*Article* **Molecular Determinants of Fibrillation in a Viral Amyloidogenic Domain from Combined Biochemical and Biophysical Studies**

**Juliet F. Nilsson <sup>1</sup> , Hakima Baroudi <sup>1</sup> , Frank Gondelaud <sup>1</sup> , Giulia Pesce <sup>1</sup> , Christophe Bignon <sup>1</sup> , Denis Ptchelkine <sup>1</sup> , Joseph Chamieh <sup>2</sup> , Hervé Cottet <sup>2</sup> , Andrey V. Kajava <sup>3</sup> and Sonia Longhi 1,\***


**Abstract:** The Nipah and Hendra viruses (NiV and HeV) are biosafety level 4 human pathogens classified within the *Henipavirus* genus of the *Paramyxoviridae* family. In both NiV and HeV, the gene encoding the Phosphoprotein (P protein), an essential polymerase cofactor, also encodes the V and W proteins. These three proteins, which share an intrinsically disordered N-terminal domain (NTD) and have unique C-terminal domains (CTD), are all known to counteract the host innate immune response, with V and W acting by either counteracting or inhibiting Interferon (IFN) signaling. Recently, the ability of a short region within the shared NTD (i.e., PNT3) to form amyloid-like structures was reported. Here, we evaluated the relevance of each of three contiguous tyrosine residues located in a previously identified amyloidogenic motif (EYYY) within HeV PNT3 to the fibrillation process. Our results indicate that removal of a single tyrosine in this motif significantly decreases the ability to form fibrils independently of position, mainly affecting the elongation phase. In addition, we show that the C-terminal half of PNT3 has an inhibitory effect on fibril formation that may act as a molecular shield and could thus be a key domain in the regulation of PNT3 fibrillation. Finally, the kinetics of fibril formation for the two PNT3 variants with the highest and the lowest fibrillation propensity were studied by Taylor Dispersion Analysis (TDA). The results herein presented shed light onto the molecular mechanisms involved in fibril formation.

**Keywords:** paramyxoviruses; Hendra virus; intrinsically disordered proteins; amyloid-like fibrils; Taylor Dispersion Analysis (TDA); negative staining Transmission Electron Microscopy (ns-TEM); Polyethylene glycol (PEG) precipitation assays; Congo Red; Small-Angle X-ray Scattering (SAXS)

#### **1. Introduction**

The Hendra virus (HeV), together with the closely related Nipah virus (NiV), is a Biosafety Level 4 (BSL-4) pathogen belonging to the *Henipavirus* genus within the *Paramyxoviridae* family [1]. Henipaviruses are zoonotic viruses responsible in humans for severe encephalitis [1]. They are enveloped viruses with a non-segmented, single-stranded RNA genome of negative polarity [2]. Their genome is wrapped by the nucleoprotein (N) within a helical nucleocapsid that is the template used by the viral polymerase for transcription and replication. The polymerase consists of the L protein, which bears all the enzymatic activities, and of the phosphoprotein (P). P serves as an indispensable polymerase co-factor as not only it tethers the L protein onto the nucleocapsid, but also keeps L in a soluble and competent form for transcription and replication [3–5].

**Citation:** Nilsson, J.F.; Baroudi, H.; Gondelaud, F.; Pesce, G.; Bignon, C.; Ptchelkine, D.; Chamieh, J.; Cottet, H.; Kajava, A.V.; Longhi, S. Molecular Determinants of Fibrillation in a Viral Amyloidogenic Domain from Combined Biochemical and Biophysical Studies. *Int. J. Mol. Sci.* **2023**, *24*, 399. https://doi.org/ 10.3390/ijms24010399

Academic Editor: Ludmilla A. Morozova-Roche

Received: 17 November 2022 Revised: 16 December 2022 Accepted: 19 December 2022 Published: 26 December 2022

**Copyright:** © 2022 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

The N and P proteins from Henipaviruses encompass long intrinsically disordered regions (IDRs) [6–8], i.e., regions devoid of stable secondary and tertiary structure [9–12]. The *Henipavirus* P protein consists of a long N-terminal intrinsically disordered domain (NTD) and a C-terminal region that possesses both structured and disordered regions (Figure 1) [7,8,13–18]. L in a soluble and competent form for transcription and replication [3–5]. The N and P proteins from Henipaviruses encompass long intrinsically disordered regions (IDRs) [6–8], i.e., regions devoid of stable secondary and tertiary structure [9–12]. The *Henipavirus* P protein consists of a long N-terminal intrinsically disordered domain (NTD) and a C-terminal region that possesses both structured and disordered regions (Figure 1) [7,8,13–18].

enzymatic activities, and of the phosphoprotein (P). P serves as an indispensable polymerase co-factor as not only it tethers the L protein onto the nucleocapsid, but also keeps

*Int. J. Mol. Sci.* **2022**, *23*, x FOR PEER REVIEW 2 of 25

**Figure 1.** Schematic illustration of the HeV particle, of the organization of the P, V and W proteins and sequence of the HeV PNT3 region shared by the P, V, and W proteins. The left panel displays a scheme of the HeV virion, with the genome encapsidated by the nucleoprotein (in green) and the P (yellow) and the Large (wheat) proteins attached onto the nucleocapsid. The central panel displays a scheme of the P, V and W protein organization, showing that they share a common N-terminal domain (NTD) and have distinct C-terminal domains (CTD). PMD: P multimerization domain; XD: X domain of P. Interaction sites with proteins associated with the host innate immune response are shown. STAT: Signal Transducers and Activators of Transcription; PLK1: Polo-Like Kinase 1; MDA5: Melanoma Differentiation-Associated protein 5; TLR3: Toll-like receptor 3. Pink bars correspond to cysteine residues within NTD. The localization and the sequence of the PNT3 region within NTD is shown in grey and its sequence is displayed in the blue, right panel. The EYYY motif is shown as a green square and is framed within the PNT3 sequence. As in many paramyxoviruses [19], the P gene from HeV and NiV also encodes the C, **Figure 1.** Schematic illustration of the HeV particle, of the organization of the P, V and W proteins and sequence of the HeV PNT3 region shared by the P, V, and W proteins. The left panel displays a scheme of the HeV virion, with the genome encapsidated by the nucleoprotein (in green) and the P (yellow) and the Large (wheat) proteins attached onto the nucleocapsid. The central panel displays a scheme of the P, V and W protein organization, showing that they share a common N-terminal domain (NTD) and have distinct C-terminal domains (CTD). PMD: P multimerization domain; XD: X domain of P. Interaction sites with proteins associated with the host innate immune response are shown. STAT: Signal Transducers and Activators of Transcription; PLK1: Polo-Like Kinase 1; MDA5: Melanoma Differentiation-Associated protein 5; TLR3: Toll-like receptor 3. Pink bars correspond to cysteine residues within NTD. The localization and the sequence of the PNT3 region within NTD is shown in grey and its sequence is displayed in the blue, right panel. The EYYY motif is shown as a green square and is framed within the PNT3 sequence.

V and W non-structural proteins. While the C protein is encoded from an alternative reading frame within the P gene, the V and W proteins (~50 kDa) result from a mechanism of co-transcriptional editing of the P messenger: the addition of either one or two non-templated guanosines at the editing site of the P messenger yields the V and W proteins, respectively (Figure 1). The editing site is located at the end of the NTD-encoding region (Figure 1). The P, V and W proteins therefore share a common NTD but have distinct Cterminal domains (CTDs) (Figure 1). While the CTD of V adopts a zinc-finger conformation [20], the CTD of W is disordered [21]. The V and W proteins are key players in the evasion of the antiviral type I interferon As in many paramyxoviruses [19], the P gene from HeV and NiV also encodes the C, V and W non-structural proteins. While the C protein is encoded from an alternative reading frame within the P gene, the V and W proteins (~50 kDa) result from a mechanism of cotranscriptional editing of the P messenger: the addition of either one or two non-templated guanosines at the editing site of the P messenger yields the V and W proteins, respectively (Figure 1). The editing site is located at the end of the NTD-encoding region (Figure 1). The P, V and W proteins therefore share a common NTD but have distinct C-terminal domains (CTDs) (Figure 1). While the CTD of V adopts a zinc-finger conformation [20], the CTD of W is disordered [21].

(IFN-I)-mediated response [22–24]. This property relies on their ability to bind to a number of key cellular proteins involved in the antiviral response (Figure 1). We previously reported the ability of the HeV V protein to undergo a liquid-to-gel The V and W proteins are key players in the evasion of the antiviral type I interferon (IFN-I)-mediated response [22–24]. This property relies on their ability to bind to a number of key cellular proteins involved in the antiviral response (Figure 1).

transition, with a region within the NTD (referred to as PNT3, aa 200-310) (Figure 1) being identified as responsible for this behavior [25]. In those previous studies, we characterized PNT3 using a combination of biophysical and structural approaches. Congo Red (CR) binding assays, together with negative-staining transmission electron microscopy (ns-TEM) studies, showed that PNT3 forms amyloid-like structures [25]. Noteworthy, Congo red staining experiments provided hints that these amyloid-like fibrils form not only in vitro but also *in cellula* after transfection or infection suggesting a probable functional role. In light of the critical role of the *Henipavirus* V and W proteins in evading the host innate immune response, we previously proposed that in infected cells PNT3-mediated fibrillar We previously reported the ability of the HeV V protein to undergo a liquid-to-gel transition, with a region within the NTD (referred to as PNT3, aa 200–310) (Figure 1) being identified as responsible for this behavior [25]. In those previous studies, we characterized PNT3 using a combination of biophysical and structural approaches. Congo Red (CR) binding assays, together with negative-staining transmission electron microscopy (ns-TEM) studies, showed that PNT3 forms amyloid-like structures [25]. Noteworthy, Congo red staining experiments provided hints that these amyloid-like fibrils form not only in vitro but also *in cellula* after transfection or infection suggesting a probable functional role. In light of the critical role of the *Henipavirus* V and W proteins in evading the host innate immune response, we previously proposed that in infected cells PNT3-mediated fibrillar aggregates could sequester key cellular proteins involved in the antiviral response. In particular, sequestration of STAT and 14-3-3 proteins would lead to prevention of IFN

signaling and abrogation of the NF-κB-induced proinflammatory response [26]. Consistent with the presence of PNT3 within their NTD, the *Henipavirus* W proteins were shown to be able to form amyloid-like fibrils as well [21].

Within PNT3, a motif encompassing three contiguous tyrosines (EYYY) was predicted as an amyloidogenic region [25]. The ArchCandy predictor [27,28] also predicted a fibril architecture in which the three contiguous tyrosines of the motif are part of the first βstrand of a β-strand-loop-β-strand motif (see Figure 4C in [25]). ArchCandy identifies amyloidogenic regions based on their ability to form β-arcades. Indeed, the core structural element of a majority of naturally occurring and disease-related amyloid fibrils is a βarcade representing a parallel and in register stacks of β-strand-loop-β-strand motifs called β-arches [27]. Substitution of the three contiguous tyrosine residues with three alanine residues yielded a variant (referred to as PNT33A) that was shown to possess a dramatically reduced fibrillation ability, thus providing direct experimental evidence for the predicted involvement of the EYYY motif in building up the core of the fibrils [25].

Here, with the aim of achieving a better understanding of the molecular determinants of HeV PNT3 fibrillation, we designed and characterized a set of additional PNT3 variants that were conceived to either further investigate the EYYY amyloidogenic motif or probe the contribution of the C-terminal half of the protein to the fibrillation process. Results, as obtained by combining various biophysical and structural approaches, show that removal of one out of the three tyrosines of the motif, irrespective of position, is sufficient to lead to a significantly reduced fibrillation ability. In addition, our results revealed that the C-terminal half of PNT3 acts as a natural dampener of the fibrillation process.

#### **2. Results and Discussion**

#### *2.1. Influence of pH on the Formation of HeV PNT3 Amyloid-like Fibrils*

We previously documented the ability of the PNT3 region of the HeV V protein to form amyloid-like structures [25]. A number of studies reported an impact of pH on both fibril structure and kinetics [29–31]. As a first step towards an in-depth characterization of PNT3 fibrils, and with the aim of selecting appropriate conditions to investigate the kinetics of fibril formation, we sought at assessing the possible impact of pH on the fibrillation process. To this end, we used a previously described method based on the titration of polyethylene glycol (PEG), a crowding agent, to quantitatively assess the relative solubility of proteins [32]. Hence, after optimization of this method (see Materials and Methods), we performed PEG precipitation assays to evaluate the relative solubility of HeV PNT3 at three different pH values, namely 6.5, 7.2 and 8.0 (Figure 2). From this assay, the PEG1/2 value, which corresponds to the PEG concentration at which 50% of the protein is still soluble, can be obtained and allows comparing protein aggregation propensities under different conditions [32]. Results display that HeV PNT3 at pH 6.5 shows less relative solubility compared to pH 7.2, indicating a higher aggregation propensity at the lower pH (Figure 2A). This behavior might be at least partly rationalized based on the isoelectric point of the protein (pI = 4.6), as proteins are well known to display minimal solubility at pH values close to their pI. The impact of pH on fibril formation, and the correlation with the PEG1/2 value, was confirmed by ns-TEM (Figure 2B). The obtained micrographs show that at pH 6.5 fibrillar aggregates are present even at time 0, while at pH 7.2 equivalent fibrillar aggregates are only observed after an incubation of 96 h at 37 ◦C (i.e., no fibrillar aggregates can be detected at time 0, Figure 2B). This trend is further confirmed at pH 8, a condition where PNT3 displays the lowest propensity to form fibrillar aggregates (Figure 2). In line with expectations, at pH 4, a value closer to the pI of the protein, the sample was found to exhibit a strongly reduced solubility (Supplementary Figure S1), and ns-TEM studies showed mainly amorphous aggregates rather than fibrillar aggregates (Figure 2B). Notably, in addition to the presence of fibrils, the micrographs obtained at all the pH values also show the presence of amorphous aggregates. These results, beyond advocating for a role of electrostatics in the aggregation process, prompted us to define pH 7.2 as the standard pH value for further studies: the rationale for this choice was that we wanted to

*Int. J. Mol. Sci.* **2022**, *23*, x FOR PEER REVIEW 4 of 25

be able to monitor the appearance and growth of fibrils over time, while at pH 6.5 fibrils were detected as early as at time zero. as the standard pH value for further studies: the rationale for this choice was that we wanted to be able to monitor the appearance and growth of fibrils over time, while at pH 6.5 fibrils were detected as early as at time zero.

pH values also show the presence of amorphous aggregates. These results, beyond advocating for a role of electrostatics in the aggregation process, prompted us to define pH 7.2

**Figure 2.** Fibrillation propensity of HeV PNT3 *wild-type* (*wt*) at different pHs. (**A**) PEG assay and relative solubility of HeV PNT3 *wt* at pH 7.2 (green), 6.5 (violet) and 8.0 (light gray). The vertical lines correspond to PEG1/2 values (with their 95% confidence intervals) as obtained after a normalization and fitting step to a sigmoid function. Note that the curve obtained at pH 6.5 exhibits poor fitting thus preventing calculation of the confidence interval. Data points at pH 8.0 could obviously not be fitted. (**B**) Ns-TEM of HeV PNT3 *wt* fibrils at four different pH values: 6.5 (at time 0, t:0 h), **Figure 2.** Fibrillation propensity of HeV PNT3 *wild-type* (*wt*) at different pHs. (**A**) PEG assay and relative solubility of HeV PNT3 *wt* at pH 7.2 (green), 6.5 (violet) and 8.0 (light gray). The vertical lines correspond to PEG1/2 values (with their 95% confidence intervals) as obtained after a normalization and fitting step to a sigmoid function. Note that the curve obtained at pH 6.5 exhibits poor fitting thus preventing calculation of the confidence interval. Data points at pH 8.0 could obviously not be fitted. (**B**) Ns-TEM of HeV PNT3 *wt* fibrils at four different pH values: 6.5 (at time 0, t: 0 h), 7.2, 8.0 (at time 0 and after 96 h of incubation at 37 ◦C, t: 96 h) and 4.0 (after 96 h of incubation at 37 ◦C, t: 96 h). White arrows indicate fibrils.

#### *2.2. Rational Design and Generation of PNT3 Variants 2.2. Rational Design and Generation of PNT3 Variants*

°C, t:96 h). White arrows indicate fibrils.

#### 2.2.1. Design of PNT3 Variants Targeting the EYYY Motif 2.2.1. Design of PNT3 Variants Targeting the EYYY Motif

*Int. J. Mol. Sci.* **2022**, *23*, x FOR PEER REVIEW 5 of 25

Within the HeV PNT3 region, we previously identified an amyloidogenic motif encompassing three contiguous tyrosine residues (EYYY) [25]. The relevance of this motif for fibril formation was experimentally confirmed through the generation of the PNT33A variant, in which the three tyrosine residues were replaced with three alanine residues, which showed a reduced ability to form amyloid-like fibrils [25]. With the goal of further investigating this amyloidogenic motif and of unveiling whether all the three tyrosine residues were critical for fibril formation or whether only a subset of them was so, we rationally designed three single-site PNT3 variants where each one of the three contiguous tyrosine residues was replaced with one alanine (PNT3A1, PNT3A2 and PNT3A3, Figure 3A). In this context, a construct encoding the corresponding PNT3 region from the NiV V protein (NiV PNT3) was also generated taking advantage of the fact that the corresponding NiV PNT3 motif encompasses only 2 contiguous tyrosine residues (EHYY) (Figure 3A and Supplementary Figure S2). Within the HeV PNT3 region, we previously identified an amyloidogenic motif encompassing three contiguous tyrosine residues (EYYY) [25]. The relevance of this motif for fibril formation was experimentally confirmed through the generation of the PNT33A variant, in which the three tyrosine residues were replaced with three alanine residues, which showed a reduced ability to form amyloid-like fibrils [25]. With the goal of further investigating this amyloidogenic motif and of unveiling whether all the three tyrosine residues were critical for fibril formation or whether only a subset of them was so, we rationally designed three single-site PNT3 variants where each one of the three contiguous tyrosine residues was replaced with one alanine (PNT3A1, PNT3A2 and PNT3A3, Figure 3A). In this context, a construct encoding the corresponding PNT3 region from the NiV V protein (NiV PNT3) was also generated taking advantage of the fact that the corresponding NiV PNT3 motif encompasses only 2 contiguous tyrosine residues (EHYY) (Figure 3A and Supplementary Figure S2).

7.2, 8.0 (at time 0 and after 96 h of incubation at 37 °C, t:96 h) and 4.0 (after 96 h of incubation at 37

**Figure 3.** (**A**) Schematic diagram of HeV PNT3 variants designed and constructed in this work. The EYYY motif of the HeV PNT3 is shows in green. (**B**) SDS-PAGE analysis of the purified proteins. **Figure 3.** (**A**) Schematic diagram of HeV PNT3 variants designed and constructed in this work. The EYYY motif of the HeV PNT3 is shows in green. (**B**) SDS-PAGE analysis of the purified proteins.

#### 2.2.2. Design of HeV PNT3 Truncated Variants Devoid of the C-Terminal Region 2.2.2. Design of HeV PNT3 Truncated Variants Devoid of the C-Terminal Region

As mentioned above, the EYYY motif plays a significant role in fibril formation. However, the fact that the substitution of the triple tyrosine motif only reduces but does not fully abrogate the ability of PNT3 to form amyloid-like fibrils [25] indicates that other motifs and/or sequence attributes, that remained to be identified, contribute to the fibrillation process. Hence, to deepen the characterization and to assess the possible contribution of the PNT3 C-terminal region, we designed a C-terminally truncated HeV PNT3 variant (PNT3 C-term truncated) that lacks the second half of the protein (Figure 3A). In addition, we also designed a C-terminally truncated HeV PNT3 variant where the three contiguous tyrosines of the EYYY motif were replaced with three alanines (PNT33A C-term truncated) (Figure 3A). As mentioned above, the EYYY motif plays a significant role in fibril formation. However, the fact that the substitution of the triple tyrosine motif only reduces but does not fully abrogate the ability of PNT3 to form amyloid-like fibrils [25] indicates that other motifs and/or sequence attributes, that remained to be identified, contribute to the fibrillation process. Hence, to deepen the characterization and to assess the possible contribution of the PNT3 C-terminal region, we designed a C-terminally truncated HeV PNT3 variant (PNT3 C-term truncated) that lacks the second half of the protein (Figure 3A). In addition, we also designed a C-terminally truncated HeV PNT3 variant where the three contiguous tyrosines of the EYYY motif were replaced with three alanines (PNT33A C-term truncated) (Figure 3A).

#### 2.2.3. Design of a HeV PNT3 Variant Bearing a Unique Cysteine 2.2.3. Design of a HeV PNT3 Variant Bearing a Unique Cysteine

The N-terminal domain, shared by the HeV P, V, and W proteins, has 3 cysteines distributed along the sequence (Figure 1). Recently, Pesce & Gondelaud et al. suggested that disulfide bridges could be involved in preventing aggregation of the W protein [21]. Thus, in this context, we reasoned that a HeV PNT3 variant bearing a cysteine residue The N-terminal domain, shared by the HeV P, V, and W proteins, has 3 cysteines distributed along the sequence (Figure 1). Recently, Pesce & Gondelaud et al. suggested that disulfide bridges could be involved in preventing aggregation of the W protein [21]. Thus, in this context, we reasoned that a HeV PNT3 variant bearing a cysteine residue could be useful to investigate the possible impact of disulfide bridge-mediated protein dimerization on the fibrillation abilities of PNT3. We targeted for cysteine substitution the unique alanine residue of PNT3 (Ala255) to yield PNT3 variant A255C (PNT3\_Cys)

(Figure 3A). The rationale for choosing an alanine, rather than a serine residue which would have enabled a more isosteric substitution, was to introduce as much a conservative as possible substitution, while preserving the content in OH groups, which might play a role in the establishment of stabilizing inter-chain interactions in the core of the fibrils.

#### 2.2.4. Expression and Purification of the PNT3 Variants

All the proteins were expressed in *E. coli* as hexahistidine tagged forms with no solubility tag. The proteins were purified from the total fraction of the bacterial lysate under denaturing conditions by Immobilized Metal Affinity Chromatography (IMAC) and size exclusion chromatography (SEC). The purity of the final purified products was assessed by SDS-PAGE (Figure 3B). The identity of all the variants was confirmed by mass spectrometry (MS) analysis of peptides resulting from the tryptic digestion of each protein (Supplementary Figure S3).

#### 2.2.5. Conformational Characterization of the PNT3 Variants

First, we evaluated the hydrodynamic properties of all the variants through analytical SEC. Table 1 shows the Stokes radius (RS) value inferred for each protein, along with the Compaction Index (CI) associated with each variant. By comparing the mean measured Stokes radius (R<sup>S</sup> OBS) with the theoretical Stokes radii expected for the various conformational states (i.e., R<sup>S</sup> NF: natively folded protein; R<sup>S</sup> PMG: premolten globule, PMG; R<sup>S</sup> U: fully unfolded form; R<sup>S</sup> IDP: IDP), all the proteins were found to have R<sup>S</sup> values consistent with a PMG state [33]. Their compaction indexes are relatively close to each other, with the notable exception of the HeV PNT3\_C-term\_truncated variant that is much more compact. Strikingly, the introduction of the triple alanine motif in the context of the truncated variant leads to a more extended conformation, a phenomenon already observed, although with a borderline significance, in the context of the full-length PNT3 protein (*cf.* HeV PNT3 *wt* and HeV PNT33A in Table 1). These results suggest that the second half of the protein is a determinant of chain expansion, with this effect being counteracted by the presence of the triple alanine motif.

**Table 1.** Stokes radii (R<sup>S</sup> OBS, Å) of the PNT3 variants as inferred from the elution volume of the major SEC peak. Shown are also the expected values for the various conformational states, along with the ratios between the R<sup>S</sup> OBS and each R<sup>S</sup> state, and compaction index (CI) values.


RS OBS: experimentally observed Stokes radius (mean value and s.d. from three independent experiments); R<sup>S</sup> NF: R<sup>S</sup> expected for a natively folded (NF) form; R<sup>S</sup> PMG: R<sup>S</sup> expected for a pre-molten globule (PMG); R<sup>S</sup> <sup>U</sup>: R<sup>S</sup> expected for a fully unfolded form; R<sup>S</sup> IDP: R<sup>S</sup> expected for an IDP based on the simple power law model; all radii are given in Å. Mass: molecular mass (Daltons) calculate d from the amino acid sequence of the recombinant protein. Compaction index (CI) mean values and s.d., as obtained from three independent experiments.

In order to evaluate the secondary structure content of each variant, we performed a Circular Dichroism (CD) analysis in the far ultraviolet (UV) region. All the variants present a spectrum typical of a disordered protein lacking any stable organized secondary structure, as judged from the large negative peak centered at 200 nm, and from the low ellipticity in the 220–230 nm region and at 190 nm (see [34] and references therein cited) (Supplementary Figure S4). These results indicate that the introduced substitutions and/or the truncation impact only marginally, if at all, the secondary structure content of the protein. They also indicate that NiV PNT3 has a secondary structure content very close to

that of its HeV counterpart. The finding that the CD spectra of PNT3 variants bearing the triple alanine motif are virtually superimposable onto those of variants bearing either the naturally occurring triple tyrosine motif or just one Tyr to Ala substitution, rules out the possibility that the expansion effect driven by the triple alanine motif, as observed in SEC studies, may arise from the presence of a transiently populated α-helix encompassing the motif. Therefore, the mechanism underlying the counteracting effect exerted by the triple alanine motif on chain compaction remains to be elucidated.

To achieve a more quantitative description of the conformational properties of the variants, we carried out Small-Angle X-ray Scattering (SAXS) studies coupled to SEC (SEC-SAXS). We selected a set of representative variants (i.e., HeV PNT33A, NiV PNT3, HeV PNT3\_C-term\_truncated, and HeV PNT33A\_C-term\_truncated) along with HeV PNT3 *wt*. Although we already previously reported SEC-SAXS studies of HeV PNT3 *wt* [25], this sample was herein again investigated under exactly the same conditions used for the other variants, so as to enable meaningful comparisons. For all the five PNT3 proteins, linearity of the Guinier region in the resulting scattering curves (Supplementary Figure S5A) allowed meaningful estimations of the radius of gyration (Rg) (Table 2). The R<sup>g</sup> values obtained for HeV PNT33A, NiV PNT3, and HeV PNT3 *wt* are very close to each other's clustering in a group with R<sup>g</sup> around 37–39 Å. As expected, the two truncated variants have smaller, and close to each other's, R<sup>g</sup> values (Table 2).

**Table 2.** *R<sup>g</sup>* and *Dmax* as obtained from SEC-SAXS studies and expected values for the various conformational states.


I(0): Intensity at zero angle as determined from Guinier approximation; R<sup>g</sup> Guinier: R<sup>g</sup> values as obtained from Guinier approximation; Dmax: maximal intramolecular distance from P(r). R<sup>g</sup> IDP: R<sup>g</sup> expected for an IDP based on the simple power-law model. R<sup>g</sup> <sup>U</sup>: theoretical R<sup>g</sup> value expected for a chemically denatured (U) protein.

Notably, all obtained experimental R<sup>g</sup> values are close to the theoretical R<sup>g</sup> <sup>U</sup>, corresponding to chemically denatured (U) proteins, and hence reflecting a highly extended conformation. Because of this, the Rg-based CI could not be computed, the numerator (R<sup>g</sup> <sup>U</sup> <sup>−</sup> <sup>R</sup><sup>g</sup> OBS) in Equation (10) being <sup>≤</sup> 0 (see Section 3). These results are in contrast with the previous SEC results, where the variants were found to adopt a PMG conformation. A possible explanation for this might be related to the differences in the buffer used in the two techniques. As expected, the five variants were all found to be disordered as judged from the presence of a plateau in the normalized Kratky (Supplementary Figure S5B) and Kratky-Debye plots (Supplementary Figure S5C). However, the two truncated variants, and particularly the HeV PNT3 C-term truncated one, showed a slight deviation from the pure random-coil regime as observed in the normalized Kratky plot. This deviation is consistent with a slightly more compact conformation, in line with the RS-based CI values discussed above.

#### *2.3. Relevance of the PNT3 EYYY Motif in Fibrillation Abilities*

#### 2.3.1. Aggregation Propensity of the EYYY Motif PNT3 Variants

With the aim of elucidating the contribution of each tyrosine in the EYYY motif to the fibrillation process, we first assessed the aggregation propensity of the set of variants bearing alanine substitutions within the EYYY amyloidogenic motif. To this end, we took advantage of the same PEG solubility assay described above [32]. For each of the five EYYY

variants (PNT33A, PNT3A1, A2, A3 and NiV PNT3), we therefore carried out PEG solubility assays which enabled ranking them based on their estimated PEG1/2 value. vantage of the same PEG solubility assay described above [32]. For each of the five EYYY variants (PNT33A, PNT3A1, A2, A3 and NiV PNT3), we therefore carried out PEG solubility

With the aim of elucidating the contribution of each tyrosine in the EYYY motif to the fibrillation process, we first assessed the aggregation propensity of the set of variants bearing alanine substitutions within the EYYY amyloidogenic motif. To this end, we took ad-

plateau in the normalized Kratky (Supplementary Figure S5B) and Kratky-Debye plots (Supplementary Figure S5C). However, the two truncated variants, and particularly the HeV PNT3 C-term truncated one, showed a slight deviation from the pure random-coil regime as observed in the normalized Kratky plot. This deviation is consistent with a slightly more com-

The PNT33A variant shows a significant increase in the relative solubility compared to PNT3 HeV *wt* (Figure 4A,B), a result in agreement with previous findings that pointed out a much lower propensity to form amyloid-like fibrils for PNT33A [25]. The three variants where only one Tyr was replaced, however, show no significant difference in the PEG1/2 values compared to HeV PNT3 *wt* (Figure 4B), suggesting that the removal of one tyrosine does not affect the aggregation propensity of PNT3. By contrast, and interestingly, NiV PNT3 (EHYY) displays an intermediate relative solubility between HeV PNT3 *wt* and HeV PNT33A (Figure 4A). In light of the results obtained with the HeV PNT3 variants bearing two tyrosines (i.e., PNT3A1, PNT3A2 and PNT3A3), the intermediate aggregation propensity of NiV PNT3 more likely arise from differences in the amino acid context rather than from the absence of just one tyrosine. assays which enabled ranking them based on their estimated PEG1/2 value. The PNT33A variant shows a significant increase in the relative solubility compared to PNT3 HeV *wt* (Figure 4A,B), a result in agreement with previous findings that pointed out a much lower propensity to form amyloid-like fibrils for PNT33A [25]. The three variants where only one Tyr was replaced, however, show no significant difference in the PEG1/2 values compared to HeV PNT3 *wt* (Figure 4B), suggesting that the removal of one tyrosine does not affect the aggregation propensity of PNT3. By contrast, and interestingly, NiV PNT3 (EHYY) displays an intermediate relative solubility between HeV PNT3 *wt* and HeV PNT33A (Figure 4A). In light of the results obtained with the HeV PNT3 variants bearing two tyrosines (i.e., PNT3A1, PNT3A2 and PNT3A3), the intermediate aggregation propensity of NiV PNT3 more likely arise from differences in the amino acid context rather than from the absence of just one tyrosine.

*Int. J. Mol. Sci.* **2022**, *23*, x FOR PEER REVIEW 8 of 25

*2.3. Relevance of the PNT3 EYYY Motif in Fibrillation Abilities*  2.3.1. Aggregation Propensity of the EYYY Motif PNT3 Variants

pact conformation, in line with the RS-based CI values discussed above.

**Figure 4.** PEG solubility assays for the set of PNT3 variants. (**A**) Soluble fraction of each variant at different PEG concentrations. HeV PNT3 *wt* (green), HeV PNT33A (red), HeV PNT3\_C-term\_truncated (yellow), NiV PNT3 (blue). Vertical lines represent PEG1/2 values including their 95% confidence intervals obtained after a normalization and fitting step to a sigmoid function. (**B**) PEG1/2 values obtained for the *wt*, the triple alanine variant and the single alanine HeV PNT3 variants. (**C**) PEG1/2 values obtained for the C-terminally truncated variants. The asterisk indicates statistically significant differences (One-way Anova test, *p*-value < 0.05). **Figure 4.** PEG solubility assays for the set of PNT3 variants. (**A**) Soluble fraction of each variant at different PEG concentrations. HeV PNT3 *wt* (green), HeV PNT33A (red), HeV PNT3\_C-term\_truncated (yellow), NiV PNT3 (blue). Vertical lines represent PEG1/2 values including their 95% confidence intervals obtained after a normalization and fitting step to a sigmoid function. (**B**) PEG1/2 values obtained for the *wt*, the triple alanine variant and the single alanine HeV PNT3 variants. (**C**) PEG1/2 values obtained for the C-terminally truncated variants. The asterisk indicates statistically significant differences (One-way Anova test, *p*-value < 0.05).

#### 2.3.2. Congo Red Binding Abilities of PNT3 EYYY Motif Variants

Congo Red is a widely used dye to document the presence of amyloids: binding of this dye to cross β-sheet structures in fact leads to hyperchromicity and a red shift in the absorbance maximum of the CR spectrum. Hence, to further characterize PNT3 EYYY motif variants, we took advantage of CR binding assays. We compared the binding abilities of the PNT3A1, PNT3A2, PNT3A3 and NiV variants to those of both PNT3 *wt* and PNT33A. We spectrophotometrically measured the red shift (from 497 nm to 515 nm) in the absorbance maximum of the CR spectrum of each sample following a four or seven days incubation at 37 ◦C. Results shown in Figure 5 indicate that all the variants promote a shift in the CR spectrum whose amplitude increases with incubation time, suggesting that all the variants are able to progressively form amyloid-like fibrils or at least structures able to bind CR.

Our previous findings that pointed out that the HeV PNT33A variant has a reduced ability to bind CR compared to HeV PNT3 *wt* [25] were confirmed here (Figure 5). Two of the three single alanine variants (PNT3A1, PNT3A3) show an intermediate behavior between HeV PNT3 *wt* and HeV PNT33A, but without significant differences with either the *wt* or the PNT33A variant, while HeV PNT3A2 displays an ability to bind CR significantly higher than that of HeV PNT33A and similar to that of HeV PNT3 *wt* (Figure 5). These results therefore suggest that the central tyrosine in the motif could be less relevant to the fibrillation process. Notably, the NiV PNT3 variant shows a significantly decreased ability to bind CR compared to PNT3 *wt*, similar to that of the triple alanine variant. Thus, as already observed for the aggregation propensity, the reduced CR binding ability of NiV PNT3 likely results from its amino acid context rather than from the fact that it lacks a tyrosine in the motif. that all the variants are able to progressively form amyloid-like fibrils or at least structures able to bind CR. Our previous findings that pointed out that the HeV PNT33A variant has a reduced ability to bind CR compared to HeV PNT3 *wt* [25] were confirmed here (Figure 5). Two of the three single alanine variants (PNT3A1, PNT3A3) show an intermediate behavior between HeV PNT3 *wt* and HeV PNT33A, but without significant differences with either the *wt* or the PNT33A variant, while HeV PNT3A2 displays an ability to bind CR significantly higher than that of HeV PNT33A and similar to that of HeV PNT3 *wt* (Figure 5). These results therefore suggest that the central tyrosine in the motif could be less relevant to the fibrillation process. Notably, the NiV PNT3 variant shows a significantly decreased ability to bind CR compared to PNT3 *wt*, similar to that of the triple alanine variant. Thus, as already observed for the aggregation propensity, the reduced CR binding ability of NiV PNT3 likely results from its amino acid context rather than from the fact that it lacks a tyrosine in the motif.

Congo Red is a widely used dye to document the presence of amyloids: binding of this dye to cross β-sheet structures in fact leads to hyperchromicity and a red shift in the absorbance maximum of the CR spectrum. Hence, to further characterize PNT3 EYYY motif variants, we took advantage of CR binding assays. We compared the binding abilities of the PNT3A1, PNT3A2, PNT3A3 and NiV variants to those of both PNT3 *wt* and PNT33A. We spectrophotometrically measured the red shift (from 497 nm to 515 nm) in the absorbance maximum of the CR spectrum of each sample following a four or seven days incubation at 37 °C. Results shown in Figure 5 indicate that all the variants promote a shift in the CR spectrum whose amplitude increases with incubation time, suggesting

*Int. J. Mol. Sci.* **2022**, *23*, x FOR PEER REVIEW 9 of 25

2.3.2. Congo Red Binding Abilities of PNT3 EYYY Motif Variants

**Figure 5.** Congo Red binding assay of the set of PNT3 variants. The ability to bind CR is represented by the fold increase in the ratio between the absorbance at 515 and at 497 nm, with respect to a sample containing CR alone, of PNT3 samples at 20 μM after 4 and 7 days of incubation at 37 °C. The error bar corresponds to the standard deviation, with n = 3. Different letters indicate statistically significant differences (*p* < 0.05) (Two-way ANOVA test; *ab* means lack of statistically significant differences with respect to *a* or *b*). **Figure 5.** Congo Red binding assay of the set of PNT3 variants. The ability to bind CR is represented by the fold increase in the ratio between the absorbance at 515 and at 497 nm, with respect to a sample containing CR alone, of PNT3 samples at 20 µM after 4 and 7 days of incubation at 37 ◦C. The error bar corresponds to the standard deviation, with n = 3. Different letters indicate statistically significant differences (*p* < 0.05) (Two-way ANOVA test; *ab* means lack of statistically significant differences with respect to *a* or *b*).

2.3.3. Propensity and Time-Dependance of Fibrillation of the PNT3 EYYY Motif Variants Using Negative-Staining Transmission Electron Microscopy (ns-TEM) 2.3.3. Propensity and Time-Dependance of Fibrillation of the PNT3 EYYY Motif Variants Using Negative-Staining Transmission Electron Microscopy (ns-TEM)

To directly document fibril formation by the PNT3 EYYY variants as a function of time and to obtain orthogonal experimental evidence corroborating the CR binding assay results, we next carried out ns-TEM studies. These analyses were performed for each of the variants at 0, 24 and 96 h of incubation at 37 ◦C (Figure 6). As shown Figure 6, although HeV PNT3 *wt* in the selected conditions does not form fibrils at time 0, after 24 h of incubation it forms short fibrils, which evolve to long fibrils after 96 h. In line with the CR binding results, ns-TEM studies confirmed that all PNT3 EYYY variants, including NiV PNT3, are able to form amyloid-like fibrils but with a significantly decreased ability compared to the *wt*. Specifically, after 24 h short fibrils can be observed in most variants, except for HeV PNT33A. In an attempt at identifying possible significant differences among the EYYY variants in spite of their overall similar behavior, we performed a comprehensive analysis of the number and length of fibrils detected for each of them (Figure 7). This

analysis revealed no significant differences in the length of the fibrils obtained after 96 h of incubation among the 3 variants with a single alanine (PNT3A1, PNT3A2 and PNT3A3). However, a slight, though significant, increase is observed in the PNT3A3 variant in the number of fibrils found *per* picture compared to the other two single alanine variants (Figure 7). This slightly higher number of short fibrils observed for the PNT3A3 variant may indicate that the last tyrosine in the motif contributes less to the nucleation process compared to the two other tyrosines. Remarkably, the same analysis showed that the NiV PNT3 variant forms significantly longer fibrils compared to the PNT3A1, PNT3A2 and PNT3A3 variants, therefore confirming that other sequence attributes, beyond the amyloidogenic motif, contribute to the fibrillation process. In other words, fibril formation would rely not only on stabilizing contacts mediated by the E(H/Y)YY motif but also on additional stabilizing interactions established by other protein regions. *Int. J. Mol. Sci.* **2022**, *23*, x FOR PEER REVIEW 11 of 25

**Figure 6.** *Cont*.

**Figure 6.** Fibril formation as a function of time by ns-TEM. Ns-TEM analysis of PNT3 variants (200 μM) at time zero and after 24 h or 96 h of incubation at 37 °C. Note that in all cases, samples were diluted to 40 μM prior to being deposited on the grid. White arrows indicate fibrils. The purple gradient represents the propensity to form fibrils. The green dashed box indicates HeV PNT3 *wt* results. **Figure 6.** Fibril formation as a function of time by ns-TEM. Ns-TEM analysis of PNT3 variants (200 µM) at time zero and after 24 h or 96 h of incubation at 37 ◦C. Note that in all cases, samples were diluted to 40 µM prior to being deposited on the grid. White arrows indicate fibrils. The purple gradient represents the propensity to form fibrils. The green dashed box indicates HeV PNT3 *wt* results.

**Figure 6.** Fibril formation as a function of time by ns-TEM. Ns-TEM analysis of PNT3 variants (200 μM) at time zero and after 24 h or 96 h of incubation at 37 °C. Note that in all cases, samples were diluted to 40 μM prior to being deposited on the grid. White arrows indicate fibrils. The purple

> **Figure 7.** Fibril length of EYYY variants. (**A**) For each variant, statistics on fibril length were obtained from analysis of the contour length of 60 fibrils after 96 h of incubation at 37 ◦C. The upper table shows the main statistical results. (**B**) Number of fibrils detected *per* picture for each variant. Upper number indicates the mean value of each variant. The analysis was done using the ImageJ software. The asterisk indicates statistically significant differences (One-way Anova test, *p*-value < 0.05). Circles: PNT3A1; squares: PNT3A2; triangles: PNT3A3, inverted triangles: NiV PNT3.

> Altogether, CR binding assays and ns-TEM enabled the documenting of slight differences among the PNT3A1, PNT3A2 and PNT3A3 variants. Although other regions beyond the amyloidogenic motif seemingly contribute to the fibrillation process, results strongly suggest that the absence of a single tyrosine in the EYYY motif leads to a significant decrease in the ability to form fibrils irrespective of the position. In light of the finding that the PNT3A1, PNT3A2 and PNT3A3 variants are able to form short fibrils, whose length does not increase under the experimental conditions herein used, we can assume that the three tyrosines mainly play a role in the elongation phase. The present results support the involvement of tyrosines in the formation of amyloid-like fibrils, with π-π stacking and H-bonding interactions between tyrosines likely allowing to form cross-β-like architectures, as previously suggested [35].

#### *2.4. Impact of the HeV PNT3 C-Terminal Region in Fibrillation Abilities* 2.4.1. Aggregation Propensity of C-Terminally Truncated PNT3 Variants

Taking into account the results presented above that lend support to a scenario where other motifs and/or sequence attributes beyond the amyloidogenic motif could be involved in the fibrillation process, we decided to investigate the impact of the PNT3 C-terminal region. We first studied the aggregation propensity of both PNT3 C-terminally truncated variants. Figure 4A shows that the PNT3 C-term truncated variant has the lowest PEG1/2 value, indicating a strikingly decreased relative solubility compared to all the full-length PNT3 variants. Notably, there are no significant differences in the PEG1/2 values between the PNT3 C-term truncated and its triple alanine mutant (PNT33A C-term truncated) (Figure 4C). These results indicate that removal of the C-terminal region has a strong impact on the aggregation propensity, with this effect being insensitive to the sequence context of the EYYY motif.

#### 2.4.2. CR Binding Ability of C-Terminally Truncated PNT3 Variants

Motivated by the results obtained by the PEG solubility assays pointing to a much higher aggregation propensity of both truncated variants, we next carried out CR binding assays. As shown in Figure 5, the HeV PNT3 C-terminal truncated variant displays a significantly increased ability to bind CR compared to full-length HeV PNT3 *wt*. In striking contrast with PEG solubility assays that detected no differences in terms of aggregation propensities between the two truncated variants, CR binding assays revealed significant differences between the two variants. In particular, the PNT33A truncated variant has a much-decreased ability to bind CR with respect to the truncated variant bearing a native EYYY motif, hence displaying an intermediate behavior between the full- length and truncated HeV PNT3 *wt* (Figure 5). Therefore, it can be concluded that the C-terminal region, far from being inert, negatively affects the ability of the protein to form CR-binding structures. Albeit the PNT33A truncated variant binds more CR than the *wt* variant, the contribution of the EYYY motif is evident when the two truncated variants are compared. These results, however, suggest that the EYYY motif would have only a marginal role in driving the formation of CR-binding structures in the context of C-terminally truncated form.

2.4.3. Propensity and Kinetics of Fibrillation of C-Terminally Truncated PNT3 Variants Using ns-TEM Studies

In order to ascertain whether the increased binding to CR and decreased solubility of the truncated variants is actually reflected in a higher fibrillation ability, we analyzed them by ns-TEM studies. In line with expectations, Figure 6 clearly shows a significantly higher abundance of fibrils, as well as an increased fibril length, in the PNT3 C-terminal truncated variant at short incubation times, thus confirming its increased fibrillation potential. Notably, the PNT33A truncated variant displays a decreased fibrillation ability compared to its *wt* counterpart, similar to the full-length HeV PNT3 *wt* (Figure 6). These findings suggest that the removal of the C-terminal region results in an acceleration of the fibrillation rate, reflecting an interaction between this region and the rest of the sequence that negatively affects the kinetics of fibril formation. Remarkably, similar results were previously documented in the case of the aggregation of α-synuclein (α-syn), an extensively characterized protein associated with neurodegeneration and whose transition from a soluble to a fibrillar form is thought to contribute to pathogenesis [36]. Compelling experimental evidence indicates that C-terminal truncation of α-syn promotes in vitro oligomer and fibril formation (see [37] and references therein cited). The middle region of α-syn, referred to as "nonamyloid component" (NAC) domain, forms the core of α-syn filaments. The C-terminal region of α-syn can adopt conformations in which the C-terminus contacts the hydrophobic NAC domain thus shielding it from pathological templating interactions [37]. The negative charge of the C-terminal region has been proposed to contribute to this self-chaperoning activity via the establishment of electrostatic interactions [37].

In an attempt at rationalizing the observed self-inhibitory effect of the C-terminal region of PNT3 on fibril formation, we analyzed the charge distribution within the PNT3 sequence using the CIDER server (http://pappulab.wustl.edu/CIDER/, accessed on 24 October 2022) [38] (Supplementary Figure S6 and Table S1). Although full-length PNT3, and its constituent N-terminal and C-terminal regions fall in very close positions in the phase diagram plot, the C-terminal region has a higher fraction of negatively charged residues compared to the N-terminal region (Supplementary Figure S6 and Table S1). In addition, the full-length form of PNT3 and the truncated variant strongly differ in their net charge at pH 7.0. Taking into account the strong impact of pH on PNT3 fibrillation, where a decrease to pH 6.5 strongly promotes fibrillation and leads to a behavior similar to that of the truncated PNT3 variant at pH 7.2, it is conceivable that the results obtained with the truncated variant could be, at least partly, accounted for by electrostatics, as in the case of α-syn. A plausible alternative scenario for the self-inhibitory effect of the C-terminal region of PNT3 on fibril formation could be the following: the disordered region downstream the fibril core may hamper fibril formation by slowing the disorder-to-order transition

expected to take place in the core of the fibril, through either a purely entropic effect or through a combination of enthalpy and entropy as already documented in the case of fuzzy appendages adjacent to molecular recognition elements (for examples see [39,40]).

Altogether, these findings advocate for a key role of the C-terminal region in regulating the fibrillation properties of PNT3. In particular, the C-terminal region may act either as a molecular shield, as in the case of α-syn [37], or by slowing down the rate of folding of the core of the fibrils, with this property, irrespective of the underlying mechanisms, being also possibly relevant to biological function in vivo. Definite answers on the precise molecular mechanisms and on the possible biological relevance await future studies.

#### *2.5. Impact of a Cysteine in the HeV PNT3 Sequence on Fibrillation Abilities*

With the aim of elucidating the possible impact of a disulfide bridge-mediated PNT3 dimerization on its fibrillation abilities, we first studied the ability of the HeV PNT3 variant bearing a cysteine residue (PNT3\_Cys) to bind CR. As shown in Figure 5, this variant shows a significant increased ability to bind CR compared to HeV PNT3 *wt*. Subsequently, the ability of this variant to form fibrils was assessed by ns-TEM in the same conditions used for PNT3 *wt*. Figure 6 shows that the PNT3\_Cys variant is able to form fibrils even at time 0, indicating a higher fibrillation propensity compared to HeV PNT3 *wt*. Notably, the PNT3\_Cys was the unique variant displaying an enrichment in shortened fibrils (Figure 6). These findings suggest that the presence of one cysteine in the sequence mainly impacts the nucleation phase. We reasoned that the peculiar fibrillation behavior of this variant could result from disulfide bridge-mediated protein dimerization. However, the addition of DTT was found to have a negligible impact, as judged from the presence at time 0 of very short fibrils and of long fibrils at 96 h (Supplementary Figure S7), resulting in an intermediate behavior between PNT3 *wt* and PNT3\_cys under non-reducing conditions. Thus, the peculiar behavior of this variant cannot be ascribed to protein dimerization and rather stems from other intrinsic properties that remain to be elucidated.

#### *2.6. Characterization of the Aggregation Process by Taylor Dispersion Analysis*

We next sought at shedding light onto the fibrillation kinetics using Taylor Dispersion Analysis (TDA) [41]. TDA is a new technique in the field of protein aggregation that has the notable advantages of being able to (i) capture intermediate species, (ii) quantify early and late-stage aggregates, and (iii) provide both kinetic and equilibrium constants. In addition, TDA is not dominated by aggregates (as opposite to scattering techniques), and is thus ideally suited to study the molecular mechanisms of protein fibrillation. Recently, this technique successfully allowed obtaining a complete quantitative picture of the aggregation process of both Aβ(1-40) and Aβ(1-42), including the size of the oligomers and protofibrils, the kinetics of monomer consumption, and the quantification of different early- and lateformed aggregated species [41–43].

In light of the fibrillation properties of all the PNT3 variants herein investigated, as revealed by the ensemble of studies described above, we decided to focus on the two variants with the most extreme phenotype, namely the variant with the highest fibrillation propensity, i.e., HeV PNT3\_C-term\_truncated, and the least fibrillogenic variant, i.e., PNT33A. Figure 8A,C show a three-dimensional overview of the obtained taylorgrams during the aggregation process of the two selected variants. In the case of the truncated variant, two major species were detected over time: the monomer (hydrodynamic radius, *Rh*, of ~3 nm) (Figure 8B) and large aggregates with an *Rh* ≥ 400 nm (see spikes at the beginning of the run in Figure 8A). Figure 8B shows that the *Rh* value of the soluble fraction increases with time, reaching about double its initial value at the end of the incubation, when this species is present in low proportions. Figure 8B also shows that the monomeric population slowly decreases with time. The decrease of the peak area (*Y*) of the monomeric population could be fitted using a first order exponential decay:

$$Y = A\_1 \times \exp\left(-\left(t/t\_1\right) + Y\_0\right) \tag{1}$$

where *Y*<sup>0</sup> = 6.13 ± 0.20; *A*<sup>1</sup> = 7.63 ± 0.34 and *t*<sup>1</sup> = 12.06 ± 1.32 h. This resulted in a good quality fit as judged from the R<sup>2</sup> value of 0.9394. From the fit, the kinetics of aggregation could be deduced, with a characteristic aggregation time of about *t*<sup>1</sup> *=* 12 h. From Figure 8B, the evolution in the spikes area gives an estimation of the quantity of large aggregated species entering the capillary. The proportion of these aggregated species increases with time until about 48 h, before decreasing because their size becomes too large to enter the capillary or because of precipitation in the sample vial. The presence of spikes at very short times of incubation suggests a fast aggregation process of the truncated form, while the absence of significant intermediate species suggests that the monomers add to the already present aggregates and elongate them. changes during the incubation time, with both *Rh* and peak area remaining constant in this period (Figure 8C,D). In conclusion, the data support a much higher fibrillation ability of the truncated variant, as corroborated by the presence of spikes in the elution profiles and by the fast consumption of the monomeric species as compared to PNT33A, where no evolution in size nor in area was observed. The aggregation process of the truncated variant follows a first order kinetics, without significant formation of intermediate species between the monomeric and the fibrillar species. In PNT33A the fibrillation process could not be detected, indicating that the fibrils observed by ns-TEM represent a very poorly populated species within the system.

of significant intermediate species suggests that the monomers add to the already present

By contrast, and in agreement with its dramatically reduced fibrillogenic abilities as unveiled by the other approaches herein used, the PNT33A variant does not show any

*Int. J. Mol. Sci.* **2022**, *23*, x FOR PEER REVIEW 16 of 25

aggregates and elongate them.

**Figure 8.** Kinetics of fibril formation by Taylor Dispersion Analysis (TDA). Three-dimensional overview of the obtained taylorgrams during the aggregation process of HeV PNT3-Cterm\_truncated (**A**) and HeV PNT33A (**C**) at different incubation times. The rainbow color code indicates the progress of the incubation time from dark violet to red. Analyses were performed using a protein concentration of 200 μM in 50 mM phosphate buffer, pH 7.2 at 37 °C. Peak area and hydrodynamic radius **Figure 8.** Kinetics of fibril formation by Taylor Dispersion Analysis (TDA). Three-dimensional overview of the obtained taylorgrams during the aggregation process of HeV PNT3-Cterm\_truncated (**A**) and HeV PNT33A (**C**) at different incubation times. The rainbow color code indicates the progress of the incubation time from dark violet to red. Analyses were performed using a protein concentration of 200 µM in 50 mM phosphate buffer, pH 7.2 at 37 ◦C. Peak area and hydrodynamic radius (*Rh*) evolution of the monomeric species and of the spikes area during the aggregation process of HeV PNT3-Cterm\_truncated (**B**) and HeV PNT33A (**D**).

By contrast, and in agreement with its dramatically reduced fibrillogenic abilities as unveiled by the other approaches herein used, the PNT33A variant does not show any changes during the incubation time, with both *Rh* and peak area remaining constant in this period (Figure 8C,D).

In conclusion, the data support a much higher fibrillation ability of the truncated variant, as corroborated by the presence of spikes in the elution profiles and by the fast consumption of the monomeric species as compared to PNT33A, where no evolution in size nor in area was observed. The aggregation process of the truncated variant follows a first order kinetics, without significant formation of intermediate species between the monomeric and the fibrillar species. In PNT33A the fibrillation process could not be detected, indicating that the fibrils observed by ns-TEM represent a very poorly populated species within the system.

#### **3. Materials and Methods**

#### *3.1. Generation of the Constructs*

The pDEST17OI/PNT3 and pDEST17OI/PNT33A expression plasmids, driving the expression of a hexahistidine tagged form of the protein of interest, have already been described [25]. For the construction of expression plasmids encoding the PNT3 variants bearing single alanine substitutions (PNT3A1, PNT3A2, PNT3A3), the pDEST17OI/PNT3 construct was used as template in two separate PCR amplifications using either primers attB1 and specific R\_ala*N*-PNT3 (PCR1), or primers F\_ala*N*-PNT3 and attB2 (PCR2), where *N* varies from 1 to 3 (see Supplementary Table S2). Primers were purchased from Eurofins Genomics Germany GmbH (Ebersberg, Germany). After DpnI treatment (New England Biolabs, Ipswich, MA, USA), 1 µL of PCR1 and 1 µL of PCR2 were used as overlapping megaprimers along with primers attB1 and attB2 in a third PCR. After purification, the third PCR product was inserted into the pDEST17OI bacterial expression vector using the Gateway® technology (Invitrogen, Carlsbad, CA, USA). This vector allows expression of the recombinant protein under the control of the T7 promoter. The resulting protein is preceded by a stretch of 22 vector-encoded residues (MSYYHHHHHHLESTSLYKKAGF) encompassing a hexahistidine tag. The DNA fragment encoding NiV PNT3 (i.e., residues 200–314 of the NiV P/V/W protein) was PCR-amplified using the pDEST17OI/NiV W construct as template [21] and primers NiV PNT3-AttB1 and NiV PNT3-AttB2. After DpnI treatment (New England Biolabs, Ipswich, MA, USA), the resulting amplicon was cloned into pDEST17OI as described above (Invitrogen, Carlsbad, CA, USA).

The expression construct encoding the truncated HeV PNT3 variant (PNT3\_C-term\_ truncated) was generated by PCR using pDEST17OI/PNT3 [25] as template and attB1 and Trunc\_PNT3\_B2 as primers. After DpnI treatment (New England Biolabs, Ipswich, MA, USA), the resulting amplicon was cloned in pDEST17OI. The same procedure was used to obtain the HeV PNT3 truncated variant bearing the triple alanine substitution (PNT33A\_C-term\_truncated) except that the pDEST17OI/PNT33A construct [25] was used as template.

The construct encoding the HeV PNT3 variant bearing a cysteine (PNT3\_Cys) was obtained using the pDEST17OI/PNT3 construct [25] as template in two separate PCR amplifications using either primers attB1 and specific PNT3\_C255\_R (PCR1), or primers PNT3\_C255\_F and attB2 (PCR2). After DpnI treatment, 1 µL of PCR1 and 1 µL of PCR2 were used as overlapping megaprimers along with primers attB1 and attB2 in a third PCR. After purification, the third PCR product was inserted into pDEST17 (Invitrogen, Carlsbad, CA, USA). The list and sequence of primers used to generate the above-described constructs is provided in Supplementary Table S2. Primers were purchased from Eurofins Genomics. All the constructs were verified by DNA sequencing (Eurofins Genomics Germany GmbH, Ebersberg, Germany) and found to conform to expectations.

#### *3.2. Proteins Expression and Purification*

The *E. coli* strain T7pRos was used for the expression of all the recombinant proteins upon transformation of bacterial cells with each of the bacterial expression plasmids described above. Cultures were grown over-night to saturation in LB medium containing 100 µg mL−<sup>1</sup> ampicillin and 34 µg mL−<sup>1</sup> chloramphenicol. An aliquot of the overnight culture was diluted 1/20 into 1 L of TB medium and grown at 37 ◦C with shaking at 200 rpm. When the optical density at 600 nm (OD600) reached 0.5–0.8, isopropyl β-Dthiogalactopyanoside (IPTG) was added to a final concentration of 0.5 mM, and the cells were grown at 35 ◦C overnight. The induced cells were harvested, washed and collected by centrifugation (5000× *g*, 15 min). The resulting pellets were resuspended in buffer A (50 mM Tris/HCl pH 7.5, 1 M NaCl, 20 mM imidazole) containing 6 M guanidium hydrochloride (GDN). The suspension was sonicated to disrupt the cells (using a 750 W sonicator and 3 cycles of 30 s each at 45% power output) and then centrifuged at 14,000× *g* for 30 min at 20 ◦C. The supernatant was first purified by IMAC by mixing it with 5 mL Nickel resin (GE Healthcare, Uppsala, Sweden) pre-equilibrated in buffer A. The affinity resin was washed with 20 column volumes (CV) of buffer A. Proteins were eluted with ~3 CV of buffer A supplemented with 250 mM imidazole. The fractions eluted were pooled and concentrated in the presence of 6 M GDN up to 1 mM using Centricon concentrators, and the proteins were then frozen at −20 ◦C. All PNT3 variants were subsequently subjected to SEC (HiLoad 16/600 Superdex 75 pg column, Cytiva, Marlborough, MA, USA), where the SEC column was equilibrated with buffer B (sodium phosphate 50 mM pH 7.2, 100 mM NaCl, 5 mM EDTA). The fractions from SEC, were pooled, supplemented with 6M GDN and concentrated (up to ~750 µM) and stored at −20 ◦C. In the case of the PNT3 cysteine variant, the GDN-containing sample was also supplemented with 10 mM DTT. Prior to each subsequent analysis, the samples were loaded onto a Sephadex G-25 medium column (Cytiva, Marlborough, MA, USA) to exchange the buffer. The proteins were eluted from G-25 columns using sodium phosphate 50 mM buffer at a pH of 7.2 unless differently specified. IMAC and SEC were performed at room temperature (RT).

Protein concentrations were estimated using the theoretical absorption coefficients at 280 nm as obtained using the program ProtParam from the EXPASY server (http://web. expasy.org/protparam/, accessed on 20 January 2021).

The purity of the final purified products was assessed by SDS-PAGE (Figure 3B). The identity of all the purified PNT3 variants generated in this work was confirmed by mass spectrometry analysis of tryptic fragments obtained after digestion of the purified protein bands excised from SDS-polyacrylamide gels (Supplementary Figure S3). The excised bands were analyzed by the mass spectrometry facility of Marseille Proteomics in the same way as previously done for the PNT3 *wt* variant [25]. Briefly, gel pieces were washed and destained using 100 mM NH4HCO3/acetonitrile (50/50). Destained gel pieces were shrunk with acetonitrile and were re-swollen in the presence of 100 mM ammonium bicarbonate in 50% acetonitrile and dried at room temperature. Protein bands were then rehydrated and cysteines were reduced using 10 mM DTT in 100 mM ammonium bicarbonate pH 8.0 for 45 min at 56 ◦C before alkylation in the presence of 55 mM iodoacetamide in 100 mM ammonium bicarbonate pH 8.0 for 30 min at room temperature in the dark, then washed twice with 25 mM ammonium bicarbonate pH 8.0 and digested with high-sequencinggrade trypsin (Promega, Madison, WI, USA). Mass spectrometry analysis were carried out by LC-MSMS using a Q Exactive Plus Hybrid Quadrupole-Orbitrap online with a nanoLC Ultimate 3000 chromatography system (Thermo Fisher Scientific™, San Jose, CA, USA). For each biological sample, 5 microliters corresponding to 25% of digested sample were injected in duplicate on the system. After pre-concentration and washing of the sample on an Acclaim PepMap 100 column (C18, 2 cm × 100 µm i.d. 100 A pore size, 5 µm particle size), peptides were separated on a LC EASY-Spray column (C18, 50 cm × 75 µm i.d., 100 A, 2 µm, 100A particle size) at a flow rate of 300 nL/min with a two steps linear gradient (2–22% acetonitrile/H20; 0.1% formic acid for 100 min and 22–32% acetonitrile/H20; 0.1% formic acid for 20 min). For peptides ionization in the EASYSpray source, spray voltage

was set at 1.9 kV and the capillary temperature at 250 ◦C. All samples were measured in a data-dependent acquisition mode. Each run was preceded by a blank MS run in order to monitor system background. The peptide masses were measured in a survey full scan (scan range 375–1500 m/z, with 70 K FWHM resolution at m/z = 400, target AGC value of 3.00 <sup>×</sup> <sup>10</sup><sup>6</sup> and maximum injection time of 100 ms). Following the high-resolution full scan in the Orbitrap, the 10 most intense data-dependent precursor ions were successively fragmented in HCD cell and measured in Orbitrap (normalized collision energy of 25%, activation time of 10 ms, target AGC value of 1.00 <sup>×</sup> <sup>10</sup><sup>3</sup> , intensity threshold 1.00 <sup>×</sup> <sup>10</sup><sup>4</sup> maximum injection time 100 ms, isolation window 2 m/z, 17.5 K FWHM resolution, scan range 200 to 2000 m/z). Dynamic exclusion was implemented with a repeat count of 1 and exclusion duration of 20 s.

Raw files generated from mass spectrometry analysis were processed with Proteome Discoverer 1.4 .1.14 (Thermo Fisher Scientific, San Jose, CA, USA) to search against a home-made database containing 20,150 human sequences, 4306 *E.coli* sequences implemented with the expected sequences (swissprot – human – reviewed – 170315 \_ 20150 \_UP\_coli\_171120\_4306\_ Patrick220922 \_ ID \_ Bandes.fast).

Database search with SequestHT were done using the following settings: a maximum of two trypsin miss cleavage allowed, methionine oxidation and N terminal protein acetylation as variable modifications, and cysteine carbamidomethylation as fixed modification. A peptide mass tolerance of 6 ppm and a fragment mass tolerance of 0.8 Da were allowed for search analysis. Only peptides with high Sequest scores were selected for protein identification. False discovery rate was set to 1% for protein identification.

#### *3.3. PEG Precipitation Assay (Relative Solubility)*

The relative solubility of each variant was evaluated at different PEG concentrations using an adaptation of the protocol recently described by Oeller et al. [32]. Briefly, PEG solutions from 0 to 30% were prepared from 50% PEG<sup>6000</sup> (Steinheim, Germany) stock solution. Then, an aliquot of corresponding protein from a stock (at 260 µM) was mixed with each PEG solutions to obtain a final concentration of 66 µM protein in a 100 µL final reaction volume. The assay was performed in 96-well plates sealed with aluminum plate sealers to prevent possible evaporation (Thermo Fisher Scientific, USA). Plates were incubated at 4 ◦C for 24 h and centrifugated at maximum velocity (4600× *g*) for 2 h. Immediately after, 2 µL of the supernatant were pipetted to quantify the soluble protein concentration using a ND-1000 Nanodrop Spectrophotometer and theoretical absorption coefficients at 280 nm as obtained using the program ProtParam from the EXPASY server (http://web.expasy.org/protparam/, accessed on 20 January 2021). Each condition was made in triplicate. The soluble fractions obtained were normalized and fitted to a sigmoid function to obtain the PEG1/2 value that reports on relative solubility (PRISM software). The error on the PEG<sup>1</sup> 2 and the quality of the fit were estimated by a 95% confidence interval analysis (PRISM software version 9.2, CA, USA).

#### *3.4. Far-UV Circular Dichroism*

CD spectra were measured using a Jasco 810 dichrograph (Jasco France, Lisses, France), flushed with N<sup>2</sup> and equipped with a Peltier thermoregulation system. Proteins were loaded into a 1 mm quartz cuvette at 0.06 mg/mL (in 10 mM phosphate buffer at pH 7.2) and spectra were recorded at 37 ◦C. The scanning speed was 20 nm min−<sup>1</sup> , with data pitch of 0.2 nm. Each spectrum is the average of ten acquisitions. The spectrum of buffer was subtracted from the protein spectrum. Spectra were smoothed using the "means-movement" smoothing procedure implemented in the Spectra Manager package. As already previously documented, a decrease in the signal spectrum was observed with increasing incubation time and ascribed to fibril formation [25]. Because the different variants have different fibrillation propensities, differences in spectra intensity might reflect differences in the fraction of fibrillar species (which are not detected by CD) rather than

*bona fide* spectral differences. We therefore normalized spectra using the maximum negative value of intensity as a normalization factor.

Mean molar ellipticity values per residue (MRE) were calculated as

$$\mathbf{f}[\theta] = \mathbf{3} \mathbf{3} \mathbf{0} \mathbf{0} \times \boldsymbol{m} \times \Delta A / (l \times \boldsymbol{c} \times \boldsymbol{n}) \tag{2}$$

where *l* is the path length in cm, *n* is the number of residues, *m* is the molecular mass in Daltons and *c* is the concentration of the protein in mg mL−<sup>1</sup> .

#### *3.5. Estimation of the Hydrodynamic Radius by SEC*

The hydrodynamic radii (Stokes radii, RS) of the proteins were estimated by analytical SEC using a HiLoad 16/600 Superdex 75 pg column (Cytiva, Marlborough, MA, USA). Buffer B was used as elution buffer. Typically, 250 µL of purified protein at 11 mg mL−<sup>1</sup> were injected.

The Stokes radii of proteins eluted from the SEC column were deduced from a calibration curve obtained using globular proteins of known R<sup>S</sup> (Conalbumin: 36.4 Å, Carbonic Anhydrase: 23 Å, RNAse A: 16.4 Å Aprotinin: 13.5 Å)

The R<sup>S</sup> (in Å) of a natively folded (*RsNF*), fully unfolded state in urea (*R<sup>S</sup> <sup>U</sup>*) and natively unfolded premolten globule (PMG) (*R<sup>S</sup> PMG*) protein with a molecular mass (*MM*) (in Daltons) were calculated according to [44]:

$$\log\left(R\_S^{NF}\right) = 0.357 \times \left(\log MM\right) - 0.204\tag{3}$$

$$
\log \left( R\_S^{\,\,\,\,} \right) = 0.521 \times \left( \log \,\, MM \right) - 0.649 \tag{4}
$$

$$
\log \left( R\_S^{PMG} \right) = 0.403 \times \left( \log \text{MM} \right) - 0.239 \tag{5}
$$

The *R<sup>S</sup>* (in Å) of an IDP with *N* residues was also calculated according to [45] using the simple power-law model:

$$R\_S \, ^{IDP} = R\_0 \mathcal{N}^{\nu} \tag{6}$$

where *R*<sup>0</sup> = 2.49 and ν = 0.509. The compaction index (*CI*) is expressed according to [46]:

$$\text{C}I = (R\_S \, ^{II} - R\_S \, ^{OBS}) / (R\_S \, ^{II} - R\_S \, ^{NF}) \tag{7}$$

This parameter, which allows comparison between proteins of different lengths, in principle varies between 0 and 1, with 0 indicating minimal compaction and 1 maximal compaction. In the case of the cysteine variant, as the SEC analysis was performed without DTT, a peak corresponding to a dimeric species was observed. However, since the latter was not well resolved, its corresponding R<sup>S</sup> was not calculated.

#### *3.6. Small-Angle X-ray Scattering (SAXS)*

In order to ensure maximal monodispersity of the sample, SAXS studies were coupled to SEC. SEC-SAXS data were collected at SOLEIL (Gif-sur-Yvette, France), as described in Table 3. In both cases, the calibration was performed with water. Sample from each PNT3 variant at 5 mg mL−<sup>1</sup> in buffer B containing 6M GDN was injected onto an AdvanceBio SEC 2.7 µm (Agilent) SEC column. Elution was carried out in buffer C (50 mM sodium phosphate buffer at pH 7.2). Data reduction and frames subtraction were done with the beamline software FOXTROT (available upon request from the SOLEIL staff). Gaussian decomposition was performed using the UltraScan solution modeler (US-SOMO) HPLC-SAXS module (https://somo.aucsolutions.com/, accessed on 6 September 2022) [47] or Chromixs (manual frames selection) [48] and the final deconvoluted scattering curves were submitted to the SHANUM program [49] to remove noisy, non-informative data at high angles.


**Table 3.** SEC-SAXS data acquisition parameters.

The data were analyzed using the ATSAS program package [49]. The radius of gyration (*Rg*) and I(0) were estimated at low angles (*q*.*R<sup>g</sup>* < 1.3) according to the Guinier approximation [50,51]:

$$Ln[I(q)] = Ln[I\_0] - (q^2 R\_{\mathcal{S}}^{-2})/3\tag{8}$$

The pairwise distance distribution functions P(r), from which the *Dmax* and the *R<sup>g</sup>* were estimated, were calculated with the program GNOM [52] and manually adjusted until a good CorMap *p*-value (α > 0.01) was obtained [52].

The theoretical *R<sup>g</sup>* value (in Å) expected for various conformational states was calculated using Flory's equation:

$$R\_{\mathcal{S}} = R\_0 \mathbf{N}^{\nu} \tag{9}$$

where *N* is the number of amino acid residues, *R*<sup>0</sup> a constant and ν a scaling factor. For IDPs, *R*<sup>0</sup> is 2.54 ± 0.01 and *ν* is 0.522 ± 0.01 [53], for chemically denatured (U) proteins *R*<sup>0</sup> is 1.927 ± 0.27 and *ν* is 0.598 ± 0.028 [53], and for natively folded (NF) proteins *R*<sup>0</sup> = √ (3/5) × 4.75 and *ν* = 0.29 [54].

As in the case of the *RS*, the *CI* allows comparing the degree of compaction of a given IDP, through comparison of the observed *R<sup>g</sup>* to the reference values expected for a fully unfolded and a folded conformation of identical mass. The *CI* referred to the *R<sup>g</sup>* can be calculated as follows [46]:

$$\text{CI} = (\text{R}\_{\text{\textS}}{}^{\text{II}} - \text{R}\_{\text{\textS}}{}^{\text{OBS}}) / (\text{R}\_{\text{\textS}}{}^{\text{II}} - \text{R}\_{\text{\textS}}{}^{\text{NF}}) \tag{10}$$

where *R<sup>g</sup> OBS* is the experimental value for a given protein, and *R<sup>g</sup> <sup>U</sup>* and *R<sup>g</sup> NF* are the reference values calculated for a fully unfolded (U) and natively folded (NF) form, as described above. Akin to the *RS*-based *CI*, this index increases with increasing compaction.

The overall conformation and the flexibility of the proteins was assessed with the dimensionless Kratky plot ((*qRg*) 2 I(*q*)/I<sup>0</sup> *vs qRg*) and the Krakty-Debye plot (*q* 2 I(*q*) *vs q*<sup>2</sup> ).

SEC-SAXS data have been deposited in the Small Angle Scattering Biological Data Bank (SASBDB) [55] under codes SASDQB7, SASDQC7, SASDQD7, SASDQE7 and SAS-DQF7 for the set of data of PNT3 *wt*, PNT33A, NiV PNT3, PNT3\_C-terminal truncated and PNT33A\_C-terminal truncated, respectively.

#### *3.7. Congo Red Binding Assays*

Quantitative measurement of Congo Red (Sigma Aldrich, Saint Louis, MO, USA) binding (CR shift assay) was carried out by using protein samples containing each PNT3 variant at 20 µM (in Buffer C) and 5 µM of CR in a final volume of 100 µL. The samples were then incubated at 37 ◦C for 4 or 7 days. The adsorption spectrum of the CR-containing samples was recorded using a PHERAstar FSX Microplate Reader (BMG LABTECH, Champignysur-Marne, France) in the 350–600 nm wavelength range. A solution of 5 µM CR in Buffer C without the protein was used as a control to normalize the analysis. Experiments were carried out in triplicate. Statistical analysis was performed using a Two-way ANOVA test implemented in the PRISM software.

#### *3.8. Negative-Staining Transmission Electron Microscopy (ns-TEM)*

All the variants, at a concentration of 200 µM, were prepared and analyzed at different times to monitor their evolution (0, 24, 96 h). Incubation was carried out at 37 ◦C in Buffer C. Prior to each measurement, the samples were diluted to reach a final concentration of 40 µM. EM grids (carbon coated copper grids, 300 mesh, Agar Scientific, UK) were exposed to plasma glow discharge for 20 s using GloQube (Quorum, UK) (Current 15 mA) in order to increase protein adhesion. Drops of 3.5 µL of the diluted protein solutions were deposited onto glow-discharged grids. After 1 min incubation with the sample, the grids were washed three times with 50 µL of buffer C, once in 35 µL 1% (*w*/*v*) Uranyl acetate solution (Laurylab, Brindas, France) and then stained for 1 min in the latter solution. Excess of uranyl was blotted and grids were left to dry for 1 h at RT. Images were collected on Tecnai 120 Spirit TEM microscope (FEI company, ThermoFisher, Illkirch-Graffenstaden France) operated at 120 kV using a Veleta 2K × 2K CCD camera (Olympus).

#### *3.9. Kinetic Protein Aggregation Study by Taylor Dispersion Analysis (TDA)*

TDA was performed as already described [41] using an Agilent 7100 (Waldbronn, Germany) capillary electrophoresis system with bare fused silica capillaries (Polymicro Technologies, USA) having 60 cm × 50 µm i.d. dimensions and a detection window at 51.5 cm. New capillaries were conditioned with the following flushes: 1 M NaOH for 30 min and ultrapure water for 30 min. Between each analysis, the capillaries were rinsed with Buffer C (2 min). Samples were injected hydrodynamically on the inlet end of the capillary (30 mbar, 6 s, injected volume is about 6.1 nL corresponding to less than 1% of the capillary volume to the detection point). Experiments were performed using a mobilization pressure of 100 mbar. The temperature of the capillary cartridge was set at 37 ◦C. The vial carrousel was thermostated using an external circulating water bath from Instrumat (Moirans, France). The solutes were monitored by UV absorbance at 198 nm. The mobile phase was Buffer C. (viscosity at 37 ◦C is 7.54 <sup>×</sup> <sup>10</sup>−<sup>4</sup> Pa.s.). Samples obtained after the desalting column were diluted to reach 700 µL at 200 µM solution and were immediately transferred to a vial and incubated at 37 ◦C in the capillary electrophoresis instrument's carrousel. The aggregation was conducted by injecting the sample every 1 h. The total number of TDA runs for each sample was about 120. The taylorgrams were recorded with Agilent Chemstation software and then exported to Microsoft Excel for subsequent data processing. Data were fitted to a first order exponential decay according to Equation (1). This resulted in a good quality fit as judged from the Reduced Chi-Sqr: 0.43723; R-Square (COD) = 0.93943, and Adj. R-Square = 0.93576. When necessary, the elution peaks were fitted with the sum of *n* Gaussian functions (in this work: *n* ≤ 3) as already described [41,42].

#### **4. Conclusions**

This study constitutes a comprehensive analysis of the molecular basis of the fibrillation process of a small region (PNT3) within the N-terminal intrinsically disordered domain shared by the HeV P/V/W proteins. Biochemical and biophysical characterization of a set of HeV PNT3 variants bearing alanine substitutions in the amyloidogenic EYYY motif, along with the characterization of the corresponding PNT3 region from the cognate NiV, revealed that each of the three tyrosines in the motif are required for the elongation step of the fibrillation process. Remarkably, the present study also unveiled a role for the C-terminal domain of PNT3 in self-inhibition of fibrillation, possibly reminiscent of the α-synuclein fibrillation model, and of potential biological significance. Noteworthy, in light of the observation that amyloid-like fibrils form not only in vitro but also the cellular context, it is tempting to hypothesize that the amyloidogenicity of V/W proteins, which

both encompass the PNT3 region, could be correlated with the pathogenic (and even encephalitogenic) properties of Henipaviruses. Therefore, the PNT3 variants that we have herein generated constitute valuable tools to further explore the functional impact of V/W fibrillation in transfected and infected cells. The present results therefore set the stage for further investigations aimed at illuminating the mechanisms underlying the disease as a preliminary step towards the rational design of antivirals.

**Supplementary Materials:** The following supporting information can be downloaded at: https: //www.mdpi.com/article/10.3390/ijms24010399/s1.

**Author Contributions:** S.L. conceived, designed and supervised the study and acquired funding; C.B. generated all the bacterial expression constructs; H.B. purified all the proteins used in the work. F.G. recorded and analyzed the SAXS data, J.C. and H.C. performed the TDA analysis; D.P., G.P. and J.F.N. performed the TEM analyses; J.F.N. carried out all the other experiments. J.F.N., F.G., G.P., A.V.K. and S.L. analyzed and interpreted the data; J.F.N. generated the first draft of the manuscript. All the authors wrote and revised the manuscript. All authors have read and agreed to the published version of the manuscript.

**Funding:** This work was carried out with the financial support of the Agence Nationale de la Recherche (ANR), specific project Heniphase (ANR-21-CE11- 0012-01). It was also partly supported by the French Infrastructure for Integrated Structural Biology (FRISBI) (ANR-10-INSB- 0005) and by the CNRS. F.G. is supported by a post-doctoral fellowship from the FRM (Fondation pour la Recherche Médicale). G.P. is supported by a joint doctoral fellowship from the AID (Agence Innovation Défense) and Aix-Marseille University. J.F.N. is supported by a postdoctoral fellowship from the Infectiopôle Sud.

**Institutional Review Board Statement:** Not applicable.

**Informed Consent Statement:** Not applicable.

**Data Availability Statement:** The data present in the current study are available from the corresponding author on reasonable request.

**Acknowledgments:** We thank Aurélien Thureau (SOLEIL) and Anton Popov (ESRF) for their help in recording SEC-SAXS data. We thank both the ESRF and the SOLEIL synchrotrons for beamtime allocation. We are also grateful to Gerlind Sulzenbacher (AFMB lab) for efficiently managing the AFMB BAG. We thank all the AFMB technical and support staff (Denis Patrat, Patricia Clamecy, Béatrice Rolland, Paul Zamboni, Chantal Falaschi and Fabienne Amalfitano). We heartily thank Patrick Fourquet for mass spectrometry analyses done using the mass spectrometry facility of Marseille Proteomics (marseille-proteomique.univ-amu.fr), supported by IBISA (Infrastructures Biologie Santé et Agronomie), Plateforme Technologique Aix-Marseille, the Cancéropôle PACA, Région Sud-Alpes-Côte d'Azur, the Institut Paoli-Calmettes, the Centre de Recherche en Cancérologie de Marseille (CRCM), Fonds Européen de Développement Régional and Plan Cancer.

**Conflicts of Interest:** The authors declare no conflict of interest. The funders had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript, or in the decision to publish the results.

#### **References**


**Disclaimer/Publisher's Note:** The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

## *Article* **Cytoophidia Maintain the Integrity of** *Drosophila* **Follicle Epithelium**

**Qiao-Qi Wang <sup>1</sup> , Dong-Dong You <sup>1</sup> and Ji-Long Liu 1,2,\***


**Abstract:** CTP synthase (CTPS) forms a filamentous structure termed the cytoophidium in all three domains of life. The female reproductive system of *Drosophila* is an excellent model for studying the physiological function of cytoophidia. Here, we use *CTPSH355A*, a point mutation that destroys the cytoophidium-forming ability of CTPS, to explore the in vivo function of cytoophidia. In *CTPSH355A* egg chambers, we observe the ingression and increased heterogeneity of follicle cells. In addition, we find that the cytoophidium-forming ability of CTPS, rather than the protein level, is the cause of the defects observed in *CTPSH355A* mutants. To sum up, our data indicate that cytoophidia play an important role in maintaining the integrity of follicle epithelium.

**Keywords:** CTP synthase; cytoophidium; *Drosophila*; epithelium; follicle cell; ingression

#### **1. Introduction**

CTP synthase (CTPS) is a glutamate aminotransferase that catalyzes the transfer of amide nitrogen from glutamine to the C-4 position of UTP. CTP, the product of CTPS, is an important nucleotide and is a component of the synthesis of RNA, DNA, and sialoglycoprotein. It also acts as an energy coupler for some metabolic reactions, such as the synthesis of glycerophospholipids and glycosylated proteins [1,2].

In 2010, CTPS was found to form filamentous structures termed cytoophidia in *Drosophila* [3]. Subsequently, CTPS has been found to form filamentous structures in bacteria [4] and *S. cerevisiae* [5]. In the following years, the existence of cytoophidia was confirmed in human cells [6], *S. pombe* [7], *Arabidopsis thaliana* [8], and archaea [9], which indicates that cytoophidia are highly conserved in evolution.

Compartmentation is the basis for the function of organelles [10]. The classical cellular compartmentation in eukaryotic cells is achieved through membrane-bound organelles, such as the endoplasmic reticulum, mitochondrion, and Golgi apparatus [11]. Compartmentation establishes a physical boundary for the biological processes within cells, enabling cells to carry out different metabolic activities at the same time, generate specific microenvironments, regulate biological processes in time and space, and determine the specific location where biological processes should occur. The formation of cytoophidia realizes the regionalization of CTPS, and its location in cells may therefore have corresponding physiological significance.

Cells in the *Drosophila* ovary exhibit vigorous anabolic activity because they need nutrients for development. Cytoophidia are observed from region 2 of the germarium to stage 10A of oogenesis. Based on the widespread presence of cytoophidia in germline cells and follicle epithelial cells of *Drosophila* ovaries [12], and the characteristics of cytoophidium observed in germline cells and follicle epithelial cells at most stages of oogenesis, *Drosophila* ovarian follicle cells have become a classic model for studying cytoophidia.

Epidermal tissues form the boundaries of organs, where they perform a range of functions, including secretion, absorption, and protection. These tissues are usually composed

**Citation:** Wang, Q.-Q.; You, D.-D.; Liu, J.-L. Cytoophidia Maintain the Integrity of *Drosophila* Follicle Epithelium. *Int. J. Mol. Sci.* **2022**, *23*, 15282. https://doi.org/10.3390/ ijms232315282

Academic Editors: Vladimir N. Uversky and Hans-Arno Müller

Received: 15 October 2022 Accepted: 25 November 2022 Published: 4 December 2022

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2022 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

of discrete cells, forming a single-cell thick sheet. Follicle epithelium is a simple epithelium. In the process of division, the cells of simple epithelium have a specific orientation of the spindle, so that both daughter cells are located in the epithelial plane. This is considered to be very important for maintaining the integrity of follicle epithelium and preventing hyperplasia [13–16].

An egg chamber consists of hundreds of follicle cells, and each follicle cell has multiple membrane domains including apical, basal, and lateral. The end adjacent to germline cells is defined as the apical side, and the end far away from germline cells is defined as the basal side [17]. Before mitosis, follicle cells will move toward the apical direction, which may be caused by the extrusion of neighbor cells. This movement results in the displacement of some cells from the tissue layer. Usually, the displaced cells need to be reintegrated to support tissue growth and maintain tissue architecture [18–20].

In a previous study, we find that cytoophidia are specifically distributed on the basolateral side of follicle cells, and this specific distribution is related to the polarity regulator of the cell membrane [21]. Therefore, we need to understand the function of cytoophidium which is specifically distributed in follicle cells.

In this study, we describe the effects of cytoophidium disassembly on follicle epithelium integrity. We are also concerned about whether these effects are directly related to the assembly of CTPS into cytoophidia, rather than to the level of CTPS protein. Our results indicate that cytoophidia play an important role in maintaining the integrity of follicle epithelium. In addition, we eliminate the influence of tissue-tissue interaction and find that cytoophidia can directly affect the integrity of follicle epithelium.

#### **2. Results**

#### *2.1. CTPS Forms Cytoophidia in Drosophila Follicle Cells*

Cells in *Drosophila* ovaries exhibit vigorous anabolic activity because they need nutrients for development. During *Drosophila* oogenesis, the follicle epithelium is a sheet of monolayer cells that encase germline cells. CTPS, as the synthase of CTP, plays an important role in the regulation of tissue growth and development. Cytoophidia exist in several different types of cells in the *Drosophila* ovary from region 2 of the germarium to stage 10A of oogenesis, including epithelial follicle cells (Figure 1A–C) and germline cells (Figure 1D–F).

## *2.2. Follicle Cells Undergo Ingression in CTPSH355A Egg Chambers*

The amino acid histidine at the 355th position, or His355, lies at the tetramer-tetramer interface of CTPS [22]. If the H355 site is mutated, the cytoophidium cannot be formed. Previous studies showed that the H355 site is essential for its polymerization, but not enzymatic function [23,24]. Our laboratory has solved the structure of *Drosophila melanogaster* CTPS (dmCTPS) and found that the H355 site lies at the tetramer–tetramer interface and does not affect the catalytic site [25]. Therefore, we constructed an H355A point-mutated knock-in *Drosophila* strain to investigate whether the disassembly of cytoophidia would affect follicle cells. Former studies found that the H355A served as a dominant negative point mutation [26] (Figure S1).

In order to find out whether the abnormality is caused by the inability of CTPS to aggregate due to H355A point mutation or the addition of mCherry tag, our laboratory constructed another *Drosophila* strain with mCherry added to the C-terminus of CTPS based on *w <sup>1118</sup>*. To determine whether the feature of cytoophidium localization was in fact introduced by protein fusion between CTPS and mCherry tag, we performed immunofluorescence microscopy and directly detected the CTPS protein of the *w <sup>1118</sup>* fly and found no difference [27]. It is proven by the observation that the knock-in mCherry tag does not affect the polymerization of CTPS protein. The morphology of the *CTPS-mCherryKI Drosophila(CTPS-mCh)* ovaries is consistent with that of the *w <sup>1118</sup>*, which implies that the *CTPS-mCherryKI Drosophila* can also be used as control in our experiment (Figure 2A–F). **2. Results** 

*Int. J. Mol. Sci.* **2022**, *23*, x FOR PEER REVIEW 2 of 17

preventing hyperplasia [13–16].

Epidermal tissues form the boundaries of organs, where they perform a range of functions, including secretion, absorption, and protection. These tissues are usually composed of discrete cells, forming a single-cell thick sheet. Follicle epithelium is a simple epithelium. In the process of division, the cells of simple epithelium have a specific orientation of the spindle, so that both daughter cells are located in the epithelial plane. This is considered to be very important for maintaining the integrity of follicle epithelium and

An egg chamber consists of hundreds of follicle cells, and each follicle cell has multiple membrane domains including apical, basal, and lateral. The end adjacent to germline cells is defined as the apical side, and the end far away from germline cells is defined as the basal side [17]. Before mitosis, follicle cells will move toward the apical direction, which may be caused by the extrusion of neighbor cells. This movement results in the displacement of some cells from the tissue layer. Usually, the displaced cells need to be

In a previous study, we find that cytoophidia are specifically distributed on the basolateral side of follicle cells, and this specific distribution is related to the polarity regulator of the cell membrane [21]. Therefore, we need to understand the function of cy-

In this study, we describe the effects of cytoophidium disassembly on follicle epithelium integrity. We are also concerned about whether these effects are directly related to the assembly of CTPS into cytoophidia, rather than to the level of CTPS protein. Our results indicate that cytoophidia play an important role in maintaining the integrity of follicle epithelium. In addition, we eliminate the influence of tissue-tissue interaction and find

Cells in *Drosophila* ovaries exhibit vigorous anabolic activity because they need nutrients for development. During *Drosophila* oogenesis, the follicle epithelium is a sheet of monolayer cells that encase germline cells. CTPS, as the synthase of CTP, plays an important role in the regulation of tissue growth and development. Cytoophidia exist in sev-

reintegrated to support tissue growth and maintain tissue architecture [18–20].

toophidium which is specifically distributed in follicle cells.

*2.1. CTPS Forms Cytoophidia in Drosophila Follicle Cells* 

that cytoophidia can directly affect the integrity of follicle epithelium.

Besides, our laboratory has used the *CTPS-mCherryKI Drosophila* as control in previous studies [28,29]. eral different types of cells in the *Drosophila* ovary from region 2 of the germarium to stage 10A of oogenesis, including epithelial follicle cells (Figure 1A–C) and germline cells (Figure 1D–F).

**Figure 1. Follicle cells maintain monolayer during** *Drosophila* **oogenesis.** (**A**–**C**) Surface view of a wild-type ovariole containing different stages of egg chambers. The ovariole is subjected to immunofluorescence analysis with antibodies against CTPS (green) and Hts (red, labelling cell membranes). DNA is labelled by Hoechst 33342 (blue). (**A**) Cytoophidia are distributed almost at different stages of each follicle cell. (**B**) The boundaies of follicle epithelia are displayed by a single projection of Hts staining. (**C**) CTPS staining shows the distribution of cytoophidia on the surface of egg chambers. (**D**–**F**) Side view of the same ovariole in (**A**–**C**). (**D**) Monolayer follicle cells envelop germline cells. (**E**) A single projection of Hts staining shows the monolayer structure of follicle epithelia. (**F**) CTPS staining shows the distribution of cytoophidia in follicle cells and germline cells. Scale bars, 50 µm.

When constructing the point-mutated *Drosophila* strain, we added a mCherry tag at the C-terminal of CTPS. Through confocal microscopy, we observed diffuse mCherry signal in *Drosophila* follicle cells, which confirmed that the CTPS could not form the cytoophidium after mutation at the H355 site (Figure 3A,F). In the egg chamber of wild-type flies, follicle cells are monolayer epidermal cells. We observed their morphological characteristics by immunofluorescence staining. The cell membrane was labeled with an antibody against Armadillo. We found that in the egg chamber of *CTPSH355A/H355A-mCherry* knock-in homozygous fly (hereinafter referred to as *CTPSH355A* strain), some follicle cells originally arranged in a monolayer migrated inward (ie. ingression), thus disrupting the monolayer arrangement. The ingression of follicle cells occurs not only in the early stages of oogenesis, such as stage 5 (Figure 3A–E), but also in the middle stages of oogenesis, such as stage 8 (Figure 3F–J).

**Figure 2.** *CTPS-mCh* **ovaries have same morphology as wild-type ovaries.** (**A**–**C**) Surface view of a stage 8 CTPS-mCh egg chamber. The CTPS signal (green) shown is obtained using mCherrytagged CTPS. Hts (red) staining marks cell membranes and Hoechst (blue) for DNA. (**A**) mCherry moiety doesn't affect CTPS assembly. (**B**) Merged panel of Hts and CTPS to display the cytoophidia location. (**C**) Single panel of the nucleus. (**D**) Lateral view of a stage 8 *CTPS-mCh* egg chamber. (**E**) Merged panel of Hts and CTPS to show the monolayer structure of follicle epithelia as well as cytoophidia distribution. (**F**) Single panel of nucleus to stress the single-layer follicle epithelia. When constructing the point-mutated *Drosophila* strain, we added a mCherry tag at **Figure 2.** *CTPS-mCh* **ovaries have same morphology as wild-type ovaries.** (**A**–**C**) Surface view of a stage 8 CTPS-mCh egg chamber. The CTPS signal (green) shown is obtained using mCherry-tagged CTPS. Hts (red) staining marks cell membranes and Hoechst (blue) for DNA. (**A**) mCherry moiety doesn't affect CTPS assembly. (**B**) Merged panel of Hts and CTPS to display the cytoophidia location. (**C**) Single panel of the nucleus. (**D**) Lateral view of a stage 8 *CTPS-mCh* egg chamber. (**E**) Merged panel of Hts and CTPS to show the monolayer structure of follicle epithelia as well as cytoophidia distribution. (**F**) Single panel of nucleus to stress the single-layer follicle epithelia. *Int. J. Mol. Sci.* **2022**, *23*, x FOR PEER REVIEW 5 of 17

**Figure 3. Follicle cell ingression in** *CTPSH355A* **mutant in early and middle stages of oogenesis.** (**A**) Cross section of a *CTPSH355A* stage 6 egg chamber. The egg chamber is labeled with CTPS (green), Armadillo (red) for apical complex in follicle cells and Hoechst 33342 (blue) for DNA. The white rectangle emphasizes the ingression of a follicle cell. (**B**–**E**) Close-up images of the ingressive follicle cell, indicated by the yellow arrow. No cytoophidium is formed due to the H355A mutation in CTPS. (**F**) Cross section of a *CTPSH355A* egg chamber at stage 9. The white rectangle emphasizes the ingression of a follicle cell. (**G)** Close-up images of the ingressive follicle cell, indicated by the yellow arrow. Scale bars, 20 μm. Our study mainly focused on stage 8 egg chambers. We demonstrated the ingression of **Figure 3. Follicle cell ingression in** *CTPSH355A* **mutant in early and middle stages of oogenesis.** (**A**) Cross section of a *CTPSH355A* stage 6 egg chamber. The egg chamber is labeled with CTPS (green), Armadillo (red) for apical complex in follicle cells and Hoechst 33342 (blue) for DNA. The white rectangle emphasizes the ingression of a follicle cell. (**B**–**E**) Close-up images of the ingressive follicle cell, indicated by the yellow arrow. No cytoophidium is formed due to the H355A mutation in CTPS. (**F**) Cross section of a *CTPSH355A* egg chamber at stage 9. The white rectangle emphasizes the ingression of a follicle cell. (**G**–**J**) Close-up images of the ingressive follicle cell, indicated by the yellow arrow. Scale bars, 20 µm.

follicle cells in stage 8 egg chambers through three-dimensional reconstruction (Figure 4A–C). Combined with the morphological changes of follicle cells observed on the surface of the egg chambers, we speculated that the integrity of follicle epithelia would be disturbed

ing 20 stage 8 egg chambers of each genotype (Figure 4D). Our results indicate that the widely and specifically distributed cytoophidia play a role in maintaining the integrity of

follicle epithelia.

Our study mainly focused on stage 8 egg chambers. We demonstrated the ingression of follicle cells in stage 8 egg chambers through three-dimensional reconstruction (Figure 4A–C). Combined with the morphological changes of follicle cells observed on the surface of the egg chambers, we speculated that the integrity of follicle epithelia would be disturbed when CTPS could not assemble into cytoophidia. Through statistical analysis, we found that approximately 20% of egg chambers at stage 8 have follicle cells ingression by counting 20 stage 8 egg chambers of each genotype (Figure 4D). Our results indicate that the widely and specifically distributed cytoophidia play a role in maintaining the integrity of follicle epithelia. *Int. J. Mol. Sci.* **2022**, *23*, x FOR PEER REVIEW 6 of 17

**Figure 4. Three dimensional view of follicle cell ingression.** (**A**) Cross section of a three-dimensional *CTPSH355A* egg chamber. A three-dimensional view of DNA stacked in layers. The interval between each layer is 0.5 μm, and a total of 12 layers are superimposed. The color from red to blue indicates the depth of DNA. (**B**) Side view of an ingressive nucleus. (**C**) On the xz plane, with the yellow dotted line marking the ingressive cell. (**D**) Quantification of the ingression frequency, 20 stage 8 egg chambers were counted per genotype. **Figure 4. Three dimensional view of follicle cell ingression.** (**A**) Cross section of a threedimensional *CTPSH355A* egg chamber. A three-dimensional view of DNA stacked in layers. The interval between each layer is 0.5 µm, and a total of 12 layers are superimposed. The color from red to blue indicates the depth of DNA. (**B**) Side view of an ingressive nucleus. (**C**) On the xz plane, with the yellow dotted line marking the ingressive cell. (**D**) Quantification of the ingression frequency, 20 stage 8 egg chambers were counted per genotype.

#### 2.3. Ingressive Follicle Cells Display Abnormal DCAD2 Pattern *2.3. Ingressive Follicle Cells Display Abnormal DCAD2 Pattern*

In a previous study, we found that cytoophidia are specifically located on the lateral and basal sides of follicle cells [21]. The polarity regulators of follicle cells and adherens junctions have certain effects on the maintenance of cytoophidia. To explore whether the cell membrane components of ingressive follicle cells would be affected when cytoophidium fails to form, we labeled the basolateral regulator Dlg of follicle cells, adherens junctions DE-Cadherin DCAD2, and cell membrane protein Hts. After immunostaining, in the follicle epithelium labeled with Hts and Dlg, there was no significant difference between the cell membrane of ingressive follicle cells and that of normal follicle cells (Figure 5A–H). In a previous study, we found that cytoophidia are specifically located on the lateral and basal sides of follicle cells [21]. The polarity regulators of follicle cells and adherens junctions have certain effects on the maintenance of cytoophidia. To explore whether the cell membrane components of ingressive follicle cells would be affected when cytoophidium fails to form, we labeled the basolateral regulator Dlg of follicle cells, adherens junctions DE-Cadherin DCAD2, and cell membrane protein Hts. After immunostaining, in the follicle epithelium labeled with Hts and Dlg, there was no significant difference between the cell membrane of ingressive follicle cells and that of normal follicle cells (Figure 5A–H).

Under the condition of DCAD2 labeling, we found that DCAD2 showed an abnormal pattern in the ingressive follicle cells compared with normally arranged follicle cells. In normal follicle cells, the end near the germline cells is defined as the apical side and the end near the muscle layer is defined as the basal side. From the cross-sectional view of the lateral side of the stage 8 egg chamber, the DCAD2 pattern should be adjacent to the germline cells. However, in the ingressive follicle cell, DCAD2 could be seen flipping in the direction rather than at the apical side (Figure 6).

**Figure 5. Ingression of follicle cells labeled with different membrane proteins.** (**A**) Lateral view of a stage 8 egg chamber. Hts (red) labels the hulitaishao protein mainly presents at the lateral of follicle cell membranes. mCherry-tagged CTPS is shown green, and Hoechst (blue) for DNA. (**B**) A stage 8 egg chamber of *CTPSH355A Drosophila*. mCherry-tagged CTPSH355A is diffused. (**C**) The Hts

ingressive follicle cells and that of normal follicle cells (Figure 5A–H).

stage 8 egg chambers were counted per genotype.

2.3. Ingressive Follicle Cells Display Abnormal DCAD2 Pattern

**Figure 5. Ingression of follicle cells labeled with different membrane proteins.** (**A**) Lateral view of a stage 8 egg chamber. Hts (red) labels the hulitaishao protein mainly presents at the lateral of follicle cell membranes. mCherry-tagged CTPS is shown green, and Hoechst (blue) for DNA. (**B**) A stage 8 egg chamber of *CTPSH355A Drosophila*. mCherry-tagged CTPSH355A is diffused. (**C**) The Hts **Figure 5. Ingression of follicle cells labeled with different membrane proteins.** (**A**) Lateral view of a stage 8 egg chamber. Hts (red) labels the hulitaishao protein mainly presents at the lateral of follicle cell membranes. mCherry-tagged CTPS is shown green, and Hoechst (blue) for DNA. (**B**) A stage 8 egg chamber of *CTPSH355A Drosophila*. mCherry-tagged *CTPSH355A* is diffused. (**C**) The Hts pattern of an ingressive follicle cell. (**D**) Yellow arrow pointed out the ingressive cell nuclear. (**E**) Lateral view of a stage 10A egg chamber. Dlg (red) labels the discs large protein, which presents in the lateral and basal side of follicle cell. CTPS-mCherry cytoophidia can be observed, and Hoechst (blue) for DNA. (**F**) Lateral view of the *CTPSH355A* egg chamber at stage 10A. (**G**) The ingressive cell nuclear is stressed by the yellow arrow. (**H**) The Dlg pattern of the follicle cell ingression. (blue) for DNA. (**F**) Lateral view of the *CTPSH355A* egg chamber at stage 10A. (**G**) The ingressive cell nuclear is stressed by the yellow arrow. (**H**) The Dlg pattern of the follicle cell ingression. Under the condition of DCAD2 labeling, we found that DCAD2 showed an abnormal pattern in the ingressive follicle cells compared with normally arranged follicle cells. In normal follicle cells, the end near the germline cells is defined as the apical side and the end near the muscle layer is defined as the basal side. From the cross-sectional view of the lateral side of the stage 8 egg chamber, the DCAD2 pattern should be adjacent to the germline cells. However, in the ingressive follicle cell, DCAD2 could be seen flipping in the direction rather than at the apical side (Figure 6).

**Figure 4. Three dimensional view of follicle cell ingression.** (**A**) Cross section of a three-dimensional *CTPSH355A* egg chamber. A three-dimensional view of DNA stacked in layers. The interval between each layer is 0.5 μm, and a total of 12 layers are superimposed. The color from red to blue indicates the depth of DNA. (**B**) Side view of an ingressive nucleus. (**C**) On the xz plane, with the yellow dotted line marking the ingressive cell. (**D**) Quantification of the ingression frequency, 20

In a previous study, we found that cytoophidia are specifically located on the lateral and basal sides of follicle cells [21]. The polarity regulators of follicle cells and adherens junctions have certain effects on the maintenance of cytoophidia. To explore whether the cell membrane components of ingressive follicle cells would be affected when cytoophidium fails to form, we labeled the basolateral regulator Dlg of follicle cells, adherens junctions DE-Cadherin DCAD2, and cell membrane protein Hts. After immunostaining, in the follicle epithelium labeled with Hts and Dlg, there was no significant difference between the cell membrane of

**Figure 6. DCAD2 distribution is disturbed in** *CTPSH355A* **follicle epithelia.** (**A**) Cross section of Drosophila egg chamber expressing CTPS labelled by mCherry. The part of the follicle cell adjacent to the nurse cell is called apical, and the DE-Cadherin labeled by DCAD2 (red) is located at the apical of follicle cells. (**B**) DCAD2 together with CTPS as control to show the normal distribution of DCAD2. (**C**) Lateral view of a *CTPSH355A–mCh* egg chamber with abnormal follicle cell. (**D**) Yellow arrow pointed to the follicle cell ingression. (**E**) Ingressive folllicle cell pointed to by yellow arrows show abnormal distribution of DCAD2. *2.4. CTPSH355A Follicle Cells Increase the Heterogeneity*  **Figure 6. DCAD2 distribution is disturbed in** *CTPSH355A* **follicle epithelia.** (**A**) Cross section of Drosophila egg chamber expressing CTPS labelled by mCherry. The part of the follicle cell adjacent to the nurse cell is called apical, and the DE-Cadherin labeled by DCAD2 (red) is located at the apical of follicle cells. (**B**) DCAD2 together with CTPS as control to show the normal distribution of DCAD2. (**C**) Lateral view of a *CTPSH355A–mCh* egg chamber with abnormal follicle cell. (**D**) Yellow arrow pointed to the follicle cell ingression. (**E**) Ingressive folllicle cell pointed to by yellow arrows show abnormal distribution of DCAD2.

The follicle epithelium of *Drosophila* consists of a monolayer of follicle cells, which surround the oocyte and 15 nurse cells. Follicle cells gradually differentiate into various subpopulations, which will undergo morphological changes. After stage 6, the follicle

mal circumstances, each follicle cell contacts six adjacent cells, most of which are hexagonal and well arranged on the surface of egg chambers. However, in *CTPSH355A* mutant, the assembly of cytoophidia was disrupted and the number of non-hexagonal cells increased. We segmented the cell by the membrane staining of Hts protein, counting the different

## *2.4. CTPSH355A Follicle Cells Increase the Heterogeneity*

The follicle epithelium of *Drosophila* consists of a monolayer of follicle cells, which surround the oocyte and 15 nurse cells. Follicle cells gradually differentiate into various subpopulations, which will undergo morphological changes. After stage 6, the follicle cells cease mitosis and are arranged in a hexagonal pattern, which means that under normal circumstances, each follicle cell contacts six adjacent cells, most of which are hexagonal and well arranged on the surface of egg chambers. However, in *CTPSH355A* mutant, the assembly of cytoophidia was disrupted and the number of non-hexagonal cells increased. We segmented the cell by the membrane staining of Hts protein, counting the different shapes cell by cell. In *CTPSH355A* mutant, we observed many pentagonal follicle cells, and the heptagonal cells increased by about 10% (Figure 7A–D). *Int. J. Mol. Sci.* **2022**, *23*, x FOR PEER REVIEW 8 of 17 shapes cell by cell. In *CTPSH355A* mutant, we observed many pentagonal follicle cells, and the heptagonal cells increased by about 10% (Figure 7A–D).

**Figure 7. Morphological comparision of follicles between wild-type and** *CTPSH355A***.** (**A**–**F**) Surface view of stage 8 egg chambers. Membrane was labeled by Armadillo (red), and CTPS is shown in green. (**A**, **C**) *w1118* egg chamber. Scale bars, 50 μm. (**B**, **D**) *CTPSH355A* egg chamber. Scale bars, 20 μm. (**E**) Zoom-in view of (**C**). (**F**) Zoom-in view of (**D**). Redlines in E and F outline a central follicle cell with neighboring follicle cells in numbers. Note that the number of neighboring cells reflects the number of sides of the central polygonal follicle cells. (**G**) Quantitative analysis of the morphological difference between the wild-type control and *CTPSH355A* follicle cells (6 egg chambers were quantified per genotype, biological repeats = 3). **Figure 7. Morphological comparision of follicles between wild-type and** *CTPSH355A***.** (**A**–**F**) Surface view of stage 8 egg chambers. Membrane was labeled by Armadillo (red), and CTPS is shown in green. (**A**,**C**) *w <sup>1118</sup>* egg chamber. Scale bars, 50 µm. (**B**,**D**) *CTPSH355A* egg chamber. Scale bars, 20 µm. (**E**) Zoom-in view of (**C**). (**F**) Zoom-in view of (**D**). Redlines in E and F outline a central follicle cell with neighboring follicle cells in numbers. Note that the number of neighboring cells reflects the number of sides of the central polygonal follicle cells. (**G**) Quantitative analysis of the morphological difference between the wild-type control and *CTPSH355A* follicle cells (6 egg chambers were quantified per genotype, biological repeats = 3).

The quantification further confirmed that compared with *w1118*, there was difference in the number of heptagonal follicle cells in stage 8 egg chambers of the *CTPSH355A* mutant. Moreover, the number of hexagonal follicle cells in *CTPSH355A* egg chambers decreased by about 15% at stage 8, while the number of polygonal cells increased by about 10% (Figure 7E–G). Con-

that when cytoophidia cannot be formed, the cell membrane will be affected, and the tight arrangement of epithelial follicle cells will not be maintained. Morphology changes indicate that cytoophidia located at the basolateral side may play a role in maintaining the

integrity of follicle epithelium.

The quantification further confirmed that compared with *w <sup>1118</sup>*, there was difference in the number of heptagonal follicle cells in stage 8 egg chambers of the *CTPSH355A* mutant. Moreover, the number of hexagonal follicle cells in *CTPSH355A* egg chambers decreased by about 15% at stage 8, while the number of polygonal cells increased by about 10% (Figure 7E–G). Considering that the more sides the polygon, the closer it is to the round circle, we speculate that when cytoophidia cannot be formed, the cell membrane will be affected, and the tight arrangement of epithelial follicle cells will not be maintained. Morphology changes indicate that cytoophidia located at the basolateral side may play a role in maintaining the integrity of follicle epithelium.

## *2.5. Follicle Epithelia Reduce Compactness in CTPSH355A Mutant*

In the wild-type flies, follicle epithelial cells at stages 4–9 of oogenesis are tightly packed, and most of the hexagonal follicle cells enclose germ cells. In the case that CTPS could not be assembled into cytoophidia, we observed that closely arranged epithelial follicle cells became relatively loose, and follicle cells of similar size in the wild-type became relatively very large or very small, which was not conducive to compact arrangement (Figure 8A–L).

(**B**) Morphology of cell membrane. The yellow rectangle highlights an area of the follicle epithelium. (**C**) Enlarged part of the framed region in B. Larger follicular cells are circled in red, and smaller follicular cells are circled in blue. (**D**) CTPS-mCherry (green) and Hts (red) of the same egg chamber shown in A. (**E**) The area of each cell was measured after dividing each cell along the cell membrane. (**F**) corresponds to (**D**). (**G**) Surface view of a *CTPSH355A* stage 8 egg chamber. CTPSH355A-mcherry knock-in (green), Hts (red, labels cell membrane), and DNA (blue, labeled with Hoechst 33342). (**H**) Morphology of the cell membrane. A yellow rectangle highlights an area of the follicle epithelium. (**I**) Enlarged part of the framed region in (**H**). Larger follicle cells are circled in red, and smaller follicle cells are circled in blue. (**J**) CTPSH355A-mCherry (green) and Hts (red) of the same egg chamber shown in (E). Surface view of the cell membrane with CTPS. (**K**) The area of each cell was measured after dividing each cell along the cell membrane. (**L**) corresponds to (**K**). Scale bars, 20 µm. (**M**) The ratio of the average area of the three largest cells to the average area of the three smallest cells in an egg chamber. N = 6, \*\*\*, *p* < 0.0001. Mann-Whitney U test. (5 stage 7–8 egg chambers/genotypes, 3 biological replicates) (**N**) Average follicle cell surface. N = 3.

In order to further clarify the observed phenomenon, we segmented the follicle cell surface based on the cell membrane and calculated the basal area of each follicle cell on an egg chamber through software. The average area ratio of the group with the largest area of three adjacent follicle cells and the group with the smallest area of three adjacent follicle cells was used as an indicator of follicle cell heterogeneity. The higher the ratio, the higher the heterogeneity of surface follicle cell. The quantitative analysis showed that the average area of follicle cells at stage 8 *CTPSH355A* was smaller than that of the wild-type, but the heterogeneity was much higher than that of the wild-type (Figure 8M,N).

## *2.6. Follicle Cell Ingression Occurs in Egg Chamber Overexpessing CTPSH355A*

In previous studies, our laboratory found that the formation of cytoophidia can prolong the half-life of CTPS protein in mammalian cells. Therefore, we want to know whether H355A point mutation affects CTPS protein level in *Drosophila* ovaries. Western blot results confirmed that the level of CTPS protein in *Drosophila* ovaries after *CTPSH355A* mutation was lower than that in the *wild-type* (Figure 9H,I). Thus, we want to investigate whether the phenotypes observed in the *CTPSH355A* strain are caused by the decrease of CTPS protein level.

To eliminate the influence of protein level, we used the *Actin-Gal4* driver to overexpress *CTPSH355A* in *Drosophila* ovaries. We constructed *Actin-Gal4-driven Drosophila* strains overexpressing *CTPSH355A* (*Actin > UAS CTPSH355A-mCherry-OE*) or *wild-type CTPS* (*Actin > UAS CTPS-mCherry-OE*). Western blot confirmed that there was no significant difference in CTPS level between the *Actin > UAS CTPS-mCherry-OE* heterozygous strain and the *Actin > UAS CTPSH355A-mCherry-OE* homozygous strain (Figure 9J,K).

We found that the distribution of cytoophidia in the basolateral side of follicle cells could be clearly observed in *Actin > UAS CTPS-mCherry-OE* heterozygous egg chambers (Figure 9A,B). Almost every follicle cell had one or two cytoophidia, and follicle cells were arranged in a single layer. In *Actin > UAS CTPSH355A-mCherry-OE* homozygous flies, the diffused distribution of CTPSH355A could be observed, and the ingressive follicle cells appeared as well (Figure 9C–G). These results indicate that loss of the cytoophidiumforming ability of CTPS, rather than its protein level, is the primary cause of follicle cell ingression in the *CTPSH355A* mutant.

**Figure 9. Follicle cell ingression occurs in egg chamber overexpressing** *CTPSH355A***.** (**A**,**B**) Side view of stage 8 egg chamber overexpressing *CTPS*. (**C**) Cross section of stage 8 egg chamber overexpressing *CTPSH355A*. The ingression is framed by a rectangle. (**D**–**G**) Zoom-in images of the ingression follicle cells in (**C**). Scale bars, 20. (**H**) Western blot is detected with antibodies against mCherry and tubulin on the ovarian lysates of *CTPS-mCh* and *CTPSH355A-mCh* mutants. Scale bars, 20 μm. (**I**) Quantitative analysis of the CTPS protein level of samples represented in (**H**), the mean and standard deviation. (**J**) Western blot of the ovarian lysates of *Actin > UAS CTPS-OE* and *Actin > UAS CTPSH355A-OE* mutants, detected with antibodies against mCherry and tubulin. (**K**) Quantitative analysis of the CTPS protein level of samples represented in (**J**), the mean and standard deviation. **Figure 9. Follicle cell ingression occurs in egg chamber overexpressing** *CTPSH355A***.** (**A**,**B**) Side view of stage 8 egg chamber overexpressing *CTPS*. (**C**) Cross section of stage 8 egg chamber overexpressing *CTPSH355A*. The ingression is framed by a rectangle. (**D**–**G**) Zoom-in images of the ingression follicle cells in (**C**). Scale bars, 20. (**H**) Western blot is detected with antibodies against mCherry and tubulin on the ovarian lysates of *CTPS-mCh* and *CTPSH355A-mCh* mutants. Scale bars, 20 µm. (**I**) Quantitative analysis of the CTPS protein level of samples represented in (**H**), the mean and standard deviation. (**J**) Western blot of the ovarian lysates of *Actin > UAS CTPS-OE* and *Actin > UAS CTPSH355A-OE* mutants, detected with antibodies against mCherry and tubulin. (**K**) Quantitative analysis of the CTPS protein level of samples represented in (**J**), the mean and standard deviation. \*, *p* < 0.05; Mann-Whitney U test.

In order to further clarify the observed phenomenon, we segmented the follicle cell surface based on the cell membrane and calculated the basal area of each follicle cell on an egg chamber through software. The average area ratio of the group with the largest area of three adjacent follicle cells and the group with the smallest area of three adjacent follicle cells was used as an indicator of follicle cell heterogeneity. The higher the ratio, the higher the heterogeneity of surface follicle cell. The quantitative analysis showed that the average area of follicle cells at stage 8 *CTPSH355A* was smaller than that of the wild-type,

In previous studies, our laboratory found that the formation of cytoophidia can prolong

the half-life of CTPS protein in mammalian cells. Therefore, we want to know whether H355A point mutation affects CTPS protein level in *Drosophila* ovaries. Western blot results confirmed that the level of CTPS protein in *Drosophila* ovaries after *CTPSH355A* mutation was lower than

but the heterogeneity was much higher than that of the wild-type (Figure 8M,N).

*2.6. Follicle Cell Ingression Occurs in Egg Chamber Overexpessing CTPSH355A*

served in the *CTPSH355A* strain are caused by the decrease of CTPS protein level.

#### ns, non-significant; \*, p < 0.05; Mann-Whitney U test. *2.7. Overexpession of CTPSH355A Increases the Heterogeneity of Follicle Cells*

Similarly, we wanted to examine whether the heterogeneity of follicle cells was affected by the level of CTPS protein. According to our study, there was a long and curly cytoophidium in each follicle cell on the surface of *Actin > UAS CTPS-mCherry-OE* heterozygous egg chamber. Compared with the wild-type egg chambers, where cytoophidia are mostly rod-shaped and distributed along the cell membrane, the elongated cytoophidia were still distributed along the cell membrane after the overexpression of CTPS (Figure 10A–C).

sion in the *CTPSH355A* mutant.

**Figure 10. The integrity of follicle epithelium is compromised when overexpressing** *CTPSH355A***.**  (**A**–**C**) Surface view of a stage 8 egg chamber with follicle cells overexpressing *CTPS*. Large cytoophidia are detectable in almost all follicle cells. (**D**–**F**) Surface view of a stage 8 egg chamber with follicle cells overexpressing *CTPSH355A*. Note that the heterogenous sizes of follicle cells and increased gaps between neighbouring follicle cells. CTPS-mCherry (green), Hts (red) and DNA (blue, labelled with Hoechst 33342). Scale bars, 20 μm. **Figure 10. The integrity of follicle epithelium is compromised when overexpressing** *CTPSH355A***.** (**A**–**C**) Surface view of a stage 8 egg chamber with follicle cells overexpressing *CTPS*. Large cytoophidia are detectable in almost all follicle cells. (**D**–**F**) Surface view of a stage 8 egg chamber with follicle cells overexpressing *CTPSH355A*. Note that the heterogenous sizes of follicle cells and increased gaps between neighbouring follicle cells. CTPS-mCherry (green), Hts (red) and DNA (blue, labelled with Hoechst 33342). Scale bars, 20 µm.

To eliminate the influence of protein level, we used the *Actin-Gal4* driver to overexpress *CTPSH355A* in *Drosophila* ovaries. We constructed *Actin-Gal4-driven Drosophila* strains overexpressing *CTPSH355A* (*Actin > UAS CTPSH355A-mCherry-OE*) or *wild-type CTPS* (*Actin > UAS CTPS-mCherry-OE*). Western blot confirmed that there was no significant difference in CTPS level between the *Actin > UAS CTPS-mCherry-OE* heterozygous strain and the

We found that the distribution of cytoophidia in the basolateral side of follicle cells could be clearly observed in *Actin > UAS CTPS-mCherry-OE* heterozygous egg chambers (Figure 9A,B). Almost every follicle cell had one or two cytoophidia, and follicle cells were arranged in a single layer. In *Actin > UAS CTPSH355A-mCherry-OE* homozygous flies, the diffused distribution of CTPSH355A could be observed, and the ingressive follicle cells appeared as well (Figure 9C–G). These results indicate that loss of the cytoophidium-forming ability of CTPS, rather than its protein level, is the primary cause of follicle cell ingres-

Similarly, we wanted to examine whether the heterogeneity of follicle cells was affected by the level of CTPS protein. According to our study, there was a long and curly cytoophidium in each follicle cell on the surface of *Actin > UAS CTPS-mCherry-OE* heterozygous egg chamber. Compared with the wild-type egg chambers, where cytoophidia are mostly rod-shaped and distributed along the cell membrane, the elongated cytoophidia were still distributed along the cell membrane after the overexpression of CTPS (Figure 10A–C).

*Actin > UAS CTPSH355A-mCherry-OE* homozygous strain (Figure 9J,K).

*2.7. Overexpession of CTPSH355A Increases the Heterogeneity of Follicle Cells* 

The diffused distribution of CTPS was confirmed on the surface of the egg chamber of *Actin > UAS CTPSH355A-mCherry-OE* homozygous fly. The changes of morphology and cell size showed that the heterogeneity of follicle cells was enhanced because there was no cytoophidium on the cell membrane. It seemed that these follicle cells could not even be arranged tightly (Figure 10D–F). The diffused distribution of CTPS was confirmed on the surface of the egg chamber of *Actin > UAS CTPSH355A-mCherry-OE* homozygous fly. The changes of morphology and cell size showed that the heterogeneity of follicle cells was enhanced because there was no cytoophidium on the cell membrane. It seemed that these follicle cells could not even be arranged tightly (Figure 10D–F).

#### *2.8. Follicle Cell-Specific Overexpression of CTPSH355A Impairs the Integrity of Follicle Epithelium 2.8. Follicle Cell-Specific Overexpression of CTPSH355A Impairs the Integrity of Follicle Epithelium*

After excluding the influence of CTPS protein level, we wanted to further eliminate the effect of the inter-tissue interaction caused by the ubiquitous expression of *Actin-Gal4*. To this end, we constructed a strain overexpressing CTPSH355A using the same *UAS CTPSH355A-*After excluding the influence of CTPS protein level, we wanted to further eliminate the effect of the inter-tissue interaction caused by the ubiquitous expression of *Actin-Gal4*. To this end, we constructed a strain overexpressing CTPSH355A using the same *UAS CTPSH355AmCherry-OE* and *UAS CTPS-mCherry-OE* strains together with *Tj-Gal4* specifically expressed in follicle cells. As a control, *wild-type CTPS* was overexpressed specifically in follicle cells using the *Tj-Gal4* driver. Our results confirmed that the integrity of follicle epithelium was impaired when *CTPSH355A* was overexpressed specifically in follicle cells (Figure 11A–N).

## *2.9. Space between Muscle Sheath and Egg Chamber Increases in CTPSH355A*

IF is a member of the integrin complex and widely exists in the muscle layer that encloses the ovarioles. We found that in the wild-type ovary, the muscle sheath tightly wrapped the ovarioles and drove their movement (Figure 12A,B), which was conducive to common life activities such as oogenesis. In *CTPSH355A* ovaries, the space between the muscle sheath and the egg chamber was significantly increased, and the egg chamber almost collapsed from the muscle sheath (Figure 12C–E), which might affect normal physiological activities.

\*\*, p = 0.002.

*mCherry-OE* and *UAS CTPS-mCherry-OE* strains together with *Tj-Gal4* specifically expressed in follicle cells. As a control, *wild-type CTPS* was overexpressed specifically in follicle cells us-

paired when *CTPSH355A* was overexpressed specifically in follicle cells (Figure 11A–N).

**Figure 11. Follicle cell ingression occurs with follicle cell-specific overexpresson of** *CTPSH355A***.**  (**A**), (**B**) Side view of a stage 8 egg chamber overexpressing CTPS in follicle epithelium. (**C**) Cross section of stage 8 egg chamber after overexpression of CTPSH355A in follicle epithelium. The ingression is framed by a rectangle. (**D–G**) Zoom-in of the ingressive follicle cell in (**C**). (**H–J**) Surface view of a stage 8 egg chamber with follicle cell-specific overexpression of CTPS. Large cytoophidia are detectable in almost all follicle cells. Note that most cytoophidia are distributed on or near the cortex of follicle cells. (**K–M**) Surface view of a stage 8 egg chamber with follicle cell-specific overexpression of CTPSH355A. No cytoophidium is detectable. CTPS-mCherry (green), Hts (red) and DNA (blue, labelled with Hoechst 33342). Scale bars, 20 μm. (**N**) Quantitative analysis of the ratio of three largest cells versus three smallest cells (5 images/genotypes, 3 biological replicates). Mann-Whitney U test, **Figure 11. Follicle cell ingression occurs with follicle cell-specific overexpresson of** *CTPSH355A***.** (**A**,**B**) Side view of a stage 8 egg chamber overexpressing CTPS in follicle epithelium. (**C**) Cross section of stage 8 egg chamber after overexpression of CTPSH355A in follicle epithelium. The ingression is framed by a rectangle. (**D–G**) Zoom-in of the ingressive follicle cell in (**C**). (**H–J**) Surface view of a stage 8 egg chamber with follicle cell-specific overexpression of CTPS. Large cytoophidia are detectable in almost all follicle cells. Note that most cytoophidia are distributed on or near the cortex of follicle cells. (**K–M**) Surface view of a stage 8 egg chamber with follicle cell-specific overexpression of CTPSH355A. No cytoophidium is detectable. CTPS-mCherry (green), Hts (red) and DNA (blue, labelled with Hoechst 33342). Scale bars, 20 µm. (**N**) Quantitative analysis of the ratio of three largest cells versus three smallest cells (5 images/genotypes, 3 biological replicates). Mann-Whitney U test, \*\*, *p* = 0.002.

the ovarioles. We found that in the wild-type ovary, the muscle sheath tightly wrapped the ovarioles and drove their movement (Figure 12A,B), which was conducive to common life activities such as oogenesis. In *CTPSH355A* ovaries, the space between the muscle sheath and the egg chamber was significantly increased, and the egg chamber almost collapsed from the mus-

*2.9. Space between Muscle Sheath and Egg Chamber Increases in CTPSH355A*

cle sheath (Figure 12C–E), which might affect normal physiological activities.

**Figure 12. Compared with wild-type control, the space increases between muscle sheath and egg chambers in** *CTPSH355A***.** (**A**,**B**) A stage 8 egg chamber of the *CTPSH355A* mutant. Integrin is labeled with knock-in GFP (green). CTPSH355A is labeled with knock-in mCherry (red). Yellow arrows point to gap between egg chamber and the muscle sheath. (**C**,**D**) A stage 8 wild-type egg chamber. Integrin is labeled by knock-in GFP. CTPS is labeled by knock-in mCherry. Scale bars, 20 μm. (**E**) The ratio of GFP intensity of the integrin to DNA from **A**–**D**. The value is normalized to the control (5 images/genotypes, 3 biological replicates). Mann-Whitney U test. \*\*\*, p = 0.0005. **Figure 12. Compared with wild-type control, the space increases between muscle sheath and egg chambers in** *CTPSH355A***.** (**A**,**B**) A stage 8 egg chamber of the *CTPSH355A* mutant. Integrin is labeled with knock-in GFP (green). CTPSH355A is labeled with knock-in mCherry (red). Yellow arrows point to gap between egg chamber and the muscle sheath. (**C**,**D**) A stage 8 wild-type egg chamber. Integrin is labeled by knock-in GFP. CTPS is labeled by knock-in mCherry. Scale bars, 20 µm. (**E**) The ratio of GFP intensity of the integrin to DNA from (**A**–**D**). The value is normalized to the control (5 images/genotypes, 3 biological replicates). Mann-Whitney U test. \*\*\*, *p* = 0.0005.

#### **3. Discussion 3. Discussion**

To explore the physiological function of cytoophidia in *Drosophila* follicle cells, we analyze the changes in follicle cells in CTPS mutant when cytoophidia cannot be formed. Our results indicate that the integrity of follicle epithelium is compromised when CTPS lose its cytoophidium-forming capability. In this study, we generate transgenic flies with a point mutation in CTPS. Mutations To explore the physiological function of cytoophidia in *Drosophila* follicle cells, we analyze the changes in follicle cells in CTPS mutant when cytoophidia cannot be formed. Our results indicate that the integrity of follicle epithelium is compromised when CTPS lose its cytoophidium-forming capability.

do not affect enzymatic activity but lead to the disassembly of cytoophidia. In the mutant flies, the integrity of follicle epithelia is impaired with two related phenotypes: (1) ingression of follicle cells and (2) heterogeneous follicle cells. We have previously discovered that cytoophidia are specifically distributed on the In this study, we generate transgenic flies with a point mutation in CTPS. Mutations do not affect enzymatic activity but lead to the disassembly of cytoophidia. In the mutant flies, the integrity of follicle epithelia is impaired with two related phenotypes: (1) ingression of follicle cells and (2) heterogeneous follicle cells.

basolateral side of follicle cells [21]. Moreover, when the polarity of follicle cells is disrupted, cytoophidia will become unstable, especially due to the disruption of apical regulators. In this study, the apical polarity of follicle cells is indeed affected by the absence of cytoophidia. Our data indicate that the cytoophidium, as a kind of membraneless orga-We have previously discovered that cytoophidia are specifically distributed on the basolateral side of follicle cells [21]. Moreover, when the polarity of follicle cells is disrupted, cytoophidia will become unstable, especially due to the disruption of apical regulators. In this study, the apical polarity of follicle cells is indeed affected by the absence of cytoophidia.

nelle, maintains its specific subcellular localization in biological processes.

Our data indicate that the cytoophidium, as a kind of membraneless organelle, maintains its specific subcellular localization in biological processes.

In these experiments, we also notice that cytoophidia play a role in maintaining the integrity of follicle epithelium. We speculate that cytoophidia located at the basolateral side of follicle cells may play a role in supporting follicle cells. In the absence of cytoophidia, the mechanical tension of the follicle cell membrane will be reduced, making it more difficult to maintain the cell morphology. Therefore, follicle cells are more likely to be drawn into polygons and expanded by surrounding cells or squeezed and reduced by surrounding cells. Similarly, due to the weakening of membrane mechanical tension, the follicle cells migrating inward after mitosis cannot be reintegrated into the follicular monolayer, resulting in the ingression. Our laboratory also found that in the male reproductive system of *Drosophila*, when CTPS cannot form cytoophidia, the main cells on the surface of the accessory gland may be difficult to maintain their cell shape, and two horizontally arranged nuclei appear to be vertically arranged. This further support our hypothesis [30].

When cytoophidia are disassembled, the observed separation of egg chamber and muscle sheath may also be due to the disappearance of the supporting force of cytoophidia. When cytoophidia cannot be formed, the internal supporting force of each follicle cell is weakened, leading to the collapse of the entire egg chamber. Considering that the follicle epithelium will develop into the eggshell of a fertilized egg in the later stages [31], it is possible that its shell hardness and the hatchability of the fertilized egg will also be affected accordingly.

However, we could not simply rescue the phenotypes found in the *CTPSH355A-mCh* mutant by expressing CTPS-mCherry protein. Our previous studies in mammals [26] and *Drosophila* [27] confirmed that the CTPSH355A point mutation is dominant-negative, which is to say that as long as the CTPSH355A protein exists, the CTPS protein would not be able to assemble into cytoophidium [26,27]. Because the H355A point mutation of CTPS would disrupt the assembly of the cytoophidia dominant negatively, we have analyzed the *CTPSH355A/TM6B-mCh* egg chambers and found that *CTPSH355A/TM6B* also have defects in follicle epithelial integrity mentioned above (Supplementary Figure S1). These results further validate our hypothesis that the cytoophidium structure plays a certain role in maintaining epithelial integrity, and the dominant negative CTPS point mutation confirmed that it is crucial for the assembly of cytoophidium.

Since the first discovery of cytoophidia in our laboratory in 2010, great progress has been made in the research on the existence of cytoophidia in different species and different types of cells. However, knowledge concerning the function of this new type of organelle widely existing in organisms is still in the initial stage. Therefore, our work has potential reference value for understanding the role of cytoophidia in *Drosophila* follicle cells. Our results indicate that forming cytoophidia is crucial to epithelial integrity.

#### **4. Materials and Methods**

#### *4.1. Fly Stocks*

All stocks were maintained at 25 ◦C on standard cornmeal. Both *w <sup>1118</sup>* and C-terminal mChe-4V5 tagged CTPS knock-in flies out of *w <sup>1118</sup>* produced in our laboratory were used as wild-type controls unless stated otherwise. The stocks used were: (1) *CTPSH335A mutated with mChe-4V5 tagged CTPS Knock-in fly*, (2) *Actin-Gal4/Cyo* (A gift from Guanjun Gao's lab, ubiquitous expression under strong promoter, a chromosome II insertion balanced over Curly of Oster [32]), (3) *Tj-Gal4* (A gift from Kun Dou's lab [33,34]), (4) *Sp/Cyo; Sb/Tm6B* (Institute of Biochemistry and Cell Biology, Chinese Academy of Sciences, Drosophila Resources and Technology Platform), (5) *UAS CTPS-mCherry-OE, and UAS CTPSH355AmCherry-OE* [27].

#### *4.2. Fly Genetics*

Transgenic flies expressing full-length CTPS (isoform C, LD25005) under the UAS promoter (*UAS-CTPS*) were generated in our lab (Back cross 5 generations before use) [35]. To generate the *Tj > UAS CTPS-OE*, we crossed the homozygous *UAS CTPS-OE Drosophila* and *Tj-Gal4 Drosophila* with the double balancer *Drosophila* for one generation, and the target generation was inbred for two generations to get the homozygous *Tj-Gal4*; *UAS CTPS-mCherry-OE Drosophila*. Same as the Tj-Gal4; *UAS CTPSH355A-mCherry-OE; Actin-Gal4; UAS CTPS-mCherry-OE* and *Actin-Gal4*; *UAS CTPSH355A-mCherry-OE* strains.

#### *4.3. Generation of Transgenic Flies*

For polymerase chain reaction (PCR), PUASTattb plasmids were used as the template, and phanta Maxa Super-Fidelity DNA Polymerase (Vazyme, #P505) as the polymerase. Sequences for primers were as below:

H355A-F: GAGCAAGTACGCCAAGGAGTGGCAGAAGCTATGCGATAGCCAT;

H355A-R: TGCCACTCCTTGGCGTACTTGCTCGGCTCAGAATGCAAAGTTT

After obtaining the required plasmids, the *CTPSH355A Drosophila* strain was constructed by microinjection.

#### *4.4. Immunohistochemistry*

Ovaries from flies were dissected in Grace's Insect Medium (Gibco) and then fixed in 4% formaldehyde (Sigma) diluted in PBS for 10 min before immunofluorescence staining. The samples were then washed twice using PST (0.5% horse serum + 0.3% Triton × 100 in PBS). For membrane staining, samples were incubated with primary antibodies at room temperature overnight, and then washed using PST. Secondary antibodies were used to incubate the samples at room temperature for another night.

Primary antibodies used in this study were rabbit anti-CTPS (1:1000; y-88, sc-134457, Santa Cruz BioTech Ltd., Santa Cruz, CA, USA), mouse anti-Discs Large (1:500, Developmental Studies Hybridoma Bank, Iowa City, IA, USA), mouse anti-D-E Cadherin (1:500, Developmental Studies Hybridoma Bank), mouse anti-HTS (1:1000, Developmental Studies Hybridoma Bank, Cat. No. AB\_528070), mouse anti-Armadillo (1:500, Developmental Studies Hybridoma Bank). Secondary antibodies used in this study were anti-mouse, rabbit, or goat antibodies that were labeled with Alexa Fluor® 488 (Molecular Probes), or with Cy5 (Jackson ImmunoResearch Laboratories, West Grove, PA, USA). Hoechst 33342 was used to label DNA.

#### *4.5. Microscopy and Image Analysis*

All images were obtained under laser-scanning confocal microscopy (Zeiss 880). Image processing was performed using Zeiss Zen. ImageJ was used to analyze the area and number of follicle cells.

We used the ImageJ SCF to segment the follicle cells by the membrane, then use ImageJ cell counter to calculate different shapes of cells to get the number of polygonal cells. We used ImageJ to measure the area of each cell. For each statistical quantification, we collected the surface images using Zeiss 880 with the interval as 0.5 µm for z-stack, 5 stage 8 egg chambers were quantified per genotype, biological repeats = 3. Mann-Whitney U test was conducted for comparison.

#### *4.6. Western Blotting*

Female adult ovaries of *Drosophila* were collected with gathered into lysis buffer RIPA (Meilunbio, Dalian, China) with protease inhibitor cocktail (Bimake, Shanghai, China) for Western blotting, and then ground with 1 mm Zirconia beads in Sonicator (Shanghai Jing Xin, Shanghai, China). The sample would then lysis on ice for up to 30 min. Samples were centrifuged for 10 min at 10,000 g at 4 ◦C. The 6× protein loading buffer was pipetted into the supernatants and boiled at 99 ◦C for 15 min to obtain protein. Then, the protein sample was run through 10% SDS-PAGE gels and transferred to PVDF membranes. At

room temperature, membrane was incubated with 5% *w*/*v* nonfat dry milk dissolved by 1× TBST for 1 h of blocking. Then, the membrane was incubated with primary antibodies in 5% *w*/*v* nonfat milk at 4 ◦C and gently shaken overnight.

The following primary antibodies were used in this study: anti-mCherry Tag Monoclonal antibodies (Cat. No. A02080, Abbkine, Beijing China), mouse anti-a-Tubulin antibodies (Cat. No. T6199, Sigma). The membranes were washed three times for 5 min per time with shaking, then incubated with secondary antibodies (anti-mouse IgG, HRPlinked antibody, Cell Signaling, Danvers, MA, USA) diluted in 5% *w*/*v* nonfat milk at room temperature for 1 h. An Amersham Imager 600 (General Electric, Boston, MA, USA) and Pierce ECL Reagent Kit (Cat. No. 32106, Thermo Fisher, Waltham, MA, USA) were adopted for the chemiluminescence immunoassay. Protein levels were quantified on ImageJ (National Institutes of Health, Bethesda, MD, USA) and normalized to tubulin. At least three biological replicates were quantified.

#### *4.7. Data Analysis*

Images collected by confocal microscopy were processed using Adobe Illustrator and ImageJ. Cell segmentation based on the cell membrane was achieved using CellPose and SCF methods. Quantitative analysis was processed by Excel and GraphPad. The Mann–Whitney U test was conducted to get the *p*-value.

**Supplementary Materials:** The following supporting information can be downloaded at: https: //www.mdpi.com/article/10.3390/ijms232315282/s1, Figure S1. *CTPSH355A/TM6B* mutants have no significant difference from *CTPSH355A* mutant.

**Author Contributions:** Q.-Q.W. and J.-L.L. conceived the studies. Q.-Q.W. and J.-L.L. performed the experiments. D.-D.Y. assisted fly genetics. Q.-Q.W. drafted the manuscript. J.-L.L. revised the manuscript. All authors have read and agreed to the published version of the manuscript.

**Funding:** This work was funded by the Ministry of Science and Technology of the People's Republic of China (Grant No. 2021YFA0804700), the National Natural Science Foundation of China (No. 31771490), the Shanghai Science and Technology Commission (20JC1410500), and the UK Medical Research Council (Grant No. MC\_UU\_12021/3 and MC\_U137788471). This research was supported by Shanghai Frontiers Science Center for Biomacromolecules and Precision Medicine at ShanghaiTech University.

**Institutional Review Board Statement:** Not applicable.

**Informed Consent Statement:** Not applicable.

**Data Availability Statement:** Not applicable.

**Acknowledgments:** We thank the Molecular Imaging Core Facility (MICF) at School of Life Science and Technology, ShanghaiTech University for providing technical support.

**Conflicts of Interest:** The authors have no relevant financial or non-financial interest to disclose.

#### **References**


## *Article* **Super-Resolution Imaging Reveals Dynamic Reticular Cytoophidia**

**Yi-Fan Fang <sup>1</sup> , Yi-Lan Li <sup>1</sup> , Xiao-Ming Li <sup>1</sup> and Ji-Long Liu 1,2,\***


**Abstract:** CTP synthase (CTPS) can form filamentous structures termed cytoophidia in cells in all three domains of life. In order to study the mesoscale structure of cytoophidia, we perform fluorescence recovery after photobleaching (FRAP) and stimulated emission depletion (STED) microscopy in human cells. By using an EGFP dimeric tag as a tool to explore the physical properties of cytoophidia, we find that cytoophidia are dynamic and reticular. The reticular structure of CTPS cytoophidia may provide space for other components, such as IMPDH. In addition, we observe CTPS granules with tentacles.

**Keywords:** CTP synthase; cytoophidium; fluorescence recovery after photobleaching (FRAP); stimulated emission depletion (STED)

## **1. Introduction**

In addition to organelles with membranes, proteins with important functions in the cell can also be compartmented into membraneless organelles. CTP synthase (CTPS), a metabolic enzyme for de novo synthesis of CTP, was found to form filament-like compartments in cells called cytoophidia [1]. Describing their shape vividly, the word "cytoophidia" means "cellular snakes" in Greek. Cytoophidia were found in many species in all three domains of life, which means cytoophidia are conserved in evolution [1–19].

A glutamine analog, 6-diazo-5-oxo-L-norleucine (DON), promotes cytoophidia formation in *Drosophila* and human cells [5]. DON binds CTPS with covalent bonds [20]. Glutamine deprivation promotes cytoophidium formation in mammalian cells [21]. IM-PDH can form cytoophidia [22] and both CTPS and IMPDH are related to glutamine and NH<sup>3</sup> metabolism. It was reported that there is an interaction between CTPS and IMPDH [23]. The function of cytoophidia may be closely related to glutamine and NH<sup>3</sup> metabolism.

In metabolic regulation, the activity of CTPS is inhibited via filament formation [24,25]. The half-life of CTPS is prolonged when forming cytoophidia [26]. Given the high-level metabolism in cancer cells, cytoophidium formation is highly related to oncogenes. Myc is required for cytoophidia assembly, and cytoophidia formation is regulated by Myc expression levels [27]. Ack kinase regulates cytoophidium morphology and CTPS activity [28]. Cytoophidium assembly was found to be regulated by the mTOR-S6K1 pathway [29]. Cytoophidia were also found in human hepatocellular carcinoma [10].

CTPS can be assembled into thin filaments in vitro, and the structures of CTPS filaments at near-atomic resolution have been solved by cryo-EM [20,25,30]. However, nanometer-scale CTPS filaments are different from the micron-scale cytoophidia observed via confocal microscopy. How CTPS filaments assemble into big micron-scale cytoophidia is still unclear. The physical properties of cytoophidia at the mesoscale remain to be explored.

To study cytoophidium properties in human cell lines, we performed fluorescence recovery after photobleaching (FRAP) microscopy to study the dynamic characteristics and

**Citation:** Fang, Y.-F.; Li, Y.-L.; Li, X.-M.; Liu, J.-L. Super-Resolution Imaging Reveals Dynamic Reticular Cytoophidia. *Int. J. Mol. Sci.* **2022**, *23*, 11698. https://doi.org/10.3390/ ijms231911698

Academic Editor: Vladimir N. Uversky

Received: 24 August 2022 Accepted: 27 September 2022 Published: 2 October 2022

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2022 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

stimulated emission depletion (STED) microscopy to study the super-resolution structure. By measuring the intensity and recovery speed of bleached ROIs, we were able to quantify the relative dynamic characteristics of cytoophidia under different treatments. STED allows fluorescence imaging to achieve a resolution of 50 to 70 nm [31,32].

#### **2. Results**

#### *2.1. Assembly of CTPS Filaments into Cytoophidia*

The cytoophidium is a compartment of metabolic enzymes, such as filamentous CTPS, which can be observed via confocal microscopy [1–3]. In vitro experiments showed that CTPS can also form filaments built from tetramer units [20,25,30]. In human cells, CTPS can be assembled into cytoophidia under DON treatment or glutamine deprivation. In this study, we did not distinguish hCTPS1 and hCTPS2. We found that after DON treatment, hCTPS1 can form granules in 293T cells with hCTPS1 over-expression (Figure 1A,B). When we observed living cells, CTPS granules existed in a small population of cells. CTPS granules can exist in the same cells as cytoophidia (Figure 1A,B).

How CTPS filaments are arranged in cytoophidia remains unclear. To conceive the arrangement model from CTPS filaments to cytoophidia, we constructed hCTPS1 overexpression vectors with different fluorescence proteins and different mutations (Supplemental Table S1). In order to show and evaluate the effect of exogenous protein overexpression in the experiment, we compared the protein levels of overexpressed hCTPS1 and endogenous hCTPS1/2 (Figure 1D) and tested the transfection efficiencies (Figure 1E,F). In transfection-positive cells, exogenous hCTPS1 expression was approximately twice as high as that of endogenous hCTPS1/2 (Figure 1G).

EGFP is a weak dimer while EGFPA206K is a monomer [33]. Overexpression of *hCTPS1- EGFP* forms cytoophidium-like condensates in 293T cells (Figure 1H). The force of forming dimer between EGFP pulls hCTPS1 together, and hCTPS is assembled into filaments. The hCTPS filaments may be compressed together by a simple force of EGFP dimerization (Figure 1C). The hCTPS1-EGFP group was a control for cytoophidium induction. CTPS1 with the H355A mutation disassembles cytoophidia. We found that overexpressed *hCTPS1H355A-EGFP* could not form cytoophidia in 293T cells (Figure 1H).

#### *2.2. Dynamic Equilibria of Cytoophidia*

To test the dynamic characteristics of cytoophidia, we performed FRAP on four groups of hCTPS1 cytoophidia (Figure 2A,B). The intensity of the FRAP ROI was normalized as *Normalized Intensity* = *<sup>I</sup>*(*non*−*bleach ROI*)*Pre*−*bleach*, 0−*I*(*background*) 0 *I*(*non*−*bleach ROI*)*n*−*I*(*background*)*<sup>n</sup>* × [*I*(*bleach ROI*)*<sup>n</sup>* − *I*(*background*)*<sup>n</sup>* ] [34]. The bleached ROIs on cytoophidia induced by 20 µg/mL for 8 h (low concentration and short time) before imaging recovered very quickly (Figure 2C). However, ROIs in cells with *hCTPS1* overexpression treated with DON in 100 µg/mL (higher concentration) recovered fluorescence slowly and ended at a lower intensity (Figure 2D; Supplemental Figure S1A).

By extending the DON treatment time to 25 h, ROIs were able to recover as quickly as possible and in a relatively short time (Figure 2E; Supplemental Figure S1B). The fluorescence intensity restored by bleaching the ROIs of cytoophidium-like condensates of hCTPS-EGFP cells was very low (Figure 2F; Supplemental Figure S1C). There was a significant difference in dynamics between hCTPS1-EGFP cytoophidium-like condensates and DON-induced hCTPS cytoophidia at low concentrations and over short time periods (Figure 2G).

This shows that DON-induced cytoophidia have very different dynamic characteristics from hCTPS-EGFP cytoophidium-like condensates. DON-induced cytoophidia seem not to be assembled by simple forces, like the hCTPS-EGFP cytoophidium-like condensates.

The bleached ROIs gradually recovered throughout, rather than from any particular side. Neither of the two ROIs moved to either side, nor were they far away or close to each other (Figure 2C). Our results showed that bleached hCTPS1 molecules in cytoophidia could exchange with free hCTPS1 molecules in the cytosol (Figure 2H).

**Figure 1.** Assembly of CTPS filaments into cytoophidia. (**A**) hCTPS1 forms granules in the same cell that coexist with hCTPS1 cytoophidia. The arrowhead points to a granule. (**B**) The trajectory of hCTPS1 granules is a random walk. For the DON treatment, 20 μg/mL DON in PBS solution was added to fresh DMEM medium 8 to 25 h before live-cell imaging. (**C**) hCTPS1-EGFP forms cytoophidium-like condensates by simple force between EGFP. The arrangement from hCTPS1 filaments to hCTPS cytoophidia is the problem that needed to be solved. (**D**) The quantities of transfected over-expressed hCTPS in 293T cells were measured. (**E**,**F**) The transfection efficiencies were quantified. (**G**) Estimated ratio of exogenous hCTPS1 to endogenous hCTPS. (**H**) hCTPS-EGFPA206K cytoophidia can be induced by DON treatment. hCTPS1-EGFP can form cytoophidium-like condensates, which are wider and larger than hCTPS cytoophidia. hCTPS1H355A-EGFP cannot form cytoophidium-like condensates. For the DON treatment, 20 μg/mL DON (PBS solution) was added to fresh DMEM medium 8 h before fixation. Scale bars, 10 μm (**A**,**B**) and 20 μm (**H**). **Figure 1.** Assembly of CTPS filaments into cytoophidia. (**A**) hCTPS1 forms granules in the same cell that coexist with hCTPS1 cytoophidia. The arrowhead points to a granule. (**B**) The trajectory of hCTPS1 granules is a random walk. For the DON treatment, 20 µg/mL DON in PBS solution was added to fresh DMEM medium 8 to 25 h before live-cell imaging. (**C**) hCTPS1-EGFP forms cytoophidium-like condensates by simple force between EGFP. The arrangement from hCTPS1 filaments to hCTPS cytoophidia is the problem that needed to be solved. (**D**) The quantities of transfected over-expressed hCTPS in 293T cells were measured. (**E**,**F**) The transfection efficiencies were quantified. (**G**) Estimated ratio of exogenous hCTPS1 to endogenous hCTPS. (**H**) hCTPS-EGFPA206K cytoophidia can be induced by DON treatment. hCTPS1-EGFP can form cytoophidiumlike condensates, which are wider and larger than hCTPS cytoophidia. hCTPS1H355A-EGFP cannot form cytoophidium-like condensates. For the DON treatment, 20 µg/mL DON (PBS solution) was added to fresh DMEM medium 8 h before fixation. Scale bars, 10 µm (**A**,**B**) and 20 µm (**H**).

**Figure 2.** Dynamic equilibria of cytoophidia. (**A**) ROIs were used for bleaching, measurement and normalization of data. (**B**) Normalized intensity curves of FRAP results in different groups were merged. (**C**) Live-cell images of FRAP in hCTPS1-mCherry cytoophidia induced by 20 μg/mL DON for 8 h. (**D**) Comparison of FRAP curves for hCTPS1-mCherry between 20 μg/mL DON for 8 h and 100 μg/mL DON for 8 h. (**E**) Comparison of FRAP curves for hCTPS1-mCherry with 20 μg/mL DON for 8 h (pink curves) and 20 μg/mL DON for 25 h (black curves). (**F**) Comparison of FRAP curves for hCTPS1-mCherry cytoophidia (pink curves) induced by DON and hCTPS1-EGFP cytoophidium-like condensates (green curves). (**G**) Analysis of the level and speed of fluorescence recovery. \*, *p*-value < 0.05; \*\*, *p*-value < 0.01; ns, no significant difference. (**H**) The model of the structure of compact or condensed hCTPS cytoophidia does not fit the results of the FRAP images. A new model **Figure 2.** Dynamic equilibria of cytoophidia. (**A**) ROIs were used for bleaching, measurement and normalization of data. (**B**) Normalized intensity curves of FRAP results in different groups were merged. (**C**) Live-cell images of FRAP in hCTPS1-mCherry cytoophidia induced by 20 µg/mL DON for 8 h. (**D**) Comparison of FRAP curves for hCTPS1-mCherry between 20 µg/mL DON for 8 h and 100 µg/mL DON for 8 h. (**E**) Comparison of FRAP curves for hCTPS1-mCherry with 20 µg/mL DON for 8 h (pink curves) and 20 µg/mL DON for 25 h (black curves). (**F**) Comparison of FRAP curves for hCTPS1-mCherry cytoophidia (pink curves) induced by DON and hCTPS1-EGFP cytoophidium-like condensates (green curves). (**G**) Analysis of the level and speed of fluorescence recovery. \*, *p*-value < 0.05; \*\*, *p*-value < 0.01; ns, no significant difference. (**H**) The model of the structure of compact or condensed hCTPS cytoophidia does not fit the results of the FRAP images. A new model of structures is needed to explain the recovery from bleached fluorescence in the cytoophidia. The intensity of the FRAP ROI was normalized as *Normalized Intensity* = *<sup>I</sup>*(*non*−*bleach ROI*)*Pre*−*bleach*,0 <sup>−</sup> *<sup>I</sup>*(*background*)<sup>0</sup> *I*(*non*−*bleach ROI*)*<sup>n</sup>* − *I*(*background*)*<sup>n</sup>* × [*I*(*bleach ROI*)*<sup>n</sup>* − *I*(*background*)*<sup>n</sup>* ]. Scale bars, 10 µm (**C**).

#### *2.3. The Reticular Structure of the hCTPS1 Cytoophidium and Its Localization with hIMPDH2*

In order to build a model to fit the dynamic-equilibrium characteristics of cytoophidia, we obtained super-resolution structures of hCTPS cytoophidia using stimulated emission depletion microscopy (STED). The images under conventional confocal microscopy could not show the structure inside cytoophidia, while the STED images revealed the superresolution structure with a resolution of 50 to 70 nm (Figures 3A and S3A,B), which implied a possible mechanism of highly dynamic cytoophidia under FRAP. We estimated the resolution by measuring the distance between two distinguishable nearby particles and it was 50 to 70 nm (Figure S3A,B).

STED revealed a heterogeneous structure for hCTPS cytoophidia. Some parts were relatively more condensed, while other parts were looser. More importantly, it seemed that there were many tiny filaments inside, which in different orientations formed reticular structures (Figure 3A; Supplemental Figure S2A).

However, the super-resolution results were homogeneous inside hCTPS1-EGFP cytoophidium-like condensates (Figure 3A). The structures of the hCTPS cytoophidia and those of the hCTPS1-EGFP cytoophidia were totally different. No condensed and loose parts or reticular structures knitted with tiny filaments could be observed in the cytoophidium-like condensates.

When performing super-resolution imaging, there might be some interfering factors against reliability of the super-resolution imaging results, such as the efficiencies of antibodies, optical properties of the fluorescent labeling, the steric hindrance of fluorescence proteins, and the influence of sample preparation. To eliminate the effects of antibodies and Cy5, we used 293T cells with hCTPS1-miRFP670nano for the live-cell imaging. We performed immunofluorescence staining on DON-treated 293T cells to eliminate the effects of potential steric hindrance of fluorescence protein tags and overexpression. Both results showed reticular structures (Figure 3A). To avoid the difference in optical properties between Cy5 and EGFP, we also performed immunofluorescence staining with Cy5 on hCTPS1-EGFP-overexpressed 293T cells, and the signal obtained was from Cy5 (Figure 3A).

In addition, we performed immunofluorescence staining on SW480 cells cultured in glutamine-free medium, which showed a reticular structure (Figure 3A). Glutamine is an NH<sup>3</sup> donor in metabolic reactions. This means that the reticular structure of hCTPS is not only a phenomenon induced by DON but also a common structure of metabolic enzymes when cells are under metabolic stress. Without changing the super-resolution structural results, deconvolution of the STED images could improve their resolution and signal-to-noise ratios (Figure 3A; Supplemental Figure S2A).

These tiny filaments appeared as subunits of hCTPS cytoophidia (Figure 3A,B). In vitro experiments showed that CTPS can be assembled into filaments [20,25,30]. Based on the in vitro and in vivo results, we envisioned a model to illustrate the reticular structure of hCTPS cytoophidia (Figure 3B). Inside cytoophidia, subunit filaments are weaved into a reticulation. The model can make FRAP results clearer. The dynamic equilibrium of assembly and disassembly occurs in the tiny filaments of hCTPS, rather than the assembly and disassembly of the whole cytoophidium. FRAP procedures performed on untreated and dispersive hCTPS1 signals resulted in fast recovery, which meant that we were unable to capture the images after bleaching, which were similar to the images before bleaching (Supplemental Figure S2B). Due to the limitation of STED resolution, it is unclear whether the subunit filament is one CTPS filament, a bundle of CTPS filaments or some other form of CTPS.

After obtaining the super-resolution reticular structures, we wanted to determine the reason for and function of this reticular structure. There might be some unknown molecules in the space between hCTPS filaments (Figure 3B). It was reported that IMPDH2 interacted with CTPS1 cytoophidia under DON treatment [23]. The cytoophidia of hCTPS1 and hCTPS2 were located together in 293T cells (Supplemental Figure S2C). IMPDH and CTPS are both part of the glutamine and NH<sup>3</sup> metabolic pathways (Supplemental Figure S2D). We overexpressed *hCTPS1-EGFPA206K*, a monomer version of EGFP, and

labeled IMPDH2 with Cy5 by immunofluorescence staining. Under DON treatment, IMPDH2 and hCTPS1 were localized spatially adjacent to each other (Figure 3C). IMPDH2 and hCTPS1 were not exclusive but were positioned mutually in each Z stack (Figure 3D). Therefore, IMPDH2 could be one of the molecules located between hCTPS1 filaments. *Int. J. Mol. Sci.* **2022**, *23*, x FOR PEER REVIEW 8 of 17

**Figure 3.** The reticular structure of the hCTPS1 cytoophidium and its localization with hIMPDH2. (**A**) Confocal STED with deconvolution and zoomed-in images of cytoophidia from the following groups: (1) hCTPS1/2 cytoophidia (Cy5 antibody-stained) induced by DON in fixed 293T cells, (2) hCTPS1-miRFP670nano cytoophidia induced by DON in live 293T cells, (3) hCTPS1/2 (Cy5-stained) cytoophidia induced by glutamine deprivation in fixed SW480 cells and (4) hCTPS1/2 (Cy5-stained)- **Figure 3.** The reticular structure of the hCTPS1 cytoophidium and its localization with hIMPDH2. (**A**) Confocal STED with deconvolution and zoomed-in images of cytoophidia from the following groups: (1) hCTPS1/2 cytoophidia (Cy5 antibody-stained) induced by DON in fixed 293T cells, (2) hCTPS1-miRFP670nano cytoophidia induced by DON in live 293T cells, (3) hCTPS1/2 (Cy5-stained) cytoophidia induced by glutamine deprivation in fixed SW480 cells and (4) hCTPS1/2 (Cy5-stained)- cytoophidium-like condensates in live 293T (hCTPS1-EGFP-overexpression) cells. For the SW480 culture, DMEM without glutamine replaced DMEM 8 h before fixation. For the DON treatment, DON (PBS solution) was added to fresh DMEM medium 8 to 25 h before live-cell imaging or 8 h before being fixed. Scale bars, 3 µm. (**B**) The model of the arrangement of hCTPS filaments into cytoophidia. (**C**) In 293T cells, hCTPS1-EGFPA206K was localized adjacently with DON-induced IMPDH2 (Cy5-stained). (**D**) In each slice along the Z stacks, hCTPS1-EGFPA206K and IMPDH2 (Cy5-stained) were localized to each other. Scale bars, 10 µm.

#### *2.4. CTPS Granules with Tentacles*

We performed live-cell imaging via confocal microscopy in 293T cells overexpressing *hCTPS1-mCherry* and found that hCTPS1 could not only form cytoophidia but also formed DON-induced granules (Figure 1A). The movement of most granules was a random walk, like that of ordinary granules in cells (Figure 1B). Fortunately, when we observed live cells treated with DON, we found a filiform structure connecting hCTPS1 granules (Figure 4A). We name these filiform structures connecting granules "tentacles", and the main part of the granules is called the "granule body". The tentacle slowly extended out of the granule body and retracted as quickly as a rubber band after reaching its longest length (Figure 4B). *Int. J. Mol. Sci.* **2022**, *23*, x FOR PEER REVIEW 10 of 17

**Figure 4.** CTPS granules with tentacles. (**A**) Tentacles connect hCTPS1 granules. Tentacles extend and retract. Arrowheads indicate the tentacle. (**B**) Tentacles extend slowly and retract rapidly after reaching the maximum length. (**C**) hCTPS1 granules move from one side to the other along the tentacles. Yellow arrowheads indicate the tentacles; white arrowheads indicate the granules. (**D**) hCTPS1 granular tentacles have three different behaviors and characteristics, bridging, retraction after extension and the trajectory of hCTPS1 granule movement. For the DON treatment, DON (PBS **Figure 4.** CTPS granules with tentacles. (**A**) Tentacles connect hCTPS1 granules. Tentacles extend and retract. Arrowheads indicate the tentacle. (**B**) Tentacles extend slowly and retract rapidly after reaching the maximum length. (**C**) hCTPS1 granules move from one side to the other along the tentacles. Yellow arrowheads indicate the tentacles; white arrowheads indicate the granules. (**D**) hCTPS1 granular tentacles have three different behaviors and characteristics, bridging, retraction after extension and the trajectory of hCTPS1 granule movement. For the DON treatment, DON (PBS solution) was added to fresh DMEM medium 8 to 25 h before live-cell imaging. Scale bars, 10 µm (**A**,**C**).

We wanted to know the function of the granular tentacles. We found that tentacles were different from small granule bodies. When the tentacle stretched out, the granule body moved from one side of the tentacle to the other side along the tentacle, and then the tentacle retracted into the granule body in the new location (Figure 4C). Granules with tentacles move with clear direction along the tentacles rather than in a random walk. The movement of tentacled granules was different from that of non-tentacled granules. Granular tentacles are tiny structures that bridge granules, move granules and retract after extension (Figure 4D).

#### **3. Discussion**

Taking advantage of multiple fluorescence tags, we study the physical characteristics of hCTPS1-containing compartments via fluorescence microscopy. We perform FRAP and STED analyses to reveal the dynamic and reticular structure of cytoophidia. In addition, we observe that hCTPS1 forms granules with tentacles.

#### *3.1. Cytoophidia Are Not Condensates*

Protein compartments, similar to droplets, can be assembled by physical forces in cells [35]. Since the discovery of CTPS-forming cytoophidia, the exact phases of cytoophidia and the arrangements of CTPS in cytoophidia are still unclear. Cytoophidia are presumed to be static bundles of filaments [4](Liu, 2011) or in a liquid phase, just like LLPS. However, when we performed live-cell imaging on CTPS cytoophidia, we found that CTPS can not only form long filamentous structures, that is, cytoophidia, but also form granules in the same cell (Figure 1A). This means that, as compartments of CTPS, cytoophidia may not be static, concentrated and rigid structures. For another hypothesis, the puzzling question is why this compartment is not a spherical droplet if it is in the liquid phase, such as LLPS.

According to previous studies, the residue 355H of CTPS (CTPS-355H) is the key site for the formation of this filamentous structure. CTPS-355H lying at the tetramer–tetramer interface plays a critical role in CTPS polymerization. In in vitro experiments, the CTPS tetramer assembly mechanism of cytoophidia is more like that of actin filaments than droplets in cells assembled by physical force [36].

To study the role of CTPS-355H, we used both dimeric EGFP and monomeric EGFPA206K tags [32]. We generated *hCTPS1H355A* mutations. hCTPS1-EGFPA206K can form cytoophidia with DON treatment. Without DON treatment, hCTPS1-EGFPA206K cannot form cytoophidia, suggesting that EGFPA206K does not promote CTPS assembly. mCherry and miRFP670nano are also monomeric tags, just like EGFPA206K. However, hCTPS1-EGFP can form filament-shaped condensates without DON treatment (Figure S2E). Since EGFP has a force to form dimer-like "sticky" features, and hCTPS1-EGFP molecules stick to each other in filament-shaped condensates, we refer to these filament-shaped hCTPS1-EGFP structures as "cytoophidium-like condensates" (to be distinguished from the term "cytoophidia") (Figure 1C).

Are cytoophidia just condensates? If they are, the key cytoophidium-forming site CTPS-355H may provide a directional force of assembly, which should be important for filamentous condensate formation. For hCTPS1H355A-EGFP, the force of assembly is provided by dimeric EGFP, since CTPS-355H has been mutated to CTPSH355A. If either CTPS-355H or EGFP can provide force for condensate formation, we would expect that both hCTPS1H355A-EGFP and hCTPS1-EGFPA206K can form condensates. Our results show that hCTPS1H355A-EGFP cannot form cytoophidium-like condensates, suggesting that CTPS-H355 is an essential site of connection rather than just there to provide a directional force (Figure 1H).

Therefore, our data argue against the idea that cytoophidia are condensates. Two factors appear to be required for assembling CTPS into cytoophidia. First, CTPS molecules are brought together by some forces of assembly. Second, CTPS tetramers need to be connected via CTPS-355H.

#### *3.2. Cytoophidia Are Dynamic*

In order to solve the problem of the physical phase of cytoophidia, we carried out FRAP assays to measure the dynamic features. We used hCTPS1H355A-mCherry as the control for complete diffusion, which recovered its intensity quickly after bleaching, so that we could not capture the difference before and after bleaching or measure its dynamic value. We used hCTPS1-EGFP cytoophidium-like condensates as static controls. The intensity recovered in cytoophidia was achieved significantly faster than in hCTPS1-EGFP cytoophidium-like condensates (Figure 2F,G). This means that the cytoophidium is not a condensate but a highly dynamic structure.

A low concentration of DON can induce cytoophidia, mimicking the stress of severe glutamine deprivation. We also measured the effects of time and concentration of DON treatment to determine whether the dynamic results were valid only in particular circumstances. Compared with these curves, the treatment time of DON had little effect on the kinetics (Figure 2E,G), while the concentration of DON had a significant effect on the kinetics of cytoophidia (Figure 2D,G). Due to the covalent bond between DON and CTPS, excessive DON may destroy the conformation of CTPS, thus damaging cells.

#### *3.3. Cytoophidia Are Reticular*

The cytoophidium is on a micrometric scale. A previous assumption was that the cytoophidium might be a bundle of CTPS filaments, similar to actin filaments. When a bleached ROI gradually recovered its intensity, no treadmill phenomenon was observed, which means that the assembly mechanism of cytoophidia is different from that of actin filaments. Moreover, recovery did not come from either side of the bleached ROI. The intensity of the entire ROI recovered steadily at the same speed.

If the cytoophidium is a bundle of CTPS filaments, how can the bleached ROI recover its intensity without a treadmill phenomenon (Figure 2H)? To solve this problem, we need to resolve the fine structure of cytoophidia. We performed super-resolution STED imaging which revealed the reticular structure of cytoophidia. In order to minimize artificial influences on the imaging results, we used miRFP670nano as a fluorescence tag to conduct STED imaging directly on live-cell samples, which gave a clear reticular structure (Figure 3A). We called it a reticular structure because small subunit filaments were interconnected, crossed and woven into the reticulation (Figure 3B).

At present, it is unclear whether the subunit filament is a CTPS filament, a bundle of CTPS filaments or some other form of CTPS. To maximize the resolution of the STED imaging (Figure 3B), all fluorescent tags were infrared-emitting. miRFP670nano is much smaller than other infrared fluorescent proteins, and it can minimize the impact of tag-space obstruction. We performed immunofluorescence staining to confirm the results and avoid the effects of spatial obstruction and overexpression. Cytoophidia with immunofluorescence dye Cy5 on hCTPS also showed the reticular structure.

We also performed immunofluorescence staining on SW480 cells. Cytoophidia in SW480 cells can be induced by glutamine deprivation and present reticular structures, which means reticular cytoophidia are common in different cell types and under glutamine metabolic stress conditions but cannot be treated artificially. Even DON is used on cells to mimic the stress of glutamine metabolic stress.

We used hCTPS1-EGFP as the structurally static control to represent the condensate assembled by physical forces. hCTPS1-EGFP cytoophidium-like condensates do not show the reticular characteristics. The cytoophidium-like condensates appear homogeneous, and their internal parts look the same. There are neither subunit filaments nor can space be observed inside the cytoophidium-like condensates. In summary, the structure of the cytoophidium is reticular, and the arrangements of CTPS assembled on cytoophidia are different from those of condensates or actin filaments.

The cytoophidium reticulation provides a structural basis for the localization of other enzymes, such as IMPDH (Figure 3C,D). Both CTPS and IMPDH are associated with glutamine and NH<sup>3</sup> metabolism. In reticular cytoophidia, there may be dynamic interaction between CTPS, IMPDH and their substrates and the microenvironment. The reticular structure can also provide elasticity for the bending or twisting of cytoophidia (Figure S1B). It was reported that cytoophidia are related to the regulation of IMPDH activity [37].

The reticular structure provides a structural basis for the FRAP results. No treadmill phenomenon has been observed. How can CTPS molecules in a free state replace those composed of the bleached ROIs? Based on the assumption of the reticular structure, the treadmill may occur in the subunit filaments rather than the whole cytoophidium. The assembly and disassembly of CTPS on subunit filaments may have a dynamic equilibrium, thus changing the CTPS molecules after bleaching. This may be a potential explanation for the dynamic equilibrium of large-scale FRAP.

Due to the limitations of STED and confocal devices, we could not achieve superresolution, high-speed live-cell imaging and low phototoxicity for the cells. In order to verify the speculations, more advanced microscope technology is required. Dynamic equilibria of assembly and disassembly in subunit filaments may contribute to the metabolic regulation of reactions in the microenvironment.

In *Caulobacter crescentus,* a small amount of CTPS forms a bundle, while a large amount of CTPS forms a splayed structure [2]. A large amount of hCTPS transforms the morphologies of CTPS bundles into complex structures. Cytoophidia in *Drosophila* female germ cells also exhibit reticular characteristics [1]. Cytoophidia are regulated by the level of molecular crowding in a cell [38]. The formation and maintenance of the reticular cytoophidia may be related to molecular crowding.

#### *3.4. CTPS Can Form Granules with Tentacles*

While using live-cell imaging to capture CTPS-containing structures, we found interesting CTPS granules with tentacles. CTPS granules move in a random-walk mode, but the granules with tentacles move in a clear direction. We assume that these two entities are different forms of compartments with similar shapes. We use the term "tentacle" because the structure slowly stretches out of a granule and quickly retracts, just like the tentacle of an octopus or snail (Figure 4B). The tentacle may be extended to find something that can be connected. The tentacles, like bridges, connect different granules (Figure 4A).

If the granules are very small, the tentacles play a role in directional movement (Figure 4C). The granules extend the tentacles to the maximal length, and the granules move quickly from one side to the other along the tentacles, just like a slingshot. The only function of tentacles we know of is related to the directional movement of granules. We still do not know the function of tentacles as bridges. Moreover, it is unclear whether the granules with tentacles have membranes, their movements being similar to the movements of mitochondria and vesicles from the Golgi.

Interestingly, CTPS granules, tentacled granules and cytoophidia appeared under the same conditions, i.e., treatment with DON. They can even exist in the same cell, with potential for interactions and transformations. The tentacles are also capable of directional movements, just like cytoophidia. However, due to their tiny size, the tentacles may not share the same reticular structure as cytoophidia.

The structure of the tentacle may be similar to the subunit CTPS filament to obtain directional characteristics, or it may be in a liquid state in the membrane vesicle. The intensities of granules with tentacles are far lower than those of cytoophidia, and the tentacles exist in fewer cells than cytoophidia do. Due to their low intensities and small volumes, it is difficult to analyze the properties and fine structures of the tentacles. Compared with cytoophidia reacting to metabolism, it is not clear whether granules with tentacles are related to glutamine metabolism or the role of DON.

#### **4. Materials and Methods**

#### *4.1. Cell Culture*

The 293T and SW480 cells were cultured in DMEM (SH30022.01, Hyclone; Cytiva; 100 Results Way, Marlborough, MA USA 01752) supplemented with 10% FBS (04–001;

Biological Industries; Kibbutz Beit-Haemek, 25115, Israel) in a humidified atmosphere containing 5% CO<sup>2</sup> at 37 ◦C. All the commercial cell lines used in this article were purchased from the Shanghai Institutes for Biological Sciences, Chinese Academy of Sciences (Shanghai, China). They were originally purchased from ATCC. DON was dissolved in PBS and was added to the culture medium as described in individual experiments. DMEM without glutamine (C11960500BT, Gibco; Thermo Fisher Scientific; 168 Third Avenue, Waltham, MA, USA 02451) replaced DMEM 8 h before imaging.

#### *4.2. Constructs and Transfection*

The pLV-hCTPS1-EGFP over-expression vector was kindly provided by Dr. Zhe Sun from ShanghaiTech University. mCherry, miRFP670nano replaced EGFP and EGFP*A206K* was mutated back into EGFP using PCR and a Gibson Assembly System (NEB). Cell transfection was performed with PEI reagent (24765-1, Polysciences; Polysciences, Inc.; 400 Valley Road, Warrington, PA 18976, USA), according to the instructions provided by the manufacturer. The sequences of oligonucleotides used in this study are listed in Supplemental Table S2.

#### *4.3. Immunoblotting*

Cells were harvested in lysis buffer (containing 20 mM Tris, 150 mM NaCl and 1% Triton X-100; P0013J, Beyotime; Beyotime Biotechnology; Building 30, Songjiang Science and Technology Entrepreneurship Center, 1500 Lane, Xinfei Road, Songjiang District, Shanghai, China 201611). Undissolved cell fractions were separated by centrifugation at 12,000 rpm for 10 min at 4 ◦C, and the supernatants were boiled in SDS-PAGE loading buffer for 10 min. Proteins in total cell lysates were separated by SDS-PAGE and transferred to PVDF membranes. Membranes were blocked in 5% nonfat milk and incubated with the appropriate primary antibodies. Protein bands were visualized using horseradish peroxidase (HRP)-conjugated secondary antibodies with ECL reagent (34577, Thermo Fisher Scientific; 168 Third Avenue Waltham, MA, USA 02451).

#### *4.4. Immunofluorescence*

Cells were fixed with 4% paraformaldehyde added into media for 25 min. Then, the fixed cells were washed in 1xPBS 3 times. Samples were incubated with appropriate primary antibodies (rabbit anti-IMPDH2, Proteintech 12948-1-AP; rabbit anti-CTPS, Proteintech 15914-1-AP; Proteintech Group, Inc.; 5500 Pearl Street, Suite 400 Rosemont, IL 60018, USA) overnight at 4 ◦C and washed in PBS 3 times. Samples were incubated with Cy5-conjugated secondary antibodies (donkey anti-rabbit Cy5-conjugated antibody, Jackson 711-175-152; Jackson ImmunoResearch Inc.; 872 West Baltimore Pike, West Grove, PA 19390, USA) at room temperature for 1 h (in the dark) and washed with PBS 3 times after incubation. The mountant used for STED imaging (Figures 3A,C,D and S2A) was ProlongTM Diamond Antifade (Invitrogen, P36965; Thermo Fisher Scientific; 168 Third Avenue, Waltham, MA, USA 02451). The mountant used for confocal imaging (Figure 1H) was HardSet Mounting Medium with DAPI (VECTASHIELD, H-1500; Vector Laboratories, Inc.; 6737 Mowry Ave, Newark, CA 94560, USA)

#### *4.5. Microscopy*

Images (Figures 1A,B, 2C, 4A,C, S1A,B,C and S2B) were acquired under 100× objectives with a confocal microscope (Nikon CSU-W1 SoRa). Confocal images and superresolution images (Figures 3A,C,D and S2A) were acquired under 100× objectives with an STED confocal microscope (Leica TCS SP8 STED 3X). Images (Figure S2C) were acquired under 63× objective with a Lattice SIM microscope (Zeiss Elyra 7) in wide-field mode. Confocal images (Figure 1H) were acquired under 63× objective with a confocal microscope (Zeiss LSM 980 Airyscan2).

#### *4.6. Live Imaging*

The 293T cells transfected with hCTPS1-mCherry and hCTPS1-miRFP670nano constructs were cultured on glass-bottom culture dishes (C8-1.5H-N, Cellvis; Vitro Scientific; Mountain View, CA 94039, USA) with medium and maintained at 37 ◦C when the live imaging was performed.

#### *4.7. Image Analysis*

Fluorescence images was analyzed with the software IMAGEJ (NIH, Bethesda, MD, USA). The ROIs of bleached regions for intensity measurement shown in Figure 2 were selected and measured manually with IMAGEJ. The FRAP curve unpaired *t*-test was analyzed with Graph Prism 8.4.0 (GraphPad Software, LLC; San Diego, CA, USA). The deconvolution of STED images shown in Figure 3 was analyzed by the lighting algorithm of the Leica LAS X. The 3D model shown in Figure 3 was analyzed with the Leica LAS X. The lengths of the tentacles shown in Figure 4 were measured manually with IMAGEJ. Quantity data were collected with Microsoft Excel.

#### *4.8. Fluorescence-Activated Cell Sorting (FACS) Analysis*

The flow cell analyzer used was an LSRFortessa X20 (BD). Flow cell data were analyzed with FlowJo 10.4 (FlowJo, LLC; Ashland, OR, USA) software. Data were collected with Microsoft Excel 16.61 (Microsoft Corporation; Redmond, WA, USA).

#### **5. Conclusions**

To sum up, the main purpose of this study is to understand the structure and arrangement of CTPS in CTPS filaments with near-atomic-resolution and micron-scale cytoophidia observed under confocal microscopy. We use dimeric EGFP tags as controls to provide aggregation viscosity and identified the connecting role of the CTPS-355H site. FRAP analysis shows that the cytoophidium is highly dynamic, while STED analysis reveals the reticular structure of cytoophidia.

According to the comparison with CTPS-EGFP cytoophidium-like condensates, the dynamic and reticular characteristics of cytoophidia are different from those of condensates (Figure 3B). Moreover, we find that the compartments of CTPS not only exist in the snakeshaped cytoophidia but also in the granules. CTPS granules move in different ways depending on whether they have tentacles. To understand the functions of CTPS granules with tentacles, further studies are required.

**Supplementary Materials:** The following supporting information can be downloaded at: https://www.mdpi.com/article/10.3390/ijms231911698/s1.

**Author Contributions:** Y.-F.F. and J.-L.L. conceived the project. Y.-F.F. designed experiments. Y.-F.F., Y.-L.L. and X.-M.L. performed the experiments. Y.-F.F. analyzed the data. Y.-F.F. wrote the original manuscript. Y.-F.F., Y.-L.L., X.-M.L. and J.-L.L. revised the manuscript. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research was funded by the Ministry of Science and Technology of China (grant number 2021YFA0804700), the National Natural Science Foundation of China (grant number 31771490), the Shanghai Science and Technology Commission (grant number 20JC1410500) and the UK Medical Research Council (grant numbers MC\_UU\_12021/3 and MC\_U137788471), and the APC was funded by ShanghaiTech University.

**Institutional Review Board Statement:** Not applicable.

**Informed Consent Statement:** Not applicable.

**Data Availability Statement:** Not applicable.

**Acknowledgments:** We thank the Molecular Imaging Core Facility (MICF) at the School of Life Science and Technology, ShanghaiTech University, for providing technical support. This work was supported by grants from the Ministry of Science and Technology of China (no. 2021YFA0804700), the National Natural Science Foundation of China (no. 31771490), the Shanghai Science and Technology Commission (no. 20JC1410500) and the UK Medical Research Council (nos. MC\_UU\_12021/3 and MC\_U137788471) (to J.L.L).

**Conflicts of Interest:** The authors declare no conflict of interest. The funders had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript; or in the decision to publish the results.

#### **References**

