Uncommon Descent Serving The Intelligent Design Community

Micro RNAs and Design inference.

Share
Facebook
Twitter
LinkedIn
Flipboard
Print
Email

http://www.nature.com/nrm/journal/v9/n9/pdf/nrm2472.pdf

“MicroRNAs (miRNAs) are known to regulate gene expression at the level of translation, but how does this affect what proteins are produced? Two recent papers have shown that individual miRNAs can affect the expression of hundreds of proteins.

One known as miR-223 seems to function as a rheostat to finely adjust protein output.

Another miRNA let-7b is can fine-tune protein production from thousands of genes.

Individual miRNAs can have an effect on global protein expression, and protein repression is likely to be mediated by a specific complementary sequence target region in the corresponding messenger RNA. The emerging picture is that miRNAs do not have just a small number of mRNA targets — they function at the transcriptional and translational level to subtly modulate what specific proteins are produced.”

Do the exact complimentary target sequences for these miRNAs being present on many mRNAs coding for different proteins, mean that all these sequences and proteins have a common ancestor, or were the target sequences derived independently?

The study of miRNAs and their target mRNAs and proteins may be good ID research, as I suspect that many unrelated proteins are controlled by the same miRNA. Because the target sequence is the same, that seems to imply strongly that design, rather than descent with modification, was involved.

http://www.nature.com/nrm/journal/v9/n9/pdf/nrm2472.pdf

Comments
sparc Amino acid was a typo. I meant to say nucleic acid. Since the subject of this post involves complementary pair bonding in non-coding RNA I'd have thought anyone involved in it would have realized that amino acid was meant to be nucleic acid. I hadn't even realized I made the mistake which is why 3'-UTR made no sense. Now I see it does if we are talking about proteins but we never were. We were talking about bonding between miRNA and mRNA.DaveScot
August 31, 2008
August
08
Aug
31
31
2008
06:38 AM
6
06
38
AM
PDT
Seems it has something to do with copy/paste. Sorry for that.sparc
August 29, 2008
August
08
Aug
29
29
2008
10:05 PM
10
10
05
PM
PDT
3´-UTR, 3'-UTR? When I copy it from a previous comments it is displayed correctly in the preview but after submitting it the prime was replaced by a question mark. Let's try it again: 3?-UTRsparc
August 29, 2008
August
08
Aug
29
29
2008
10:05 PM
10
10
05
PM
PDT
sparc Your reply makes no sense. For bonding to occur there must be aligned bases.
Indeed, but you were talking about amino acids:
Some of us knew about complementary pair bonding between amino acids and how this works in determining whether any given miRNA sequence will bind with any given mRNA sequence.
In another comment you've stated:
Thus the order of the amino acid sequence is indeed critical.
and later
How exactly does 3? -UTR defeat the need for sequence alignment in complementary pair bonding?
Is this what you read from my comment? I've said
Unfortunately, the order of amino acid is irrelevant since in animals functional miRNA pairing sequences are found in 3?-UTR of mRNAs. Actually this fact is used to detect miRNA targets in mRNAs by programs like miRinda and miRacle.
Obviously, I didn't refer to base pairing but just mentioned that miRNA targets tend to be located in the 3' untranslated region of target mRNAs. By definition the 3'-UTR begins downstream of an mRNA's stop codon and ends at the polyadenylation site. Thus, by definition the
order of the amino acid sequence
or the coding sequence (CDS) as one would rather say is irrelevant for the function of miRNAs discussed here. When you look at figure 1B of the Zeng et al. paper you've cited you will realize that the miRNA target sequences analyzed by Zeng and colleagues have indeed been introduced downstream of the luciferease CDS. This is clearly stated in the Materials and Methods part:
Indicator plasmids pCMV-luc-Target [Target being miR-30(B), miR-30(AB), miR-30(P), miR-30(AP), miR-21(B), miR-21(P), dNxt(B), dNxt(P), or random; AB, anti-miR-30 bulge; P, perfect; AP, anti-miR-30 perfect (Fig. 1A)] were made by combining oligos encoding two copies of the Target sequence and inserting them after the luciferase (luc) stop codon in pCMV-luc (29).
In addition, the luciferase ORF is just used as a reporter which is completely different from natural miRNA target mRNAs discussed in the paper. The fact that combining of a given miRNA target site with a unrelated CDS doesn't affect its action further further demonstrates that the
order of the amino acid sequence
is indeed not critical.sparc
August 29, 2008
August
08
Aug
29
29
2008
09:56 PM
9
09
56
PM
PDT
10_3_3 introduced some confusion by switching the subject from miRNA to siRNA. After an hour or so of pouring through the literature I found the key differences. Both siRNA and miRNA are ~20-25 -nt. siRNA induces RNAi at the transcription level by binding to mRNA and causing the mRNA molecule to be cleaved by accessory proteins. In order for the accessory proteins to complete the cleavage a strong bond must be formed between the siRNA and the mRNA target. Almost perfect complementarity at each of the 20-odd bases is required. siRNA is thus highly specific to inhibition of a single target protein. They are not endogenous in DNA and are believed to be synthesized outside the nucleus and their purpose is the destruction of viral RNA. miRNA on the other hand are endogenous in the DNA and while they were at first thought to have the same function as siRNA in combating viral invasions they have since been found to have a significant and crucial role in development. The extent of their role is not fully understood yet. miRNA can match imperfectly to mRNA and while this does not induce cleavage it still represses protein expression by inhibiting translation at the ribosome. The mismatches in complementarity produce bulges in the mRNA molecule which are thought to play some role in interfering with translation. Thus miRNA by being less specific can repress, to varying degress depending on how inexact the complementarity is, expression in a large number of proteins. That said the complementarity must still be far more exact than 10_3_3 orginally indicated else a single miRNA would be so non-specific in targeting as to create havoc instead of some carefully orchestrated function. To get to the point the id_net_au was trying to make about miRNA and a design inference is that because a single miRNA can effect a plethora of proteins he's wondering how all the complementary sites on genes from different families came to be targeted by a unique miRNA molecule. Such coordinated action in the formation of complementary target sites on just the right set of proteins so that a single miRNA can act on them in concert is difficult to imagine by way of RM+NS when the proteins in the set are from different families which presumably had different ancestral genes. Mass convergent molecular evolution? Yeah, right. This is design staring you doubters in the face yet again. Write that down. DaveScot
August 29, 2008
August
08
Aug
29
29
2008
03:10 PM
3
03
10
PM
PDT
10_3_3 Here ya go. Here's the experimental results of two synthetic miRNA molecules where one bears a sequence mismatch used to determine how much the mismatch effects protein expression: miRNA binding Tell me what part you don't understand.DaveScot
August 29, 2008
August
08
Aug
29
29
2008
01:34 PM
1
01
34
PM
PDT
sparc Your reply makes no sense. For bonding to occur there must be aligned bases. What chemical or mechanical forces do you think causes bonding absent aligned base pairs? Do they just somehow know each other intimately, grow arms, and hug like long lost lovers? DaveScot
August 29, 2008
August
08
Aug
29
29
2008
12:38 PM
12
12
38
PM
PDT
10_3_3 genes containing as few as eleven contiguous nucleotides of identity to the siRNA Thanks for link. Your problem is obviously that you don't understand that 11 contiguous nucleotides of identity means that a sequence of 11 nucleotides in the probe had an exact complementary match with another 11 base sequence in the target. Since the probes were 17 bases long a few extra non-matching bases on either or both ends of the probe was not enough to inhibit bonding. So this is even less likely than perfect 8 nucleotide complementary match. If it was any 11 out of 17 bases that must match the author would not have used the word "contiguous". Is english not your first language? You've also confused the probe (siRNA) with the target (mRNA). The probes were 17 base pairs in length. The targets are manyfold larger. What the article is actually saying is that if 11 continguous base pairs in the probe match any 11 contiguous base pairs in the target then a bond of sufficient strength to derail the mRNA expression into a protein will result. The conclusion is that siRNA molecules can be somewhat promiscuous because they contain more than one contiguous sequence of 11 bases. An siRNA of length 17 will have 7 different possible matching sequences in any mRNA molecule. Fewer than 11 contiguous matches is evidently a bond too weak to derail the mRNA expression into protein as it can be too easily broken. It's unclear to me where you got the 8 base length in previous comments. I suspect that absent an unmatched head & tail hanging off the bonded molecules a perfect match of an 8 base siRNA with an 8 base sequence in mRNA is too difficult to knock off. If there's a loose tail wagging around like a hanging chad then external mechanical and electrostatic forces can use the chad for leverage to knock loose the entire siRNA beginning at the first good bond beyond the hanging chad. DaveScot
August 29, 2008
August
08
Aug
29
29
2008
12:26 PM
12
12
26
PM
PDT
tenstrings I liked your reply. In a private listserv this article was mentioned. Here was my reply to the listserv subscribers: Quote from article: “And Mickey gets cuter because Walt Disney makes more money that way. That is ‘selection.’ ” My reply:
Actually, sir, unless you're prepared to argue that Walt Disney operated his business by throwing darts at a decision matrix while blindfolded what you've presented to the class is a perfect example of intelligent design.
DaveScot
August 29, 2008
August
08
Aug
29
29
2008
12:06 PM
12
12
06
PM
PDT
Paul, Here is the citation for non-specificity of RNAi target sites: http://www.nature.com/nbt/journal/v21/n6/full/nbt831.html relevant text: "Here, we used gene expression profiling to characterize the specificity of gene silencing by siRNAs in cultured human cells. Transcript profiles revealed siRNA-specific rather than target-specific signatures, including direct silencing of nontargeted genes containing as few as eleven contiguous nucleotides of identity to the siRNA. These results demonstrate that siRNAs may cross-react with targets of limited sequence similarity." This is from a 17 basepair target, showing that as many as 6 bases can be mismatched and still have repression effect. Also, the 8 base pair region that we're discussing is a target site, not the RNAi it self. There is no need for a start and stop, and no need for complementary sequence to form a hair-pin. Also, as I said before, this sequence can be located almost anywhere in the 3' region (that is a 5000 basepair window as defined by the authors of the paper) of the gene that is being suppressed.10_3_3
August 29, 2008
August
08
Aug
29
29
2008
11:11 AM
11
11
11
AM
PDT
DaveScot, if this is single-stranded RNA, without other information it could still be transcribed off of either side of the DNA ladder, and there is still frameshifting to consider, so that the true probability of a given base pair being part of the desired sequence is closer to 1/4,096, as I detailed in (26). However, without some reasonable showing that 7, or 6, or less, correct base pairings are adequate, (claimed by 10_3_3 in (27) but not backed up by cited research and seemingly denied by idnet.com.au--"the exact complimentary target sequences for these miRNAs"), the raw probability of this happening is much less than the 1/280 he claimed, and your way of doing the numbers is a more biologically reaslistic one than his. I cited problems with his analysis that he did not deal with. We may still be premature. I looked at your reference, and according to the text and animation there, it seems that there are additional requirements for micro-RNA. The animation shows a complementary sequence upstream (in order to create a hairpin), and it must be located a reasonable number of base pairs away from the original sequence. The animation shows that there are multiple matches, but some areas that do not match. Exactly how long each of these regions must be is not clear. Finally, no information regarding requirements for stop or start signals for the transcription of these micro-RNA's has been given here, so the true random probability of a micro-RNA being produced by chance is likely to be much less than the 1/65,536 you gave. But without more information, it is very hard to calculate that probability. We need more information.Paul Giem
August 29, 2008
August
08
Aug
29
29
2008
10:13 AM
10
10
13
AM
PDT
sparc This is the last time I'm going to warn you to explain yourself. How exactly does 3' -UTR defeat the need for sequence alignment in complementary pair bonding? This knowledge may not be very helpfull when you’re dealing with nucleic acids. Thanks for spotting the typo. I don't care what everyone else says about you, you're demonstrably not completely useless. DaveScot
August 29, 2008
August
08
Aug
29
29
2008
09:47 AM
9
09
47
AM
PDT
Thus the order of the amino acid sequence is indeed critical. Moving one miRNA base out of alignment with its mRNA complement will prevent binding.
Unfortunately, the order of amino acid is irrelevant since in animals functional miRNA pairing sequences are found in 3'-UTR of mRNAs. Actually this fact is used to detect miRNA targets in mRNAs by programs like miRinda and miRacle.sparc
August 29, 2008
August
08
Aug
29
29
2008
09:31 AM
9
09
31
AM
PDT
Some of us knew about complementary pair bonding between amino acids and how this works in determining whether any given miRNA sequence will bind with any given mRNA sequence.
This knowledge may not be very helpfull when you're dealing with nucleic acids.sparc
August 29, 2008
August
08
Aug
29
29
2008
09:15 AM
9
09
15
AM
PDT
Dave, I have made it clear that I am just talking about the math, not the biology. That doesn't mean that I don't know some biology - it just means that I have addressed a purely mathematical problem: if you have 4 x's, 3 y's and 1 z, there are 280 ways to arrange them in order. This says nothing about what the x, y and z might represent. I said nothing at all about whether 10_3_3 was right about the biology. Your remark that "that might work to convince people like you he knows what’s he talking about but it falls flat on its face with those of us better informed" is not accurate because I made no statement one way or another as to whether I think his biological arguments are correct. All I did was explain the math behind the fact that if you assume you have 4 x's, 3 y's and 1 z, there are 280 ways to arrange them in order. We all agree that this math is correct, and I'm not involved in the biological discussion at all, so I'll retreat from the thread.Jack Krebs
August 29, 2008
August
08
Aug
29
29
2008
07:59 AM
7
07
59
AM
PDT
Jack We all understood the math behind both answers. Some of us knew more than just the math. Some of us knew about complementary pair bonding between amino acids and how this works in determining whether any given miRNA sequence will bind with any given mRNA sequence. Order of the bases is critical. No one on my side disputed that 8!/(4!*3!) did not yield 280. We disputed the calculation's applicability to the underlying issue. Evidently 10_3_3 is just throwing around technical buzzwords and equations he reads in the literature without any basic understanding of underlying concepts. That might work to convince people like you he knows what's he talking about but it falls flat on its face with those of us better informed. DaveScot
August 29, 2008
August
08
Aug
29
29
2008
06:47 AM
6
06
47
AM
PDT
id_net Relating to your design hypothesis our level of difficulty in unraveling the genetic and epigenetic secrets of life is going to be largely determined by whether life is the result of design or random coding. In computer programming there is structured code, which is easy to read and comprehend, and comes from planning the overall design in advance of writing the individual pieces. Conversely we also have what is called spaghetti code which result from no planning. When reverse engineering any given program code it is far easier to do with structured code. Reverse engineering of spaghetti code is a nightmare. So the competing heuristics of design and no design become critical. If there is design then we can expect rhyme and reason underlying the code. We can expect to find rules of structure followed in code and once those rules are discovered the reverse engineering job becomes much easier. If not design then we can expect that there are no structural rules being followed, none to discover, and the path of execution in any given thread is unique and gives no insight into how others are structured. The universal genetic code is an example of structure and since it was really the first such structure discovered it was hoped that it would be the keystone making everything else easily understood. As it turned out it was just one keystone. The question is whether or not there are other similar rigid structures like it that explain what is still unexplained or is the rest of molecular biology just one big hopeless mass of spaghetti code. Personally I'm of the opinion that the one bit of rigid structure represented by the genetic code is compelling evidence that there are others waiting to be discovered and once they are understood then with it comes a flood of answers just like there was a flood of answers generated once the operation of the genetic code was understood.DaveScot
August 29, 2008
August
08
Aug
29
29
2008
06:33 AM
6
06
33
AM
PDT
Dave, I am just discussing the math, not the biology. I have not been following this thread, and therefore don't know anything about the biological content of what you guys are discussing. I'm just explaining the math: if one has 4 of one thing, 3 of another thins, and 1 of a third thing, there are 280 ways to make distinct arrangement of the three things. That's all - I was explaining 10-3-3's math. I am making no judgments and no statements about whether this is biologically relevant to what you guys are talking about.Jack Krebs
August 29, 2008
August
08
Aug
29
29
2008
06:19 AM
6
06
19
AM
PDT
Jack A quick sanity check confirms I wasn't wrong about the miRNA binding mechanism. http://www.rosettagenomics.com/?CategoryID=174
MicroRNA regulate protein production in a process known as base pairing – in which complementary codes found on microRNA bind to the corresponding mRNAs much like a lock and key. This process leads to inhibition of protein translation and, in some cases, to degradation of the mRNA itself.
Thus the order of the amino acid sequence is indeed critical. Moving one miRNA base out of alignment with its mRNA complement will prevent binding. In fairness your math was correct. It was your presumptions that were wrong. DaveScot
August 29, 2008
August
08
Aug
29
29
2008
05:56 AM
5
05
56
AM
PDT
Jack You're saying an exact match isn't required. That of course changes the odds and I explicitely stated it would change the odds. If the number to be matched is 22221110 and order is not significant then out of the set of 65536 possibilities many more than one would meet the requirement. 01112222 20111222 22011122 and etcetera. From where are you getting the criteria that any reordering of the digits will cause a match? To the best of my knowledge that isn't how nucleic acid binding works but I could be wrong. DaveScot
August 29, 2008
August
08
Aug
29
29
2008
05:34 AM
5
05
34
AM
PDT
Dave, I understand perfectly well that if any one of the four bases can occupy any of the 8 positions, then there are 4^8 = 65,536 possibilities. But if you know that that you have to have 4 C's, 3 G's, and 1 T, then there are only 280, as I explained. Think of it this way. Suppose you listed all 65,536 possibilities: AAAAAAAA AAAAAAAC AAAAACAA etc. How many of those sequences would contain 4 C's, 3 G's and a T? There would be 280 of them. Conversely, get yourself 8 cards - 4 jacks, 3 queens and a king. Start writing down the number of unique ways you can arrange them in order. There are 280, not 65,536.Jack Krebs
August 29, 2008
August
08
Aug
29
29
2008
05:05 AM
5
05
05
AM
PDT
Jack No, you're quite wrong. Let's do this with base 4 numbers to eliminate confusion dealing with letters. In any base 4 number of 8 digits there are 65536 possible numbers. It's a simple whole number counting sequence: 00000000 00000001 00000002 00000003 00000010 00000011 ... 33333333 i.e. the set of all whole base 4 numbers between 0 and 33333333. There are 65536 numbers in the set. Every last one of them, by definition, is different by at least one digit from all the rest. If an exact match between any two arbitrary 8 digit numbers is required then the odds of it occurirng are exactly 1 out of 65536. If an exact match isn't required and/or the 8 digit numbers are not arbitrary then the odds will of course change but otherwise they don't. Your calculation presumed an exact match between two arbitrary 8 digit numbers. Maybe you should be on the other side of the lectern if you don't understand that. DaveScot
August 29, 2008
August
08
Aug
29
29
2008
04:18 AM
4
04
18
AM
PDT
Paul, Your calculation above does not account for the allowed degeneracy in the binding sequence. This will again make it easier to achieve the sequence randomly, as some bases in the sequence are free to be any base.10_3_3
August 29, 2008
August
08
Aug
29
29
2008
03:33 AM
3
03
33
AM
PDT
It looks to me like both estimates are off. The high estimate assumes that the sequence can be read in only one direction, and if that is not the case, then the probability is multiplied by 2 (yielding 1/32,768). It also assumes that the genome is arranged in 8-bit blocks, whereas for any arbitrary bit, it could be at the beginning, in the middle, or at the end of a block, yielding a reduction, IIUC, of a factor of 8, giving a probability of 1/4,096 for any given base pair being part of the desired sequence. On the other hand, 1/280 is way too large a probability. In a given sequence, we have no a priori reason to suspect that it should consist of precisely 3 A-T pairs and 5 G-C pairs. Perhaps it is even more G-C rich, so that there are 2 A-T pairs and 6 G-C pairs, whereas conversely. areas that are rich in A-T pairs would have less binding sites. Furthermore, there is nothing to prevent a given 8-pair sequence, even if it has the requisite 5 G-C pairs, from having them all with G on the same side, or roughly evenly divided. Perhaps the A-T pairs split with two T's on one side and one on the other, thus making the desired code impossible. IMO, the best estimate using a sheerly random sequence would be 1/4,096. However, certain areas of DNA, such as genes, have their sequence dictated by their function, and are not in fact random. Whether this raises or lowers the estimate for the probability of the desired sequence would have to be determined on a case-by-case basis. This difficulty is one of the reasons that we can only estimate the order of magnitude when doing such probability calculations.Paul Giem
August 29, 2008
August
08
Aug
29
29
2008
01:50 AM
1
01
50
AM
PDT
Unless we know the mechanistic reason that the regions in question, the targets of miRNAs, are already CG rich, we cannot start to calculate probabilities from what we already know. If we already know that a certain hand has been dealt in cards, the probability that it will be dealt is 1. The sort of calculations that are being suggested by some here are sneaking in favourable probabilistic resources without justification.idnet.com.au
August 28, 2008
August
08
Aug
28
28
2008
11:28 PM
11
11
28
PM
PDT
Let me see, Jack's proposal that the 65535 is somewhat high is reasonably believable. However, your number was 280. So is the actual number more likely to be above 10,000, or below? Oh, by the way, if the genic regions are GC rich, where are the Ts going to come from. Being GC rich is surely a benefit, but not all that much.bFast
August 28, 2008
August
08
Aug
28
28
2008
09:49 PM
9
09
49
PM
PDT
bFast, The problem is that it is not just as likely to have As, Cs, Ts, and Gs. As I mentioned before this binding site falls in a genic region. In humans, genic regions are highly GC rich. So, there is a greater chance of finding mostly Gs and Cs. This is why I suggest that the probability is much smaller than 65536. My point is that if you have mostly Cs and Gs there are much fewer ways to arrange them. The calculation above is an upper bound on the probability and the 65536 is a lower bound. I just happen to think that given the location and the fact that the binding site can be degenerate and the fact that it was not found in front of all the genes that were down regulated suggests that the probability is closer to my upper bound than your lower bound. Jack, Thanks! That is exactly right.10_3_3
August 28, 2008
August
08
Aug
28
28
2008
07:10 PM
7
07
10
PM
PDT
Duh - "four bases", not "for bases". At least I didn't write "for basses." :-)Jack Krebs
August 28, 2008
August
08
Aug
28
28
2008
06:57 PM
6
06
57
PM
PDT
10_3_3, If we only have 4 Cs 3 Ts and a G in our kit of parts, I would agree with you, but we've got a bin of Cs a bin of Ts a bin of Gs and a bin of As to choose from. If we look at a random sequence of DNA it is highly unlikely to have exactly 4 Cs, 3 Ts and a G in it. It is just as likely to have all As.bFast
August 28, 2008
August
08
Aug
28
28
2008
06:55 PM
6
06
55
PM
PDT
I can explain the discrepancy between what 10_3_3 is calculating and what the rest of you are calculating. If I have for possible bases A, T, C, and G, and I can pick any one of them for each position, then there are 4 ^ 8 = 65,536 ways to arrange those bases. However, if we already know that there must be 4 C's, 3 T's and a G, the situation changes. Since we are constrained to just using the bases given, and since the 4 C's are indistinguishable from each other, as are the 3 T's, the number of possible ways to arrange these are considerably less than in the first case where any base can be in any position. For the second case, consider this reasoning. We have 8 bases which have to be arranged in order, and there are 8! ways of doing that. However, since the 4 C's are indistinguishable from each other, there are 4! ways to arrange them that all alike: that is, the sequence CCCCTTTG would actually be 24 possibilities because of all the ways the 4 C's could be arranged. Likewise there are 3! ways to arrange the 3 T's, and of course 1! ways to arrange the 1 G. Therefore, the number of ways to arrange the 4 Cs 3 Ts and a G in a sequence is 8!/(4!*3!*1!) = 280 ways, as 10_3_3 says. I teach these concepts in beginning stats, so I'm pretty sure this is correct. Hope that helps.Jack Krebs
August 28, 2008
August
08
Aug
28
28
2008
06:49 PM
6
06
49
PM
PDT
1 2

Leave a Reply