Bactrim, Cyclospora, and the pathway you don’t have
“Antibiotic” is a clinical filing category, not a biochemical one. Humans sort drugs by the kind of organism we point them at — antibacterial, antifungal, antiparasitic — but drugs don’t act on organisms. They act on molecules. And a molecule that appears in bacteria may also appear in protozoa, in fungi, and in plants, while being entirely absent from humans.
Trimethoprim-sulfamethoxazole is not really an antibacterial that happens to work on a parasite. It is an antifolate that happens to work on anything which must build its own folate. Bacteria qualify. So does Cyclospora cayetanensis.
Part One: The mechanism
The vulnerability is a pathway humans have lost
Folate — vitamin B9 — is required by every living cell. Its reduced form, tetrahydrofolate, is the carrier for one-carbon units in the reactions that build thymidylate, purines, methionine, and glycine. Block it and DNA synthesis stops. The cell dies in what the literature calls, memorably, “thymine-less death” (Cold Spring Harbor Perspectives review of antibacterial antifolates, PMC4968165).
Here is the exploitable asymmetry: humans cannot make folate. We lost the biosynthetic machinery and became dependent on dietary intake, which is precisely why it is classified as a vitamin. We take folate in preformed and transport it into cells.
Plants, most bacteria, and many unicellular eukaryotes retain the biosynthetic pathway. Some parasitic protozoa, including Plasmodium and Toxoplasma, can both synthesize folate de novo and salvage it from host plasma — for reasons that remain unclear (Trends review on alternative folate biosynthesis, PMC2720532).
An organism that must run its own folate factory has a target. A human who buys folate retail does not.
Two drugs, two consecutive steps
Bactrim is a fixed combination, and the combination is the point.
Sulfamethoxazole is a sulfonamide — a structural mimic of para-aminobenzoic acid (PABA). It competitively inhibits dihydropteroate synthase (DHPS), the enzyme that condenses PABA into the growing pterin molecule. Sulfonamides act as PABA analogues and either inhibit or act as suicide substrates in the folate biosynthetic pathway (ScienceDirect topic review on Cyclospora cayetanensis).
Human selectivity here is absolute rather than a matter of degree: humans do not possess DHPS at all. There is no human enzyme for sulfamethoxazole to inhibit. Whatever toxicity sulfa drugs carry — and they carry real toxicity, particularly hypersensitivity — it does not come from on-target folate inhibition in human cells.
Trimethoprim is a 2,4-diaminopyrimidine. It inhibits dihydrofolate reductase (DHFR), the next enzyme along, which reduces dihydrofolate to the active tetrahydrofolate. Humans do have DHFR, so selectivity here is a matter of binding affinity — and the margin is enormous. Trimethoprim has roughly 10⁵-fold higher affinity for prokaryotic than eukaryotic DHFR (analysis of DHFR in Rickettsiales, PMC11765544). We also have the escape hatch of dietary folate; the parasite does not.
The two drugs hit sequential steps in a single pathway, which is why the combination is synergistic rather than merely additive. Pyrimethamine with a sulfonamide works the same way against Toxoplasma, providing inhibition of two sequential steps in folate metabolism (ScienceDirect review on pyrimethamine).
Why this class crosses kingdoms
Once you see the target rather than the taxonomy, the drug family’s scattered clinical applications snap into a single picture. The DHFR inhibitors used against parasites include pyrimethamine, trimethoprim, proguanil, and trimetrexate (ScienceDirect, Cyclospora topic review). Methotrexate — a DHFR inhibitor — treats cancer. Dapsone, a DHPS inhibitor, treats leprosy and malaria. And trimethoprim-sulfamethoxazole is the treatment for Pneumocystis jirovecii pneumonia, caused by an atypical fungus (Candida DHFR study, PMC7316490).
One pathway. Bacteria, protozoa, fungi, and human cancer cells. The clinical labels were always downstream of the biochemistry.
Why Cyclospora in particular
Cyclospora cayetanensis is an apicomplexan coccidian, and genome sequencing established that its organization, metabolic capabilities, and invasion machinery closely resemble those of Eimeria tenella (comparative genomics study, PMC4851813). Its relatives — Toxoplasma gondii, Cystoisospora belli, Plasmodium — are all antifolate-susceptible. It sits inside a clade defined, pharmacologically, by this vulnerability.
And the exception proves the rule. Cryptosporidium parvum is a coccidian too, superficially similar, similar oocyst, similar clinical picture. It is refractory to conventional therapy including TMP-SMX (Springer review on treating intestinal protozoa). Same drug, same drug class, adjacent organism, no effect. Whatever Cryptosporidium is doing with folate, it isn’t doing the thing that trimethoprim interrupts. That single failure is strong evidence that the drug’s success against Cyclospora is genuinely about folate metabolism rather than some nonspecific antiparasitic property.
Part Two: How it was discovered
There are two discovery stories here, separated by sixty years, and they were made in opposite directions.
The drug: from a red dye to rational design
1932–35 — Prontosil. Gerhard Domagk at Bayer, screening synthetic dyes for antibacterial activity, found that Prontosil protected mice against streptococcal infection. The earliest antibacterial discovery programs centered on activity observed in synthetic dyes (PMC4968165). Domagk received the 1939 Nobel Prize, though the Nazi government forced him to decline it; he accepted the medal in 1947.
Prontosil had a strange property: it worked in animals but not in the test tube. By 1937 it became clear the compound was broken down in the body to release sulfanilamide, which was the actual active agent (Pharmacology review of sulfonamides). Prontosil was a prodrug. Sulfanilamide itself was already off-patent, and the sulfa era opened.
1940 — The antimetabolite principle. Donald Woods and Paul Fildes independently published the idea that some substances block cellular replication by interfering with metabolites required for normal function — competing with, replacing, or blocking the true metabolite (Embryo Project Encyclopedia, Hitchings entry). Woods showed that sulfonamides, being structural analogues of PABA, competitively block the incorporation of PABA into folate (Pharmacology review).
This is the intellectual hinge of the whole story. For the first time, a drug’s action was understood as rational metabolic antagonism rather than empirical poison. The antimetabolite theory turned out to apply not just to sulfonamides but to trimethoprim and to the anticancer agents 5-fluorouracil, 6-mercaptopurine, and methotrexate (Basicmedical Key, sulfonamide chapter).
1940s–60s — Hitchings and Elion. George Hitchings and Gertrude Elion at Burroughs Wellcome built an entire research program on the principle, studying purines and pyrimidines on the rationale that interfering with nucleic acid metabolism would selectively harm rapidly dividing pathogens (Basicmedical Key, trimethoprim chapter). During the 1940s they determined that substituted 2,4-diaminopyrimidines interfere with folate metabolism (PMC4968165).
Their output is remarkable: pyrimethamine (antimalarial), trimethoprim (antibacterial), 6-mercaptopurine (leukemia), allopurinol (gout), and later acyclovir (herpes). Trimethoprim emerged from the diaminopyrimidine series in the mid-1950s (Springer, Trimethoprim and Pyrimethamine). Hitchings and Elion shared the 1988 Nobel Prize with James Black for establishing principles of rational drug treatment.
Note what this means for our question. Pyrimethamine — an antiparasitic — and trimethoprim — an antibacterial — came out of the same chemical series aimed at the same enzyme. The cross-kingdom activity wasn’t a later surprise. It was baked in at the bench.
~1968 — The combination. Trimethoprim was paired with sulfamethoxazole, chosen partly for compatible pharmacokinetics, and marketed as Bactrim and Septra.
The parasite: fifteen years of not knowing what it was
Cyclospora’s story runs the other way — from clinical mystery backward to identity, and only then to treatment.
1979 — R.W. Ashford described an undescribed coccidian in humans in Papua New Guinea. It was thought to be an Isospora species and received little attention (CDC Emerging Infectious Diseases, “Cyclospora: An Enigma Worth Unraveling,” 1999).
1985–1991 — The “cyanobacterium-like bodies” period. The organism resurfaced in New York and Peru, then in AIDS patients (Hart et al., Lancet, 1990) and in outbreaks in Chicago and Nepal, where it was reported as “cyanobacteria (blue-green algae)-like bodies” (Kocka et al., MMWR, 1991). Clinicians were seeing something under the microscope that looked vaguely algal and could not classify it. Under phase contrast the oocyst has an algal-like morula appearance (CDC EID, 1999) — hence the confusion. Its recognition owed a good deal to the increased use of staining methods introduced for detecting Cryptosporidium (CDC EID, 1999).
1993–94 — Identification. Ynés Ortega, Charles Sterling, Robert Gilman, Vitaliano Cama, and Fernando Díaz established it as a novel coccidian in the New England Journal of Medicine (1993), followed by the formal species description in the Journal of Parasitology (1994). The species name honors Universidad Peruana Cayetano Heredia in Lima. In parallel, Bendall and colleagues in the UK characterized the same organism as a new coccidian enteritis of man (Lancet, 1993). Molecular phylogenetics later confirmed close kinship with Eimeria (Relman et al., 1996).
This classification is the pivot point. The moment the organism was established as a coccidian rather than an alga, an obvious therapeutic hypothesis existed: coccidia respond to antifolates. Cystoisospora belli did. Toxoplasma did. Physicians had a taxonomic reason to reach for a drug already sitting on the shelf.
1994–95 — Confirmation. A case report suggested co-trimoxazole might work (noted in the abstract of Hoge et al., 1995). Pape and colleagues documented treatment and secondary prophylaxis in HIV-positive adults in Haiti (Annals of Internal Medicine, 1994).
Then the definitive trial: Hoge et al., Lancet, March 1995. Expatriates in Kathmandu with diarrhea and confirmed Cyclospora in stool were randomized, double-blind, to co-trimoxazole (160 mg trimethoprim / 800 mg sulfamethoxazole) twice daily for seven days, or placebo. Of 40 patients, 21 received drug and 19 placebo. After seven days, Cyclospora was detectable in 1 of 16 treated patients who submitted specimens, versus 15 of 17 on placebo (p<0.0001). Eradication tracked clinical improvement, and no relapses occurred over a further week of follow-up (Hoge et al., Lancet 1995;345:691–3).
Forty patients. That was enough, because the effect size was enormous and the prior probability — a coccidian, an antifolate — was already high.
The honest gap
There is a loose end worth naming, and it connects back to a theme from the on-farm piece.
Cyclospora cayetanensis still cannot be cultured in the laboratory, and no animal model exists — which is why the discovery of alternative treatments remains hampered (Springer review on treating intestinal protozoa). The 2016 genome paper was undertaken precisely because outbreak investigation had been held back by a lack of genetic data and poor understanding of the organism’s biology (PMC4851813).
So consider what actually underwrites the claim that Bactrim kills Cyclospora by blocking folate synthesis:
- The organism is a coccidian, genomically close to Eimeria — inference from taxonomy and sequence.
- Coccidia possess de novo folate biosynthesis and are antifolate-susceptible — inference from relatives.
- Patients given TMP-SMX clear the parasite and get better — direct clinical evidence.
What is missing is the middle link done directly: cultured Cyclospora, its DHFR isolated, inhibition constants measured against trimethoprim. You cannot easily run that experiment on an organism you cannot grow. The mechanism is very probably exactly as described — the biochemistry is coherent, the Cryptosporidium contrast is informative, and the clinical effect is unambiguous — but it rests on a triangulation rather than a demonstration.
That is not unusual in medicine. It is worth knowing about anyway, because it is the same methodological bottleneck that leaves growers without a validated on-farm control protocol and leaves investigators with a detection assay that misses most contaminated samples. An organism you cannot culture is an organism you keep having to reason about indirectly.
A note on the clinical side
Current reviews report TMP-SMX at 160/800 mg twice daily for 7–10 days as effective, with low recurrence, and as the treatment of choice in HIV-positive patients as well. Ciprofloxacin is less effective but is used for patients intolerant of sulfonamides; nitazoxanide has been reported successful in only small numbers of patients (Cambridge review of C. cayetanensis, PMC10317703; Verdier et al., Annals of Internal Medicine, 2000).
This is background on how the drug works, not guidance on taking it. TMP-SMX is prescription-only, carries meaningful hypersensitivity and interaction risks, and sulfa allergy is common. Anyone who suspects cyclosporiasis — particularly with an active outbreak underway — should be evaluated by a clinician, and should specifically request Cyclospora testing, since it is not detected on routine stool ova-and-parasite examination unless sought.
