Tagatose is probably not metabolized by human cells

[Devlin-Moyer]

Current behavior:

Human-GEM currently has this short pathway for metabolizing tagatose:

ID	Reaction	Genes
MAR09417	D-tagatose [e] ⇔	None
MAR08760	D-tagatose [e] ⇒ D-tagatose [c]	None
MAR01472	ATP [c] + D-tagatose [c] ⇒ ADP [c] + D-Tagatose 1-Phosphate [c] + H+ [c]	KHK
MAR01038	D-Tagatose 1-Phosphate [c] ⇔ DHAP [c] + glyceraldehyde [c]	ALDOB
MAR08761	ATP [c] + D-tagatose [c] ⇒ ADP [c] + D-tagatose-6-phosphate [c] + H+ [c]	KHK
MAR08762	D-tagatose-6-phosphate [c] ⇔ DHAP [c] + glyceraldehyde [c]	ALDOA or ALDOB or ALDOC
MAR04774	CDP [c] + H+ [c] + tagatose-1,6-bisphosphate [c] ⇔ CTP [c] + D-tagatose-6-phosphate [c]	PFKP or PFKL or PFKM
MAR04775	D-tagatose-6-phosphate [c] + ITP [c] ⇔ H+ [c] + IDP [c] + tagatose-1,6-bisphosphate [c]	PFKP or PFKL or PFKM

I noticed this because I was looking for pairs of nearly-identical reactions that might form internal loops, and the last two do exactly that. They're also a dead-end; no other reactions involve tagatose-1,6-bisphosphate. The fact that they're both reversible also seemed suspicious, and reminded me of #547, so I tried looking to see if there was any evidence that they should or shouldn't be reversible.

I checked the references for MAR01472, MAR04774 and MAR04775 (the only reactions in the above table that had references), and none of them were about human enzymes. I found this paper that found that administering tagatose to cultured human kidney cells induced production of MCP-1 (a protein, not a metabolite), and while they suggested this was because it was phosphorylated by KHK, they did not directly demonstrate this, instead citing a paper that showed that tagatose is phosphorylated by KHK in rat hepatocytes. I also found this paper that says (in the second paragraph under "Metabolism" subheading under the "Tagatose" main heading):

Higher animals, including humans, have no tagatose-6-phosphate pathway, and tagatose does not break down spontaneously. [...] Tagatose is also a malabsorption sugar for humans, meaning it is a non-absorbing nutrient, and it is supposedly partially absorbed in the form of short-chain fatty acids by the normal flora of the human large intestine.

Expected feature/value/output:

There doesn't seem to be particularly solid evidence supporting the existence of any of these reactions in human cells. It is maybe the case that the KHK reactions happen (and thus also the transport and exchange reactions), but since tagatose monophosphate is apparently "poorly metabolized further" (in rat hepatocytes), I highly doubt the aldolase or PFK reactions are real. I'm leaning towards removing the whole pathway.

Proposed changes:

(updated based on comments below)

• Remove MAR04774, MAR04775 and MAM02955c
• Add PMID:33715524 as a reference for MAR09417, MAR08760, MAR01472, MAR01038, MAR08761, and MAR08762

[JHL-452b]

This is a very interesting finding. Howerver, I got some different tips. This paper describing that tagatose can be absorbed by liver:

Carriers transporting fructose in the small intestine do not show an affinity for d-tagatose therefore, only 20% of consumed d-tagatose is absorbed in the small intestine and then metabolized in the liver by the same route as fructose but at a slower rate > (Muddada Citation2012). The rest of the undigested d-tagatose, like other low-digestible carbohydrates (dietary fiber), is fermented in the large intestine by indigenous microflora, resulting in the production of short-chain fatty acids (SCFA), specifically butyrate (reported to combat colon cancer) (Wu et al. Citation2018). Next, SCFAs are absorbed into the bloodstream from the digestive tract as a source of energy. Furthermore, d-tagatose limits the digestion and absorption of sucrose and maltose by inhibiting the activity of sucrases and maltases in the small intestine (Lu, Levin, and Donner Citation2008). Therefore, it causes a slight or > no increase in postprandial blood glucose.

So I think these reactions can continue to be preserved in the model. Maybe, we should focus on whether MAR04774 and MAR04775 could be reversible to solve the internal loops.

[Devlin-Moyer]

that paper says tagatose is "metabolized in the liver by the same route as fructose", and a quick look through papers that mention fructose metabolism in livers (e.g. this one) makes it seem like people are pretty confident that fructose is metabolized independently of phosphofructokinase in livers, i.e. fructose -> fructose-1-phosphate -> DHAP + glyceraldehyde. So I still think MAR04774, MAR04775 and tagatose-1,6-bisphosphate (MAM02955c) should be removed (since the two reactions are associated with the phosphofructokinases and they're the only two reactions tagatose-1,6-bisphosphate participates in), but the rest now seem plausible (although we should update their references with that paper you found).

[JHL-452b]

Thank you for your reply.

I also noticed the concentrations of tagatose-1,6-bisphosphate in HMDB is not available. So, the suggestion about removing MAR04774, MAR04775 and MAM02955c looks fine to me.