Paperonium

← Back to the plain-language case

The abstract reverses the open/closed conformational preference reported in the Results and figures

Nat Commun (2017) · doi:10.1038/ncomms15744 · PMC5458555

Summary

The abstract states that the GTPase domain–BSE region of MxA prefers the open conformation in the nucleotide-free, GDP·AlF4−-bound and GDP-bound forms, and that GTP loading shifts the preference to the closed state. The Results define high FRET (~0.72) as closed and low FRET (~0.35) as open. They report the high-FRET (closed) state at ~70% occupancy in those three forms, and a shift toward low FRET (open) with the GTP analogue GMPPCP. The Fig. 3c legend, the Discussion and the Fig. 6 model all use the Results direction, so the abstract is the only place where the direction is reversed. This matters because the abstract carries the paper's headline result: a reader of the abstract alone gets the main smFRET finding, and the direction of the hydrolysis-driven step in the proposed torque model, the wrong way round.

What the paper reports

Abstract:

Whereas the open conformation is preferred in nucleotide-free, GDP·AlF4--bound and GDP-bound forms, loading of GTP activates the relative movement between the two domains and alters the conformational preference to the ‘closed' state.

Results, "Conformational preference for GD-BSE region" (paragraph 1):

Surprisingly, MxAMO2-AS displayed two steadily distinguished FRET states (∼0.35/low and ∼0.72/high, Fig. 3a and Supplementary Fig. 8a). Prior structural studies identified two GD-BSE conformations: MxA/dynamin molecules in apo, GDP·AlF4--bound and GDP-bound forms display a closed conformation, while MxA/dynamin bound with non-hydrolysable GTP analogue guanosine-5′-[(β,γ)-methyleno]triphosphate (GMPPCP) adopts an open conformation where the BSE swings ∼70° against the GD

Thus, the high- and low-FRET states represent the closed and open conformations of stalkless-MxA, respectively. The FRET traces in apo, GDP·AlF4--bound and GDP-bound forms are very similar and the high-FRET state is dominant with an occupancy of ∼70%, whereas loading of GMPPCP shifts the FRET distribution towards the low-FRET state (Fig. 3c).

Figure 3 legend, panel (c):

(c) Fraction variation in the low-FRET open state (red circle) and the high-FRET closed state (green circle) during GTP hydrolysis.

Discussion (paragraph 1):

This in turn releases the intrinsic potential for conformational recovery, thereby producing a mechanistic force that drives the transition from ‘open' to ‘closed' conformation at the GD-BSE region (Figs 3c and 6a).

Figure 6 legend, panel (b):

(I) Loading of GTP promotes domain movement. The GDs of molecule 4 and A associate at open conformation. (II) Stimulated GTP hydrolysis releases the potential of conformational alternation in the GD-BSE region of molecule 4 and A, and the GDs move back to closed conformation, thereby generating latitudinal shear force.

All quotes were checked word for word against the article as archived in PubMed Central (src/show.py PMC5458555 --grep). The archive writes "AlF4-" where the publisher's page has a subscript 4 and a superscript minus.

What the paper's own numbers show

Method. I took the state mapping the Results define (high FRET = closed, low FRET = open) and applied it to the occupancies and shifts the Results report. I then compared the result with the abstract sentence, and with the Discussion and Fig. 6 statements about the GTP-loaded and post-hydrolysis states. As an independent check of the mapping, I converted the two reported FRET efficiencies to dye-to-dye distances with the standard Förster relation E = 1/(1 + (R/R0)^6). The paper does not state R0, so I used 50–60 Å, the range usually quoted for Cy3/Cy5. All numbers below are printed by recompute.py.txt.

Quantity Location Reported Recomputed from the paper's own statements
FRET values of the two GD-BSE states Results; Fig. 3a ~0.35 (low), ~0.72 (high) Förster distances: high FRET 42.7–51.3 Å, low FRET 55.4–66.5 Å (R0 = 50–60 Å). High FRET is the shorter distance, so it is the compact (closed) state
State mapping Results; Fig. 3c legend high FRET = closed, low FRET = open Same; consistent with the Förster check and with the cited crystal structures (apo, GDP·AlF4−, GDP closed; GMPPCP open)
Occupancy in apo, GDP·AlF4−, GDP Results high FRET (closed) ~70% closed 70%, open 30% (closed:open = 2.33:1), so closed is preferred
Preferred state in apo, GDP·AlF4−, GDP Abstract open Results imply closed; the Abstract is reversed
Effect of GTP / GMPPCP loading Abstract preference altered "to the ‘closed' state" Results: shift toward low FRET, i.e. toward open; the Abstract is reversed
Direction of the hydrolysis-driven step Discussion; Fig. 6b (I–II) GTP-loaded GDs open; hydrolysis drives open → closed Agrees with the Results (GTP state open, post-hydrolysis closed); disagrees with the Abstract
Distance gap between the two states Results ~20 Å (crystal structures, residues 149–639) 12.7–15.3 Å from the FRET values (R0 = 50–60 Å). Same direction; the size differs, plausibly because of dye linkers and the use of residue 639 for 634. Not part of the finding

Figure

Abstract and Results state opposite conformational preferences for the same nucleotide conditions

Figure 1. The abstract sentence (left, orange) next to the Results statements for the same conditions (right, blue). Under the mapping the Results define (high FRET ≈ 0.72 = closed, low FRET ≈ 0.35 = open), the high-FRET state at ~70% occupancy means the nucleotide-free, GDP·AlF4− and GDP forms prefer the closed state, and GMPPCP loading shifts the distribution toward open. The abstract states the opposite in both cases. The Discussion and the Fig. 6b model follow the Results. All excerpts are quoted from the paper; no numbers in the figure are new.

Förster relation places the high-FRET state at the shorter distance

Figure 2. A physics check of the Results mapping. The gray curve is the Förster relation E = 1/(1 + (R/R0)^6) for R0 = 54 Å; the shaded band covers R0 = 50–60 Å. The dashed orange lines are the two FRET values reported in the Results. The blue bars are the distances recomputed from those values across the R0 band (dots at R0 = 54 Å). The high-FRET state lies at about 43–51 Å and the low-FRET state at about 55–67 Å. The high-FRET state is therefore the more compact one, which supports the Results' "closed" label and gives no support to the abstract's reversed wording.

Why it matters

The abstract is the only part of the paper most readers see, and it is what indexing services and citing papers pick up. Here it states the direction of the paper's main smFRET result the wrong way round on two counts: which conformation dominates without GTP, and which way GTP loading moves it. Read literally, it also implies that hydrolysis drives a closed-to-open change, the opposite of the open-to-closed step on which the Discussion builds the shear-force and torque model (Fig. 6). It also suggests that the full-length smFRET data disagree with the earlier stalkless-MxA and dynamin crystal structures, when the Results say they agree ("These results demonstrate that the GMPPCP-dependent BSE rearrangement observed in stalkless-MxA structures also occurs in full-length MxA").

The measurements, figures and model are consistent with one another, so the paper's actual conclusion is not affected. The rest of the abstract sentence, that GTP loading "activates the relative movement between the two domains", agrees with the body (Fig. 3g: "enhanced dynamics for GMPPCP-loading condition"). The error is confined to the two words "open" and "closed" in one sentence of the abstract.

Explanations considered

None of these explains the abstract sentence. It is an internal inconsistency in the text; the data and the body of the paper agree with each other.

Status

Suggested correction

Swap the two conformation labels in the abstract sentence so that it matches the Results, Fig. 3c, the Discussion and Fig. 6:

Whereas the closed conformation is preferred in nucleotide-free, GDP·AlF4−-bound and GDP-bound forms, loading of GTP activates the relative movement between the two domains and alters the conformational preference to the ‘open’ state.

No change to the data, figures, Results or Discussion appears to be needed.