Circularly permuted LOV domain as an engineering module for optogenetic tools

Lequn Geng, Jiaqi Shen, Wenjing Wang

Preprint posted on May 19, 2021

The circle of LOV: Circularly permuted LOV domain enables photo-control of proteins that require a free N-terminus and expands the applicability of optogenetic tools

Selected by Soni Mohapatra

Categories: synthetic biology


Optogenetics tools are increasingly being used for the on-demand spatiotemporal control of various cellular processes using light. The light sensitive LOV (light-oxygen-voltage) domain is one of the most commonly used protein domains for modulating a variety of proteins (1). The LOV protein consists of a core, called the PAS (Per-Arnt-Sim) domain and a C-terminal helix, named Jα-helix. Normally, the protein of interest is fused to the Jα-helix of the LOV domain. In the dark state, Jα-helix is stacked against the protein core which cages the protein of interest via stearic hindrance. Upon blue light irradiation, the Jα-helix unfolds which exposes the protein of interest and restores its function (2). Currently, the usage of LOV domain is limited to caging the N-terminus of the proteins and involves optimizing linker lengths between the LOV domain and the protein of interest. Having a more versatile version of the LOV domain can expand the current optogenetic toolkit.

Circular permutation (3) is a method that is adopted to engineer proteins with different connectivity between the amino acids while retaining the three-dimensional structure of the protein. In this preprint, the authors used the circular permutation strategy to re-engineer a version of LOV domain with an accessible C-terminus that can be fused to a wide variety of proteins.

Key findings

Design of a circularly permuted LOV domain

Geng et al. developed a circularly permuted version of LOV (cpLOV) while retaining the light sensing capability of the LOV domain. hLOV1 (4), a version of LOV with a better caging propensity in the dark state was chosen as the candidate for circular permutation. cpLOV was generated by permuting hLOV1 using a glycine-serine linker. hLOV1 was split between L250 and H251 and a new terminus was introduced at the original helix region connecting the Jα-helix and the PAS core. The GSGS linker is sufficiently long to serve as a connector between the original N- and C-terminus of the LOV domain.

cpLOV as a modular domain for controlling protein activity

To test the usability of cpLOV as a caging domain, a hexapeptide SsrA (5) was fused to cpLOV. The choice of the cpLOV fusion site to ssrA has a significant effect on the photo-sensing ability of the SsrA-cpLOV fusion construct. The authors tested ten different fusion sites along the Jα-helix. Yeast surface display was used to evaluate the binding ability of these different SsrA-cpLOV fusion construct to SspB with and without light exposure. The four fusion constructs that displayed light-induced SspB binding were fused away from the central region of the Jα-helix. While the best cpLOV fusion constructs showed better binding ability under light conditions compared to hLOV1, they also displayed undesirable higher background activity in the dark.

The authors also employed cpLOV in a multi-component system to achieve photo-activated transcription in HEK293 cells. They were able to achieve nearly six-fold enhancement of transcriptional activation upon exposure to light.

Improving the peptide caging efficiency of LOV domain using cpLOV

One of the major drawbacks that several photo-controlled proteins suffer from is the lack of sufficient contrast between the light and dark states of these proteins. In this work, hLOV1 is partnered with the engineered cpLOV to act as a dual protein cage and provide tighter caging in the absence of light. As a prototype, tobacco etch virus protease cleavage site (TEVcs) is shielded through a dual cage of hLOV1 and cpLOV1. This doubly caged TEVcs is integrated into a highly sensitive transcriptional assay. In dark conditions, TEVcs is caged, inaccessible to the protease and therefore cannot be cleaved. Upon photo-stimulation and availability of dopamine, TEVcs is uncaged and cleaved by the protease which leads to the downstream expression of the reporter mCherry gene. Geng et al. note a 90% reduction in background activity and demonstrated that dual caging is an effective approach to minimize undesirable background activities of photosensitive proteins.

Why I like this preprint

I am interested in designing photo-controlled helicases and have employed LOV domain as a means to introduce structural modulations to proteins in a light-based manner. This preprint as well as a recently published article by He et al. (6) have introduced circularly permuted versions of LOV that improves the functionality of LOV as a photo-sensing domain. These new circularly permuted versions of LOV can be especially useful for proteins that require a free N-terminus to function. I would be interested in inserting my protein of interest between the PAS domain and the Jα-helix of cpLOV and explore that aspect of cpLOV. The dual caging module that combines hLOV1 and cpLOV to improve the contrast between the light and dark states of the photo-controlled proteins addresses one of the limitations of various optogenetics tools.

Questions for the authors

  1. How does the lifetime of light/dark states of cpLOV compare to that of hLOV1?
  2. For the dual caging, a combination of cpLOV and AsLOV2 was used to simultaneously cage the N- and C- terminus of the protein of interest. Is it possible to use two domains of cpLOV instead to achieve similar levels of dual caging effects?


  1. Losi et al. (2018). Blue-light receptors for optogenetics. Chemical Reviews 118,
  2. Halavaty et al. (2007). N- and C-Terminal Flanking Regions Modulate Light-Induced Signal Transduction in the LOV2 Domain of the Blue Light Sensor Phototropin 1 from Avena sativa. Biochemistry 46, 49.
  3. Baird et al. (1999). Circular permutation and receptor insertion within green fluorescent proteins. PNAS 96, 11241
  4. Kim et al. (2017). Time gated detection of protein-protein interactions with transcriptional readout. eLife 6, e30233.
  5. Song et al. (2003). Structural basis of degradation signal recognition by SspB, a specificity-enhancing factor for the ClpXP proteolytic machine. Molecular cell 12,1.
  6. He et al. (2021). Circularly permuted LOV2 as a modular photoswitch for optogenetic engineering. Nature Chemical Biology

Tags: circular permutation, lov, optogenetics, protein engineeering

Posted on: 1st July 2021 , updated on: 13th July 2021


Read preprint (No Ratings Yet)

Author's response

Lequn Geng and Jiaqi Shen shared

How does the lifetime of light/dark states of cpLOV compare to that of hLOV1?
We haven’t directly measured the on-off kinetics of our cpLOV. However, we noticed from the SsrA-SspB binding assay that if the light samples were allowed to sit in the dark for several minutes before APEX labeling, there was a decrease in signal. This indicates that cpLOV likely reverts back from light state to dark state within minutes. A recently published cpLOV design ( similar to ours measured the photo-responsiveness of their cpLOV and found the activation half-life and deactivation half-life to be 10 s and 1 min, respectively. These measurements are in line with the non-circularly-permuted AsLOV2 domain.
For the dual caging, a combination of cpLOV and AsLOV2 was used to simultaneously cage the N- and C- terminus of the protein of interest. Is it possible to use two domains of cpLOV instead to achieve similar levels of dual caging effects?
With the direct fusion strategy (not embedding) used by our work, using two domains of cpLOV in the same construct would require the N-terminal cpLOV to fuse to the effector via the non-J-alpha helix end, and we don’t expect this to offer good caging. To create the most caging effects, prior literature almost exclusively fuses effectors with the J-alpha helix region on AsLOV2. This is because the helix docks to the LOV core in the dark state and unwinds from the LOV core in the light state. The only published attempt (, to our knowledge, to use the non-J-alpha helix end of AsLOV2 to cage an effector found this strategy ineffective. For similar reasons, we believe it would be best to use the J-alpha helix region on the cpLOV to cage peptides/proteins of interest. Therefore, the best dual caging strategy is theoretically to use AsLOV2 to the N-terminus and cpLOV to the C-terminus of the target.

Have your say

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.

Sign up to customise the site to your preferences and to receive alerts

Register here