Current Microbiology

, Volume 65, Issue 2, pp 195–201

Biochemical and Microbial Analysis of Ovine Rumen Fluid Incubated with 1,3,5-Trinitro-1,3,5-triazacyclohexane (RDX)

Authors

    • Department of Environmental and Molecular ToxicologyOregon State University
  • A. Morrie Craig
    • Department of Veterinary MedicineOregon State University
Article

DOI: 10.1007/s00284-012-0144-1

Cite this article as:
Perumbakkam, S. & Craig, A.M. Curr Microbiol (2012) 65: 195. doi:10.1007/s00284-012-0144-1
  • 153 Views

Abstract

In this study, the rumen was assessed for its potential to detoxify RDX using molecular microbial ecology as well as analytical chemistry techniques. Results indicated significant loss (P < 0.05) of RDX in <8-h post incubation, and qualitative LC-MS/MS analysis showed evidence for the formation of 1-NO-RDX (M–O + HCOO) and methylenedinitramine metabolites. A total of 1106 16S rRNA-V3 clones were sequenced, and most sequences associated with either the phyla Bacteroidetes or Firmicutes. A LibCompare analysis for the RDX treatment showed an enrichment (P < 0.01) of the genus Prevotella. From these results, it can be concluded that the rumen is capable of detoxifying RDX, and the members of the genus Prevotella are linked to this detoxification.

Introduction

Hexahydro-1,3,5-trinitro-1,3,5-triazine (RDX; royal demolition explosive) is a high-energy trinitrated cyclic compound developed during World War II. RDX is also used with other munition mixtures and formulations with worldwide impact [11]. Owing to its extensive use, RDX is present as a contaminant in several military sites as munition waste and unexploded ordnance. The munition runoff causes potential harm to man and the environment. In humans, accidental ingestion of RDX-containing compounds has occurred several times and induces rash, nausea, as well as neurological effects including headaches, dizziness, and seizures [2].

Structurally, RDX is unabsorbed in the soil making it easy for its transportation into ground water aquifers [29, 44]. The present US Environmental Protection Agency (EPA) guideline for RDX quantity in drinking water is set at 2 ppm, but actual data suggest RDX prevalence as high as 36 ppm in the United States [2]. This suggests that the native microflora in soil and ground water do not have the capacity of detoxifying RDX and an introduction of an external microbial source is needed to complete remediation. With the use of genetically modified organisms (GMOs) debatable, and with the expense of ex situ remediation, many researchers are looking to develop affordable, practical, and safe technologies for remediation [37].

Anaerobic microbial transformation of nitroaromatics is also receiving increased attention because of the increased susceptibility of nitroaromatics molecules to degrade under anaerobic conditions [1, 8]. The animal rumen has been relatively well characterized in terms of its physiology [22]. Ruminants have tremendous potential in degrading or detoxifying numerous substrates ranging from grass-based carbohydrates to munitions [12, 15, 31] due to their anaerobic ecosystem and strong metabolic potential due to the diversity of microbes. Moreover, ruminants could be used as “bio-reactors on hooves” when combined with phytoremediation, using cool season grasses, thus presenting an easy accesses to in situ-based remediation [15]. The grasses would help in the transport of munitions from the soil to be available for the sheep to graze. Munitions also seem to have less toxic effect on ruminants. In a previous study [35], sheep were fed radiolabeled TNT, and the results indicate that the rumen converted most of the grass-bound TNT into organic matter without any toxic effect to the sheep. Tissue samples from the sheep showed no signs of radioactive damage suggesting that the microbes in the rumen play an important part in the detoxification. Such a novel idea of using phytoremediation combined with ruminal degradation can be helpful in bioremediation of nitroaromatic compounds [15].

In this study, the rumen was assessed for its potential to detoxify RDX by using classic molecular microbial techniques such as cloning of the 16S rRNA gene and phylogenetics. LibCompare and Libshuff analysis [42] tested differences between the samples (treatment and control) at the community level. Further, the metabolites of RDX detoxification were identified using LC-MS/MS analysis.

Materials and Methods

Whole Rumen Fluid Collection, Sheep Diet, and Experimental Design

Ovine whole rumen fluid (WRF) was collected as described before[30]. Sheep diet consisted of grass-hay. The experimental design consisted of three conditions: WRF with RDX (RDX-T), WRF without RDX (RDX-C), and autoclaved WRF biomass to which RDX was added after sterilization (RDX-SC). Balch tubes containing the WRF samples were sealed with sterile butyl rubber stoppers and aluminum crimp caps. RDX concentration of 25 μg mL−1 was used for all incubation experiments. All treatments were done in triplicate. Experiments were undertaken in an anaerobic glove box (Coy, Grass Lake, MI) with a mixed atmosphere of CO2 and H2 (9:1). The Balch tubes were incubated at 39 °C in the dark under constant rocking.

High-Performance Liquid Chromatography (HPLC) and LC-MS/MS Analyses

Time point samples were processed for both DNA and metabolite identification. The samples were spun at 10,000×g for 3 min to pellet the cells, and the supernatant was used for the HPLC and LC-MS/MS analyses. HPLC analyses were carried out using the US Environmental Protection Agency (EPA) method 8330 [41].

LC-MS/MS was used to qualitatively assess the presence of the parent molecule and metabolites in the rumen samples. Analysis was performed on an ABI/SCIEX QTRAP 3200 LC-MS/MS system (Applied Biosystems, Foster City CA) using a turbo spray interface in negative ion mode. Samples were separated on an HPLC system (Perkin Elmer Series 200 Micropump) using an Ultracarb ODS 250 × 4.6 mm, 5-μm particle size column (Phenomenex, Torrance CA). The method used for the separation of HMX involved a 15-μL injection volume followed by a mobile phase gradient program with A (methanol) and B (200 mM formic acid dissolved in ultrapure H2O) pumps. The HPLC was set to equilibrate at 100 % B for 5 min followed by a linear increase to 100 % A in 20 min. The column was re-equilibrated for 10 min with 100 % B. The flow rate was set at 300 μL min−1. The method was optimized by running the LC-MS/MS in the infusion mode with the parent molecule HMX and two important metabolites: methylenedinitramine (MEDINA) and 4-nitro-2,4-diazabutanal (NDAB). Data were acquired using multiple reaction monitoring (MRM) as the survey scans to generate MS/MS spectra within the Analyst 1.4.2 software package (Applied Biosystems). Based on these optimization runs, the final method had the following parameters: curtain gas (nitrogen) set at 30 psi, temperature at 450 °C, dwell time of 60 ms, gas 1 (GS1) = 45.00, gas 2 (GS2) = 45.00, and a scan range of 50–400 Da. Declustering potential, entrance potential, and collision energy were dependent on the ion being scanned (Supplemental data, Table S1).

Isolation of Genomic DNA, V3 Region Amplification, PCR Conditions, Cloning, and Plasmid Extraction

Genomic DNA was extracted from the cell pellets, as mentioned in the previous section, using the Gentra puregene kit (Qiagen, Valencia, CA) combining the extraction procedure for Gram-positive and Gram-negative bacteria. Tubes were left at room temperature to hydrate overnight and run on a 1 % agarose gel stained with ethidium bromide. Samples were quantified using a Nanodrop (Thermo Fisher, Waltham MA), stored at −20 °C and used for all subsequent PCR reactions.

The hypervariable region, V3, of the 16S rRNA was used as a gene marker. The primers and PCR amplification protocol used in this study have been described previously [27]. PCR thermocycling was carried out using recombinant AmpliTaq Gold polymerase (Applied Biosystems, Foster City, CA) in a PTC-200 thermocycler (MJ Research Inc., Watertown, MA). Each 50-μL PCR reaction contained approximately 75 ng of purified bacterial genomic DNA, 200 μmol of each dNTP, 5 μL of 10× PCR Buffer, 5 μL of 25 mM MgCl, 20 ng of bovine serum albumin (BSA), primer concentration at 25 pmol (each primer), and 0.25 U polymerase; the remaining volume was made up with sterile water. All PCR reactions were setup in triplicate, and the products were visualized on a gel and pooled before purifying using the QIAquick PCR purification kit, according to the manufacturer’s recommendations (Qiagen Inc., Valencia, CA). PCR products were quantified, cloned, and transformed into competent E. coli cells using the TOPO® TA Cloning Kit for Sequencing (Invitrogen Corporation, Carlsbad, CA), according to the manufacturer’s recommendations. Transformants were spread onto petri dishes containing Luria–Bertani (LB) agar (EMD Chemicals Inc., Gibbstown, NJ) supplemented with 50 μg mL−1 kanamycin sulfate (EMD). Plates were incubated at 37 °C overnight. Clones were picked and grown for 36 h in TYGPN [17] supplemented with 50 μg mL−1 kanamycin. Sterile glycerol was added to the colonies at a final concentration of 40 % and stored at −20 °C. The colonies were transferred into TYGPN media with 50 μg mL−1 kanamycin, and plasmid DNA was extracted as per previously published protocol [18].

Sequencing, RDPII, and Phylogenetic Analyses

Sequencing was performed using the BigDye® Terminator v. 3.1 Cycle Sequencing Kit (Applied Biosystems, CA) using an ABI Prism® 3730 Genetic Analyzer at the Center for Genomic Research and Biocomputing (CGRB) at Oregon State University. Single reads utilizing the T7 promoter were used to determine the nucleotide sequences. The sequences were imported into the Geneious computer program [14] extracted and checked for chimeras, and the resulting FASTA file was used for further analysis. The RDP II Classifier [42] was used to sort sequences into their respective operational taxonomic units (OTUs) at a confidence interval of 50 % [9]. The LibCompare software of the RDPII database was also used to compare significance of community changes at a confidence interval of 50 % [42]. The MOTHUR software package [34] was used for the analysis of the data and estimation of collectors and rarefaction curves. All sequence data from this study were submitted to the GenBank database [6] under accession numbers HQ159874–HQ161053.

Results and Discussion

Although ruminants consume forage as primary mode of nutrition, they possess diverse microflora that has been used to help detoxify several plant toxins such as pyrrolizidine alkaloid (PA) [32], oxalate [3], nitropropanol [5], dihydroxypyridine [13], and munition such as TNT [12, 31] and RDX [15].

In this study, based on the HPLC data, 98 % decrease in parent RDX molecule (25 μg mL−1 to below detection limit) occurred within 4 h of incubation in WRF (Fig. 1a). There was no parent molecule present at 8-h post incubation. Such quick reduction times have been seen with other incubation studies involving WRF and munitions such as TNT [12], RDX [15], and HMX (Unpublished data). To determine if there was a complete breakdown of RDX at 4 h, samples were injected into LC-MS/MS (Fig. 1b). LC-MS/MS results showed that there were trace amounts of RDX present at 4-h time point, but complete removal of RDX occurred before 8-h sampling period. Such difference in quantification using the two methods is likely due to the detection limits of the instruments (HPLC vs LC-MS/MS). Although complete RDX degradation occurred in 8 h, the rate of degradation is still quick compared with other reports, thus, making the rumen one of most efficient and quick bioreactors [7, 20, 21].
https://static-content.springer.com/image/art%3A10.1007%2Fs00284-012-0144-1/MediaObjects/284_2012_144_Fig1_HTML.gif
Fig. 1

Degradation of RDX by ovine rumen fluid. Panel A RDX (25 μg mL−1) was incubated in Balch tubes and time point samples analyzed using HPLC. Shown are RDX detoxification in WRF (filled square) and (open square) an autoclaved WRF biomass control at times 0-, 4-, and 8-h. Values indicate the RDX concentration remaining in the samples at the sampled time points. Error bars represent ± 1 standard deviation of three replicate analyses. Panel B LC-MS/MS analysis of the parent RDX at the time points 0-, 4-, and 8-h post incubation with RDX. Collected time point samples were injected into a LC-MS/MS and analyzed as mentioned in the “Materials and methods” section. Peaks indicate the loss of parent molecule at 0-, 4-, and 8-h post incubation

Based on earlier research, three mechanisms of RDX transformation have been proposed: two-electron reduction, single-electron reduction/denitration, and direct enzymatic cleavage [11]. The qualitative LC-MS/MS results correlate well with those of the previously published pathway and metabolite analysis of RDX [26]. LC-MS/MS analysis of the 4-h time point identified the parent RDX ion at a m/z of 266.95 Da and mononitroso RDX intermediate hexahydro-1-nitroso-3,5-dinitro-1,3,5-triazine (MNX) with formic acid adduct at 250.9 Da (Supplemental data, Table S1). There was no presence of dinitroso (DNX) and trinitro (TNX)—the other metabolites of RDX degradation. One other important intermediate, methylenedinitramine (O2NNHCH2NHNO2, MEDINA), was found in the analysis. MEDINA was identified at 4 h after the incubation with RDX at 134.90 Da [45]. These intermediates form the “classic” metabolites that were a part of a previous discovery. The most published literature show the presence of nitroso derivatives (MNX, DNX, and TNX) or MEDINA, but not both metabolites. Owing to the presence of both indicator metabolites, one could say that rumen breaks down the RDX molecule into more than one process. This is mainly due to the diversity of microbes in the rumen and the substrate utilization capable to derive energy or breakdown pollutants. The absence of nitroso products could either be due to the transient nature of the metabolites [11], production in less quantity, or other degradation mechanism (Fig. 2).
https://static-content.springer.com/image/art%3A10.1007%2Fs00284-012-0144-1/MediaObjects/284_2012_144_Fig2_HTML.gif
Fig. 2

Rarefaction plot of RDX (T) and RDX (C) conditions with the 16S rRNA-V3 primers as calculated by the MOTHUR program. The estimates plotted represent 3 % cutoff numbers of observed OTUs in relation to the number of sequences sampled. All filled symbols represent 3 % cutoff; (filled circle) represents the observed OTUs at 3 % difference in the RDX (C) 0-h sample; (filled square) at 3 % difference in the RDX (C) 8-h sample; (filled triangle) at 3 % difference in the RDX (T) 0-h, (filled inverted triangle) represents the 3 % difference in the RDX (T) 4-h, and (filled diamond) represents the 3 % difference in the RDX (T) 8-h post incubation

These results present a comprehensive study of prokaryotic 16S rRNA-V3 gene diversity found within the rumen, in relation to the degradation of RDX. A total of 1,180 non-chimeric sequences were obtained for both the control and treatment groups. The prokaryotic diversity present in the ovine whole rumen fluid is impressive. The sheep consume grass-hay to derive energy and so have anaerobic microorganisms represented by all three phyla, namely, bacterial, archea, and fungi [22, 33]. The two major bacterial phyla represented in the rumen are also prevalent in other gut ecosystems such as humans [40], birds [19], and mouse [39]. From the results of the RDP II Classifier analysis [42], the predominant phyla were determined to be Bacteroidetes and Firmicutes (Table 1). The percentage of sequences ranged from 66.79 to 91.93 % either in control or treatment samples. One interesting observation is the similarity of diversity (phyla associations) between ruminants and humans. Although ruminants are foregut fermenters and human’s hindgut fermenters, both have the two major phyla: Bacteroidetes/Firmicutes. The minor phyla consisted of clones from Actinobacteria, Proteobacteria, TM7, and Synergistetes. These results correlate well with the other published research that has characterized the rumen using the 16S-based gene marker [16, 30, 38, 43]. Evaluations of ruminal microbial communities in cattle [16, 38, 43], sheep [24, 31] and some wild ruminants [4, 28] have shown them to be dominated by low GC Gram-positive bacteria (LGCGPB), particularly those related to the broad genus Clostridium.
Table 1

Tabulated classification of the 16S rRNA-V3 clones associated with RDX (T) and RDX (C) sampled at 0-, 4-, and 8-h at the phylum level using the Naïve Bayesian classifier of the RDPII website at a confidence interval of 50 %

Treatment

RDX-T (0)#

RDX-C (0)

RDX-T (4)#

RDX-T (8)#

RDX-C (8)

No of clones

Total (%)

No of clones

Total (%)

No of clones

Total (%)

No of clones

Total (%)

No of clones

Total (%)

Phylum

Actinobacteria

3

(1.21)

1

(0.41)

1

(0.40)

    

Bacteroidetes

24

(9.71)

137

(56.61)

27

(10.84)

114*

(54.02)

107

(46.32)

Firmicutes

141

(57.08)

86

(35.53)

141

(56.62)

80*

(37.91)

91

(39.39)

Proteobacteria

6

(2.42)

  

12

(4.81)

    

Synergistetes

2

(0.80)

1

(0.41)

7

(2.81)

  

2

(0.86)

Tenericutes

      

1

(0.47)

  

TM7

42

(17.00)

5

(2.06)

27

(10.84)

8

(3.79)

12

(5.19)

aUncla bacteria

29

(11.74)

12

(4.95)

34

(13.65)

8

(3.79)

19

(8.22)

 

247

100.00

242

100.00

249

100.00

211

100.00

231

100.00

aUnclassified bacteria

#Libshuff significance between treatments (P < 0.0001)

* LibCompare significance between treatments (P < 0.01)

Despite having high-quality sequences with which to search RDPII database, 42.35 % or 499.73 of the 1,180 sequenced clones did not associate to any well-characterized bacterial groups in the RDPII database. Neither of the cultured representatives is available or the region chosen for amplification, and in this case, V3 of the 16S rRNA gene is not appropriate for rumen-associated studies. Most researchers have either used the V6 region [23, 36] or V3 region, with a few researchers claiming both are equally good in differentiating communities [10].

The MOTHUR analysis [34] rarefaction curves showed no trend toward reaching a plateau, which also indicates that we have not yet come close to sampling enough clones to observe all the diversity present in this system. The results in the study also are indicative of many potentially uncharacterized microbes that exist within the rumen. The total observed OTUs for all treatments were from 131 to 142. The ChaoI estimated OTUs for all groups were between 260.11 and 376.39 (Table 2). These results are comparable to the previously sampled rumen data [16].
Table 2

Phylogenetic analyses of RDX (T) and RDX (C) treatments at sampling time point’s 0-, 4-, and 8-h. Observed and estimated OTUs were calculated using the MOTHUR program. The OTUs common between treatments are also tabulated. All data represented OTU classification at a cutoff at 3 %

Treatment

Observed OTUs

Common OTUs

Estimated OTUs

Common OTUs

RDX

RDX

RDX

RDX

RDX

RDX

RDX

RDX

RDX

RDX

T (0)

C (0)

T (4)

T (8)

C (8)

T (0)

C (0)

T (4)

T (8)

C (8)

RDX–T (0)

131

 

29

4 7

28

36

260.11

 

65.82

74.25

78.85

91.83

RDX–C (0)

142

  

34

50

56

362.50

  

73.63

111.95

166.45

RDX–T (4)

136

   

27

33

304.00

   

117.83

61.34

RDX–T (8)

139

    

44

376.39

    

170.14

RDX–C (8)

134

     

308.41

     
To estimate statistical differences between OTUs that were enriched because of treatment with RDX, the Libshuff and Libcompare tools of the RDPII database [42] were used (Table 3). Libshuff results showed that there were statistically significant differences (P < 0.0001) between the three clonal libraries at time points 0-, 4-, and 8-h post incubation. The LibCompare analysis done for the same three times points showed a decrease in trend for clones associated with TM7 and Clostridia. Clones associated to Prevotella sp. saw an increase from 12 to 85 clones, which were significant (Table 3). Such enrichment of single bacteria has been seen before in rumen samples incubated with TNT [31].
Table 3

Libcompare comparison of the RDX clonal libraries at 0-, 4-, and 8-h post incubation. The sequences were submitted to LibCompare software of the RDPII database and significant shifts in populations among time points were estimated

Bacterial classification

RDX (O hr)

RDX (4 h)

RDX (8 h)

aSignificance

Phylum

Firmicutes

 

141

85

4.13E−5

Class

Clostridia

 

134

74

4.13E−5

Family

Ruminococcaceae

 

45

18

2.62E−3

Genus

Lachnobacterium

8

  

4.13E−3

Phylum

TM7

 

27

8

3.98E−3

Genus

TM7 genera Incertae sedis

 

27

8

3.98E−3

Phylum

Bacteroidetes

 

26

114

6.00E−14

Class

Bacteroidia

 

20

101

6.00E−14

Family

Prevotellaceae

 

14

98

6.00E−14

Genus

Prevotella

 

12

85

6.00E−14

aLibCompare significance at (P < 0.01)

Prevotella strains are Gram-negative, non-motile, rod-shaped, and singular cells that thrive in anaerobic growth conditions. Members of the genus Prevotella are regarded as the most dominant bacterium in the rumen (Stevenson and Weimer 2007). Prevotella sp. are among the most numerous microbes cultivable from the rumen and hind gut of cattle and sheep, where they help in the breakdown of protein and carbohydrate derived from plant material [25]. They are also present in humans as opportunistic pathogens. Two strains of Prevotella sp. (Prevotella intermedia 17 and Prevotella ruminicola 23) have been completely sequenced by The Institute for Genomic Research (TIGR). In silico analysis (data not shown) of sequenced Prevotella genomes indicated the presence of a nitroreductase enzyme system that is responsible for the breakdown of TNT. Such an enzyme system has also been shown to be active with RDX [11].

In conclusion, this study shows the potential of rumen fluid to degrade RDX. Further research has presently been undertaken in establishing the microbial population using stable isotope probing (SIP) in combination with next generation sequencing (NGS). We are also fine tuning our analytic procedure to tease apart this complex matrix to better understand the metabolites. Finally, we are incubating pure cultures of Prevotella sp. with RDX to decipher the metabolites and regulation of detoxification.

Acknowledgments

The research was supported by a jointly funded grant by the Oregon Agricultural Experiment Station project ORE00871 and by the U.S. Department of Agriculture under project number 6227-21310-007-00D agreement nos. 58-6227-8-044 and 58-1265-6-076. Any opinions, findings, conclusion, or recommendations expressed in this publication are those of the author(s) and do not necessarily reflect the view of the U.S. Department of Agriculture. The authors would like to thank Ms. Karen Walker for her help with HPLC, Lia Murty during the LC-MS/MS analysis, and Ms. Zelda Zimmerman for editorial assistance.

Supplementary material

284_2012_144_MOESM1_ESM.doc (44 kb)
Supplementary material 1 (DOC 43 kb)

Copyright information

© Springer Science+Business Media, LLC 2012