Antibiotics and Their Mechanisms of Action
A patient with a Streptococcus pyogenes throat infection swallows a penicillin tablet. She recovers within days. Meanwhile, the trillions of bacteria in her own gut are barely disturbed. How can one molecule kill an invading pathogen? At the same time, it leaves the human cells it passes through completely unharmed. The answer is selective toxicity. Antibiotics work by exploiting molecular differences between bacterial and eukaryotic cells. They strike structures or processes that bacteria possess but human cells lack. Alternatively, they target things that differ enough in bacteria to be druggable without collateral damage. This is not a vague notion of "killing germs." It is precision biochemistry. Understanding it requires knowing exactly which bacterial machine each drug class jams. It also requires knowing why that jam is lethal or growth-arresting. This article builds on the Bacterial Cell Wall and Peptidoglycan and the Gram Staining and the Gram-Positive/Negative Divide. It explains how four great structural and physiological differences become drug targets. These differences are a peptidoglycan wall, bacterial-type ribosomes, distinct membrane lipids, and bacterial-specific enzymes of DNA/RNA metabolism. The logic matters for two reasons. It explains why antibiotics work so well when they do. It also explains why bacteria can evolve around them so readily, foreshadowing Antibiotic Resistance: Mechanisms and Spread. The organizing principle is Paul Ehrlich's century-old dream of a "magic bullet." This is a chemical toxic to the parasite, harmless to the host. Modern antibiotics realize this dream imperfectly but powerfully. Each class occupies a distinct biochemical niche. This article works through wall synthesis inhibitors (β-lactams, glycopeptides), membrane-disrupting agents, protein synthesis inhibitors (aminoglycosides, tetracyclines, macrolides, oxazolidinones), and nucleic acid synthesis inhibitors (fluoroquinolones, rifamycins, sulfonamides). Each mechanism is paired with the classic experiments that revealed it. By the end, you should be able to predict some things from a drug's target alone. You should know whether it will be bactericidal or bacteriostatic. You should understand why it affects some bacteria and not others. You should also know what a resistant mutant would need to change to survive it.
- Antibiotics achieve selective toxicity by hitting bacteria-specific targets absent or divergent in human cells.
- Cell wall inhibitors (penicillins, cephalosporins, vancomycin) block peptidoglycan cross-linking; cells lyse from unchecked autolysins and osmotic pressure.
- Protein synthesis inhibitors exploit the structural difference between bacterial 70S and eukaryotic 80S ribosomes.
- Nucleic acid inhibitors target bacterial DNA gyrase/topoisomerase IV, RNA polymerase, or folate synthesis enzymes absent in humans.
- The Hodgkin/Abraham era clarified β-lactam structure; Gause's discovery of gramicidin and Waksman's screening programs revealed whole new classes.
- Bactericidal vs. bacteriostatic action, spectrum (broad vs. narrow), and Gram-positive/negative permeability all follow directly from mechanism.
Antibiotics work by binding bacterial-specific molecular targets tightly enough to block an essential process, while sparing the analogous human structure because it differs enough in sequence, structure, or existence to escape the drug.
The core idea
Selective toxicity is the principle that a chemical can be lethal to one organism (a bacterium) while being innocuous to another (its human host), because the drug's target is either absent from the host, structurally distinct enough that the drug does not bind it, or present at concentrations low enough that the therapeutic dose does not affect host cells. Antibiotics are a subset of antimicrobial agents — specifically those produced (originally) by microorganisms themselves, as opposed to fully synthetic antimicrobials, though the term is now used loosely for both. They are classified along two independent axes: by mechanism of action (what cellular process they block) and by spectrum of activity (how many kinds of bacteria they affect — narrow-spectrum drugs like vancomycin hit mainly Gram-positives, broad-spectrum drugs like tetracycline hit both Gram-positive and Gram-negative organisms). A second crucial distinction is between bactericidal agents, which kill bacteria outright (most cell wall and DNA-damaging agents), and bacteriostatic agents, which merely halt growth and replication, relying on the host immune system to clear the now-static population (most protein synthesis inhibitors at standard doses). This distinction has real clinical consequences: bacteriostatic drugs are inadequate in immunocompromised patients or in anatomically protected sites like heart valve vegetations, where host defenses cannot finish the job. The four canonical target classes, each covered in turn below, are: (1) cell wall synthesis — unique to bacteria because human cells have no peptidoglycan at all; (2) cell membrane integrity — exploiting differences in membrane lipid composition and phospholipid charge; (3) protein synthesis — exploiting the biophysical difference between the bacterial 70S ribosome (30S + 50S subunits) and the eukaryotic cytoplasmic 80S ribosome (40S + 60S subunits); and (4) nucleic acid synthesis and folate metabolism — exploiting bacteria-specific enzymes such as DNA gyrase, a bacterial RNA polymerase with different subunit structure, and dihydropteroate synthase, an enzyme in folate synthesis that humans lack entirely because we obtain folate from the diet.
Mechanism
1. Cell wall synthesis inhibitors. Peptidoglycan is a mesh of alternating N-acetylglucosamine (NAG) and N-acetylmuramic acid (NAM) chains, cross-linked by short peptide bridges. The final cross-linking step is catalyzed by transpeptidases, historically called penicillin-binding proteins (PBPs), which form a peptide bond between the terminal D-alanyl-D-alanine of one peptide stem and a neighboring stem, releasing the terminal D-Ala. β-lactam antibiotics (penicillins, cephalosporins, carbapenems, monobactams) contain a four-membered β-lactam ring that structurally mimics the D-Ala-D-Ala terminus. The transpeptidase's active-site serine attacks the β-lactam carbonyl as if it were the natural substrate, forming a stable, essentially irreversible acyl-enzyme complex — the PBP is permanently inactivated. Cross-linking stops, but the cell's own autolysins (peptidoglycan-remodeling hydrolases active during normal growth) keep cutting existing bonds unopposed. The wall weakens catastrophically at growing points; with the cell's high internal osmotic pressure unchecked, water rushes in and the cell lyses. This is why β-lactams are bactericidal only against actively growing, dividing cells — a resting bacterium making no new wall is largely untouched (the basis of drug antagonism seen when a bacteriostatic drug is combined with a β-lactam).
Glycopeptides (vancomycin, teicoplanin) act one step earlier and by a completely different chemistry: rather than binding an enzyme, vancomycin binds directly to the D-Ala-D-Ala terminus of the peptidoglycan precursor itself (the lipid II intermediate), sterically blocking both transglycosylation and transpeptidation. Because vancomycin is a bulky glycopeptide, it cannot cross the outer membrane porins of Gram-negative bacteria, making it essentially Gram-positive-only — a direct consequence of the Gram Staining and the Gram-Positive/Negative Divide.
2. Membrane-disrupting agents. Polymyxins (e.g., colistin/polymyxin E) are cationic cyclic peptides that bind the negatively charged lipid A phosphate groups of lipopolysaccharide in the Gram-negative outer membrane, then insert into and permeabilize both outer and inner membranes like a detergent, causing leakage of cellular contents. Because human cell membranes are built from zwitterionic phospholipids (phosphatidylcholine) rather than the anionic lipids enriched in bacterial membranes, polymyxins show some selectivity — though imperfectly, which is why they are reserved as last-line drugs against multidrug-resistant Gram-negatives, given real nephrotoxicity.
3. Protein synthesis inhibitors. The bacterial ribosome (70S = 30S + 50S subunits, sedimentation coefficients not additive because shape matters) is structurally distinct from the cytoplasmic eukaryotic ribosome (80S = 40S + 60S), giving several drug classes a genuine selectivity window (mitochondrial ribosomes are more bacteria-like, which explains some off-target toxicities such as aminoglycoside ototoxicity).
- Aminoglycosides (streptomycin, gentamicin) bind the 30S subunit at the A site, distorting the mRNA decoding region so that near-cognate tRNAs are accepted, causing misreading of codons and production of nonfunctional, misfolded proteins that get inserted into the membrane, increasing permeability — a bactericidal, concentration-dependent mechanism unusual among protein synthesis inhibitors.
- Tetracyclines bind the 30S subunit and sterically block incoming aminoacyl-tRNA from docking at the A site, halting elongation — bacteriostatic.
- Macrolides (erythromycin, azithromycin) bind the 50S subunit near the exit tunnel through which the nascent polypeptide emerges, blocking its passage and causing premature dissociation of the peptidyl-tRNA — bacteriostatic.
- Oxazolidinones (linezolid) bind the 50S subunit at the peptidyl transferase center itself, blocking formation of the initiation complex — a distinct site from macrolides, useful against macrolide-resistant strains.
- Chloramphenicol binds the 50S subunit and directly inhibits the peptidyl transferase reaction that forms the peptide bond.
4. Nucleic acid and folate synthesis inhibitors. Fluoroquinolones (ciprofloxacin, levofloxacin) inhibit DNA gyrase (topoisomerase II) in Gram-negatives and topoisomerase IV in Gram-positives — enzymes that relieve supercoiling ahead of the replication fork by transiently cutting both DNA strands. The drug traps the enzyme-DNA cleavage complex, converting a transient nick into a permanent double-strand break, which is bactericidal. Human topoisomerase II is structurally different enough (and the drug concentrations used are low enough) that host DNA is largely spared. Rifamycins (rifampin) bind the β subunit of bacterial RNA polymerase, blocking the initiation of transcription; the bacterial and human RNA polymerases are different enough in this region that rifampin does not touch host transcription at therapeutic doses. Sulfonamides and trimethoprim attack sequential steps of folate biosynthesis: sulfonamides are structural mimics of para-aminobenzoic acid (PABA) that competitively inhibit dihydropteroate synthase, while trimethoprim inhibits dihydrofolate reductase. Because humans obtain folate from the diet and lack these bacterial synthesis enzymes altogether, this pathway offers essentially perfect selective toxicity — the reason sulfa drugs were the first broadly effective synthetic antibacterials, and why the two drugs are combined (co-trimoxazole) for sequential, synergistic blockade of one pathway.
The evidence
Penicillin's mechanism — Park and Strominger, and the "Park nucleotides." In the 1950s, Jack Strominger and colleagues, building on work by J. T. Park, discovered that penicillin-treated Staphylococcus aureus accumulated small soluble nucleotide precursors of peptidoglycan (UDP-linked muramyl-peptides, the "Park nucleotides") instead of incorporating them into finished wall. This showed penicillin blocked a late, specific step of wall assembly rather than precursor synthesis generally. Subsequent biochemical work (Tipper, Strominger, and others in the 1960s) demonstrated directly that penicillin structurally resembles the D-Ala-D-Ala terminus of the pentapeptide — the "suicide substrate" hypothesis — and that purified transpeptidase enzymes become covalently and irreversibly acylated by penicillin, exactly as they would transiently be acylated by their true substrate during normal cross-linking. This structural mimicry, confirmed by X-ray crystallography of the β-lactam ring superimposed on D-Ala-D-Ala, is now textbook and directly explains why bacteria missing PBPs of appropriate affinity (or that alter them) become resistant.
Fleming's discovery and Chain/Florey's proof of therapeutic efficacy. Alexander Fleming observed in 1928 that a stray Penicillium mold contaminating a Staphylococcus culture plate created a zone where no bacterial colonies grew — the classic zone of inhibition that remains the basis of the disk-diffusion susceptibility test today. Fleming published the observation in 1929 but did not purify or test the substance therapeutically. It was Howard Florey and Ernst Chain (with Norman Heatley) at Oxford who, in 1940, purified enough penicillin to test it in vivo: in a landmark controlled experiment, they infected eight mice with lethal doses of Streptococcus, then treated four of them with penicillin. All four untreated mice died within hours; three of the four treated mice survived. This side-by-side controlled comparison — the direct forerunner of a modern efficacy trial — provided the first unambiguous evidence that a mold-derived compound could cure a lethal bacterial infection in a living animal, not merely inhibit bacteria on a plate. Fleming, Florey, and Chain shared the 1945 Nobel Prize in Physiology or Medicine.
The ribosome-target hypothesis — cell-free translation systems. In the 1950s–60s, using cell-free extracts that could translate synthetic mRNA (poly-U) into polypeptide (poly-phenylalanine) — the same systems Nirenberg and Matthaei used to crack the genetic code — researchers showed that streptomycin and tetracycline blocked this reaction specifically when bacterial ribosomes (not mammalian ribosomes) were used, and that the block could be localized to the 30S subunit by reconstituting hybrid ribosomes from bacterial and (where possible) altered subunits. Julian Davies' work in the 1960s further showed that specific point mutations in the ribosomal protein S12 (part of the 30S subunit) conferred streptomycin resistance by altering the drug's binding site directly — genetic proof that the 30S subunit, and this protein specifically, is the physical target.
Vancomycin's mode of binding — NMR and biochemical footprinting. Direct biochemical and later NMR structural studies (Perkins, and later Walsh and colleagues) showed vancomycin forms hydrogen bonds directly with the D-Ala-D-Ala dipeptide of peptidoglycan precursors — confirmed decisively when resistant Enterococcus strains were found (in the 1980s–90s) to synthesize precursors ending in D-Ala-D-lactate instead, which vancomycin binds roughly 1,000-fold more weakly because it loses one hydrogen bond. This resistance mechanism is itself powerful confirmatory evidence for the proposed binding mode, since altering exactly the predicted contact residue abolished drug activity in a quantitatively predictable way.
Quinolone mechanism — gyrase mutants and cleavage complex trapping. Gellert and colleagues (1976) discovered DNA gyrase and showed nalidixic acid (the prototype quinolone) inhibited its supercoiling activity in cell-free assays. Later work demonstrated that quinolones do not simply block the enzyme — they trap it mid-reaction, stabilizing the transient double-strand cleavage complex that gyrase normally forms and reseals. Adding protein denaturants to quinolone-treated reactions released permanent double-strand breaks, directly visualized on gels, proving the drug converts a reversible catalytic intermediate into a lethal lesion.
Worked examples
Example 1 — Predicting drug class from a resistance mutation. A clinical E. coli isolate is found to have acquired a mutation in the gyrA gene encoding the A subunit of DNA gyrase, at the codon corresponding to serine-83, changing it to leucine (a mutation in the so-called "quinolone resistance-determining region"). Reasoning: gyrase is the target of fluoroquinolones (ciprofloxacin, levofloxacin) in Gram-negative bacteria. A mutation in the drug-binding region of gyrase would be expected to reduce the drug's ability to trap the enzyme-DNA cleavage complex without necessarily destroying the enzyme's normal supercoiling function (which must be preserved, or the bacterium would be inviable). Prediction: this strain should show elevated minimum inhibitory concentration (MIC) to ciprofloxacin specifically, while remaining fully susceptible to unrelated drug classes such as β-lactams, aminoglycosides, or sulfonamides, since gyrA mutations do not alter PBPs, ribosomes, or folate enzymes. Clinically, if the reported MIC for ciprofloxacin rises from a susceptible 0.06 μg/mL to a resistant 8 μg/mL while the ceftriaxone MIC stays at 0.5 μg/mL, this pattern is exactly consistent with a target-site mutation restricted to gyrase — confirming target specificity is a testable, quantitative prediction, not just a qualitative story.
Example 2 — Bactericidal synergy calculation (a checkerboard/kill-curve scenario). A patient with enterococcal endocarditis is treated with ampicillin (a β-lactam) alone; despite an MIC that shows the organism is "susceptible" (MIC = 4 μg/mL, below the achievable serum concentration), the bacterial count in blood cultures falls only slowly and the drug is called bacteriostatic-like against this organism at achievable doses (a known property of enterococci, which have low-affinity PBPs). Adding gentamicin (an aminoglycoside) — even though gentamicin alone is nearly ineffective against enterococci because it cannot cross their cell envelope efficiently — produces a synergistic bactericidal effect. Mechanistic reasoning: ampicillin's inhibition of cell wall synthesis compromises the integrity of the enterococcal envelope, allowing gentamicin — normally excluded — to penetrate to the ribosome, where its bactericidal, concentration-dependent mechanism (mistranslation and membrane-damaging aberrant proteins) can proceed. Quantitatively, a synergy (checkerboard) assay might show the fractional inhibitory concentration (FIC) index: FIC index = (MIC of ampicillin in combination / MIC of ampicillin alone) + (MIC of gentamicin in combination / MIC of gentamicin alone) If ampicillin's combined MIC drops to 1/8 of its alone-MIC and gentamicin's combined MIC drops to 1/8 of its alone-MIC, FIC index = 0.125 + 0.125 = 0.25, well below the conventional synergy threshold of ≤ 0.5 — confirming true synergy, not mere additivity (which would require FIC index ≈ 1.0). This combination strategy — a cell-wall drug given specifically to let a ribosome-targeting drug in — is a direct clinical application of the layered-target logic developed above.
Common misconceptions & traps
| Misconception | Correction |
|---|---|
| "Antibiotics work on viruses too, or at least don't hurt to try." | Viruses have no cell wall, ribosomes, or independent metabolism for antibiotics to target; they are intracellular parasites of host machinery. Antibiotics are inert against them. |
| "Bactericidal drugs are always better than bacteriostatic ones." | Bacteriostatic drugs are often equally effective clinically because the immune system clears the growth-arrested population; bactericidal action matters most when host defenses are impaired (neutropenia, endocarditis). |
| "Broad-spectrum antibiotics are simply 'stronger.'" | Spectrum refers to how many bacterial types a drug affects, not its potency against any one organism; broad-spectrum use also causes more collateral damage to the microbiome. |
| "Penicillin kills all bacteria regardless of growth state." | β-lactams require active peptidoglycan synthesis; non-growing or dormant cells (persisters) are largely unaffected, which is why prolonged or combination therapy is often needed. |
| "Gram-negative bacteria are just 'more resistant' bacteria in general." | Their intrinsic resistance to many drugs (vancomycin, some macrolides) is a specific structural consequence of the outer membrane barrier and porin size exclusion, not a general toughness. |
| "Sulfa drugs work the same way in bacteria and humans, just weaker in us." | Humans lack the entire bacterial folate-synthesis pathway (we absorb folate from food); the target enzyme is simply absent, not just less sensitive — this is closer to perfect selective toxicity than most drug classes achieve. |
The scientist behind it
Alexander Fleming trained as a bacteriologist at St Mary's Hospital, London, and had already discovered the antibacterial enzyme lysozyme in tears and mucus (1922) before his 1928 penicillin observation — meaning he was primed by prior experience to recognize a zone of bacterial killing around a contaminant as biologically significant rather than a ruined experiment. He was, by his own account and by later historians' assessment, a meticulous but not especially charismatic bench scientist who lacked the chemical resources at the time to purify and scale up penicillin himself; that task fell a decade later to the Oxford team of Howard Florey (an Australian pathologist) and Ernst Chain (a German-Jewish biochemist refugee), whose systematic purification, animal testing, and — remarkably, during wartime Britain — extraction of penicillin from the urine of treated patients to conserve the scarce drug, turned a laboratory curiosity into a life-saving medicine reaching mass production by 1943–44 with American pharmaceutical partners.
Origins · 1 minSir Alexander Fleming and how this was discovered
Where this matters today
Understanding antibiotic mechanism is not academic trivia — it drives real prescribing and public-health decisions daily. Clinicians choose a β-lactam plus β-lactamase inhibitor (e.g., amoxicillin-clavulanate) specifically because the inhibitor protects the drug's active moiety from a bacterial enzyme that would otherwise hydrolyze the β-lactam ring before it reaches its PBP target. Combination therapy for tuberculosis (rifampin plus isoniazid plus others) exploits multiple independent targets — RNA polymerase and mycolic acid synthesis — specifically to reduce the probability that a single spontaneous resistance mutation lets the organism survive, since resistance to each drug arises independently and the joint probability of simultaneous resistance to both is the product of two small numbers. In agriculture, understanding which drug classes overlap with human-use antibiotics (versus classes reserved for animals) informs stewardship policy aimed at slowing the spread of resistance genes between animal and human microbiomes. In drug design, structural knowledge of the ribosomal exit tunnel and peptidyl transferase center has enabled rational design of newer oxazolidinones and ketolides to overcome existing macrolide resistance. And in diagnostic microbiology, the entire disk-diffusion and MIC testing framework used every day in clinical labs to guide antibiotic choice is a direct, unbroken methodological descendant of Fleming's original zone-of-inhibition observation.
Ask the tutor: How do bacteria's resistance mechanisms (efflux pumps, target mutation, drug-inactivating enzymes) map back onto each of these specific mechanisms of action?
Why don't antibiotics harm human cells if they block protein or DNA synthesis, which humans also need?
Because the specific molecular machines differ enough between bacteria and humans that the drug binds one but not the other. Bacterial ribosomes are 70S particles built from different rRNA sequences and proteins than the human cytoplasmic 80S ribosome, so aminoglycosides, tetracyclines, and macrolides bind bacterial ribosomal RNA/protein contacts that simply are not present in the human 80S structure. Similarly, bacterial DNA gyrase and RNA polymerase are structurally distinct enough from human topoisomerase II and RNA polymerase II that quinolones and rifampin bind the bacterial versions with much higher affinity at therapeutic doses. This is why side effects (like aminoglycoside effects on the structurally similar mitochondrial ribosome) tend to occur specifically where the bacterial-like target does exist in humans, proving the selectivity is structural, not magical.
What determines whether a drug is bactericidal or bacteriostatic?
It depends on whether blocking the target causes cell death directly or merely halts growth reversibly. Cell wall synthesis inhibitors are usually bactericidal because blocking cross-linking, combined with ongoing autolysin activity, causes irreversible osmotic lysis. Fluoroquinolones are bactericidal because they convert a transient gyrase-DNA cleavage intermediate into permanent double-strand breaks. Most protein synthesis inhibitors (tetracyclines, macrolides, chloramphenicol) are bacteriostatic because halting translation simply pauses growth without directly destroying the cell; the notable exception is aminoglycosides, which are bactericidal because mistranslated, misfolded membrane proteins actively damage the cell envelope. The distinction is drug- and even organism-specific, not an absolute law.
Why do some antibiotics work against both Gram-positive and Gram-negative bacteria while others work against only one group?
The outer membrane of Gram-negative bacteria is the deciding factor. It acts as a size- and charge-selective barrier studded with porins that admit only small, often hydrophilic molecules below roughly 600 daltons; bulky glycopeptides like vancomycin cannot pass through it at all, restricting them to Gram-positive organisms whose thick peptidoglycan is directly exposed. Broad-spectrum drugs like tetracyclines and fluoroquinolones are small enough, or specifically evolved/optimized, to cross porins and reach their intracellular targets in both groups. Some β-lactams (like ampicillin) were specifically modified with added side chains to improve porin penetration and extend their spectrum into Gram-negative territory, directly illustrating how structural chemistry, not just target biology, determines spectrum.
How does the discovery of the D-Ala-D-Ala mimicry mechanism explain vancomycin resistance in enterococci?
Once biochemists established that vancomycin's entire mechanism depends on hydrogen-bonding to the terminal D-Ala-D-Ala of the peptidoglycan precursor, resistance became predictable in principle: any change to that terminus that preserves transpeptidase function but removes a vancomycin contact point would confer resistance. This is precisely what vancomycin-resistant enterococci (VRE) evolved — acquisition of genes (the van gene cluster) that reprogram precursor synthesis to end in D-Ala-D-lactate, removing one hydrogen bond donor and reducing binding affinity roughly a thousandfold, while the altered precursor is still an adequate transpeptidase substrate. The discovery validated the proposed mechanism precisely because it predicted, correctly, which single chemical change would break drug binding without breaking the cell wall itself.
Why is it important that antibiotics were originally natural products made by other microorganisms?
Fungi like Penicillium and soil bacteria like Streptomyces (the source of streptomycin, tetracyclines, and many others, discovered via Selman Waksman's systematic soil-screening program in the 1940s) evolved these compounds over millions of years as chemical weapons in microbial competition for resources in soil and other crowded microbial communities. This means the targets these molecules hit were already validated, essential, and vulnerable — nature had already done the "drug discovery" work of finding molecules that bind bacterial machinery with high affinity. It also means resistance mechanisms (efflux pumps, modifying enzymes) frequently already existed in the environment before clinical use began, since producing organisms need self-resistance genes and neighboring soil bacteria faced the same selective pressure for eons, which is directly relevant to understanding the ancient environmental reservoirs of resistance genes.
If a drug's target is bacteria-specific, why do side effects still occur?
Selectivity is rarely absolute. Aminoglycosides damage the inner ear and kidneys because mitochondria, being of bacterial evolutionary origin, retain ribosomes closer in structure to the bacterial 70S type than to the cytoplasmic 80S ribosome, so at sufficient drug concentration mitochondrial protein synthesis is also disrupted in metabolically active tissues like the cochlea and renal tubule. Polymyxins damage host cell membranes to some degree because the electrostatic selectivity for anionic bacterial lipids over zwitterionic human lipids is real but imperfect. Broad-spectrum antibiotics also disrupt the human gut microbiome, which is composed of bacteria sharing the very targets the drug was designed to hit, causing dysbiosis and complications like Clostridioides difficile overgrowth. Selective toxicity is a matter of degree and dose, not a binary switch.
How does understanding antibiotic mechanism connect to the coming problem of antibiotic resistance?
Every mechanism described here defines exactly the molecular change a resistant mutant or horizontally acquired gene must make to survive: altering the target (mutated PBPs, altered gyrase, modified ribosomal proteins), destroying the drug (β-lactamases), pumping the drug out (efflux pumps), or bypassing the blocked pathway (acquiring a drug-insensitive enzyme, as in some sulfonamide-resistant folate pathways). Because resistance is target-specific, a bacterium resistant to one class (say, a β-lactam via altered PBPs) remains fully susceptible to unrelated classes (aminoglycosides, quinolones) unless it separately acquires resistance genes for those too — which is exactly why multidrug resistance requires accumulation of multiple independent mechanisms, and why understanding the mechanism of each drug class is the essential foundation for understanding how resistance mechanisms, in turn, evolve and spread.