ENDOLITHIC PHOTOTROPHS: EXAMPLES FROM CAVE-LIKE ENVIRONMENTS

Endoliths are more frequently found in extremely harsh environments, but they can inhabit mesic climate and subterranean habitats as well. Whether they are adapted to life inside rocks or do they seek refuge in this way, remains largely unknown for caves and pits. Consequently, subterranean habitats in limestone areas, were explored: three caves in Serbia (Cerjanska, Petnička and Potpeć) and one pit in Croatia (Keranova Golubinka). Sampling of rock substratum containing endoliths was performed at various distances from the entrance, at sites characterized by a lower level of light intensity. Phototrophs were separated from the rock matrix using 10% hydrochloric acid and analyzed using light microscopy. In general, a low diversity of endoliths was observed, but representatives from Cyanobacteria, Chlorophyta and Xanthophyta were found. Cyanobacteria, more precisely coccoid forms (genera Aphanocapsa, Chroococcidiopsis, Gloeocapsa, Gloeocapsopsis, Synechococcus), were dominant. The most interesting cyanobacterial taxon, Leptolyngbya-like taxon, representing a true endolith (euendolith), was found in the Petnička Cave. This taxon and its relation to the substratum, as well as chasmoendolithic community from the Keranova Golubinka Pit, were observed using Scanning electron microscopy (SEM). It was observed that chasmoendoliths inhabit already existing surface rock cracks and fissures, while a true endolith actively bores into the substratum. Non-metric multidimensional scaling (NMDS) was performed to observe the similarity between the sampling sites. Endoliths can grow inside the rocks, making them very successful biodeteriogens, which is not a desirable trait for caves and pits that are characterized by various attractive geomorphological features that need to be preserved. Thus, considering these habitats, more attention should be paid to the endolith community in the future.


INTRODUCTION
Cyanobacteria and algae, as phototrophic microorganisms, are best known and studied from water habitats, but they can also inhabit various other niches (ALBERTANO, 2012), such as rock substrates exposed to air, that represent a specific and extreme environment for their growth and development (HAUER et al., 2015). Biofilms developing on these substrates represent an assemblage of different microorganisms embedded together in extracellular polymeric substances (ALBERTANO, 2012). The phototrophic component in biofilms growing on substrates that are exposed to light is usually very well developed, fixing carbon dioxide and providing other biofilm constituents with organic matter (ALBERTANO, 2012). The development of microorganisms on these substrates is the result of a combination of different environmental factors, substrate characteristics, as well as the mutual interaction of the microorganisms. Sometimes, owing to the seasonal variation and specific combination of environmental factors, different, physically and chemically distinctive microhabitats can be formed in the same place for one year (CASAMATTA et al., 2002;ALBERTANO, 2012;PENTECOST and WHITTON, 2012;HAUER et al., 2015). One of the special and extreme aerophytic environments for life are subterranean ecosystems (CZERWIK-MARCINKOWSKA et al., 2015).
Phototrophic microorganisms that inhabit rocks are primarily divided into epilithic and endolithic Cyanobacteria and algae, where endoliths are further divided into the following groups (according to GOLUBIĆ et al., 1981): chasmoendoliths, cryptoendoliths and euendolits. Epilithic Cyanobacteria and algae inhabit the surface of the rock substratum. Chasmoendolithic representatives inhabit rock cracks and fissures, i.e., live in already existing surface cracks that are connected with the outside environment. Cryptoendolithic Cyanobacteria and algae grow inside the porous rocks that already contain a network of open channels between the grains. The presence of empty spaces in rocks enables additional water and nutrient retention, thus providing a favourable habitat for these microorganisms. Euendoliths, on the other hand, are capable to actively, by physical and mechanical forces, deteriorate the rock and bore deeply into the substrate (ALBERTANO, 2012; KESHARI and ADHIKARY, 2014). Nevertheless, many researchers primarily focus on the epilithic community, as it is easy to observe and easy to sample, while endolithic growth of phototrophic microorganisms on rocks can often be overlooked (ASENCIO and ABOAL, 2000).
Endolithic microorganisms have gained scientific interest for several reasons: they have been considered as possible analogs of extraterrestrial life (i.e. life on Mars), they are studied due to their unique adaptation mechanisms to extreme and hostile conditions owing to the processes of weathering and mineral dissolution or phylogenetic reasons (HORATH and BACHOFEN, 2009). Some authors believe that endoliths play the main role in rock deterioration (undesirable changes in the quality and structure of rocks caused by microorganisms), because they provide a wider range of microhabitats that could be colonized by other organisms, sometimes even more dangerous biodeteriogens (KESHARI and ADHIKARY, 2014). This is especially important when cultural heritage or geoheritage (caves included) is observed (PANDEY, 2013 (GERRATH et al., 2000), and Maastricht limestone that was used for the construction of the 14th century O.L. Basilica in Tongeren, Belgium (SAIZ-JIMENEZ et al., 1990). Endoliths from cave-like environments are examined by ASENCIO and ABOAL (2000), while a detailed study from Leontari Cave was given by LAMPRINOU et al., (2009). However, the community of microorganisms in subterranean environments is still not sufficiently explored, especially in the sense of true endoliths or euendoliths.
Therefore, the goal of this study was to provide an insight into the endolithic community from subterranean environments which include three caves in Serbia and one pit in Croatia. A special attention is paid to one sample containing a true endolithic representative and its relationship with the substratum.

Study area
The sampling was performed in three caves in Serbia (Potpeć, Cerjanska and Petnička) and one pit in Croatia (Keranova Golubinka).
Potpeć Cave, located in the Potpeć village near Užice (N 43° 47' 45,1"; E 19° 56' 00,8") in western Serbia, is characteristic for its monumental 50 m high entrance. The cave is separated into two levelsupper and lower, which is the active level (ĐUROVIĆ, 1998). Part of the lower level of the cave called Hangar, which has a typical horseshoe shape 22 m wide at the top and 12 -16 m at the bottom, with a depth of about 50 m (LAZAREVIĆ, 1981), was chosen for sampling. Cerjanska Cave, located near Cerje, Niš, in southeastern Serbia (N 43˚ 27' 07,74''; E 21˚ 55' 22,81'') is an active cave, rich in various formations, more than two million years old. The main part of the Cerjanska Cave system is the Provalija ponor (often used as the synonym for Cerjanska Cave) where 6131 m of the cave system has been explored so far, putting this cave in the third place by length in Serbia. It has a relatively simple pattern of passages, consisting of one main passage and several lateral ones. However, this cave system consists of other parts too, of which one is Gornjekravljansko Vrelo The explored pit was Keranova Golubinka situated in Marići, near Oklaj (N 43˚ 56' 03,9''; E 16˚ 03' 31,6''), Croatia. The pit is approximately 60 m deep, with the wide entrance area which allows light to penetrate to the deeper parts of this subterranean habitat.
All localities ( Figure 1) are in limestone areas.

Sampling procedure
Samples were taken during autumn 2017 from the Potpeć Cave, during summer months 2015 from Cerjanska and Petnička caves, and during spring 2020 from the Keranova Golubinka Pit.
Two samples have been collected in the Potpeć Cave, marked P1 and P2. The samples were taken at the most distant point from the entrance, approximately at the distance of 50 m. 127 The place at the cave entrance from which both samples were taken is shown in Figures 1a and 1b. Two samples were also taken from the Cerjanska Cave (C1 and C2) (Figure 3c). The sampling was performed at the distance of 5 m from the entrance. One sample (PE) was taken from Petnička Cave, from the lower part of the cave (active one), at the distance of 20 m from the cave entrance ( Figure 1d). Also, one sample named KG was taken in Keranova Golubinka Pit (Figure 1e), from the depth of approximately 25 m, from a shaded place.
At all sampling sites where endolithic growth of phototrophs was observed, changes in substrate color were obvious, the substrate surface was rough, but epilithic growth of phototrophs (epilithic biofilm) was absent and nothing could be sampled from the stone surface using a scalpel, as it was done in the previous works by POPOVIĆ et al. (2015,2017). One such rock sample is shown on Figure 2. For this purpose, sites with noticeable deterioration of rock substratum were selected for sampling, and rock samples were collected. In such places, broken parts of rock substratum characterized by present endolithic growth usually lay on the floor of the cave. By collecting these samples without sampling from the cave wall itself, causing additional damage to the cave structures was avoided. The samples are stored in the Institute of Botany and Botanical Garden "Jevremovac" at the Faculty of Biology, University of Belgrade.

Separation of phototrophs from the rock matrix
The separation of endoliths from the rock matrix was done according to the method described by MATTHES-SEARS et al. (1999). Rock samples with observed endolithic growth were well washed, rinsed with distilled water and left at room temperature to dry. A nail file was used to subtly grind the rock surface, to eliminate epilithic algae if present. For each sample 100 ml of hydrochloric acid in a concentration of 10% was poured into 250 ml beakers. 10% HCl was used because this concentration dissolves limestone rocks, but according to authors, it does not harm algal and cyanobacterial cells. Rock samples were put in beakers so that the side on which endolithic growth was noticeable was immersed in the solution of hydrochloric acid. The contact of rock and acid solution resulted in an intense reaction as a consequence of dissolving the rock matrix (limestone), with the release of carbon dioxide. Samples were left in the acid solution for 24 h, to ensure that the rock was dissolved and phototrophs released. Only one treatment with HCl was performed, and no more treatments were needed. All organic material that was released from the rock precipitated at the bottom of the glass after 24 h. The acid solution was carefully removed using Pasteur pipette, to avoid the disruption of the precipitate at the bottom of the beaker glass. After that 100 ml of distilled water was added to the beaker to wash the sample from acid, and once 128 again left for 24 h for phototrophs to precipitate. The procedure was repeated after 24 h. Finally, samples were transferred to sterile 100 ml plastic bottles and fixed with formaldehyde to a final solution concentration of 4%. Using a centrifuge was avoided when rinsing the samples with distilled water, to prevent possible damage of the cyanobacterial and algal cells already treated with acid. The whole procedure was done in the digester using protective equipment.

Algological analyses
Wet mount and/or semi-permanent microscopic slides for all six samples were made in order to perform qualitative analysis of phototrophs. Wet mount microscopic slides were made when water was added and mixed with the sample, while glycerine was used for the semi-permanent slides. All was covered with the cover glass and fixed using nail polish. For the observation of microscopic slides, microscope Carl Zeiss AxioImager M.1. equipped with AxioCam MRc5 camera was used. Taxa dimensions and other parameters necessary for the proper identification were assessed in the AxioVision 4.8 software, after which identification was performed according to the standard literature (ETTL, 1978(ETTL, , 1980ANAGNOSTIDIS, 1998, 2005;JOHN et al., 2003).

Scanning electron microscopy
Scanning electron microscope (SEM) observation of the two samples (from Petnička Cave (PE) and Keranova Golubinka Pit (KG)) was performed using TESCAN Mira3 XMU at the Faculty of Technology and Metallurgy, University of Belgrade and JEOL JSM-6610LV located at Faculty of Mining and Geology, University of Belgrade. The rock samples were prepared in a standard way and covered with gold using BALTEC-SCD-005 (ANONYMOUS, 2011). This non-destructive technique allowed a close view of the substratum surface, as well as phototrophs/substrate relation and connection.

Light intensity
Light intensity (LI) was measured at sampling sites in all caves using the Velleman DMV 1300 Luxmeter.

Data analysis
Non-metric multidimensional scaling (NMDS) based on Bray-Curtis distances was performed to observe the similarity between sampling sites when documented taxa are taken into account.

RESULTS
Considering all samples, a total of 12 taxa were recorded. Cyanobacteria were dominant (10 recorded taxa), with coccoid forms prevailing (7 recorded taxa). Simple trichal cyanobacterial forms were less present (3 taxa), while heterocytous were completely absent. Only one green algal representative was recorded, as well as one from Xanthophyta division.
The first sample from Potpeć Cave (P1) was characterized by the presence of three coccoid cyanobacterial taxa from genera Aphanocapsa, Chroococcidiopsis, and Gloeocapsa and one taxon belonging to the genus Leptolyngbya. Lower diversity was recorded in the sample P2, with exclusively coccoid cyanobacterial taxa belonging to genera Chroococcidiopsis, Gloeocapsa, and Synechococcus. In Cerjanska Cave, cyanobacterial representatives of the coccoid genus Aphanocapsa and simple trichal Leptolyngbya foveolarum were recorded in the first sample. Additionally, one green algae, Desmococcus olivaceus was documented too. Higher diversity was observed in the sample C2. Three coccoid cyanobacterial representatives were found, two from the genus Aphanocapsa and one belonging to the genus Gloeocapsopsis. A simple trichal representative, Leptolyngbya sp. has also been observed. The same green algae has been documented, as well as one representative from the Xanthophyta division. Only one representative belonging to simple trichal Cyanobacteria (Leptolyngbya-like taxon) was observed in the Petnička Cave, while exclusively coccoid Cyanobacteria were recorded in the sample taken from the Keranova Golubinka Pit (Chroococcidiopsis and Gloeocapsopsis representatives) ( Table 1).  The most interesting taxon was the cyanobacterial simple trichal representative (Leptolyngbya-like taxon) documented in the sample from the Petnička Cave ( Figure 3). According to micrographs obtained using light microscopy, this taxon reminds of the genus Leptolyngbya and it is characterized with olive green trichomes constricted at cross-walls, cells that are slightly longer or longer than wide (1,71-2,86 x 1,55-1,79 μm) and colorless, to bright brown sheaths that are well-developed (Figure 3 a-d). Sheaths can be seen in all micrographs, with the exception of Figure 3d. The relation of this taxon and substratum was better observed using SEM.
Sample from the Keranova Golubinka Pit observed by SEM shows chasmoendoliths in connection with the substratum, while one from the Petnička Cave demonstrates a true endolith (Figure 4). On the SEM micrographs of the sample from the Keranova Golubinka Pit (Figure 4, a-b), a close connection between the coccoid cyanobacterial representatives (Chroococcidiopsis spp. and Gloeocapsopsis sp.  Light intensity (LI) values were similar among sampling sites. At sites where P1 and P2 samples in Potpeć Cave were taken, light intensity of 149 Lux and 192 Lux was measured. In Cerjanska Cave, LI of 77 Lux and 68 Lux was measured at places where C1 and C2 samples were taken, respectively, while in Petnička Cave the value of LI at sampling site was 98 Lux. We do not have the data from the Keranova Golubinka Pit.
According to the NMDS (Figure 5), and taking into account the recorded taxa, the most similar sampling sites are P2 and KG, as well as C1 and C2. Even though P1 has separated from these two groups, PE has shown to be the most distant, when compared to all other sampling sites.

DISCUSSION
When discussing aerophytic Cyanobacteria and algae in general, identification of these taxa can be rather problematic. This applies also to those representatives which inhabit various stone surfaces, where different life forms, such as epiliths and endoliths are included. Morphology of the endoliths could be slightly different when compared to the morphology of the same taxa living as epiliths, due to their adaptations on endolithic life. This can lead to problems in the proper identification to the species level, based on morphological and ecological traits. Consequently, the identification of endolithic Cyanobacteria and algae based on identification keys was mostly possible only to the genus level in this study. Futhermore, the treatment of samples with an acid solution has also contributed to this. Even though MATTHES-SEARS et al. (1999) reported that 10% acid solution used for the separation of phototrophs from the rock matrix does not affect the cell morphology of Cyanobacteria and algae, slight changes which have influenced proper identification have been evident in the representatives identified only to the genus level (Table 1).
Cyanobacteria are found to be the dominant phototrophs, with coccoid forms prevailing (genera Aphanocapsa, Chroococcidiopsis, Gloeocapsa, Gloeocapsopsis, Synechococcus), with considerably less simple trichal cyanobacterial forms present (mostly genus Leptolyngbya) ( Table 1). ASENCIO and ABOAL (2000) have recorded Gloeocapsa, Chroococcidiopsis and Leptolyngbya to be the most diversified chasmoendolithic genera in By taking into account endolithic communities from subterranean environments explored in this study, we can conclude that an overall low diversity of endoliths is observed, compared to, epilithic communities explored in cave environments (i.e. POPOVIĆ et al., 2015, 2017). Low and rather restricted diversity of endoliths is frequently observed in places with harsh climatescold and hot deserts and semideserts (HORATH and BACHOFEN, 2009; QUESADA and VINCENT, 2012). Nevertheless, endolithic communities should be more diverse in non-desert environments (GERRATH et al., 2000). Despite the overall low diversity recorded in this study, in addition to Cyanobacteria, the presence of Chlorophyta and Xanthophyta is also noted. Contrary to the general opinion that diversity of endoliths is still low overall, some authors demonstrate astonishingly high diversity of endolithic Cyanobacteria, with representatives of the genera Cyanosaccus, Hormathonema, Hyella, Solentia, Plectonema, Kyrthutrix, Schizotrix, Iyengariella, and Mastigocoleus being considered true euendoliths (RAMÍREZ-REINAT, 2010).
Water and nutrient availability, temperature, light, UV radiation, and predation are Light is a driving factor for the development of phototrophic organisms. Even though according to some authors there is no evidence of the relationship between the depth of the endolithic colonization and light penetration (ANTONY et al., 2012;QUESADA and VINCENT, 2012), light penetration into the rock, which is dependent on the rock type, is still considered to be closely correlated with the depth on which phototrophs develop. Often, if the light intensity is high, as a consequence of the irradiance transmission, different colored bands (composed of different groups of microorganisms at varying depths) that run parallel to the rock surface can appear (SAIZ-JIMENEZ et al., 1990;HORATH and BACHOFEN, 2009;QUESADA and VINCENT, 2012). Light intensity at cave entrances is variable during the day, but it has considerably lower values than in outside environments. Light intensity measured at the sampling sites in this study was the lowest in Cerjanska Cave, and the highest in Potpeć Cave. However, the differences between these values were not high and significant and in general there was sufficient light for the development of phototrophs. Due to a lower light intensity when compared to the outside environment, our samples were characterized by green coloration at the surface part of the substratum. It is interesting that DE LOS RÍOS et al. (2014), observing sample with different colored bands, reported that the lowest layer was composed of Cyanobacteria. The presence of Cyanobacteria in the lowest layer may point to their ability to grow at lower light levels, which can be related to subterranean habitats explored in this study, where Cyanobacteria dominated as endoliths. According to some authors, the type of substrate can also be of importance in the development of microorganisms (MAZINA and MAXIMOV 2011), while according to others, rock chemistry is usually not an essential driver and whichever substrate is colonized, physical rock properties, space available for colonization, and water retention capabilities are the most important factors which play a significant role in shaping the endolithic community (MESLIER et al., 2018). However, limestone is favourable for the development of Cyanobacteria due to higher pH (PANTAZIDOU and ROUSSOMOUSTAKAKI. 2005), which coincides with cyanobacterial dominance in this study.
Scanning electron microscopy is a non-destructive method used for detailed observation of the sample surface, morphology, texture, size and shape of particles, their orientation and type of connection. The sample from Keranova Golubinka Pit (Figure 2) has demonstrated a chasmoendolithic community related to the surface cracks of the rock substratum. On the other hand, endolithic cyanobacterial representative, Leptolyngbya-like taxon, that was found growing inside the limestone substrate of the Petnička Cave most likely belongs to true euendolithsendoliths that actively bore into the rock. The record of euendolith cyanobacterial representatives is not unexpected in limestone areas (LAMPRINOU et al., 2009). Compared to others, carbonate substrate is soft and gives satisfying protection to microorganisms that are capable to inhabit it. Few mechanisms of microorganisms boring into the rock have been suggested. According to one assumption (RAMÍREZ-REINAT, 2010), euendoliths secrete compounds that are characterized with low pH value. However, autotrophs use CO2 in photosynthesis, thus increasing pH of their environment. Therefore, it is believed that they bore into the rock only during the night, or that they photosynthesize only when they are outside the carbonate matrix. A suggested mechanism could be interesting, but authors agree that it would be very energy-demanding for the microorganism itself. The second mechanism worth mentioning is based on the removal of calcium ions by transport enzymes. Microorganisms can adsorb calcium ions from interstitial space (space between cells and minerals), leading to lower quantities of calcium in substrate and its weakening. They adsorb calcium ions on one side of the cell, move them through the cell and secrete 134 them on the opposite side. This mechanism of calcium transport is exactly what allows euendoliths to actively bore into the substratum (RAMÍREZ-REINAT, 2010).
NMDS diagram separated a group of samples based on the recorded taxa. First group is composed of P2 and KG that are characterized by the presence of only coccoid cyanobacterial forms (representatives of Chroococcidiopsis, Gloeocapsa, Gloeocapsopsis and Synechococcus). P1 was separated, as in this sample, coccoid, as well as simple trichal cyanobacterial representatives, were recorded (genera Aphanocapsa, Chroococcidiopsis, Gloeocapsa and Leptolyngbya). C1 and C2 that form a separate group are not only characterized by the presence of coccoid and simple trichal Cyanobacteria (genera Aphanocapsa, Gloeocapsopsis, Leptolyngbya), but are also the only sites where green algae Desmococcus olivaceus was found. PE was the most distant of all, since only one sample of a trichal cyanobacterial taxon whose identification was not certain was found here.

CONCLUSIONS
Endolithic phototrophs are not explored enough, especially in habitats where they are expected to develop to a lesser extent, such as subterranean ones. In this study, they are explored from three caves in Serbia (Cerjanska, Petnička and Potpeć) and one pit in Croatia (Keranova Golubinka), i.e. all from places characterized by lower level of light intensity and a little bit more uniform environmental conditions. In general, low diversity of endoliths was observed, with Cyanobacteria (mostly coccoid forms, genera Aphanocapsa, Chroococcidiopsis, Gloeocapsa, Gloeocapsopsis, Synechococcus) being dominant. The most interesting cyanobacterial taxon was one euendolith (Leptolyngbya-like taxon) which association with the substratum, together with chasmoendolithic community from other locality, is observed using SEM. Even though a small number of endolithic samples was found at examined localities, this community should not be neglected. Considering subterranean environments that are rich in various morphological features attractive to tourists, development of phototrophic community, especially euendolithic one, is not desirable. In this study, endoliths are explored mainly from the cave/pit entrances, but they can be also present inside of attractive and illuminated tourist caves. Thus, more attention should be paid to this community in the future and monitoring of such sites needs to be performed continuously.