16S rRNA: Structure, Variation, Phylogenetics, and Resistance

Every bacterium carries the same basic molecule at the heart of its protein-making machinery, a strand of RNA called 16S. Because all bacteria have it but each lineage’s version differs a little, it has become biology’s most widely used barcode for naming bacteria and drawing the tree of life.¹ The same molecule is also a target for antibiotics, so it sits where identifying germs and fighting them meet. Its usefulness has limits, though: a barcode that changes slowly can blur closely related species together.²

What 16S rRNA Is and Where It Sits in the Cell

A ribosome is the cell’s protein-building machine. It reads genetic instructions and strings together proteins, and bacteria build it from two halves. The small half, called the 30S subunit, is built around a single RNA strand about 1,500 letters long. That strand is 16S rRNA, which from here on is just 16S. About twenty proteins attach to it as the small half is assembled.³

The small half holds and reads the message, while the large half joins the building blocks together. 16S is less a passive scaffold than an active part that helps check each reading. Humans have ribosomes too, but ours differ enough that some antibiotics hit bacteria and spare us.

The way 16S folds onto itself is now mapped, with about 90% of its paired letters confirmed by comparing many species.⁴ Some proteins grip it in more than one place, such as one that binds two separate regions at once.⁵ Proteins binding the RNA nudge its shape, which helps the whole particle assemble step by step and fine-tunes how it moves.⁶

A common mix-up is gene versus molecule. The 16S gene is the DNA instruction, and 16S RNA is the working copy made from it. Lab tests usually read the gene, which is why reports say “16S gene sequencing.”

The “S” in 16S is odd too. It stands for Svedberg, a unit describing how fast a particle sinks when spun at high speed. Because that depends on shape as well as weight, the numbers don’t add up like weights: the 30S and 50S halves join to form a 70S particle, not an 80S one.

Why 16S Works as a Bacterial Barcode

Because 16S must keep its shape to work, much of its sequence is nearly identical across bacteria. Labs design short matching “grab” sequences, called primers, against these steady stretches, and some match both bacteria and archaea.⁷ Because primers stick to these steady stretches, they can latch onto 16S from bacteria the lab has never seen. The steady stretches act as handles on both ends of the gene.

Between those handles lie nine variable regions, labelled V1 to V9, where the letters differ between lineages. They get copied along with the handles. So one test can pick up unknown bacteria and then tell them apart by their variable regions.

In practice, DNA is pulled out of a sample such as stool, a skin swab, or fluid. The 16S gene, or part of it, is copied millions of times using the primers. A sequencing machine reads the copies, and each read is matched against a reference library of known bacteria.

Many common methods read only a short stretch such as V3–V4 rather than the full gene. Which region is read changes the result, and in airway samples V1–V2 identified bacteria best.⁸ For skin samples, V1–V3 came closest to reading the whole gene.⁹ Reading the full gene gives a sharper picture than any one region.¹⁰

No single region is best, because different bacteria differ in different variable regions. Newer long-read machines can read the whole gene, which removes the choice.

Where 16S Identification Falls Short

Across 15 major bacterial groups, more than 175 pairs of clearly separate species carried essentially identical 16S.¹¹ A 16S match can therefore point to a genus while leaving the species uncertain. Even the full gene struggles to tell apart strains and species in some plant-associated bacteria.¹² Still, full-length reading did correct some misidentifications that the short V3–V4 read made.¹³

“Identical” 16S doesn’t mean identical bacteria. Two species can share 16S yet differ greatly in the rest of their DNA, and one may cause disease while its look-alike is harmless. Telling them apart takes other tests, such as whole-genome reading.

A species name in a report is often the best match the software found, not a confirmed identification, and a close relative with a near-identical sequence could be the real one. Some software reports only down to genus for exactly this reason.

Every read is named by comparing it to a reference library, so the result is only as good as that library.¹⁴ Two widely used libraries have not been updated since 2013 and 2016.¹⁵ When known bacteria were run through common libraries, roughly 1 in 10 to nearly 1 in 5 were misnamed.¹⁶ The same sample can get different names depending on the library a lab uses.

For a microbiome report, that means trusting genus names more than species names. Results from different companies may not match even for the same sample, so if you’re comparing results, ask which region and library were used.

Copies, Counts, and Variation Within One Bacterium

Bacteria carry anywhere from 1 to 37 copies of the 16S gene, and archaea 1 to 5.¹⁷ Before sequencing, the gene is copied millions of times from whatever DNA is present, and a cell with several copies contributes several times as much starting material. A bacterium with many copies therefore looks more common than it really is.¹⁸ The method counts copies of a gene rather than cells, so two species present in equal numbers can produce very different numbers of reads.

In simulations, the most abundant microbe was picked correctly only about 13% of the time from raw counts.¹⁹ Correcting for copy number raised that to about 65% when close reference genomes existed.²⁰ But correction tools are often inaccurate for bacteria without close known relatives.²¹ Copy numbers are known only for bacteria whose full genomes have been read, so unknown bacteria are guessed from relatives, and the guesses get worse the more distant the relative.

That affects the “percent” numbers in a microbiome report. A “30% of your gut” figure usually means 30% of gene reads, not of cells. Numbers are best for comparing samples run the same way over time, and a big swing in one group may reflect method, not your body.

Copies inside one cell can also differ. About 60% of bacterial genomes carry 16S copies that differ slightly from each other, mostly by under 1%.²² One bacterium can then be counted as several different types, inflating apparent diversity.²³ E. coli, with 7 copies, needed a fairly loose matching rule to be kept as one type.²⁴

How Antibiotics Attack 16S

Aminoglycosides, a group that includes gentamicin, tobramycin, and amikacin, bind a pocket in 16S at the ribosome’s reading site.²⁵ There they hold two key letters flipped outward, the position used when checking each genetic word.²⁶ During protein building, the ribosome checks each three-letter genetic word against the incoming building block, and 16S helps confirm they fit. Aminoglycosides disturb this check, so the ribosome accepts wrong building blocks and makes faulty proteins.²⁷ Those proteins can damage the cell, including its membrane, which helps explain why aminoglycosides kill bacteria rather than only slow their growth.

Spectinomycin grabs two specific letters in a different part of 16S.²⁸ It may lock the molecule in an inactive shape.²⁹ In living cells it barely disturbs its own pocket but over-stiffens contacts reaching across the ribosome.³⁰ It stops the ribosome’s moving parts rather than causing misreading.

These drugs largely spare human cells because human ribosomes differ in the exact letters the drugs grab. Aminoglycosides still carry known risks to hearing and kidneys, which doctors monitor.

How Bacteria Become Resistant Through 16S

Changing a single letter in the drug’s pocket can block aminoglycosides from binding.³¹ In E. coli, certain changes gave resistance to nine different aminoglycosides at once.³² In Campylobacter, a common cause of food poisoning, one letter change raised the dose needed for kanamycin over 64-fold.³³

Because many bacteria carry several 16S copies, they make ribosomes from all of them, so a change in one copy leaves the other copies’ ribosomes vulnerable. The drug can still bind those and halt the cell, so resistance only becomes strong when the change reaches most or all copies. Species with fewer copies can acquire resistance this way more easily.

Some bacteria carry enzymes that attach a small chemical tag to 16S, preventing the drug from binding.³⁴ Those enzymes are the bigger clinical threat. They now drive high-level aminoglycoside resistance in hospital bacteria.³⁵ In one country’s hospital Klebsiella samples, about 1 in 5 had high-level resistance to amikacin and gentamicin, all carrying such enzymes.³⁶

The enzymes spread easily. Their genes can sit on mobile DNA passed between bacteria, and they protect every 16S copy at once, unlike a single-letter change.

Bacteria also add their own tags to 16S as a normal part of life, and removing them in lab E. coli made it more sensitive to several antibiotics.³⁷ Removing one tag, though, made it more resistant to kasugamycin.³⁸

For a patient, lab testing shows which drugs a bacterium resists before treatment is chosen. Resistance to one aminoglycoside often means resistance to others, so if testing shows that, doctors switch to a different drug family that targets another part of the cell.

16S Tests for Hard-to-Diagnose Infections

Sometimes a sample from a suspected infection grows nothing in the lab. This is called a culture-negative infection. Standard culture needs living bacteria that can grow on a dish, and antibiotics already in the patient’s body may have killed or weakened them. Some bacteria are also slow or fussy about what they grow on. A 16S test looks for bacterial DNA rather than living cells, so it can find traces of bacteria that no longer grow.

In such infections the test found the cause in about 43% of cases, and when positive it was right every time.³⁹ About 1 in 5 negative results still came from patients who were infected.⁴⁰ Newer portable sequencing makes these tests faster and cheaper for routine hospital use.⁴¹

If your doctor orders one, it is usually run on fluid or tissue from joints, heart valves, bone, or spinal fluid. Results may take days because samples often go to a specialist lab. A positive names the likely germ so treatment can be targeted. A negative is weighed with symptoms and scans, not taken alone.

The test can miss infections for a few reasons. Too little bacterial DNA may be left in the sample, mixed infections blur the signal, and the test doesn’t show which antibiotics will work.

16S and the Tree of Life

Comparing ribosomal RNA across organisms showed that life splits into three domains: bacteria, archaea, and organisms with nuclei like us.⁴² This overturned the old two-way split between simple and complex cells.⁴³

A slow-changing molecule makes a good family-tree marker. It changes slowly enough to compare even very distant life forms, and its steady parts let scientists line up sequences reliably. That same slowness is why it fails for close relatives.

The picture has wrinkles. Some bacteria hold a 16S copy that differs by more than 10% from their other copies and looks borrowed from a relative.⁴⁴ In Enterobacter, a group that includes some hospital germs, 16S genes have been reshuffled between genomes at least five times.⁴⁵ In one archaeon, different 16S copies suit different temperatures.⁴⁶

Heat-loving microbes leave a trace in 16S as well. Those from extreme heat carry more tightly knotted structures in the molecule, which likely help the RNA hold its shape at high temperatures. Cold-loving microbes lack the most stable knots.