Introduction

Archaeological remains are limited and non-renewable resources of valuable information about ancient civilizations; consequently, they may potentially contribute to a deeper understanding of human societies in previous times. Archaeological sites are important not only for education and research purposes, but also from tourism, cultural and recreational perspectives. Both archaeological monuments and historical buildings are exposed to multiple environmental factors that make them vulnerable to deterioration and destruction, such as humidity, high temperatures, precipitation, anthropogenic activity and the effect of micro- and macro-biological communities (^{Videla, Guiamet & Gómez de Saravia, 2000}).

Biodeterioration is a term that collectively refers to the damages caused by microbiological communities in archaeological sites and historic buildings as a consequence of the metabolism of endo- and epilithic microorganisms that degrade the rocky substrates where they thrive as a result of the production of various metabolites such as organic and inorganic acids (^{Ascaso, Wierzchos, Souza-Egipsy, De los Ríos & Delgado Rodrigues, 2002}).

Microbial communities frequently produce biofilms, i.e., structured communities that include bacteria, algae, cyanobacteria, fungi and protozoa, embedded in a polymer matrix that provides support and protection and serves as a nutrient reservoir (^{Romaní, Found, Artiagas, Schwartz, Sabater & Obst, 2008}). Microorganisms that are capable to colonize and survive in inhospitable environments play a key role in the establishment of microbial populations that contribute to the formation of soils from the substrates on which these grow (^{McNamara, Perry, Bearce, Hernández-Duque & Mitchell, 2006}).

Recently, the family Mycobacteraceae has been subdivided into five different genera: Mycobacterium, Mycolicibacter, Mycobacteroides, Mycolicibacillus, and Mycolicibacterium (^{Gupta, Lo & Son, 2018}). These mycobacteria are short aerobic bacilli, straight or slightly curved, and non-motile. The high G+C content coupled with a cell envelope rich in mycolic acids and other lipids make them resistant to discoloration by acid-alcohol when stained with the Ziehl-Neelsen technique. According to their 16S rRNA sequences and the disease they can cause, mycobacteria have been classified in two groups: the Mycobacterium tuberculosis complex (MTC) and nontuberculous mycobacteria (NTM). Some mycobacteria cause infections in humans and/or animals, while others are colonizing organisms that form biofilms. Studies on these organisms have focused primarily on the clinical and veterinary areas; consequently, the search for them has been done on soil and drinking water samples, as well as in man and other mammals (^{Falkinham, 2009a}; ^{Johnson & Odell, 2014}).

Nontuberculous mycobacteria are ubiquitous microorganisms that inhabit terrestrial and aquatic environments and have been isolated from various sources, some of which are directly linked to human environments (^{Falkinham, 2015}), as shown by the report of the isolation of five strains of mycobacteria that might be involved in the deterioration of monuments in Angkor, Cambodia from 2011; according to the author, all five strains isolated were able to use elemental sulfur (S⁰) for chemolithoautotrophic growth, and organic substances for chemoorganoheterotrophic growth (^{Kusumi, Shu-Li & Katayama, 2011}). To the best of our knowledge, the latter is the first report of NTM isolated from such environments; therefore, the aim of the present work was to search, isolate and identify NTM from biofilms grown on stone monuments of various archaeological sites in Mexico.

Materials and methods

Sample Collection

Biofilm samples were collected aseptically using a sterilized spatula and placed in a sterile polypropylene container with screw cap. Samples were collected mainly in the rainy season (July, August, and September), from deteriorated buildings located in nine Mexican archaeological sites (Figure 1), namely Guachimontones (Jalisco), Atetelco and Malinalco (Estado de México), Zaachila (Oaxaca), Tulum (Quintana Roo), EkBalam (Yucatán), Tizatlán and Xochitécatl (Tlaxcala), and La Antigua (Veracruz). Sampling was conducted with proper care to avoid damaging the monuments (non-invasive method). During the sampling, the most deteriorated points of monuments were chosen, where well-consolidated biofilms or a light layer of soil (not thicker than 2 cm) were observed; the characteristics of all collection sites were recorded (Table I).

Figure 1 Location of archaeological sites under study. A: Atetelco, Estado de México; B: Malinalco, Estado de México; C: Guachimontones, Jalisco; D: Tulúm, Quintana Roo; E: Cobá, Quintana Roo; F: EkBalam, Yucatán; G: ChichénItzá, Yucatán; H: Cacaxtla, Tlaxcala; I: Xochitécatl, Tlaxcala; J: Ocotelulco, Tlaxcala; K: Tizatlán, Tlaxcala: L: Uxmal, Yucatán; M: Yohualichan, Puebla; N: Monte Albán y Zaachila, Oaxaca; O: Mitla, Oaxaca y Q: La Antigua, Veracruz. Figure designed by the authors. 

Table I Characteristics of the archaeological zones studied. 

Archaeological sites	State	Altitude (msnm)	Climate	Humidity	Mean annual precipitation (mm)	Mean annual temperature (°C)	Construction materials
Atetelco	Estado	2,270	Semi-dry with	Low	600 - 700	14 - 16	"Tepetate" or volcanic
(Teotihuacán)	de México		summer rainfall				tolba with and without caliche (calcium carbonate)
Malinalco		1,750	Semi-warm subhumid with summer rainfall	High	800 - 1,300	14 - 28	Extrusive igneous rock, and limestone
Guachimontones	Jalisco	1,260		Medium	800 - 1,100	16 - 22	Stone, adobe, and lime
Tulúm	Quintana Roo	5	Warm subhumid with summer rainfall	High	0 - 1,300	24 - 28	Limestone
Cobá		15
Ek Balam	Yucatán	22		Medium	1,000 - 1,200
Chichén Itzá		20			1,100 - 1,200
Uxmal		31			1,000 - 1,100	24 - 28
Cacaxtla	Tlaxcala	2,200	Temperate subhumid with summer rainfall		800 - 1,000	14 - 16	"Tepetate", limestone, adesitas, "tezontle", and adobe.
Xochitécatl
Ocotelulco		2,340					Stone slabs of xalnene, adobe, stone, tepetate, and limestone.
Tizatlán		2,240					"Tezontle"and limestone
Mitla	Oaxaca	1,680	Temperate subhumid with summer rainfall		600 - 1,500	14 - 22	Limestone
Monte Albán		1,580	Semi-dry semi-warm	Low	600 - 700	18 - 22	Stone, adobe, and stucco
Zaachila		1,490			600 - 800	16 - 22	Limestone
La Antigua	Veracruz	20	Warm subhumid with summer rainfall	High	1,100 - 1,300	24 - 26	Coral, river stone, brick quarry, flat tile, and stucco
Yohualichan	Puebla	600	Semi-warm wet with rain all year		1,900 - 4,100	18 - 26	Limestone

Results

Fifty acid-fast microorganisms were isolated from the nine archaeological sites. Table II shows the distribution of these organisms; the largest number of isolates was obtained from Guachimontones and Atetelco. In contrast, the sites with the lowest number of isolates were Tizatlán, La Antigua, Tulum, and EkBalam. The remaining isolates were distributed among Xochitécatl, Zaachila and Malinalco.

The comparisons of environmental factors with the number of AFB microorganisms through the univariate analysis of variance (general linear model) was significant only for climate (sig. = 0.127, Eta = 0.239) and for the precipitation-humidity interaction-temperature (sig. = 0.193; Eta = 0.180); that is, these were the factors with the greatest influence on the presence of AFB microorganisms in archaeological zones.

The morphological descriptions of the isolates obtained were conducted in Middlebrook 7H10 medium after 6 and 15 days of incubation at 28 °C. Colonies were either circular or irregular, measuring 1.5-15 mm in diameter, and white, yellow or light yellow; most were flat or umbonate and soft, with no other distinctive characteristics. All isolates were acid-fast bacilli, measuring 1-3 μm in length. The Mycobacteriaceae-specific PCR assay tested positive for 45 of the 50 AFB isolates, i.e., 90% of isolates were mycobacteria. The results of the three molecular markers (rrs, hsp65, and rpoB genes) serve to define 39 species of the 45 NTM isolated; five consensus species were determined, distributed as follows: 21 isolates were Mycobacteroides chelonae; seven, Mycobacteroides abscessus; five, Mycolicibacterium flavescens; four, Mycobacterium alvei; and two, Mycobacterium fortuitum; the six that could not be identified to species through the comparison of the three markers used were reported as Mycobacterium sp. (Table III). The Gene Bank access numbers are listed on Table IV, these species were isolated from eight of the nine archaeological sites. The largest number of NTM (25) was isolated from Guachimontones (Jalisco), corresponding to at least two different species; the site that ranked second in number of NTM isolated was Atetelco (Estado de México), with at least three different species.

Discussion

We investigated the biofilms that grow on monuments of different archaeological sites of Mexico searching for cultivable mycobacteria; this required performing decontamination and isolation procedures during which mycobacteria were isolated from the microbial consortium where they were immersed. The course of these techniques required microscopic observations to identify mycobacteria based on acid fastness. However, this trait is not unique to mycobacteria, and a number of other acid-fast microorganisms of different shapes, which coexist with NTM in their natural environment, were observed throughout the procedures.

Microorganisms do not live in isolation; they rather form complex communities (such as biofilms) that allow them to with stand the fluctuating environmental conditions. In addition, microbial assemblages display a differential response to climatic variations, so that their composition is constantly changing over time and space (^{Garret, Bhakoo & Zhang, 2008}). A particular case was observed in a sample from Guachimontones, where Mycobacterium sp. was found along with a filamentous acid-fast microorganism with which the former was so closely associated that it was necessary to incubate both microorganisms together in liquid medium followed by their subsequent subculture by cross streak method in Middlebrook 7H10 agar. This was not the only close association observed during the isolation process; in general, mycobacteria grew with in a microbiota assemblage that made it difficult to obtain pure cultures, mainly due to the time required to observe growth.

Most mycobacterial isolates developed after 6 days of incubation at 28 °C (fast-growing NTM); however, this generation time is too long relative to the one of other bacteria that, when cultivated together, displayed a faster growth compared to mycobacteria. This issue was solved by adding either antibiotics (PANTA^TM) or malachite green (0.4 g/L) to the culture medium.

In nature, both the impermeability of the cell envelope and the time taken for mycobacteria to grow give them an advantage relative to other coexisting organisms. The synthesis of this highly complex envelope made of mycolic acid requires mycobacteria to invest a longer time, thus affecting growth rate. This key adaptation, due to the hydrophobicity it confers, enables mycobacteria to attach to surfaces and remain attached to the substrate even despite water runoff. Furthermore, this impermeable barrier protects them from dehydration in dry seasons, as well as from a number of chemicals under lab conditions, such as antibiotics and dyes (^{Falkinham, 2009b}; Falkinham, 2015).

Most isolates came from sampling points where moss was also present, forming a consolidated biofilm on the stones that make up the foundation of the buildings under study; this is consistent with other investigations that report mycobacteria in association with bryophytes (^{Pavlik, Kazda & Falkinham, 2010}; ^{Thorel, Falkinham & Moreau, 2004}). Apparently, the life cycle and seasons of the year when the various species of bryophytes grow are directly related to the establishment and development of mycobacteria, i.e., during the growth of mosses (early spring to late autumn) a green hygroscopic layer develops that serves as habitat for various species of microorganisms. Each year, moss regrows to form a new layer on top of the one of the previous year, leaving the latter deprived from light; as this accumulates over time, the deeper layers are broken down by pectinases from mycorrhizal fungi. The products released from the metabolism of fungi and the decomposition of mosses (amino acids and carbohydrates) are dissolved in the medium, producing a suitable environment for the development of mycobacteria (^{Pavlik, Kazda & Falkinham, 2010}). Mycobacteria were also found on biofilms devoid of moss and with low humidity (Atetelco and Zaachila), a finding that can be explained by the same adaptations developed over the course of the evolutionary history (hydrophobicity, impermeability, and adherence) that allows them to survive under extreme conditions such as desiccation or low nutrient levels.

The 45 colonies identified as NTM showed the typical characteristics of the genus Mycobacterium; however, during the isolation process some colonies (later identified as the same species), developed into two different morphologies (usually flat or umbonate), leading to uncertainty about their identity. It has been reported that the chemical structure of the mycobacterial cell envelope changes under stressful conditions (e.g., exposure to antibiotics), by rearranging the peptidoglycan layer or during an infection (bacteria-host interaction), by increasing the production of mycolic acids and other cell-envelope lipids (^{Yang et al., 2013}). Some studies have shown the production of mutants with different morphological characteristics after exposure to antibiotics such as rifampicin and ethambutol or to sanitizing agents such as glutaraldehyde. In these cases, changes in the composition of fatty acids and the arabinogalactan/arabinomannan ratio in the mycobacterial cell envelope have been detected (^{Manzoor, Lamber, Griffiths, Gill & Fraise, 1999}; ^{Sareen & Khuller, 1990}). In addition, it has been reported that these microorganisms display a high plasticity in the synthesis of the cell envelope, thus facilitating the development of mutants that are more resistant to adverse environmental conditions (^{Kieser & Rubin, 2014}). Our observations during the isolation of NTM from archaeological sites are consistent with the hypothesis that the differentiation of colony morphology occurs in response to stress, as this phenomenon was observed when the antibiotic mixture was added to the culture medium of some samples or upon exposing them to malachite green. This likely led to the selection of mutant strains of the same species, since colonies did not return to their original morphology upon re-inoculation in medium devoid of antibiotics; alternatively, these colonies may have developed other morphotypes of the same NTM species when the growth of other microorganisms was inhibited. Furthermore, our observations regarding growth in liquid medium revealed that flat morphotypes were more hydrophobic than the umbonated isolates. All strains that differentiated in the same medium shared the same identification at species level (M. abscessus, M. chelonae, and M. flavescens).

Conclusions

In this pilot study, five different species of NTM were successfully isolated from several archaeological sites in Mexico. Most were part of biofilms grown on stones of monuments and were associated to other microorganisms, a relationship that deserves further investigation in order to gain a deeper insight on the role of these NTM in such habitats.