Re: Nicholas Bennett and the "HIV-1 infectious molecular clone" 8 September 2004
Previous Rapid Response Next Rapid Response Top
Nicholas Bennett,
Infectious Disease Postdoc/Clinician
Department of Pediatrics, University Hospital, Syracuse NY

Send response to journal:
Re: Re: Nicholas Bennett and the "HIV-1 infectious molecular clone"

I'm intrigued more than anything at the Perth Group's audacity. They question my comparison of cloning a virus genome versus an infectious molecular clone by using self-created definitions! Note that I did make clear that I was talking about the virus _genome_ being an infectious molecular clone - and anyone with any experience at all with molecular genetics would be hard-pushed to disagree. The Perth Group have confused virus isolation with an infectious molecular clone, whether through an inability to read or comprehend, or purposeful misdirection remains to be seen.

The Perth Group then go on to describe in laborious detail some history about DNA cloning, which doesn't really contribute much after the sentence:

"DNA cloning-the production of identical copies of a DNA fragment, any DNA fragment, from an ancestral DNA fragment by splicing it into a suitable cloning vehicle"

That's it. Therefore, by isolating a viral genome and placing it into a DNA vector, we have our infectious molecular clone, and lo and behold it is practically synonymous with DNA cloning.

The Perth Group then go onto to invent criteria for virus isolation that are not necessary or have been done already. 100% pure preparations are not required if one has a method for controlling for uninfected cultures (i.e. by transfecting control DNA or infecting with control blood samples) and the detected genetic material is only in the infected sample. Molecular methods such as PCR can amplify plasmid genes from unpurified bacterial colonies (straight off the petri dish!), so detecting a viral gene or genome from a filtered, clarified and phenol-extracted sample of a culture supernatant can hardly be considered heretical.

As I have repeatedly stated, supported by various references, the HIV genome does encode the proposed viral proteins. GenBank alone should tell you this. Richieri et al (1998) characterised the protein constituents of the virus particles, to name but one.

Countless experiments in the literature have detailed HIV integration, protein expression and used appropriate controls, as suggested by the Perth Group. As such, any reasonable critique of the literature has no choice but to conclude that HIV has been isolated and characterised at the molecular level. I am sure they are perfectly capable of looking this up themselves.

Brian Foley is quite right that there is no need to specify an origin: in fact, as the Perth Group themselves acknowledge, if a retrovirus is integrated then there can be no logical objection to starting with a proviral clone, and using that as a basis for the infectious molecular clone. One must then prove of course that the virus particles produced are then capable of integration and further replication, but fundamentally whether you start with an RNA sequence AUCG and reverse transcribe it to ATCG to clone it, or start with the provirus ATCG, it makes no difference. Their example of the blood bank is demonstrative of their need for a quick-fix single-paper answer to all the questions. They seem unable to understand that later experiments will confirm or refute the original thought (i.e. comparing your cloned genes to those of Dolly). They are also simplifying things by comparing a genomic isolation to grabbing a random bag of blood: far more thought and background knowledge goes into these things, inreality one would have the patient ID on a piece of paper to cross-reference with the ID tag on the blood bag.

There really is no conceptual difference between the genetic code existing as a provirus, an RNA genome, a cDNA's just a sequence of nucleotide bases.

Finally, the Perth Group misrepresent the Weiss paper. Full length genomes were obtained from PBMC, described in the very first Results paragraph no less.

"Because of the high proviral concentration in PBMC from this patient, we were able to amplify a 9.0-kb fragment by using 200 ng of genomic DNA. This fragment was cloned into pGEM13-LTR." Constraints of detection (the reason why people like Gallo in the past, using inferior techniques, couldn't always detect HIV sequences) made a two-step approach necessary to obtain genomes from other tissues. Are the Perth Group going to argue that detection of the 5`-end of the genome encoding the Group Antigen is not sufficient to state detection of HIV!?

Hyrbidisation techniques can not in no way, ever, at all, be used to isolate a genetic sequence. They can only be used to detect a sequence. There is no "standard" hybridisation technique in this context. Specificity of the technique is implicit in the results: the fact that lower target concentration (with the same overall DNA concentration) did not result in amplification of DNA suggests that their primers were appropriate. The genome was of the correct length, partially sequenced, and several clones were replication competant in second-pass culture! That's an extraordinary achievement. The Perth Group can state that "Nowhere in the study can one find evidence" of the results, and yet Figure 2, Figure 3 and Figure 4 could not exist without some kind of virus having been isolated with at least some genetic material and protein production identical to HIV: a virus that behaves much like that of control HIV constructs.

Specificity was also tested as described in the materials and methods section:

"The fidelity of this prolonged PCR approach was checked by digesting the 9.0- and 4.5-kb fragments which were PCR products obtained by using HXB2-containing plasmid as a target. Digestion with 17 selected restriction enzymes, divided throughout the proviral genome, revealed no mutation at least in the restriction enzyme recognition sequences"

I will agree that there are no formal controls as the Perth Group would like there to be (i.e. uninfected cultures in the replication studies), but note that the authors state that the samples do replicate with 10-fold lower maximal p24 antigen amounts relative to lab-adapted strains, and that in the tropism experiment (in which the authors state they "expected productive infection of MDM and BDC by these viruses") no detectable virus was found from the experimental samples, in contrast to the control viruses.

The statement above highlights the uselessness of performing "blind" lab research. No lab research is performed blindly, it is only in the context of clinical trials where that is important due to the psychological placebo effect. Even with certain expectations you cannot always predict the outcome of an experiment.

Additionally, the ELISAs performed in the paper are standardised, so there are external controls to the data at that level.

The Perth Group have once again failed to appreciate that the world has moved on from trying to isolate HIV and prove it exists: the paper in question is discussing the isolation of various different HIV genomes from different organs. A similar paper is Salminen et al, from 1995.

Salminen et al, Virology 213, 80-86 (1995) Recovery of virtually full -length HIV-1 Provirus of diverse subtypes from primary virus cultures using the polymerase chain reaction.

I urge the Perth Group to look at the entire field of HIV characterisation and isolation, both viral and genetic. There really are no inconsistencies, hence no need to question it. HIV has been scrutinised more than possibly any other pathogen in existence, and surely if it were simply a side effect of culture stimulation, gene rearrangemnt or cell-line artifacts someone would have noticed it by now? So many experiments under so many different conditions, and yet the model of an exogenous retrovirus still works, however you choose to look at it. Can the Perth Group come up with a resonable explanation why my own HIV second -passage cultures in unstimulated Jurkat T cells (infected with virus cultured from cells transfected with proviral plasmids) expressed reverse- transcriptase activity, had full-length HIV genomes by RT-PCR and expressed HIV capsid proteins by Western Blot, and the uninfected and untransfected cells did not?

I'm beginning to appreciate that the mind-numbing awfulness of performing sucrose centrifugation on RT-negative samples was worth the effort just to make the point here...

Competing interests: None declared