Wednesday, March 6, 2013

Grainy Westerns and Fuzzy Logic: A Review of Shi, et al 2011

A review of Shi et al. Pyrazinamide inhibits trans-translation in Mycobacterium tuberculosis. Science (2011) vol. 333 (6049) pp. 1630-2
PMCID: PMC3502614

This study was recently published in the prestigious journal Science attempted to explain the mode of action for an important antibiotic, Pyrazinamide, used in treating tuberculosis. Due both to the high profile of this work, and the conclusion they draw regarding trans-translation, I will provide below a review of this article.

The interested reader should be aware of one important fact when approaching this article: there are different versions of it, depending on how you access it. The open access version at Pubmed is outdated; a more recent version (with updated Figure 3) is available from the journal Science itself. 

Manuscript Highlights:
  1. Demonstrates binding of ribosomal protein S1 to POA (active form of Pyrazinamide drug)
  2. Suggests mechanism of action for POA: binding to S1 inhibits trans-translation.
  3. Quality of trans-translation related data (western blots) make the conclusion drawn by the authors questionable.
  4. Alternative explanations to trans-translation rescue of stalled ribosomes in their in vitro system are not ruled out, undermining the conclusions drawn.


In this study, Shi and colleagues sought to identify the target and antibiotic mechanism of the anti-tuberculosis drug Pyrazinamide. This drug is important in treatment of tuberculosis, particularly in clearing persister cells through combination with other compounds. In order to accomplish this task, the authors used affinity chromatography to capture M. tuberculosis proteins that are capable of interacting with the drug. Through this approach, ribosomal protein S1 is identified as the primary target.

Using ribosomal protein S1 (RpsA) as a lead, the authors attempt to explain the mode of action for PZA. Binding studies using isothermal titration demonstrate more conclusively that PZA is capable of binding RpsA Furthermore, a PZA resistant strain with mutant RpsA genes that cannot bind the drug is identified. Finally, the authors use an in vitro translation system to assess the effect of PZA on translation and trans-translation. The conclusion drawn from these studies is that PZA specifically inhibits trans-translation, but only in the context of M. tuberculosis ribosomes.

A Critical Review:

If correct, the conclusions drawn by the authors of this study carry significance for several reasons. Due to the importance of PZA in the treatment of tuberculosis, insights into the target and mechanism of action of this drug will expedite development of more potent anti-tuberculosis agents (not to mention arming researchers with knowledge which might help combat drug resistance in M. tuberculosis.) Secondly, PZA would be the first drug identified to specifically inhibit trans-translation. The ribosome rescue process of trans-translation is known to be important for the virulence of several pathogenic bacteria (Okan, 2006), and drugs that inhibit this process could make ideal antibiotics (which would have limited side effects since they would not indiscriminately kill all bacteria as traditional antibiotics do.)

Unfortunately, poor quality data was used to draw the second set of conclusions regarding trans-translation. This is primarily a problem in figures 3C through 3F (PMC version). This set of western blot data is grainy and loading appears uneven in some cases. For example, comparison of the background levels of protein in lanes 1 and 4 from Figure 3E would suggest either 1) unequal loading or 2) POA increases translation from M. smegmatis ribosomes in vitro. The authors do not address this problem, but instead conclude simply that trans-translation on M. smegmatis ribosomes is not inhibited by POA. In other panels from the same figure, the grainy texture makes it hard to evaluate the relative levels of a protein band between lanes. Indeed, large error bars are present in the quantitation of one of these results (Figure 3F in the version from Science).

Also lacking in figures 3C, 3E, and 3F are controls to demonstrate that the products indicated as tagged DHFR are indeed the result of trans-translation. An appropriate control to include here would be the same in vitro translation system without tmRNA and SmpB. It is important to note why such a control is necessary; in addition to tmRNA-tagging, the higher molecular weight product observed can be accounted for by alternative explanations. One such possibility is that trans-translation is not occurring at all, but rather ribosomes are miscoding their way to completion on templates with rare codons.

Bacterial ribosomes have multiple mechanisms by which to resolve unproductive pauses at tandem rare codons. Translation elongation can continue if ribosomes either frameshift or miscode. Either of these explanations are sufficient to account for full length DHFR: all three reading frames downstream of the tandem rare codons are devoid of stop codons. Furthermore, this extra portion of the reading frame is a similar length to the tmRNA ORF, and therefore would migrate on the gel at the same level as tmRNA tagged DHFR. l Although both miscoding and frameshifitng are thought to occur at slower rates than normal translation events, they sufficiently compete with trans-translation in vivo (Lainé, 2008; Calderone, 1996). Furthermore, enough time (2hrs) is given for either miscoding or frameshifting to occur and drive protein synthesis to completion. This can be demonstrated through a simple calculation:

If we assume a translation rate of 10 amino acids a second (a slow estimate; Bionumbers database), and an error rate of 1x10e-4 incorrect amino acids per sampling (Kramer, 2007), then it would take 17 minutes to incorrectly decode one of the rare codons. Under these generous assumptions, decoding all 8 would take just over 2 hours (133 minutes).

Another ribosome rescue pathway for resolving pauses at tandem rare codons involves cleavage of the mRNA in the A-site of the ribosome and subsequent binding by tmRNA and SmpB to carry out trans-translation. The authors fail to note that this cleavage is a prerequisite for tmRNA activity, as the tmRNA-SmpB complex recognizes only ribosomes with an A-site devoid of mRNA (Pech, 2012; Shimizu, 2012).  Furthermore, other studies using the same PURExpress in vitro translation system featured in this manuscript have observed no evidence of cleavage. For example, rescue of ribosomes stalled in a similar fashion cannot be rescued by ArfA, which also requires an empty A-site (Shimizu, 2012). The presence of tmRNA and SmpB is not believed to be sufficient to promote cleavage, as they are dispensable in vivo (Garza-Sánchez, 2009). If this study did a better job of demonstrating that the higher molecular weight product is indeed the result of cleavage and tagging, the search for the as of yet unidentified A-site nuclease would be greatly facilitated (Genes common to E. coli, M. smegmatis, and M. tuberculosis would be excellent candidates).

Despite methodological issues (prolonged incubation time, lack of controls), there is one final aspect of the data which is in fact inconsistent with a trans-translation role for PZA. In Figure 3C, the disappearance of the high molecular weight band is taken to be evidence that trans-translation is being inhibited. Studies of trans-translation from other organisms demonstrate that in the absence of tmRNA, untagged protein products accumulate (Dulebohn, 2007). In this study, the disappearance of the higher band is not accompanied by accumulation of a smaller band. There may be potential explanations for this, especially if ribosomes are limiting in the in vitro translation mixture. Still, this raises further questions as to what step during translation of these mRNAs is actually being inhibited. 

Concluding Remarks:

Although this paper makes an interesting contribution to the study of PZA and the treatment of M. tuberculosis infections, the insufficient controls and poor quality data make it difficult to agree with their conclusions regarding inhibition of trans-translation. This is unfortunate, for this conclusions would be important if they could be demonstrated with more scientific rigor. As it currently stands, PZA appears to have a tandem rare codon specific effect on translation by M. tuberculosis in vitro, but this could be caused by an inhibition of either trans-translation, A-site cleavage, miscoding, frameshifting, or possibly even another event.

Surprisingly, none of these issues were even touched upon in the discussion by either the authors or a complementary perspective article (Cole, 2011) Addressing these concerns would constitute an improvement that might make this manuscript worthy of the journal in which it was published.

There are a couple of other interesting observations and discrepancies in this paper.
  1. Lane 4 from Figure 3A and Lane 3 in Figure 3C are not consistent; why has "tagged" DHFR completely disappeared if binding of tmRNA to RpsA is relatively unaffected.
  2. Why is so much more untagged protein produced from E. coli ribosomes (Figure 3F). It is possible that the combination of the PURExpress components (all derived from E. coli) with M. tuberculosis or M. smegmatis ribosomes results in inefficient or restrictive translation?
  3. Ribosomal protein S1 has been shown to be dispensable for trans-translation in vitro, raising questions regarding the exact role it plays in this system.


Calderone et al. High-level misincorporation of lysine for arginine at AGA codons in a fusion protein expressed in Escherichia coli. Journal of Molecular Biology (1996) vol. 262 (4) pp. 407-12

Cole. Microbiology. Pyrazinamide--old TB drug finds new target. Science (2011) vol. 333 (6049) pp. 1583-4

Dulebohn et al. Trans-translation: the tmRNA-mediated surveillance mechanism for ribosome rescue, directed protein degradation, and nonstop mRNA decay. Biochemistry (2007) vol. 46 (16) pp. 4681-93

Garza-Sánchez et al. RNase II is important for A-site mRNA cleavage during ribosome pausing. Molecular Microbiology (2009) vol. 73 (5) pp. 882-97

Kramer and Farabaugh. The frequency of translational misreading errors in E. coli is largely determined by tRNA competition. RNA (2007) vol. 13 (1) pp. 87-96

Lainé et al. Ribosome can resume the translation in both +1 or -1 frames after encountering an AGA cluster in Escherichia coli. Gene (2008) vol. 412 (1-2) pp. 95-101

Okan et al. A Role for the SmpB-SsrA system in Yersinia pseudotuberculosis pathogenesis. PLoS Pathog (2006) vol. 2 (1) pp. e6

Pech and Nierhaus. Three mechanisms in Escherichia coli rescue ribosomes stalled on non-stop mRNAs: one of them requires release factor 2. Mol Microbiol (2012) vol. 86 (1) pp. 6-9

Shi et al. Pyrazinamide inhibits trans-translation in Mycobacterium tuberculosis. Science (2011) vol. 333 (6049) pp. 1630-2

Shimizu. ArfA recruits RF2 into stalled ribosomes. Journal of Molecular Biology (2012) vol. 423 (4) pp. 624-31

No comments:

Post a Comment