Reverse transcriptase inhibitors: NRTIs vs NNRTIs

Oct. 24, 2019
“The single biggest threat to man’s continued dominance on this planet is the virus.” – Dr. Joshua Lederberg, Nobel Prize for Physiology or Medicine, 1958

Whether you agree with Dr. Lederberg’s famous quote or think that in the intervening 60 -odd years we’ve discovered (or created) even bigger threats to our survival as a species, there’s little debate that viral diseases are a serious health problem. On top of the ones we already recognize as infecting humans and causing disease, new ones are constantly being added to this list—either newly recognized or actually newly emerging.

For bacterial and fungal pathogens, we can often identify and take advantage of cellular or biochemical differences between human and microorganism to create targeted drugs—antibiotics—which can selectively inhibit or kill microorganisms with limited toxicity to the host.

By nature, however, viruses are genetically minimalist, acting solely as an intracellular parasite and often having only a very few of its own unique enzymes or pathways distinct from those of the cell it infects. This provides less opportunities to target selective pharmacological treatments against, and indeed in lay parlance it’s usually described as “there are no antibiotics for viruses.”

Ignoring any semantic argument that viruses aren’t alive and thus "antibiotic" would be the wrong term in any case, it’s true that there really are no general, broad-spectrum, antiviral drugs. Even innate responses such as interferons and Protein Kinase R (PKR) are relatively narrow with regard to what viruses they work against and can have significant off-target deleterious responses.

Against one broad (and important) class of viruses, however, we do have a readily differentiable target to exploit. In this month’s column, we’re going to look at that viral group (retroviruses), what this convenient target of opportunity is (reverse transcriptase) and what the two major classes of drugs attacking this target are (NRTIs and NNRTIs).


Retroviruses are characterized by having infectious particles carrying RNA genomes, which replicate by a unique mechanism. Contrary to the Central Dogma (DNA begets RNA, which codes for proteins), this family of viruses takes its RNA genome and uses a self-encoded enzymatic function called a Reverse transcriptase (often shortened to RT, not to be confused with RT for Real Time in PCR) to make a single-stranded DNA copy of the RNA genome. This unusual enzyme is synthesized by the host cell ribosomal machinery, utilizing the infecting RNA as an mRNA. Once produced, the RT does what its name implies: something exactly analogous to RNA transcription with the same Watson-Crick base pairing rules applying, but it’s “reverse” in that the viral RNA acts as the template strand and a single DNA complement is synthesized. Since this step is the key to our topic, we’ll simplify the subsequent steps in the retroviral replication cycle to just say this single-stranded DNA gets converted to double-stranded DNA, then (usually) gets inserted (“integrated”) somewhere in the host genome, where normal cell division processes will replicate it at every cell division, passing this integrated virus to daughter cells. The integrated virus in turn can be transcribed both to generate any virus-specific coding regions such as capsid proteins or integrase enzymes, and transcribed as full length viral RNA genomes, which package in their capsid proteins, leave the cell and begin the cycle anew in a new host cell.

The key here is that our cells follow the Central Dogma, and we wouldn’t expect any uninfected cell to express an RT enzymatic function. (Aside: as with many topics in this column, biology is incredibly diverse and absolute blanket statements often have exceptions. Trace levels of RT activity can sometimes be detected in some cells, but it’s thought these all arise from ancestral integrated retroviruses or retrotransposons; this isn’t a part of normal cellular activity, and for most practical purposes, we can act as though any significant RT activity will occur only in those cells with ongoing retroviral infection).

The second key point is that this RT function is essential to the propagation of the retrovirus. In combination, we’re at least in theory presented with a perfect Achilles’ heel by which to treat retrovirus infection. If we could design a drug which is widely taken up by all our cells, and specifically and effectively blocks RT activity, then we can imagine how such a drug would both be well tolerated (no impact on uninfected cells) and yet specifically block retroviral propagation in infected cells. In fact, two general classes of drugs have been developed and are in widespread use with exactly this intent, but as we’ll see it’s not a perfect world and while these are incredibly useful therapeutic agents, they’re sadly not retroviral miracle cures.


The first class of these drugs is the Nucleoside Reverse Transcriptase Inhibitors (NRTIs). To understand how these work, consider the mechanism of RT action. The enzyme ‘grips’ the RNA template strand hydrogen bonded to the 3’ end of the growing DNA reverse transcript and allows 2’-deoxynucleoside triphosphates (dATP, dGTP, dCTP, or dTTP) to diffuse into the active site.

When one of these has appropriate Watson-Crick base pairing to the RNA base in the active site, the enzyme catalyzes the nucleophilic attack by the growing DNA strand’s 3’ -OH on the a phosphate of the incoming dNTP. The b-g pyrophosphate group is displaced (providing a thermodynamically linked driving energy for the reaction), while the growing DNA chain is now longer by one nucleoside monophosphate. Its 3’ -OH is now the growing end of the DNA, and the enzyme slides down one nucleotide long the template to line this new end up for another cycle of the same process. Now, imagine if instead of natural nucleoside triphosphates, you could make a molecule which “looks”—or more correctly, “feels,” since the interaction is based on physical contacts—like one or more of dATP, dGTP, dCTP, or dTTP, but was lacking the 3’ -OH. 

Such a molecule would have potential to be incorporated into a growing DNA strand, but it would be what’s called a chain terminator. That is, there’s no way to grow it further; replication of the strand is blocked. That’s literally the end of the line for whatever’s being replicated, so a natural concern should be whether such a drug would do the same to our own cellular DNA polymerases with undesirable side effects. The short answer is yes, they might, but cellular DNA polymerases (working from a DNA template) and viral RTs (working from an RNA template) have enough differences in their enzyme active sites to make it possible to develop NRTIs which don’t bind well to host DNA polymerases but are pretty effective decoys for viral RTs. The more effective this selection is, the less side effects these drugs have and the more efficient they are at blocking retroviral replication.

While there are additional nuances to this—some drugs of this type are administered as non-phosphorylated prodrugs which are activated by intracellular kinases, while others are administered in a phosphorylated state—the differences relate more to biological stability and uptake than to core mechanism of action. NRTIs are as a group any of these molecules which mimic a dNTP and selectively fool the RT into incorporating a chain terminator in its product; at least 10 are in current clinical use.

Unfortunately, not all retrovirus RT enzymes are highly conserved, so while there is some spectrum of activity of these drugs, they tend to work best when screened or developed against specific retroviruses —  they’re not a magic bullet against all retroviruses. Further, retroviruses tend to have ‘sloppy’ replication; that is, the RT enzymes are not very accurate, leading to high mutation rates. This is in effect a rapid evolutionary strategy common to most RNA-based viruses, allowing a single infecting particle to create a quasi-species swarm of variants in a host; any of these better adapted to their host environment than their peers becomes the dominant population. This high mutation rate equates to an ability to develop and select resistance to NRTIs.

In the case of HIV, for example, there are six amino acid residues in the RT whose mutation can commonly lead to at least reduced resistance to one or more NRTIs, mostly by selectively reducing chain incorporation efficiency of NRTIs compared to natural dNTPs. Other mutations can lead to an ability for the RT to effectively “stall” over the incorporated chain terminator and drive the reverse reaction, in essence recapitulating what’s referred to as a proofreading function in cellular DNA polymerases.


If nucleotide decoy molecules aren’t a perfect way to exploit this biochemical difference between disease and host, are there other ways to selectively interfere with the same step? There are, and not surprisingly they’re referred to as Non-Nucleoside Reverse Transcriptase Inhibitors (NNRTIs). In essence, these are allosteric inhibitors of the RT, binding at a site on the RT enzyme and inducing conformational changes which reduce its activity and thus, rate of viral replication. As mentioned above, however, RTs from different retroviruses are each unique so even more so than NRTIs, NNRTIs would only reasonably be expected to work against the particular virus RT for which they’re developed.

HIV has been the driving need in this space, and at present at least five different anti-HIV NNRTIs are available. All of these at present work by binding in the same general area of the HIV RT enzyme; unfortunately, this means that particular mutations in the RT, such as ones that effectively block this allosteric binding pocket, can confer resistance to this entire class of drugs. On the other hand, since NNRTIs are specific to the viral RT and not expected to interact with host polymerases such as NRTIs might, they have somewhat less potential for side effects than NRTIs.

Combinations – the HAART origin

Since NRTIs and NNRTIs work through different mechanisms on the same target, combination therapies, where one or more drugs of each class are administered simultaneously, should be expected to—and do—provide a synergistic effect, not only through combined reduction in net RT activity, but also by making mutational escape from inhibition dramatically less likely. It’s possible that a single mutation in RT can confer some level of resistance to either a specific NRTI or NNRTI, but for an RT to be resistant to a multidrug cocktail containing both, that RT must simultaneously gain both mutations.

That’s suddenly a vastly harder statistical challenge to meet, and the basis for what’s more generally known in HIV treatment circles as Highly Active Antiretroviral Therapy (HAART). In practice, HAART strategies can include not just mixtures of NRTIs and NNRTIs, but drugs against other steps in the HIV cycle, such as the viral integrase and a specific viral protease.

A basic understanding of the mechanism and function of NRTIs and NNRTIs is not only of direct relevance to a major medical issue facing the world today—HIV infection—but also a simple vignette of how therapeutic responses to novel pathogens are often rooted in an understanding and leveraging of a subtle yet differentiable biochemical path between host and pathogen.