Vitamin K2 in Foods: A Closer Look
Vitamin K2 in foods comes either from the conversion of other K vitamins to MK-4 in animals or from bacterial production of various MKs. A good example that ties these concepts together is cheese. A cow eats grass that contains K1. The cow converts a portion of that K1 to MK-4. Both K1 and MK-4 are found in the milk. Humans take the milk and ferment it into cheese. During the fermentation process, bacteria proliferate that synthesize a variety of MKs, mainly MK-7 through MK-10, and especially MK-8 and MK-9.
A comparison of different cheeses illustrates the importance of the specific type of bacteria used in the fermentation. For example, in each 100 gram serving, Jarlsberg contains 74 μg while blue cheese contains 36, cheddar contains 21, Swiss contains 8, and mozzarella only contains 4. This variation can also be seen among fermented plant foods. For example, sauerkraut has only 5 μg, compared to nearly 1000 for natto.
Within a particular type of cheese, ripening has little effect. For example, gamalost increases from 38 to 51 in the first ten days of ripening, but this level remains mostly stable over the course of 20, 30, and 60 days. This is probably because the bacteria that produced the K2 during the initial stage of fermentation die off during the ripening (Hojo, 2007).
The data for cheese also provide a window into the possibility that some of our current food data are gross underestimates. For example, most cheeses are made with lactic acid bacteria that produce mostly MK-8 and MK-9, but some cheeses are made with proprionibacteria that also produce tetrahydro-MK-9 (Hojo, 2007), which has a structure that is the same as MK-9 except it lacks some double bonds in its side chain. These include the Swiss cheeses Emmental and Gruyère, the French cheese Comté, and the Norwegian cheese Jarlsberg. Whether tetrahydro-MKs might be present in other foods is somewhat unclear because virtually all analyses of vitamin K in foods have ignored them. No analysis has yet evaluated both tetrahydro-MKs and all the regular MKs in any food at the same time, strongly suggesting that the total K2 in foods that contain tetrahydro-MKs is grossly underestimated. To take Jarlsberg as an example, Hojo (2007) showed that, per 100 grams, it contains 8 μg MK-4 and 65 μg tetrahydro-MK-9, and cited evidence that it also contains another ~50 μg of MK-8 and MK-9. In our database, we only report values that were measured in a single scientific paper for any given sample, so our data for Jarlsberg reflects what was actually measured in the Hojo paper, 74 μg, but the true value may be over 130 μg.
Our own gut microbiota also synthesize K2: Bacterioides synthesize MK-10 and MK-11, Enterobacteria synthesize MK-8, Veillonella synthesize MK-7, and Eubacterium lentum synthesizes MK-6 (Shearer, 2014). However, this probably makes little if any contribution to our own vitamin K status for two reasons: first, most of this occurs in the large intestine, which is well past the sites of vitamin K absorption in the small intestine, and all the K2 is stuck in bacterial membranes that would have to be digested to release it.
MKs produced during the fermentation of foods such as cheese or natto are also bound in bacterial membranes, but when we eat them we digest those membranes to release the K2 in the small intestine where it can be absorbed. Some animals eat their own feces, a practice known as coprophagia, and this allows the the microbiota-derived K2 to be released and absorbed in the same way as when we eat cheese or natto. This may explain the recent finding that pork products are extremely rich in MK-10 and MK-11 (Fu, 2016). The meat was obtained from supermarkets in Eastern Massachusetts, so it presumably came from commercial farms. Perhaps pigs on those farms whether by instinct, necessity, or accident, consume feces.The only other possibility would seem to be that the pigs are fed rotten or fermented food.
The question arises whether MK-10 and MK-11 provide similar bioavailability to the MKs in other foods, which are generally much richer in MK-4 (animal foods), MK-7 (natto) or MK-8 and MK-9 (cheese) than in MK-10 or MK-11. In humans, MK-10 and MK-11 tend to predominate in the liver rather than in other tissues, and in the mitochondria rather than in the endoplasmic reticulum where vitamin K-dependent carboxylation takes place (Thijssen, 1996). Thus, we should be cautious before making a conclusion about how interchangeable the MKs in pork products are with the MKs in most other foods. Ultimately this can be resolved with studies comparing the abilities of the different MKs to support different biological functions of vitamin K.