Astronomer Carl Sagan once said that “extraordinary claims require extraordinary evidence” and that basis I was primed for scepticism when I first glanced at this paper in mBio, which reports apparently human DNA sequences (specifically from the Line 1 retrotransposon) within a bacterial genome (that of the obligate human pathogen, Neisseria gonorrhoeae). But the most exciting scientific reports are those that overturn one’s assumptions and force one to realise that “there are more things in heaven and Earth than were ever dreamt of in your philosophy”. Examples for me include the discovery of reverse transcriptase in bacteria, group II introns in bacteria, the non-essentiality of the ATPase in flagellar protein export and evidence of Neanderthal admixture in non-African human genomes. After careful reading of the paper, I cannot find any obvious flaws, so I have to accept that human DNA does sometimes get incorporated into bacterial genomes.
But as a professional sceptic, I wanted to check whether these authors had not told the whole story, so I had a look for myself at the evidence for Line 1-like sequences in bacteria. The easiest place to start is to look for homology at the protein level. There are two CDSs within Line 1. Let’s start with ORF1. A BlastP search of the NR database does indeed find the gonococcal sequences:
But when the search results are limited to bacterial sequences, one also finds these
Which at first glance provides additional evidence that Line-1-like sequences have found their way into other bacterial genome sequences. However, given that Beggiatoa is not intimately associated with humans and the sequence sits in its own little contig, one has to assume that contamination of the sequencing library is the most likely explanation. Ditto for the TM7 sequences.
What if we try similar searches with the second Line 1 CDS, the one that encodes reverse transcriptase. Well, here we turn up a dozen or more hits to bacterial proteins:
All of these are in dinky little contigs, again suggesting contamination of the bacterial DNA with human DNA rather then genuine HGT.
All that is, apart from the gonococcal Line-1 RT sequence!! This sits slap bang in the middle of the TCDC-NG08107 gonococcal chromosome, next to an invertase-related gene (just like in the paper!), providing clear evidence that there is more going on here than the authors report.
Now, I was prepared to leave it there, but for some reason, I decided to look for hits to the RT CDS just in Neisseria and then discovered over a dozen matches to the Line-1 RT among proteins from the Neisseria polysaccharea ATCC 43768 genome. These all appear to be in small contigs, so look like contamination!
So the take away message is that, yes, it seems plausible that Line-1 DNA has made its way into gonococcal genomes, but given that Line-1 DNA has also contaminated many bacterial genomes, I think we should say that the verdict is so far rests on the balance of probabilities, rather than being beyond all reasonable doubt.
P.S. Oh, and I just checked and the marine metagenome in the environmental DNA archive is also chock full of Line 1 sequences!