Influenza A and B Detection: Multiplex RT-PCR

Influenza virus was targeted in RT-PCR, to evaluate the application of RT-KTq 2 in clinical diagnostics. As a genetically highly dynamic virus known for considerable antigenic variation,^[187] influenza viruses have been the cause for several pandemics.^[188] Consequently, rapid detection methods are required for identifying new outbreaks in time and subsequent medical intervention. Along these lines, RT-PCR methods were established for the detection of influenza viruses, as they provide the ability to distinguish between influenza A and B, enable high sample throughput and are easily adapted to novel targets.^[189] In quantitative real-time PCR, TaqMan probes are commonly used as they ensure high reaction specificity.^[14] The TaqMan Assay relies on the 5´–3´ nuclease activity of a DNA polymerase to cleave a dual-labelled TaqMan probe during hybridization to the complementary target sequence and fluorophore-based detection (for further details see chapter III 3.1.). TaqMan-based multiplex real-time RT-PCR was selected in this experiment to detect influenza viruses A and B simultaneously. As 5’-nuclease activity is required for the hydrolysis of dual-labelled TaqMan probes, an enzyme blend of Taq DNA polymerase wild-type and RT-KTq 2 was employed in this assay.

Extracted RNA was obtained from respiratory swab samples from patients known to be either influenza A or B positive. These were kindly provided by Labor Dr. Brunner, Konstanz (Germany). The samples were analysed in the FAM and HEX channel using specific TaqMan probes consisting of an oligonucleotide conjugated to a 5’ reporter dye 6-carboxyfluorescein (FAM) and a 3’ minor groove binder (MGB) for influenza A detection and an oligonucleotide with a 5’ reporter dye hexachlorofluorescein (HEX) and a 3’ black hole quencher (BHQ-1) for influenza B detection (Figure 33). TaqMan probes conjugated to minor groove binders,^[190] as well as optimized primers recommended for influenza A detection, were employed to further increase the specificity of the reaction. Each reaction set-up contained both primer pairs targeting influenza A and B, respectively, both dual-labelled probes and either extracted RNA from influenza A or B positive samples. Cycling was performed with 60 s initial denaturation at 95 °C and 50 cycles of two-step cycling with denaturation for 15 s at 95 °C and combined annealing/extension for 60 s at 60 °C. An additional reverse transcription step was omitted, further promoting time reduction. As shown in Figure 33, detection of both viruses, influenza A and B, was highly specific since only a signal increase in the expected fluorescence channel was observed (detection of influenza A and B in the FAM and HEX channel, respectively).

Furthermore, template dilutions ranging from undiluted to 100-fold diluted extracts resulted in a positive fluorescence signal, although the signal for influenza B detection decreased from higher to lower amounts of RNA (Figure 33).

The observed decrease in total fluorescence for influenza B detection indicates unspecific priming. However, further optimization of primers used for influenza B detection might enhance the reaction specificity. The differentiation between influenza A and B was still highly specific as described above. The experiment demonstrates the potential of RT-KTq 2 for diagnostic applications, as it enables rapid and specific detection of influenza A and B detection in multiplex RT-PCR.

1.3 Discussion

In summary, thermostable DNA polymerases combining reverse transcriptase and PCR activity in one protein scaffold were developed, thus facilitating their use in various applications in molecular biology or clinical diagnostics. The new DNA polymerase variants were discovered by DNA shuffling of a previously reported variant showing some reverse transcriptase activity with a variant that has an enhanced substrate spectrum.

The recombination of both enzymes via DNA shuffling was based on already existing protocols reported in the literature.[137, 139, 181, 182] As additional point mutations were not desired, the protocol was modified in order to reduce the error-rate during recombination. Based on reports,^[139] we employed a DNA polymerase (Phusion^™ DNA polymerase) with high fidelity in the gene preparation, reassembly and post-amplification step. In addition, the purification step after DNase I digestion was omitted, due to the small size of the fragments and resulting difficulties during purification with commercially available kits. An amplification based on undigested parental DNA, resulting from omitting the purification step after digestion as suggested in the literature, was not observed. Rather, a wide range of mutation combinations was found after sequencing various clones of the existing library. A recombination rate was

Figure 33. Detection of influenza viruses A and B from RNA extracts in multiplex RT-PCR. A) Detection of influenza A in the FAM channel. B) Detection of influenza B in the HEX channel. A template dilution series was employed in both cases: undiluted (1), 1:10 (2), 1:100 (3) and no RNA as control (4).

not determined, but a total number of one to four mutations were detected when sequencing various clones from the library. KlenTaq expression, lysis and heat-denaturation of the host proteins as well as the screening system were already established in our group and were adopted.[132, 148] DNA amplification was followed in real time. As a large number of variants showed a reverse transcriptase activity superior to the parental enzymes, we subsequently increased the selection pressure in the screening system, such as in reducing the reverse transcription time or the RNA concentration. Sequencing of the nine most promising variants revealed four different mutation combinations. Although a preceding sequencing of the original library revealed mainly clones with a lower total number of mutations (maximum four), the identified hits comprised at least four mutations, with the exception of one variant.

On the one hand this shows high library coverage of all mutation combinations and on the other hand it suggests a general importance of combining a higher number of mutations for efficient reverse transcriptase activity. The mutations L459M, S515R, I638F and M747K were present in all clones, except RT-KTq 1, indicating a general significance of these mutations.

Additionally, this variant was found in four out of nine clones sequenced. One could argue that this could be due to a bias in the recombination step, but additional mutations as shown in RT-KTq 3 and 4 result in no gain in reverse transcriptase activity, whereas the removal of L459M as depicted in RT-KTq 1 seems to lead to an overall loss in activity; thus supporting the notion that the before mentioned mutation combination is the minimal number of mutations required for the superior reverse transcriptase activity inherent in all variants identified.

Interestingly, the S739G mutation was not present in any of the identified clones. A comparison with KTq M1/M747K, the mutant generated by site-directed mutagenesis and the only enzyme exhibiting all possible mutations, including S739G, showed no change in reverse transcriptase activity. But the absence of this mutation in the identified hits might be due to a loss in thermostability as seen in thermal denaturation experiments conducted with the mutants KTq M1/M747K and RT-KTq 2.

Furthermore, the combination of mutations has synergistic effects in the resulting mutant enzymes compared to the parental enzymes M1 and M747K. The mutations seem to interact with each other not in an additive but rather in a non-additive manner.^[186] Recently, the phenomenon of non-additive interactions between mutations has gained increasing awareness in evolutionary biology.[191, 192] So far, adaptation which underlies C. Darwin’s theory of natural selection was considered as the major determinant for evolution. An organism adapts to the environment and ecology and thus facilitates evolution. The process of adaptation was always thought to depend on the context-independent (additive) genetic effects that exist in a population. Epistasis as context-dependent (non-additive) genetic effects was only recently suggested as a key player in evolution besides adaptation.^[192] Indeed, a statistical analysis of proteins performed by Breen et al. suggests that functional interactions between amino acids are a primary factor of protein-sequence evolution as well.^[191]

Most DNA polymerases evolved towards reverse transcriptase and PCR activity in one enzyme scaffold belong to the family A DNA polymerases.[147-150, 155, 157] Only mutants of Tgo DNA polymerase, belonging to the family B archaeal DNA polymerases, were reported to possess the ability to accept both DNA and RNA as substrate.^[193] Although tremendous efforts have been invested in the evolution of DNA polymerases with increased reverse transcriptase activity, few of them were shown to be applicable in RT-PCR and few are commercially available. The existing one-step RT-PCR kits generally contain an enzyme blend comprising a reverse transcriptase and a DNA polymerase. Thus, the properties of both enzymes have to be considered when optimizing analysis of new RNA targets for clinical diagnostics. But few enzymes are presently on the market, which are capable of both reverse transcription and subsequent amplification. Recently, Moser et al. published a thermostable DNA polymerase isolated from a viral metagenome library exhibiting innate reverse transcriptase activity;

commercially available under the brand name PyroPhage.^[151] Alternatively, the use of Tth polymerase, isolated from the thermophilic eubacterium Thermus thermophilus,^[147] in one-step RT-PCR can be considered, although several drawbacks, such as lower sensitivity compared to two-enzyme mixtures, arise from the fact that Mn²⁺ is required as divalent cation.^[142]

The identified enzymes represent a promising alternative to two-enzyme mixtures by minimizing work and time costs, avoiding time consuming optimization and inhibitory effects present between a reverse transcriptase and a PCR-competent DNA polymerase.^[152-154]

Particularly, in certain important fields of clinical diagnostics such as point of care testing minimized time consumption and reliable detection is highly important. In addition, performing the reverse transcription step at high temperatures prevents unspecific priming and biased RNA detection based on secondary structure formation of specific RNA molecules.^[142] In summary, the variants exhibit very promising properties, which allow applications such as fast detection and quantification of RNA in RT-PCR, thus facilitating their use in transcriptome analysis, pathogen as well as disease-specific marker detection.

2. Crystallization Studies with RT-KTq 2