Analysis of the pressure requirements for silk spinning reveals a pultrusion dominated process

Silks are remarkable materials with desirable mechanical properties, yet the fine details of natural production remain elusive and subsequently inaccessible to biomimetic strategies. Improved knowledge of the natural processes could therefore unlock development of a host of bio inspired fibre spinning systems. Here, we use the Chinese silkworm Bombyx mori to review the pressure requirements for natural spinning and discuss the limits of a biological extrusion domain. This provides a target for finite element analysis of the flow of silk proteins, with the aim of bringing the simulated and natural domains into closer alignment. Supported by two parallel routes of experimental validation, our results indicate that natural spinning is achieved, not by extruding the feedstock, but by the pulling of nascent silk fibres. This helps unravel the oft-debated question of whether silk is pushed or pulled from the animal, and provides impetus to the development of pultrusion-based biomimetic spinning devices.


Reviewer #1
Comment: |This is a very well written and detailed study on the fibre spinning of silk fibroin using a computational model and I applaud the authors for taking a well thought approach in the development of the COMSOL model. Silk fibre spinning is still challenging and biomimetic devices to date haven't been able to produce fibres that are comparable to native fibres. The development of the model is based on several earlier studies that provide the basis for the models. The COMSOL model provides fodder for development of artificial silk spinning devices. As such this is only a model and would require experimental validation.
Technically I have no issues with the paper but would suggest that either the authors revise this manuscript with experimental data to validate their predictive models or submit to a more specialized journal. Experimental validation is critical for impact of this study and as such I am sceptical till experiments are carried out.
We thank the reviewer for their kind comments. With regard to their request to see experimental validation to widen the impact, we have now provided further references to several previous studies in our discussion to better support our findings. See line 82 (Knight & Vollrath 1999;Asakura et al. 2007;Davies et al. 2013b;Davies et al. 2013a), and lines 105-107 (Kataoka & Uematsu 1977;Terry et al. 2004;Moriya et al. 2009;Ochi et al. 2002;Holland et al. 2012;Holland et al. 2006;Laity et al. 2015;Laity & Holland 2016;Moriya et al. 2008;Kojic et al. 2006;Boulet-Audet et al. 2014), and

Reviewer #2:
Comment: |The authors have used finite element simulations to explore the role of internal pressure, spinning duct geometry, zero shear viscosity and wall friction on the flow rate of the generated fibre. From their results they conclude that silk being pulled out of the duct plays a significant role in the mechanism of silk formation. The paper is well written and easy to read with a comprehensive list of references.
Response: |We thank the reviewer for their positive appraisal of our work. Response: |L273 -we apologise for the lack of clarity, and have re-worded this statement to discuss the likelihood (or lack thereof) of peristalsis being a method for transportation within the gland. See Lines 273-282 and thus the peristaltic effect would need to be transmitted through the haemocoel onto the entire gland at once from the external musculature of the body. Although this may be appropriate for a simple, linear gland, it seems unlikely that such a contraction could create peristaltic flow, due to the twisted, folded nature of the gland itself, with pressure applied at the haemocoel unable to be directed at specific sections of the gland due to the unsegmented nature of the body cavity. Another common suggestion is that continuous production of silk proteins in the rear of the gland creates a concentration gradient which drives flow through osmosis. However, this system would require production throughout the spinning process, which is at odds with the single spinning event that occurs postproduction in B. mori, and suggests that this is not an appropriate mechanism for driving flow. This is substantiated by the knowledge that fibroin concentration increases in the flow direction (Percot et al. 2014), in direct contradiction to the principles of osmotic flow. Comment: |I have made some, mostly minor, comments directly on the manuscript pdf.
Response: |L8 -we have changed explore to review.
Response: |Lines 121-123: the lack of dots/dashes in the figures is regrettably our mistake and an artefact of the PDF conversion process. These have been updated accordingly.
Response: |L134 -use of word models: We agree that this was poorly phrased and have altered this to be clearer.

Now lines 133-136:
The simplified form of the CY-model allows the viscosity to be described in terms of its zero shear viscosity, (readily estimated from low shear measurements, and the rate of shear thinning.) which seems appropriate, given that the primary difference between individual specimens lies in the reported values of the zero-shear viscosity Response: |Lines 214 to 233 -Unclear labelling in figure 6 and 7. We have removed the superfluous arrows, replacing them with clear labels and provided further details in the caption. See lines 217-219 was earlier shown to have little effect on the pressure requirements, this study employed the same geometry as Breslauer (Breslauer et al. 2009) to allow comparison with previously published work. The biomimetic spinning domain is split into natural (dark green) and forced (light green) spinning domains. Fig. 1) brings the lower portion of Laity's ranges (green, mean-1722 Pa.s (Laity et al. 2015;Laity & Holland 2016), Orange, 400-5000 Pa.s (Laity & Holland 2016), Blue 200-10000 Pa.s (Laity & Holland 2016

Reviewer #3
Comment: |The paper is well written and based on the authors' approach. It seems that spinning is achieved by pulling as opposed to pushing.
Response: |We would like to take the opportunity to thanks the reviewer for their detailed feedback.
Comment: |However, the paper could be improved if two topics were discussed in more detail: If both mechanismspush and pulling-actuated simultaneously, will there be any difference on the pushing pressure?
Response: | This is a really interesting point which we are glad to have been given the opportunity to raise in the revised manuscript. As a result we have added the following paragraphs and references to Lines 294-298 in the manuscript.
Of course, these effects are not mutually exclusive, and thus several could be in use within the system. This means that, although we cannot conclusively write off pushing in the system, we can state with certainty that it cannot be considered as the dominant force acting within the system, and that instead, our data shows the maximum rates achievable through pulling alone, hence the difference between this and the natural system represents a minimum force requirement for pultrusion. Comment: |More information on how the silkworm is pulling the fibre would be welcome.
Response: |We are happy to do so. We have now added the following text to Lines 290-292 in the manuscript.
allowing the fibres to be pulled by the worm from itself against their anchorage (either the cocoon site or internal cocoon wall, depending on the extent of cocoon construction).
Comment: |Additional information on the boundaries of the biomimetic spinning domain would also be welcome, particularly on available realistic spinning speeds.
Response: |We apologise for this omission and have now included values and references in text for spinning speeds to supplement those already listed for pressures.  Response to reviewers' comments:

Reviewer #1 (Remarks to the Author):
Comment: | I would prefer to see experimental validation rather than references to support the model.
We thank the reviewer for their consideration of our manuscript. Whilst we originally believed that reference to previous studies were sufficient to support our simulation, upon reflection and discussion we are genuinely grateful for the reviewer's comment, as it has led us towards what we feel is a more rigorous approach to the topic.
We have spent the past few months conducting validation experiments detailed in the new subsection titled "experimental validation". In order to validate our simulation work, we have considered the twin aspects of the question we posed as to whether silk is pushed or pulled, and have developed experiments which approach this problem from both the perspective of pushing, and of pulling.
We are pleased to report that both new methodologies have provided evidence which supports our hypothesis. The extrusion rig has demonstrated that the pressures required are both far in excess of what silkworms have the capacity to generate, and are also in line with both our, and previous, simulation studies. Our reeling rig has provided more support for our closing hypothesis, showing that through pultrusion, it is possible to generate stresses that are required to spin fibres in the same order of magnitude as those predicted through our simulations.
We believe both experimental approaches, in their non-typical use of a tensile testing machine and a rheometer as measurement devices, display experimental novelty that is in itself useful to the field, opening doors for further exploration. For the extrusion device we believe it may aid those in developing silk feedstock spinning devices. In the case of the rheometer-reeling device we can now perform practically unlimited reeling with greater control over processing conditions than the previous systems such as those used by Mortimer (Mortimer et al. 2013;Mortimer et al. 2015).
In summary we hope that our new experimental section broadens the impact and novelty of the manuscript and that we have addressed any concerns the reviewer may have held about the validity of simulation work.