Update: Please see the article available at this link for an important correction and revision to the argument entertained here.
L-gulonolactone oxidase (GULO) is the final enzyme of the biosynthetic pathway leading to the production of vitamin C (ascorbic acid) from glucose and galactose. Many animals (and most plants) possess a GULO gene that produces a functional protein. Others (such as teleost fish) don’t possess any GULO-like sequences. In still others — including humans and several other mammals such as primates, guinea pigs, and the bat genus Pteropus — the GULO gene is present but appears to have been “broken” by inactivating mutations. Hence, these organisms need to obtain their vitamin C through their diet. GULO pseudogenes in these organisms possess multiple indel mutatons and premature stop codons, meaning that they are not translated into a functional protein.
Evidence For Endogenous Production of Vitamin C
Now, there are in fact some data to suggest that vitamin C may in fact be produced endogenously in utero. In other words, humans seem to be producing their own vitamin C, as evidenced by increased levels of vitamin C in humans that cannot be explained by dietary intake alone. This suggests the vitamin C “pseudogene” may not necessarily be “pseudo” after all.
Evidence indicates that ascorbic acid in the human fetus and in the neonate is not wholly explicable by the mother’s intake of vitamin C in her diet. For example, a study conducted by Andersson et al. (1956) documented, over a period of five years, no more than two reports of infantile scurvy in malnourished South African Bantu infants. It was found that ascorbic acid plasma concentrations were comparable among infants who had been well nourished.
A further study, by Adlard et al. (1974), found substantially heightened vitamin C concentrations in the fetal human brain as compared to that of the adult. This concentration was found to fall with increased gestational age.
Salmenpera (1984) examined the levels of plasma vitamin C in infants who had been breast-fed as compared with controls who had been supplemented with vitamin C. They reported that the concentration of plasma vitamin C was the same or higher in the former compared to the latter. In fact, the concentration was roughly double the maternal concentration. The author reported, “Surprisingly, the infantile plasma concentration, which was already high compared with maternal concentration, continued to rise despite the decreasing concentration in milk … the significance of this phenomenon is unknown.”
More recently, a study was published in Nature reporting on the importance of vitamin C as “a direct regulator of Tet [ten eleven translocation enzyme] activity and DNA methylation fidelity in [mouse embryonic stem] cells,” (Blaschke et al., 2013). The Tet enzymes are thought to mediate global DNA demethylation in the germ line and developing embryo by converting 4-methylcytosine into 4-hydroxymethylcytosine (Guo et al., 2011). Vitamin C is a potential cofactor for the Tet enzymes. Blaschke et al. (2013) report that,
[A]dditon of vitamin C to mouse [embryonic stem] cells promotes Tet activity, leading to a rapid and global increase in [5-hydroxymethylcytosine]. This is followed by DNA demethylation of many gene promoters and upregulation of demethylated germline genes. Tet1 binding is enriched near the transcription start site of genes affected by vitamin C treatment.
The authors also suggest that “vitamin C may also have a role in human [embryonic stem] cells, as it has been shown that they accumulate DNA methylation after several passages in the absence of vitamin C, although the underlying mechanisms were not addressed.” Of course, if humans could produce vitamin C in utero, that would explain this evidence.
A Potential Reason For Turning Off Vitamin C Production
We know that, in developed humans, vitamin C is not produced endogenously. In fact, we have to obtain vitamin C from our diet. Clearly, then, if this proposal is correct, vitamin C biosynthesis must be downregulated–i.e., turned off. What advantage might this serve? Calabrese (1982) describes a probable survival advantage of having a broken GULO gene:
Considerable controversy exists over the role of ascorbic acid in maintaining health and resisting a wide variety of diseases including cancer. It has been contended that the evolutionary loss of ascorbic acid synthesis capability in man has enhanced the occurrence of numerous chronic diseases and was essentially a maladaptive alteration which initially was well tolerated because early man lived in a habitat which supplied foods with amounts of vitamin C equivalent to what they normally may have synthesized. However, as humans migrated into habitats with less availability of vitamin C, the adverse aspects of the loss of ascorbic acid synthesizing capability came to be demonstrated (1,2). In contrast, this paper proposes that the loss of an ability to synthesize ascorbic acid in humans, far from being a neutral or totally negative mutation, may have been a critical preadaptation which markedly enhanced the survival of early man with a G-6-PD deficiency living in a malarial infested environment. [emphasis added]
What survival advantage might this have imparted? The paper explains,
It is proposed that the loss of ability by humans to synthesize ascorbic acid may have markedly enhanced the survival opportunities of early man living in a malarial infested environment. This hypothesis is based on biomedical evidence which indicates that glucose-6-phosphate dehydrogenase (G-6-PD) deficient individuals display enhanced sensitivity to ascorbic acid induced hemolysis which has been fatal at sufficiently hiqh doses and that the G-6-PD deficient trait has been selected for in malarial environments.
What if the GULO gene is in fact functional in humans in utero and subsequently downregulated for good biochemical reasons like those described above? Indeed, in addition to these reasons, the highly toxic hydrogen peroxide is also known to be a biproduct of ascorbic acid biosynthesis (Banhegyi et al., 1997). Such a hypothesis, however, still requires us to propose a mechanism for turning on and off the GULO “pseudogene” as required. To put it more clearly, how can we explain how this GULO “pseudogene” becomes functional during human development? It is to this question that I now turn.
A Hypothesis for the Upregulation of the Human GULO Gene
As I mentioned previously, the GULO gene in humans is rendered inactive by multiple stop codons and indel mutations. These prevent the mRNA transcript of the gene from being translated into a functional protein. If the GULO gene really is functional in utero, therefore, presumably it would require that the gene’s mRNA transcript undergo editing so that it can produce a functional protein. It’s not at all difficult to understand how this could occur.
Editing of mRNA transcripts is a widely recognized phenomenon in biology today. The mRNA transcript would presumably need to have the stop codons replaced with amino-acid specifying codons (as well as other edits of course, for instance to correct frame shifts caused by indels). The occurrence of such a process is not unheard of. For instance, the transcripts of tRNA and rRNA genes typically undergo editing to render them functional. And the mRNAs for mitochondrial proteins (e.g. NADH dehydrogenase 7; cytochrome c oxidase III) in Trypanosoma brucei — the parasite that causes sleeping sickness — are known to undergo extensive editing, in many cases being substantially rewritten to produce functional transcripts (Ochsenreiter et al., 2008; Piller et al., 1996).
There are examples of known cases in the literature where RNA editing can result in the creation of premature stop codons, thereby stopping translation. Apolipoprotein B is one example of this (Morrison et al., 1996; Chan, 1995; Scott et al., 1989). It is made in the liver and intestine. The two mRNA transcripts in humans are almost identical, but the latter protein is smaller than the former. The only difference in the transcript is a single base change (cytosine in liver version; uracil in intestine version). This means that the intestinal transcript of this protein has the stop codon UAA instead of CAA (which codes for glutamine). Translation is thus stopped prematurely and a shorter protein produced.
Another example is the adenine deaminase enzyme ADAR2 where creation of a premature stop codon can silence the gene. Interestingly, this case of RNA editing is mediated by the adenine deaminase enzyme itself (Feng et al., 2006)! Adenine deaminase is responsible for conversion from adenine (A) to inosine (I). The ADAR2 transcript contains an AA sequence 47 bases upstream of the AG sequence ordinarily used as a 3′ splice site. When AA is edited to AI, it mimics AG, thereby adding a novel 3′ splice site, resulting in the retention of the 47 base sequence in the mRNA transcript (Rueter et al., 1999). Since the mRNA now possesses a premature stop codon, the result is a nonfunctional ADAR2 protein. This editing process is controlled in a tissue-specific manner such that it takes place frequently in the brain and lung but much less frequently in the heart.
In the case of GULO, however, stop codons are found in the DNA itself. DNA editing is known to occur at some level. For example, the mouse and human gene AICDA codes for a DNA cytidine deaminase enzyme — activation-induced cytidine deaminase — which is involved in somatic hypermutation, gene conversion, and class-switch recombination of immunoglobulin genes, and is also known to target non-immunoglobulin loci (Mu�oz et al., 2013; Gu et al., 2012; Kato et al., 2012; Morgan et al., 2004). I am not aware, however, of any examples of the kind of editing required to explain the existence of premature stop codons taking place in DNA. The most plausible model, therefore, seems to be that stop codons are expunged from the mRNA transcript when the gene is required to produce a functional protein.
Are there any documented cases where RNA editing leads to the removal of stop codons? Indeed, there are. An example would be the mitochondrial mRNAs of the lycophyte Isoetes engelmannii. A recent study reported that “mRNA editing affects 1782 sites, which lead to a total of 1406 changes in codon meanings. This includes the removal of stop codons from 23 of the 25 mitochondrial protein encoding genes,” and that “a total of 89 U-to-C editings are necessary to remove genomically encoded stop-codons within reading frames” (Grewe et al., 2011). If premature stop codons can be edited out of the mRNA transcript in this manner, thereby rendering functional what is otherwise a “pseudogene”, this seems to me to have profound implications regarding not only GULO, but our understanding of pseudogenes in general.
If the hypothesis entertained here turns out to be correct, it will be yet another example of where a paradigm of design has led to fruitful scientific research. Of course, for ethical reasons, the hypothesis cannot be easily tested on humans. Guinea pigs also possess a GULO pseudogene with premature stop codons, however, and would make a good model organism for the testing of this hypothesis.
Implications of this Evidence for the Debate over Darwinian Evolution
This article has presented a relatively simple model for how the vitamin C GULO “pseudogene” might be functional. While the gene normally yields an RNA transcript that cannot produce a protein product (because of premature stop codons and other mutations), RNA editing could modify the transcript so it could be translated into a functional protein. RNA editing is a widely recognized process that affects the transcripts of many genes. Given that vitamin C is found in elevated levels in human embryos (levels too high to be explained by dietary intake alone), and given that vitamin C is known to be used in the development of other mammals, this model could help explain how the GULO pseudogene is functional during human development. After embryogenesis is complete, the RNA editing is turned off, and the “pseudogene” no longer yields a functional protein. What are the implications of this hypothesis?
The GULO pseudogene has become one of the favorite arguments of Darwinian evolutionists against intelligent design. In their book The Language of Science and Faith, Francis Collins and Karl Giberson call this pseudogene “broken” DNA (p. 49). Because, in their view, it is “not remotely plausible” that “God inserted a piece of broken DNA into our genomes” (p. 49), the existence of pseudogenes has supposedly “established conclusively that the data fits a model of evolution from a common ancestor” (p. 43). Over at BioLogos, Dennis Venema has written article after article in the past few years, making pseudogenes his primary argument for evolution–one which he even suggests may be a “silver bullet” argument for evolution. The GULO pseudogene seems to be his favorite example.
Altogether, these results indicate that the discovery of pseudogene functions depends on the skills and methodologies used. Many expressed pseudogenes remain to be annotated. We believe that more and more functional pseudogenes will be discovered as novel biological technologies are developed in the future.
Another favorite pseudogene example of evolutionists has been the beta-blogin pseudogene. Earlier this year Casey Luskin reported that this icon evolution — the beta-globin pseudogene — now has evidence of function. Is it hard to believe that the same thing could happen to the GULO pseudogene, especially as we develop new technologies in the future?
- New technology suggests we’re just standing on the shores of understanding a vast ocean of potentially functional pseudogenes,
- It’s highly feasible to understand how this pseudogene might have function,
- Evidence suggests it might indeed have function during human development,
it would seem that Venema’s claim that the GULO pseudogene is a “silver bullet” argument for evolution is highly premature. An ID-based paradigm of biology can help us understand the functions of pseudogenes, whereas a neo-Darwinian paradigm leads us to wrongly assume these genes like the GULO “pseudogene” are “broken” genetic “junk.”