Group members: Wen-Yi, Xinyin (Sing)
Wen-Yi and I (Sing) used fused deposition modeling (FDM) to print two spCas9 molecules. In this blog, you will learn how we made the spCas9s step by step, stories about our experimentations with different sizes and supports, and lessons we found worth sharing. Let’s begin!
- Background
As lovers of challenges and newbies to 3D printing, we chose the open-ended prompt of printing two identical medical-related molecules. We chose spCas9 because it is one of the coolest proteins in my opinion. The CRISPR-Cas9 system is a game changer that revolutionizes gene editing and gives hope to many patients. Among Cas9, spCas9 is the most widely used, where “sp” stands for Streptococcus pyogenes1, a bacterial pathogen that infects humans2. Due to its structural complexity with 7 domains, spCas9 is very fitting for 3D printing. Additionally, 3D-printed Cas9 will not only be a visually appealing artwork, but also a great teaching tool for introductory to biology classes.
There are different types of 3D printing available in OEDK, including fused deposition modeling (FDM), stereo-lithography (SLA), (and sintering (SLS) in the wet lab?). We printed both molecules on FDM, which can use melting polylactic acid (PLA) as “ink” and essentially does 2D printing in layers (Fig. 1).
Figure 1. Fused deposition modeling (FDM)3.
We chose FDM out of two considerations. First, FDM is cheaper than SLA, which uses resin as the “ink”, whose price is above $50 per kilogram4. In contrast, PLA for FDM is around $20 per kilogram (Table 1). Second, the post-processing of FDM printed spCas9 is easier and has a higher likelihood of success than SLA. The post-processing of SLA products includes washing, drying, and curing, which is time-consuming. In addition, the supports in the pockets of spCas9 are difficult to access for a normal-size cutter, but removing supports in FDM can be easy (as you will see). Although SLA can create products with finer surfaces than FDM, we consider FDM a more suitable method for our application.
2. Steps to make the spCas9s
In this section, I will describe three steps to our final products in details – file preparation, printing, and post-processing.
2.1 File preparation
First, we downloaded the pdb file of spCas9 from Protein Data Bank (PDB) (Fig. 2).
Figure 2. Crystal structure of spCas9 (4CMP) from PDB.
Next, we imported the pdb file into Pymol. The file contains chains A and B, each representing one spCas9 (Fig. 3). spCas9 functions as monomers. The presence of the dimer could be an artifact from protein crystallization, where proteins were crystallized at a very high concentration.
Figure 3. spCas9 dimers in Pymol.
To print scientifically accurate monomers, we learned to separate the two chains on Pymol on YouTube (around 4:02). After selecting the “show surface” option and removing water molecules, we got the two spCas9 monomers almost ready for the printer (Fig. 4). Then, we exported the image as STL and opened it in the Bambu Studio slicer.
Figure 4. Two spCas9 monomers.
Next, we adjusted parameters and created a printer-understandable file using the Bambu Studio slicer (Fig. 5). Based on our previous experimentation (which you are welcome to read more below under section 3), tree support is more suitable for our application. Additionally, for the infill, we chose the honey cone style, which was shown to be the most economical style with regard to strength compared with other styles available5.
Figure 5. Setting for the bambu 1X carbon printer.
2.2 Printing
Including the prepare time, the whole printing for the two spCas9 monomers process takes 7 hours and 46 minutes. Thankfully, we did not enter an error in this run. We monitored the printer until the first layer was done smoothly. If you are not familiar with how the Bambu 1X carbon printer works, check the video below that we took (Video 1).
Video 1. A busy Bambu 1X carbon printer.
The spCas9s with supports are shown in Fig. 6.
Figure 6. spCas9s with supports.
2.3 Post-processing
Post-processing includes removing the support, coloring, and paint spraying.
To remove the support, we used a cutter, tweezers, and a metal stick (Fig. 7). The cutter was used to separate the branches, tweezers were used to drag the branch out, and the metal stick was used to push supports out of hole-like structures. It took Wen-Yi and I half an hour to remove all the support. The support accounts for 37% of the total material.
Figure 7. Products without supports and tools (left). Waste (right).
Before painting the spCas9s, we color-coded the 7 spCas9 domains on Pymol according to the domain information provided by Jinek et al6 (Fig. 8). 
Figure 8. spCas9 (4CMP) domains6.
The envisioned product is shown in Video 2. We learned to create a video on Pymol from this tutorial, which is really cool!
Video 2. Color-coded spCas9 (Chain A) on Pymol. Red: PAM interacting domain, yellow: RuvC domain, green: HNH domain, blue: REC I, gray: REC II, cyan: REC III, magenta: bridge helix (consistant with the paper).
We used the Vallejo Paints and our watercolor brushes. Since the Vallejo Paint set does not have cyan or magenta, we replaced them with brown (skin color) and purple, respectively. To depict the domains as accurate as possible, it took Wen-Yi and I 2 hours to paint, which was the longest (and fun) labor period (Fig. 9). Surprisingly, we found that Chain A and Chain B have minor differences. Upon examining the amino acid sequence on PDB, we found that neither Chain A nor Chain B captures the complete sequence of spCas9, and they miss different amino acids. If you examine closely enough, you can tell that the two molecules are not perfectly identical. Although this result deviates from the assignment requirement (please do not take off points from us), we were excited about this minor discovery.
Figure 9. Wen-Yi hand-painted the scCap9 (Chain B).
After finishing painting, Wen-Yi and I applied clear spray paint to the spCas9s, following the instructions on the bottle. The paint endorsed the spCas9s a glossy appearance. Tada~ Here are our final products(Fig. 10)!
Figure 10. Final spCas9 final products.
- Side stories
Before we realized that the PDB file contains two chains, we printed 3 spCas9 dimers to test the suitable size and support (Table 2).
| spCas9 dimer | Longest dimension (inch) | Support |
| #1 | 3 | tree |
| #2 | 4 | tree |
| #3 | 4 | normal |
Table 2. Three spCas9 dimers.
We first printed #1 and #2 to figure out a suitable size. Although the 3-inch one just looked a bit smaller than the 4-inch one (Fig. 11), it was ultra-difficult for our tools to reach the supports in the small holes and took us one more hour to remove its support. In comparison, it was easier to remove the tree support from the 4-inch molecule.
Figure 11. spCas9 dimers without support, 3-inch and 4-inch.
To finish our assignment, we could just print another 4-inch molecule with tree support. However, out of curiosity, we decided to compare the tree support with the normal support, as the former is recommended by our TA, and the latter is recommended by some classmates.
We printed #3 with normal supports, which otherwise had an identical setting to #2. It turns out that it was a pain in the butt to remove normal support, and we gave up (Fig. 12). Therefore, for our monomer spCas9, we stuck with tree support, which worked so well.
Figure 12. 4-inch spCas9 dimer with normal support.
- Discussion
From the monomer-dimer story, we learned an important lesson – check the molecule files carefully before printing! However, the three dimers we experimented with provided useful information about size and supports, so I deem them as our steps to success.
For our final products, we observed valleys and grooves on the surface. Additionally, we observe that the sides of the products are smoother and more detailed (Fig. 10 bottom left image), and the filament is apparent on the plane of printing (Fig. 10 bottom right image). Due to the intricacy of the proteins, they are unsuitable for sanding, as sanding can easily erase an amino acid. A potential approach is to reduce the layer height. The current value we used is 0.15 mm. However, there is a tradeoff between details and printing time.
Finally, I want to say that I am so proud of us! We were newbies, but we can now comfortably use the Bambu Studio slicer, with experience with different settings. Moreover, the challenge of the open-ended project gave us the opportunity to explore protein crystal structures and various of functions on Pymol. Additionally, we know the structure of spCas9 better than most people now, because we covered every mm2 of the surface with paint stroke by stroke.
Cleaned table
| Cost Type | Cost | Price | Source | Quantity | Total |
| Materials | PLA filament | $19.99 / kg | Bambu Lab | 0.0845 kg | $1.69 |
| Vallejo Paints | $4.13 / Fl Oz | Amazon.com | 0.12 Fl Oz* | $0.50 | |
| Spray paint | $0.5 / Oz | Homedepot | 0.1 Oz | $0.05 | |
| Labor | 3D printing Lab Assistant | $17.5 / hour** | Glassdoor.com | 0.25 hr*** | $4.38 |
| Prototyping Engineer (Intern) | $17 / hour | ZipRecruiter.com | 6.5 hr | $113.90 | |
| Overhead | Bambu 3D printer (rental price) | $19/ day | Fatllama.com | 8 hr | $6.33 |
| Electricity (for the light, the computer, and the printer) | $0.13 per kWh | Raise3D | 2.3 kWh | $0.3 | |
| Post-processing tools (tweezers and brushes) | Tweezer $5.29 each Brushes $3.99/ set |
Amazon.com Amazon.com |
1 set | $9.28 | |
| Quality control | 27.5/ hour****** | LinkedIn.com | 0.5 hr | $13.75 | |
| Design | Engineering and Development***** | $17 / hour | ZipRecruiter.com | 1 hr | $17.00 |
| Iterations
***** |
$17 / hour | ZipRecruiter.com | 1 hr | $17.00 | |
| Misc. | Waste and Scrap | $0.66/Gal= $0.17/L | Houstontx.gov | 0.025L | $0.0042 |
Sum: $184.1842
Table 1. Cost table for the TWO 3D-printed spCas9.
*We used 70 drops of Vallejo Paints, and we estimate each drop to contain 0.0016907 Fl Oz.
**We took the average of $14 and $21.
***The lab assistant replaced the filament, addressed error messages, and took out the products. The hours they spend on the products are at most 15 minutes.
****0.5 hours for setting up the printing and monitoring. For each of two people, 0.5 hours for removing the support, and 2.5 hours for coloring and spraying, 0.1 hours for cleaning. 0.5+2x(0.5+2.5+0.1) = 6.7 hours
*****We treat it as a part of the intern time.
******We took the average of $25 and $30.
