This question is based upon a wrong inference about the work that forms the basis of the National Geographic article, which includes this statement:
All species in all three domains share 23 universal proteins, though the proteins' DNA sequences—instructions written in the As, Cs, Gs, and Ts of DNA bases—differ slightly among the three domains (quick genetics overview).
The 23 universal proteins perform fundamental cellular activities, such as DNA replication and the translation of DNA into proteins, and are crucial to the survival of all known life-forms—from the smallest microbes to blue whales.
The article is referring to a paper published in Nature in 2010.
http://www.ncbi.nlm.nih.gov/pubmed/20463738
This paper uses a set of 23 proteins to test the idea that certain genes arose only once.
The data set consists of a subset of the protein alignment data from ref. 27, containing 23 universally conserved proteins for 12 taxa from all three domains of life, including nine proteins thought to have been horizontally transferred early in evolution27. The conserved proteins in this data set were identified based on significant sequence similarity using BLAST searches, and they have consequently been postulated to be orthologues.
There is no suggestion that there are only 23 such proteins, it's just that the author chose this subset to test his ideas.
And here they are:
alanyl-tRNA synthetase, aspartyl-tRNA synthetase, glutamyl-tRNA synthetase, histidyl-tRNA synthetase, isoleucyl-tRNA synthetase, leucyl-tRNA synthetase, methionyl-tRNA synthetase, phenylalanyl-tRNA synthetase β subunit, threonyl-tRNA synthetase, valyl-tRNA synthetase, initiation factor 2, elongation factor G, elongation factor Tu, ribosomal protein L2, ribosomal protein S5, ribosomal protein S8, ribosomal protein S11, aminopeptidase P, DNA-directed RNA polymerase β chain, DNA topoisomerase I, DNA polymerase III γ subunit, signal recognition particle protein and rRNA dimethylase