Automated dead-end ultrafiltration of large volume water samples to enable detection of low-level targets and reduce sample variability.

AIMS: A Portable Multi-use Automated Concentration System (PMACS) concentrates micro-organisms from large volumes of water through automated dead-end ultrafiltration and backflushing. The ability to detect microbial targets from ground, surface and cooling tower waters collected using standard methods was compared with samples from the PMACS in this study. METHODS AND RESULTS: PMACS (100 l) and standard grab samples (100-500 ml) were collected from sites in Florida and South Carolina, USA. Samples were analysed for the presence of faecal indicator bacteria (FIB; ground and surface water) or Legionella pneumophila (Lp; cooling tower water). FIB were enumerated by growth on selective media following membrane filtration or in IDEXX defined substrate media. Lp cells were detected by direct fluorescence immunoassay using FITC-labelled monoclonal antibodies targeting serogroups 1, 2, 4 and 6. FIB were found in PMACS samples from ground and surface waters when their concentrations were below detection limits in grab samples. The concentrations of Lp in cooling tower samples collected over 5 months were more consistent in PMACS samples than grab samples. CONCLUSIONS: These data demonstrate that PMACS concentration is advantageous for water monitoring. FIB were detected in PMACS samples when their concentrations were below the detection limits of the standard methods used. PMACS processing provided more representative samples of cooling tower waters reducing sample variability during long-term monitoring. SIGNIFICANCE AND IMPACT OF THE STUDY: This study highlights the utility of PMACS processing for enhanced monitoring of water for low-level microbial targets and for reducing sample variability in long-term monitoring programmes.

Trichloroethene reductive dehalogenase from Dehalococcoides ethenogenes: sequence of tceA and substrate range characterization.

The anaerobic bacterium Dehalococcoides ethenogenes is the only known organism that can completely dechlorinate tetrachloroethene or trichloroethene (TCE) to ethene via dehalorespiration. One of two corrinoid-containing enzymes responsible for this pathway, TCE reductive dehalogenase (TCE-RDase) catalyzes the dechlorination of TCE to ethene. TCE-RDase dehalogenated 1,2-dichloroethane and 1, 2-dibromoethane to ethene at rates of 7.5 and 30 micromol/min/mg, respectively, similar to the rates for TCE, cis-dichloroethene (DCE), and 1,1-DCE. A variety of other haloalkanes and haloalkenes containing three to five carbon atoms were dehalogenated at lower rates. The gene encoding TCE-RDase, tceA, was cloned and sequenced via an inverse PCR approach. Sequence comparisons of tceA to proteins in the public databases revealed weak sequence similarity confined to the C-terminal region, which contains the eight-iron ferredoxin cluster binding motif, (CXXCXXCXXXCP)(2). Direct N-terminal sequencing of the mature enzyme indicated that the first 42 amino acids constitute a signal sequence containing the twin-arginine motif, RRXFXK, associated with the Sec-independent membrane translocation system. This information coupled with membrane localization studies indicated that TCE-RDase is located on the exterior of the cytoplasmic membrane. Like the case for the two other RDases that have been cloned and sequenced, a small open reading frame, tceB, is proposed to be involved with membrane association of TCE-RDase and is predicted to be cotranscribed with tceA.

An algorithm for automated bacterial identification using matrix-assisted laser desorption/ionization mass spectrometry.

An algorithm for bacterial identification using matrix-assisted laser desorption/ionization (MALDI) mass spectrometry is being developed. This mass spectral fingerprint comparison algorithm is fully automated and statistically based, providing objective analysis of samples to be identified. Based on extraction of reference fingerprint ions from test spectra, this approach should lend itself well to real-world applications where samples are likely to be impure. This algorithm is illustrated using a blind study. In the study, MALDI-MS fingerprints for Bacillus atrophaeus ATCC 49337, Bacillus cereus ATCC 14579T, Escherichia coli ATCC 33694, Pantoea agglomerans ATCC 33243, and Pseudomonas putida F1 are collected and form a reference library. The identification of test samples containing one or more reference bacteria, potentially mixed with one species not in the library (Shewanella alga BrY), is performed by comparison to the reference library with a calculated degree of association. Out of 60 samples, no false positives are present, and the correct identification rate is 75%. Missed identifications are largely due to a weak B. cereus signal in the bacterial mixtures. Potential modifications to the algorithm are presented and result in a higher than 90% correct identification rate for the blind study data, suggesting that this approach has the potential for reliable and accurate automated data analysis of MALDI-MS.

Use of an internal control for matrix-assisted laser desorption/ionization time-of-flight mass spectrometry analysis of bacteria.

A method to aid in the analysis of bacterial samples of unknown concentration by matrix-assisted laser desorption/ionization (MALDI) mass spectrometry is demonstrated. It is shown that in MALDI analysis of bacteria, the intensities of resulting peaks in spectra are sensitive to the microbial concentration. At the high and low ends of the concentration range, no signal can be obtained, leaving very concentrated or very dilute samples indistinguishable. The addition of cytochrome c as an internal control allows the differentiation of these concentrated and dilute samples. The presence of the internal control causes only a 20% to 30% decrease in signal intensity when the bacterial concentration is optimum. However, the signal quality is improved when the internal control is added to some low concentrations of bacteria.

Reproducibility of matrix-assisted laser desorption/ionization time-of-flight mass spectrometry for replicate bacterial culture analysis.

Matrix-assisted laser desorption/ionization time-of-flight mass spectrometry (MALDI-TOFMS) was used to demonstrate the reproducibility of bacterial spectra collected on different days. The reproducibility of analysis by MALDI-MS of intact Escherichia coli and Bacillus atrophaeus is presented as a replicate culture study in which spectra were collected on ten different occasions over a three-month period and by two different operators. The analysis resulted in the detection of specific biomarkers in the m/z 2000-20 000 range. Some of the peaks in the Escherichia coli spectra are identified by comparison with other published work. All of the spectra obtained are reproducible over the course of the experiment, but operator variability does exist. The Escherichia coli spectra show operator variability while the Bacillus atrophaeus spectra do not. This work demonstrates the utility of MALDI in obtaining consistent spectra from bacteria over a period of time.

The opsX locus of Xanthomonas campestris affects host range and biosynthesis of lipopolysaccharide and extracellular polysaccharide.

Xanthomonas campestris pv. citrumelo strain 3048 is the causal agent of citrus bacterial leaf spot disease and has a wide host range that includes rutaceous and leguminous plants. A spontaneous prototrophic mutant of strain 3048 (strain M28) that had lost virulence on citrus but retained virulence on bean plants was recovered. Growth studies in planta showed that M28 cells died rapidly in citrus leaves but grew normally in bean leaves. In addition to the loss of citrus-specific virulence, M28 displayed the following mutant phenotypes in culture: decreased growth rate, reduction of the amount of exopolysaccharide (to ca. 25% of the amount in 3048), loss of capsules, and significant alterations of the two 3048 lipopolysaccharide (LPS) bands visualized by silver stain on polyacrylamide gels, consistent with a defect(s) in LPS assembly. A 38-kb DNA fragment from a 3048 total DNA library that complemented the mutant phenotypes of M28 was identified. The 38-kb fragment did not hybridize to two similarly sized fragments carrying different hrp (hypersensitive response and pathogenicity) genes cloned from 3048. Subcloning, DNA sequence analyses, and gene disruption experiments were used to identify a single gene, opsX (for outer-membrane polysaccharide), responsible for the mutant phenotypes of M28. At least one other gene downstream from opsX also affected the same phenotypes and may be part of a gene cluster. We report here the DNA sequence and transcriptional start site of opsX. A search of protein sequence data bases with the predicted 31.3-kDa OpsX sequence found strong similarity to Lsi-1 of Neisseria gonorrhoeae and RfaQ of Escherichia coli (both are involved in LPS core assembly). The host-specific virulence function of opsX appears to involve biosynthesis of the extracellular polysaccharide and a complete LPS. Both may be needed in normal amounts for protection from citrus, but not bean, defense compounds.

An Xanthomonas citri pathogenicity gene, pthA, pleiotropically encodes gratuitous avirulence on nonhosts.

The pathogenicity gene, pthA, of Xanthomonas citri is required to elicit symptoms of Asiatic citrus canker disease; introduction of pthA into Xanthomonas strains that are mildly pathogenic or opportunistic on citrus confers the ability to induce cankers on citrus (S. Swarup, R. De Feyter, R. H. Brlansky, and D. W. Gabriel, Phytopathology 81:802-809, 1991). The structure and the function of pthA in other xanthomonads and in X. citri were further investigated. When pthA was introduced into strains of X. phaseoli and X. campestris pv. malvacearum (neither pathogenic to citrus), the transconjugants remained nonpathogenic to citrus and elicited a hypersensitive response (HR) on their respective hosts, bean and cotton. In X. c. pv. malvacearum, pthA conferred cultivar-specific avirulence. Structurally, pthA is highly similar to avrBs3 and avrBsP from X. c. pv. vesicatoria and to avrB4, avrb6, avrb7, avrBIn, avrB101, and avrB102 from X. c. pv. malvacearum. Surprisingly, marker-exchanged pthA::Tn5-gusA mutant B21.2 of X. citri specifically lost the ability to induce the nonhost HR on bean, but retained the ability to induce the nonhost HR on cotton. The loss of the ability of B21.2 to elicit an HR on bean was restored by introduction of cloned pthA, indicating that the genetics of the nonhost HR may be the same as that found in homologous interactions involving specific avr genes. In contrast with expectations of homologous HR reactions, however, elimination of pthA function (resulting in loss of HR) did not result in water-soaking or even moderate levels of growth in planta of X. citri on bean; the nonhost HR, therefore, may not be responsible for the "resistance" of bean to X. citri and may not limit the host range of X. citri on bean. The pleiotropic avirulence function of pthA and the heterologous HR of bean to X. citri are both evidently gratuitous.

Extracellular Polysaccharide Is Not Responsible for Aluminum Tolerance of Rhizobium leguminosarum bv. Phaseoli CIAT899.

Strain UHM-5, a pSym Exo derivative of the aluminum-tolerant Rhizobium leguminosarum bv. phaseoli strain CIAT899, was equally tolerant of aluminum (Al) as the parental culture. Dialyzed culture supernatants of the wild-type cells grown in YEM broth (10 cells ml) contained 185 mug of glucose equivalents ml whereas UHM-5 culture supernatants yielded 2 mug of glucose ml. The Exo derivative and the parental strain gave essentially similar growth in medium containing from 0 to 300 muM Al, indicating that the pSym of CIAT899, and extracellular polysaccharide, were not involved in the aluminum tolerance of this strain. However, increasing the level of Al from 80 to 150 muM increased the lag phase, induced a slight killing of the inoculum, and depressed the final populations by about fivefold. Doubling the aluminum concentration from 150 to 300 muM presented a severe aluminum stress to CIAT899 and UHM-5: the inoculum level dropped 10-fold, indicating killing of the inoculum, and remained depressed for ca. 4 days before continuing to grow slowly; the final population was decreased 15-fold relative to that of cultures grown in medium containing 80 muM Al. The production by CIAT899 of other extracellular or intracellular aluminum tolerance factors was investigated in culture by using aluminum-sensitive rhizobia as stress indicators. These experiments, conducted at 80 muM Al, demonstrated that CIAT899 produced neither extracellular nor intracellular products that could alleviate toxicity for the Al-sensitive indicator rhizobia.

Release of Rhizobium spp. from Tropical Soils and Recovery for Immunofluorescence Enumeration.

Limitations associated with immunofluorescence enumeration of bacteria in soil derive largely from the efficiency with which cells can be separated from soil particles and collected on membrane filters for staining. Many tropical soils fix added bacteria tightly, resulting in low recoveries. Eight soils, representative of three of the major soil orders found in the tropics (oxisols, vertisols, and inceptisols), were tested for recovery of added Rhizobium strains. All except one Hawaiian andept (Typic Eutrandept) yielded recoveries ranging from <1 to 13%. Recovery from the andept was 100%. In soil-sand mixtures, addition of only a small amount of soil caused a dramatic decrease in recovery of added rhizobia. Increasing the soil content of the mixture from 0% (10 g of sand) to 50% (5 g of soil-5 g of sand) reduced recoveries from >90 to <1%. Varying the ionic strength and pH of the extracting solution did not cause marked increases in recovery. Protein solutions, ethylenediaminetetraacetate, and NaHCO(3), on the other hand, improved release of bacteria. We report a modification to the usual membrane filter immunofluorescence procedure which yielded consistently high and reproducible recovery (coefficient of variation, 30%) of rhizobia from several tropical soils. In the modified procedure, partially hydrolyzed gelatin, diluted in ammonium phosphate, was used to suspend the soil. This caused dispersion of the soil and release of the bacteria from soil flocs. The efficiency of recovery of Rhizobium spp. from several tropical and two temperate soils remained high as the content of these soils in soil-sand mixtures was increased from 0 to 100%. The modified membrane filter immunofluorescence procedure was used to follow the growth of a strain of chickpea (Cicer arietinum) Rhizobium in a sterilized oxisol. The results showed a close agreement with viable counts at different stages during the growth cycle. Diluent for the hydrolyzed gelatin also had a marked effect on recovery. The efficiency of release of Rhizobium spp. from an oxisol was in the following order for the diluents used: 0.1 M (NH(4))(2)HPO(4) > 0.1 M Na(2)HPO(4) = 0.1 M sodium-phosphate-buffered saline (pH 7.2) > 0.2 M NH(4)Cl > 0.2 KCl > NaCl = LiCl > water.