I thank Kevin Young for kindly calling this paper to my attention and thinking that it ought to be discussed in this blog.
Caves captivate everybody. They are totally unlike our usual landscape, bewildering and challenging in design. No wonder that they attract casual visitors and explorers alike. Among the latter, we count speleomicrobiologists who search such secluded environments for unusual microbiota. The subject is well developed and even the topic of specialized meetings (see for example here). But do not think that a cave is just a cave. Caves have a great variety of niches, each with its own microbiome. Here we will deal with one kind of cave formation, the 'moonmilk speleothems.' Qu'est-ce que c'est, you ask.
Speleothems, a term that I had not encountered before. It refers to secondary mineral deposits in caves (from the Greek σπήλαιον for cave, "them" from θέμα, for deposit). Think stalagmites and stalactites. According to Wikipedia, there are 319 known varieties of speleothems, most made of calcium carbonate in the form of calcite or gypsum.
Moonmilk is the strange name given to white deposits on cave surfaces (suggesting frozen moon beams ?). This material is rather soft and gooey when wet, something like cream cheese (moonmilk is also the name given to a milk-based remedy for sleeplessness). When dry, it becomes crumbly and disintegrates into a powder. Moonmilk consists of unusually fine crystals of calcium carbonate in various chemical forms. It is quite common in karst caverns, those formed by the dissolution of carbonate rocks such as limestone or gypsum. We discussed their microbiology in a prior post: In a Cavern, in a Canyon... Prehistoric artists used their fingers to make patterns on moonmilk on cave walls and, much later, it was used as a remedy for acid stomach reflux and to heal cuts in cattle.
Microbially-induced carbonate precipitation (calcification) takes place in several ways. It may be due to the action of either autotrophs (for example, photosynthesizers or methanogens) or heterotrophs. Autotrophs are not often found in caves, leaving heterotrophs as the likely agents of moonmilk genesis. How they do it is not fully understood, but raising the local pH to about 8.0 seems to be a good contender. CaCO3 precipitates under alkaline conditions that ensue from the microbially-induced accumulation of ammonia due to the breakdown of urea, the catabolism of amino acids, or nitrate reduction. A lot of bacterial biochemistry may be involved here, as witnessed by the intricacies seen in Fig. 3. Additionally, a prevailing notion of moonmilk formation includes a second possibility, that bacteria and fungi act as nucleation centers for calcification.
Microbial processes are thought to act during the early stages of moonmilk formation, and are eventually taken over by abiotic processes. Later growth of moonmilk formations probably takes advantage of the incessant drip of carbonate-containing water into a cave. Notice that moonmilk deposits can get to be one meter thick.
Previous work proposed that a large variety of microbes may be involved in moonmilk-o-genesis, including Bacteria, Archaea, and Fungi, with filamentous forms being most prominent. The Actinobacteria loom large in this list, with their cell walls possibly acting as the nucleation centers. Because our understanding of this process is incomplete, the authors of this paper were prompted to undertake a systematic investigation on the origins of moonmilk.
Are Streptomycetes Involved ?
The source of material in this study was a 70 m long and 20 m wide Belgian limestone cave known as 'Grotte des Collemboles' (Cave of the Springtails). The researchers isolated 31 distinct Streptomyces phylotypes and looked in their genomes for genes involved in ammonia formation. For example, genes for urea breakdown, nitrate/nitrite reduction, and others. On the surface of the samples, the authors found filaments known as needle-fiber calcite and compact nano-fibers, possibly consisting of thin Actinobacteria, their unusual thinness probably being due the oligotrophic or nutrient poor environment (see here for substantiation). Unlike common streptomycetes, these structures did not show much branching.
Just being there and providing nucleation centers probably does not suffice for moonmilk formation. How about raising the pH? About half the isolates had ureolytic activity in the lab, two of them being extra-strong at it. Most of the strains carried the corresponding genes in their genomes, some in several copies. Also common were genes for carbonic anhydrase, an enzyme that converts bicarbonate into the CO2 needed for biomineralization. So were other genes involved in inorganic ion transport and metabolism that may be involved. In addition to messing with the pH, bacteria may also be directly involved in calcium carbonate mineralization by actively transporting calcium ions. In fact, they found the genes for a Ca2+/2H+ antiporter in half their isolates.
Moonmilking in the Lab
Now for making moonmilk in the lab. Can the streptomycetes do it ? The authors selected strains that had strong ureolytic and ammonification activities, probably the two most significant activities involved in the process. After growing them for several months on urea-containing medium, they found abundant calcite deposits on the colonies. No urea, no deposits. The nature of the deposits was confirmed by looking under polarized light plus by backscatter electron and elemental X-ray analysis.
The authors propose that the streptomycetes they isolated have the metabolic potential to promote calcification, although perhaps not all by the same pathway. They say: "These findings, supported by microscopy observations of bacteria-like filaments in moonmilk deposits and crustal formed by individual moonmilk-originating bacteria confirm its biogenic origin and the importance of filamentous Actinomycetes in its genesis. However, the metabolic activities evaluated in vitro were not always related to the genetic predisposition of individual strains as some isolates with grater genetic potential remained metabolically silent. This may suggest that they were grown under conditions too different from those encountered in their original niche to trigger the investigated activity or that the specific environmental cues are required for their activities. Consequently, whether these processes are indeed active in situ also remains an open question." They also point out that the uncultivated microbes endemic to moonmilk, which are the large majority, were not included in these or other studies. They think that a proteomic analysis will tell. Perhaps they will come up with the answer to the fundamental question of why do carbonates become moonmilk and not other shapes and forms?
I, for one, appreciate how complex the study of the microbiome of a natural sample can be. Lots of factors, lifeless and living, come into play and their interactions are staggering in number. This study is a good example of how hard it really is to figure out even something as simple sounding as making 'milk' in a cave.
This paper was discussed in the podcast This Week in Microbiology (TWiM), episode 157.