Nutrient availability in tropical caves influences the dynamics of microbial biomass

Abstract Few studies have evaluated the trophic level in tropical caves, and none related the microbial biomass dynamics in the immobilization of carbon and nitrogen. Here, four tropical caves of Terra Ronca State Park, Brazil, were studied: Angélica, São Bernardo, Terra Ronca I, and Terra Ronca II caves. Physical, chemical, and microbiological parameters (microbial biomass and respiration) were estimated in the dry and wet seasons. São Bernardo, Terra Ronca I, and Terra Ronca II caves presented higher nitrogen and microbial biomass nitrogen (MBN) values in the wet season than in the dry season. On the other hand, the Angélica cave showed larger amounts of nitrogen and lower MBN values in the dry season. These results indicate that caves can be adjusted in two ecological theories known as “stoichiometric decomposition” and “microbial nitrogen mining”—to the effects of nutrient availability on organic matter decomposition. The caves studied showed different environmental dynamics in relation to organic matter decomposition, which allows them to be considered unique and possess specific characteristics. Microbial biomass dynamics can be an important parameter to evaluate the availability of nutrients and ecological dynamics of the trophic network in subterranean environments.


| INTRODUC TI ON
Limestone caves are habitats in karst landscapes where surface water sinks into soluble rock in the subsurface and flows in a network of subterranean stream passages (Ford & Willians, 2013).
Many researchers consider temperate caves to be extreme oligotrophic environments (<5 mgC/L) (Engel, 2007). The partial or total absence of light in areas farther from the entrance and the limitation of resources contribute to the uniqueness of these places, resulting in microbial characteristics, such as the absence of phototrophic organisms. However, little is known about the flow of nutrients in subterranean systems (Gilbert, 1986;Simon, Benfield, & Macko, 2003), thus restricting the discussion about the limitation of resources in this ecosystem. Therefore, recent research in tropical caves shows that these subterranean environments may not be limited by the energy input (Paula, Montoya, Rodrigues, Bichuette, & Seleghim, 2016) since caves can receive a large input of organic matter through dripping water, floods of subterranean rivers and animals, and their excreta entering the caves (Jurado et al., 2010).
Microorganisms are important components of all subterranean ecosystems (Chapelle, 2000;Griebler & Lueders, 2009). In these environments, the microorganisms are key agents in nutrient flux dynamics (Simon et al., 2007) and can also regulate chemical reactions that cause mineral dissolution and precipitation (Engel & Randall, 2011;Lian, Yuan, & Liu, 2011). Some studies in aquatic and marine caves showed high microbial activity in the early stages of organic matter decomposition that correlated with increased microbial respiration (Fichez, 1991;Graening & Brown, 2003). Other studies highlighted the importance of the microbial community as key organisms at the base of the subterranean food web and in decomposition processes (Graening & Brown, 2003;Paula et al., 2016). Microbial parameters, such as biomass and microbial respiration, can be used as sensitive indicators to monitor changes in subterranean environments (Andrews, Karlen, & Cambardella, 2004;Matsuoka, Mendes, & Loureiro, 2003).
Microbial biomass is a living component that generally comprises 1%-5% of the total soil organic matter and plays a critical role in soil fertility because of its relatively fast turnover rate (Jenkinson & Ladd, 1981). The physicochemical characteristics of soil have a great impact on its microbial biomass and can be used to measure environmental quality (Parr & Papendick, 1997), but it may take years for these parameters to result in significant changes in the environment. Also, microbiological and biochemical changes are very sensitive to small changes in environmental conditions and thereby provide more accurate and immediate information on environmental quality because the microbial activity has a direct influence on ecosystem stability (Schloter, Nannipieri, Sørensen, & Elsas, 2018;Spohn, 2015).
Microbial respiration can be considered another important attribute in ecosystems, such as subterranean ones. Large amounts of organic carbon (C) are transformed, stored, and respired by microorganisms. Hence, new insights into the factors controlling the respiration rate per unit of microbial biomass are crucial for understanding the terrestrial C cycle (Zhou et al., 2014).
The studied area of this work (Terra Ronca State Park) has a complex system of superficial and subterranean drainage, with great potential for the transport of organic matter, causing debris accumulation in some cavities. Due to this, their cavities are highly rich in subterranean terrestrial and aquatic macrofauna taxa (Bichuette & Trajano, 2003;Pinto-da-Rocha, 1995;Rheims & Pelegatti-Franco, 2003;Simões, Ferreira, & Bichuette, 2013;Trajano, Majer, Santos, & Basile, 2003). However, there are no studies on the dynamics of microbial biomass in this system and no other tropical cave. Thus, the objective of the present work was to study the dynamics of microbial parameters (biomass and respiration) and its relationship with nutrient availability (carbon and nitrogen) in four tropical caves of Terra Ronca State Park.  (Guyot, Auler, Oga, Obstancias, & Appay, 1996). PETeR is a karst area crossed by parallel streams running westwards to join the Paraná River, a tributary of the Upper Tocantins River, in Amazonas

| Study site
Basin. The major streams and some of their tributaries cross the surface and enter a cave through a sinkhole, pass hundreds to thousands of meters through subterranean conduits, and surface through a resurgence. These are typical headwater streams, with rocky bottoms with gravel, pebbles, boulders, and transparent and well-oxygenated waters.
The study area is inserted in the Cerrado phytogeographical domain (South America savannah-like). The climate of the study area is tropical semi-humid with a mean annual precipitation of approximately 1,270 mm/yr registered at the meteorological station in São Domingos-GO (Moquet et al., 2016) close to the PETeR. The wet season extends from November to April, and rainfall essentially ceases between May and October (dry season).
Floods may render some caves partially or entirely inaccessible in the wet season (November to April). Due to the high carrying capacity of these streams and the fact that the caves in the area represent stream sinkholes, substantial amounts of organic matter from vegetal debris and associated fauna accumulate inside these caves.
Four caves were studied: Angélica Cave ("A" cave), São Bernardo Cave ("S" cave), Terra Ronca I Cave ("T" cave), and Terra Ronca II Cave ("TR" cave). Angélica Cave (13°31'29" S and 46°23'07" W) is crossed by the Angélica River and is one of the largest caves of  Figure A2). São Bernardo cave entrance is a doline located at an altitude of 631 m.a.s.l. This cave has a great aquatic and terrestrial macrofauna diversity with endemic and troglomorphic species, and it is considered a hot spot of biodiversity (Bichuette et al., 2015;Trajano et al., 2003). Terra Ronca I and Terra Ronca II caves are part of the Terra Ronca-Malhada system (Appendix Figure A3). Thousands of years ago, a landslide caused its division into two parts. Currently, a canyon (approximately 900.0 m in diameter) separates the caves. Terra Ronca I extends 700 m and has a cave entrance that is 96 meters high and 120 meters wide, with a religious altar measuring 76 meters long and 100 meters high where the religious ceremony of "Bom Jesus da Lapa" occurs every year in August. Terra Ronca II extends approximately 5.1 km, has an entrance of approximately 70 meters. Terra Ronca I and Terra Ronca II caves harbor several endemic animals and troglomorphic species.

| Sampling
Samplings were conducted in April and October-2016 (license no 28992-11 (ICMBio/SISBIO) and no 14886/2010 (Secima, Goiás)). At least two samples (replicates), composed of squares of approximately 0.25 m 2 , were collected in three areas of the caves: surface environment (outside of the cave), entrance cave (transition zone), and subterranean area (dark zones). Sampling zones were classified according to Culver and Pipan (2019). Approximately 300 g of the substrate (soil or cave sediment) was collected at five different points in each square to form a composite sample. The substrate was collected from a depth of 0 to 10 cm with the aid of a gardening shovel and stored in sterile plastic bags. The samples were transported to the laboratory in coolers, homogenized, sieved (2 mm mesh), and stored in a refrigerator at 4°C.

| Physical and chemical parameters
Temperature (°C), air humidity (%), and luminosity (Lux) were measured at each sampled area with a minimum time interval of 1 min between measurements (Thermo-hygrometer Instruntherm THAL-300, 0.1 resolution, and ±5.0% accuracy). Rainfall data over 30 days before sampling were obtained from the Meteorological Database for Teaching and Research of the National Institute of Meteorology (Posse-GO meteorological station, approximately 90 km from PETeR). Substrate pH was measured at a substrate: water ratio of 1:2.5 (w/w). The moisture in the soil and cave sediment samples was estimated by the gravimetric method, which involved drying at 105°C for 20 hr to 7 days after sampling, and the results were expressed as the dry weight percentage. Organic carbon (OC) concentration data were measured colorimetrically as described by Bartlett and Ross (1988). Total nitrogen (TN) was determined by Kjeldahl digestion followed by ammonia distillation (indophenol blue method) (Bremmer & Mulvaney, 1982). The composition of the subterranean substrate was carried out by scanning electron microscopy (SEM), together with chemical analysis by energy dispersive spectroscopy (EDS). An Oxford EDS coupled to a FEI Quanta 250 SEM was used to examine the chemical composition of the samples. The substrate was adhered to a double-sided copper tape mounted on an aluminum stub to be observed (Zhai et al., 2012). About 15 ESEM images and corresponding EDS spectra of elements were acquired for each sample.
Microbial respiration was determined by the quantification of the CO 2 released by samples incubated in Bartha respirometric flasks (Bartha & Pramer, 1965). Samples (50.0 g substrate previously adjusted to 40% moisture content) were incubated in triplicate, and the CO 2 was trapped in 30.0 ml of 0.5 N NaOH. For each analysis, there were two control flasks without substrate and only NaOH solution.
The flasks were sealed and incubated in the dark at 25 ± 2°C for 48 hr. After this, microbial respiration was determined by the addition of saturated BaCl 2 to the NaOH solution, followed by titration of the nonconsumed NaOH with 0.2 N HCl.

| Statistical analysis
Data were analyzed by basic descriptive statistics (Shapiro-Wilks).
Analysis of variance and Student's t tests with a 5% probability threshold were also applied to verify the significance of the differences among the results. To analyze the relationship between variables, Pearson's correlation coefficient was used, considering variables positively correlated with r ≥0.90. RStudio (2018) was used for the statistical analyses and production of graphs.

| RE SULTS
The total precipitation in April 2016 and October 2016 was 129.80 mm and 19.00 mm, respectively. There was no significant difference in the temperature and air humidity of the subterranean environments between the seasons. In general, subterranean environments had higher air humidity than surface areas. Substrate moisture was higher in the wet season than in the dry season at all sampling sites (Table 1). Surface areas showed higher substrate neutrality (pH close to 7.0). In contrast, the pH values of the entrance area and subterranean environments exhibited a greater degree of variation (6.93 to 9.02).
TA B L E 1 Mean values and standard deviations of physical parameters (substrate moisture, air temperature, air humidity, and luminosity) in Angélica (A), São Bernardo (S), Terra Ronca I (T), and Terra Ronca II (TR) caves on the surface, entrance, and subterranean sample sites during wet (April 2016) and dry (October 2016)  All caves had more organic carbon (OC) on the surface than in the other sampled areas, except the "T" cave, which had high OC values in the entrance area. Also, all caves (except the "T" cave) had significant differences in the amount of OC between wet and dry seasons. There was no pattern in the variation of the nitrogen amount in the sampled sites. Subterranean environments ("S", "T", and "TR" caves) showed a lower amount of nitrogen (N) in the dry season than in the wet season. Instead, the subterranean area of the "A" cave had a greater amount of N in the dry season. However, there was a significant difference in N concentration in all the subterranean areas between the wet and dry seasons ( Table 2). The main chemical composition of the substrate was silica (Si), but it was not observed large concentrations of phosphorus and sulfur in subterranean samples ( Table 3).

| D ISCUSS I ON
The literature reports a wide pH range, from near neutral to slightly alkaline, in the karst (Zhang, Li, Pan, & Ren, 2006) that was also observed in this study. Highly alkaline areas (such as the "SB" cave) show the influence of water on the dissolution of the carbonate rock,     (Ortiz et al., 2014).
Although the caves studied presented nitrogen, no significant amounts of sulfur and phosphorus were found in the subterranean samples. Several nutrients, such as nitrogen, sulfur, and phosphorous, are considered limiting factors in subterranean environments (Goldscheider, Hunkeler, & Rossi, 2006). Areas farther from the entrance zone of these caves showed smaller variations in the nitrogen amount, as observed in Lapa Nova cave (Pellegrini & Ferreira, 2013). The presence of high amounts of nutrients in the caves (nitrogen, phosphorous, and sulfur), from endogenous (autochthonous) or exogenous (allochthonous) sources, can have a profound impact on the fauna, microbial growth, and community structure (Jonhston, Muench, Banks, & Barton, 2012). Sulfur and phosphorus may be present, originating from the mineral matrix of the rock, depending on the cave geochemistry, but any intrinsic source of nitrogen in caves is rare (Klimchouk, 2000). In our work, the amount of nitrogen was higher in the wet season than The surface area also had higher microbial biomass values than those inside the cave. All the caves showed microbial biomass values similar to previous studies in subterranean environments worldwide (Carmichael et al., 2015;Paula et al., 2016). There was a strong positive correlation between microbial biomass and organic matter, indicating a great influence of the quantity (amount of C) and quality (amount of N) of organic matter in microbial community biomass.
Previous studies have demonstrated that microbial biomass within caves is generally low compared with surface habitats (Barton & Jurado, 2007). The putative reasons for this include the low rate of inputs and low diversity of the autochthonous and allochthonous organic substrates (Chelius et al., 2009).
Recent studies suggest the use of soil microbial biomass as an ecosystem limitation indicator (Schloter et al., 2018;Xu, Thornton, & Post, 2013) and affirm that this parameter reflects the degree of immobilization of carbon and nitrogen. The relationship between MBC and organic carbon or MBN and nitrogen can be interpreted as substrate available and the portion of carbon and nitrogen immobilized in microbial cells. Therefore, the dynamics of microbial biomass can be a useful parameter to monitor the availability of organic matter that is less recalcitrant (Anderson & Domish, 1993).  (Li et al., 2015), such as the "A" cave.
Microbial respiration represents CO 2 release by microbial organic matter and litter decomposition (Schindlbacher, Zechmeister-Boltenstern, & Jandl, 2009;Zhou et al., 2014). Higher microbial respiration was found on the surface area of the caves. Microbial respiration inside the caves varied positively according to organic carbon availability and microbial biomass variation. In general, the amount of organic matter (carbon and nitrogen) and microbial biomass was higher in the wet season, with an increase in microbial respiration. All these factors agree with a greater immobilization of organic matter into microbial biomass and a reduction of the microbial growth in the dry season, as described above. so that the substrate meets their nutritional demands (Sinsabaugh et al., 2013). Overflow respiration in ecosystems, that is, respiration without the production of energy, has recently been questioned by several studies. First, it has been found that it may be more beneficial for the microorganisms to eliminate excess carbon by releasing as dissolved organic carbon than to expend energy in the respiratory chain (Spohn, 2015). Second, it has been pointed out that the energy lost by the disposition of carbon could be invested in storage or other processes, which increase the fitness of the microorganism (Hessen, Elser, Sterner, & Urabe, 2013).
The dynamics of the microbial biomass in the subterranean areas of the "A", "S", "T", and "TR" caves could be framed in two theories-"stoichiometric decomposition" and "microbial nitrogen mining"-regarding the impacts of nutrient availability on organic matter decomposition (Craine, Morrow, & Fierer, 2007). The "S", "T", and "TR" caves could fit the "stoichiometric decomposition" theory, where the microbial activity is highest, and the decomposition rates are maximal if the carbon and nitrogen inputs with substrate match the microbial demands. The "A" cave would fit into the "microbial nitrogen mining" theory, which assumes that the N-acquiring microbes use labile carbon as an energy source to decompose recalcitrant organic matter, which contains nitrogen (Craine et al., 2007;Moorhead & Sinsabaugh, 2006). This means that low-N pool availability for microorganisms facilitates the decomposition of recalcitrant organic matter to acquire nitrogen. In short, according to "stoichiometric decomposition," high-N availability (nutrient-rich condition) is likely to be beneficial for organic matter decomposition, while according to "microbial N mining," low-N availability (nutrient-poor condition) is likely to facilitate recalcitrant organic matter decomposition (Chen et al., 2014).
However, further studies on the stock and flow of carbon and nitrogen in these caves are needed to improve the understanding of the metabolism of these compounds.

| CON CLUS IONS
Although physical and chemical parameters are widely used as terrestrial and subterranean environmental indicators, microbiological parameters in caves, such as respiration and microbial biomass, can design the environmental quality scenario and respond more quickly to changes in caves. The microbial parameters in this study were good indicators for evaluating the nutrient dynamics in the studied environments, and with this knowledge, we can perform better health monitoring to follow the changes in these habitats.
Differences in the availability of C and N found between the "A", "S", "T", and "TR" caves may have shaped the microbial community strategies for organic matter decomposition and the incorporation of these elements into their biomass. Therefore, the caves studied showed different dynamics of microbial biomass, which allows them to be considered unique and possess specific characteristics.
Knowledge of the subterranean environments and local ecosystem processes is still very incipient, with few studies on nutrient flux in subterranean environments. Considering the specificity of each subterranean habitat, the study of each cave is extremely important to understand the local community (microorganisms and fauna) and the dynamics of energy flow in food webs. Thus, research developed in this scope would contribute to better use and management of this environment, allowing us to indicate areas more fragile to economic and/or ecotourism exploration and help adopt safer decisions about the conservation of these sites.

ACK N OWLED G EM ENTS
This study was financed in part by the Coordenação de

CO N FLI C T O F I NTE R E S T S
None declared.

AUTH O R CO NTR I B UTI O N S
Caio Paula involved in conceptualization, formal analysis, funding acquisition, investigation, methodology, and writing-original draft; Maria Bichuette involved in conceptualization, investigation, writing-review, and editing; Mirna Seleghim involved to conceptualization, funding acquisition, investigation, writing-review, and editing.

E TH I C S S TATEM ENT
None required.

DATA AVA I L A B I L I T Y S TAT E M E N T
All data generated or analyzed during this study are included in this published article. On request, additional data can be obtained from the corresponding author.

F I G U R E A 3
Terra Ronca System highlighting Terra Ronca I and Terra Ronca II caves (frame I), and the sites sampled: surface sites, entrance cave (E and TZ) and subterranean (Sub) environments. The images present the entrance of "T" cave (a) and the entrance of "TR" cave (b) (a) (b)