All posts by Brian LM Cheng

SPOTLIGHT ARTICLES: Control of allosteric signaling switch (Dueber et al. 2003) REVIEW

Today I will briefly discuss one of the landmark papers in protein dynamics. This paper is called Reprogramming Control of an Allosteric Signaling Switch Through Modular Recombination, by John Dueber and Wendell Lim from UCSF, published in Science in 2003.

Allosteric signaling switch has been a focus in synthetic biology, as proteins that belong to the signaling pathways are often regulated by allosteric gating mechanisms. Some of these mechanisms are governed by simple binding domains, in which the switch is under autoinhibited state when proper ligand is absent. In this paper, Dueber and his group proposed that combinations of domains could give rise to a diversified number of gating behaviours in response to nonphysiological inputs. This story is highly significant because their results indicate that simple catalytic and interaction domains within a single polypeptide could give rise to different proteins of complex properties. It also advances our knowledge to how proteins of complex functions could be derived.


To approach the hypothesis, Dueber had developed two key major experimental procedures: (i) switch proteins; and (ii) actin polymerization assays: pyrene-labeled and carboxylated polystyrene bead. There were three types of switch proteins developed: single heterologous ligand; chimeric switch; and heterlogous switch. To develop switch proteins, the ligand motifs and the linker region were PCR-amplified from existing plasmids. These proteins were then expressed with an affinity tag (either 6xHis tag or GST tag), followed by purification and tag removal by Ni beads and TEV protease or glutathione agarose resin. Next, actin polymerization assays were performed. In the case of pyrene-labeled actin assays, pyrene-labeled actin was incubated with Arp2/3, switch proteins and ligand proteins to initiate polymerization. Then the rate of actin polymerization could be measured by fluorescene measurement. Next, bead actin polymerization assay was performed. Carboxylated polystyrene beads were coated with GST fusions to input ligands and PDZ ligand. After incubation, the beads were washed and incubated with Xenopus oocyte extractsand rhodamine-labeled actin, followed by microscopic analysis on rhodamine.

Dueber developed a signaling switch gated by single ligand. This switch was consisted of N-WASP output domain, GBD and B-motif. At the autoinhibited state, GBD and B motif interacted with output and Arp2/3 to suppress output to interact with actin filament. To initiate actin polymerization, Cdc42 and PIP2 bind to GBD and B motif respectively to relieve autoinhibition. This model was consistent with the observation that increasing PDZ ligands could activate the switch. For the second designs, they developed AND-gate switches by linking two domains with an output. There were two classes of designs developed: chimeric and heterlogous switches. Chimeric switch design was regulated by PDZ ligand and Cdc42. The modular domains of this particular design were PDZ and GBD domains, in which Cdc42 could disrupt interaction of GBD and output. Heterologous switch was regulated by PDZ and SH3 ligands. The modular domains were composed of PDZ and SH3 domains. Both classes of switches were subdivided into behavioural classes. Out of 34 switches, 2 showed antagonistic gating, 2 showed OR gate behaviour, and 5 showed AND gate behaviour. From analyzing the behaviours, the basic design principle was identified, in which the affinity of autoinhibition was correlated with the basal repression and input sensitivity. This principle was supported by 2 observations: (1) reducing the PDZ ligand-affinity for switch C11, the switch then resembled AND gate; and (2) increasing PDZ ligand affinity turned H15 switch from behaving like OR gate to AND gate. Their results also indicated the importance of linker region to switch behaviour, as the increase of linker region from 5 to 20 residues could increase the sensitivity of inputs. Furthermore, in the case of antagonistic switches, PDZ ligand was found to act as an activator while SH3 increased the basal level of repression. PDZ was responsible for the autoinhibitory interaction with output, and SH3 interaction was responsible for destabilizing the PDZ interaction. In the case of positive integrating switches, both domains worked together to downregulate the output activity. Interestingly, disruption of both was required to activate the output function.

Dueber and his group have effectively addressed the main question. The results from simple binding domains and corresponding ligands had supported the notion that complex signaling circuits could be derived from simple domains, and more importantly, domain recombination could potentially derive proteins with novel regulated functions. For future direction, the geometry of the output domain must be investigated. In order for the output domain to be regulated by other domains and to carry its native function, certain geometry must be preserved. To understand the preservation, X-ray crystallography and mass spectroscopy should be conducted on samples: output domain bound to GBD and output domain bound to actin filament.

Back to home page

Reference: Dueber JE, Yeh BJ, Chak K, Lim WA. Reprogramming Control of an Allosteric Signaling Switch Through Modular Recombination. Science. 2003 Sep 26;301(5641):1904-8.

Spotlight articles: Trading Places on DNA (Yuzhakov et al. 1999) Review

Today I will briefly discuss one of the landmark papers in molecular biology. This paper is by Alexander Yuzhakov, Zvi Kelman and Mike O’Donnell, from the Rockefeller University and the HHMI in New York City.

DNA duplex has an elegant structure, yet its replication relies on a sophisticated and complex network of factors. During DNA replication, helicase unwinds the DNA duplex. SSB binds to this single stranded region to maintain and stabilize the structure. On the lagging strand, primase is required to produce a series of RNA primers. The β sliding clamp is then loaded onto the primed site to enable the extension by DNA polymerase III. This landmark paper, by O’Donnell group, tries to understand how primase is displaced from DNA. They hypothesize that there is a component in the DNA replication that removes primase after RNA primer is made. This paper is highly significant as (i) it shows that there is a competition between components to enable efficient and regulated DNA replication; (ii) and due to this competition, components that are present at a low concentration, such as primase, can be recycled.

This paper started from trying to understand the temperature-sensitive replication defect on phage DNA when it is coated with SSB-113 mutant. When phage G4 ssDNA is primed with synthetic RNA, both SSB- and SSB-113-coated ssDNA can replicate at permissive and restrictive temperatures. However, at restrictive temperature, SSB-113-coated ssDNA can not be replicated. This defect is not due to the malfunction of primase. Knowing that the β-clamp is loaded after primase priming, they then continue to see if the loading is defective. Using 32P-labeled β-clamp, they realize that the clamp loading is defective in SSB-113-coated ssDNA at restrictive temperature, which suggests at restrictive temperature, the SSB-113 prohibits the clamp loading, which may consequently leads to replication defect. They then continue to ask what causes the displacement of primase. β-clamp alone does not displace primase, but DNA Pol III and the β-γ complex does. Interestingly, χ subunit of Pol III alone displaces primase effectively. Because Pol III does not displace primase at restrictive temperature when ssDNA is coated with SSB-113, they then hypothesize χ’s interaction with SSB is temperature sensitive. χ does not interact with primase. When ssDNA is coated with SSB-113, χ shows decreased interaction comparing to SSB-coated ssDNA. In contrast, primase does not show this interaction discrepancy with SSB-113. Interestingly, as χ increases, primase bound to the ssDNA decreases, which indicates that there is a competition between χ subunit and primase for SSB interaction. To confirm the competition, they compare the level of synthesized DNA and bound primase in the presence and absence of χ. In the absence of χ, primase remains bound to the ssDNA, and replication of lagging strand is defective. In the presence of χ, primase is displaced, and lagging strand is replicated. Last but not the least, they show that the displacement by Pol III is polymerase-specific, as other polymerases do not replicate. This primase-polymerase switch is also species-specific as human clamp PCNA does not load when phage primase is present. The primase also protects the RNA primer from RNAse. Taken together, primase-polymerase switch depends on the competition between χ subunit of RNAP and primase for interacting with SSB. Primase also protects the RNA primer, and its displacement from SSB-coated ssDNA is species- and polymerase-specific.

Mike O’Donnell’s group has presented an important mechanism in DNA replication with well-supported evidence. The paper’s successful attempt to bring the switch concept in phage G4 ssDNA into a broader and general concept, such as replication fork, is admirable. This brilliant attempt is shown in Fig. 5, with well thought-out designs to prove the displacement function of χ subunit. O’Donnell’s approach to answer the questions is logical, and I believe that this paper has answered the questions posed. However, the author has yet addressed why the primase is not displaced at restrictive temperature. On 154, the author states, “the χ subunit of Pol III holoenzyme interacts much more weakly with SSB-113 than with wild-type SSB”. It is possible that the res. 113 is essential for interacting with χ. And at restrictive temperature, interaction with res. 113 is lost possibly due to conformational change. SSB-113 carries Pro (non-polar) to Ser (polar) mutation. To see if the polarity of this residue is important for interaction with χ, we can mutate the Pro113 to Cys113 (slightly polar). By immobilizing the Cys113 mutant in SPR, we can see if this slightly-polar residue would result in intermediate interaction. Fig. 1A illustrates how the primase produces RNA primer in a 5′-3′ direction. How does χ detect the primase-bound SSB? As χ does not interact with primase (Fig. 4C), does primase binding change the structure of SSB providing a signal for χ subunit recognition? To address this question, we have to first understand where χ binds to in ssDNA. In other words, does χ bind to the opposite or the same face of primase? To approach, we can perform DMS methylation footprinting to probe for MG/mg regions that are protected by χ and primase. By comparing the binding regions, we will have a better understanding on how the displacement works.



Back to home page


Yuzhakov A., Kelman Z., O’Donnell M. (1999). Trading places on DNA — a three point switch underlies primer handoff from primase to the replicative DNA polymerase. Cell (96, 153-163).


A visit to the current controversy: manmade mutant H5N1 influenza virus (1)

A researcher, named Ron Fouchier, claimed that he has created a new variant (or you may call it the mutant) of H5N1 influenza virus. Normally, H5N1 influenza virus can not be transmitted between human and human. And it rarely transmits from an avian source to human. But when it does, the fatality rate is approximately 50%. Dr. Fouchier, who is an investigator at the Erasmus Medical Center in the Netherlands, claimed that he has created a new and airborne H5N1 influenza virus. This mutant is transmissible between mammals. Although his research finding is yet published and opened to the public, his research has created a controversy in the balance of academic freedom and protecting the nation from bioterrorism. As a member of the general public, I, too, have no idea what the introduced mutation Dr. Fouchier made to the original H5N1 strain. As a nascent virologist here at UC, here I will make some speculation and discuss about the ongoing situation.

Redefining the risk of mutant virus

One of the key aspects circulating in the news is the fact that this mutant is “airborne”. In my opinion, almost all viruses that disseminate within the respiratory tract are airborne. If viruses are able to replicate in our lung, it is highly possible that the mucus or droplets from an infected individual contain a considerable amount of viruses. The reason why some viruses is not able to disseminate effectively is possibly due to the failure to replicate efficiently within the host, resulting in a low concentration/titer of viruses which may not be enough to infect another susceptible individual.

“TONY EASTLEY: Scientists who have produced an airborne mutated of the killer bird flu virus H5N1 say it’s essential their research be published, but opponents say the biosecurity risks are too high and would be a “how to” manual for terrorists.” ABC NEWS

The public is now informed that this mutant virus is airborne. But honestly, the concern should be focus on the fact that it has now adapted to be spread between human and human. In the past, humans are the accidental host of H5N1 viruses. In other words, it is rare when the H5N1 viruses infect humans. So the concern should focus on the fact that the specific mutation can change the targeting species and disseminating efficiency. For an effective virus replication cycle, the virus must find a susceptible host cell that expresses a certain protein on its membrane. These proteins are known as receptors. These receptors are the key for the virus to recognize the proper cells to enter. The next barrier the virus needs to overcome is whether the host cells can provide the necessary factors for it to replicate.

In the next issue, we will continue to discuss how the viral surface proteins can be mutated to recognize different hosts (which I think it’s one of the key barriers to derive a new species in viruses). We will go over some current literature on CHIKV mutation and how the mutation on the viral surface protein can change the selectivity and the replication efficiency.


Back to home page