MERS-CoV partial spike gene sequences do not implicate viral change in April's Jeddah human case cluster

Stars highlight difference in scale at left-hand side,
 (x-axis) numbers. Seasons based on info [2]
Click on image to enlarge.

With a new article at ScienceInsider written by Kai Kupferschmidt (@kakape on Twitter)[1], it seems that the idea of a Spring start to human detections of MERS-CoV in Saudi Arabia is gaining some support from other scientists.

Springing into action...

The "pattern" of MERS-CoV detection seen in the chart still need some fleshing out wit more time. They might be heavily and unrealistically biased by the enormous number of recent hospital detections. However, it's worth noting that in April 2013, it was the Al-Ahsa cluster that added to detection numbers. In April 2014 it is the Jeddah outbreak started with a hospital cluster. What's that about? I say "detection" rather than cases so we don't get bogged down in whether they were ill or not at the time of sampling. In fact the pattern seems to be one that begins with a hospital cluster. 

I've tried to track down when Spring is in this region, and the chart - a rework of one I've been showing for a while now - overlays the seasons of interest.[2]

http://www.flickr.com/photos/xikita/48647105/in/photostream/
CC BY 2.0

Camels have a very long gestational period, up to 15-mths, which makes it difficult to know when a 1-2-year old camel, the ones most likely to acquire MERS-CoV and to shed virus, are around and mixing with humans. Jennifer Yang reported that most camel births occur between Nov and Jan and so their birthdays would fall on...well you can do the maths.

NOTE: The fatal case numbers in the past month, as with most data in the past month, are lacking detail which means I cannot plot them. Pleeease post scrubbed data from all April and late March detections soon WHO.

A pattern of animal to human to hospital to human..?

So could how about this summary of MERS-CoV movement among animals and people?

Young camels become infected within a year or so of birth > virus spreads to humans in contact with camels, a rare few of whom become ill > seriously ill cases shed virus and infect other humans in a healthcare setting > humans spread to other humans, severe illness showing up mostly in those who are already diseased due to something else (e.g. diabetes or cardiac issues)

Hospitals are an obvious starting point for this human-to-human spread to occur because of increased level of close contact and the concentration of already ill people.

But my question today is, as spread of MERS-CoV among humans has likely been occurring since at least 2012 (I suspect earlier), why such a big spike in detections in Jeddah now? Also, lest we forget, among that temporally-linked large cluster of paramedics in Abu Dhabi (being able to link that to Jeddah would answer an important question for me).

Part of the spike gene sequenced but not a sign of viral change...

We learned today, at very long last (I know it's been less than a month but these clusters seem to have been going on for years!), that a part of the MERS-CoV genome from 30/31 samples from Jeddah have been able to be amplified by Christian Drosten's group in Germany (collaborating with Dr Memish and the Kingdom of Saudi Arabia's Ministry of Health) using a separate assay that that used in screening (a safe assumption as spike is not a diagnostic screening target recommended by the WHO). That means that laboratory contamination is a less likely cause of case climb. It doesn't exclude it though. We call this a possible "false positive". It can be due to a reverse-transcription polymerase chain reaction (RT-PCR) being accidentally contaminated ;by some DNA from a previous test and showing up as positive result, when in fact there was no MERS-CoV RNA in that person. Contamination can also occur at the purification step, when (viral) RNA is extracted from the patient material and DNA (or RNA from another high viral load specimen) again contaminates the process. The latter will always be positive while the former may appear virus-positive in 1 reaction run and negative in another.

Sample selection for sequencing...

We don't know how the 31 samples tested by Drosten's lab were selected for packaging and transport to Germany. It is possible that only repeatedly positive samples were sent.

Schematic of a coronavirus virion.
Click to enlarge

In Kupferschmidt's article, Drosten informs us that there is nothing different or distinctive about the 2014 MERS-CoV partial spike gene sequences he's generated thus far when compared to sequences from previously sequenced MERS-CoV genomes in 2012 and 2013. This is not the end of that story though as we are not yet sure which part of the spike was sequenced and still have a lot of the genome to see before we can feel assured that the virus has not changed. We also don't know if sample selection has occurred from cases during the beginning, middle and later stages of the Jeddah cluster and thus whether these new sequence data accurately represent all the viruses circulating among detected in the Jeddah outbreak. But this is a good start.

Severe acute respiratory syndrome (SARS) and spike..

The ~4,000nt, region of the MERS-CoV genome 
that encodes the ~1,300aa spike gene is highlighted in 
pink. Schematic derived from the EMC-2012 variant 
sequence of MERS-CoV.
Click to enlarge.

With the SARS-CoV, evolutionary changes in the spike gene (especially between amino acids 75-1025 or nucleotides 224-3075) could be used retrospectively as a marker of the virus adapting though time,[5] perhaps towards a better transmitting variant among humans. Along with ORF1a, these two regions showed changes suggestive of adaptation from animal to human.[5] I'm not sure that we've seen that with the human/camel genomes sequenced to date have we? Does this reflect that the virus has already "settled in" to humans? Or does it mean that the spike region is just less informative for MERS-CoV than it was for SARS-CoV? To answer that we'd need older camel MERS-CoV sequence. With the new data we have for MERS-CoV, we begin to wonder whether we can exclude viral change. Hopefully that information will be forthcoming when compete genomes are sequenced and when we know more about which "part of the spike gene" was sequenced. Those details will probably be delivered via a rapid scientific publication.

So, a day full of new data and a mad day on Twitter. 

Thanks to @nika7k, @yasnot and @AB_Algaissi for very helpful Twitter exchanges on this topic this morning.

Sources...

1. Soaring MERS Cases Cause Pandemic Jitters, but Causes Are Unclear
   http://news.sciencemag.org/health/2014/04/soaring-mers-cases-cause-pandemic-jitters-causes-are-unclear
2. http://www.iexplore.com/travel-guides/middle-east/saudi-arabia/weather
3. Tracking MERS-CoV through time: a spikey problem
   http://virologydownunder.blogspot.com.au/2013/08/tracking-mers-cov-through-time-spikey.html
4. Saudi Geography and Climate
   http://fanack.com/en/countries/saudi-arabia/basic-facts/geography-and-climate/
5. Molecular evolution of the SARS coronavirus during the course of the SARS epidemic in China
   http://www.ncbi.nlm.nih.gov/pubmed/14752165
6. Camels likely source of deadly coronavirus, study shows
   http://www.thestar.com/news/world/2014/02/25/camels_likely_source_of_deadly_coronavirus_study_shows.html

Tracking MERS-CoV through time: a spikey problem

This morning on Twitter, Helen Branswell (@HelenBranswell) asked this question, with a comment...

How is this so? 16 months after the 1st known #MERS cases, there are only 9 viral sequences in the public domain. So much for collaboration.
— Helen Branswell (@HelenBranswell) August 6, 2013

So I thought a little perspective might be nice. 

The SARS epidemic had its origins around Nov 16th 2002, although the major activity started in Feb of 2003. 

• 64 human SARS-CoV genomes had been produced by September 2003 ([UPDATED:] see Science paper). That is by 317-days later, or 10-months, 13-days (perhaps less given that the genome sequences were possibly sequenced well before the paper was submitted e.g. late phase genome s seem to have been submitted to GenBank by July 2003). 
• For MERS-CoV we currently have 9 genomes at 505-days (give or take), or 1-year, 4-months.

Not that anyone needs to be reminded, but 80% of MERS-CoV cases come from the Kingdom of Saudi Arabia. The world is relying on them, or their collaborators, to turn the nucleic acid extracts used to define these cases (PCR-POSs hopefully kept in a -80'C freezer), into templates for gene or genome sequencing.

I personally don't believe we need to have complete genomes right now in order to fulfil the fairly urgent public health need to monitor the virus and notice if it changes, or is changing, or is not changing. These changes tell us whether the virus is still adapting or has settled in - perhaps having done so prior to this outbreak's indicator, severe disease. 

What else to use to track adaptation?

Perhaps the 4,000nt Spike (S) gene, or some smaller but suitably variable portion of it, could be a target for sequencing? 

Zhang and colleagues have data showing it could be used to track an animal coronavirus's adaptation to humans, through its 3 pandemic phases. This was done using phylogeny (a way to show how one sequence relates to another through time and space) of nucleic acid sequences and alignments of the translated version of these sequences. All we need is primer sequences that could be used to reliably amplify the S gene of the MERS-CoV. If anyone has those already perhaps they could publish them...if they haven't already. A very brief look at the 9 MERS-CoV genomes already shows some variety. Perhaps unsurprisingly, there is very little change among the 4 Al-Ahsa genomes; their collection dates are separated in time by 17-days.

This shows a schematic of the aligned Spike genes. The black lines within the grey boxes represent nucleotides that differ from the consensus. More differences are obvious in the earlier sequences. The oldest MERS-CoV isolate is at the bottom, the most recent, at the top (detailed below). See the full version here at VDU.

Interestingly, the phylogeny of the complete Spike genes looks  similar to that of the complete MERS-CoV genomes. However  its doe snot place the isolates in order of increasing time to the extent that the full genomes do. I also looked at a 900bp fragment of the 3' of the Spike gene - easier to amplify but a very similar tree to that of the complete Spike.

All 9 complete MERS-CoV spike protein genes (nt). Alignment in Geneious Pro, tree in MEGA 5.10.
Full version will be here at VDU.

All 9 complete MERS-CoV genomes (nt). The arrow indicates moving forward in time; the oldest MERS-CoV isolates at the bottom, the most recent at the top. Alignment in Geneious Pro, tree in MEGA 5.10.
Full version will be here at VDU.

So where does that leave us?

Adaptive pressures on the SARS-CoV drove its genome towards settling down in the late stage of the 3-phase outbreak (defined by the Chinese SARS Molecular Epidemiology Consortium), with changes in the Spike gene occurring before that. Complete genomes are clearly the gold standard - so I dial down that personal belief from earlier.

The Spike gene still seems a useful target for MERS-CoV too, although not as accurate at plotting the time of virus isolation as complete MERS-CoV genomes were in my example above. Still, it, or some part of it, is still of use as an early-warning system to alert us to viral change and it will prove easier to amplify by smaller or less genomics-focussed laboratories. Something we need to consider in order to get some information, which is far better than none.

While we've seen predictive modelling for the age of MERS-CoV, we don't actually know when the virus came to be or when it started spilling over to humans. More full genome sequences would certainly help address that question. And finding its origin.

However, perhaps we should make the trade off and use the 3' end of the Spike gene now, in an effort to keep some sort of eye on how the MERS-CoV is travelling? Anyone else have a good region that fits the bill?