Inactivation of CMP-N-acetylneuraminic acid hydroxylase occurred prior to brain expansion during human evolution

Abstract

Humans are genetically deficient in the common mammalian sialic acid N-glycolylneuraminic acid (Neu5Gc) because of an Alu-mediated inactivating mutation of the gene encoding the enzyme CMP-N-acetylneuraminic acid (CMP-Neu5Ac) hydroxylase (CMAH). This mutation occurred after our last common ancestor with bonobos and chimpanzees, and before the origin of present-day humans. Here, we take multiple approaches to estimate the timing of this mutation in relationship to human evolutionary history. First, we have developed a method to extract and identify sialic acids from bones and bony fossils. Two Neanderthal fossils studied had clearly detectable Neu5Ac but no Neu5Gc, indicating that the CMAH mutation predated the common ancestor of humans and the Neanderthal, approximately 0.5-0.6 million years ago (mya). Second, we date the insertion event of the inactivating human-specific sahAluY element that replaced the ancestral AluSq element found adjacent to exon 6 of the CMAH gene in the chimpanzee genome. Assuming Alu source genes based on a phylogenetic tree of human-specific Alu elements, we estimate the sahAluY insertion time at approximately 2.7 mya. Third, we apply molecular clock analysis to chimpanzee and other great ape CMAH genes and the corresponding human pseudogene to estimate an inactivation time of approximately 2.8 mya. Taken together, these studies indicate that the CMAH gene was inactivated shortly before the time when brain expansion began in humankind's ancestry, approximately 2.1-2.2 mya. In this regard, it is of interest that although Neu5Gc is the major sialic acid in most organs of the chimpanzee, its expression is selectively down-regulated in the brain, for as yet unknown reasons.

Figure 1

Sias from bones and fossilized bones. Sias were released and purified, derivatized with DMB, and analyzed by HPLC. The elution positions of standard Neu5Ac and Neu5Gc are indicated. (A) Contemporary bones of humans and great apes. (B) Pleistocene and Miocene fossils. 1, mammoth tusk, Holocene, ≈1,000 years ago (kya); 2, cave bear jaw, Pleistocene, 40–80 kya; 3, deer leg bone, Pleistocene, 40–80 kya; and 4, dugong femur, Miocene, ≈20 mya.

Figure 2

Example of mass spectrometric proof of Sias isolated from fossils. DMB-derivatized Sias from a Pleistocene cave bear jaw (B2 in Fig. 1) were separated by C18 HPLC, peak areas corresponding to the expected elution times of Sias collected, concentrated, and analyzed by mass spectrometry. (Upper) m/z profiles obtained at the elution times of the DMB derivatives of Neu5Ac and Neu5Gc. The ions corresponding to hydrated DMB derivatives of Neu5Ac and Neu5Gc (426 and 442) are not easily seen in the presence of contaminating ions of higher intensity. (Lower) The outcome of secondary mass spectrometry performed on the above ions, which each lose one molecule of water, giving DMB-Neu5Ac (408) and DMB-Neu5Gc (424). Variable amounts of the parent ion are also seen in the secondary spectrum.

Figure 3

Sias from Neandertal fossils. Sias were released, purified, and characterized as in Fig. 2. (A) HPLC profile of extract from the original Neandertal-type specimen. The wash through blank is a control for human handling. The column was deliberately overloaded to take the Neu5Ac peak off scale, highlighting the absence of Neu5Gc. (B) HPLC profile of Sias from Georgian Neandertals. The absence of Sias from sample ID no. 2024 was confirmed by MS analysis. (C) MS of Neu5Ac peak from sample ID no. 486 showing m/z corresponding to the hydrated (426) and dehydrated (408) DMB derivatives of Neu5Ac. There were no detectable ions corresponding to Neu5Gc derivatives (442 or 424).

Figure 4

Unrooted tree of AluYb8s. Eighty-six intact AluYb8 elements that are human-specific and fixed in human populations (29) were analyzed as described in Materials and Methods. SahAluY is the Alu that inactivated the CMAH gene, and msAluY is most similar to it.

Figure 5

Phylogenetic tree of hydroxylase sequences built by the maximum parsimony method. Individual nucleotide substitutions on each branch (based on ancestral sequence deduced from rhesus monkey) are indicated and designated S or N, depending on whether the nucleotide substitution is a synonymous or nonsynonymous change, respectively (chronological order is not implied in the order of each listing). The rhesus monkey sequence (GenBank accession no. AB013814) was used as an outgroup to deduce an ancestral sequence of orangutan, gorillas, chimpanzees, and humans. A change that induced a new stop codon at the C terminus of the gorilla Beta sequence was observed and excluded from the analysis.