Abstract
The phage amplified biologically assay is a new method for rapid and low-cost phenotypic determination of the drug sensitivities of Mycobacterium tuberculosis isolates and the detection of viable organisms in patient specimens. Infection of slowly growing mycobacteria with phage (phage D29) was followed by chemical virucide destruction of extracellular phage. Infected mycobacteria were mixed in culture with rapidly growing sensor cells, which the phage can also infect; i.e., lytic amplification of phage occurs. The aims of the present study were to optimize the speed and sensitivity of the assay and reduce its cost for developing countries by using an M. tuberculosis-spiked sputum model with (i) identification of inhibitory components of sputum and optimization of decontamination methods; (ii) simplification of the washing and development steps; (iii) reduction of the use of high-cost components, e.g., oleate-albumin-dextrose-catalase (OADC) supplement; and (iv) optimization of virucide treatment. The following results were obtained. (i) An inhibitory factor in sputum which could be removed by treatment of the sample with sodium dodecyl sulfate or NaOH decontamination was identified. (ii) A microcentrifuge-based approach with thixotropic silica as a bedding and resuspension agent was developed as an alternative to conventional centrifugation medium exchange. The yield was increased 228-fold, with increased speed and reduced cost. (iii) At present, after extracellular inactivation of phage, the ferrous ammonium sulfate (FAS) virucide is sequestered by dilution with an expensive supplement, OADC. Sodium citrate with calcium chloride was found to be a cost-effective after treatment with the FAS protectant and offered greater protection than OADC. Kinetic-lysis experiments indicated that an infection time of 1 to 3 h prior to FAS addition was optimal. (iv) Amplification of the signal (which corresponded to the burst size) was shown by allowing lysis prior to plating in a spiked medium model (up to 20-fold) and a spiked sputum model (up to 10-fold). A liquid culture detection method capable of detecting approximately 60 viable M. tuberculosis organisms in 1 ml of sputum was developed. Taken together, these improvements support the routine application of the assay to sputum specimens.
Tuberculosis (TB) remains the single greatest cause of mortality due to an infectious agent worldwide. It was estimated in 1991 that one-third of the population is infected with Mycobacterium tuberculosis, the causative organism, leading to 2 million to 3 million deaths annually (7, 20) The increasing prevalence of M. tuberculosis isolates resistant to first-line antibiotics makes the rapid detection and the rapid determination of drug susceptibility ever more urgent (13). The primary aim of prompt diagnosis and appropriate treatment of pulmonary disease is to render an individual noninfectious and, thus, break the chain of transmission (5).
Conventional culture-based techniques for M. tuberculosis detection take weeks from the time of receipt of a clinical specimen (3). Liquid culture analysis with the semiautomated BACTEC 460 instrument and automated liquid culture systems such as the MBBacT system (Organon Teknika, Cambridge, United Kingdom) have significantly reduced turnaround times but require expensive monitoring equipment and costly reagents (18). Additionally, the BACTEC method necessitates handling of radioactive material.
The phage amplified biologically assay (the PhaB assay) was first described as a rapid (turnaround time, 2 to 4 days) and sensitive phenotypic method for the drug susceptibility testing of M. tuberculosis isolates (8, 9, 21). The relative simplicity of the assay makes it ideal for use in both developed and developing countries. The method is based on the ability of infected mycobacteria, e.g., M. tuberculosis, to protect internalized mycobacteriophages from chemical inactivation and to support their replication. After sequestration of the chemical virucidal agent, progeny mycobacteriophages released by lysis from mycobacteria can be detected rapidly via infection and subsequent lysis of fast-growing Mycobacterium smegmatis sensor cells.
The method has also been described in connection with detection of M. tuberculosis isolates directly from primary patient specimens, but without refinement (Fig. 1) (12). In this respect, the test offers a complement to smear and standard culture approaches. This study set out to examine the application of the PhaB assay to the detection of viable mycobacteria from sputum with a M. tuberculosis-spiked sputum model. Specifically, the role of sputum inhibitory factors and removal of sputum inhibitory factors were addressed, along with the critical process of buffer or medium exchange after decontamination treatment of sputum. Additionally, end-point detection approaches were studied, including the sequestration of virucide.
FIG. 1.
Schematic flow representation of sputum processing and PhaB assay execution by the methods published by McNerney et al. (12) (a) and suggested on the basis of the findings of this study (b). D29, infection of mycobacteria with mycobacteriophage D29; FAS, inactivation of extracellular D29 with the virucide FAS following infection of mycobacteria; OCG, OADC, calcium chloride, glycerol 7H9 medium supplement; CC, 7H9 medium supplemented with sodium citrate and calcium chloride. Steps 1 to 5 (a) and steps 1 to 4 (b) relate to washes after sputum decontamination and neutralization with medium involving a number of centrifugation steps, which result in the resuspension of mycobacteria in 7H9-OCG prior to overnight incubation at 37°C (see Materials and Methods). Steps 6 to 10 in the scheme in panel a refer to the collection of a 0.2-ml aliquot for D29 infection (step 6), 3.5 h of incubation at 37°C prior to addition of FAS (step 7), 10 min of incubation prior to FAS inactivation by the addition of 1 ml of 7H9-OGC (step 8), 4 h of incubation at 37°C to enable lytic burst (step 9), and plating of a 0.1-ml aliquot with 7H9-OCG-agar and M. smegmatis sensor cells (step 10). Steps 5 to 7 in the scheme in panel b refer to D29 infection (step 5), 2 h of incubation at 37°C prior to addition of FAS (step 6), and 10 min of incubation prior to FAS inactivation by plating with 7H9 medium supplemented with sodium citrate, calcium chloride, agar, and M. smegmatis sensor cells (step 7). Plaques result on the sensor-cell lawn when mycobacteria harboring D29 lyse following plating.
MATERIALS AND METHODS
Sputum spiked with M. tuberculosis.
M. tuberculosis-spiked smear-negative sputum was used as a model to investigate various PhaB assay parameters and the effects of different processing methodologies on assay performance. One microliter of freshly grown M. tuberculosis (MT14323) was scraped from a Löwenstein-Jensen (LJ) medium slope with a 1-μl loop (approximately 106 organisms), mixed with 2 ml of Middlebrook 7H9 medium (7H9) containing glass beads, and vortexed vigorously for 30 s. Excess smear-negative sputum from a clinical microbiology laboratory was spiked with different volumes of a preparation suspended in medium containing glass beads according to experimental requirements. Standard PhaB assay processing was as described previously (8, 9, 21).
Heat treatment of sputum.
An experiment investigating the effects of heat treatment of sputum on PhaB assay performance was designed with a view to removing or characterizing the activity inherent in sputum inhibitory to the PhaB assay. One milliliter of mixed, smear-negative sputum or 7H9-10% (vol/vol) OCG (Middlebrook oleate-albumin-dextrose-catalse [OADC], 10 mM calcium chloride, 2% [(vol/vol] glycerol; donated by Biotec Laboratories, Ipswich, United Kingdom) was spiked with M. tuberculosis either before or immediately after heat treatment and subsequent cooling. The samples were immersed in a water bath at room temperature or 50, 55, 60, 65, or 80°C for 10 min prior to incubation at room temperature for 20 min. The samples were then centrifuged at 3,000 × g for 20 min, the supernatants were discarded, and the pellets were resuspended in 10 ml of 7H9. The samples were then centrifuged at 3,000 × g for an additional 20 min, the supernatants were discarded, and the pellets were resuspended in 1 ml of 7H9-10% (vol/vol) OCG for overnight recovery at 37°C prior to application of the PhaB assay (see below).
Universal method of sputum processing by tube centrifugation.
A universal method of sputum processing by tube centrifugation was used to represent a standard sputum processing regimen in a pathology laboratory for comparison with a novel microcentrifuge-based system (see below) with respect to the influence of the processing method on PhaB assay performance. All chemicals were supplied by Sigma Aldrich (Poole, United Kingdom) unless stated otherwise. An equal volume of 2% (wt/vol) NaOH-1.45% (wt/vol) trisodium citrate · 2H2O was added to spiked sputum, the components were mixed well, and the mixture was incubated at room temperature for 20 min. Subsequently, 20 ml of 67 mM phosphate buffer-0.015% (wt/vol) phenol red (pH 6.8) was added, and the components were mixed prior to centrifugation at 3,000 × g for 20 min in an MSE bench-top centrifuge. The supernatant was discarded, and 20 ml of 7H9 was added prior to resuspension. After further centrifugation at 3,000 × g for 20 min, the supernatant was discarded prior to resuspension in 1 ml of 7H9 supplemented with 10% (vol/vol) OCG and antibiotics from the MBBacT process bottle (Organon Teknika). The samples were then incubated overnight at 37°C prior to application of the PhaB assay.
Sputum processing by microcentrifugation of tubes with TS.
Processing of sputum by microcentrifugation of tubes with fine, fumed thixotropic silica (TS) was introduced in an attempt to limit the loss of buoyant M. tuberculosis during centrifugation steps and also to improve the speed and handling aspects of the assay compared with those of the universal method outlined above. An equal volume of 2% (wt/vol) NaOH-1.45% (wt/vol) trisodium citrate · 2H2O was added to spiked sputum in the original container, and the components were mixed well. After incubation at room temperature for 20 min, 1 ml of the agitated mixture was transferred to a 2-ml microcentrifuge tube containing 0.25 ml of well-agitated TS (25 mg) prior to centrifugation at 12,000 × g for 30 s in an aerosol-protected IEC Micromax microcentrifuge. The supernatant was discarded. The pellet was resuspended in 1.5 ml of 67 mM phosphate buffer-0.015% (wt/vol) phenol red (pH 6.8) by flicking the base of an inverted tube (to dislodge the matrix), followed by vigorous shaking. A second step of centrifugation at 12,000 × g for 30 s was performed, followed by resuspension of the matrix in 1.5 ml of 7H9. A final step of centrifugation at 12,000 × g for 30 s was performed prior to resuspension of the matrix in 1 ml of 7H9-10% (vol/vol) OCG-MBBacT antibiotics. The samples were then incubated overnight at 37°C prior to application of the PhaB assay.
PhaB assay.
Following sputum processing and overnight recovery of M. tuberculosis, the PhaB assay was routinely performed as follows. When the method of sputum processing by microcentrifugation of 2-ml tubes with TS was used, a 5-s pulse in the microcentrifuge was included prior to the reagent addition steps; 100 μl of 109 PFU of mycobacteriophage D29 (donated in lyophilized form by Biotec Laboratories and freshly suspended in 7H9 before use) per ml was mixed with a sample prior to incubation at 37°C for 2 h. The chemical virucide (200 μl of 100 mM ferrous ammonium sulfate [FAS; donated by Biotec Laboratories], which was freshly dissolved in distilled H2O [dH2O]before use) (12) was added; and, importantly, prior to incubation at room temperature for 10 min, the contents of the vessel were repeatedly mixed by turning and inversion to ensure that all surfaces of the vessel came into contact with the FAS. The samples were then mixed in triple-vented petri dishes with 1 ml of a stationary-phase M. smegmatis ATCC 607 suspension (donated in lyophilized form by Biotec Laboratories and freshly suspended in 7H9 before use), 1 ml of OCG (or trisodium citrate and calcium chloride to yield final concentrations in the plate of 8 and 10 mM, respectively), and 9 ml of molten 7H9-1.5% (wt/vol) agar (52°C). Set plates were incubated and inverted overnight at 37°C, and the numbers of plaques were recorded on the following morning.
Screen for progeny phage D29 protectants after FAS treatment.
Following infection of M. tuberculosis with phage D29 and virucide treatment to inactivate external phage, the virucide itself needs to be inactivated prior to mixing of M. tuberculosis with the M. smegmatis sensor cells. The published methodology has thus far relied on an ill-defined and expensive approach in the form of dilution in OADC-rich medium. In this study, a series of alternative prospective virucide inactivants was screened. Two hundred D29 PFU was mixed with 1 ml of a stationary-phase M. smegmatis suspension, FAS to a final concentration in the plate of 0.5 or 2 mM, various test progeny D29 protectants (thiourea, N-acetyl cysteine, potassium iodide, sodium benzoate [pH 7], EDTA [pH 7], diethylenetriaminepentaacetic acid [DTPA; pH 7], dipyridyl, trisodium citrate [pH 7], and calcium chloride) and 9 ml of 7H9-1 mM calcium chloride-1.5% (wt/vol) agar. The agar was held at 52°C and swirled to suspend the calcium precipitates immediately prior to use. Separate aliquots of all components were pipetted onto plates and mixed only upon pouring of the molten agar. The plates were incubated by inversion overnight at 37°C, and the numbers of plaques were recorded on the following morning.
Overlay experiment with a fraction washed with SDS.
The overlay experiment with a fraction washed with SDS was designed to demonstrate the ability to fractionate and detect any activity in sputum inhibitory to the assay upon washing of the samples with sodium dodecyl sulfate (SDS) solution. Fractions were represented by soaked filter paper squares overlaid on a plate with a mixture of M. smegmatis and phage D29 containing the antibiotic cocktail from the MBBacT system. M. smegmatis growth zones (transluscent zones) against a background of confluent lysis correspond to inhibition of D29 infection. Smear-negative sputum (0.5 ml) was mixed vigorously with 0.5 ml of dH2O, and a 30-μl sample was taken. After incubation for 20 min at room temperature, 1 ml of the suspension was centrifuged at 12,000 × g for 1 min and 30 μl of supernatant was taken from near the top of the specimen. The remaining supernatant was then discarded, and the pellet was resuspended in 1 ml of 0.05% (wt/vol) SDS-1% (wt/vol) NaCl. After 20 min of incubation at room temperature the 1-ml suspension was centrifuged at 12,000 × g for 1 min, and 30 μl of supernatant was taken from near the top of the specimen. Then, 30 μl of dH2O, 30 μl of 0.05% (wt/vol) SDS-1% (wt/vol) NaCl, 30 μl of the sputum suspension, 30 μl of supernatant washed in dH2O, or 30 μl of supernatant washed in 0.05% (wt/vol) SDS-1% (wt/vol) NaCl was added to filter paper squares (1.5 by 1.5 cm). The squares were then overlaid onto freshly set plates with 1 ml of a stationary-phase M. smegmatis suspension, 1 ml of OCG, 100 μl of 7H9 containing 2,000 PFU of phage D29, and 9 ml of 7H9-1.5% (wt/vol) agar (52°C) with antibiotics from the MBBacT system. The plates were incubated by inversion overnight at 37°C, and the lysis profile was recorded on the following morning.
Lysis-time course experiments.
The kinetics of infection and lysis and the timing of virucide addition and plating are important considerations for the PhaB assay. A sufficient time of infection prior to FAS addition is required to achieve good results, but the time of infection should not be so long that released progeny phages are destroyed. Furthermore, assay signal amplification relating to lytic burst could conceivably be obtained following FAS treatment and FAS inactivation by allowing additional incubation prior to plating of aliquots sampled following agitation. As such, the samples were processed as described above for the PhaB assay, except that the times of FAS addition prior to plating were varied, and in a separate experiment, after FAS addition, 5 ml of 7H9-10% (vol/vol) OCG or 5 ml of 7H9-8 mM trisodium citrate-10 mM calcium chloride (which was agitated to suspend the precipitates) was added and the samples were incubated at 37°C for various times prior to plating of aliquots sampled following agitation.
Phage detection in liquid culture with MTT.
Thus far, the method described for the final step in the PhaB assay protocol, detection of progeny phage from lysed M. tuberculosis cells, has been plate based, in which zones of lysis are observed on a background of a sensor-cell lawn. An experiment for detection of phage in liquid culture after 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyl-2H-tetrazolium bromide (MTT) treatment set out to determine the possibility of using a simple liquid culture detection method with a view to reducing technical demand and increasing the processivity. One milliliter of mixed sputum specimens spiked with different concentrations of M. tuberculosis suspended in solution with beads (based on the approximation that a 1-μl loop of fresh growth from an LJ medium slope harbors 106 M. tuberculosis organisms) was processed by addition of 1 ml of 2% (wt/vol) NaOH-1.45% (wt/vol) trisodium citrate · 2H2O, followed by application of the universal method of sputum processing. After the addition of FAS, 5 ml of 7H9-10% (vol/vol) OCG was added along with 0.5 ml of a stationary-phase M. smegmatis suspension. This mixture was incubated at 37°C overnight without shaking. On the following morning, 5 μl of 10 mg of MTT per ml was added to 1 ml of agitated suspension. After further incubation at 37°C for 10 min, the color was recorded. MTT undergoes a color change from yellow to blue in the presence of the succinate-tetrazolium reductase of metabolically active cells. If lysis of the M. smegmatis culture occurs, insufficient metabolically competent cells remain to convert MTT, such that the yellow color remains to indicate a positive result. Conversely, a blue color indicates a negative result.
RESULTS
Removal of sputum-associated activity inhibitory to the PhaB assay.
Unprocessed sputum or sputum subjected to multiple washing steps with various agents (dH2O, 7H9, 0.2 M guanidinium chloride, 0.2 M urea, 0.25% [wt/vol] N-acetyl cysteine, and combinations thereof) prior to washing with 7H9 and final resuspension in 7H9-10% (vol/vol) OCG-MBBacT system antibiotics was found to have activity inhibitory to the PhaB assay. Heat treatment removed this activity, but under conditions which killed the organisms under detection (Fig. 2a). However, treatment of sputum with the decontamination agent NaOH removed the inhibitor, while it allowed the viability of M. tuberculosis to be retained. A final NaOH concentration of 1% (wt/vol) was established as optimal with respect to balancing of inhibitor removal, M. tuberculosis viability, and contaminant suppression (Table 1).
FIG. 2.
(a) Ability of heat treatment to remove activity inherent in sputum inhibitory to the PhaB assay and M. tuberculosis survival after heat treatment. One milliliter of mixed, smear-negative sputum or 7H9-10% (vol/vol) OCG was spiked with M. tuberculosis either before (open bars) or after (solid bars) heat treatment (slashed bars, medium spiked with M. tuberculosis before heat treatment). After immersion in a water bath at different temperatures for 10 min and cooling to room temperature, an equal volume of 0.25% (vol/vol) NaOH-0.5% (vol/vol) N-acetyl cysteine was added to the samples. These were incubated at room temperature for 20 min, prior to centrifugation at 3,000 × g for 20 min. The supernatants were discarded, and after washing with 10 ml of 7H9 and a centrifugation step, the samples were resuspended in 7H9-10% (vol/vol) OCG for overnight recovery, prior to application of the PhaB assay (see Materials and Methods). (b) A plate after incubation of a mixture of M. smegmatis and phage D29. The plate was overlaid with filter paper squares bearing sputum (filter paper a), a sputum fraction washed with dH2O (filter paper b), a sputum fraction washed with 0.05% (wt/vol) SDS-1% (wt/vol) NaCl (filter paper c), dH2O (filter paper d), or 0.05% (wt/vol) SDS-1% (wt/vol) NaCl (filter paper e). This experiment was designed to demonstrate the ability to fractionate the activity of sputum inhibitory to the assay upon washing with SDS solution. The soaked filter paper squares were overlaid on a plate with a mixture of M. smegmatis and D29 containing the MBBacT antibiotic cocktail. M. smegmatis growth zones (translucent zone, e.g., within filter paper c) against a background of confluent lysis correspond to inhibition of D29 infection (see Materials and Methods).
TABLE 1.
Plaque yields from PhaB assay after processinga
| Sputum | Presence of M. tuberculosis | No. of plaques in the presence of final sodium hydroxide concn of
|
||||
|---|---|---|---|---|---|---|
| 0% | 0.25% | 0.5% | 1% | 2% | ||
| Mixed | − | 0b | 0 | 0 | 0 | 0 |
| 1 | + | 0b | 79 | 209 | 383 | 342 |
| 2 | + | 0b | 203b | 400b | 408 | 317 |
| 3 | + | 0b | 56 | 47 | 201 | 457 |
| 4 | + | 0b | 342 | 413 | 435 | 420 |
| 5 | + | 0b | 543b | 224 | 344b | 348 |
| 6 | + | 0b | 103 | 163 | 359 | 112 |
| 7 | + | 0b | 547 | 604 | 588 | 565 |
| 8 | + | 0b | 412b | 487 | 406 | 359 |
| 9 | + | 0b | 0b | 3 | 84 | 28 |
| 10 | + | 0b | 35 | 423 | 449 | 304 |
Spiked sputum specimens (0.5 ml; approximately 10,000 M. tuberculosis organisms) with different concentrations of NaOH, followed by application of the microcentrifuge method with TS (see Materials and Methods).
Plates were processed on which growth of contaminants was recorded.
During the search for alternative processing agents, a sputum fraction washed with a low concentration of SDS (0.05% [vol/vol], which was below the level required for visible mucin matrix breakdown or contaminant suppression) was shown to have activity inhibitory to the assay (Fig. 2b). Moreover, addition of an equal volume of 1% (vol/vol) SDS to sputum efficiently removed the inhibitory activity, with the effect being similar to that of the optimal NaOH treatment (data not shown). The methods with both NaOH and SDS required multiple follow-up processing and centrifugation steps to enable efficient phage D29 infection of M. tuberculosis. Extensive dilution of SDS to levels below 0.005% (wt/vol) was required, whereas neutralization of the NaOH with phosphate buffer followed by medium exchange was required. Following phosphate buffer neutralization and dilution, washing steps with 20 ml of 7H9 or 20 ml of 7H9-10% (vol/vol) OCG prior to resuspension in 7H9-10% (vol/vol) OCG with MBBacT system antibiotics enabled comparable yields.
The possibility of removing centrifugation steps from the protocol was considered (see below). However, attempts to titrate NaOH by the drop-wise addition of HCl in the presence of dilute phenol red pH indicator (0.0015% [wt/vol]) resulted in the heavy precipitation of materials which persisted throughout the assay, resulting in its inhibition. Multiple subsequent washing steps with medium did not alleviate this problem. Conversely, direct dilution and neutralization of NaOH with 7H9 or 7H9-10% (vol/vol) OCG caused medium components to precipitate, which resulted in false-positive breakthrough plaques.
Microcentrifuge tube system with TS as a gel trap and resuspension agent.
Conventional centrifugation methodologies for the sedimentation of M. tuberculosis from sputum specimens use large, bench-top centrifuges which are compatible with vessels such as 30-ml universal containers, and it is recommended that they be operated at between 2,000 and 3,000 × g for 20 min (3, 14). This represents a balance between the sedimentation rate of the relatively buoyant organism, practical time constraints, and the problem of overheating in nonrefrigerated centrifuges. A substantial loss of viable M. tuberculosis cells can result. For reasons of time, yield, materials, equipment cost, and technical demand, multiple conventional centrifugation steps prior to application of the PhaB assay appeared to be unattractive. A method which introduced TS as a bedding and resuspension agent to 2-ml microcentrifuge tubes was devised. The application of a high relative centrifugal force (12,000 × g) enabled consistently high yields with rapid (30-s) spin steps. TS was demonstrated to have a beneficial effect on the microcentrifuge system (Fig. 3a), producing higher yields and more homogeneous resuspension. Improvements in yields of 1.32- to 228-fold compared to those obtained by conventional centrifugation with an array of spiked 0.5-ml sputum specimens resulted (Fig. 3b).
FIG. 3.
(a) Effect of TS addition to 2-ml microcentrifuge tubes during processing of 0.5 ml of M. tuberculosis (TB)-spiked sputum prior to application of the PhaB assay. An equal volume of 2% (wt/vol) NaOH-1.45% (wt/vol) trisodium citrate · 2H2O was used for processing, followed by application of the microcentrifugation method with TS. The lines above the bars are standard errors. (b) Bar chart comparing the yields from the PhaB assay after processing of 0.5 ml of sputum specimens spiked with M. tuberculosis by the universal method with 30 ml of supernatant (slashed bars) compared with the yields after processing by the method with a 2-ml microcentrifuge tube with TS (solid bars) prior to application of the PhaB assay (open bars, M. tuberculosis spiked with 7H9-10% [vol/vol] OCG). Specimens were processed with an equal volume of 2% (wt/vol) NaOH-1.45% (wt/vol) trisodium citrate · 2H2O, as outlined in Materials and Methods. +, positive control (7H9-10% [vol/vol] OCG spiked with M. tuberculosis with no prior processing); −, negative control (unspiked mixed sputum).
Overnight recovery in supplemented 7H9.
Since application of the PhaB assay to primary sputum specimens requires exposure to relatively harsh chemical agents, the effect of medium supplementation on M. tuberculosis recovery was investigated. M. tuberculosis was shown to require an overnight recovery incubation following treatment with NaOH or SDS to enable efficient phage D29 infection. Overnight recovery in 1 ml of 7H9 supplemented with 20 or 50% OCG (vol/vol) enabled yields from the PhaB assay higher than those obtained by OCG supplementation following NaOH processing of spiked 7H9 (Fig. 4a). The increased calcium chloride concentration could not account for this, since supplementation of the infection medium with calcium chloride at concentrations in the range of 1 to 10 mM did not apparently affect assay performance. The observation was not replicated with an M. tuberculosis-spiked sputum model, in which supplementation with 10% OCG yielded results similar to those achieved by supplementation with 20 and 50% (vol/vol) OCG. Of note, neither model yielded false-positive breakthrough plaques for the negative controls, even when 50% (vol/vol) OCG was applied (see below). Supplementation with a less well defined medium in the form of egg white and/or egg yolk proved to be inhibitory to the assay. Interestingly, TS was demonstrated to improve yields, in the absence of centrifugation steps, when M. tuberculosis suspended in fresh medium from an old LJ medium slope (age, 6 months) was applied to the PhaB assay without prior recovery. This was not evident when overnight recovery incubation was allowed (Fig. 4b).
FIG. 4.
(a) Effects of various OCG contents in the 7H9 used for overnight recovery following mock NaOH treatment (treatment with an equal volume of 2% [wt/vol] NaOH-1.45% [wt/vol] trisodium citrate · 2H2O) of 0.5 ml of unspiked (−) or M. tuberculosis-spiked (+) 7H9 on the PhaB assay. The samples were then processed by the microcentrifugation protocol with TS prior to overnight incubation at 37°C in 7H9 supplemented with different OCG concentrations, as indicated. The asterisk indicates the control that received no NaOH treatment. (b) Effect of TS on performance of the PhaB assay with old cultures. A 6-month-old suspension from an LJ medium slope (in 7H9-10% [vol/vol] OCG) was applied directly to the assay or was applied following overnight recovery (o/n rec.) incubation at 37°C in the presence or absence of TS, as indicated. −, unspiked medium; +, M. tuberculosis-spiked medium; the lines above the bars indicate standard errors.
Protection of progeny phage D29 after trisodium citrate and calcium chloride virucide treatment.
Following treatment with the FAS virucide, the methods described to date have used dilution with OADC-supplemented 7H9 to protect progeny phages upon plating with M. smegmatis sensor cells (12). This study confirmed that sixfold dilution of 20 mM FAS in 7H9-10% (vol/vol) OADC-1 mM calcium chloride effectively protected progeny D29 at the end-point detection step. However, OADC serves as an ill-defined and very expensive protectant after virucide treatment. Its presence as a supplement is not required for efficient growth of M. smegmatis or infection of M. smegmatis by phage D29. This led to the search for a more specific and cost-effective protectant after FAS treatment. Initially, a series of oxidative protectants (thiourea, N-acetyl cysteine, potassium iodide, and sodium benzoate) were examined for their abilities to protect D29 from the virucidal activity of FAS upon direct plating with M. smegmatis (see Materials and Methods), but it was demonstrated that they were not effective. As an alternative approach, chelating agents (DTPA, EDTA, and trisodium citrate [the pH of each of which was adjusted to 7]) supplemented with calcium chloride were subsequently investigated by using the same system. Although DTPA proved inhibitory to M. smegmatis viability, EDTA and trisodium citrate were found to offer progeny D29 protection from FAS. Moreover, trisodium citrate was demonstrated to be fully compatible in the PhaB assay and offered improved protection over that offered by OADC in the context of a spiked sputum model (see Materials and Methods) at a fraction of the cost (Fig. 5b).
FIG. 5.
(a) Yields from the PhaB assay by using various infection times prior to addition of virucide and plating. Aliquots from an agitated culture of 1 ml of M. tuberculosis-spiked 7H9-10% (vol/vol) OCG at 37°C were taken at different time points after addition of phage D29. These were treated with FAS for 10 min and then plated (see Materials and Methods). (b) Time courses of lytic yield from the PhaB assay after phage D29 infection under different conditions (see Materials and Methods). At the indicated times, 0.5 ml of sputum was processed by the addition of 0.5 ml of 2% (wt/vol) NaOH-1.45% (wt/vol) trisodium citrate · 2H2O, followed by the microcentrifugation method with TS. Symbols: +, unspiked sputum with TS and OCG; ○, unspiked sputum with TS and CC; ▾, medium with OCG spiked with M. tuberculosis; ×, medium spiked with M. tuberculosis with CC; •, medium with TS and OCG spiked with M. tuberculosis; *, sputum with TS and OCG spiked with M. tuberculosis; and ▵, sputum with TS and CC spiked with M. tuberculosis, where medium refers to 7H9-10% (vol/vol) OCG, CC is 5 ml of 7H9-8 mM trisodium citrate-10 mM calcium chloride, and OCG is 5 ml of 7H9-10% (vol/vol) OCG added after FAS addition. Overnight recovery was allowed for each sample prior to application of the PhaB assay. Phage D29 was added at time zero. FAS was added after 2 h of infection. Aliquots were plated at different time points after further incubation at 37°C.
Lysis kinetics of phage D29-infected M. tuberculosis.
Previously published methods for the PhaB assay have described the plating of samples with M. smegmatis 7.5 h after initial phage D29 infection (12, 21) in an attempt to incorporate signal amplification through D29-associated lytic burst. Theoretically, the increase in signal should relate to the burst size per D29-infected M. tuberculosis organism. In this study, the kinetics of infection and lysis were recorded from a series of time course experiments with medium spiked with actively growing M. tuberculosis and 0.5 ml of sputum spiked with M. tuberculosis and processed with NaOH prior to overnight recovery. The results from the experiments with spiked medium demonstrated that the optimal infection time prior to addition of FAS was between 1 and 3 h (Fig. 5a). Plating at 7.5 h after initial introduction of D29 resulted in a fivefold increase in the signal relative to that achieved by plating prelysis (the prelysis point was determined by plating immediately following FAS treatment and subsequent sequestration) (Fig. 5b). The signal could be increased 20-fold if this was extended to 24 h after the initial introduction of D29. However, the results of the experiments with sputum processed with NaOH were less favorable. Approximately 2-fold amplification of a signal could be expected when plating was done 7.5 h after the initial infection, and a maximum of 10-fold amplification could be attained if this was extended to 24 or 48 h. Residues from processed sputum apparently reduced the half-life of progeny D29 in the extracellular medium. Trisodium citrate supplemented with calcium chloride appeared to offer protection from this phenomenon.
Liquid culture end-point detection.
It was reasoned that a liquid culture end-point detection method for the PhaB assay might offer a simple and highly sensitive alternative to the plate-based approach. As such, reagents from the Promega CytoTox96 tetrazolium dye-based lactate dehydrogenase detection system (which were used to monitor cell lysis directly) and the cell viability indicator MTT (which was used to monitor metabolic activity) were included with overnight M. smegmatis liquid cultures. This was done with a view to creating a continuous monitoring arrangement. Unfortunately, these approaches yielded such high background levels that no results could reliably be interpreted. However, when MTT was added to the samples following overnight liquid culture in the presence of M. smegmatis, the method proved capable of detecting approximately 60 M. tuberculosis organisms in 1 ml of sputum (Table 2).
TABLE 2.
Sensitivity of the liquid culture end-point detection method with MTTa
| Result for sputum | Presence of M. tuberculosis | Approx. no. of M. tuberculosis isolates | Result a | Result b |
|---|---|---|---|---|
| − | + | 3,000 | Yellow | Yellow |
| + | − | 0 | Blue | Blue |
| + | + | 3 | Blue | Blue |
| + | + | 30 | Blue | Blue |
| + | + | 60 | Yellow | Yellow |
| + | + | 300 | Yellow | Yellow |
| + | + | 3,000 | Yellow | Yellow |
See Materials and Methods. Results a and b refer to experimental duplicates.
DISCUSSION
McNerney et al.
(12) have already demonstrated that NaOH-processed smear-positive sputum specimens yield positive results by the PhaB assay. They did not describe, however, an inhibitory activity inherent to sputum. The assay can be successfully applied to cerebrospinal fluid specimens without the requirement for decontamination (6). In this study, an inhibitory activity inherent to sputum was identified which led to the requirement for suitable processing methods prior to application of the assay to sputum specimens. Heat treatment or processing with NaOH or SDS was shown to inactivate or wash away the inhibitory activity, suggesting that the activity is of a proteinaceous nature. However, the heat treatment required was beyond the temperature permissive for M. tuberculosis survival, nullifying its value as a sample processing method. Optimal sample processing results were achieved with a 1% (wt/vol) final concentration of NaOH. This successfully balanced the requirement for M. tuberculosis survival against the need to suppress contaminating organisms and remove the inhibitor of the assay inherent in sputum. Inhibitory activity could be detected in the supernatant of a fraction washed with a low concentration SDS but not the supernatant of a fraction previously washed with dH2O, suggesting that the causative agent(s) may form noncovalent associations with the mucin matrix in the native state. SDS should be noted as a potential agent for processing regimens in the future.
Following processing of the specimens with NaOH or SDS, neutralization and/or washing steps were required prior to resuspension in overnight recovery and infection medium. SDS had to be diluted to below 0.005% (wt/vol) to prevent inhibition of the assay. This was consistent with previous observations that the detergent Tween 80 is inhibitory to mycobacteriophage infection (15). Putative mycobacteriophage receptors may be disrupted or washed from the surface of M. tuberculosis in the presence of detergent and/or detergent may directly impair the integrity of phage D29. After processing of the specimens with NaOH, a phosphate buffer neutralization step and subsequent exchange of 7H9 were required. When 7H9 or 7H9-10% (vol/vol) OCG was added directly to NaOH, precipitates formed, which resulted in false-positive breakthrough plaques. This was possibly due to residues sequestering ferrous ions at the virucide (FAS) treatment step.
The method of buffer or medium exchange is of profound importance to the practicality and sensitivity of the PhaB assay; hence, a focus of the study was to address this. The rationale behind the use of the 2-ml microcentrifuge tube with TS was that the combination of a trap matrix and brief, high-speed centrifugation steps would result in high yields rapidly and consistently. Furthermore, upon resuspension, the matrix would be dispersed with greater homogeneity. Indeed, the 2-ml microcentrifuge tube system with TS performed consistently better than the 30-ml universal container-based system in a model consisting of 0.5 ml of sputum spiked with M. tuberculosis. Improvements in yields ranging from 1.32- to 228-fold across a range of sputum specimens demonstrated the benefits of this approach.
Higher yields were observed following mock NaOH processing of M. tuberculosis-spiked 7H9 when 20 or 50% (vol/vol) OCG instead of OCG at concentrations 10% (vol/vol) or lower were used to supplement 7H9 during overnight recovery, without false-positive results. This may reflect a shortened metabolic lag phase or the enhanced regeneration of putative phage D29 receptors following NaOH stress. That the profile was not replicated with a sputum model suggests that sputum offers M. tuberculosis protection from NaOH. This may be due to buffering activity and/or physical shielding. Of interest to the question of putative receptor regeneration was the observation that old-growth M. tuberculosis (6-month-old growth from LJ medium slopes) suspended in 7H9-10% (vol/vol) OCG prior to immediate application of the PhaB assay produced 10-fold higher plaque counts in the presence of TS than in its absence. This observation was not reproduced when such a suspension was allowed overnight recovery at 37°C prior to application of the assay. It is possible that TS abridges damaged putative receptor sites or mediates D29 docking in the absence of a specific receptor. This could account for some of the improvement observed by the microcentrifugation method with 20-ml tubes and TS.
The mechanism of activity of the virucide FAS has not been elucidated. Unpublished preliminary data suggesting that oxidative damage does not constitute the mode of action have been reported previously (12). This was corroborated by the findings of this study. This study also discovered that the chelating agent trisodium citrate supplemented with calcium chloride is a specific, efficient, and cost-effective alternative to OADC as an agent that protects D29 from FAS. Chelation by citrate may sequester ferrous ions to prevent destabilization of D29. Ferrous citrate chelates have been demonstrated to be protected from oxidative activity (11). The protection may be due to a combination of effects.
The inclusion of a lysis step prior to plating (7.5 h after initial infection) with the intention of amplifying the signal by a factor equivalent to the burst size has been described (12). Completion of the D29 lytic cycle in M. tuberculosis after 13 h has been described, however (4). In our study, a model with medium spiked with actively growing M. tuberculosis tested by the previously described method (12) produced a fivefold signal amplification. Twentyfold greater amplification was possible by incorporation of an overnight incubation step prior to plating (plating was done 24 h after initial infection). Under conditions in which sputum was processed after the samples were spiked with M. tuberculosis, however, this performance tailed off dramatically. It appeared that sputum residues could reduce the half-life of progeny D29 released into the extracellular environment. Metabolic lag may also have contributed to this. Trisodium citrate supplemented with calcium chloride appeared to offer a degree of protection, suggesting a role for salts or a metal ion-dependent enzyme in decreasing the half-life of progeny D29. These data also indicate that there may be a practical threshold in terms of the volume of sputum that can be processed and effectively applied to the PhaB assay. In the context of sputum, an extended period of incubation to enable lysis of D29-infected M. tuberculosis prior to plating is probably not justified on the basis of these findings.
The apparent burst size decreased in the presence of TS, possibly due to the sequestration of progeny phage D29 and/or as a result of premature induction of lysis. However, the dual ability of TS to behave as a D29 receptor bridge and as a D29 sequestration agent would also be consistent. TS did not appear to decrease the half-life of progeny D29 released into the extracellular medium.
The liquid culture end-point detection method described in this paper, which is capable of detecting M. tuberculosis from 1 ml of sputum harboring approximately 60 organisms, is important. The method used conventional centrifugation, but application of the microcentrifugation approach with TS and modification of other parameters may further enhance this sensitivity. The ability of progeny D29 to infect M. smegmatis sensor cells immediately following lytic release reduces the level of exposure to the destabilizing effects of sputum residues after processing. Rapid infection and lysis of M. smegmatis enable an exponential markup of the signal and, therefore, an extremely sensitive system. However, the method described here is limited. Growth of contaminating bacteria would cause false-negative results, since the MTT color change is dependent on metabolically active cells, but not exclusively M. smegmatis cells. Conversely, the method would yield false-positive results in cases in which extracellular D29 phage survives treatment with the virucide FAS. This has been observed in a small number of cases, in which one or two breakthrough plaques result. At this stage, it is not clear whether insufficient exposure to FAS prior to plating, an excess volume of sputum, the spontaneous reassociation of disrupted D29 particles, or the presence of mycobacteria in smear-negative sputum is the cause. The reassociation of disrupted D29 or the presence of mycobacteria in smear-negative sputum would appear to be consistent with the very low number of breakthrough plaques observed when such cases arose. In a plate-based system, a cutoff number of plaques could be determined, above which a positive result would be returned. By using a liquid system, continuous monitoring could resolve the problem, in which the basis for the cutoff could be determined by use of a graph profile. Conceivably, the problems of breakthrough D29 and contaminant growth could be resolved by using an approach with dual reporter phages. Each phage might code for the production of one component of a reporter protein (similar to the β-lactamase, α-peptide system, but with a reporter not encoded by potential contaminants, e.g., green fluorescent protein or firefly luciferase [1, 10, 16, 17, 19]). Upon dual infection of host M. tuberculosis, a recombination event could produce chimeric phages encoding both required components. In combination with the FAS and M. smegmatis aspects of the PhaB assay, this could represent both a highly specific and a highly sensitive method with direct, positive signal detection. In principle, neither contaminants nor breakthrough D29 would compromise the final outcome. Phage breeding to produce an enhanced burst size and enhanced extracellular stability could also prove beneficial (2).
The approaches discussed in this paper require validation with clinical specimens. However, they do enable a greater understanding of the PhaB assay and indicate how the method might be best applied to sputum in a routine clinical setting. Besides offering a rapid alternative to culture as a method for detection of mycobacteria and drug susceptibility testing of isolates, the ultimate goal for the PhaB assay would be the direct susceptibility profiling of smear-positive primary specimens.
REFERENCES
- 1.Carrière, C., P. F. Riska, O. Zimhony, J. Kriakov, S. Bardarov, J. Burns, J. Chan, and W. R. Jacobs. 1997. Conditionally replicating luciferase reporter phages: improved sensitivity for rapid detection and assessment of drug susceptibility of Mycobacterium tuberculosis. J. Clin. Microbiol. 35:3232-3239. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Chatterjee, S., M. Mitra, and S. K. Das Gupta. 2000. A high yielding mutant of mycobacteriophage L1 and its application as a diagnostic tool. FEMS Microbiol. Lett. 188:47-53. [DOI] [PubMed] [Google Scholar]
- 3.Collins, C. H., J. M. Grange, and M. D. Yates. 1997. Tuberculosis bacteriology: organisation and practice. Butterworth Heinemann, Oxford, England.
- 4.David, H. L., S. Clavel, and F. Clement. 1979. Absorption and growth of the bacteriophage D29 in selected mycobacteria. Ann. Virol. 131:167-184. [Google Scholar]
- 5.Drobniewski, F. A., A. Pablos-Méndez, and M. C. Raviglione. 1997. Epidemiology of tuberculosis in the world. Semin. Respir. Crit. Care Med 18:419-429. [Google Scholar]
- 6.Drobniewski, F. A., and S. M. Wilson. 1998. New biomolecular assays must be tested by direct study in the developing world. BMJ 316:940.. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Dye, C., G. P. Garnett, A. Sleeman, and B. G. Williams. 1998. Prospects for worldwide tuberculosis control under the W.H.O. DOTS strategy. Lancet 352:1886-1891. [DOI] [PubMed] [Google Scholar]
- 8.Eltringham, I. J., F. A. Drobniewski, J. A. Mangan, P. D. Butcher, and S. M. Wilson. 1999. Evaluation of reverse transcription-PCR and a bacteriophage-based assay for rapid phenotypic detection of rifampin resistance in clinical isolates of Mycobacterium tuberculosis. J. Clin. Microbiol. 37:3524-3527. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Eltringham, I. J., S. M. Wilson, and F. A. Drobniewski. 1999. Evaluation of a bacteriophage-based assay (phage amplified biologically assay) as a rapid screen for resistance to isoniazid, ethambutol, streptomycin, pyrazinamide, and ciprofloxacin among clinical isolates of Mycobacterium tuberculosis. J. Clin. Microbiol. 37:3528-3532. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Jacobs, W. R., R. G. Barletta, R. Udani, J. Chan, G. Kalkut, G. Sosne, T. Kieser, G. J. Sarkis, G. F. Hatfull, and B. R. Bloom. 1993. Rapid assessment of drug susceptibilities of Mycobacterium tuberculosis by means of luciferase reporter phages. Science 260:819-822. [DOI] [PubMed] [Google Scholar]
- 11.Krishnamsurti, G. S. R., and P. M. Huang. 1991. Influence of citrate on the kinetics of Fe(II) oxidation and the formation of iron oxyhydroxides. Clays Minerals 39:28-34. [Google Scholar]
- 12.McNerney, R., S. M. Wilson, A. M. Sidhu, V. S. Harley, Z. Al Suwaidi, P. M. Nye, T. Parish, and N. G. Stoker. 1998. Inactivation of mycobacteriophage D29 using ferrous ammonium sulphate as a tool for the detection of viable Mycobacterium smegmatis and M. tuberculosis. Res. Microbiol. 149:487-495. [DOI] [PubMed] [Google Scholar]
- 13.Pablos-Méndez, A., M. C. Raviglione, and A. Laszlo, et al. 1998. Global surveillance for antituberculosis drug resistance, 1994-1997. N. Engl. J. Med. 338:1641-1649. [DOI] [PubMed] [Google Scholar]
- 14.Ratnam, S., and S. B. March. 1986. Effect of relative centrifugal force and centrifugation time on sedimentation of mycobacteria in clinical specimens. J. Clin. Microbiol. 23:582-585. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Riska, P. F., and, W. R. Jacobs. 1998. The use of luciferase-reporter phage for antibiotic-susceptibility testing of mycobacteria, p. 431-455. In Mycobacteria protocols. Humana Press Inc., Totowa, N.J. [DOI] [PubMed]
- 16.Riska, P. F., W. R. Jacobs, B. R. Bloom, J. McKitrick, and J. Chan. 1997. Specific identification of Mycobacterium tuberculosis with the luciferase reporter mycobacteriophage: use of p-nitro-β-hydroxypropiophenone. J. Clin. Microbiol. 35:3225-3231. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Riska, P. F., Y. Su, S. Bardarov, L. Freundlich, G. Sarkis, G. Hatfull, C. Carriere, V. Kumar, J. Chan, and W. R. Jacobs. 1999. Rapid film-based determination of antibiotic susceptibilities of Mycobacterium tuberculosis strains by using a luciferase reporter phage and the Bronx box. J. Clin. Microbiol. 37:1144-1149. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Roberts, G. D., N. L. Goodman, L. Heifets, H. W. Larsh, T. H. Lindner, J. K. McClatchy, M. R. McGinnis, S. H. Siddiqi, and P. Wright. 1983. Evaluation of the BACTEC radiometric method for recovery of mycobacteria and drug susceptibility testing of Mycobacterium tuberculosis from acid-fast smear-positive specimens. J. Clin. Microbiol. 18:689-696. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Sarkis, G. J., W. R. Jacobs, and G. F. Hatfull. 1995. L5 luciferase reporter mycobacteriophages: a sensitive tool for the detection and assay of live mycobacteria. Mol. Microbiol. 15:1055-1067. [DOI] [PubMed] [Google Scholar]
- 20.Snider, D. E., M. Raviglione, and A. Kochi. 1994. Global burden of tuberculosis, p. 3-11. In B. R. Bloom (ed.), Tuberculosis: pathogenesis, protection and control. American Society for Microbiology, Washington D.C.
- 21.Wilson, S. M., Z. Al-Suwaidi, R. McNerney, J. Porter, and F. Drobniewski 1997. Evaluation of a new rapid bacteriophage-based method for the drug susceptibility testing of Mycobacterium tuberculosis. Nat. Med. 34:465-468. [DOI] [PubMed] [Google Scholar]