For this project we really wanted to print out a molecule. With visualization software (such as Chimera and Pymol), the process of obtaining a 3D printable, digital representation of a biological molecule, specifically proteins, has been incredibly streamlined. These programs already have built-in systems that allow the user to fetch a desired protein using its PDB (Protein Data Bank) ID, modify it and enhance it, and then save it as a 3D printable object (in our case, an STL file). Trying to take advantage of this workflow, we set out to find an appropriate protein for our project.
Some of the key features we were looking for in this molecule were:
- It had to be compact and possess a well-defined domain: no long sidechains that would be impossible to contain within the volume of the 3D printer or undefined regions that would create holes in our print.
- It had to possess a distinctive shape: something visually interesting and appealing that would be easy to handle and orient during post-processing.
- It had to be of biological importance: something that a person would actually want to 3D print out of interest or curiosity.
Looking for a molecule that fit all these criteria was difficult. We finally landed on the C-terminal domain of spindle and kinetochore-associated protein 1 (Ska1, PDB ID: 2LYC). It had a globular, clear, and well-defined surface, with multiple polarity regions; it possessed two central folds that manifested themselves into a distinctive arched shape, with multiple regions of interest on the surface; and it plays a central role in chromosome segregation during cell division (it’s necessary for proper kinetochore-microtubule attachment and mitotic progression).
We loaded the protein into Chimera and started with some pre-processing so that the STL object would be as high definition as possible for the print. We started by hiding all ribbons and atoms, and only turning on the surface view. We then changed the color space from standard to CMYK, turned on the spectrum display, set the rendering to use maximum memory, set the display to maximum quality, and exported the whole model as an STL file.
Figure 1. Ska1 CAD file.
The two biggest considerations for printing molecules like this are the orientation of the molecule and the supports. The intrinsic asymmetry of the molecule means that orientation is important in order to allow the print to be as big as possible and still fit in the printer. We also want to provide the widest base possible, in order for the vertical curvatures of the surface to have greater support and avoid the problems associated with extruding in mid-air. In this regard, we needed to balance wanting to have the biggest molecule possible (which would imply orienting it vertically) and wanting to provide the most amount of support for the details on the surface (implying a horizontal orientation). Lastly, another potential issue that could arise was that FDM is not particularly good a printing small, multiple, and irregular curvatures (which this print was full off).
We decided to favor the stability and support of the print to offset the problems associated with using FDM for this kind of print. We oriented the molecule so as to achieve the widest base possible, with the peak of the arch on top. We also oriented the molecule at a 45 degree angle, along the diagonal of the base of the printer. This allowed us to maximize the print’s size given our current orientation. For all other variables we used the standard values (initial layer height: 0.2; infill percentage: 15%; supports: active).
This first print came out looking incredibly well, but it was too heavy and it took longer than necessary. Our print ended up weighing 871.35 grams and took 26 hours and 12 minutes. For our second print we decided to change some variables in order to reduce the amount of filament needed and the amount of time required to print it. The infill percentage was reduced to 5% and the infill pattern was changed to concentric.
Figure 3. Protein in Bambu Studio.
The resulting print was much lighter (402.40 grams) and faster to print (20 hours and 10 minutes), even though it was noticeably larger, and didn’t appear to be structurally or aesthetically inferior, to the first print. Overall the weight was reduced by 54% and the print time by 23%.
Figure 4. Comparing both prints.
After finishing our prints, our first objective was to remove all the supports. For this we mostly used our hands, while occasionally being aided by pliers in order to reach supports in harder to reach areas (particularly the support inside a hole in the molecule).
Figure 5. Removing the supports.
We then decided to sand the pieces to smoothen out the surface. Because the prints were so large and had a very complex surface, we used a Dremel sanding tool, and then went over remaining imperfections with sanding paper.
On debating on how to paint the proteins, we chose to lean into the their biologic aspect, and paint them in a way that would convey information about the molecules themselves. We landed on painting the Coulombic, electrostatic surface interactions of the molecule (red for negatively charged regions and blue for positively charged regions). This would provide insight into the ways the protein is able to bind to other molecules, but would also be simple to color, as it only requires two colors. In order to have a model of these interactions, we decided to load the protein into Chimera and use its surface analysis tools.
Figure 7. Coulombic surface analysis
In order to achieve this, we first had to apply primer to the pieces. We spray painted 2 in 1 (filler and sandable) primer, and we applied two coats to ensure all the nicks and crannies of the prints were covered.
Figure 8. Applying primer.
We then began with the painting process. We hand painted the blue and red regions of the molecules using acrylic paint.
Figure 9. Painting the molecules.
Coulombic surface interactions are always displayed in shades of red and blue to showcase the increased and decreased polarity of the protein, similar to a heatmap. In the interest of time we chose to do a plain, single color to highlight general regions of positivity or negativity. This, obviously, means our models are less accurate than digital representations or 3D printing methods that can print in color. Moreover, errors in the painting lines are noticeable, we tried to correct them by painting over them with grey paint, but painting on such an irregular surface definitely presented a challenge. Lastly, small differences between the two pieces can be appreciated, although efforts to make both pieces as identical as possible were made.
Lastly, we wanted to add a shiny, glossy finish to the pieces.
Figure 10. Sprayed with clear-coat.
The finished pieces ended up looking incredibly well! Even though there is definitely room for improvement, we tried our hardest to ensure they looked as professional and detailed as possible.
Cost Analysis:
| Cost Type |
Cost |
Price |
Source |
Quantity |
Total |
| Materials |
PLA Filament |
$19.99/kg |
BambuLabs |
0.87 kg (first print)
0.40 kg (second print) |
$25.39 |
| Automotic Primer Spray |
$7.97/btl |
Amazon |
1 btl |
$7.97 |
| Clear Gloss Spray Paint |
$6.19/btl |
Amazon |
1 btl |
$6.19 |
| Acrylic Paint |
$9.99/set |
Amazon |
1 set |
$9.99 |
| Labor |
3D Printing Engineer |
$41/hr |
ZipRecruiter |
10 hrs |
$410 |
| Total |
$459.54 |
Figure 11: Clean workspace.
