Wikipedian

I love Wikipedia. No, you shouldn’t use it for academic research, but that doesn’t mean it should be avoided altogether. I adored my printed encyclopaedias as a child and read them back-to-back, sucking in as much knowledge as I could. Our collective knowledge changes over the years. Science is especially fluid. I’ve looked back at my old encyclopaedias and found many incorrect “facts”. This pleases me greatly as it means I absorbed the knowledge at the time and kept up-to-date on those cases. Wikipedia is the perfect encyclopaedia in my opinion because it’s as fluid as our knowledge. Sure, someone might incorrectly edit a page, but there are plenty of obsessive individuals (Wikipedians) willing to research the changes and maintain the quality of the content. I’ve finally became one of those obsessives.

I use Wikipedia as a dictionary. When I don’t know what something is, I search for it on Google. The Wikipedia article is usually among the first results and I’m usually happy with that. Even if there are mistakes somewhere on a whole Wikipedia page, the entry usually does a better job than any dictionary when it comes to quickly understanding the meaning of a word or name. Who the hell was Émile Combes? Oh, he was the Prime Minister of France. That’s what I use Wikipedia for and it’s usually a brief visit. But occasionally, I stay for longer during some free time or procrastination and see where the links take me. Prime Minister of France? Who succeeded him? Oh, what’s this about the 1905 French law on the separation of Church and State? It was on one of these convoluted trips through Wikipedia that I briefly ended up on the Spliceosome page and something caught my eye…

 

Hold up. What’s this? Located in the cytosol? Hmm. I predicted the reference before I had even scrolled down. I was right, a very controversial paper by König et al (2007). There are so many problems with that claim, including the grammar, and it isn’t as straightforward as it seems. König et al found that the minor spliceosome functions outside the nucleus, speculating that pre-mRNA splicing can continue to occur during mitosis. But there are many reasons to be skeptical about this claim.

Our DNA alone doesn’t just build proteins. There are many steps along the way from your genes to proteins. Your DNA acts as a template for transcribing mRNA. Ultimately, the mRNA leaves the nucleus to be translated into proteins. But it’s far more complicated than that. The pre-mRNA has to be processed before it can leave the nucleus as mature mRNA. There are many steps during this post-transcriptional processing, but we’re focusing on splicing. The pre-mRNA contains introns, non-coding sequences, that are removed by splicing machinery known as the spliceosome.

A spliceosome is a complex containing snRNAs and associated proteins. There are different types of introns, sometimes referred to as the minor- and major-class introns. These are removed by two different splicing systems, the minor and major spliceosomes. The major spliceosome contains U1, U2, U4, U5 and U6 snRNAs. The biogenesis of the snRNPs required for the spliceosome involves sending some of these snRNAs briefly outside the nucleus where they are modified with various proteins. Importantly, they are given a m₃G cap and other modifications that allow the snRNAs to be reimported back into the nucleus. This major spliceosome removes the major-class intros from the pre-mRNA, bringing it a step closer to being mature mRNA that can leave the nucleus and be translated.

The minor spliceosome gets its name because it is only responsible for splicing 1 in 300 introns. The splicing rate of the minor spliceosome is also far slower than the major spliceosome. But at the end end of the day, it’s remarkably similar. The main difference is that it is built with minor-class snRNAs: U11, U12, U4atac, U6atac, and U5 (common to both spliceosomes). Here’s where I get to my point. Previous research has found that minor splicing occurs in the nucleus, just like major splicing, except for one paper by König et al… the very paper cited on the Wikipedia article.

When researchers look for the minor spliceosome, it’s in the nucleus. The minor-class snRNAs? In the nucleus. Proteins associated with the minor spliceosome? In the nucleus. Independent researchers have found these results time and time again, in different species and tissues. But König et al contradicted the large body of prior research to find that minor splicing occurs in the cytoplasm. Your first question might be, what kind of crap journal published it though? Cell. So what the hell is going on?

König et al performed various experiments, and it all seems ok at a first glance. They claim that the two spliceosomes have distinct subcellular localisation and that this may allow minor splicing to occur during mitosis. They make some sense and the logical progression through the paper is sound, if you accept each of their steps. The first thing they did was in situ hybridisation (ISH) experiments on zebrafish tissues at various developmental stages to detect distinct subcellular localisation of minor- and major-class snRNAs. Their results revealed strong nuclear staining for the major snRNAs, but probes for the minor snRNAs showed cytoplasmic or perinuclear staining, with weak nuclear staining. At first glance it seems legit, they used controls, the data and figures seem fairly convincing at first (though very representative). They then went on to perform ISH on primary human fibroblasts and murine NIH 3T3 cells, again finding minor snRNAs in the cytoplasm.

Ok, so far so good. Or is it? Firstly, they only showed data for two minor-class, one major-class, and the U5 snRNAs. What about the others? Far more importantly, no Northern analyses were provided to confirm the specificity of the probes. Also, the control anti-U2 and anti-U5 probes should selectively stain the nucleus, excluding the nucleolus. But in their data, the nucleolus is also stained. They’re doing something wrong and at least some of these results are certainly not reliable.

Next, they performed cell fractionation experiments and found U2 in the nucleus, but U6atac in both the nucleus and the cytoplasm. They only looked at those two, and only briefly addressed the U6atac findings by speculating the snRNA associates with outer nuclear structures that may copurify with the nuclear fraction.

You may have already guessed the next experiment. If minor splicing occurs outside the nucleus, then transcripts with unspliced minor-class introns must be able to leave the nucleus. König et al inhibited minor-class splicing, predicting that these transcripts will begin to accumulate in the cytoplasm. Using fractionated NIH 3T3 cells, p120 was tested for transcripts containing minor- and major-class introns. Unspliced major-class introns were weak in the nucleus (due to their rapid splicing) and weren’t found in the cytoplasm. Unspliced minor-class introns were mostly found in the nucleus, and König et al explained this as due the lack of minor-class splicing in the nucleus. They overlooked something that seems obvious to me. If the spliceosomes splice at different rates, wouldn’t that alone cause differences in accumulations? König et al claimed their results couldn’t be the result of leakage, as major-class introns didn’t leak. But if they are spliced at different rates, there could be more minor-class introns remaining in the nucleus potentially ready to leak out, compared to major-class introns.

Many of the results are explained poorly and very briefly before moving onto the next experiment. The paper is filled with statements followed by “data not shown”. Every paper I’ve ever read that has used ISH and found minor splicing to occur in the nucleus has performed Northern analyses to confirm the specificity of the probes that were used. The only paper I’ve ever seen on the subject that didn’t include Northern analyses is this one by König et al, and they are the only people to get results that contradict all previous research. I think there’s good reason to doubt.

The next year, Pessa et al published a paper claiming that both the major and minor spliceosomes act in the nucleus, as expected. This paper was a direct rebuttal to König et al and covered every stone that they had left unturned. The experiments were specifically designed to copy König et al but be more rigorous. They highlighted the obvious problems with the controversial paper, such as the biogenesis of the snRNPs. Earlier I mentioned that the major snRNAs are exported from the nucleus briefly, and modified in special ways so that they can be reimported into the nucleus. The minor snRNAs get the same modifications. If they function in the cytoplasm, why do they receive everything they need for reimport into the nucleus? Also, U6 (major) and U6atac (minor) are exceptions in that they are transcribed by RNA Pol III (rather than RNA Pol II) and don’t leave the nucleus at all. They both acquire a γ-monomethyl phosphate cap and Lsm proteins, which are found in the nucleus. They both do. Both. The U6atac snRNA, part of the minor spliceosome, forms a complex with proteins in the nucleus. Basically, the biogenesis of the various components of the minor spliceosome parallels the biogenesis of those for the major spliceosome. It seems reasonable to suggest they function similarly as well.

Pessa et al did ISH experiments too, on embryonic and mouse tissues, but they used multiple detection methods and confirmed the specificity of the probes using Northern hybridisations. They found both the minor and major snRNAs in the nucleus. They also performed fluorescence in situ hybridisation (FISH) using HeLa SS6 cells finding that the major and minor snRNAs colocalise in the nucleus. Again, the probes were confirmed by Northern analysis.

To confirm the ISH/FISH results, Pessa et al assayed for spliceosomal components in cytoplasmic nuclear fractions prepared from the same batch of HeLa cells. RNA and protein was isolated from an identical cell-equivalent amount of each extract and analysed using Northern or Western blotting. The majority of U1, U2, U4 & U5 snRNAs were found in nuclear extract. U6 and U6atac were also almost equal between nucleus and cytoplasm. This seems surprising, but is consistent with previous studies. U6 & U6atac must differ from the others or somehow more readily leak out of the nucleus during cell fractionation. Western analysis confirmed localisation of major & minor snRNPs predominantly in the nuclear extract. In contrast, Lsm1 was found predominantly in the cytoplasm as you would expect. Admittedly, the paper by Pessa et al is designed to counter the controversial claims and does avoid some of the more genuinely interesting results from the König paper, including the similarities between embryos with suppressed minor splicing and the early-arrest mutants from previous studies (Kane et al, 1996). Pessa et al might have been very biased, and ignored a truly meaningful result, but for the most part they did highlight all the experimental problems made by König et al.

So, doing the experiments properly, researchers seem to agree on nuclear localisation of both splicing systems. One paper made a drastic departure from the scientific consensus, and sometimes this results in leaps forward, but this happens to be the paper containing many problems. The paper appears rushed. Data is missing. Essential experimental procedures aren’t followed. Anomalies in the data are barely addressed, some that actually could be taken as evidence for the consensus view. The results were undeniably interesting, but we’ll need to see more rigorous research from König before accepting his results.

So my first major edit on Wikipedia was fairly simple. The page originally claimed that the minor spliceosome functions in the cytosol. My revision explains that both function in the nucleus. But I decided to add some new information to the page too, because of some weird exceptions that might just influence your view of König’s paper, and to make things interesting. There are a few strange exceptions where splicing does occur outside the nucleus in highly specialised cells. Denis et al (2005) found that splicing can take place in anucleate platelets. They used a model of human megakaryocyte differentiation and proplatelet formation and found spliceosomal components in the cytoplasm, which are distributed to platelets during thrombopoiesis. Analysis of primary human platelets revealed that snRNAs and spliceosomal proteins accumulate in these anucleate cytoplasts. Mature platelets were shown to splice interleukin-1β (IL-1β) into a mature message in response to cellular activation, resulting in synthesis of the IL-1β protein. Weird!

There’s another exception. The dendroplasm of neuronal cells is known to contain cellular machinery required for RNA & protein metabolism (link to PDF). Glanzer et al (2005) conducted several experiments detecting localisation of pre-mRNA-splicing factors. More directly, the researchers introduced pre-mRNA constructs into intact, isolated dendrites and observed splicing. Some of the dendritically spliced mRNAs were translated, resulting in detectable protein. Awesome!

So, keep an open mind on splicing. I felt these exceptions should be mentioned on the Wikipedia entry to make it clear that splicing can technically occur in the cytoplasm of some cells. But understand that even if you put the splicing machinery in the cytoplasm of other cells, it doesn’t work. Actually, the only way we can force splicing to occur in the cytoplasm of other cells is by adding various components from the nucleus. What does that tell you? But in some weird places, splicing can occur outside the nucleus. Do consider those examples. Neuronal cells need to perform activities far from the nucleus that would normally take place in the nucleus of other cells. As for other cells, in various tissues, in species including mice, zebrafish and humans, at different developmental stages, all research has found that both spliceosomes function in the nucleus. Well, all except for a single a paper. Judge the quality of the paper for yourself. But to let Wikipedia state that minor splicing occurs in the cytosol, we have to ignore 99.9% of the research. We also have to ignore the various problems with the paper by König et al.

After digging through the edit logs and history of the page, I was surprised to find that the incorrect statement had been sitting there for 5 years! The date actually explains a lot… when the paper was published, someone added the information to the Wikipedia page, assuming it was correct. Here’s how the page looks now:

Keep an open mind, we still don’t know everything about splicing. But when it comes to the evidence, there’s no good reason to think the minor and major spliceosomes show different subcellular localisation. So this is my first adventure into the world of editing Wikipedia. Do I get some kind of pedantry award?