Design of a single-tube, endpoint, linear-after-the-exponential-PCR assay for 17 pathogens associated with sepsis



Lawrence J. Wangh, Department of Biology, Brandeis University, Waltham, MA 02453-2728, USA. E-mail:



The goal of this study was to construct a single-tube multiplex molecular diagnostic assay using linear-after-the-exponential (LATE)-PCR for the detection of 17 microbial pathogens commonly associated with septicaemia.

Methods and Results

The assay described here detects 17 pathogens associated with sepsis via amplification and analysis of gene-specific sequences. The pathogens and their targeted genes were: Klebsiella spp. (phoE); Acinetobacter baumannii (gyrB); Staphylococcus aureus (spa); Enterobacter spp. (thdF); Pseudomonas aeruginosa (toxA); coagulase-negative staphylococci (tuf), Enterococcus spp. (tuf); Candida spp. (P450). A sequence from an unidentified gene in Lactococcus lactis, served as a positive control for assay function. LATE-PCR was used to generate single-stranded amplicons that were analysed at endpoint over a wide range of temperatures in four fluorescent colours. Each target was detected by its pattern of hybridization to a sequence-specific low-temperature fluorescent probe derived from molecular beacons.


All 17 microbial targets were detected in samples containing low numbers of pathogen genomes in the presence of high levels of human genomic DNA.

Significance and Impact of the Study

This assay used new technology to achieve an advance in the field of molecular diagnostics: a single-tube assay for detection of pathogens commonly responsible for septicaemia.


Linear-after-the-exponential polymerase chain reaction (LATE-PCR) is an advanced form of nonsymmetric PCR in which amplification of double-stranded DNA is followed by efficient production of single-stranded amplicons. LATE-PCR also allows for a high degree of multiplexing followed by endpoint detection by hybridization of low-temperature fluorescent probes in multiple colours over a wide range of temperatures (Sanchez et al. 2004, 2006; Hartshorn et al. 2005a,b, 2007; Pierce et al. 2005; Salk et al. 2006; Rice et al. 2007). This highly multiplexed single-tube approach is not achievable using conventional symmetric PCR which only generates double-stranded amplicons.

We have previously described relatively simple LATE-PCR assays for foot-and-mouth disease virus (Pierce et al. 2009), African swine fever virus (Ronish et al. 2010) and mitochondrial DNA mutational load (Osborne et al. 2009). The highly multiplexed assay described here can amplify and detect any one of 17 possible bacterial and fungal targets commonly associated with septicaemia, a condition that must be diagnosed in a matter of hours because of its high rate of mortality (Balk et al. 2001). The assay was constructed by first identifying specific gene sequences unique to each bacterium or fungus. The limiting and excess primers and a fluorescent probe for each of these targets were then combined stepwise to build the multiplex assay. The complete assay described here has undergone preliminary evaluation at the University of California at Davis, School of Medicine and the results of that study are being published as an accompanying paper (Gentile et al. 2012).

Materials and methods

Configuration of the sepsis gene specific LATE-PCR multiplex assay

The scheme shown in Table 1 depicts the location of each of the pathogens in a two dimensional detection space comprised of fluorescent colour and temperature. Low-melting temperature (Tm) molecular beacons, 64–43°C, in four fluorescent colours were used to detect their corresponding targets with high specificity. Subdivision of the colour/temperature space in this manner made it possible to detect 17 different pathogens plus an internal control. Table 2 shows the pathogens by dye channel, as well as the target genes and the GenBank accession ID numbers from which the primers and probes were designed.

Table 1. The linear-after-the-exponential-PCR multiplex assay configuration
Temperature, °CFAMCal OrangeCal RedQuasar
image_n/jam12061-gra-0001.png Klebsiella pneumoniae Staphylococcus aureus Pseudomonas aeruginosa Candida albicans
Acinetobacter baumannii Enterobacter aerogenes Staphylococcus haemolyticus Candida glabrata
Klebsiella oxytoca Enterobacter cloacae Staphylococcus saprophyticus Enterococcus faecalis
Lactococcus lactis   Staphylococcus epidermidis Enterococcus faecium
   Staphylococcus hominis Candida parapsilosis
    Candida tropicalis
Table 2. Reference sequences for primer probe design in linear-after-the-exponential-PCR multiplex assay
ChannelTargetGeneAccession IDAcquisition
FAM Klebsiella pneumoniae phoEM28295ATCC: BAA-1706D -5
Acinetobacter baumannii gyrB NC_011595 ATCC: 17978D-5
Klebsiella oxytoca phoEX68022ATCC: 700324D-5
Lactococcus lactis Unknown NC_008527 Aliquot Edgewood
CalOrg Staphylococcus aureus spa AM076292 Aliquot PHRI
Enterobacter aerogenes thdF EU569333 ATCC: 15038D-5
Enterobacter cloacae thdF EU569339 ATCC: 13047D-5
CalRed Pseudomonas aeruginosa toxA NC_008463 ATCC: 17933D-5
Staphylococcus haemolyticus tuf AF298801 ATCC: 29970D-5
Staphylococcus saprophyticus tuf AF298804 ATCC: 15305D-5
Staphylococcus epidermidis tuf AF298800 ATCC: 12228D-5
Staphylococcus hominis tuf AF298802 ATCC: 700236D-5
Quasar Candida albicans P450 AF153846 ATCC: 14053D-5
Candida glabrata P450 AY942647 ATCC: 15545D-5
Enterococcus faecalis tuf AE016830 ATCC: 700802D-5
Enterococcus faecium tuf NZ_ACHL01000063ATCC: BAA-472D-5
Candida parapsilosis P450 GQ302972 ATCC: 22019D-5
Candida tropicalis P450 AY942643 ATCC: 750D-5

Design features of LATE-PCR multiplex assays

LATE-PCR is an advanced form of nonsymmetric PCR. Each monoplex reaction utilizes a limiting primer (L) and an excess primer (X) whose initial Tm fit the formula Tm0− Tm0X ≥ 0 (Sanchez et al. 2004). Primers of this design efficiently generate double-stranded amplicons until the limiting primer runs out and the reaction switches to linear amplification of a single-stranded amplicon driven by extension of the excess primer. The accumulated single-stranded targets are measured at endpoint, by hybridization to low-temperature sequence-specific fluorescent probes.

Construction of a highly multiplexed assay for sepsis began with bioinformatic identification of specific primer pairs that would amplify gene-specific sequences characteristic of a particular organism. Each gene sequence was scrutinized using Basic Local Alignment Search Tools (Blast) and was then compared with all other sequences in the NCBI Genbank. The chosen sequences were then further analysed to identify relatively conserved organism-specific targets. Each of these sequences, in turn, was downloaded into Visual OMP (VOMP) software (DNASoftware, Inc., Ann Arbor, MI, USA) for primer and probe design.

LATE-PCR primers and probes were designed with multiplexing in mind using the following criteria. All limiting primers were picked to have average concentration-dependent Tm's of 71·1°C (range 73·6–68·0°C), and all excess primers of 68·2°C (range 71·4–66·0°C). In addition, the Tm of each excess primer was designed to have a lower Tm than the Tm of its paired limiting primer. The design process also circumvented possible primer-primer or primer-probe interactions. Each pair of prospective primers and its corresponding probe sequences were then run through another Blast search to confirm that they had the appropriate homology to identify the target sequences, and were also sufficiently divergent from related targets to prevent amplification or detection of the wrong target. Collectively these precautions greatly reduce the chances of false positives.

Table 3 describes the final primers, probes and amplicons in terms of their sequences and in silico Tm's as determined using VOMP software. Conserved gene regions were selected on either the species level, or across several species, as shown in Table 3. Some of the chosen pathogen targets had large numbers of strain sequences to examine which aided in design. A few of the pathogens, namely Enterobacter cloacae, had very little sequence data available, so the amplicon was designed from the single available sequence. Also, the dye channel and temperature space were chosen to detect each of the specific products at endpoint with a low Tm mismatch tolerant probe. Primers and probes were designed in an effort to conserve, and make efficient use of the available fluorescent colour and temperature space leading to all species being detected.

Table 3. Primer and probe sequences in linear-after-the-exponential-PCR multiplex assay
NameSequence 5′–3′Bases numberTm °C
  1. Coag Negative, coagulase-negative staphylococcus; Candida APT, C. albicans, C. parapsilosis, and C. tropicalis.

Kl. pneumoniae excess primerTCGCTGAAGTAGTGCATGGCTTTGATC2769·5
Kl. pneumoniae limiting primerAAAGCTTTGTGGCTTCAACGGCGACG2671·3
Kl. pneumoniae probeBHQ-1-TAAAGAACGCGAACAAGCTGGTA- FAM2364·4
Ac. baumannii excess primerGCATTGCAATGGAACGATAGTTAC2466·3
Ac. baumannii limiting primerAGATACTGGTTTAAGCCACGTGTTAAAGCTGCG3372·0
Ac. baumannii probeBHQ1 – TGTGTGAAACAGCGAACATTCA – FAM2258·1
Kl. oxytoca excess primerAGGCTCATATAGTGCATCGTTTTGACT2768·9
Kl. oxytoca limiting primerGATGATGGGTCTTATTGCTTCTTCGGCTACC3171·8
L. lactis probeBHQ 1 – ATAAACCTTTCTTAAAAT – FAM1843·9
Staph. aureus excess primerTGAATATGCCTAACTTGAACGAAGAACAACG3171·4
Staph. aureus probeCal Orange – AATGGGTCATCTTTTAAGCTTTGGTT – BHQ12662·4
Ent. aerogenes excess primerGCTTCAATTTTACCATCGGAGAGGAA2668·2
Ent. aerogenes limiting primerGTGGAAGCGCTTACTCACCTGCGCATCT2872·0
Enterobacter probeCal Orange – TCTTCCCGGATGAAGAAATGA – BHQ12155·6
Ent. cloacae excess primerGCTTCAATTTTACCGTCAGAGAGAAA2665·9
Ent. cloacae limiting primerTCTTGTGGAAGCACTTACTCACCTCAGGATCT3271·0
Ps. aeruginosa excess primerTGAACTGGCTGGTATCGATCGG2269·9
Ps. aeruginosa limiting primerCTGAGCTGGTTACCGGCGTTCAGTTC2671·2
Ps. aeruginosa probeCal Red – AAGGATGAACACCTTGATGTTCGATT – BHQ22663·1
Coag Negative probeCalRed- AACTGTTACTGGTGTAGAATT-BHQ22159
Enterococcus probeQuasar 670 – AAGCATCATGAATGGTTTGTT – BHQ22155·5

Target-specific sets of primers and probes were first tested in monoplex assays using genomic DNA of the relevant pathogen as the target. All primer pairs and all probes were combined into a multiplex assay which was tested separately against the genomic DNA of each pathogen. All such reactions were carried out in triplicate along with many replicates of no-template-controls (NTC). This iterative process allows for optimization of all components in such a multiplex reaction. Reaction mixtures that were efficient using high number of genomic copies were then tested even more rigorously by 10-fold serial dilution of the genomic DNA from 105 to 10−2 copies per reaction. Only those reactions that have very low levels of nonspecific hybridization are reliable at high dilution. Genomic DNA used to determine the limit of detection for each pathogen was first quantified via real-time monoplex LATE-PCR analysis with SYBR Green.

Assay composition

Every experiment had a final volume of 25 μl and contained the same reaction component, mixture comprised of: 1× PCR buffer (cat. no. 10966-034; Invitrogen, Carlsbad, CA, USA), 3 mmol l−1 MgCl2 (cat. no. 10966-034; Invitrogen), 250 nmol l−1 dNTPs (Invitrogen), PCR Grade Water ( BP2819-1; Fisher Scientific), 100 nmol l−1 of each probe (Biosearch Technologies, Novato, CA, USA), 50 nmol l−1 of each limiting primer (Sigma-Aldrich, St Louis, MO, USA) and 1000 nmol l−1 of each excess primer (Sigma-Aldrich). All reactions contained 1·25 units of Platinum Taq DNA polymerase (cat. no. 10966-034; Invitrogen) and 1 μl of target genomic DNA. Target genomic DNA (Table 2) was received from American Type Culture Collection (ATCC, Manassas, VA, USA) as a dry reagent (5 µg) and was suspended in 500 μl of 10 mmol l−1 Tris-Cl pH 8·3 (Sigma-Aldrich) for a final genomic DNA concentration 106 genomes per µl or as previously characterized suspended purified genomic DNA from University of California Davis, Smiths Detection Diagnostics or Public Health Research Institute (PHRI). SYBR Green dilution series were used to verify the starting genomic DNA copy number for all pathogens (Dhanasekaran et al. 2010). All multiplex mixture and limit of detection reactions contained 10 000 copies of human genomic DNA (cat. no. G304A; Promega, Madison, WI, USA) which was included to verify the robustness and specificity of the reaction.

LATE-PCR protocol and thermal amplification parameters

Each reaction was run in triplicate in either a Bio-Rad IQ5 Multicolor Real-Time Detection System (Bio-Rad Laboratories, Hercules, CA, USA) or in the Stratagene Mx3005P (Agilent Technologies, Santa Clara, CA, USA). No significant endpoint differences were observed for the two machines. Three NTCs were also run in parallel in every experiment. An initial denaturation step at 95°C for 3 min was followed by 50 amplification cycles of 95°C for 10 s, 65°C for 15 s and 72°C for 30 s. Next, the temperature was dropped from 72 to 25°C. Beginning at 25°C, the temperature was increased in 56 cycles of 30-s steps and 1°C increments to 80°C. Empirical observation suggests that this down-up process improves the homogeneity of molecules in the reaction (data not shown). Once the molecules have reached this level the reproducibility of the data is enhanced. Finally, endpoint data collection was carried out by decreasing the temperature, annealing in 1°C decrements at 30 s intervals for 56 steps between 80 and 25°C. Fluorescence was measured in the FAM, CalOrg560, CalRed610 and Quasar670 channels at each anneal step.

Data analysis

The fluorescent contours were generated as follows: The raw fluorescent endpoint data for each temperature step and each dye channel collected on either the Bio-Rad IQ5 or the Stratagene Mx3005P were exported to Microsoft Excel 2007. Each set of data was normalized by dividing all values by the value of the signal at 70°C in that data set. This was because no probe was bound to any target at this temperature and the fluorescence intensity in each colour was at the background level for all probes of that colour. Next, each data set was corrected for background temperature-dependent levels of fluorescence due to unbound probes in that colour. This was accomplished by subtracting the background temperature-dependent level of fluorescence of an average NTC sample from each corresponding experimental sample in that colour. At this point all positive fluorescent signals above background for each pathogen in their respective colour were re-scaled by dividing all fluorescent values in a data set by the highest fluorescent value in that data set. The resulting fluorescent values describe a temperature-dependent fluorescent contour in relative fluorescent unit (RFU) values on a scale of 0-to-1. The fluorescent signatures illustrated in this article are the first derivatives of the raw fluorescent data collected by either the Bio-Rad IQ5 or the Stratagene Mx3005P, and were generated using the algorithms built into each machine.


The results presented here describe the design of a highly multiplexed LATE-PCR single-tube assay for detection of 17 microbial pathogens commonly associated with septicaemia. The complete LATE-PCR sepsis multiplex assay was comprised of 24 primers and 11 probes distributed into four fluorescent colours as shown in Table 1. The success of the assay rests on the bioinformatic analysis of the chosen DNA sequences which were both target specific and compatible with properties of LATE-PCR limiting and excess primers.

Figure 1 displays the results of the fluorescent contours and their corresponding fluorescent signatures for each of the genomic DNA targets analysed one-by-one in triplicate in a master mix of all primers and probes. The data in Fig. 1 were generated using 106 copies of each genome. The results established that each target pathogen had its own unique fluorescent pattern, except for a group of four coagulase-negative staphylococci (CoNS) (Fig. 1 Panel c1) that had a unique fluorescent signature, but cannot be distinguished individually. These results were expected and were a consequence of probe design. The probe for the four CoNS was designed not to distinguish on the species level. Consequently, as shown in Fig. c1, the CoNS species (i.e., Staphylococcus haemolyticus, Staphylococcus saprophyticus, Staphylococcus epidermidis and Staphylococcus hominis) have the same CalRed610 fluorescent signals and were detectable only as CoNS.

Figure 1.

The detection of 17 pathogens associated with sepsis. Anneal curves (a1, b1, c1, d1) are fluorescent contours. All 1st derivative anneal curves (a2, b2, c2, d2) are fluorescent signatures. (a1) FAM: Klebsiella pneumoniae – black dotted, Acinetobacter baumannii – black solid, Klebsiella oxytoca – black dashed and Lactococcus lactis – grey solid. (b1) CalOrg560: Staphylococcus aureus – black dashed, Enterobacter cloacae – solid grey and Enterobacter aerogenes – solid black (c1) CalRed610: Pseudomonas aeruginosa – black dotted, coagulase-negative staphylococci; including Staphylococcus hominis, Staphylococcus saprophyticus, Staphylococcus epidermidis and Staphylococcus haemolyticus – grey and black solid and dashed. (d1) Quasar670: Candida albicans – black solid, Candida glabrata – black dotted, Enterococcus faecalis – black dashed and Enterococcus faecium – grey dashed, Candida tropicalis – grey solid and Candida parapsilosis – black hashed.

Some probes were designed to differentiate between similar pathogens while still conserving the number of probes used in the multiplex because of dye channel and temperature detection space restrictions. A single probe can be designed for detection and differentiation of multiple pathogens by ensuring that each probe/target hybrid has its own Tm. As shown in Fig. d1, the probe that bound to Candida albicans at 63°C was the same probe that bound to Candida tropicalis at 50°C (refer to Table 3), and thus generated pathogen specific patterns. A second Candida probe was designed for multiple pathogen detection for differentiation of two other Candida pathogens to ensure that the probe to amplicon binding produced distinctly specific contours through nucleotide mismatching. Candida parapsilosis and C. tropicalis both bound their detection probe at 50°C, however, the contours of their anneal curves looked distinctly different, as shown in Fig. d1. At progressively lower temperatures the C. parapsilosis signal dips below the C. tropicalis, thus distinguishing the contours of these two species.

Limit of detection of genomic DNA samples

The limit of detection of each genomic DNA target is critical to this assay's clinically relevance, as in septicaemia the concentration of pathogens in a blood sample varies from very low to very high depending on the extent of infection. As shown in Table 4, the limit of detection was dependent on several factors: the sample preparation, the pathogen, the fluorescent dye channel, the detection temperature, starting copy number, and the amount of single-stranded DNA produced during LATE-PCR amplification. A balance was achieved between the efficiency of target amplification by optimization of primer design, and the number of cycles used for single-stranded amplicon accumulation. The more efficient the primers or the larger the number of initial genomes, the smaller the number of amplification cycles required to reach a detectable concentration of a single-stranded product. In this assay, the high level of primer specificity and low level of nonspecific interactions allowed for detection of small numbers of initial genomes. Positive detection of a particular pathogen was made when all three replicates were above the NTC background adjusted for each fluorescent colour. Typically, the signal to background ratio in the FAM dye channel was lower in comparison with the other dye channels utilized in this assay. The CalRed610 and Quasar670 dye channels produced higher fluorescent signals relative to background fluorescence. Lower temperature probes were more affected by background fluorescent levels because of increasing secondary structure of the target at lower temperatures. When background fluorescent levels were high, the same low temperature probes do not generate a high enough signal for the lower copy numbers to be distinguishable from the background fluorescence.

Table 4. Limit of detection of genomic DNA targets by copy number in linear-after-the-exponential-PCR multiplex assaya
Temperature, °CFAMCal OrangeCal RedQuasar
  1. a

    Genomic DNA dilution series were previously quantified in monoplex using SYBR-Green (Applied Biosystems, Foster City, CA, USA).

image_n/jam12061-gra-0002.png Klebsiella pneumoniae Staphylococcus aureus Pseudomonas aeruginosa Candida albicans
Acinetobacter baumannii Enterobacter aerogenes Staphylococcus haemolyticus Candida glabrata
Klebsiella oxytoca Enterobacter cloacae Staphylococcus saprophyticus Enterococcus faecalis
Lactococcus lactis   Staphylococcus epidermidis Enterococcus faecium
1000 10100
   Staphylococcus hominis Candida parapsilosis
    Candida tropicalis

The genomic DNA targets with the highest limits of detection were Klebsiella oxytoca, Lactococcus lactis, Enterobacter aerogenes and Ent. cloacae with a limit of detection of 1000 copies. The Kl. oxytoca and the L. lactis, annealing at the temperatures of 54·3 and 43·9°C, respectively, have a relatively high limit of detection because they were in the FAM dye channel with their respective probes expected to anneal at a lower temperature than the Klebsiella pneumoniae at 64·4°C (limit of detection 10 copies) and Acinetobacter baumannii at 58·1°C (limit of detection 100 copies). The use of four FAM probes in the full multiplex of this assay detrimentally contributed to background fluorescence. In the absence of target, the unbound probes greatly increased the background making the lower annealing temperature probes (Kl. oxytoca and the L. lactis) difficult to discern from background fluorescence at low copy number. Acinetobacter baumannii (limit of detection 100 copies) was also affected by background fluorescence in the FAM channel but not to the same extent because its probe had a higher annealing temperature relative to the Kl. oxytoca and L. lactis.

In the CalOrg560 dye channel, Ent. aerogenes and the Ent. cloacae shared a single probe and had a limit of detection of 1000 copies. The Enterobacter probe, although perfectly matched to both amplicons, gave species specific probe contours because of the design of the probe stem. The four nucleotides, which make up the probe stem, were typically designed to bind to each other and not the amplicon. The probe stem of the Enterobacter probe was different for two reasons. The first reason was that one half of the stem binds to the Ent. cloacae sequence and not to the Ent. aerogenes sequence, providing for species specific probe contours. The second reason was that the stem was composed of the nucleotides TC and GA. Typically stem design for low temperature molecular beacons uses only A's and T's to allow the probe to readily open. The presence of the C/G nucleotide binding slowed the opening of the probe stem and affected the probe's binding efficiency, making it less efficient at the lower copy number, but still provided discrimination between the two targets. The other pathogen that was detected in the CalOrg560 dye channel was Staphylococcus aureus, which had a limit of detection of 10 copies with its probe binding at 62°C.

Another factor that induces a higher limit of detection was the use of mismatched tolerant primers and probes because they do not bind perfectly to the target and could delay amplification. If a primer is mismatched but the mismatch was not destabilizing, the effect on amplification delay can be minimized. All pathogens detected in the CalRed610 dye channel had a limit of detection of 10 copies or less. Pseudomonas aeruginosa was detected with a high-temperature probe (63·1°C). A slightly lower temperature probe (59°C) detected pathogens as CoNS, which included Staph. haemolyticus, Staph. saprophyticus, Staph. epidermidis and Staph. hominis. It is significant to note that there was a single primer pair that amplifies all four CoNS. Each target was mismatched differently to both the limiting and excess primers but all were perfectly matched to the probe. However, although both primers were mismatched, the mismatches were not destabilizing and the effect on amplification delay was minimized and the limit of detection remained low.

Of the four fluorescent dyes used here, Quasar670 has the most capacity for pathogen detection because its signal intensity over background is greatest, even at low temperatures. For this reason, Quasar670 was chosen for analysis of the largest number of pathogens (six), as compared to four in FAM, three in CalOrg560 and five in CalRed610. The six pathogens analysed in Quasar can be considered in three groups: (i) C. albicans, C. parapsilosis, C. tropicalis; (ii) Candida glabrata and (iii) Enterococcus faecalis, Enterococcus faecium. C. albicans, C. parapsilosis and C. tropicalis were amplified by a single mismatch tolerant primer pair to the P450 gene target. As shown in Table 3, the probe was perfectly matched to the C. albicans amplicon allowing for detection at 63°C. It was intentionally designed to be mismatched to C. parapsilosis and C. tropicals allowing for detection of all three species down to a level of 100 copies. The C. glabrata P450 gene target was not sufficiently conserved to be amplified by the primers used for group 1 and was therefore amplified by its own pair of primers and a probe which achieved a limit of detection of 10 copies. The species in group 3, Ent. faecalis and Ent. faecium were amplified with a set of mismatch tolerant primers that amplified both species and were distinguished using a perfectly matched probe whose binding was impacted by an adjacent sequence that had a distinctly different secondary structure in the two species. This novel approach allowed the targets to be distinguished all the way down to a limit of detection of 100 copies.

Mixtures of genomic DNA samples

LATE-PCR primers detected low copy levels of a pathogen in a large background of other pathogens. This is important in the clinical setting, in which mixtures of pathogens can be present either through true polymicrobial infection or contamination of sample during acqusition. As shown in Table 1, pathogen detection was divided into four different dye channels, such that some pathogens were detected in a shared dye channel. For this reason, LATE-PCR endpoint multiplex analysis of mixtures was tested between different dye channels and within a single dye channel. In every multiplex experiment, when a single pathogen was present, only one dye channel had a positive signal above background. When a mixture of pathogens between dye channels was present, two dye channels had separate positive signals above background. This established that each pair of primers functioned independently even though all the primers and probes were present in the reaction mixture.

Mixtures of pathogens in this LATE-PCR multiplex assay were carried out by preparation of genomic DNA targets at a ratio of 99 : 1 using 105 and 103 copies of the chosen genomes. All assays also contained a background of 104 copies of human genomic DNA to rigorously test primer and probe specificities. Figure 2 illustrates the normalized fluorescent contours for a representative set of results for the many possible combinations of pathogen mixtures across the four dye channels. In this case, Staph. aureus was detected by virtue of its signals in CalOrg560, which was present in each two-pathogen mixture. The other dye channels FAM, CalRed610 and Quasar670 were used for the detection of the other variable pathogen in the two-pathogen mixture: Kl. pneumoniae visible in FAM, Ps. aeruginosa visible in CalRed610 and Ent. faecalis visible in Quasar670 (see Table 1 for assay configuration).

Figure 2.

The use of two dye channels for the detection of 1–99% mixtures. Solid black lines 100% Staphylococcus aureus for all panels. (a) CalOrg560 and FAM fluorescent contours show the Staph. aureus at the 1% level (grey a1) in a background of 99% Klebsiella pneumoniae (grey a2); the reverse mixture is shown by Kl. pneumoniae at the 1% level (black dashed a2) in a background of 99% Staph. aureus (black dashed a1). (b) CalOrg560 and CalRed610 fluorescent contours of Staph. aureus at the 1% level (grey b1) in a background of 99% Pseudomonas aeruginosa (grey b2); the reverse mixture is shown by Ps. aeruginosa at the 1% level (black dashed b2) in a background of 99% Staph. aureus (black dashed b1). (c) The CalOrg560 and Quasar670 fluorescent contour shows the detection of the mixture of Staph. aureus at the 1% level (grey c1) in a background of 99% Enterococcus faecalis (grey c2); the reverse mixture shows detection of Enterococcus faecalis at the 1% level (black dashed c2) in a background of 99% Staph. aureus (black dashed c1).

In Fig. 2a, the FAM and CalOrg560 dye channels were used to detect Kl. pneumoniae (FAM) at the 1% level in a background of 99% Staph. aureus (CalOrg560) and 1% Staph. aureus (CalOrg560) in a background of 99% Kl. pneumoniae (FAM). In the CalRed610 dye channel (Fig. 2b), Ps. aeruginosa (CalRed610) was detected at the 1% level in a background of 99% Staph. aureus (CalOrg560) and 1% Staph. aureus (CalOrg560) in a background of 99% Ps. aeruginosa (CalRed610). In the Quasar670 dye channel (Fig. 2c), Ent. faecalis (Quasar670) was detected at the 1% level in a background of 99% Staph. aureus (CalOrg560) and 1% Staph. aureus (CalOrg560) in a background of 99% Ent. faecalis (Quasar670).

As a mixture of pathogens might also be present within the same dye channel, tests were carried out to analyse this possibility a well. In Fig. 3a, 1% Staph. aureus, was mixed into a background of 99% Ent. aerogenes and both were detected in the CalOrg560 dye channel. Positive controls of 100% Staph. aureus and Ent. aerogenes were run in parallel. In another experiment (Fig. 3b), 1% Ps. aeruginosa was mixed in a background of 99% CoNS (Staph. hominis) and was detected in the CalRed610 channel. Positive controls of 100% Ps. aeruginosa and Staph. hominis were run in parralel. Referring back to Table 1, other combinations of mixtures within a single dye channel were possible; Figure 3 provides only two representative examples. The differentiation of mixtures in a single dye channel is only limited by the resolution of specific fluorescent probes to specific pathogens. Therefore, CoNS mixtures would not be resolvable nor would mixtures of Ent. facalis and Ent. faecium, as in these cases similar fluorescent contuors and fluorescent signatures result.

Figure 3.

Shows the detection of complex mixture in a single dye channel. Fluorescent contours (a1, b1) were normalized as described in Materials and methods. The corresponding fluorescent signatures (a2, b2) show that the inflection points of the mixtures are the same as the positive controls. (a1) In CalOrg560 the detection of the positive controls: Staphylococcus aureus (solid black line) and Enterobacter aerogenes (solid grey line) and the mixture of 1% Staph. aureus, in a background of 99% Ent. aerogenes (dotted black line). (b1) In CalRed610 the detection of the positive controls Pseudomonas aeruginosa, (solid black line) and of Staphylococcus hominis (coagulase-negative staphylococcus – solid grey line) and the mixture of 1% Ps. aeruginosa, in a background of 99% Staphylococcus hominis (dotted line).

Even more complex mixtures of pathogens were analysed by mixing three pathogens. Figure 4, through fluorescent signatures, illustrates the representative use of two dye channels for the detection of mixtures of three pathogens. Pseudomonas aeruginosa and Staph. hominis which were detected in the CalRed610 dye channel were mixed with C. glabrata which was detected in the Quasar670 dye channel (see Table 1 for further clarification of assay configuration). Equal mixtures of the three ATCC genomic DNA targets were made at both the 100 000 copy level and the 1000 copy level to serve as positive controls. Mixtures were also made so that individual genomic DNA targets were at the 1000 copy level while the other two remained at 100 000 copies. All mixtures were run in parallel to allow comparison for detection of the mixture. As shown in Fig. 4, the mixture of Ps. aeruginosa (CalRed610), Staph. hominis (CalRed610) and C. glabrata (Quasar670) were detectable at the 1% level in a background of two other pathogens. Complex mixtures of clinically relevant pathogens were correctly detected.

Figure 4.

Full Multiplex mixtures of three genomic DNA targets, Pseudomonas aeruginosa, Staphylococcus hominis, in CalRed610 (a1) and Candida glabrata in Quasar 670 (a2). The fluorescent signatures are shown. The solid lines are positive controls. The black line represents an equal mixture at 1000 copies each, while the solid grey line represents an equal mixture of 100 000 copies. (a1) In the CalRed610 channel the dashed black line illustrates a mixture of 1% Ps. aeruginosa in a background of 99% Staph. hominis and C. glabrata, the dashed grey line – a mixture of 1% Staph. hominis in a background of 99% Ps. aeruginosa and C. glabrata. (a2) In the Quasar670 channel, the dashed black line – 1% C. glabrata in a background of 99% Staph. hominis and Ps. aeruginosa.


The single-tube LATE-PCR multiplex assay described above can detect 17 microbial and fungal pathogens using genomic DNA. These 17 pathogens were chosen based on discussions with our clinical collaborators at the University of California at Davis and took the clinical prevalence of pathogens and mixtures associated with septicaemia into account. Each of the 17 targets was detected by its specific pattern of hybridization to sequence-specific low temperature probes. This multiplex assay was also shown to resolve complex mixtures constructed with genomic DNA.

Currently, there are several septicaemia assays that have been described and evaluated for clinical relevance (Andrade et al. 2008; Lehmann et al. 2008; Louie et al. 2008; Dierkes et al. 2009; Westh et al. 2009; Zhao et al. 2009; Chakravorty et al. 2010; Bravo et al. 2011). The three most pertinent assays are the SeptiFast assay (Lehmann et al. 2008), a real time iso-thermal assay (Zhao et al. 2009) and a LATE-PCR Molecular Beacon Assay (Chakravorty et al. 2010). All three of these assays provide useful clinical information for bacterial and fungal pathogen detection, but all of these tests use multiple tubes and none is highly multiplexed into a single-tube assay as described here. Moreover, as in the case of many other assays for sepsis that have been described, all three of these assays have limitations in terms of ease of use and clinical breath.

The SeptiFast (Lehmann et al. 2008) assay uses three separate symmetric PCR reaction tubes to differentiate gram-positive bacteria, gram-negative bacteria and fungal pathogens using the internal transcribed spacer region between the 16S and 23S ribosomal DNA sequences for bacteria and the 18S and 5.8S ribosomal sequences for fungi. The assay is prone to the detection of reaction mixture contamination components from unwanted bacteria and the assay uses a cutoff point in cycle threshold (Ct) value to limit detection of low background contamination. This limitation leads to low sensitivity vs blood culture analysis of certain pathogen targets (Lehmann et al. 2008).

The isothermal assay (Zhao et al. 2009) uses the 16S and 28S ribosomal DNA in two reaction tubes for bacterial and fungal detection, and is prone to the same risk of contamination and sensitivity levels as the SeptiFast assay. The LATE-PCR molecular beacon assay (Chakravorty et al. 2010) also targets the 16S ribosomal DNA region for bacterial targets and uses six tubes to accomplish the discrimination of over 100 bacterial signatures, where each tube is designated to resolve a single fluorescent probe. This assay has similar issues as other 16S assays with regard to reaction mixture contamination detection and sensitivity. The need to split a blood sample of low pathogen concentration into six tubes will cause sensitivity issues.

The LATE-PCR single-tube multiplex assay described here expands on the three assays described above and also shows some advantages and limitations. Unlike the 16S based assays, this assay only achieves designed results. Consequently, the assay will not pick up unknown pathogens outside of the genera or species that the assay was designed to detect. This result is actually an advantage in that low levels of contamination in reaction mixture components are not picked up. It is also a limitation in that adding more pathogens to be detected in these four dye channels is not prudent. However, expanding the number of dye channels would be a possibility for the inclusion of more pathogen detection. As shown in the results section, the assay detects mixtures of pathogens cleanly, whether they occur in a single dye channel or in multiple dye channels.

The greatest advantage of the multiplex assay described here is that all bacterial and fungal pathogens are detected in a single tube using only four fluorescent dye channels. This saves dividing precious amounts of targets over several tubes for the analysis. We have also shown that we can easily analyse related species under a single genus for example Candida spp. (P450). The assay still could be further optimized to lower the limit of detection values that are shown in Table 4. Further clinical samples also need to be tested to determine the breath of detection of this assay. The accompanying article discusses a preliminary evaluation of this LATE-PCR single-tube multiplex assay. We conclude that the multiplex assay described here holds promise as a rapid and accurate molecular diagnostic screening method for septicaemia.


We are grateful to Dr Barry Kreiswirth of PHRI, Newark N.J. for genomic DNA from bacterial cultures of 12 strains of Staph. aureus. We also thank the UC Davis POC Technologies Center for providing ATCC genomic DNA samples and Anna Dillier and Samaan Mahmoudzadeh of the Kost Lab for their thorough input and review of this article. This research was supported by the NIH grants (1RC1EB010543-01 and 5RC1EB010643-02) and also by Smiths Detection Diagnostics, Inc., Watford, UK.