Detecting Zika viruses using PCR: a look at published assays...

[NOTE: I'll update this post with other assays and information as I find time]

You may be helped - hopefully anyway - by reading the earlier post in this series - Primers prime PCR unless past their prime... [1]

Also hopefully helpful - the Zika virus (ZIKV) lineages I'll be referring too can be seen in this (very basic) tree. The viruses currently circulating in the Americas fall into the "Asian lineage".

Basic tree showing the majority genome sequences of ZIKV variants form discovery to 2015.

PCR and ZIKV...

Because flaviviruses are "RNA viruses" and because ZIKV is a flavivirus, we'll be talking about the range of RT-PCRs that have been described in the literature for ZIKV. This will include both RT-rtPCR (Reverse Transcript real-time Polymerase Chain Reaction) but also the non-real-time versions which I'll be calling "conventional" RT-PCRs or RT-conPCRs.

This will not be all the variants and home brew concoction of RT-PCR out there for ZIKV, but it is the ones that have been published and some of the articles that reference them.

I'll be counting nucleotide positions from 5' to 3' (or "left" to "right") so nucleotide #1 will be furthest form the pointy end of the green arrow-no, not that Green Arrow).

I'm using Spondweni virus, a related but different virus (usually) at the top of the alignment and in the tree. This is to provide something that shows how different a very near cousin is, even when "zoomed in" and compared to the brothers and sister variants of ZIKV. 

In some alignments I have marked, in the left most panel, some - but not all - examples of Asian or African lineage variants. See if you can spot patterns among the nucleotides that may be specific to each. 

------------------------

One-Step RT-PCR for detection of Zika virus
Faye et al. J Clin Virol 2008 43:96-101 [2]
PCR Aim: To detect ZIKV in human serum
Subsequent publications using this assay:

Faye et al assay. (A) The forward primer and (B) the reverse primer aligned against a range of African and Asian lineage ZIKV variant sequences.

• Both primers have degenerate positions and look as though they should work well against African and Asian lineages on paper
• There is a mismatch in the reverse primer at position #16 that is not accounted for and may be destabilising for a number of Asian lineage variant templates

----------------------

Genetic and Serologic Properties of Zika Virus Associated with an Epidemic, Yap State,Micronesia, 2007
Lanciotti et al. Emerg Infect Dis 14(8):1232-9 [3]
PCR Aim: To rescreen sera from the 2007 Yap Island epidemic
Subsequent publications using this assay:[3]

Assay 1...

Lanciotti et al assay #1. (A) The forward primer (853), (B) the real-time PCR probe (860-FAM) and (C) the reverse primer (911c) aligned against a range of African and Asian lineage ZIKV variant sequences.

• The forward primer must interact with a 2-3 mutations in some of the the African variants, but is a good match with all Asian lineages variants.
• The probe also has 3 mismatches with the same variant as above and 1-2 with African lineage variants but is again a good  match for most Asian lineage variants
• The reverse primer is a great match to Asian variants but not very good if faced with trying to detect an African variant, especially KF383117

Assay 2...

Lanciotti et al assay #2. (A) The forward primer (1086), (B) the real-time PCR probe (1107-FAM) and (C) the reverse primer (1162c) aligned against a range of African and Asian lineage ZIKV variant sequences.

• The forward primer works well against most variants and the mismatches it does face are at the less destabilising 5' end, against some African lineage variants.
• The probe is a good match for most variants of either lineage, having a couple of issues with 2 Central African Republic variants and KF383118 poses a problem at the 5' end of the probe which could be a issue for the polymerase which comes charging down the same strand as the which the forward primer is hybridised to, looking to chop the probe up.
• The reverse primer is a great match to Asian variants except for KF383117 and will suffer between 1 to 5 mismatches with African lineage variants - also impacting across the primer landing site. 

Generally these 2 assays should be good for detecting the Asian variants, will suffer varying degrees of performance issues against African variants.

Imported Zika Virus Infection from the Cook Islands into Australia, 2014
Pyke et al...PLoS Curr. 2014 Jun 2;6. pii
PCR Aim:
Subsequent publications using this assay

E gene assay...

Coming up...

Quantitative real-time PCR detection of Zika virus and evaluation with field-caught Mosquitoes
Faye et al. Virol J 2013 10:311

References...

1. Primers prime PCR unless past their prime...
   http://virologydownunder.blogspot.com.au/2016/03/primers-prime-pcr-unless-past-their.html
2. One-step RT-PCR for detection of Zika virus
   http://www.ncbi.nlm.nih.gov/pubmed/18674965
3. Genetic and serologic properties of Zika virus associated with an epidemic, Yap State, Micronesia, 2007
   http://dx.doi.org/10.3201/eid1408.080287
4. Zika Virus Infection in Pregnant Women in Rio de Janeiro — Preliminary Report
   http://www.nejm.org/doi/pdf/10.1056/NEJMoa1602412
5. Quantitative real-time PCR detection of Zika virus and evaluation with field-caught Mosquitoes
   http://www.virologyj.com/content/10/1/311
6. Imported Zika Virus Infection from the Cook Islands into Australia, 2014
   http://currents.plos.org/outbreaks/article/imported-zika-virus-infection-from-the-cook-islands-into-australia-2014/

Primers prime PCR unless past their prime...

The PCR is a tough old bird... 

The polymerase chain reaction (PCR) is a hugely powerful tool for detecting virus. It can still do this more quickly and sensitively, from any human specimen, than any other diagnostic method that came before it.

Amazingly, the method of DNA copying by extension hasn't changed so much since the 1980s. The major variations have been to the machines that control the temperatures, the wide range of kits and the way PCR product is detected.

In the 1990s this detection became about measuring fluorescence within the same tube as the reaction as it took place - in real-time. But prior to this method, detection of a successful PCR was about opening the tube after the run had finished and performing a multi-step lengthy process of identifying the successfully amplified DNA by electrophoresis - end-point detection. This change underpinned what became known as of real-time PCR (rtPCR).

Quite amazing to think that anything could remain so integral to virology - or anything - these days. Rather than go into the details of what real-time PCR is, was and could have been, you can get a pretty good background understanding in our 2002 open access review of it as used in virology, aptly named "Real-time PCR in virology".[1]

PCR in the laboratory with viruses...

Today, rtPCR is integral to detecting the presence of virus and a successful detection is usually taken to mean that the person who provided that sample had a current infection at the time of sampling. 

That's not technically a correct assumption, nonetheless decades of research and epidemiology have largely been based on it. It's a pretty safe bet that there is a good correlation - in my expert opinion - in most of the instances in which we use PCR in a diagnostic laboratory setting. Detection by PCR does not however, mean we have disease or that the signs and symptoms a patient may have at the time of sampling are caused by that virus. The laboratory results along with any relevant travel history, contact history, signs,  symptoms and other clinical investigations and background knowledge all combine to help a medical doctor make that call. It's a team effort but laboratory testing, including PCR, biochemistry and immunology, provide crucial information that very often cannot be supplanted by a diagnosis based on how a patient looks and feels.

When we use PCR we use it to either detect DNA from viruses that have a DNA-based genetic make-up, or RNA from those viruses with RNA-based genetic material. It is of course used for bacteria and fungi and humans and many other things too, but this is not "all manner of other things down under".

For a PCR to work on an RNA virus, given the enzymatic process that is at its core (a DNA dependent polymerase), we first have to make a DNA copy of the RNA - a process called "reverse transcription". This term comes from the need to copy RNA back into DNA  - transcription being the process of copying DNA into RNA which occurs in our cells and elsewhere.

What a primer looks like when you align it to a reference sequence..

This probably deserves its own explanatory blog but we'll see what we can get away with for now.

A nucleotide sequence alignment of Zika virus (ZIKV) sequences downloaded from GenBank aligned with a primer from a published RT-rtPCR assay.[2] 1-the region of a primer that, if poorly "matching" to a target region of viral sequence, will most likley negatively impact the efficiency of the resulting PCR amplification, possibly causing the PCR to perform poorly, which can be of concern when there is only lower amount of of virus present in a sample, or fail completely; 2-in this example,the primer from Lanciotti et al. does not match the target at these marked positions - there is one or more "mismatches" between the primer and some of the viruses represented here which will affect the stability of the primer's binding, or hybridisation, with the viral genetic material during the PCR. Generally-this primer binds very well to Asian lineage Zika viruses, but much less well to African lineage Zika viruses.
Alignment made using Geneious v 8.1.5. The primer, 911c is the reverse primer in a pair used for an RT-rtPCR described by Lanciotti et al.[2]
Click on image to enlarge.

In the figure above we can see a few important items to consider when looking at primers and how they may work when put it into use. I say 'may' because even rubbish primers can sometimes do the job in PCR'ville, while the best designs may sometimes fail. One has to test to know for sure whether a primer pair does or does not perform the way you intend. Much like in infectious disease'ville before you try and link an infection with a disease.

Some key things worth noting about the figure above:

1. This is called a nucleotide sequence alignment - or just an alignment. In this version, I've used dots to highlight whenever one of my reference sequences has the same nucleotide - an A, C G or T - at the same position as the primer I randomly chose for this example.
   
   If there is an entire column of dots, then at that nucleotide position we have a perfect match between all the viruses I've used in this alignment and that primer. In theory the primer will bind to those viruses if the one of them is in the patient's samples and we've successfully purify the virus's genetic material away from the other unwanted stuff in the samples - like proteins, inhibitory chemicals and carbohydrates.
   
   I won't be covering this "extraction" or purification process here.
   
2. The enzyme in the PCR that makes PCR work relies on the primer to be matched up with the viral genetic material - especially at one end - the end at which the enzyme is going to be extending. I've highlighted that direction with the green arrow (no, not the Green Arrow).
   
   As the enzyme moves, it adds new nucleotides which it pulls in from the PCR chemical soup we use (in kits these days). This extension eventually results in a copy of the virus sequence. The primer shown is the reverse primer. There is also a forward primer which makes a copy in the other direction of the newly made strand. There is a bit more to this strand thing, but also not covering that here.
   
   As the copies get copied we eventually end up with just the region spanned by the primer pair amplified up - cloned if you like. Within that region, copied millions of times, is a sequence to which we previously designed a probe labelled with fluorescence molecules that has also been added to the chemical soup.
   
   As the copies were being made, the probe bound to its target and was destroyed, results in the signal which allows us to see those clones pile up....in real time.
   
3. The end of the primer from which the enzyme starts must be well bound, or hybridised, to the viral sequence or else the polymerase may just fall off and not have anything to copy. This is really important to the success of the PCR, and if it's not a stable hybridisation (As hybridise to Ts, Cs hybridise to Gs), then the PCR won't work, or will work poorly.
   
   If hybridisation is poor then the early cycles don't make those first few copies which act as the template for more copies, and it is these early cycles which make or break the success of a PCR.
   
   In the figure, the important end is highlighted by the orange pill shape.
   
4. I've added some info to the sequence names used in the alignment above. This indicates where and when some of them came from.
   
   The Asian lineage (see the basic tree below-also made using a nucleotide sequence alignment as a starting point)  is the one currently circulating in the Americas and previously in the Yap Island outbreak in 2007 and in French Polynesia.
   

   

   Zika virus complete or near-complete genome sequences aligned using Geneious v8.1.5.
   Neighbor-joining p-distance tree mad in Mega v6; 500 bootstraps 

   
   The primer example in the  alignment does not have any, but sometimes a PCR primer can have one or more "degenerate" nucleotide positions in it. These introduce variations to the primer sequence to accommodate one or more different nucleotides in the viral genetic target sequence.
   
   Sometimes, one has no choice but to design the primer to a variable viral region and so a degenerate position or two is added into your primer to help the primer hybridise and cope with that.
   
   We use codes for these degenerate positions - see the Table below excerpted from Chapter 2, written by Andreas Nitsche in a book I edited about a decade ago and in an extended version of the International Union of Pure and Applied Chemistry (IUPAC) code.[3,4]
   
   
   This is all a bit of a misnomer though because it means that when we order a primer from a company, they actually make a mix of primers - some have the full sequence and one nucleotide of the degenerate position, and the rest of the primer with the full sequence and the other nucleotide at that position - up to 4 different primers in the mix if we use an "N" degenerate position.

So with all that in mind - the next post - because this one is now too long - will examine some of the PCR-based diagnostics for Zika virus (ZIKV).

References...

1. Real-time PCR in virology
   http://nar.oxfordjournals.org/content/30/6/1292.full
2. Genetic and serologic properties of Zika virus associated with an epidemic, Yap State, Micronesia, 2007
   http://wwwnc.cdc.gov/eid/article/14/8/08-0287_article
3. Real-Time PCR in Microbiology: From Diagnosis to Characterization
   http://www.horizonpress.com/rtmic
4. An extended IUPAC nomenclature code for polymorphic nucleic acids
   http://www.ncbi.nlm.nih.gov/pmc/articles/PMC2865858/

New MERS-CoV positives - is there a brevity competition going on...a shortage of electrons perhaps?

2 MERS-CoV POS, 26 and 19 years, released from hospital.

That is the only relevant text from 68 Google translated press release words on these 2 new cases posted on the Arabic language Ministry of Health website, Saudi Arabia. 56 words on the English-language site. I used half as many just explaining where it came from.

2 other cases were announced as positive and already discharged (the children of the 38M from Harf Al Batin, from where 1 of the cases above also tested positive) as well.

So much for being hopeful of a new era of data sharing on Friday. Must have thought better of it after the weekend.

So we're at 110 cases, 50 deaths (1 still unaccounted for) resulting in a fatal case proportion of 46.4%. That's about 16 cases reported since August 20th after a nearly 3-week pause in new cases being described. That's a rate of 1.2 lab confirmations reported per day. 

Not exactly setting the world on fire but it is worth considering what may be happening among other contacts; those perhaps not defined as close enough to test with PCR. If there are any of course. Hospital-based cases can be a great sentinel of an increase in mild respiratory viruses infections out in the broader community. 

My first source: Avian Flu Diary's article

T.perforatus MERS-CoV strain sequence, and others, online...

If you're a bit of a sequence collector/hoarder/nut then you'll be interested to know that the recent bat CoV RNA-dependent RNA polymerase (RdRp) sequences are now online on GenBank.

• Bat coronavirus Pikuh/CII_KSA_001/Riyadh/Saudi Arabia/2013
• Bat coronavirus Rhhar/CII_KSA_002/Bisha/Saudi Arabia/2013
• Bat coronavirus Rhhar/CII_KSA_003/Bisha/Saudi Arabia/2013
• Bat Coronavirus Taper/CII_KSA_287/Bisha/Saudi Arabia/2012
• Bat Coronavirus Rhhar/CII_KSA_004/Bisha/Saudi Arabia/2013

These seem to include the primer regions judging by their length. Consider that when using them.

The MERS-CoV strain from T.perfortaus is CII_KSA_287 - highlighted in bold. Please note, that at writing, it is erroneously identified on GenBank as originating from Rhinopoma hardwickii. It should be Taphozous perforatus.

Why only 181 nucleotides of T.perforatus MERS-CoV sequence?

In some of the many articles written about the new discovery this week, there were comments along the lines of  its amazing any sequence could be obtained from the samples cause they had sat for 48-hours at US customs and thawed. A more precise quote could be found here for example.

I have some thoughts on that - and these come from me, someone who has worked with a lot of clinical human specimens from which I've been able to amplify viral bits and pieces on a regular basis. Many small (200-600 basepairs[bp] fragments) but also longer pieces of >1,000bp, assembling small viral genomes from them. These samples may >10-years old, having been freeze-thawed numerous times after spending various amounts of time in courier vans, planes or sitting at room temperature before having nucleic acids extracted, tested and eventually (extracts may also sit around during testing and preparation and be freeze-thawed etc) frozen at -20°C or -80°C.

Keeping in mind that this issue of thawing might simply be a case of "hold your horses people". The EID paper was an early and quick report announcing the discovery of this MERS-CoV strain. So, my thoughts:

1. Because the materials that yielded the sequence (collected in October 2012) were described as "thawed" we can presume that the dry ice they were shipped with ran out during the transport to, or waiting time at, US customs. Once the refrigerant is all gone, the samples would come to room temperature as fast as the cardboard box and plastic receptacle it held, allowed. The publication described them as having been thawed for 48-hours.
2. How warm are we talking? The average temperature of Bisha (where the Taphozus perforatus bat was found, in an old date orchard outdoors) in October ranges from 15-20°C to 30-35°C. I don't know where the US customs site was so don't know that temp range - but expect it's less. So let's make some wholly unfounded assumptions:
• That this MERS-CoV strain can spread via the virus found in faecal pellets or other bat excreta. Perhaps as wind-blown dust or to other animals via a faecal-oral route. Even if the bats are hanging from a cave ceiling, but certainly when they are hanging outdoors, the virus must be capable of surviving in faecal pellets at a very high "room temperature" to complete a transmission event. If they can survive, that means intact virus - RNA genome + proteins + capsid + lipid envelope - the whole lot. For RT-PCR - you only need the RNA bit, not infectious virus. So, you're already lowering your expectations for what's required of a "successful" shipment.
• To confirm bat species, a genetic test was used which required the amplification of another piece of DNA - a region of the cytochrome B gene was amplified and sequenced. How large this fragment was, I'm not sure. However, a relatively large fragment of this gene can be used to differentiates bats, useful when you can't tell them apart by looking at physical features. Other work on opossums by the collaborator who helped sequence this region (Dr George Amato) in bats, employed >800bp of sequence. Why did this fragment amplify so well if the viral RNA did not? Perhaps because DNA is more hardy (various reasons) or because the bat blood or skin that it was amplified from, better protected the DNA from the thawing than bat faeces did for the viral RNA? Or...
3. Perhaps the primers used for other regions of the T.perforatus MES-CoV strain failed because the virus was too genetically distinct. I've had a look at the alignments and the primer binding sites can be found so it's probably not that. However, some of these primers that produce larger products are very degenerate (primers specially designed to account for nucleotide variation in a range of subtly different viruses or viral strains). 
• Degenerate PCR primers generally have much decreased sensitivity compared to 100% target-specific primers. This drop in ability to detect low amounts of RNA is the case even when using nested PCR - sorry if this has become to PCR technical! 
• The primers that did work for the T.perforatus bat MERS-CoV, Nested CII-MERS-RdRp, were much more target specific with only 1 degenerate base in 4 primers. That, combined with a drop in viral RNA amount, may well be why this 1 assay worked, worked where the others did not.
4. There was no mention in the EID paper of the use of an internal control RT-PCR target - a region of a gene in bat faeces (or blood or tissue depending on what was tested) that might allow some quality monitoring to see if there was truly decreased amounts of intact RNA in the October 2012 batch compared to that in the April 2013 batch of samples. That would be helpful to know which course to follow next.

So what does all this mean? Just me thinking in print I guess.

It's always important to maintain the cold chain from sample collection through to nucleic acid extraction and template addition to an RT-PCR/PCR tube. But I think we should look elsewhere for reasons why the T.perforatus MERS-CoV-positive sample has not yielded more than 1 fragment from the few assays used. 

I wouldn't be surprised if there was more sequence coming soon from this sample.