About 10,000 of these proteins are present in all cells (Wilhelm et al., 2015) and somewhere between 1500 and 2000 are derived from genes that are essential in the average human cell (Blomen et al., 2015; Wang et al, 2015; Hart et al., 2015) [see How many human protein-coding genes are essential for cell survival?].
Let's assume there are about 10,000 protein-coding genes that are expressed in a typical human cell. Does this mean that there are only 10,000 different proteins in those cells? The answer is "no" but the differences are often subtle. The activities of some proteins, for example, are regulated by covalent modification so a typical cell will contain different versions of the protein: some are modified and some are not (e.g. phosphorylated and non-phosphorylated). These would be genuine versions of different proteins although you probably wouldn't want to make a fuss about it.
In some cases, there are various intermediates produced during protein synthesis. For example, some proteins destined for the mitochondria have an N-terminal tag that's cleaved when the proteins reach their destination. There are two different versions of the protein but, again, this isn't really a big deal. We should really only count the steady-state, terminal, stage of processing and modification.
Similarly, there are proteins that are glycosylated in various ways and cells will always contain intermediates including non-glycosylated versions that have just entered the ER. These don't count as different versions of the protein.
Some genes are alternatively spliced to give proteins that have different internal amino acid sequences. These are genuinely different proteins produced from the same gene.
If you add up all the genuinely different versions of proteins produced from 10,000 protein-coding genes, how many proteins are present in a typical human cell?
Here's a standard answer given in a recent news article in Nature (Savage, 2015)
The human body contains roughly 20,000 genes that are capable of producing proteins. Each gene can produce multiple forms of a protein, and these in turn can be decorated with several post-translational modifications: they can have phosphate or methyl groups attached, or be joined to lipids or carbohydrates, all of which affect their function. “The number of potential molecules you can make from one gene is huge,” says Bernhard Küster, who studies proteomics at the Technical University of Munich in Germany. “It's very hard to estimate, but I wouldn't be surprised to have in one cell type 100,000 or more different proteins.”I suspect that Küster is one of those scientists who think that almost all human protein-coding genes are alternatively spliced to yield several different proteins in each cell. He has to imagine that there are, on average, ten different versions of a protein produced from each gene that's expressed in a typical cell.
That means ten different versions of each of the subunits of pyruvate dehydrogenase and RNA polymerase. It means ten different versions of triose phosphate isomerase and each of the ribosomal proteins. There should be ten different versions of actin and ten different versions of cytochrome c.
This seems very unlikely to me.
Discuss.
(There may be a few genes that have thousands of different variants, although I'm skeptical. In that case there may be 100,000 different proteins in a human cell but surely this is misleading even if it's accurate?)
Blomen, V.A., Májek, P., Jae, L.T., Bigenzahn, J.W., Nieuwenhuis, J., Staring, J., Sacco, R., van Diemen, F.R., Olk, N., and Stukalov, A. (2015) Gene essentiality and synthetic lethality in haploid human cells. Science, 350:1092-1096. [doi: 10.1126/science.aac7557 ]
Hart, T., Chandrashekhar, M., Aregger, M., Steinhart, Z., Brown, K.R., MacLeod, G., Mis, M., Zimmermann, M., Fradet-Turcotte, A., and Sun, S. (2015) High-Resolution CRISPR Screens Reveal Fitness Genes and Genotype-Specific Cancer Liabilities. Cell 163:1515-1526. [doi: 10.1016/j.cell.2015.11.015]
Kim, M.-S., Pinto, S.M., Getnet, D., Nirujogi, R.S., Manda, S.S., Chaerkady, R., Madugundu, A.K., Kelkar, D.S., Isserlin, R., Jain, S., Thomas, J.K., Muthusamy, B., Leal-Rojas, P., Kumar, P., Sahasrabuddhe, N.A., Balakrishnan, L., Advani, J., George, B., Renuse, S., Selvan, L.D.N., Patil, A.H., Nanjappa, V., Radhakrishnan, A., Prasad, S., Subbannayya, T., Raju, R., Kumar, M., Sreenivasamurthy, S.K., Marimuthu, A., Sathe, G.J., Chavan, S., Datta, K.K., Subbannayya, Y., Sahu, A., Yelamanchi, S.D., Jayaram, S., Rajagopalan, P., Sharma, J., Murthy, K.R., Syed, N., Goel, R., Khan, A.A., Ahmad, S., Dey, G., Mudgal, K., Chatterjee, A., Huang, T.-C., Zhong, J., Wu, X., Shaw, P.G., Freed, D., Zahari, M.S., Mukherjee, K.K., Shankar, S., Mahadevan, A., Lam, H., Mitchell, C.J., Shankar, S.K., Satishchandra, P., Schroeder, J.T., Sirdeshmukh, R., Maitra, A., Leach, S.D., Drake, C.G., Halushka, M.K., Prasad, T.S.K., Hruban, R.H., Kerr, C.L., Bader, G.D., Iacobuzio-Donahue, C.A., Gowda, H., and Pandey, A. (2014) A draft map of the human proteome. Nature, 509:575-581. [doi: 10.1038/nature13302]
Savage, N. (2015) High-protein research. Nature 527:S6-S7. [doi: 10.1038/527S6a]
Wang, T., Birsoy, K., Hughes, N.W., Krupczak, K.M., Post, Y., Wei, J.J., Lander, E. S., and Sabatini, D.M. (2015) Identification and characterization of essential genes in the human genome. Science, 350:1096-1101. [doi: 10.1126/science.aac7041]
Wilhelm, M., Schlegl, J., Hahne, H., Gholami, A.M., Lieberenz, M., Savitski, M.M., Ziegler, E., Butzmann, L., Gessulat, S., Marx, H., Mathieson, T., Lemeer, S., Schnatbaum, K., Reimer, U., Wenschuh, H., Mollenhauer, M., Slotta-Huspenina, J., Boese, J.-H., Bantscheff, M., Gerstmair, A., Faerber, F., and Kuster, B. (2014) Mass-spectrometry-based draft of the human proteome. Nature, 509:582-587. [doi: 10.1038/nature13319]
How many alternatively-spliced proteins are actually known (and the alternatives known to be functional)? I only know of one, MYC, which has two forms, but I only know about it because I happen to have sequenced it. Is there a review of functional alternative splicing in some taxon or taxa?
ReplyDeleteGenes are not spliced in the course of gene expression.
ReplyDeleteJohn, perhaps the first comment to this post has a tiny bit of what you are looking for.
ReplyDeleteInteresting. Are you a big fan of Arabidopsis? Better yet, is there a review of alternative splicing in Arabidopsis so I don't have to wade through those various references? Finally, any estimate of the average number of functional transcripts per gene?
DeleteI don't know about plants but in mice it's supposed to be one major transcript per gene.
DeleteGonzàlez-Porta M, Frankish A, Rung J, Harrow J, Brazma A. Transcriptome analysis of human tissues and cell lines reveals one dominant transcript per gene. Genome Biol. 2013 Jul 1;14(7):R70
Sorry, the aboven mentioned paper is on human tissues as the title actually states. Working with mice let me forget men sometimes.
DeleteEven that paper has a significant "miss" in the results section when they get hung up on the fact that "Genes with major non-coding transcripts are expressed at higher levels in the nucleus, compared to those with major coding transcripts, while this trend is inverted in the cytosol"
DeleteSaying an one form of an mRNA is "expressed" in the nucleus and a different form is "expressed" in the cytosol is biochemical nonsense. They do address this somewhat better in the discussion, but it shouldn't have needed discussion! Of _course_ they find more retained introns when they look at immature hnRNA in the nucleus compared to mature transcripts in the cytosol where they are translated.
There are something like 150 different post-translational modifications known to be found among the core histones alone, i.e. around 30 different marks for each of histones 2A, 2B, 3 and 4. If these marks were completely independent of each other, then the potential complexity is 2^30 ~= 1 billion different potential isoforms of _each_ of the core histones.
ReplyDeleteObviously the true complexity is not that high since many marks will be mutually exclusive, and other marks will be strongly correlated with each other.
However, given the explosive variety afforded by _combinatorial_ post-translational modification, I would not find it remotely surprising to see thousands or even tens of thousands of subtly different variant histones present in the cell in varying proportions at different times and in response to different stimuli.
Ironically, this is the same mistake made in the opposite sense by people like John Mattick - they assume that a huge number of genes is necessary to specify the uncountable number of of different heritable phenotypes, missing the point that with just a handful of different transcription factors working together you could trivially end up with a unique combination for every individual cell in the body of every individual animal on the planet.
Human intuition is REALLY BAD at maths, particularly combinatorics and exponential functions.
Now, in regard to the histone code, the question of how much of the variation is functionally relevant is a very different story. Sticking an acetyl group onto one lysine or the next door lysine is, strictly speaking, a different modification. However, the net effect of adding a charged residue to a particular region of the histone surface may well be much the same. Likewise for crotonylation versus acetylation - do the readers/writers/erasers really care about the difference, or do they just say "stick a negative charged group of about this size somewhere near here"?
At the moment I'd say we're not doing that well at working out which differences are meaningful and which are not, and that's because cells sometimes care about single atom differences, and sometimes they don't! (Don't remind me of the time I mistakenly ordered labelled UTP instead of labelled dUTP as a student...)
I'm focusing on the histone code there because Larry says that PTMs are "genuine versions of different proteins although you probably wouldn't want to make a fuss about it", and also because unlike Larry's example of N-terminal signal peptides, there is no final terminal "steady state" for histones.
DeleteTo the "wouldn't want to make a fuss about it" part all I can say is that H2AX and gamma-H2AX do very different jobs, as do H3K9 mono- or di-methylated isoforms. Every modification potentially has a different meaning, but we don't yet know what they are.
Realistically, I guess you could set an upper bound on the meaningful diversity of the histone code by working out how many "readers" there are and what their specificities are (no I don't know the number and nor does anyone else). Then add on a few for the modifications that don't have direct readers but are meaningful in regulating nucleosome removal, replacement etc.)
In the case of histones these modifications may have been selected for. However, what if the machinery responsible for certain modifications alters other proteins which just happen to contain modifiable sequences/motives but the modification doesn't interfere with their functions? I don't know how likely this scenario is but wouldn't that be noise which is not selected against and thus retained by the cell?
DeleteOh certainly. I'm sure there are whole fields of new noise waiting to be discovered. Just because kinase A phosphorylates target B, that doesn't mean there's any functional relevance to it. The "real target" (i.e. with a selected function) could be protein C, and the effect on B just an irrelevant byproduct. Unfortunately there's no easy way to tell the difference - checking if the PTM site on B is conserved across species is one useful approach, but won't tell you about species-specific biology.
DeleteAnd then there's the perennial thorn of what counts as a 'function'. Many phenotypes will have little to no selective effect even if there is a causal role for particular gene variants. While the "selective effect" definition of gene function is appropriate for evolutionary biologists, physiologists and medics need to use the "causal role" definition - under which many more protein variants will be classed as functionally different.
If a PTM is reversible, should we really consider them a different version of a protein?
ReplyDeleteI think one should not expect too many functional alternative transcripts. Otherwise conditional gene targeting which is based on Cre mediated deletion of single exons to inactivate won't work and we would have seen hundreds of publications describing how murine phenotypes were altered by alternative transcripts. There are such cases but by heart I remember only one:
ReplyDeleteCaution! Analyze transcripts from conditional knockout alleles
I guess I should have further explained why I think why data from conditional mouse mutants rather contradict the importance of alternative transcripts encoding many different variants of single proteins.
ReplyDeleteTo inactivate a gene conditionally one flanks an exon of the gene with loxP sites. Mice carrying such "floxed" alleles are then bred with mice that contain transgenes of Cre recombinase which will recognize the loxP sites and delete the sequence which resides between the later i.e., the floxed exon. Depending on the promoter used to express Cre one can limit expression of the recombinase to certain tissues resulting in tissue specific inactivation of the floxed alleles.
The critical step in designing conditional alleles is the decision which exon should be flanked by loxP sites i.e., later deleted to reliably inactivate the gene. Rather than looking for functionally important exons such as those encoding active centers of enzymes exons flanked by differently phased introns are chosen (intron in phases 0, 1 and 2 interrupt the reading frame of the mature mRNA between codons or after the first or the second nucleotide of a codon). Deleting such an exon will result in frame shift in the mRNA which in many if not most cases leads to the formation of a pre-mature termination codon (PMTC) in one of the downstream exons. If the PMTC is located more than 50 nucleotides upstream of the next splice donor site it will be recognized by the NMD machinery and the mRNA will undergo non-sense mediated mRNA decay i.e., it will disappear completely. Thus, one prevents the proper maturation and expression of the mRNA of the gene. Since hardly any mRNA of the gene is detectable it seems unlikely that alternative spliced transcripts that could remove the PMTC by introducing a curative phase shift are expressed in relevant amounts. I have to admit though, that researchers usually rather look for protein than mRNA expression. However, if alternative splicing would give rise to alternative peptides frequently there should be hundreds of reports of such observations.
As I mentioned above the decision which exon will be deleted is mostly based on the intron phases of the gene of interest (the approach obviously doesn’t work if all introns are in the same phase). Thus, the selection of the exon to be floxed is rather arbitrary and doesn’t relate to the function of the peptide sequence encoded by this exon. Thus, counter arguments stating that the chosen exon is indispensable and thus its deletion doesn’t say much about the relevance of alternative transcripts appear not convincing.