Revealing Microbial Population Complexity on Hospital Textiles Using DHPLC

Hospital environment provides an important ecological niche that can serve as a reservoir for potentially pathogenic microorganisms. Hospital textiles together with moisture and heat create the right conditions for growth, dissemination and long-term survival of many microorganisms, which may results in transmission of nosocomial infections. Therefore, understanding of total microbial population on hospital textiles is an important part for assessment of public health risks. In the present work, we tested the suitability of DHPLC for estimation of total microbial population on hospital textiles. All of the identified bacteria belonged to the phyla Actinobacteria, Firmicutes and Proteobacteria and up to 63 bacterial genera/species were identified. Mostly abundant species important in perspective of HCAI were Acinetobacter spp., Corynebacterium spp., Staphylococcus spp., Sphingomonas mucosissima and Stenotrophomonas maltophilia. Microbial diversity on hospital textiles was extensive and partly coincides with the intensive care units bacterial communities and pathogens obtained from clinical samples. This is the first study describing total microbial population complexity on hospital textiles, which are important factors contributing to the hospital microbiome.


Introduction
Understanding microbial populations in hospital environments is crucial for improving human health [1], since indoor microbial communities are an important component of everyday human health [2,3]. Characteristics of the hospital environments are very specific where several major factors need to be considered: patients and medical workers, infections and resistant bacteria and also hospital inanimate environments including hospital textiles. Hospital textiles that come frequently into contact with hands are often contaminated with nosocomial pathogens and may serve as vectors for cross transmission and can be a source of healthcare associated infections [4]. Therefore, hospital textiles can contribute to the transmission of pathogens among hospital staff, patients and visitors. A single hand contact with contaminated textiles is enough to transfer the pathogens [5]. The patients are susceptible to colonization [6] and often underlying severe diseases [7,8] that can accelerate infection. Therefore, many patients get HCAI [healthcare associated infection] that implies in prolonged hospital stay, long-term disability, which are a massive additional financial burden for health systems, high costs for patients and their families, and excess deaths [9]. On the basis of the Report on the Burden of Endemic Health Care-Associated Infection Worldwide that included data from results of systematic reviews of the literature on endemic HCAI from 1995 to 2010 in high-and low/ middle-income countries is estimated that 4.131.000 patients are affected by approximately 4.544.100 episodes of HCAI every year in Europe. In Europe, HCAIs cause 16 million extra-days of hospital stay, 37.000 attributable deaths, and contribute to an additional 110.000 deaths every year. Annual financial losses are estimated at approximately €7 billion, including direct costs only. In the USA, approximately 99.000 deaths were attributed to HCAI in 2002 and the annual economic impact was 16 Revealing Microbial Population Complexity on Hospital Textiles Using DHPLC estimated at approximately US$ 6.5 billion in 2004 [9]. Another important fact regarding HCAI is increased resistance of microorganisms to antimicrobials. In hospital environments selective pressure for microorganisms is extreme and largely oriented for survival in hostile environments rather than for traits providing fitness in slowly evolving populations. Over the last century, the phenomenon of antibiotic resistant microorganisms that can survive in the presence of a broad spectrum of antimicrobial agents has developed [10]. The hospital environment may act as filter selecting specific bacterial populations and creating special and confined areas where microbiota clearly differs from that present in nature. Deep studies are needed to describe the microbial diversity associated to hospital environments to allow the identification of potential microorganisms causing HCAI [1].
For estimation of total microbial population complexity and studying of microbial diversity the microbial ecology has undergone a profound change in the last two decades. Emphasis has shifted from culturing to the molecular-based approaches [11] since most of the bacteria isolated from natural environments are not represented in culture collections [1]. On the other hand uncultured organisms comprise the vast majority of the microbial world [12][13][14], but most of the time encountered in complex environments [11] that also includes hospital environment. It is well known that cultivation-dependent techniques capture only a small part of the microbiome [15,16] and that molecular methods can reveal high bacterial diversity [17], besides being faster and more reliable possibility. Although new generation sequencing techniques are currently becoming a standard for analysis of complex microbial populations, some older methods could still be used for populations with low or moderate complexity. One of such methods is denaturing high-performance liquid chromatography (DHPLC). DHPLC has been primarily used for mutation detection, mapping of genes, genotyping of known polymorphisms [18][19][20][21] and to detect DNA sequence variations such as insertions, deletions, genomic single nucleotide polymorphisms (SNPs), and microsatellites [22]. Limitations of DHPLC could be attributed to conserved nature of 16S rRNA genes, DNA extraction inherent limitations, formation of chimera and heteroduplex, preferential bias during the PCR [22] and the occurrence of mixed sequences in one fraction.
The aim of our survey was to test efficiency of the DHPLC method for the estimation of total microbial population of human pathogens on hospital textiles. We used a nondestructive sampling method that included a compact test device called Morapex A (SedoTreepoint GmbH), which is based on forced desorption by pressing the microorganisms through the fabric, for nondestructive sampling the hospital textiles [23] and afterwards study microbial population complexity which is an important aspect in preventing nosocomial infections.

Sampling
To test the efficiency of a nondestructive elution method with the Morapex A device [23] two hospital sheets and one hospital pajama were sampled. Sheets and pajama were collected from patient at University Clinical Center Maribor at Department of Infectious Disease and Febrile Conditions in routine disposal of hospital laundry after being used for one day. Sheets were sampled at eight evenly spaced spots and pajama at three different spots (i.e. end of a sleeve, armpit and collar). The testing material was placed between two metal plates; 20 ml test liquid (0,9% NaCl + 0,2% Tween 80) was pressed through the fabric in three cycles of 30 seconds and collected in a tube. The eluate was stored in a refrigerator until the DNA extraction. Sampling was conducted at room temperature.

DNA Extraction
Mixed bacterial genomic DNA was extracted from the suspension of microorganisms retrieved from textiles with a nondestructive the elution method described above. Extraction was performed with PrepMan Ultra Sample Preparation Reagent (Applied Biosystems) for each sampling spot on hospital textiles in accordance with the manufacturer's instructions. Extracted DNA was stored at -20⁰C prior to 16s rRNA gene amplification.

16S rRNA Gene Amplification
The amplification of the targeted variable region V6 -V8 of 16S rRNA gene was carried out with primers described by Domann et al. [24]. Forward primer 0933F (5′-GCACAAGCGGTGGAGCATGTGG -3′) and the reverse primer 1407R (5′-GACGGGCGGTGTGTACAAG-3′) were used. The amplicons obtained from mixed bacterial DNA were checked for the correct size [470 bp] by agarose gel electrophoresis. Heteroduplex DNA molecules were reduced by reconditioning as described by Thompson et al. [25]. Finally the PCR products were purified using the PCR Purification Kit from Qiagen.

Separation of Amplified 16S rRNA Gene Fragments with DHPLC
Purified PCR products were separated by DHPLC on a DNASep® HT cartridge using the WAVE Microbial Analysis System (Transgenomics, USA). For the particular analysis and separation of the mixed bacterial species we generated the gradient "mixed species" and used it at a column temperature of 62.0°C as described by Domann et al. [24]. The gradient was formed by buffer A, consisting of 0.1 M triethylammonium acetate (TEAA), pH 7.0, and buffer B, consisting of 0.1 M TEAA and 25% acetonitrile, pH 7.0. Buffer C, consisting of 25% water and 75% acetonitrile, was used for washing the column [24]. Fractions representing chromatogram peaks were collected in different retention time intervals (Appendix). Percentage of representation in the sample was calculated based on area of each peak on DHPLC chromatogram [26].

Sequencing of DHPLC Peaks and Cloning of Fraction
From DHPLC chromatograms 188 most outstanding peaks were chosen and sent to Sanger sequencing at Macrogen (Netherlands). Nucleotide sequences were analyzed with the program RipSeq MIXED (Isentio). For fraction 7 in sample 8 cloning was performed using pGEM-T Vector System (Promega, Germany) for further identification. Altogether 28 cloned fragments were obtained and further analyzed in the WAVE system. Out of those, 12 that differed in retention times were selected, and processed for nucleotide sequences.

Data Processing
Data matrix based on the presence of the peak and the median value of retention time for each collected peak from DHPLC chromatograms was used as input for Principal Component Analysis (PCA). PCA cluster analytic method and visualization was performed with Canoco and CanoDraw programs [27].

Results
Data analyzes described below are based on the chromatograms obtained by DHPLC method. Each point of the chromatogram was defined as the retention time (min) on the x axis and the signal intensity (mV) on the y-axis. One chromatogram represents one sampling spot on textiles from the hospital environment. Each chromatogram was composed of many peaks (Figure 1 -3), which differed by retention times and represented different taxa (Appendix).
From the different number of peaks per sample it can be seen that the complexity of the bacterial population varied between samples and that microbiota detected on sheet 1 and 2 was more complex in comparison to the pajama (Figure 1 -3). On the other hand, the complexity of the bacterial population within single samples was quite similar, a certain deviation can be observed only in sample 6 from sheet 1. Some of the peaks are attributed to several different taxa of bacteria, since the relatively short nucleotide sequences (420 bp), which was used for analysis by DHPLC did not allow precise identification of individual amplified fractions. In the case of mixed sequences the presence of three different bacterial genera/species in the samples was determined by RipSeq MIXED (Isentio) algorithm. Only sequences whose resemblance to a referenc sequence was higher than 99.5% were taken into account.   Revealing Microbial Population Complexity on Hospital Textiles Using DHPLC The composition of microbial populations included three bacterial phyla: Actinobacteria, Firmicutes and Proteobacteria. The phyla Actinobacteria was present at low levels and mainly represented by genus Corynebacterium spp. Second, regarding the percentage of representation was the phylum Firmicutes mainly represented by genus Bacillus spp., Enterococcus spp., Staphylococcus spp. and Streptococcus spp. The most abundant phylum across all samples (e.g. collected peaks) was Proteobacteria mainly represented by genus Cupriavidus spp., Ralstonia spp., Sphingomonas spp. and Stenotrophomonas spp.
On the sheet 1 a total of 59 different groups were identified, those were composed of various combinations of 56 bacterial genera/species (Appendix). The highest percentages among each of eight sampling spots on the sheet 1 are represented in Table 1. The peak of the sample 6 ( Figure 1) indicates the presence of bacteria from family Enterobacteriaceae.
On the sheet 2 a total of 50 different groups were identified, those were composed of various combinations of 63 bacterial genera/species (Appendix). The over represented groups reflected more than half of the bacterial population in the individual sampling spots. The highest percentages of representation among eight sampling spots on the sheet 2 are represented in Table 2. The correlation between the bacterial community and environmental variables (pajama and sheets) was studied using Canoco. PCA using the entire retention time list clearly separated all identified species in four distinct clusters, visible in the 3-dimensional subspace spun by the first three principal components ( Figure 4).

Discussion
Different studies have been performed to describe microbial population in hospital environments (mostly intensive care units -ICU), but have been focusing on floors, medical devices, and workplaces [1,17] and left out the hospital textiles as an important factor of the hospital's and the patient's environment. Phyla Actinobacteria, Firmicutes and Proteobacteria that were found in our study were also obtained for microbiota communities found in hospital ICU [1] and in the human body [28][29][30]. Also the finding in our study, that phylum Proteobacteria was most abundant, is consistent with finding of Obernauer et al. [17] in the ICU microbiome study.
According to a specialty of the Department of Infectious Disease and Febrile Conditions, which treated patients with severe, sometimes multiple infections and complications of infectious diseases as well as patients who require isolation,

22
Revealing Microbial Population Complexity on Hospital Textiles Using DHPLC the high diversity of bacterial populations was expected. Regarding the percentage of the representation in the sampling spots at sheet 1 and sheet 2 the most common types, which are also important in terms of infections in humans and in terms of HCAI are the following: Acinetobacter baumannii, Acinetobacter junii, Bacillus circulans, Bordetella holmesii, Corynebacterium tuberculostearicum Edwardsiella spp., Pandoraea sputorum, Proteus spp., Sphingomonas mucosissima, Staphylococcus spp., Stenotrophomonas maltophilia, Sutterella wadsworthensis and Yersinia spp. Of the above-listed species on sheet 1 two types of bacteria were dominant. The first one is Sphingomonas mucosissima, for which there is a report on the cause of bacteremia, that's why it needs to be regarded as potentially pathogenic. Especially in impaired patients and patients undergoing medical procedures it could cause exceptional cases of opportunistic infections [31]. Sphingomonas members are metabolically versatile and its abundant presence in hospital textiles could be explained because this genus has a good ability to utilize a wide range of organic compounds and to grow and survive under low-nutrient conditions, including toxic compounds [32]. Actually S. paucimobilis has been previously detected in hospital equipment [33] and in respiratory therapy items, bedside water bottles, sinks etc. and has been described as a cause of minor infections in humans although it is also capable of causing serious infections [34]. The second one is Stenotrophomonas maltophilia, which is increasingly emerging as a cause of nosocomial infections, especially in immunocompromised patients, where the treatment of such infections can be problematic because of the resistance of different strains to many antibiotics [35]. On sheet 2 four types of the above-listed species were dominant. The first one is Corynebacterium spp. which is commonly present in humans but can also cause some diseases in both healthy and immunocompromised hosts [36]. The second one is Acinetobacter spp. that was also the dominant genus of the floor sample and present on almost all devices in ICU [17]. In aspect of nosocomial infections antibiotic-resistant Acinetobacter infections have become an increasingly common nosocomial problem [37,38]. The most important type is A. baumannii, which is resistant to a wide range of antibiotics [39]. The third and fourth outstanding species on sheet 2 were Sphingomonas mucosissima and Stenotrophomonas maltophilia described above.
Pajama is the part of the patient's environment that is in closest contact with the patient's skin; hence the microbial flora in pajama is quite different from that of the sheets. Highly over represented species are the one of the genus Staphylococcus spp. and species Staphylococcus aureus that can be found on human skin and mucosae and are therefore expected on samples that are in close skin contact. Obernauer et al. [17] describe Staphylococcus spp. as skin-associated genera that were highly abundant on medical devices and working surfaces, which are just like textiles touched by hands of hospital staff. Mentioned bacteria may induce a broad spectrum of infections in immunocompromised humans [40]. As an extremely versatile pathogen, this genus has developed a broad spectrum of mechanisms leading to resistance to our most powerful antimicrobial agents [1]. S. aureus is, in fact, one of the major causes of hospital acquired infection, being normal flora of ICUs [41] and as our results show, also normal flora of patient's pajama.
Textiles were partially colonized with human associated bacteria that were also found in other indoor microbial communities [3]. Species that have facultative pathogenic and nosocomial character, e.g. Pseudomonas spp., Haemophilus spp., Enterobacter spp., Citrobacter spp., Bordetella holmesii, Escherichia spp., Yersinia spp. and Proteus spp. are common pathogens obtained from clinical samples [1,42,43] and were also present on textiles investigated in this study. This finding could classify hospital textiles as a part of environment where these pathogens are presumably existing, besides already recognized areas of colonization like mechanically controlled ventilation systems and catheters. Regarding issues in clinical settings Pseudomonas spp. find the hospital and intensive care unit environments accommodating and even more so hospital textiles as wet, body-temperature environment [44][45][46]. On the other hand families such as Corynebacteriaceae or Sphigomonadaceae do not usually cause infectious diseases therefore they do not appear in the clinical samples but are broadly overrepresented in ICU samples [1] and also in textiles samples from our study. Hospital bacterial communities are composed of bacteria closely related to human pathogens, but also of taxa known for their beneficial interaction with eukaryotes [17]. Therefore, whole bacteria community should be considered, since diversity within these communities often correlates with the ecosystem function of disease suppression [47].
Regarding the methodology some difficulties have emerged in the analysis of microbial populations on the hospital textiles by DHPLC. Identification to one bacterial species was not possible, and as a result there was only the probability of occurrence of up to three different bacterial genera/species, for which the percentage of representation cannot be determined. More accurate identification of individual fractions can be achieved with a shorter interval for collecting fractions on DHPLC chromatograms [48]. Škraban et al. [47] have collected fractions in 0.2min interval and were successful in differentiation between selected groups in fecal specimens. Wagner et al. [26] have also been successful in distinguishing the individual fragments from DHPLC analysis of mixed microbial populations from complex environmental samples such as sludge, compost and soil. Since collecting fractions of distinct peaks did not yield satisfactory results in terms of separation, they have used cloning approaches. Similarly, in our study, we did in the repetition of sample 8 from sheet 1. Our finding is that microbial diversity on hospital textiles is very extensive and partly coincides with the ICU bacterial communities from another study [1], but still the fact that DHPLC diversity profiles do not necessarily reflect the true diversity in the field [22] must be taken into account. The limitations could be attributed to different factors such as the conserved nature of 16S rRNA gene, DNA extraction inherent limitations and formation of chimera and heteroduplex. Closely similar DNA may result in clamping or clustering into a single or closely overlapping peak [22]. We believe that DHPLC can be used to monitor the diversity of microbial populations in a process, for example in the process of laundering hospital textiles or to compare the complexity of bacterial populations on textiles from different hospital departments. Similarly were the findings of Goldenberg et al. [49] that refer DHPLC as a suitable method for studying bacterial population dynamics in fermentation, bioremediation or environmental ecology [16]. Apart from that, Wagner et al. [26] describes the DHPLC method as suitable for a broad range of applications in microbial ecology. It is superior to other fingerprinting techniques in microbial ecology with easy application of conventional PCR, a high degree of automatization, and the possibility of quantifying fragments. A wide microbial diversity on hospital textiles described in this study points to the need for further research on microbial population's complexity on hospital textiles, since controlling the hygiene of hospital textiles is part of the surveillance programs for nosocomial infection prevention. Further research should especially be focusing on the possibility for transmission of pathogenic and multidrug resistant bacteria and virulent/resistant genes via contaminated textiles.

Conclusions
This study validates that DHPLC can be used to monitor the diversity of microbial populations on hospital textiles. Our research shows that the microbiota of hospital textiles is very complex and to some extent consistent with the microbiota of other hospital environments like devices, floor and workplaces. Therefore, hospital textiles are an important factor that contribute to the hospital microbiome and that should not be overlooked.