Introduction

Metal removal fluids (MRFs) are used to lubricate machining contact surfaces, cool the working zone, prevent corrosion, and remove metal chips in metalworking processes, i.e., metal deformation and removing. They play a pivotal role in modern industries related to metal cutting and metal forming, which includes the automotive, aerospace, defense, and construction sectors. They have high annual consumption volume, with over 2 billion gallons used per year. About one million US workers are exposed annually to MRFs ((NIOSH) 1999). During their lifecycle, water-based MRFs are prone to microbial growth, impacting fluid integrity as well as giving rise to the emergence of opportunistic pathogens, creating risk of health problems, including respiratory illnesses, cancers, and dermatitis (Bennett 1972; Bennett and Bennett 1987; Mattsby-Baltzer et al. 1989; Mattsby-Baltzer et al. 1990; Tolbert et al. 1992; Savonius et al. 1994; Rossmoore 1995; Eisen et al. 1997; Simpson et al. 2003). Some reports have suggested that exposure to the aerosolized microbial antigens present in used fluids may cause respiratory disorders such as asthma, hypersensitivity pneumonitis, bronchitis, and other respiratory symptoms (Kennedy et al. 1989; Greaves et al. 1997; Gupta and Rosenman 2006). The presences of Pseudomonads and Mycobacteria in MRFs have been implicated in respiratory disorders such as hypersensitivity pneumonitis in the exposed machine workers (Mattsby-Baltzer et al. 1989). Pseudomonads and their endotoxins were initially considered the main cause of HP (Bernstein et al. 1995), whereas later reports linked Mycobacteria with HP(Kreiss and CoxGanser 1997; Shelton et al. 1999; Wilson et al. 2001; Wallace et al. 2002). In the past decade, more than two hundred hypersensitivity pneumonitis (HP) cases have been reported among workers in MRF processes and most of them have been associated with Mycobacterium immunogenum (MI) (Moore et al. 2000; Wilson et al. 2001; Gordon 2004; Gordon et al. 2006; Thorne et al. 2006). These reports highlight the need for more detailed analyses of these fluids and their aerosols for total microbial load including both culturable and nonculturable microorganisms.

Monitoring and regulation of the microbiology relevant to MRFs has relied primarily on culture-based analysis that specifically target classical pathogens. Although culture can be successful for assessment of some microbes, a large body of DNA sequence-based studies has shown that standard enrichment techniques significantly underestimate the actual quantity and diversity of microorganisms in a wide variety of environments (Pace 1997). Therefore, the goal of this study was to better understand the microbial composition in MRFs, with special focus Mycobacteria including M. immunogenum and the role of biofilms, using molecular techniques in addition to traditional culture-based approaches.

Materials and methods

MRFs sampling

A set of fifteen samples were collected on the same day from six different MRF systems in an automobile engine plant including ten bulk fluid samples and five biofilm samples (list of 15 samples are summarized in Table 1). Two biofilm samples (biofilm 1 and biofilm 4) were collected by scraping biofilms from the floor close to the machining and other three biofilm samples were collected by scraping biofilms from inside surfaces of MRF tanks. Two types of soluble oils, semi-synthetic, and synthetic oil, were used in those systems. An isothiazolone class of biocide had been added to all fluids. The collected samples were kept on wet ice and transported to the laboratory at the University of Michigan for immediate analysis within 24 h of collection.

Table 1 Summary of data on the microbial characterization of fluid and biofilm samples collected from different systems in an automotive engine plant using plate counting and PCR amplification methods
Full size table

Quantification of microorganisms using culture-based method

MRF samples were serially diluted with 1 × PBS buffer and seeded onto agar plates. The total bacterial viable counts were estimated as CFU on nutrient-rich Tryptic Soy Agar plates (0.1 g liter−1 cycoheximide to inhibit fungal growth) after aerobic incubation at 30 °C for 5 to 7 days. The total viable Mycobacteria counts were estimated as CFU on Middlebrook 7H9 agar plates after incubation at 30 °C for 2 weeks.

DNA Extraction and PCR Amplification

One milliliter of each MRF sample was taken for DNA extraction using UltraClean Soil DNA Kit (Cambio Ltd., Cambridge, UK) according to the manufacturer’s protocol. DNA was finally eluted with 50 μL of water. The quality and concentration of DNA was measured spectrophotoscopically using Nanodrop ND-100 (NanoDrop Technologies, Wilmington, Delaware, USA). DNA extracted from MRFs samples were amplified with the universal primer pairs 515F (5′-GTGCCAGCMGCCGCGGTAA-3′) and 1391R (5′-GACGGGCGGTGWGTRCA-3′) targeting small ribosomal RNA gene of bacteria and fungi (Angenent et al. 2005). PCR reactions were carried out in 50 μL volumes containing 400 nM of each primer, 1 μL of DNA samples, 10 μL of 5 × PCR buffer, 4 μL of dNTPs mix (2.5 mM of each dNTPs), and 0.25 μL of GoTaq DNA polymerase (Promega, Madison, WI). PCR was conducted with a Mastercycler gradient Machine (Eppendorf, Westbury, NY) by running 20 cycles with a gradient from 65 to 45 °C (92 °C for 30s, 65 °C to 1 °C/cycle for 30 s, 72 °C for 90 s) and an additional 20 cycles at a 45 °C annealing temperature (92 °C for 30 s, 45 °C for 90 s, 72 °C for 90 s) before a final extension at 72 °C for 20 min. Mycobacteria genus-specific amplification were conducted with the primer pair targeting Mycobacterial 16S rRNA gene, 515F and 1027R (5′-GCACACAGGCCACAAGGG-3′) (Angenent et al. 2005). MI species-specific amplification was conducted with the primer pair targeting MI genes pMycImmF (5′-GGGGTACTCGAGTGGCGAAC-3′) and pMycImmR (5′-GGCCGGCTACCCGTTGTC-3′) (Veillette and Thorne 2005). PCR reaction mixtures were the same to that mentioned above. These two sets of PCR were conducted with an initial denaturation at 94 °C for 2 min followed by 30 cycles of the following incubation pattern: 94 °C for 30 s, 63 °C for 30 °C, and 72 °C for 40 s. A final extension at 72 °C for 10 min concluded the reaction.

16S rRNA gene clone library construction and sequencing

The PCR products amplified with universal primers were checked by electrophoresis on a 2 % agarose gel, stained with SYBR Green I (Cambrex, Rockland, ME. USA), and visualized with a UV transilluminator. The approximately 0.8 and 1.2-kb heterologous 16S rRNA and 18S rRNA gene products were excised from the gel, and the PCR products were purified from the gel slices by using a Wizard PCR prep kit (Promega, Madison, Wis.). The gel-purified PCR products were cloned into the pGEM-T easy cloning vector as specified by Promega (Promega, Madison, WI) and transformed into calcium chloride-competent Escherichia coli JM109 cells according to the manufacturer’s instructions and standard techniques (Promega, Madison, WI). Plasmid DNA was isolated from individual clones by using a Wizard Plus SV minipreps DNA purification system (Promega, Madison, WI). Aliquots from a subset of purified plasmid DNA were digested with the restriction enzyme EcoR I (TaKaRa, Otsu, Shiga, Japan) for 4 h at 37 °C, and the digested products were separated by electrophoresis on a 2 % agarose gel. After being stained with SYBR Green I, the bands were visualized using a UV transilluminator to select clones containing the appropriately sized insert. The clones with the correct plasmid insert were then used for DNA sequencing. Sequencing was performed by The University of Michigan’s DNA Sequencing Core using ABI PRISM dye terminator cycle sequencing ready reaction kit with AmpliTaq DNA polymerase FS and an ABI model 3730 DNA sequencer (Applied Biosystems, Foster, CA. USA). All sequences were deposited in the NCBI sequence database (http://www.ncbi.nlm.nih.gov/) under access numbers from EU118986 to EU119153.

Sequence analyses and phylogenetic tree analysis

Primary screening of these clone libraries for the presence of chimeric sequences was done using Mothur program (Schloss et al. 2009). The screening results were further analyzed with MEGA (Molecular Evolutionary Genetics Analysis, version 3.1) (Kumar et al. 2004) to dereplicate the libraries of 16S rRNA and 18S rRNA gene sequences for subsequent analyses by comparing all the sequences in a data set to each other, grouping sequences with ≥97 and ≥90 % identity together, and outputting a representative sequence from each group. The all-rRNA gene sequences of each group were first compared with those in the GenBank database using the basic local alignment search tool BLAST (http://www.ncbi.nlm.nih.gov/BLAST/BLAST.cgi). A 97 to 100 % match of the unknown sequence with the GenBank data set was considered an accurate identification to the species level, 93 to 96 % similarity was accepted as genus-level identification, and an 86 to 92 % match was considered an accurate identification of a related organism (Stackebrandt and Goebel 1994). The sequence data for reference strains were obtained from the GenBank database. Phylogenetic trees were constructed using neighbor-joining method and graphically represented with MEGA software.

Microbial community similarity comparison

Microbial community similarity comparison was conducted using UniFrac program on the website (http://bmf.colorado.edu/unifrac/index.psp). Briefly, sequences from the four libraries were aligned and a single phylogenetic tree containing all sequences was created using MEGA software and exported in Newick format. Then, a text file was created that contained the name of each sequence, the name of the environment that it came from, and the number of times each clone was observed. Then, the PCoA and Jackknife Cluster option were selected to perform UniFrac analysis.

Real-time PCR quantification of total bacteria and MI

Total bacteria quantification was carried out by using real-time PCR according to Einen et al. (Einen et al. 2008). PCR reactions mixture were in 20 μL volumes containing 500 nM of primer Eu338 (5′-ACTCCTACGGGAGGCAGCAG-3′) and Eu518 (5′-ATTACCGCGGCTGCTGG-3′), about 10 ng of template DNA, and 10 μL of 2 × SYBR Green PCR Master Mix (Applied Biosystems, Foster City, CA). PCR program was as follow: 95 °C for 15 min, 40 cycles of denaturing (15 s at 94 °C), annealing (30 s at 61 °C), and extension (30 s at 72 °C). The cycling was followed by a final extension at 72 °C for 7 min, and a melting curve analysis from 65–95 °C with a plate read every 0.5 °C. Quantification of MI was accomplished by a TaqMan assay according to Veillette et al. (Veillette et al. 2008). Primers pMycImmF/pMycImmR and a dual-labeled probe (FAM-5′-CCGCATGCTTCATGGTGTGTGGT-3′-BHQ1) were used for specifically detection of MI. PCR mixture comprises 1 × QuantiTect probe PCR buffer (Qiagen, Valencia, CA), 400 nM of each primer and 100 nM of probe. The program consisted of 15 min hot start at 94 °C, followed by 40 cycles of 3 s at 94 °C and 1 min at 61.5 °C. Fluorescence was acquired by detection of FAM (excitation, 495 nm; emission, 520 nm). As calibration standard, a dilution series of quantitative standard DNA was used as template (Ochsenreiter et al. 2003; Kemnitz et al. 2005). It was prepared from strain E. coli K-12 or M. immunogenum ATCC 200506, respectively. The standard DNA solution was then diluted to 108 copy of target DNA molecule per microliter and stored at −20 °C.

Results

Viable bacteria and Mycobacteria from MRFs samples

In 11 out of 15 samples bacterial colonies formed on the TSA plates. Five samples (Biofilm-2, Biofilm-4, Fluid-9, Fluid-10 and Biofilm-5) formed colonies on the Mycobacteria cultivation agar 7H9 suggesting the likely presence of Mycobacteria in these five samples (Table 1). The number of colony-forming units on the TSA plates and 7H9 plates were not counted because of inconsistencies of these numbers with serial dilutions of the original fluid samples, which may be due to some components in the fluids may have inhibitory factors.

PCR-based detection of Mycobacteria and M. immunogenum in MRFs samples

Detection of Mycobacteria and M. immunogenum in the collected MRF samples were performed using the Mycobacteria genus-specific primers 515F/1027R and species-specific primiers pMycImmF/pMycImmR, respectively. Among the 15 samples, the presence of Mycobacteria in the five samples detected by using culture-based method was confirmed with PCR amplification of 16S rRNA gene fragment using Mycobacteria genus-specific primers (data not shown). Two additional samples, Biofilm-1 and Biofilm-3, were positive in PCR amplification with expected 513 bp DNA fragment for Mycobacteria, although no colonies were formed on the 7H9 cultivation media. All seven samples that had positive PCR amplicons of expected 220 bp DNA fragment using Mycobacterium genus-specific primers also had positive amplicons using M. immunogenum species-specific primers (confirmed by DNA sequencing). PCR analyses based on genus-specific PCR were summarized in Table 1.

Bacterial and fungal diversity in MRFs

To understand the composition of microorganisms in MRFs samples, partial SSU rRNA gene libraries of four samples (Fluid-7 and Biofilm-3 from System 5 and Fluid-9 and Biofilm-5 from System 6) were prepared by cloning PCR products amplified with universal primers 515F/1391R targeting SSU rRNA gene (Angenent et al. 2005). Analysis of libraries based on universally conserved primers provided an estimate of the overall phylogenetic diversity in the samples. Fifty clones were randomly picked up from each SSU rRNA gene library for further DNA sequencing. Five small Mycobacteria SSU rRNA gene libraries using Mycobacteria genus-specific primers were constructed from Mycobacteria-genus PCR positive samples (Biofilm-1, Biofilm-2, Biofilm-3, Biofilm-4, Biofilm-5, Fluid-9, and Fluid-10) and 10 clones from each clone library were randomly selected for DNA sequencing.

Sequences were compared with sequences in the public database, and phylogenetic trees were generated using neighbor-joining method for each sample (Fig. 1). DNA sequencing of 50 clones from each library revealed that Paecilomyces sp. and Acremonium sp. and Acremonium-like were dominant fungal genera in all samples. More than 50 different bacterial species from the main phylum of proteobacteria were found in these four samples, and Pseudomonas was the dominant bacterial genus in these samples (>11 %). The microbial compositions of the different samples were highly variable. Among the clones sequenced, 185 (92.5 %) showed ≥97 % rRNA sequence identity to a database sequence, indicating a species-level relationship between the environmental organisms and the database reference organisms.

Fig. 1
figure 1

Neighbor-joining phylogenetic tree based on nearly complete SSU rRNA gene sequences from four samples (A, Fluid-9, B, Fluid-7, C, Biofilm-5, and D, Biofilm-3). Bootstrap values were shown at the major branches. The scale at the bottom indicates sequence divergence. The numbers of clones that have the same sequences are indicated

Full size image

Similarity comparison of microbial community in MRFs samples

UniFrac analysis was used to determine whether or not these microbial compositions are significantly different (Lozupone and Knight 2005; Lozupone et al. 2006). Figure 2 shows a UniFrac clustering diagram comparing the communities from four different samples and two different systems. Sample Fluid-7 and Biofilm-3 from System5 clearly cluster together and samples Fluid-9 and Biofilm-5 from System 6 are separated from the above two samples. The data suggest that microbial compositions in the suspended fluid samples and biofilm samples from the same system are more similar to each other than those from different systems.

Fig. 2
figure 2

Distribution of community samples in UniFrac principal coordinates analysis (PCoA). Unweighted UniFrac was used in the comparison. Samples Fluid-7 and Biofilm-3 from System5 clearly cluster together and samples Fluid-9 and Biofilm-5 from System 6 are separated from the above two samples from System 6

Full size image

Quantification of M. immunogenum and bacteria in MRFs samples

In order to monitor the abundance of M. immunogenum in different MRFs samples, a quantitative PCR assay that allows estimating the relative amount of 16S rDNA of this species in the collected samples was developed. For this purpose, two methods (Taqman and SYBR real-time PCR) specific for MI and general bacterial domain, respectively, were used. Based on the calibration curves generated for MI and bacteria from known amounts of template DNA, the copy number of MI and bacterial 16S rRNA genes were calculated in each collected sample. This allowed determination of the relative abundance of MI in the different MRFs samples. The results are summarized in Table 1. The data show that the number of copies of total bacterial 16S rRNA genes in all samples ranges from 106 to 1010 copy/ml, and the number of copies of M. immunogenum 16S rRNA gene in 7 out of 15 samples ranges from 4.3 × 104 copy/ml to 4.6 × 107 copy/ml. The ratio of M. immunogenum to total bacteria, indicating the relative abundance of M. immunogenum, ranges from 0.002 to 4‰. The relative abundance of M. immunogenum in biofilm samples is high, especially in samples Biofilm-1 and Biofilm-4 collected from the floor adjacent to the MRF system. The relative abundance of M. immunogenum in biofilm sample (Biofilm-5) is 5 or more times higher than that of M. immunogenum in the corresponding fluid samples (Fluid-9 and Fluid-10).

Discussion and conclusion

The present study, though limited in the number of samples collected, documents that traditional culture-based methods underestimate the full diversity of the microbial composition in MRFs. Culture-based and PCR methods were used to check the presence of microorganisms in MRFs samples. Although four out of 15 samples had no bacterial growth on TSA plates, all samples had bacterial and fungal detected by PCR method when the universal primers, targeting 16S rRNA gene of bacteria or 18S rRNA gene of fungi, were used. In addition, five samples formed colonies on Mycobacteria cultivation agar 7H9 suggesting the likely presence of Mycobacteria in these five samples. The presence of Mycobacteria in these five samples was confirmed with PCR amplification of 16S rRNA gene fragment using Mycobacteria specific primers. Two additional samples, Biofilm-1 and Biofilm-3, were positive in PCR amplification for Mycobacteria, yet no colonies formed on the 7H9 cultivation media. These five Mycobacteria-contaminated samples also contained M. immunogenum cells (which confirmed by DNA sequenceing of clone library). In addition, analysis of libraries based on Mycobacteria-specific primers provided an estimate of the overall phylogenetic diversity of Mycobacteria in the samples. The results showed all the DNA sequences amplified with Mycobacteria-specific primers had high identities to the 16S rRNA gene of M. immunogenum (>99 %) (total of 50 clones were sequenced, data not shown). These data imply that M. immunogenum was the dominant Mycobacteria species in all five Mycobacteria-contaminated samples.

Real-time PCR was used to quantify the total bacteria and the abundance of M. immunogenum in these samples. The data show that the number of copies of total bacterial 16S rRNA genes in all samples ranges from 106 to 1010 copy/ml and the number of copies of M. immunogenum 16S rRNA gene in 7 out of 15 samples ranges from 4.3 × 104 copy/ml to 4.6 × 107 copy/ml. The ratio of M. immunogenum to total bacteria, indicating the relative abundance of M. immunogenum, ranges from 0.002 to 4 ‰ (Table 1). The relative abundance of M. immunogenum in biofilm samples is high, especially in samples Biofilm-1 and Biofilm-4 collected from the floor. The relative abundance of M. immunogenum in biofilm sample 5 (Biofilm-5) is five or more times higher than that of M. immunogenum in the corresponding fluid samples (Fluid-9 and Fluid-10). Compared to the data from the culture-based analyses, the much higher concentration of M. immunogenum in these samples revealed by real-time PCR indicates that real-time PCR is significantly more sensitive than culture-based techniques, and/or a certain fraction of M. immunogenum in MRFs might be uncultivable or dead.

The rDNA sequences reveal the basic characteristics of environmental microorganisms related to known organisms. Our analyses detected and identified a broad diversity of microbes, more than 50 different strains and species. Most of them belong to fungi and alpha-, beta-, gamma-, and other proteobacteria (Fig. 1). Among these strains, Paecilomyces sp. and Acremonium sp. and Acremonium-like strains were dominant in fungi (>67 %), and Pseudomonads were the dominant gamma-proteobacteria strains (>23 %). Only 2 M. immunogenum DNA sequences were detected in Fluid-9 while no M. immunogenum sequences were found in Biofilm-5 although Biofilm-5 was shown to have M. immunogenum revealed by PCR analysis. This observation indicates that many more clones should be sequenced in order to obtain a better representation of microbial diversity in the MRF environment, especially to detect microorganisms of low abundance. It was interesting that there were eight clones identified as uncultured bacterium clones in Fluid-7 (16 % of total microbes).

Moreover, in sample Fluid-7, we found some clones were linked to geochemical recycle related strains, such as a DNA sequence with high identity to SSU rRNA gene of Rhodocyclus sp. HOD 5 (AY691423, 99 %), which was isolated from hydrogen-oxidizing-denitrifying bioreactor for treatment of nitrate-contaminated drinking water (Smith et al. 2005); a DNA sequence with high identity to SSU rRNA gene of Desulfuromonas acetexigens (DAU23140, 99 %), which has the ability to reduce Fe(III) and S0 (Lovley et al. 1995); and a DNA sequence with high identity to SSU rRNA gene of an uncultured bacterium clone TANB77 (AY667263, 93 %), which was detected in a dechlorinating community (Macbeth et al. 2004). These results suggest that some bacteria in MRF bacterial communities have the ability to use the ingredients in MRFs, such as mineral oil, monoethanolamine, and surfactant, as their nitrogen and energy source to support their growth and further supply nutrients for the growth of other strains.

This study enables several important observations. First, Mycobacteria (more specifically M. immunogenum) were found in all biofilm samples, including samples from inside the MWF systems and samples from adjacent environmental surfaces. In addition, the relative abundance of M. immunogenum in biofilm samples was much higher than that in the corresponding fluid samples. These findings strongly support that biofilms play an important role in the microbial ecology of MRFs, which confirmed previous studies also suggested that biofilms play a critical role in the microbial ecology of MWFs (Mattsby-Baltzer et al. 1989; Thorne et al. 1996; Skerlos et al. 2001). Biofilms may provide a shield or sheltered microenvironment for the growth and survival of Mycobacteria and act as seeding sources for circulating MRFs. To control growth of Mycobacteria in MRFs, the primary target may need to be biofilms.

Second, in one particular system in this study, S6, using one type of soluble oil, growth of Mycobacteria was found in both fluid and biofilm samples and the relative abundance of M. immunogenum in biofilm samples was much higher than that in its corresponding fluid samples. It had been previously reported by the company that Mycobacteria contamination was present in this MRF system about 6 months prior to obtaining the present samples. The old MRF was discharged; the whole system was cleaned and recharged with this new type of MRF (soluble oil, which is designed for effective control of microbial growth including Mycobacteria). This system would be a valuable target system in which to study specific components in this MRF that may promote the growth of Mycobacteria in fluids or associated biofilms (Veillette et al. 2004; Selvaraju et al. 2008), or other factors in the system, e.g., certain locations where biofilms are abundant in the system may provide seeding of Mycobacteria for the new fluid.