A Wonderland of RNA

A snoRNA mates with the 7SL RNA and mRNA to promote protein secretion.

As molecular biologists wander through the wilderness of the cell, they keep stumbling across RNAs. From early on, the ribosomal RNA (rRNA) and amino acid transfer RNA (tRNA) were obviously incredibly abundant, in their somewhat inefficient job of carrying on translation. Messenger RNAs (mRNA) were less abundant, but recognized from the start for their key role relaying information from the genome. But over the decades, more and more types of RNA kept popping up. Here is one tabulation of genes by type in humans:

• 22,700 protein-coding (along with 19,000 derelict "pseudogenes")
• 820         rRNA
• 659         tRNA
• 1,960 miRNA
• 2,100 snRNA
• 1,390 snoRNA
• 51,000 ncRNA
• 65         scaRNA, scRNA, piRNA

One big step in the realization of the prevalence of RNA was the ENCODE project, done as part of the human genome project. They found that most of the genome is transcribed to RNA, one way or another. Not all those products are important, or abundant, but just the fact that all this RNA is floating around was startling. This does not mean that there isn't junk DNA, (or junk RNA), but it does mean that a lot of potential function lurks waiting to be found. And the last couple of decades have seen many such finds. 

From the list above, microRNAs are small fragments that bind to matching mRNAs and repress their translation to protein. They have wide-ranging networks of regulation, mostly of a fine-tuning nature, but sometimes quite decisive and relevant to human biology and pathology. snRNAs are small nuclear DNAs, some of which function in RNA splicing. snoRNAs are small nucleolar RNAs, some of which mate with various sections of the ribosomal RNA as it is being assembled in the nucleolus, and guide chemical modifications made by enzymes, such as attachment of methyl and uridine groups. The non-coding (nc) RNAs are typically products of protein coding genes that, due to splicing or altered start sites, happen to not code for anything, and occasionally have significant regulatory roles. 

In general, RNAs may have a few different mechanisms of action: guide characteristics, where they mate with their antisense sequence in a target RNA and direct some other process like sequestration, cleavage, or chemical modification. Or they may bind to specific proteins, such as the RNAs that bind to chromatin and regulate X-linked dosage compensation. Or they have structural, even catalytic roles, like the ribosomal and spliceosomal RNAs. 

What should be clear that there are many more genes are recognizable by sequence than we understand. Only a couple hundred snoRNA genes are understood by their targets and activity. But there are well over a thousand in the genome. What do the rest do? A recent paper took on this quest, devising a novel way to isolate these snoRNAs and their partners from the welter of other material. They did this by crosslinking everything, ligating the RNAs locally to each other (which linked the snoRNAs to their targets) and then reverse-transcribing the RNAs before trying to capture them individually by custom anti-sense DNA probes, one per gene. It was a complicated procedure, but far more productive than trying to capture them directly as RNA with antisense RNA probes, since these snoRNAs are intensely structured (lots of hairpins and other duplexes) and expected to be tightly bound to other things.

Taking the most abundant snoRNAs, these researchers then looked for novel partners and functions. After seeing that they recovered plenty of the known interactions, the most interesting novel interaction they came up with was of a gene called SNORA73. This was found linked to two other RNAs, 7SL RNA and various mRNAs. 

Just another holdover from the RNA world. The SRP particle (in red) is built around the 7SL snRNA (helix). This particle detects the signal peptide (green) of the nascent protein emerging from the ribosome (beige, blue), and clamps on (right) to arrest translation. Translation is later resumed after the whole complex has successfully docked with the membrane receptor, allowing the SRP to be released, and the peptide to be threaded through the membrane. 

Funny story ... 7SL RNA is yet another snRNA that has a key role in translation. It is the core of the signal recognition particle (SRP), which binds to "signal" sequences in proteins as they come off the ribosome. These are a special code segement at the start that says "I want to be secreted across (or into) a membrane, not just located in the cytoplasm". The SRP captures this signal segment, and then sticks its head into the ribosome, stalling its translation. Then the whole mess goes off to the membrane (endoplasmic reticulum in eukaryotes, or plasma membrane in bacteria) where it docks with the SRP receptor complex. This is the signal for translation to restart, the SRP to come off, and the nascent protein to thread its way through the membrane to the other side. 

Incidentally, it is notable also that SRP is scaffolded by a large RNA, with a few proteins stuck on for decoration / specificity. This makes sense as an echo of early evolution, where not only did RNAs likely arise before proteins ever existed, but those RNAs had gotten quite large while the earliest proteins were still relatively small. The genetic code appears to have started as a two letter code, before the third letter was munged onto the end, vastly expanding the chemical repertoire of proteins and making them premier catalysts. 

A few results, indicating that knockdown of SNORA73 (with the anti-RNAs LNA-1 and LNA-2) dramatically decreases secretion of the proteins CLU and LGAL3BP. On left  are signals from proteins isolated from inside and outside the cells, as indicated. On right is a graph of the same data. The mRNA levels are not changed nor the protein levels. Only the level of secretion is altered.

So the implication of all this was that SNORA73 affects protein translation/secretion. This is indeed the case, when these authors assayed the secretion of one of the SNORA73-bound mRNA-encoded proteins in the presence of an inhibitor of SNORA73 (above). The mechanism is that SNORA73 serves as a special glue between the 7SL snRNA and the translating mRNA, with parts of its RNA sequence complementary to both a segment of the 7SL snRNA, and also to a small 10 base-long segment of the mRNA. The mRNA segment is hanging off the ribosome while the beginning of the message is being translated. The whole setup helps the SRP find these mRNAs efficiently and hold on to them effectively, increasing not their translation rate, but their secretion rate.

Models of the structures of SNORA73 (which is made by a pair of similar genes, A and B), as they bind to the 7SL snRNA, and the target mRNAs. These binding areas are far apart, to allow the mRNA tail (that is not yet in the ribosome) to reach the MBM binding site. The psi pocket is of uncertain function, but in other snoRNAs directs the uridine addition to target rRNA.

The mRNAs that have this 10 base (MBM) signal that binds to SNORA73 are a subset of those that express secreted proteins, though it is not really clear from this work what kind of a subset this is. Perhaps this mechanism makes up for weak signal sequences, or some other defect in the protein's access to the secretion machinery. Whatever that logic, we have here a conjunction of four RNAs, (7 SL snRNA, the SNORA73 snoRNA, the mRNA target, and the ribosomal RNA structure) all collaborating to promote the secretion of a target protein. This is just one of thousands of uncharacterized and conserved RNAs visible in our genome. It is startling to think what else might be going on.

• PLOS goes on a mole hunt.
• We are going to be without key vaccines quite soon.
• Is lithium a vital nutrient?
• Doctors are fed up with a system they created.
• Credit isn't free- Krugman on predatory finance.
• Like a battered wife, Trump keeps going back for more abuse.