Discoveries

This adventure (2001 – 2023) was a perfect example of the difference between following the data and adherence to dogma for the augmentation of some imagined prestige. What began as a typical gene discovery and characterization process, namely a gene that regulates telomere length in mice, became colonized by a priori thinking and then subjected to adherence to a model that had, and still has, no experimental basis. The discoveries that resulted from my approach indicate that the very real telomeric effects of human RTEL1, and the pathogenic consequences of mutations in RTEL1 or haploinsufficiency, derive from processes that are elsewhere, and for which a curious researcher would be well-challenged to explain using current models of telomere biogenesis and homeostasis.

It should be noted that, apart from the identification of pathogenic mutations in RTEL1 in patients with HHS (Hoyeraal-Hreidarsson syndrome) or IPF (Idiopathic pulmonary fribrosis), there has been no finding about human RTEL1 that has aided my research in any appreciable way. This is in itself is condamnation enough for their approach.

The beginning of the compounding of errors was the adoption of a model by the community which was perfectly valid for another gene (dog-1 in C.elegans; see below under discoveries). This model which was devised (by Dr. Peter Lansdorp and myself) to explain the observed behaviour of a gene which maintains the stability of guanine-rich sequences which are capable of forming higher-order structures (G -quadruplexes) and which can be impediments to DNA replication. This function was validated in the human system by its homolog FANCJ/BRIP1 (Brosh). Telomeres also form G-quadruplxes and so the model was applied to RTEL1, which is a helicase in the same family as FANCJ. However, the ability of RTLE1 to unwind G-quadruplexes has never been demonstrated, nor has the preferred substrate(s) for it been determined. Hence, it is a bad model, which should not come as a surprise since it is descriptive of an entirely different gene. It should be noted that a hypothes or model such as this that has never been tested, only assumed, speaks eloquently and definitively for its scientific value.

The utility of a concept can also be an obstacle to further knowledge if its provisional nature is forgotten.

The primary impediment to scientific advance is not religion, or political interference, or cultural ignorance, but obsolete science. And science is almost always obsolete.

Error is compounded when the interpretation of an observation is subjugated to an unfounded assumption. Numerous such errors dependent upon the same assumption result in a network of false induction which appears coherent but whose coherence is only that of the logical relations of unjustified conclusions. For such errors to thrive they must resist conditions that will question their validity. In this way a closed system is created which must remain closed. The necessity of incorprating such a system within the larger framework of cell biology for instance will ultimately fail, revealing the inadquacy of the conception. And so, such efforts of incorporation will be resisted and the system will remain closed, and this, even if the system seems to be growing (which it must if the only way to publish or communicate one’s results is in reference to the closed system).

In other words, a system of self-referential errors must be resistant to a transformative observation and so, it is not surpriing that a common criticism to apparently radically new observations is, if this is true we would have seen it already. A criticism that avows an incapacity to see, which is conditioned by the adherence to the system of self-referential erros in question.

I can certainly be accused of harbouring preconceptions. If there is one I am unwilling to relinquish it is the idea that all biological processes are linked. In my concerns I am speaking about the limited realm of cellular biology. It is of course necessary to limit one’s area of concern (replication, mitosis, DNA repair) in order to study what is key to such processes, but it must not be forgotten that the pragmatic intellectual simplifications must eventually be removed and our models re-integrated into the complexity of the cellular process of which they are a part. It is this re-integration which is the measure of the robustness of the knowledge we have gained and the very real resistance of scientists to accept the fusion of their domain of interest with that of another is due to psychological reasons and not to scientific justifications

Hoyeraal-Hreidarsson syndrome (HHS) patient-derived fibroblasts with mutations in te RING domain of RTEL1 have NE deformations and ruptures.

RTEL1-dependent deformations and ruptures require entry into S-phase (CDC7-dependent) but do not require DNA synthesis (aphidicolin-insensitive).

Deformations and ruptures are depleted for LMNB1 but not LMNA/C, which is contrary to HGPS (progeria).

RTEL1 protein is imported by the classical import pathway involving KPNA2/KPNB1 and the NLS of RTEL1, and also by KPNB1 alone. This latter import pathway is reponsible for NE integrity.

The KPNB1 import pathway of RTEL1 involves an interaction between RTEL1 and NUP153 and an intact and functional RING domain of RTEL1.

Inhibition of KPNB1 import with importazole results in NE deformations and ruptures indistinguishable from those caused by RTEL1 dysfunction. RTEL1 overexpression can rescue these defects, presumably by binding KPNB1 and blocking binding acces of the drug..

RIF1 interacts with the C-terminus of RTEL1 and the delpletion of RIF1 results in NE deformations and ruptures indistinguishable from those induced by RTEL1 dysfunction. RTEL1 does not rescue these defects indicating that RTEL1 acts upstream of RIF1.

NE deformations are present in pre-senescent fibroblasts and this corresponds to a diminution of RTEL1 and LMNB1 protein expression. 

Identification of a peptide in RTEL1 (ITD, or import targeting domain) which can target small molecules (< 60 kDa) to the nuclear pore and promote their import in a non-energetic manner (no RANGTP required). 

RTEL1 is required for the biogenesis of U2 snRNA (and other spliceosomal RNAs) via XPO1 and RANGTP, and this requires transport of the preU2 RNA to the cytoplasm. RTEL1 dysfunction results in widespread splicing defects.

RTEL1 is required for the biogenesis of TERC (RNA component of telomerase). RTEL1 dysfunction results in the cytoplasmic aggregation of TERC.

RTEL1 is involved in the splicing of TERT (catalytic component of telomerase).

RTEL1 regulates its own splicing.

RTEL1 is involved in the processing of primary ribosomal RNA transcripts. RTEL1 dysfunction results in grossly deformed nucleoli and reduced levels of primary transcripts.

RTEL1 dysfunction has no effect on polyadenylated mRNAs.

The overexpression of a C-terminus deletion mutant of RTEL1 results in complete loss of XIST RNA.

RTEL1 is present in centrosomes and certain RTEL1 mutations result in mispostioned centrosomes and multiple centrosomes.

This gene is the worm homolog of FANCJ/BRIP1. It was the first demonstration of the mechanism of a eukaryotic helicase ensuring the stability of gauanine-rich sequences of a length sufficient to form higher order structures such as G-quadruplexes. This mechanism was eventually validated for mammalian FANCJ (Brosh). The model for the action of dog-1 was adopted to explain the action of RTEL1. Although RTEL1 is a member of the Rad3 family of helicases as is FANCJ (XPD/ERCC2 and DDX11 are the other two members) its preferred substrate is unknown and has never been shown to unwind G-quartets. The dog-1 model of helicase action has resulted in no predictive success in the RTEL1 context, nor any verified experimental observations based on it; as such the model has far outlived any usefullness it may have had and should, like all such models, be retired.

As all research should do, my work on human RTEL1 has led to a hypothesis; moreover, it is one that has clinical implications. Patients defective in RTEL1 function can present two phenotypically distinct diseases: Idiopathic pulmonary fibrosis (IPF), and Hoyeraal-Hreidarsson syndrome (HHS). The latter syndrome is fatal and presents in the first years of life as sever bone marrow failure. The hypothesis is: the phenotypic differences between IPF and HHS may be the difference between haploinsufficiency and obligate expression of pathogenic RTEL1 alleles.


During routine non-directional sub-cloning of DNA into plasmid vectors and subsequent propogation in E.coli I noticed that the expected 1:1 ratio of the orientation of the insert was severely affected when the insert contained guanine-rich (G-rich) sequences capable of forming secondary structures that can impeded replication. In addition, for directional cloning, certain G-rich constructs were impossible to clone. For historical reasons, commonly used cloning vectors have the bacterial origin of replication sequence close to the multiple cloning site (MCS), which results in an asymmetry in the probablity of replication of the inserted DNA by leading or lagging strand synthesis. It so happened that G-rich sequences in the inserted sequences were preferentially replicated by leading strand sequence (as expected), which resulted in a bias in the observed orientation of resultant sub-clones, often up to 100%. After re-engineering the cloning vector so that the origin of replication was in the middle plasmid with respect to the MCS, the expected ratio was re-established, meaning that leading strand synthesis was possible no matter the orientation of the insert. Importantly, in these vectors directional cloning of G-rich constructs that were impossible to clone previously, were successfully generated.