WO2004097046A1

WO2004097046A1 - Stable internal standard for pcr and rt- pcr reactions

Info

Publication number: WO2004097046A1
Application number: PCT/GB2004/001921
Authority: WO
Inventors: Kate Dingle
Original assignee: Isis Innovation Limited
Priority date: 2003-05-01
Filing date: 2004-04-30
Publication date: 2004-11-11

Abstract

The present invention relates to use of a nucleic acid molecule having intra-molecular base pairing as an internal control in nucleic acid amplification. The invention relates to a method of nucleic acid amplification comprising having components which amplify target nucleic acid if it is present in a sample and comprising a nucleic acid molecule having intra-molecular base pairing as an internal control. The invention also relates to a kit for use in the method.

Description

STABLE INTERNAL STANDARD FOR PCR AND RT-PCR REACTIONS

The present invention relates to use of a nucleic acid molecule having intra-molecular base pairing as an internal control in nucleic acid amplification technologies, and kits for such use.

BACKGROUND

Over the last decade, PCR has promised the rapid, sensitive and accurate detection of many pathogens, e.g. viruses, even when present in low numbers. The promptness and accuracy of this approach has been shown to be clinically useful in both making a positive diagnosis and ruling out infection. It also allows appropriate treatment to be implemented and costly interventions to be avoided.

Commercially produced RT-PCR diagnostic kits are available for relatively few pathogens, hence many "in house" assays have been developed for additional targets, using both gel-based and real time detection systems. Such assays are economical, but generally omit appropriate internal, external and negative controls e.g. "reaction specific" internal controls to monitor the extraction, reverse transcription, amplification and detection steps. This is due to the technical difficulty of designing and implementing robust internal controls. Such controls allow negative results due to inhibition or human error to be distinguished from true negatives, increasing confidence in the technique. These controls also ensure that the tests are reliable and reproducible and that technical failures or inhibition are identified.

Few clinical diagnostic laboratories routinely control for inhibition or technical failure of in house RT-PCR assays. Without such internal tube-specific control, false negative and genuine negative results are indistinguishable, and both laboratory workers and clinicians may lack confidence in the technique. An ideal internal RNA control for RT-PCR would be (i) non-competitive with the target virus amplicon, (ii) economical and easily produced and standardised, (iii) stable on storage and use, being resistant to RNAses and freeze thawing, (iv) co-extracted with the clinical sample, (v) suitable for use in many different assays, (vi) non-infectious, (vii) economical and (viii) its sequence should be absent from the clinical samples in which it is used. The results of diagnostic assays including the control should also be simple to interpret. Few (if any) internal controls described to date have all of these features.

Many internal DNA controls for PCR have been described which have been either assembled from short artificial DNA fragments or consist of a cloned target amplicon modified by insertions, deletions or 5' and 3' extensions. Controls of this nature are differentiated from the target amplicon on the basis of size. However, when the same primer pair is used to detect both the control and the target, the possibility of competition arises. Also, if used in assays to detect RNA, DNA controls do not confirm the reverse transcription step was performed efficiently.

Several groups have designed competitive DNA and RNA internal controls, using the same primer pair to detect both the control and the target microbe. The two amplicons were then differentiated by size or using heterologous probes. A method for the production of competitive RNA controls has been described which proved useful for the identification of RT-PCR inhibitors. However, if the same primers are used to detect both the control and the target amplicon, the overall detection limit of the assay may be compromised, especially if the target microbe is present at low levels. Competitive controls are incompatible with multiplex PCR in which several primer pairs are required. Multiplex PCR is attractive for molecular diagnostics since multiple pathogens producing similar symptoms can be screened simultaneously in a single reaction. An internal control of a universally expressed cellular RNA contained within the clinical specimen has been considered, for example the housekeeping genes such as β-actin or 18 S ribosomal RNA. However, even if it is expressed at a constant level, the RNA concentration may vary widely among samples depending on the number of cells in the specimen, and some RNA degradation during transport and storage is inevitable. In addition, contaminating cellular DNA may be difficult to remove if a control for the reverse transcription step is required. Such a control may be provided by RNA transcribed in vitro, using plasmid DNA as the template. For example, an internal control RNA encoding a tomato transcription factor has been described. A criticism of in vitro transcribed RNA controls is their sensitivity to nuclease degradation and the reliability of plasmid template DNA removal.

Non-competitive internal controls have also been described, with separate primer pairs to detect the control and the microbe(s) of interest. Such controls often comprise endogenous RNA or DNA contained in the sample, for example β-actin or 18 S ribosomal RNA. Unfortunately, the concentrations of these RNAs vary widely among clinical samples.

Alternatively, a clinical sample may be spiked with a known amount of an animal virus such as non-human seal herpes virus type 1 (DNA) or phocine distemper virus (RNA). The clinical specimen was spiked with a low and fixed amount of virus to control the extraction, amplification and detection steps. A "tube-specific" internal control was lacking in the case of RNA viruses as a second reaction was performed to detect the control virus. Although attractive, the use of live viruses as internal controls may raise issues of safety and consistency between preparations.

Positive controls for various RT-PCR assays have been developed commercially ("Armored RNA" Ambion [Europe Ltd], Cambridgeshire, UK), consisting of an RNA fragment assembled into phage-like particles. A commercially available positive control for diagnostic RT-PCRs has been developed, named "armored RNA". It consists of an RNA fragment with the same sequence as a target virus which is packaged into phage- like particles by complexmg with MS2 bacteriophage coat protein (I). This coat provides protection from RNAse degradation. Armored RNA is an effective overall extraction, amplification and detection control for a specific assay, but not an "in-tube" reaction- specific internal control. It has also been adapted to provide such an internal control for real time PCR by providing an RNA distinct from the target, allowing the use of heterologous probes. The major disadvantages are the expense and a lack of flexibility. This technology has also been adapted to provide an internal control for a real time PCR assay. Their cost and a lack of versatility have so far prevented the widespread adoption of Armored RNA controls.

A significant number of the many RT-PCR assays described to date are for respiratory viruses. These pathogens cause a considerable burden of illness, particularly during winter. Since different viruses often cause similar symptoms, appropriate patient management can be assisted by rapid laboratory tests with a low limit of detection. Respiratory viruses have been traditionally detected by cell culture and more recently using commercially available direct immuno-fluorescence tests. RT-PCR is faster and more sensitive than cell culture, and offers greater sensitivity than immunofluorescence.

INVENTION

In a first aspect, the present invention provides a use of a nucleic acid molecule having intra-molecular base pairing as an internal control in nucleic acid amplification.

The nucleic acid may be DNA or RNA. Also included within the invention are chemically modified nucleic acids such as RNA incorporating 2' modified ribonucleoside triphosphates or others as known to the person skilled in the art which will have greater stability than naturally occurring nucleic acids.

The RNA may be rod shaped or any other shape which an RNA molecule may form when there is a high level of intra-molecular base pairing. The RNA may have chemical moieties attached to it to increase its stability such as replacement of the 2' OH by fluorine, e.g. RNA containing 2'-fluoro-2' deoxyuridine triphosphate or 2'-fluoro-2' deoxycytidine triphosphate or others as are known to the person skilled in the art as suitable for stabilising nucleic acids.

Alternatively or in addition, the nucleic acid molecule may be self-stabilising due to complementary base pairing. The complementary base pairing may be GC intramolecular base pairing or any other form of intra-molecular base pairing as known to the person skilled in the art.

The intra-molecular base pairing may be greater than 50%, preferably greater than 60%, or may even be greater than 70%. The GC intra-molecular base pairing may have one of these numerical values. The molecule may have a secondary structure with around 60% - 80% intra-molecular base pairing and around 50% - 70%GC content, optionally around 65% - 75% intra-molecular base pairing and around 55% - 65% GC content.

Preferably the nucleic acid is a virus genome or a modified virus genome. More preferably the nucleic acid is a modified hepatitis delta virus genome (HDV). Alternatively, the nucleic acid may be a man-made e.g. sequence synthesised on a nucleic acid synthesiser or by other chemical synthesis method. The nucleic acid maybe 500 nucleotides in length, preferably 700 nucleotides in length, more preferably above 1000 nucleotides in length. If the nucleic acid is a modified hepatitis delta virus the genome preferably comprises 1.2 copies of the virus genome with the ORF encoding the delta antigen inactivated by a 2 nucleic deletion such as is encoded by the pTW107 plasmid.

Preferably a T7 promoter is used to transcribe the nucleic acid from the pTW107 plasmid. Other promoters suitable for in vitro transcription may also be used, such as SP6 and these are known to the person skilled in the art.

Preferably the viral genome is unable to replicate in the absence of a helper virus, preferably in the case of the HDN virus this is the hepatitis B virus.

The RΝA control (e.g. in vitro transcribed) may be stored in any suitable storage solution e.g. such as Ambion's 'RΝA storage solution' sodium citrate or other biochemical buffers or in water. The nucleic acid solution may be freeze thawed at least twice without the nucleic acid showing a decrease in its concentration as detected by RT-PCR. The integrity of the RΝA can be confirmed by electrophoresis through a 1% formaldehyde MOPS gel. The modified HDN in plasmid pTW107 is 1864 nucleotides in length and has a secondary structure predicted to contain 70% intra-molecular base pairing and has a 60% GC content.

Key features of the nucleic acid are that the control is GC rich and has a complex secondary structure making it a difficult template for RT-PCR; the control is sufficiently large to allow the design of smaller target virus PCR amplicons. The GC richness and secondary structure give the modified HDN genome an unusually high level of stability during freeze/thawing and increased resistance to nucleases.

Its high level of secondary structure also ensures that it is a poor template for reverse transcription and its cDΝA is a relatively inefficient template for PCR. This ensures that if cDNA is synthesised successfully from such a difficult template cDNA will also be synthesised from any target virus present in the clinical sample.

The control would also be predicted to be a poor PCR competitor with the target virus amplicons. The amplification of target viruses would also be favoured over the amplification of the internal control. This can be enhanced by supplying PCR primers for the control at lower than the concentration used for the target virus PCR primers (e.g. around 50% concentration). Other highly structured viral and subviral genomes are suitable for this use and include viroids, virusoids and viroid-like satellite RNA's, such as potato spindle tuber viroid, chrysanthemum stunt viroid or others known to the person skilled in the art. Their small genomes (300 -400 nucleotides) may make them especially useful for real time DNA amplification analyses. Preferably the nucleic acid amplification technique is a polymerase chain reaction (PCR) based method or other methods as known to the person skilled in the art. Alternatively the nucleic acid amplification may be reverse transcription polymerase chain reaction (RT-PCR) e.g. with gel-based detection. The nucleic acid may be used in real time analysis or as a quantitative control, in ELIS A-based systems or in hybridisation probe detection. Such detection includes gel based detection and fluorescent probe detection but its not limited to such methods of detection.

Preferably the nucleic acid amplification is performed for clinical diagnosis of viral infection such as e.g. respiratory viral infections including those viruses described in the examples. Alternatively the amplification may be performed for bacterial infections, for mycoplasma or other microbiological agents which are present at a low level within a sample. The sample may comprise eukaryotic or prokaryotic organisms. Preferably the subject from which a sample is taken is mammalian, more preferably a human, cat, dog (particularly the domestic cat or domestic dog), horse (equine), cow (bovine), pig (porcine) or fish. In a second aspect the present invention provides a kit comprising the nucleic acid molecule as described for use according to the first aspect and instructions for using the nucleic acid molecule as an internal control in a nucleic acid amplification assay. Preferred features of the nucleic acid molecule are as described for in the first aspect. The kit may also include primers for the control and/or for a molecule to be detected from a sample as well as other standard kit components, such as buffers, dyes etc.

In a third aspect there is provided a method of nucleic acid amplification comprising providing components which amplify target nucleic acid if present in a sample and further comprising a nucleic acid molecule having intra-molecular base pairing as an internal control for the assay. Preferably the nucleic acid molecule is as described for use according to the first aspect of this invention. All preferred features of the nucleic acid molecule are as described for use according to the first aspect of the invention and apply to the nucleic acid molecule for the method of the third aspect in the invention. The component for the method may have been provided by the kit according to the second aspect of the invention.

In a fourth aspect there is provided a modified hepatitis delta virus genome such as is produced by the plasmid TW107 as described in the present application figure 5A or 5B.

The fourth aspect of the invention may be used according to the first aspect, as part of the kit according to the second aspect or the nucleic acid molecule in the method of the third aspect.

BRIEF DESCRIPTION OF FIGURES AND TABLES

FIG. 1. Shows the modified HDN genome used as an internal control. (A) Representation of the predicted rod like secondary structure of the HDN genome and (B) a linear representation of the sequence used for in vitro transcription of the control RΝA (1.2 copies of genome), and the PCR amplicons used as an internal control. The sequences of the vector pTW107, including the T7 promoter and Hind III site used to linearise the vector are shown in bold.

FIG. 2. Shows the detection of in vitro transcribed HDN control RΝA by RT-PCR and confirmation that the DΝA plasmid template was efficiently removed. (A) Lanes 1 to 10; RT-PCR amplification products from a ten fold serial dilution of in vitro transcribed HDV RΝA. The dilution series extended from 0.1 μg RΝA per RT-PCR in lane 1 to O.lfg in lane 10. The amplification primers were delta 1 and delta 2, giving a product of 761bp. Lanes 11 to 20; as lanes 1 to 10, but without reverse transcription of the template RΝA, to detect any contaminating plasmid template DΝA. Lane 21, negative control. L:lkb ladder (Biorad), sizes indicated (bp). lOμl of each 50μl PCR was analysed.

(B) As A, but showing the amplification products of a nested PCR performed using the templates from A and primers delta 3 and delta 4 to obtain a product of 506bp. The detection limit is shown in lane 8 at lOfg RΝA, corresponding to approximately 10,800 molecules of RΝA added to the extraction.

FIG. 3. Shows an overview of the internally controlled nested RT-PCR assay for the detection of nine respiratory viruses. The flow chart illustrates the use of the modified hepatitis delta virus internal control RΝA in three multiplex RT-PCRs for the detection of nine clinically important respiratory viruses.

FIG. 4. Shows the HDV internal control did not compromise (i.e. decrease the sensitivity of) the assay through competition with the RΝA of the target virus. Nine plasmids were constructed containing each of the first round PCR amplification products. A dilution series was made containing 10 to 10^" molecules of each plasmid per μl and lμl of each dilution was amplified in each nested multiplex PCR. (A) RSV A, RSV B and human metpneumo virus, (B) influenza virus A and B, (C) PIV types 1-4. To determine whether the HDV internal control compromised the limit of detection (i.e. inhibited the sensitivity of target virus detection), the PCR also contained the HDV internal control cDNA at the same concentration used in the diagnostic assay. The virus amplicons are indicated by an arrow head and the HDV control is indicated by *. As the concentration of plasmid containing target virus sequences was increased, the yield of the control amplicon decreased or it disappeared completely. For comparison, a PCR containing HDV cDNA alone was performed and is indicated by δ. The negative controls lacking target DNA are indicated by Neg.

Note: no difference in the limit of detection for each assay was observed whether or not the HDV internal control was present.

FIG. 5. Shows the results of screening of 20 respiratory specimens for nine target viruses by internally controlled nested multiplex RT-PCRs. Lanes 1 to 20 refer to all three gels shown and each lane corresponds to a single specimen. (A) Assay for PIV types 1 to 4, (positive detection is indicated by PI, P2, P3 or P4). (B) Assay for influenza virus A and B (positive detection indicated by FA or FB). (C) Assay for RSV A, RSV B and human metapneumovirus (positive detection indicated by RA, RB or M). A dual infection is shown in lane 5. * Indicates the internal control which was detected when target viruses were not detected confirming the assay has performed correctly and the PCR inhibitors are not present in the sample. L, lkb DNA ladder (Biorad); +ve, co- amplified positive control plasmid DNA for each of the target viruses, -ve, negative control for contamination: no internal control or target virus detected.

Lane 1 : Influenza A detected.

Lane 2: Human metapneumovirus detected.

Lane 3 : RSV B detected.

Lane 4: No target viruses detected. Lane 5: Dual infection: Parainfluenza virus 4 and RSV A detected. Lane 6: Parainfluenza virus 2 detected. Lane 7: No target viruses detected. Lane 8: Influenza B detected. Lane 9: No target viruses detected. Lane 10 RSV A detected. Lane 11 RSV B detected. Lane 12 Parainfluenza virus 3 detected. Lane 13 No target viruses detected. Lane 14: Human metapneumovirus detected. Lane 15 Influenza B detected. Lane 16 Parainfluenza virus 1 detected. Lane 17 RSV A detected. Lane 18 Influenza A detected. Lane 19 RSV B detected. Lane 20 No target viruses detected.

Table 1. PCR primers used in the three internally controlled, nested multiplex RT-PCR assays to detect nine respiratory viruses. The primers Delta-1 to 4 were used to detect the internal control.

All primers are shown in the 5' to 3' direction. Primers 1 and 2 were used in the first amplification reaction and primers 3 and 4 in the second, nested amplification. Forward primers are indicated by the odd numbers and reverse primers by even numbers.

Nucleotide positions where degenerate sequence was used are indicated in brackets. The parainfluenza primers for types 1, 2 and 4 were either identical to or slightly modified from Aguilar et al, 2000. * Paral-1; as Pipl+, with a 3' extension. Paral-3; as PipSl+. Paral-4; as PiSl- with modified 5' sequence. As Aguilar et al., 2000 Journal of Clinical Microbiology 38 1191-1195. All other primers are unique to this study. Other primers modified slightly from this study include; PIVl-1 = as Pipl+ with a 3' extension, PrVI-4 = as PiSl with a modified 5' sequence. The primers Delta-1 to 4 were used to detect the internal control.

Table 2

Virus genes targeted in nested multiplex RT-PCRs to detect nine respiratory viruses. The three assays are designated "Influenza Assay", "RSV and HuMV Assay" and "PIV Assay". The sizes of PCR amplicons (bp) obtained in the first and second round (nested)

PCRs are indicated. The control amplicon was designed to be larger than all the virus amplicons in both rounds of the nested PCR enhancing the amplification of the smaller target virus amplicon (if present) over the internal control. Sizes of the PCR amplicons in the three internally controlled nested multiplex RT-PCR assays.

Table 3. PCR Primer Verification

A total of 191 known positive samples containing virus which was up to 20 years old, were tested to confirm that the PCR primer sequences are genetically conserved among isolates and stable over time. The samples comprised NPAs (confirmed as positive by immunofluorescence, cell culture or an alternative PCR assay) and cell culture supernatants archived at -80 °C for up to 20 years.

Table 4. Application of the assay to routine clinical testing. A total of 324 clinical samples were tested: NPA, Nasopharyngeal aspirate; TS, throat swab; BAL, broncho alveolar lavage; NS, nasal swab; ETT, endotracheal tip; LB; lung biopsy. The present invention is now described with reference to the following non-limiting examples:

EXAMPLE 1

Construction of plasmid pT W107. The plasmid pTW107 was constructed in the vector pGem3z (Promega UK, Southampton, UK). It contained 1.2 copies of a modified hepatitis delta virus (HDV) genome flanked by the T7 and SP6 promoters within the vector. The HDV sequence extended from the SaE restriction site at nucleotide positions 962-967, through the remaining full length of the genome, ending at the Xbaϊ restriction site (nucleotide positions 781-786) of the second genome (Fig. 1). The construct contained a 2 nucleotide deletion at positions 1435-1436 (near a unique Eco RI restriction site) within the open reading frame encoding the delta antigen, rendering the RNA incapable of autonomous replication. It is also unable to replicate in the absence of the helper virus hepatitis B virus and this was not present, and at no time was the HDV RNA introduced into eukaryotic cells.

In vitro transcription of modified HDV RNA and removal of DNA template. pTW107 was linearised using the restriction enzyme Hind HI, the reaction was terminated by the addition of 0.5M EDTA and the plasmid DNA was precipitated using 1/10 volumes 3M sodium acetate and 2 volumes of absolute ethanol. HDV RNA (genomic sense) was transcribed in vitro from the T7 promoter using 1 μg of linearised plasmid DNA and a commercially available kit according to the manufacturer's instructions (Megascript T7 transcription kit; Ambion [Europe Ltd], Cambridgeshire, UK). The template plasmid DNA was removed by incubation with DNAse 1 for 15 minutes (Ambion [Europe Ltd], Cambridgeshire, UK) at 37 °C for 15 minutes. Further precautions were taken to remove the template DNA by extraction using Tri reagent LS (or Trizol LS reagent)(Sigma Aldrich Company Ltd, Dorset, UK) according to the manufacturer's instructions. The resultant HDV RNA pellet was resuspended in RNA storage solution containing sodium citrate (Ambion [Europe Ltd], Cambridgeshire, UK) to inhibit RNA hydrolysis, dispensed into aliquots and stored at -80°. Its integrity was confirmed by electrophoresis through a 1% formaldehyde MOPS gel, containing ethidium bromide which revealed a single band of RNA. The total yield of RNA from a single in vitro transcription reaction was around lOOμg, as expected from the manufacturer's information. Aliquots of RNA for single use were stored at -80°C.

Application of internal control to three multiplex nested RT-PCR assays for nine respiratory viruses, (i) Co-extraction of internal control and RNA from clinical samples. Clinical samples including nasopharyngeal aspirates (NPA), throat and nasal swabs (TS, NS), bronochoalveolar lavage (BAL), samples from endotracheal tips (ETT) and lung biopsies (LB) were diluted by the addition of l-5ml virus transport medium (depending on specimen volume). Cellular debris were removed by centrifugation at

1,000 rpm for 5 minutes. Total RNA was extracted from the supernatant using a silica column based kit (Viral RNA mini kit, Qiagen Ltd., Crawley, UK) according to the manufacturer's instructions, with the following modifications. Prior to RNA extraction, 1 μl 100 fg/μl modified HDV RNA (internal control) was added to 280 μl AVL extraction buffer (Qiagen Ltd., Crawley, UK). This was followed by 69 μl respiratory sample supernatant. The RNA was eluted from the column in 40μl AVE buffer (Qiagen Ltd., Crawley, UK), (ii) cDNA synthesis. Reverse transcription was performed in a 20 μl reaction containing 10.5 μl RNA and 0.5 μl (250 ng) random hexamers (Promega UK, Southampton, UK). This was denatured at 70 °C for 10 minutes followed by immediate transfer to ice on which the remainder of the reaction was assembled: 4 μl first strand buffer (Invitrogen Ltd., Paisley, UK), 2 μl 0.1M dithiothreitol (Invitrogen Ltd., Paisley, UK), lμl dNTPs (lOmM) (Invitrogen Ltd., Paisley, UK), 1 μl RNAsin (Promega UK, Southampton, UK) and 1 μl M-MLV reverse transcriptase (Invitrogen Ltd., Paisley, UK). Incubation was at 37 °C for 1 hour, followed by inactivation at 70 °C for 10 minutes, (iii) PCR amplification. The oligonucleotide primers used are shown in Table 1. For each first round amplification a 50μl PCR contained 5μl cDNA, 5 μl lOx hotstarTaq DNA polymerase reaction buffer (Qiagen Ltd., Crawley, UK) containing 15 mMMgCl₂, lμl oligonucleotide primer mix (containing each target virus primer at 10 μM [as in Table 2] and the internal control primers at 5 μM), 1 μl lOmM dNTPs, and 0.25 μl (1.25 units) HotstarTaq DNA polymerase (Qiagen Ltd., Crawley, UK). The nested second round amplification reaction (50μl) included 1 μl of the first PCR as template and an additional 3 μl 25mM MgCl₂ (4 mM final concentration). The following amplification conditions were used; denaturation and enzyme activation at 95 °C for 15 minutes, followed by 40 cycles of denaturation at 94 °C for 20 seconds, annealing at 55 °C for 20 seconds and extension at 72 °C for 45 seconds in the first PCR or 30 seconds in the nested PCR, followed by a final extension at 72 °C for 5 minutes. The reaction products were analysed by 2 % agarose gel electrophoresis.

Preparation of plasmid DNA for establishing assay detection limits and use as positive controls. PCR amplicons generated during the first amplification round of each target virus were ligated directly to the vector pGem T easy (Promega UK, Southampton, UK) and electrocompetent E. coli were transformed. Colonies containing plasmid DNA with the required insert were identified by PCR and DNA was prepared by miniprep (Spin miniprep kit, Qiagen Ltd., Crawley, UK). The OD₂₆o of each plasmid solution was determined and, using Avogadro's number, stock solutions were prepared containing 1 x 10¹⁰ molecules plasmid DNA per μl. A ten fold dilution series was made for each plasmid from 1 x 10¹⁰ to 1 x 10^"2 molecules per μl and was used to determine the detection limit of the PCRs. These plasmids were also used to prepare positive control mixtures, one for each of the three multiplex assays. The positive control plasmids were co-amplified, providing a single tube positive control for each of the three assays. As a precaution against contamination, the positive control plasmids were prepared in a separate laboratory on a different floor to the laboratory where the diagnostic RT-PCRs were carried out. RESULTS

Clinical diagnostic tests based on nucleic acid amplification offer the potential for the rapid and sensitive detection of viral infections. However, the design of appropriate controls for such assays has proven difficult. Here, we describe the use of a rod-like RNA molecule, self-stabilized by secondary structure involving 70% intra-molecular base pairing, as an internal control for diagnostic RT-PCR. The secondary structure of the control RNA confers both stability and resistance to nucleases and inhibits competition of the control with the target amplicon. We have implemented the control in an assay to detect nine clinically important respiratory viruses.

The RNA control molecule consists of a modified hepatitis delta virus (HDV) genome which was transcribed in vitro from a plasmid template. The template DNA was efficiently removed by treatment with Trizol LS reagent (Sigma) and DNAse (Invitrogen); the DNA was not detectable in the RNA control (by RT-PCR) at the concentration used in diagnostic assays, or at concentrations lxlO⁷ fold higher. A known quantity of the RNA molecule was 'spiked' into every RNA extraction, controlling individual clinical samples at each step of the process. Detection of the control at the end of the assay confirmed it was performed correctly, and that the specimen lacked inhibitors. The assay was performed on 324 respiratory specimens in a routine diagnostic laboratory.

Three features of the RNA control and assay design enhanced the preferential amplification of the target virus RNA over the control, (i) The control is GC rich, making it a difficult template for RT-PCR. (ii) The control is sufficiently large to allow the design of smaller target virus PCR amplicons. (iii) The PCR primers for the control were present at half the concentration of the primers for the target virus. In our experience, if the target viral RNA was present, the control RNA was either not detected or reduced in intensity. Conversely, if the target virus was absent, the control molecule was detected. This stable RNA molecule could be used to control any reverse transcriptase or nucleic acid amplification based diagnostic test. The aims of the present study were two fold. Firstly, to develop a universal internal control for diagnostic RT-PCRs meeting the criteria described above and secondly to apply this control in molecular diagnostic assays for the routine detection of nine clinically significant respiratory viruses. The RNA molecule chosen was a modified genome of hepatitis delta virus (HDV). The genome is 1679 nucleotides in length (1.2 copies used = 1864 nucleotides), and unusual in having an extremely high level of intramolecular base pairing, such that 70% of the molecule effectively consists of double stranded RNA. HDV RNA is also GC rich at 60%. These characteristics give it an unusually high level of stability during freeze thawing and resistance to nucleases. Its high level of secondary structure ensures that it is a poor template for reverse transcription, and its cDNA a relatively inefficient template for PCR. If cDNA is synthesized successfully from such a difficult template, cDNA would also be synthesised from any target virus present in the clinical sample. The control cDNA would also be predicted to be a poor PCR competitor with the target virus amplicons. The amplification of the target viruses would therefore always be favoured over the amplification of the internal control.

We describe three internal and externally controlled nested multiplex RT-PCR assays for the routine detection of nine respiratory viruses clinical specimens. The modified HDV RNA was used as an internal control in all three assays for the detection of (i) influenza A and B, (ii) parainfluenza 1, 2, 3 and 4, and (iii) respiratory syncytial virus (RSV) A and B and human metapneumovirus. Choice of RNA internal control molecule

The HDV genome was identified as an unusual RNA molecule which could be used as a reaction-specific internal control in RT-PCR assays. The HDV genome is 1679 nucleotides in length, and its GC rich RNA (60%) forms a rod-like structure with 70% intra-molecular base pairing. This unusual secondary structure was central to its choice as an internal control. It enhances stability during freeze thawing, increases resistance to nucleases and renders the molecule a difficult template for RT-PCR. The latter was intended to ensure preferential amplification of the target virus amplicons, thus maintaining a low limit of detection (1 to 10 target molecules) for the viruses of interest.

In vitro transcription of HDV internal control RNA

Plasmid pTW107 was constructed containing 1.2 copies of a modified HDV genome, with a 2 nucleotide deletion in the open reading frame encoding the delta antigen (Fig. 1). This safety consideration rendered the RNA transcribed from the plasmid incapable of replication. After transcription, the integrity of the RNA was confirmed by gel electrophoresis and it was standardised using a spectrophotometer. The number of molecules present were calculated using Avogadro's number. RNA hydrolysis was minimised by storage at -80 °C in sodium citrate solution (RNA storage solution, Ambion [Europe Ltd], Cambridgeshire, UK).

To confiπn that the control RNA was free of contaminating plasmid DNA, a ten fold dilution series of the RNA was made from 0.1 μg to O.lfg per μl. Each aliquot was extracted as a "clinical sample" and the RNA was divided into two. Half was used for cDNA synthesis and the other was stored on ice prior to amplification. All aliquots were then amplified by nested PCR using primers delta-1 and delta-2 in the first round (Fig.

2A) and delta-3 and delta-4 in the second (Fig. 2B). These primers were designed from one side of the predicted rod-like structure of the HDV genome (Fig. IB, Table 1). The detection limit after the first PCR amplification corresponded to the 10 pg RNA dilution, and after the nested PCR, to 10 fg. The latter corresponds to approximately 10,800 RNA molecules. No residual plasmid DNA was detected under these conditions in the RNA lacking reverse transcription.

When Q solution (Qiagen Ltd., Crawley, UK) (a PCR additive which relaxes secondary structure) was included in the PCR, the detection of limit of the HDV cDNA increased 100 fold. To ensure that the amplification of the internal control remained sub-optimal compared to the target viruses, Q solution was not used in the PCR assays. As 100 fg RNA was detected reproducibly in nested PCR, this concentration was used as an internal control in routine RT-PCR assays.

Internally controlled assay for nine respiratory viruses

The nine viruses were chosen on the basis of clinical need and potential benefits for patient management. They were influenza A, influenza B, respiratory syncitial virus (RSV) A, RSV B, human metapneumovirus and parainfluenza virus (PIV) 1 to 4. Although RT-PCR assays have been previously described for these viruses, the assays were incompatible with the internal control due to amplicon sizes and/or oligonucleotide primer design. Furthermore, όur aim was to detect all nine viruses in three nested multiplex RT-PCRs (Fig. 3) using one cDNA synthesis reaction followed by identical thermocycling conditions. This would simplify the routine use of the assay. New multiplex assays were therefore designed, incorporating the internal control.

The viral genes chosen as targets were selected on the basis that (i) they included highly conserved nucleotide sequences, (ii) sequences of a large number of representative strains were available in public databases (Genbank and the Los Alamos Influenza database rhttp://www.flu.lanl.gov1) and (iii) they had been amplification targets in previously published assays (Table 2). New primers were designed for six of the nine target viruses with the exception of PrV-2 and PIV-4 which were published previously and the primers for PIV-1 which were modified slightly from this study. The primers for the internal control were designed to amplify a DNA fragment larger than all the target virus amplicons, to enhance the preferential amplification of the latter.

Assay verification: Amplification of known positive samples and limit of detection Firstly, the nested PCR primers for the detection of each target respiratory virus were confirmed using cDNA synthesised from an infected cell culture supernatant. Amplicons were produced using the first and nested primer pairs in separate single round PCRs, to confirm successful amplification by both primer pairs for each virus. Cell culture supernatant was not available for PIV-4, but these primers were validated previously and amplicons later obtained from clinical samples for PIN-4 were confirmed by nucleotide sequencing.

Secondly, the sequences chosen for primer design were confirmed to be conserved among isolates and genetically stable over time. A total of 191 isolates obtained up to 20 years ago were tested (Table 3). These isolates were contained in either known positive clinical samples (identified previously by direct immunofluorescence, cell culture or an alternative PCR assay) or in archived cell culture supematants. A single contradictory result was obtained; a ΝPA which was previously positive for influenza A by cell culture was negative by PCR. One avian influenza A virus (H7Ν7) isolated from a human conjunctival swab was detected by the influenza A RT-PCR from archived, frozen cell culture supernatant.

To determine whether the internal control compromised the detection limit of the PCRs, reactions were performed both with and without the control. The template was a 10 fold dilution series of positive control plasmid DNA, both with (Fig.4) and without the internal control HDN cDΝA. No difference in the detection limit of the PCRs was observed whether or not the control was present. As the number of target plasmid molecules increased, the control amplicon decreased or disappeared. The detection limits were 1 molecule of plasmid DNA per 50 μl PCR for influenza A and B, and 10 molecules for RSN A, RSN B, huMV and PIV-1 to 4.

Routine detection of respiratory viruses in clinical samples by internally controlled RT-PCR. A total of 324 respiratory samples taken between 29/10/2002 and 14/4/2003 were available for testing. The samples consisted of nasopharyngeal aspirates (n = 231), throat swabs (n = 33), bronchoalveolar lavage (n = 17), nasal swabs (n = 12), ETT (n = 11), lung biopsies (n =8), and others (n = 12). RΝA from at least one target virus was detected in 150 samples (46.3%) and the RT-PCR was inhibited in only two (Table 4). Each of nine target viruses were identified at least once (Fig. 5 and Table 4). RΝA from two different viruses was detected in 3 specimens (0.1%), indicating co-infections of PIV-4 and RSV A (Fig. 5 lane 5), PTV-3 and RSVB, and PIV-4 and RSVB. When RΝA from the target virus was detected, the internal control RΝA was not detected (Fig. 5). When a target virus was not detected, the internal control provided confirmation that the assay was performed correctly and the sample lacked inhibitors.

DISCUSSION

The HDV genome was used successfully as a reaction specific internal control in routine clinical diagnostic RT-PCRs. Its extensive secondary structure has a self-stabilising effect which together with its high GC content makes it a difficult template for reverse transcription and amplification. HDV RNA is also known to have enhanced resistance to nucleases. As a consequence, the RNA was sufficiently stable for routine use and did not compromise the detection limit of the three nested multiplex RT-PCRs in which it was used. The control RNA was modified by the deletion of two nucleotides within the ORF encoding the delta antigen, a safety consideration designed to eliminate the ability to replicate autonomously. The internal control was used in three nested multiplex RT- PCRs for the routine detection of nine respiratory viruses in clinical samples. These three assays were designed to be performed under identical thermocycling conditions using cDNA from a single randomly primed synthesis reaction (Fig. 3). The advantages of this approach to the diagnostic laboratory are improved interpretation of assays due to the internal control together with savings in cost and time.

The control RNA was simple to produce and standardised by spectrophotometry. High yields of the control RNA were easily transcribed in vitro and contaminating plasmid template DNA was not detectable at the concentrations tested in nest RT-PCRs. The RNA was stored at -80 °C in concentrated stocks and single use aliquots, and it proved sufficiently stable during storage for routine use (at least 8 months at -80 °C was tested and two freeze-thaw cycles). The RNA was co-extracted with each clinical sample, providing a control for the extraction, reverse transcription, PCR and detection steps of the assay. The modified HDV sequence may provide a suitable control for other assays based on nucleic acid amplification. Our data suggest that HDV cDNA (plasmid or linear HDV sequence) may also be a suitable control for assays to detect pathogens with DNA genomes, and that it could also be adapted for use in real time PCR.

The HDV genome sequence was absent from the respiratory specimens tested using the assay. However, the control sequence may be present very rarely in certain samples such as blood from patients with hepatitis B virus infection who are co-infected with HDV. The structural characteristics which make HDV an attractive control are shared by its relatives, the plant viroids which may provide an alternative to HDV. The viroids have much smaller genomes than HDV (246 to 401 nucleotides compared to 1,700 nucleotides) and could be particularly useful controls for assays based on real time PCR.

The control RNA was modified by the deletion of two nucleotides within the open reading frame encoding the delta antigen. This renders the HDV RNA incapable of replication, which also requires the presence of hepatitis B virus. The control was therefore suitable for routine use, since it is non-infectious. Infectivity may be a possible disadvantage for the routine use of animal viruses as internal controls, and many laboratories lack the facilities for their production. However, animal viruses have the advantage of an intact capsid, providing an authentic control for the protein disruption phase of the extraction step. A protein coat could be added to modified HDV RNA, possibly using phage proteins as in Armored RNA (Ambion [Europe Ltd], Cambridgeshire, UK). This would combine the structural advantages of the HDV RNA, with the capsid advantages of the animal viruses.

The assay was designed so that the presence of the internal control did not detectably compromise the detection limit of the RT-PCRs (Fig. 4). When the target viruses were detected in clinical samples, the control amplicon disappeared or was reduced in intensity

(Figs. 4 and 5). This may be due to the HDV rod-like secondary structure and 60% GC content, which make it a sub-optimal template for RT-PCR. The difficult nature of the control sequence for amplification was illustrated experimentally using Q solution (Qiagen Ltd., Crawley, UK) to reduce template secondary structure. In the presence of Q solution, the limit of control sequence detection was enhanced 100 fold. Preferential amplification of the target viruses over the control was also enhanced by providing target virus primers at 0.2 μM, and HDV primers at 0.1 μM, and by designing target amplicons smaller than the internal control (Table 2) to be more readily amplified (Fig. 5).

The only conditions identified to date in which the internal control compromised the detection limit of a PCR, was when it was multiplexed with highly degenerate primers for the target virus (data not shown). Under these conditions, at low concentrations of target virus, the control amplicon was amplified in preference to the virus amplicon. Therefore degeneracy was avoided or minimised in the assay primers, being restricted to a single wobble base per primer (Table 1) where unavoidable.

The PCR primers were validated using a large number of known positive clinical samples and cell culture supematants (Table 3). The detection limit of the PCR was confirmed using a ten fold plasmid dilution series (Fig. 4). The difference in detection limit of the influenza A and B RT-PCR (1 molecule detected) compared to the RSV, HuMV and PIV RT-PCRs (10 molecules detected) may be due to the higher numbers of primers included in the latter. The numbers of viruses identified in clinical samples between 29/10/2002 and 14/4/2003 (Table 4) agreed with the UK national trends for RSV and Influenza during winter 2002-2003 (Health Protection Agency Data, available at http://www.phls.co.uk/topics_az/seasonal/menu.htm). Three co-infections were identified, representing 0.92 % of the total, which is fewer than reported in many studies using RT-PCR. However, the co-amplification of all target amplicons from plasmids containing cloned amplicons (Fig. 5 lane +ve) indicates that co-infections can be detected. The low detection limit of nested RT-PCR (as little as a single molecule of target virus) raises the possibility of detecting viral RNA which is not relevant to a patient's current condition. An objective measurement of the data collected in the present study together with patient symptoms is underway to address this issue and provide quantification of the performance of these assays as diagnostic indicators of viral disease.

Claims

1. Use of a nucleic acid molecule having intra-molecular base pairing as an internal control in nucleic acid amplification.

2. Use according to claim 1 where the nucleic acid molecule is RNA.

3. Use according to any preceding claim where the nucleic acid molecule is self- stabilising due to GC intra-molecular base pairing.

4. Use according to any preceding claim where the nucleic acid molecule has greater than 50% intra-molecular base pairing.

5. Use according to any preceding claim where the nucleic acid molecule is a modified hepatitis delta virus genome.

6. Use according to any preceding claim where the nucleic acid molecule is a modified hepatitis delta virus genome comprising 1.2 copies of the genome with an ORF encoding the delta antigen inactivated by a 2 nucleotide deletion.

7. Use according to any preceding claim where the nucleic acid amplification is PCR.

8. Use according to any preceding claim where the nucleic acid amplification is RT- PCR.

9. Use according to any preceding claim where the nucleic acid amplification is clinically diagnostic RT-PCR.

10. Use according to any preceding claim where the nucleic acid amplification for RT-PCR detection of viral particle in a sample.

11. A method of nucleic acid amplification comprising having components which amplify target nucleic acid if it is present in a sample and comprising a nucleic acid molecule having intra-molecular base pairing as an internal control.

12. A kit comprising the nucleic acid molecule of any preceding claim and instructions for using the nucleic acid molecule as an internal control a nucleic acid amplification.