Cdc42 couples septin recruitment to the axial landmark assembly via Axl2 in budding yeast

Cdc42 plays a central role in polarity establishment in yeast and animals. In budding yeast, Cdc42 is involved in the selection of a bud site and polarized organization of the actin and septin cytoskeletons (Bi and Park, 2012). Septins are recruited to an incipient bud site and then converted into a ring at the mother–bud neck after bud emergence (Iwase et al., 2006; Okada et al., 2013), and the septin organization depends on Cdc42 polarization (Caviston et al., 2003; Gladfelter et al., 2002; Iwase et al., 2006). Early studies suggested a linear morphogenetic hierarchy from spatial landmarks to the cytoskeletons via the Rsr1 and Cdc42 GTPase modules (Chant and Pringle, 1995; Kang et al., 2001; Park et al., 1997). However, temporal interactions among the landmark components, Cdc42 and septins suggest that there is more complex crosstalk between the proteins involved in polarity establishment. The septin ring recruits the axial landmark components Bud3 and Bud4 (Chant et al., 1995; Sanders and Herskowitz, 1996), and Bud3 and Bud4, in turn, are necessary for septin ring integrity during cytokinesis (Eluère et al., 2012; Kang et al., 2013; Wloka et al., 2011) and play distinct roles during the transition of septin filaments from an hourglass to a double ring (Chen et al., 2020). Bud3 also plays a regulatory role in Cdc42 polarization as a GDP-GTP exchange factor (GEF) for Cdc42 (Kang et al., 2014).

How are septins recruited to an incipient bud site? Cdc42 polarization occurs stepwise upon its activation by its GEFs, Bud3 and Cdc24, during the two temporal steps (T1 and T2) in the G1 phase (Kang et al., 2014). This biphasic Cdc42 polarization is likely linked to the recruitment and assembly of a septin ring. Notably, Cdc42 polarization during T1 is necessary for the axial landmark assembly and septin recruitment in a and α cells (Kang et al., 2018). Gic1 and Gic2, two related Cdc42 effectors, are involved in septin recruitment (Iwase et al., 2006; Sadian et al., 2013). However, overexpression of CDC42 suppresses the lethality of cells lacking Rsr1 and both Gics, and rescues the defects in septin recruitment (Kang et al., 2018), suggesting that Cdc42 can recruit septins directly or indirectly via a protein other than the Gic proteins.

Previous genetic studies have suggested that Axl2 is involved in septin organization. AXL2 was identified as a dosage suppressor of the lethality of the spa2Δ cdc10-10 mutant defective in Cdc10 and Spa2, a polarisome component (Roemer et al., 1996). Overexpression of AXL2 also suppresses temperature-sensitive (ts) growth of gic1 gic2 (Gandhi et al., 2006) and a cdc42 mutant defective in the septin ring assembly (Gao et al., 2007). Although it has been reported that Axl2 interacts with Cdc42-GDP and Bud4 even in the absence of Bud3 (Gao et al., 2007), we found that Axl2 fails to associate with Bud4 in cells lacking Bud3 but expressing these proteins involved in bud-site selection at endogenous levels (Kang et al., 2012). This apparent inconsistency and other outstanding questions led us to investigate further the function of Axl2 in axial budding. Here, by combining live-cell imaging with genetic analyses, we found that Axl2 interacts with Bud3, Cdc42-GTP and Cdc10. We also identified a cdc42 mutation that disrupts its interaction with Axl2 and causes compromised septin recruitment. Our findings suggest that Axl2 promotes efficient septin recruitment to the axial bud site and that Axl2 may also function in septin recruitment in non-axially budding cells.

To test whether Axl2 interacts with Bud3, we performed pull-down assays using Axl2–TAP in wild-type (WT) or mutant strains expressing Bud3–Myc or Bud3ΔN–Myc (which lacks the N-terminal 259 residues). Bud3–Myc, but not Bud3ΔN–Myc, was associated with Axl2–TAP (Fig. 1A). This Axl2–Bud3 association was diminished by ∼50% in an axl1 mutant and almost completely abolished in a bud4Δ mutant. Given that Axl2 fails to interact with Bud4 in bud3Δ cells (Kang et al., 2012), these results indicate that Axl2 associates with the Bud3–Bud4 complex and support the idea of the stepwise assembly of the axial landmark: first, the interaction between Bud4 and Bud3, and then the association of Axl2 and Axl1 with the Bud4–Bud3 dimer.

Is such ordered assembly of the axial landmark functionally significant? Given that Bud3 is a Cdc42 GEF (Kang et al., 2014), we considered two possible scenarios. First, Bud3 might become an active Cdc42 GEF upon its interaction with Axl2. If so, Axl2 could associate with Cdc42-GDP in a manner mediated by Bud3. Second, upon activation by Bud3, Cdc42-GTP might interact with Axl2 and bring it for the axial landmark assembly. To distinguish these, we tested the interaction between Axl2 and Cdc42 using a yeast two-hybrid assay. Expression of the LEU2 reporter indicated that Axl2 interacts with WT Cdc42 and Cdc42^G12V, which has a GTP-locked state in vivo. In contrast, Axl2 failed to interact with Cdc42^G15A and Cdc42^D118A, which are likely in a GDP-locked or nucleotide-free state (Daubon et al., 2011; Davis et al., 1998) (Fig. 1B). These results suggest that Axl2 interacts with Cdc42-GTP. These findings also imply that Axl2 may not be necessary for Bud3 GEF activity and that Bud3 is unlikely to mediate the Axl2–Cdc42 interaction. Indeed, when we examined the interaction between Axl2 and Cdc42 in a bud3Δ mutant by a two-hybrid assay (Fig. 1B) and Axl2–TAP pulldown assays (Fig. S1), Axl2 interacted with Cdc42 similarly in WT and bud3Δ strains. Collectively, these data suggest that Axl2 interacts with Bud3 and with Cdc42-GTP independently of Bud3.

To determine the functional significance of the interaction between Axl2 and Cdc42-GTP, we looked for a cdc42 allele that encodes Cdc42 that has a defective interaction with Axl2. Among the cdc42^ts alleles that arrest in G1 at 37°C (Kozminski et al., 2000), cdc42-101^K5A is specifically defective in the axial budding pattern at 30°C, a semi-permissive temperature (Kang et al., 2014). Interestingly, Cdc42^K5A interacted normally with Bud3 but poorly with Axl2 in a two-hybrid assay (Fig. 1C). We then examined how this cdc42 mutation impacts the localization of Axl2 at 30°C by time-lapse imaging. In WT cells, Axl2–GFP localized to the bud neck, forming a double ring, and in a patch next to the ring soon after cell division (93%, n=36). In cdc42-101 cells, however, Axl2–GFP frequently appeared as a patch at the pole distal to the cell division site (60%, n=48; Fig. 1D). Despite the presence of all axial landmark proteins, this single residue substitution in Cdc42 seems to cause its poor interaction with Axl2, resulting in inefficient recruitment of Axl2 to the axial bud site. This also leads to the poor association of Axl1 with Bud4 in cdc42-101, given that the assembly of Axl2 with Bud3–Bud4 is necessary for the subsequent interaction of Axl1 with the complex (Kang et al., 2012, 2014). These findings suggest that Cdc42-GTP promotes the assembly of the axial landmark by bringing Axl2 near the cell division site.

As axl2Δ mutants are viable, the lethality of cdc42-101^K5A at 37°C suggests additional defects that might be related to (or distinct from) its poor interaction with Axl2. Interestingly, temperature-sensitive growth of cdc42-101 was suppressed uniquely by overexpression of Gic1 among Cdc42 effectors (Kozminski et al., 2000). Based on this genetic interaction and the role of Gic1 in septin recruitment (Iwase et al., 2006; Sadian et al., 2013), we postulated that cdc42-101 might be defective in septin recruitment. To test this idea, we examined the localization of septin Cdc3–GFP in cdc42-101 cells by time-lapse imaging after temperature upshift to 37°C. Septin recruitment occurred within 16 min after cytokinesis in WT mother cells (90%; n=22). In contrast, less than 40% of the cdc42-101 mother cells (n=24) showed septin recruitment at 1 h after cytokinesis, and even in those cells, new septins often failed to develop into a ring at 37°C (Fig. 2A; Fig. S2). We next examined the localization of Cdc3–GFP at 30°C, together with PBD (p21-binding domain)–RFP, a marker for Cdc42-GTP (Tong et al., 2007). New septins appeared next to the old septin ring at the division site in WT cells (95%, n=60) but frequently at the pole distal to the division site in cdc42-101 cells (59%, n=68) (Fig. 2B). Collectively, these observations suggest that Cdc42^K5A, which poorly interacts with Axl2, is defective in the recruitment of septins to the axial bud site at 30°C and more severely defective in septin recruitment at 37°C or above.

In both WT and cdc42-101 cells, new septins started to appear soon after the development of the first wave of Cdc42-GTP polarization in G1 at 30°C. We quantified the intensity of the PBD–RFP cluster from the time-lapse images and compared the peak intensities in two time windows: (1) 15 min before cytokinesis, and (2) from the onset of cytokinesis to the first appearance of new septins (see Materials and Methods). The PBD–RFP level was similarly high in WT and cdc42-101 cells before cytokinesis, indicating that Cdc42^K5A is not defective in its interaction with Cdc42 effectors (that contain the PBD) or with its GEFs. However, the PBD–RFP cluster intensity level in early G1 was lower in cdc42-101 than in WT (Fig. 2C), likely because of the compromised positive feedback loop of Cdc42 polarization due to inefficient coupling of the axial landmark (lacking Axl2) to the Rsr1 module (Kang et al., 2014; Lee et al., 2015). This defect is likely to result in inefficient septin recruitment to an axial bud site at 30°C.

Given that Axl2 interacts with Cdc42-GTP but poorly with Cdc42^K5A, which is compromised in septin recruitment, we hypothesized that Cdc42-GTP promotes septin recruitment via Axl2. To test this idea, we determined whether Axl2 interacts with any septin subunit. A two-hybrid assay suggested that Axl2 interacts with Cdc10 among the septin subunits Cdc3, Cdc10, Cdc11 and Cdc12 (Fig. 3A). We next examined the Axl2–Cdc10 interaction by a bimolecular fluorescence complementation (BiFC) assay. Expression of the N- and C-terminal fragments of Venus (V_N and V_C) fused to Axl2 and Cdc10, respectively, resulted in YFP signals at the mother–bud neck and the division site (Fig. 3B). In contrast, co-expression of Axl2–V_N and Cdc11–V_C (or Cdc12–V_C) did not show any YFP signals. Unlike other septin subunits, Cdc10 does not have a C-terminal extension (CTE). The proximity of V_C (at the C terminus of septins) to V_N might potentially result in more efficient recovery of YFP with Cdc10–V_C compared to other septin–V_C fusions even when Axl2–V_N interacted with a common domain in septin subunits. To test this possibility, we expressed Cdc11ΔC–V_C, which lacks its CTE [amino acids (aa) 357–415]. Expression of cdc11ΔC-V_C (as a sole copy of CDC11) did not cause any growth defect, consistent with a previous report that the Cdc11 CTE is dispensable for the septin assembly (Versele et al., 2004). Notably, co-expression of Cdc11ΔC–V_C and Axl2–V_N also did not recover any YFP signal (Fig. 3B). Collectively, these data suggest that Axl2 interacts closely with Cdc10.

What is the functional significance of the Axl2–Cdc10 interaction? The Cdc10 dimer is the core of the septin protofilaments and associates with one of two trimers, containing either Cdc11 or Shs1, to form an octamer (Weems and McMurray, 2017). We hypothesized that Axl2 might recruit septins by interacting with Cdc10 at the incipient bud site where Cdc42 is polarized. To test this idea, we first examined how Axl2 affects the localization of septins. Cdc10–GFP, which displayed no noticeable defects in WT, frequently mislocalized in axl2Δ cells, appearing diffused in the cytoplasm or at the bud tip (Fig. 4A). In contrast, Cdc3–GFP or Cdc11–tdTomato localized similarly in WT and axl2Δ cells (Fig. S3A,B). These observations suggest that Axl2 interacts mainly with Cdc10, although GFP tagging of Cdc10 might have a minor impact on its function (see below).

Next, we performed BiFC assays to test whether Axl2 affects the interaction between Cdc10 and Cdc3. Expression of Cdc3–V_N and Cdc10–V_C showed YFP signals at the bud neck and cell division site in WT cells but additionally at the bud tip in axl2Δ cells (Fig. 4B). The YFP signals at the division site in axl2Δ cells often appeared as a less tightly organized ring (see Materials and Methods). These observations suggested that Axl2 functions in the efficient recruitment of septins but also raised a question of whether some molecules of Cdc3 mislocalize in axl2Δ cells at least transiently. Indeed, Cdc3–mCherry mislocalized to the bud tip in some axl2Δ cells expressing Cdc10–GFP, even though no such mislocalization of Cdc3–GFP was observed in axl2Δ cells (carrying untagged CDC10) (see Fig. S3A). GFP (or V_C) tagging of Cdc10 is thus likely to cause a subtle defect that is exacerbated in the absence of Axl2, as also indicated by a small percentage of these axl2Δ cells having abnormal cell shapes (see Fig. 4A,B). We also examined the homotypic interaction of Cdc10 by BiFC assays using a diploid strain expressing Cdc10–V_N and Cdc10–V_C. YFP signals appeared at the bud tip and also diffused at the division site in some axl2Δ cells (Fig. S4). Collectively, these results suggest that Axl2 is involved in the efficient recruitment of septins to the cell division site in both haploid and diploid cells. Although this is consistent with similar localization patterns of Axl2 to the bud neck and the division site in all cell types, it is yet to be determined whether Axl2 plays any role in septin organization in diploid cells.

Our findings in this study suggest that Axl2 couples septin recruitment to the axial landmark assembly by interacting with Bud3, Cdc42-GTP and Cdc10. Yet several questions remain, including whether these interactions are direct or whether these interactions occur sequentially or simultaneously. It also remains an open question whether Axl2 interacts with Cdc10 as a monomer, dimer or within septin protofilaments. Axl2 is a transmembrane protein, whose glycosylation is critical for its function (Powers and Barlowe, 1998; Roemer et al., 1996; Sanders et al., 1999), but direct evidence is limited regarding when and where Axl2 interacts with Cdc42 or Cdc10.

Despite these limitations, we note that Axl2 shares similar binding partners with Gic1, a yeast analog of the mammalian Borg proteins involved in the septin organization (Farrugia and Calvo, 2016; Iwase et al., 2006; Sadian et al., 2013). Gic1 binds to Cdc42-GTP and Cdc10 and scaffolds septin filaments into long, flexible cables, whereas Cdc42-GDP binds to Cdc10 and dissociates septin filaments (Sadian et al., 2013). Axl2 may share a partially redundant role with Gic1 in septin recruitment, although the mechanisms of their actions are likely different (Fig. 4C). In a or α cells, this role of Axl2 is likely linked to the axial landmark via its interaction with the Bud3–Bud4 complex. After the joining of Axl2 and Axl1, this complex becomes the functional axial landmark that interacts with the Rsr1 GTPase module (Kang et al., 2012). Given that Axl2 can interact with Cdc42 in bud3Δ cells (this study), Axl2 may also function in septin recruitment independently of the axial landmark, as previously suggested (Gao et al., 2007). However, this is also likely dependent on interaction between Axl2 and Cdc42-GTP, which is polarized by a default mechanism in the absence of a spatial cue (Kang et al., 2018). Remarkably, a cdc42 mutant that has a defective interaction with Axl2 displays abnormal septin recruitment (this study) and is suppressed uniquely by overexpression of Gic1 among Cdc42 effectors (Kozminski et al., 2000). These data further support the functional link between Axl2 and Gic1. We thus propose that Cdc42 orchestrates both axial landmark assembly and septin recruitment via Axl2 in haploid cells. Further studies are required to fully understand these interactions and to delineate the mechanisms of septin recruitment to the incipient bud site.

Standard methods of yeast genetics, DNA manipulation and growth conditions were used (Guthrie and Fink, 1991). Yeast strains used for imaging express tagged proteins under their native promoters from the chromosomes. Yeast strains were grown in rich yeast medium (YPD; yeast extract, peptone and dextrose) or synthetic complete (SC) containing 2% dextrose as a carbon source unless stated otherwise. The strains and plasmids used in this study are listed in Tables S1 and S2, and brief descriptions of construction methods are provided below the Tables.

Cells were grown in SC medium (with 2% dextrose) overnight and then freshly subcultured for 3–4 h before mounting on a slab containing the same medium and 2% agarose. Time-lapse imaging was performed essentially as previously described (Kang et al., 2014; Lee et al., 2015) using a spinning-disk confocal microscope (Ultra-VIEW VoX CSU-X1 system; PerkinElmer) equipped with a 100×1.45 NA Plan Apochromat objective lens (Nikon), 440-, 488-, 515- and 561-nm solid-state lasers (Modular Laser System 2.0; PerkinElmer), and a back-thinned electron-multiplying charge-coupled device (CCD) camera (ImagEM C9100-13; Hamamatsu Photonics) on an inverted microscope (Ti-E; Nikon). Images were captured (nine Z-stacks, 0.3 µm Z-steps) every 3 or 4 min at 30°C or after shifting to 37°C (as indicated), and maximum intensity projections of Z-stacks were used to make Figs. 1D, 2 and Fig. S2. Throughout the paper, numbers in time-lapse images indicate time (in min) from the onset of cytokinesis (t=0), which was identified based on the splitting of the septin ring, and scale bars indicate 3 µm.

To analyze Cdc42 polarization, the PBD–RFP clusters were quantified by a threshold method using an ImageJ (National Institutes of Health) macro, as previously described (Okada et al., 2013, 2017). Briefly, average intensity projection images were generated from five best-focused Z stacks, and then a threshold method was used after background subtraction to quantify the PBD–RFP clusters in mother cells at each time point. The peak PBD–RFP level was identified in each period: (1) from t=−15 min to t=0 (the onset of cytokinesis) or (2) from the onset of cytokinesis until the appearance of new septins. These peak values were plotted after normalizing against the average of the WT peak values during the second time window (i.e. early G1) in Fig. 2C.

To compare the localization of septin subunits in WT and axl2Δ cells, cells were grown as described above. Slides were prepared on an agarose slab as described above, and static images were captured under the same conditions (13 Z-stacks, 0.3 µm Z-steps) at ∼24°C using an inverted widefield fluorescence microscope (Ti-E; Nikon) equipped with a 100×1.45 NA Plan Apochromat Lambda oil immersion objective lens, YFP, FITC/GFP, and mCherry/TexasRed filters from Chroma Technology, DIC optics, an EM CCD (Andor iXon Ultra 888) (Andor Technology) and the software Nis elements (Nikon). Maximum intensity projections of Z-stacks (without deconvolution) were used to make Fig. 4A and Fig. S3.

BiFC assay is based on the fluorescence recovery by the interaction of two proteins, each of which is fused to the N- or C-terminal fragment of Venus (V_N or V_C) (Kerppola, 2006). Yeast strains expressing a combination of test proteins fused to V_N or V_C were grown, and slides were prepared with an agarose slab (as described above). Images were captured (five Z-stacks, 0.4 µm Z-steps) at ∼24°C using an inverted widefield fluorescence microscope (Ti-E; Nikon) and the YFP filter (see above). Localization patterns of BiFC signals were analyzed by applying the same threshold to all images (to highlight the localized YFP signals) after subtracting background using the ‘rolling ball’ method in ImageJ or Fiji software. Abnormal patterns of BiFC signals (such as a less tightly organized ring) in axl2Δ cells were more clearly identified by highlighting each pixel above a threshold, and the same threshold was applied to compare WT and axl2Δ cells. These patterns were analyzed in unbudded cells or cells with large buds from three independent image sets, excluding a small percentage of axl2Δ cells that had abnormal shapes or a cytokinesis defect. To make figures, YFP images were deconvolved by the Iterative Constrained Richardson–Lucy algorithm (Nis Elements), and a single best-focused Z-slice was overlaid with DIC (Figs. 3B, 4B; Fig. S4).

All tagged proteins were expressed in yeast strains at their chromosomal loci (see Table S1), except GST–Cdc42. These strains were grown to the mid-log phase [optical density at 600 nm (OD₆₀₀)∼1.0] in YPD at 30°C. GST or GST–Cdc42 was expressed from a plasmid (pRD56 or pEGKT-CDC42) (Gao et al., 2007; Park et al., 1993), and strains carrying these plasmids were initially grown in SC (with 2% sucrose) lacking Ura and then for additional 5 h after adding 2% galactose. TAP pull-down assays were carried out at 4°C, as previously described (Kang et al., 2014). Briefly, ∼100 OD₆₀₀ units of cells were used to prepare cell lysates using a lysis buffer (50 mM HEPES pH 7.6, 300 mM KCl, 1 mM EGTA, 1 mM MgCl_2, 10% glycerol and 1% Triton X-100) with a cocktail of protease inhibitors. The crude cell lysates were centrifuged for 10 min at 10,000 g, and the supernatant (S10 fraction) was used for subsequent assays. For TAP pull-down assays, the S10 fraction was incubated with 25 µl of IgG-Sepharose (Cytiva, Marlborough, MA, USA) for 1 h at 4°C with rocking. After washing the beads with the same lysis buffer, proteins were eluted from the beads and then subjected to immunoblotting. To determine the Cdc42–Axl2 interaction by TAP pull-down assays, cell lysates were prepared from ∼100 OD₆₀₀ units of cells expressing GST or GST-CDC42 using a lysis buffer (50 mM HEPES pH 7.6, 100 mM KCl, 1 mM EGTA, 1 mM MgCl_2, 10% glycerol and 0.33% Triton X-100) with a cocktail of protease inhibitors. These cell lysates were then incubated with TAP–Axl2 bound IgG-Sepharose beads for 1 h at 4°C by rocking, followed by subsequent steps of washing, elution and immunoblotting. About 1% of each extract was loaded for input blots, and 30% of each pull-down fraction was loaded for pull-down blots. Myc-, GST- and TAP-tagged proteins were detected using anti-Myc antibody 9E10 (a gift of J. Michael Bishop, University of California-San Francisco, CA, USA; used at a dilution of 1:1000), rabbit anti-GST antibody (Santa Cruz Biotechnology, cat. #sc-459, Santa Cruz, CA, USA; used at a dilution of 1:500), and rabbit monoclonal anti-calmodulin binding protein antibody (Upstate Cell Signaling Solutions, cat. #05-932, Temecula, CA, USA; used at a dilution of 1: 5000), respectively. Protein bands were then detected with Alexa Fluor 680 goat anti-rabbit IgG (Invitrogen, cat. #A32734; used at a dilution of 1:10,000) or IRDye^R 800CW conjugated goat anti-mouse-IgG secondary antibodies (LI-COR Biosciences, cat# 926-32210, Lincoln, Nebraska; used at a dilution of 1:10,000) using the LI-COR Odyssey system (LI-COR Biosciences, Lincoln, Nebraska). These assays were repeated with two independent protein preparations, and the relative recovery of Bud3 compared to the input in each TAP pulldown assay was normalized against the same from the WT BUD3-Myc strain (Fig. 1A). The average percentage of GST–Cdc42 pulled down relative to input was shown from two independent protein preparations (Fig. S1). A set of the original immunoblots is shown in the blot transparency in Fig. S5.

Yeast two-hybrid assays were performed as previously described (Kang et al., 2014) by expressing activation domain (AD) fusions using pJG4-5 and DNA-binding domain (DBD) fusions using pEG202 (Gyuris et al., 1993). Two-hybrid assays were performed by patching three independent transformants of the host strain – WT (EGY48) or bud3Δ (HPY3657) – with a combination of AD and DBD plasmids on SGal plates lacking His and Trp (+Leu) and SGal plates lacking His, Trp and Leu (−Leu). Expression of the LEU2 reporter was monitored based on the growth on SGal −Leu for 3–4 days at 30°C. The cytoplasmic domain (C) of Axl2 (aa 529–823) and Bud3 (aa 1–656), which carries the GEF domain, were fused to AD. Plasmids and strains used in two-hybrid assays are listed in Tables S1 and S2.

Data analysis was performed using Prism 8 (GraphPad Software). Error bars in bar graphs (in Fig. 4 and Fig. S4) indicate the s.e.m. In the box graph (Fig. 2C), quartiles and median values are shown together with the mean (marked with ×). The Tukey method was used to create the whiskers, which indicate variability outside the upper and lower quartiles. Any point outside those whiskers (indicated as dots) is considered an outlier. A two-tailed unpaired Student's t-test was performed to determine statistical differences between two sets of data. Data is denoted as *P<0.05, **P<0.01 and ns (not significant) for P≥0.05.

We thank K. Kozminski, M. Peter, M. A. McMurray, E. Bi, and M. Farkasovsky for strains and plasmids.

The peer review history is available online at https://journals.biologists.com/jcs/lookup/doi/10.1242/jcs.261080.reviewer-comments.pdf