iiinova — where your digital self keeps evolving

Brian Kelch @briankelch.bsky.social · 07/07/2025

RFC is conceptually analogous to a GTPase, which has dedicated Nucleotide Exchange Factors and NTPase Activating Partners. What makes RFC unique is that its own substrates are the NEF (PCNA) and NAP (DNA). (13/16)

120 2

Brian Kelch @briankelch.bsky.social · 07/07/2025

Based on our data, we propose a model whereby fast ATP binding to some subunits readies RFC to bind PCNA, and then PCNA completes RFC’s nucleotide exchange en route to clamp loading. We propose this prevents futile ATP hydrolysis and helps ensure RFC binds PCNA before DNA. (12/16)

100 1

Brian Kelch @briankelch.bsky.social · 07/07/2025

Josh performed more MD sims to try to understand how PCNA could promote nucleotide exchange in RFC. His sims predict that the D/E interface weakly bridged by ADP can be pried open as D binds PCNA, accelerating release. (11/16)

100 1

Brian Kelch @briankelch.bsky.social · 07/07/2025

So, what catalyzes nucleotide exchange? Josh developed a FRET assay using a labeled ATP analog to monitor ATP binding to RFC. RFC alone binds some ATPs fast but then binds more ATP slowly, consistent with A&B (and maybe C) exchanging faster than D. PCNA speeds up these slow binding steps! (9/16)

100 1

Brian Kelch @briankelch.bsky.social · 07/07/2025

To test how interface opening plays a role in ADP release, Josh performed tRAMD(doi.org/10.1021/acs....), which is a technique that predicts relative release rates. He found that the half-off rate for C or D is slower than A or B, consistent with his cryo-EM. (8/16)

100 1

Brian Kelch @briankelch.bsky.social · 07/07/2025

Next, we wondered how the dynamics of RFC influence ADP release? So, Josh performed MD sims. His sims predict that ATP-interfaces are tighter than ADP ones, and that the B/C interface can open farthest, offering an explanation why this interface is apo in our cryo-EM reconstructions. (7/16)

100 1

Brian Kelch @briankelch.bsky.social · 07/07/2025

First, we looked at how RFC subunits respond to nucleotide binding by comparing Josh’s apo subunits to our lab’s previous ATPgS-bound subunits and found that the B&C subunits’ lids rotate in opposite directions! We had never seen this in a AAA+ before (if you know of any, please let us know!) (6/16)

100 1

Brian Kelch @briankelch.bsky.social · 07/07/2025

Imagine our surprise when we saw that the C and D subunits were ADP-bound! Because he never added nucleotide, these ADPs must have co-purified with RFC over a 3 day prep, meaning that RFC releases ADP very slowly from these subunits. (4/16)

100 1

Brian Kelch @briankelch.bsky.social · 07/07/2025

To get insights into RFC recycling, postdoc @joshuapajak.bsky.social determined the structure of RFC in the apo state. He got great overall resolution of the core RFC complex, but the AAA+ domain of the A subunit is missing, suggesting it is flexibly tethered when apo. (3/16)

100 1

Brian Kelch @briankelch.bsky.social · 07/07/2025

Recently, our lab showed how an ATP-bound RFC loads PCNA onto DNA during replication and repair (doi.org/10.7554/eLif...). But RFC loads PCNA extremely fast, so that when RFC has done its job once, how does it prep for the next round? (2/16)

100 1

Brian Kelch @briankelch.bsky.social · 07/07/2025

Every day we make enough DNA to go to the moon and back. So, all our DNA replication enzymes work fast, right? In our latest preprint, we show that a key enzyme, the clamp loader ATPase RFC, releases ADP much slower than it loads PCNA. What gives? (1/16) doi.org/10.1101/2025...

180 8

Brian Kelch @briankelch.bsky.social · 18/04/2025

Putting that together with decades of research into phage genome packaging, we propose a speculative model for how the Portal switches from DNA-pumping mode to Tail attachment mode, using the pressure-induced piston motion of Portal to displace the motor. /12

130 3

Brian Kelch @briankelch.bsky.social · 18/04/2025

So how does the Portal know it’s time to close the pore? We got a clue by comparing the position of the Portal in capsids before and after DNA has been pumped in. The pressure from DNA pushes the Portal lower in the capsid aperture like a piston! /11

130 3

Brian Kelch @briankelch.bsky.social · 18/04/2025

On the left, Bayfield et al. showed the pore through Portal in an early, immature capsid. A specialized motor pumps the genome into the capsid through this open, hydrophilic pore. Then, it constricts in the mature phage, becoming narrow and hydrophobic. /10

130 3

Brian Kelch @briankelch.bsky.social · 18/04/2025

In 2020, Ollie Bayfield and Fred Antson predicted that one component of the Neck complex, the Portal protein (dark blue), would be constricted to hold DNA in the capsid of this phage. And they were right! /9 Bayfield et al elifesciences.org/articles/55517

130 3

Brian Kelch @briankelch.bsky.social · 18/04/2025

We know that the Neck of this phage is sufficient to hold its DNA inside the capsid because we also isolated phage that have broken tails, and therefore nothing left but the Neck to retain DNA. /8

150 5

Brian Kelch @briankelch.bsky.social · 18/04/2025

But the phage has another problem—it needs a channel through the Neck and Tail to inject its genome into the host when it infects, but it can’t shoot its shot until it finds the right host. How does it keep that channel closed? /7

130 3

Brian Kelch @briankelch.bsky.social · 18/04/2025

Instead of beefing up electrostatic or hydrophobic interactions between these proteins, the phage intertwines different subunits to create topological linkages. This is the same strategy used to make a wicker basket strong! /6

192 9

Brian Kelch @briankelch.bsky.social · 18/04/2025

With 105 protein chains and 5 DNA segments, across 4 different symmetries, this is the largest structure the Kelch Lab has ever built (2.4 MDa)! The proteins in pinks/reds are part of the capsid shell, and they form an aperture that the rest of the Neck sits in. /4

170 7

Brian Kelch @briankelch.bsky.social · 18/04/2025

Postdoc @emma.sedivy.bsky.social determined the structure of the Neck complex that joins the capsid head to the tail. We only knew the identity of 4 out of 8 of the proteins in the Neck, but we were able to identify the remaining 4 using ModelAngelo by @kjamali.bsky.social. /3

151 5

Brian Kelch @briankelch.bsky.social · 18/04/2025

This phage not only lives in hot springs but also has the longest tail of any known phage. Here you see 3 viruses infecting their host, T. thermophilus. One has released its genome into the host, but two still have DNA-filled heads. /2

191 9

Brian Kelch @briankelch.bsky.social · 18/04/2025

Did you know that there is so much DNA packed inside a phage capsid that the pressure is 10X higher than in a bottle of champagne? In our latest preprint, we wondered how that is contained in a phage that lives at extremely high temperatures. /1 www.biorxiv.org/content/10.1...

Meme describing types of headaches, with the most all encompassing headache being the stress of holding copious amounts of DNA in a phage head.

69225 92

Brian Kelch @briankelch.bsky.social · 04/01/2024

This leads to a mechanistic hypothesis that Copia and Arc antagonize each others’ capsid assembly by competing for RNA. The functional relevance of Copia in synaptic dev also suggests that we might be watching Copia in process of ‘domestication’. /6

A) Image showing mutual antagonism between Copia and Arc. Arc is pro-plasticity, while Copia is inhibitory. B) Proposed mechanism for mutual antagonism, in which Arc and Copia capsids compete for RNA.

100 1

Brian Kelch @briankelch.bsky.social · 04/01/2024

Amazingly, knockdown of Copia results in increased bouton development and synaptic plasticity, the opposite of an Arc knockout. Copia and Arc are antagonizing each other. 5/

A) Images of Drosophila NMJ showing that Copia knockdown causes increased boutons. B&C) quantification of bouton increase

100 1

Brian Kelch @briankelch.bsky.social · 04/01/2024

We find that Copia requires RNA for proper assembly of the capsid structure. The structure shows pores that presumably allow small molecules to enter the capsid, presumably for reverse transcriptase activity. Arc lacks these pores. /4

Image of the Copia hexamer, showing the pore in the center is ringed w postiviely charged residues.

100 1

Brian Kelch @briankelch.bsky.social · 04/01/2024

Here, they report that the retrotransposon Copia forms capsids that are also packaged into EVs that transport across the NMJ. We determined the structure of these virus-like particles, which are larger than Arc capsids, presumably to bind bigger Copia RNA. /3

A) images of Drosophila NMJs, showing presence of Copia protein at the NMJ. B) Negative stain EM micrograph of Copia capids. C) Structure of the Copia capsid. D) Structure of the dArc1 capsid (from the Briggs group).

100 1

Brian Kelch @briankelch.bsky.social · 04/01/2024

I'm back to do an in-depth dive into our most recent work up on biorxiv. Led by the Thomson Lab in the Neurobiology dept, w Yumeng Liu of the Kelch Lab. We find that the retrotransposon Copia forms capsids that antagonize Arc for controlling synaptic plasticity. /1 www.biorxiv.org/content/10.1...

140 4

Brian Kelch @briankelch.bsky.social · 24/12/2023

Just in time for Festivus, another paper up on biorxiv. Led by Thomson Lab in Neurobiology dept @ UMass Chan Med School, w Yumeng Liu of the Kelch Lab. We find that the retrotransposon Copia forms capsids that antagonize Arc for controlling synaptic plasticity. www.biorxiv.org/content/10.1...

020 2

Brian Kelch @briankelch.bsky.social · 04/12/2023

So why the different mechanisms between euks and bacs? We suspect it has to do w arrangement of binding sites on the clamp. Euk clamp has strong-weak-strong-weak-strong-weak arrangement. Bac clamp is strong-weak-weak-strong-weak-weak, leading to different mechanisms. 8/

100 1

Brian Kelch @briankelch.bsky.social · 04/12/2023

We also find an alternate state in the loader in which a pore opens at the same position as seen in the euk loader, which we & others had shown is important for unwinding DNA. Similar unwinding activity in Ecoli? We don’t see evidence. Jury is still out on pore’s function. 7/

Figure showing that the Alternate_collar state adopts an unanticipated conformation where a pore opens between the A and B subunits of the clamp loader.

100 1

Brian Kelch @briankelch.bsky.social · 04/12/2023

Then the primer/template junction binds, which forces the assembly to adopt a spiral that tightly hugs the p/t. This state is nearly identical with an RNA or a DNA primer. The clamp then closes and the loader undergoes a step-wise disengagement from the clamp. 6/

100 1

Brian Kelch @briankelch.bsky.social · 04/12/2023

Clamp opening in bacteria is driven by a single-pivoting motion, similar to chopsticks. Whereas eukaryotes open the clamp in with multiple pivoting motions, broadly resembling a plumeria flower blooming. 5/

Schematic illustrating that E.coli loader opens using a single pivot point, whereas yeast loader pivots at multiple points.

100 1

Brian Kelch @briankelch.bsky.social · 04/12/2023

The clamp is then cracked into a Semi-Open intermediate that is not observed in eukaryotes, before it pivots into a planar, wide-open state. We had previously seen that euks open into a spiral shape that is not as wide. 4/

100 1

Brian Kelch @briankelch.bsky.social · 04/12/2023

We used cryoEM to determine structures of the E.coli clamp loader in action. We then compare these structures against our recent structures of eukaryotic clamp loading. elifesciences.org/articles/74175 The Ecoli loader initially binds using a different interface than euks. 3/

Comparison of the initial binding states of bacterial (this study) and eukaryotic (Gaubitz et al 2022) clamp loaders. The bacterial loader uses a different interface to initiate binding.

100 1

Brian Kelch @briankelch.bsky.social · 04/12/2023

All life require sliding clamps that encircle DNA where they activate DNA polymerases. Accordingly, all life use a pentameric ATPase clamp loader to place sliding clamp rings onto DNA. Our comparative study examines how conserved this mechanism is across evolution. 2/

111 1

Brian Kelch @briankelch.bsky.social · 04/12/2023

I’m happy to announce a new paper from the lab. We find numerous differences in the mechanisms of clamp loading in bacteria vs eukaryotes, despite this machinery being amongst the most conserved components of the replication fork. 1/ www.biorxiv.org/content/10.1...

183 8