to identify the binding site and most likely mode of action. 28 In 2008, two reported X-ray co-crystal structures of linezolid bound to 50S ribosomal subunits confirmed the previously established site of action and suggested a mode of action (Figure 4). 29, 30

A) Co-crystal structure of linezolid (1 C green, F purple) in the 50S ribosomal subunit from Haloarcula marismortui (PDB ID: 3CPW). B) A model of linezolid (1) in the 50S ribosomal subunit from H. marismortui that has been methylated by Cfr at A2503Ec. The surface clash is highlighted in red. Escherichia coli numbering in parentheses. Model was generated with Chimera 1.10.1 according to K. J. Shaw et al. 14, 35

Linezolid binds to the A-site pocket of the 50S subunit at the peptidyl transferase center (PTC) in actively translating bacterial ribosomes and interferes with binding of the charged aminoacyl tRNA. Specifically, linezolid binds to a pocket formed by eight RNA residues, one of which, U2585Ec, is stabilized in a distinct conformation. By stabilizing U2585Ec in a nonproductive conformation, linezolid affects the binding and/or positioning of the initiator-tRNA and prevents the binding of tRNA at the A site, thereby halting the translation sequence (Figure 4 A). 29

Resistance to oxazolidinones is still relatively rare. So far, three classes of oxazolidinone resistance mechanisms have been characterized. 31 The first involves mutations in the 23S rRNA central loop of domain V, the peptidyl transferase center. While some of the mutated residues interact directly with the oxazolidinone, many do not but are instead used to stabilize the region surrounding the oxazolidinone. 28 Mutations in these residues lead to small conformational changes of the linezolid binding pocket, which adversely affects drug binding. The second mechanism, which is less common, involves mutations in the genes rplC and rplD, which encode ribosomal proteins L3 and L4, respectively. 31

Beyond these chromosomally encoded point mutations, the last mechanism involves acquisition of the ribosomal methyltransferase gene cfr (chloramphenicol-florfenicol resistance). This resistance is more worrisome than the mutation-based mechanism since it is horizontally transferable and carries a low fitness cost. 32, 33 Mechanistically, the methyltransferase Cfr, through C8 methylation of the key residue A2503Ec in the 23S rRNA, greatly reduces susceptibility to a wide range of ribosome-targeting antibiotics, including amphenicols, lincosamides, pleuromutilins, streptogramin A, 16-membered macrolides, and linezolid. 34 As seen in Figure 4 B, the addition of a methyl group on A2503Ec leads to a steric clash with the acetamide group of linezolid, thereby causing a two- to eight-fold increase in the minimum inhibitory concentration (MIC). 36 Since the discovery of linezolid, over 30 companies have advanced more than a dozen candidates in clinical development. Unfortunately most of these have failed owing to issues related to pharmacokinetic (PK) properties, safety profile, solubility, or lack of improvement of antimicrobial activity over linezolid. Therefore the two main challenges for successful second-generation oxazolidinones are minimizing the myelosuppression safety signal and achieving adequate activity against linezolid-resistant strains of bacteria. 31 Two second-generation oxazolidinones that attempt to solve these problems in a rational way will be presented herein.

2.1.1 Tedizolid

Tedizolid phosphate (3 previously torezolid phosphate, TR-701, DA-7218) is the inactive prodrug of tedizolid (4 previously torezolid, TR-700, DA-7157, which was discovered by Dong-A Pharmaceuticals, and developed by Trius Therapeutics and Cubist Pharmaceuticals). 37, 38 After two successful phase III trials, tedizolid was approved by the Food and Drug Administration (FDA) in June 2014 under the trade name Sivextro for the treatment of MRSA skin infections (Figure 5).

Structures of tedizolid phosphate, tedizolid, and radezolid compared to linezolid.

Structurally, tedizolid presents two main differences compared to linezolid: substitution of the acylaminomethyl group at C5 by a hydroxymethyl moiety and introduction of the C/D ring system, here a 6-(2-methyl-2H-tetrazol-5-yl)pyridine (Figure 5). The increase in lipophilicity brought by the addition of the C and D rings obliged the medicinal chemists to find an adequate prodrug that would solve the low aqueous solubility and oral bioavailability problems. 37 A series of formulations was evaluated and the monophosphate ester was found to have the best properties (high water solubility and improved bioavailability) while also masking the primary alcohol, which provided a greatly improved monoamine oxidase inhibition profile. 31

The effect that these structural modifications bring to tedizolid can be clearly seen in Figure 6. At the bottom end of the binding site, the methylated A2503Ec is able to accommodate the sterically compact hydroxymethyl group while still maintaining the hydrogen bonding with G2505Ec. Additionally, the proposed binding model predicts two new stabilizing hydrogen bonds between the C/D ring system and the backbone ribose sugars of A2451Ec and U2584Ec.

Models of tedizolid (4) (C green, F purple) in A) the 50S ribosomal subunit from H. marismortui (PDB ID: 3CPW) and B) the 50S ribosomal subunit from H. marismortui that has been methylated by Cfr. Even with methylation at A2503Ec, tedizolid (4), unlike linezolid, is able to bind to the ribosomal RNA. E. coli numbering in parentheses. Models generated with Chimera 1.10.1 and AutoDock Vina 1.1.2 according to K. J. Shaw et al. 14, 16, 35

In a study of its activity against linezolid-resistant staphylococci, tedizolid showed a more than 16-fold improvement compared to linezolid. 39 It also maintained activity against most of the tested isolates, including multidrug-resistant ones, thus clearly demonstrating the benefit of the structure-based approach.

2.1.2 Radezolid

Radezolid (5 previously RX-1741 and Rx-01_667 Figure 5) is a fully synthetic oxazolidinone (developed by Melinta Therapeutics, formerly Rib-X) that has completed two phase II clinical studies. 40 Radezolid is the result of a program developed to expand the spectrum of oxazolidinones to Gram-negative bacteria and optimize drug-like properties. 41 The program started with the observation that linezolid and sparsomycin (6, a non-selective antibiotic) had overlapping binding sites within the peptidyl transferase center of the 50S ribosomal unit (Figure 7). 42, 43

Overlay of sparsomycin (6) (C orange) and linezolid (1 C green, F purple) in the 50S ribosomal subunit of H. marismortui (PDB IDs: 1M90, 3CPW).

The strategy was therefore to link the two molecules together via an adequate bridging element and make the required structural modifications on the sparsomycin side to increase potency and selectivity. 41, 44 Figure 8 shows two (7 and 8) of the numerous bridged antibiotics that were synthesized on the road to radezolid and clearly shows the evolution of the western half and the bridging unit.

Structures of sparsomycin, linezolid, and molecules developed on the way to radezolid.

Once the optimal structural elements were determined, the activity of radezolid was assessed and it was found to bind with higher affinity to the ribosome than linezolid, which gave it enhanced antibacterial activity (2–8-fold improvement over linezolid) against various Gram-positive pathogens. 45 Structure–activity studies of diverse oxazolidinones revealed that despite the presence of a C5 acylaminomethyl group in radezolid, it retains activity against the clinical cfr-positive CM05 strain of S. aureus with a minimal inhibitory concentration (MIC) value of 2 μg mL −1 , which is between those of tedizolid (0.5 μg mL −1 ) and linezolid (8 μg mL −1 ). This is most likely due to additional binding interactions of the C/D ring system with the PTC, far from the ribosomal modifications that lead to resistance to linezolid, including the cfr-mediated methylation of A2503Ec. 36, 40, 46 Most noteworthy is the fact that the spectrum could be expanded to Gram-negative organisms such as Haemophilus influenzae and Moraxella catarrhalis. Additional features of radezolid, besides overcoming the ribosomal mutation resistance, include a 100-fold decreased activity in inhibiting translation in rabbit reticulocytes than in S. aureus ribosomes and interaction with U2585Ec, as verified by biochemical assay. 40

2.2 Macrolides

Macrolides are a family of 14-, 15-, and 16-membered polyketide lactone rings with one or more neutral or amino sugar substituents at various positions (Figure 9). 47 Erythromycin (9), the first and prototypical macrolide, was isolated from actinomycete bacteria in 1949. 48 It was first used clinically in the early 1950s, but owing to its acid instability, second-generation semisynthetic macrolides such as clarithromycin (10) 49 or azithromycin (11) 50 were developed (Figure 9). Subsequently, owing to the emergence of resistance to first- and second-generation macrolides, a third generation, coined the ketolides, was developed. In these 14-membered lactone rings, the cladinose sugar at C3 is replaced by a ketone moiety, C11 and C12 are now part of an oxazolidinone ring, and an alkyl-aryl side chain is appended to the macrolactone core. Telithromycin (12) is the only registered ketolide to date (Figure 9).

Structures of the three macrolides erythromycin, clarithromycin, and azithromycin, as well as the ketolide telithromycin.

Like many other antibiotic classes, macrolides are bacteriostatic. They bind to the 50S ribosomal subunit in the vicinity of the peptidyl transferase center just above the constriction formed by the extended loops of ribosomal proteins L4 and L22 and were originally thought to inhibit protein synthesis by completely obstructing the ribosomal tunnel. 51 In fact, modeling studies have shown that even with a bound macrolide, the tunnel can still accommodate a nascent peptide chain. The macrolide nevertheless greatly hinders the progression of the peptide, which is usually dissociated by the peptidyl-tRNA drop-off mechanism before reaching its full size. 51, 52 Macrolides have also been shown to block the formation of the large 50S ribosomal subunit by binding to its precursors. 53

A number of co-crystal structures of macrolides bound to the ribosome have been published and they show that the main component in the binding pocket is A2058Ec. 54-56 Key interactions include polar contacts between the functional groups of the C5 desosamine sugar and residues A2058Ec and A2059Ec, as well as hydrogen bonds between the three lactone hydroxy groups and the 50S ribosomal subunit (Figure 10).

Overview of the binding mode of erythromycin (9 C green) to the 50S ribosomal subunit (PDB ID: 1JZY) from Deinococcus radiodurans. The nucleotides are labeled according to D. radiodurans (E. coli in parentheses).

Bacteria have developed a number of mechanisms of resistance to macrolides. 57 Even though the majority of them involve alterations at the ribosomal target site, substrate inactivating enzymes (mainly esterases) and efflux mechanisms have also been reported. Alteration of ribosomal binding sites usually leads to failure of the antibiotic to bind, which in turn disrupts its ability to inhibit protein synthesis. In the case of macrolide resistance, these alterations mainly take place in one of two ways. The first, which has the smallest effect, is modification of the ribosomal proteins L4 and L22 at the end of their hairpin structures, close to the binding site. 58 Interestingly, mutations at L4 lead to large reductions in the binding affinity while the L22 mutants show no change in the binding constant but still confer resistance. This is explained by a widening of the tunnel, which allows passage of the peptide without affecting the binding. 59 The second and most common resistance mechanism is modification of the rRNA. This takes place, unsurprisingly, on A2058Ec, the key component of the binding pocket. Mono and dimethylation of A2058Ec are carried out by erythromycin ribosome methylation (Erm) methyltransferases. 60 While monomethylation confers only moderate resistance to macrolides and none to ketolides, dimethylation leads to complete blocking of the binding site and high resistance to both macrolides and ketolides. 18, 61 Mutation of A2058Ec into G2058Ec also induces resistance owing to the similar increase in steric bulk. Notably, in archaea and eukaryotes, position 2058Ec is naturally a guanine, which explains the selectivity for bacteria. 62

As mentioned earlier, ketolides were developed in an effort to counter resistance to macrolides. Their key features include a 14-membered macrolactone, the replacement of the C3 cladinose sugar by a keto group, a cyclic carbamate and an extended alkyl-aryl side chain (Figure 11). The C3 keto group gives rise to potent activity against strains with Erm-mediated inducible resistance and surmounts resistance through efflux. 63 It also removes some of the steric hindrance around the desosamine sugar, which allows it to reposition itself when binding to monomethylated ribosomes. 64 The cyclic carbamate introduces additional interactions with the ribosome, which stabilize the conformation of the core macrolide and lead to potent antibacterial activity. 65 Crystal structures of ketolides in both D. radiodurans and in H. marismortui have been published and show notable differences in their binding modes. 51, 56, 66, 67 Indeed, when complexed to the ribosome of H. marismortui, the side chain of telithromycin resides over the plane of the macrolactone ring (PDB ID: 1YIJ) 56 whereas with the ribosome of D. radiodurans, the side chain points away from the macrolactone core (PDB ID: 1P9X). 66 This clearly highlights the importance of crystallographic data and is a good reminder to exercise caution when using models. Nevertheless, the common feature is that the ketolides not only bind to domain V in a similar fashion as macrolides, but their elongated side chains engage in additional interactions in domain II. This results in tighter binding and allows them to compensate for modifications in domain V resulting from mutation or methylation.

Structures of the ketolides telithromycin and solithromycin.

Telithromycin (12) is currently the only ketolide on the market, but following safety controversies, it has been partially withdrawn. 68 Its use has been associated with severe hepatotoxicity along with blurred vision and serious cases of liver failure. 69 These are believed to result from inhibition of the nicotinic acetylcholine receptors. The pyridine ring in the extended side chain of telithromycin has been suggested to be the culprit. 70 New ketolides lacking this problematic pyridine ring are currently being investigated.

2.2.1 Solithromycin

Solithromycin (13 previously CEM-101, developed by Cempra) is a 2-fluoroketolide currently undergoing phase III clinical trials. As seen in Figure 11, solithromycin is very similar to telithromycin, with only two minor modifications. The first is replacement of the aforementioned problematic imidazolyl pyridine with a triazolyl aniline, and the second is the introduction of a fluorine atom at the C2 position. Removal of the imidazolyl pyridine resulted in a 30-fold reduction in the inhibition of nicotinic acetylcholine receptors compared to telithromycin. 70 As mentioned earlier, previous X-ray structures of ribosome-bound ketolides showed very different orientations of their alkyl-aryl side chains depending on the bacterial species. Therefore, questions remained as to the actual binding mode of solithromycin and telithromycin in pathogenic bacteria. In 2010, Cate, Mankin, and co-workers crystallized solithromycin as well as telithromycin in complex with the E. coli ribosome and were able to indicate that the placement of these ketolides most likely reflects the binding in S. aureus ribosomes given the sequence conservation of the A752:U2609 base pair among many eubacteria. 71, 72

These crystal structures delivered revealing insight into various aspects of the interaction (Figure 12). A stacking interaction can be seen with the A752:U2609 base pair, which is present in the ribosome of E. coli and many other pathogenic bacteria. The aniline moiety, which replaces the detrimental pyridine is shown to form additional hydrogen bonds, notably to A752 and G748, and results in tighter binding. Finally the fluorine atom at the C2 position was shown to contribute to drug binding as well as chemical properties such as solubility and cellular uptake. 73 Indeed, in comparison to non-fluorinated analogues, the fluorinated versions showed stronger inhibition of the growth of streptococci carrying the erm gene. Interestingly, for solithromycin, weak binding to ribosomes dimethylated at A2058Ec could be detected by chemical probing. 71 Key structural features of the ketolides are summarized in Figure 13 (for an example of a ketolide with a 6-O-attached side chain, see cethromycin). 74, 75

Overlay of telithromycin (12 C green, PDB ID: 4V7S) and solithromycin (13 C yellow, F purple) in the 50S ribosomal subunit (PDB ID: 4WWW) from E. coli. Nucleotides are labeled using the E. coli numbering system.

Structure–activity relationships of ketolides.

2.3 Thiopeptides

The thiopeptides are a family of highly modified sulfur-rich macrocyclic peptides (Figure 14, numbering according to thiomuracin A), 76 which have been isolated from diverse sources such as soil bacteria and marine samples. 77 While there are now over 100 known thiopeptides, the first member, micrococcin P1 (14), was isolated in 1948. 78 These molecules are of ribosomal origin, are highly posttranslationally processed, and feature a characteristic macrocyclic core consisting of multiple thiazoles and a 6-membered nitrogen-containing heterocycle, which can be found in different oxidation states. While thiopeptides are a new class of antibiotics with a novel mechanism of action, their use as an antibiotic treatment option has been hampered by their very large molecular size and poor aqueous solubility. Thiopeptides are inhibitors of protein synthesis, but their mode of action differs depending on the size of the macrocycle. Thiostrepton A (15), 79 the archetypal thiopeptide, and micrococcin P1 possess 26-membered macrocyclic cores and are known to bind to the GTPase-associated region of the ribosome/L11 protein complex. 80 Thiopeptides with 29-membered macrocyclic cores such as GE2270 A (16), on the other hand, interact with GTP-bound bacterial EF-Tu, preventing the formation of the ternary complex with aa-tRNA. 81

Structures of thiopeptides.

GE2270 A was isolated in 1991 by scientists at Lepetit Research Institute. 82 Although the in vitro activity of this compound was shown to be excellent against MRSA, VRE, and streptococci, poor aqueous solubility prevented further development. 83 Two derivatives are currently being investigated.

2.3.1 LFF571

In order to increase the water solubility of GE2270 A, the unstable oxazoline side chain was replaced with solubilizing functional groups (Novartis). The 4-aminothiazolyl moiety was chosen as a starting point and a wide variety of amines and acids linked via different spacers were synthesized. 83-85 Guided by co-crystal structures with EF-Tu, this search led to the discovery of two potent analogues with cyclohexylcarboxylic acid side chains residing in proximity to the Arg223 residue of EF-Tu.

These compounds were then improved by the addition of a second solubilizing group. 86 Taking into account the position of Arg262, a large variety of differently linked acids were synthesized. It was found that a pentanoic acid residue appended onto the aminothiazole was the best since it placed the acid in proximity to this residue (Figure 15). The resulting compound, named LFF571 (17), also showed very high aqueous solubility (>10 mg mL −1 ). Triacid-containing analogues were also synthesized but these both failed to yield additional benefits and greatly increased the synthetic complexity, and they were therefore not pursued. LFF571 is currently undergoing phase II clinical trials against C. difficile infections.

Co-crystal structure of LFF571 (17 C green) and EF-Tu (PDB ID: 3U2Q) from E. coli. The carboxylic acid groups of LFF571 are in proximity to Arg223 and Arg262.

2.3.2 NVP-LDU796

In 2009, the isolation of thiomuracins, a novel class of antibiotic thiopeptides, was reported (Novartis). 76 These secondary metabolites, which are produced by a strain of Nonomuraea, are structurally related to GE2270 A and share the same mechanism of action, namely binding to EF-Tu. The thiomuracins show potent antibiotic activity against MRSA and VRE, with minimum inhibitory concentrations below 1 μg mL −1 . Unfortunately, as with GE2270 A, these novel molecules are plagued by solubility and stability problems. In 2012, the same group reported the synthesis of a novel derivative of thiomuracin A (18), termed NVP-LDU796 (19), which retains antibacterial activity and shows improved chemical stability and physiochemical properties (Figure 14). 87 Key modifications include removal of the C2–C10 side chain and conversion of the C84 epoxide into an N70–C84 pyrrolidine ring. In co-crystal structures of NVP-LDU796 (19) with EF-Tu, the conformation adopted is very similar to the those of GE2270 A and LFF571, with key interactions still present (PDB ID: 4G5G). 87

2.4 Tetracyclines

Tetracyclines are broad-spectrum antibiotics with activity against both Gram-positive and Gram-negative bacteria. 88 Chlortetracycline (20), the oldest member of this class, was discovered in 1945 and first used clinically in 1948 (Figure 16). 89 Since then, only three other naturally occurring tetracyclines have been discovered (tetracycline (21), 90 oxytetracycline (22) 91 and demethylchlortetracycline (23) 92 ), while countless others have been derived semisynthetically, including doxycycline (24) 93 and minocycline (25). 94 Tetracyclines are easily recognizable by their four highly oxygenated fused rings and they show favorable antimicrobial properties along with an absence of major side effects, which has led to their extensive use in both human and animal infections. The third generation of tetracyclines, the glycylcyclines (Figure 17), were introduced 10 years ago, with the first and so far only clinically used member being tigecycline (26, previously GAR-936, discovered at Wyeth-Ayerst Research). 95

Structures of representative tetracycline antibiotics.

Structures of representative glycylcycline and aminomethylcycline antibiotics.

Glycylcyclines can be recognized by the glycylamido substituent on the C9 carbon. The tert-butylglycylamido moiety of tigecycline engages in stacking interactions with C1054 of the 16S rRNA, which leads to increased potency of tigecycline compared to tetracycline (Figure 18). 96

Overlay of tetracycline (21 C yellow PDB ID: 4V9A) and tigecycline (26 C green, Mg 2+ ions light brown PDB ID: 4V9B) bound to the 30S ribosomal subunit (PDB ID: 4V9B) from Thermus thermophilus. Tigecycline makes additional stacking interaction with C1054 compared to tetracycline. T. thermophilus numbering is used for the nucleotides.

Tetracyclines are primarily bacteriostatic. They penetrate the outer membrane of Gram-negative bacteria by passive diffusion through the OmpF and OmpC porin channels as divalent metal-ion chelates. 88 Once inside the periplasm, the liberated neutral tetracycline diffuses through the inner membrane in the same way that it penetrates Gram-positive bacteria and other organisms, namely by energy-dependent active transport. Once inside the cytoplasm, the higher pH and metal-ion concentration lead to the regeneration of a metal-ion/tetracycline complex that is postulated to be the active species. This complex then binds reversibly to the A site on the head of the 30S subunit (Figure 18). 96, 97 This binding site has a slight overlap with the anticodon stem-loop, which correlates well with previous observations that tetracycline prevents binding of the aforementioned aa-tRNAEF-TuGTP ternary complex to the A site. 98

Resistance to tetracyclines can occur through four different mechanisms. 88 The proteins responsible for three of these mechanisms are encoded by the tet (tetracycline) and otr (oxytetracycline) genes, over 40 of which have been characterized. Of these, only three, namely tet(X), tet(34), and tet(37), lead to an enzymatic-alteration resistance mechanism where the encoded proteins chemically modify tetracycline. 88 Of the remaining genes, approximately two thirds encode efflux proteins and the others encode ribosomal protection proteins (RPPs). The efflux proteins occur in both Gram-positive and Gram-negative bacteria, but are more prominent in the latter. They are membrane-associated proteins and export tetracyclines from the cell, which effectively protects the ribosome by reducing the intracellular concentration of the antibiotic. Ribosomal protection proteins are cytoplasmic proteins that protect the ribosome from the action of tetracyclines by reducing their susceptibility. They are also found in both Gram-positive and Gram-negative bacteria, but they are usually more common in Gram-positive organisms, and tet(M) and tet(O) are the two most common genotypes. Mechanistically, these proteins cause allosteric disruption of the primary tetracycline binding site, which leads to the release of bound tetracycline molecules. 99 The ribosome is then able to return to its productive conformation and resume protein synthesis. The last and most recent mechanism of resistance involves mutations in the vicinity of the tetracycline binding site. 100

The presence of the tert-butylglycylamido moiety in tigecycline (26) interferes with the binding of TetM to the ribosome and thereby enables the drug to overcome TetM-mediated resistance. As shown in Figure 19, tigecycline and TetM overlap in the ribosome owing to the bulky tert-butylglycylamido substitution of the drug molecule. 96

Overlay of TetM (blue ribbon PDB ID: 3J25) and tigecycline (26 C green, Mg 2+ ions light brown PDB ID: 4V9B) bound to the 30S ribosomal subunit (PDB ID: 4V9B) from T. thermophilus. The superimposition was generated by means of the cryo-electron microscopy density map EMD-2183 of the TetM-70S complex from E. coli according to Jenner et al. and Dönhöfer et al. 96, 99

2.4.1 Omadacycline

Omadacycline (27 previously PTK-0796, Paratek Pharmaceuticals, Figure 17) is a semisynthetic tetracycline derivative and the first member of the novel aminomethylcycline class. It is currently undergoing phase III clinical trials for the treatment of acute bacterial skin and skin structure infections (ABSSSI), community-acquired bacterial pneumonia (CABP), and complicated urinary tract infection (cUTI). Omadacycline has been shown to be active against strains that express either efflux proteins (tet(K)) or ribosome protection (tet(M)), while also exhibiting moderate inhibition of peptidoglycan synthesis. 101, 102 The exact mechanisms by which omadacycline evades both efflux and ribosome protection are not fully known, but it is believed that omadacycline is a poor substrate for efflux transporters and that it binds in a unique way that circumvents the action of ribosome protection proteins. 101 Both of these factors almost certainly arise from the structural modification at C9 of the tetracycline core.

2.5 Aminoglycosides

Aminoglycosides are hydrophilic molecules comprised of a central aminocyclitol core linked to one or more amino sugars. In most cases, the aminocyclitol is streptamine or 2-deoxystreptamine (28 and 29, Figure 20). Depending on the substitution pattern, aminoglycosides can be grouped into four different subfamilies: monosubstituted (such as neamine (30)), 103 atypical, 4,5-disubstituted, and 4,6-disubstituted (Figure 20). 104 The aminoglycoside family of antibiotics is one of the oldest, with its first member, streptomycin (31), being discovered more than 70 years ago in 1944. 105 Despite their long history, widespread resistance, and possible safety issues such as nephrotoxicity and ototoxicity, aminoglycosides are still widely used. 106

Structures of typical and atypical aminoglycosides. The aminocyclitol core is highlighted in blue.

The aminoglycosides have two distinct main mechanisms of action. They are not only potent inhibitors of translocation but are also able to induce misreading by stabilizing the binding of near-cognate tRNAs and promoting their incorporation into peptide chains. 104 The high fidelity of translation (error frequencies ranging from 10 −3 to 10 −4 per codon) 107 is achieved by the ability of the ribosome to select the proper (cognate) tRNA over a wrong (non-cognate) one at the A site.

The structure of the 30S ribosome shows a “decoding” site within helix 44 (h44) of the 16S rRNA. In this asymmetric internal loop, two universally conserved adenine residues, A1492 and A1493, are directly involved in the decoding process, during which they flip out of the helix to analyze the codon–anticodon complex. 108 The energy needed for this flip is thought to be compensated by stabilizing interactions between the nucleotides and the codon–anticodon complex, but only in the case of a cognate tRNA. When a non- or near-cognate tRNA binds, the energy compensation is insufficient and the tRNA dissociates. Upon binding within the internal loop of h44, aminoglycosides induce a local rearrangement that flips A1492 and A1493 out of the helix and stabilizes them in this open conformation (Figure 21). 109 This results in near-cognate tRNA being fully accommodated in the A site, with the consequence that incorrect amino acids are incorporated into the peptide chain.

Kanamycin A (37 C green) in complex with a decoding A-site oligonucleotide (PDB ID: 2ESI). A1492 and A1493 are kept in the flipped-out position by kanamycin A. E. coli numbering is used for the nucleotides.

When the altered proteins are inserted into the cell membrane, the permeability is modified, which in turn leads to an increase in aminoglycoside uptake (hence their high bactericidal and concentration-dependent activity). Co-crystal structures of a variety of bound aminoglycosides have been published, including with streptomycin (31, PDB ID: 1FJG), 110 spectinomycin (32, PDB ID: 1FJG), 110 hygromycin B (33, PDB ID: 1HNZ), 97 neomycin B (34, PDB ID: 4V52), 111 paromomycin (35, PDB ID: 1FJG, 1IBK), 110, 112 tobramycin (36, PDB ID: 1LC4), 113 and kanamycin A (37, PDB ID: 2ESI). 109

Bacteria have developed three different mechanisms of resistance to aminoglycosides: uptake inhibition or efflux, ribosome modification, and aminoglycoside modification. 114 The first two mechanisms are relatively rare and have not been targeted yet. Aminoglycoside-modifying enzymes (AMEs), on the other hand, are widespread and the most common mechanism of resistance. These enzymes are described according to the modification they promote, along with the carbon atom at which the modification takes place. 115 The three known types of enzymes are aminoglycoside N-acetyltransferases (AACs), aminoglycoside O-nucleotidyltransferases (ANTs), and aminoglycoside O-phosphotransferases (APHs). Figure 22 shows the co-crystal structure of kanamycin A (37, Figure 20) with the aminoglycoside-modifying enzyme ANT(2′′)-Ia. 116 The 2′′ hydroxy group is complexed to the magnesium ion, ready to be modified by the enzyme.

Co-crystal structure of kanamycin A (37 C green) and adenylyltransferase ANT(2“)-Ia (Mg 2+ ions light brown PDB ID: 4WQL) from Klebsiella pneumoniae. Kanamycin A is kept in place by a hydrogen-bonding network.

A second generation of aminoglycosides called neoglycosides, 117 which maintain the potency of first-generation aminoglycosides while evading modification enzymes, is currently being researched.

2.5.1 Plazomicin

Two aminoglycosides, sisomicin 118 and amikacin 119 (38 and 39 Figure 23), were used as inspiration for the development of the semisynthetic plazomicin (40, formerly ACHN-490, developed by Achaogen), which is the first molecule of this new family (Figure 24). 117

Structures of sisomicin and amikacin.

Structure of plazomicin and AME modification sites. The HABA chain (blue) blocks modifications at the 3-N and 2′′-O positions, the hydroxyethyl chain (red) blocks modifications at the 6′-N position, and the absence of hydroxy groups at C3′ and C4′ prevents modification at these positions. Only the 2′-N position remains unblocked to AACs.

The difference between sisomicin and plazomicin is the presence of two side chains on the nitrogen atoms at C1 and C6′ in plazomicin. As highlighted in Figure 24, the hydroxyethyl chain shown in red blocks AAC(6′), while the amikacin-derived hydroxyaminobutyric acid (HABA) chain shown in blue blocks AAC(3) along with ANT(2′′) and APH(2′′). 120, 121 Compared to kanamycin A, plazomicin is also protected from APH(3′) and ANT(4′) by the absence of hydroxy groups at positions C3′ and C4′. The only aminoglycoside-modifying enzymes that plazomicin is still vulnerable to are AAC(2′) enzymes, but so far the expression of these enzymes has been detected only in Providencia stuartii. In MDR Enterobacteriaceae, including carbapenem-resistant Enterobacteriaceae (CRE), plazomicin remains active where most other antibiotics, including the commercially available aminoglycosides, show limited potency owing to resistance.


Database implementation

DRAGdb comprises of a single table where each mutation entry is uniquely identified with DRAGDB_ID as the primary key. The NUCLEOTIDE_POSITION, NUCLEOTIDE_CHANGE, AMINOACID_POSITION, AMINOACID_CHANGE define the mutation point at both levels. The PUBMED_ID provides PubMed identifier, hyperlink to PubMed database and ENSEMBL_BACTERIA_ID provides the gene identifier.

DRAGdb was developed using the Apache HTTP 2.2.15 web server and MySQL 5.1.69. The PHP 5.3.3, HTML, JavaScript and CSS were used to build the web interfaces of the database. The PHP-based web interfaces execute the SQL queries dynamically. It is freely accessible at

Data curation

The PubMed database (till March 2018) was searched for studies that reported at least one mutation in rpoB, pncA, inhA, katG, embA, embB, embC, gidB, rpsL, rrs, gyrA and gyrB associated with resistance to rifampicin, pyrazinamide, isoniazid, ethambutol, streptomycin and fluoroquinolones respectively in MTB, ESKAPE and other bacterial species. The literature was searched using advance search option of PubMed with the terms: “Gene name (Abstract/title) AND Resistance (Abstract/title) AND mutation (Abstract/title) AND/ NOT tuberculosis (Abstract/title)”. The combination of search terms helped to obtain instances with cross resistance and multiple resistances. In total, 2548 unique publications were obtained from this search. The publications that were missing full English text in public domain, or lacked relevant data or had ambiguous data were filtered out. Around 604 publications were systematically reviewed to obtain mutational information. All the mutations described in drug resistant bacterial strains in the literature were manually read, further curated and compiled in the database. The devised methodology is given as workflow in Supplementary File 1: Figure S3.

Mutation data analysis with reference to MTB H37Rv

All the gene mutations reported in the literature across bacterial species have different numbering systems (NS) thus leading to genetic location inconsistency and conflict. One of the examples of NS discrepancy is of gyrA in MTB, for which 4 different NS were found in the literature 54 . For better understanding and comparison across species of a single gene, Mycobacterium tuberculosis H37Rv was selected as reference organism, further multiple sequence alignment (MSA) was performed at amino acid codon level for each drug resistance gene to have single numbering system across all organisms. MSA was performed on on-line Clustal Omega platform using default iterated mBed-like Clustering Guide-tree 55,56 . The rational for choosing MTB as reference genome was due to the fact that exposure of 3–6 antibiotics including broad spectrum antibiotics during TB treatment for 6 months results in known multiple drug resistance phenotypes. The MSA of the regions of interest for genes such as gyrA, gyrB, rpoB, and rpsL were shown in Supplementary File 1: Figures S1(A–D). The common reference number at the amino acid codons level of drug resistance genes across bacterial species helped in calculating frequency of mutated codon positions in DRAGdb. The frequency percentage was calculated using the following formulae –

where (_) is the frequency percentage of (^) codon or nucleotide position in a gene of (^) group, (x) can be all organisms, Mycobacterium, ESKAPE pathogens or other bacteria. (_) is the number of mutation entries of (^) codon or nucleotide position in the gene of (^) group in DRAGdb. (mathoplimits_^_) is the total number of mutation entries in the drug resistance determining region (DRDR) of the gene, (j) is the starting codon or nucleotide position and (t) is the end codon or nucleotide position of DRDR. The number of mutation entries was calculated based on report of a single mutation across various PubMed literature. We assume that the higher the number of publications reporting a particular drug resistance determining gene mutation, the higher is the confidence of that mutation entry.

Functional effects of the mutations

The functional effects of the unique SNPs in drug-resistance genes in different bacteria were predicted using PROVEAN webserver with Score thresholds for prediction as of −2.5. The variants with score equal to or below of −2.5 were considered “deleterious”, and the variants with score of above −2.5 were considered “neutral” 57 .

Blast search

A customized BLAST database was created with wild type and mutated small nucleotide stretches of drug resistance determining regions of associated genes. The mutated sequences were modified wild type sequences with incorporation of single mutations enlisted in DRAGdb. blastall, a package for BLAST search was used 58 . formatdb utility from that package was used for converting nucleotide FASTA sequences to BLAST database. blastn program was used to find similar sequences to query sequences in the BLAST database.

DRAGdb user interface

The ‘HOME’ page of DRAGdb web interface provides two different search options: 1) keyword search: a single keyword can be searched specific to bacteria, resistant drugs, genes, geographical location or ‘ALL’ option to search in any category. 2) Advance search: three fields are present where bacteria and gene name are mandatory and drug name is optional. Both the search options will generate a table giving details of the mutations related to the search and also provide the number of specific entries. The DRAGdb result pages also contain hyperlinked Ensembl Bacteria IDs, PROVEAN score and PubMed IDs. To keep with the open access policy, the result table can be downloaded by the users. The ‘BROWSE’ page allows users to browse DRAGdb data in three categories: 6 drugs, 12 genes, and 126 bacterial species. It shows the comparison of DRAGdb data with other tuberculosis databases namely, TBDReaMDB and MUBII-TB-DB. The ‘Organisms’ section is further divided into 3 parts: ‘Mycobacterium tuberculosis’, ‘ESKAPE’ and ‘others’ which includes other bacterial species. The entries within the three categories are linked to DRAGdb table and provide specific results with details of the gene mutations. The nucleotide BLAST search with customized BLAST database is incorporated in the ‘TOOL’ page to determine whether the users input bacterial gene sequence is drug resistant. Users can define the ‘E-value’ for BLAST operation. The output page shows the user input sequence, the DRAGDB_ID of the best hit, the BLAST score and E-value of the hit. ‘OTHER LINKS’ page is also included to help users find popular TB and antibiotic resistance related databases and webservers. To guide users through DRAGdb, a ‘HELP’ page is also presented in the online web server.

Data visualization

The bar plots for representation of frequency % of various codon level mutations of drug resistance genes across bacterial species were drawn using Microsoft office excel. The circular plots for representations of homoplasy and pleiotropy were drawn using ‘circlize’ R package 59 .

Watch the video: Deinococcus radiodurans (June 2022).


  1. Tarleton

    What are all these people talking about in the comments? o_O

  2. Praza

    Anything similar.

  3. Ausar

    Excuse me, I have removed this phrase

Write a message