Credit: Freepic

Fire is horrible especially if flame spreads over. Fire retardant chemicals can stop the fire and resist for up to 30 minutes. However, the chemicals contain halogen elements and organophosphate that have adverse effect on wildlife. Natural products have been reported as fire retardant biomaterials such as milk and wool because of high amount of phosphorus and nitrogen contents, respectively. We mined data from protein database and found a protein candidate, SR protein (serine/arginine splicing factor) with fire retardant properties. Cotton and cloth soaked with SR protein-producing bacteria have been proved as fire retardant materials on flame test. Furthermore, the wooden house painted with our bio-coatings are resistant to fire up to 45 seconds. We demonstrated the prototype to DLuB Fire Retardant Paint Inc. and plan to transfer the technology to KraigLabs, a Spider Silk Technology Company using genetically modified silkworm to produce fire retardant fabrics.



In June 2015, flammable dust explosion occurred at Formosa Fun Coast in Taiwan. The accident killed 10 people and injured more than 500 visitors. It reminds us how harmful a fire could be. We studied fire retardant chemicals and amazed at the 30-min fire resistant on the wooden house painted with fire retardant coatings.


Credit: The Inquisitr News


However, the chemicals are composed of organophosphates and/or halogen element such as chlorine and bromine. Both of which are confirmed toxic to human body and further make poisonous gasses when burning. U.S. Environmental Protection Agency warn the adverse effect on wildlife with exposure to such chemicals. Any natural products can replace the chemicals and be safe materials with fire retardant properties? We want to find the answer.




What’s fire? Fire is an oxidation-reduction chemical reaction. Combustion occurs between oxygen and fuel with heat over ignition temperature. The three elements are known as fire triangle. Free radicals formed in the burning process could spread the fire. Fire extinguished if one, or more, of four of them are removed or isolated.

In addition to fire retardant chemicals, there’s some natural products have been recognized as fire retardant bio-materials.




Milk is considered as the next-generation materials with fire retardant properties. Proteins in the milk have been studied and proved with fire retardant effect. Casein is the major components and estimated 80% in the cow milk. Federico Carosio, et al. studied the effectiveness of thermal stability and flame retardancy on the polyester and cotton fabrics treated with casein and published the data on the journal of Industrial & Engineering Chemistry Research in 2014. The data showed as below that the casein on the fabrics strongly reduced the burning rate and blocked the fire propagation. Casein is a phosphoprotein, which has up to 7 phosphorylation sites on the serine residues. The high content of phosphorus (0.8%) is believed to give casein fire retardant efficacy.

The picture of cotton (COT_C, left) and polyester  (PET_C, right) treated with casein at the end of flame spread test.

Credit: Ind. Eng. Chem. Res., 2014, 53 (10), pp 3917–3923




Why the high contents of phosphorus in milk and the nitrogen in wool play an important role in fire retardancy? Flame combustion requires three essential elements, heat, oxygen and fuel (fire triangle) and fire needs free radicals to spread. Removing one of them can stop the fire. Phosphorus and nitrogen can both react with free radicals and eliminate them. Fuels burned with materials containing phosphates can rapidly release water and form carbon layer. This charring process can isolate the fuel and protect it from burning. Ammonia burned in the air may react with oxygen to produce water and nitrogen gas, which could be a kind of inert gas to replace or isolate oxygen during combustion process.


Our goal of iGEM project this year is to find a novel fire retardant bio-material in addition to casein and wool. Furthermore, we’d like to make a fire retardant bacterium and produce fire retardant proteins. Finally, the prototype we create could be qualified through flame test and applied to the industry.


In order to find fire retardant protein candidates, we collected protein sequence data from Uniprot database and ranked with highest contents of serine (phosphoserine) and/or arginine (the highest nitrogen percentage among 20 amino acids). We chose a possible candidate (SR protein) in the top list of highest contents of both serine and arginine. Next, we built biobricks (pSB1C3) carrying SR gene and SRPK kinase gene which encodes proteins to phosphorylate SR protein, followed by putting the genes on the exprssion vectors of pGEX-2T and pET-29b. Then, we transformed and cultured bacteria (E. coli BL21) carrying the expression vectors. Finally, we analyzed the protein expression and performed flame test on cloth and cotton soaked with the bacteria.

Flow chart of finding and testing novel proteins with fire retardant properties



In order to obtain a novel bio-material with the best fire-retarding effect, we’re searching a protein with the highest nitrogen and phosphorus contents through data mining approach.


We performed four steps to find the fire retardant protein.

Protein Data Mining



Determining Serine and Arginine with fire retardant properties.

As mentioned earlier, phosphorous and nitrogen play important roles to stop fire propagation. We analyzed the contents of phosphorus and nitrogen among amino acids (AAs).


Phosphorus exists as the form of phosphate. Phosphate group was only added on the AA by phosphorylation through a post-translational modification (PTM). Among all 20 AAs, only serine, threonine and tyrosine could be phosphorylated. Phosphoserine is the most common modification in phosphorylated proteins. Serine was chosen as the major source of phosphorus in proteins.


Nitrogen content analysis showed that arginine contains the highest nitrogen rate of 33% by weight among AAs. (American Journal of Plant Sciences, 2011, 2, 287-296)

Thus, we chose it as the major source of nitrogen in proteins.

Nitrogen contents in amino acids. Credit: American Journal of Plant Sciences, 2011, 2, 287-296


Analyzing Proteins by C programming

In order to find protein candidates with fire retardant properties, C programming was written for data mining. (For collaboration with NCTU-Formosa and the code programming data, see here)


We collected protein sequences from UniProt and downloaded data as fasta format on February 13, 2015.


In the process, the sequence of a protein was read and the amount of each AA was added up, after that the AA amount was calculated into the rate of each AA by “ AA rate = (each AA amount)/(total AA in a protein)” by the C program automatically.

Ranking & Sorting

Next, the list of the AA rates of each protein was ranked by the (1) serine rate, (2) arginine rate, and (3) serine plus arginine rate.

Manual Curation

Finally, a manual curation was performed to remove the unknown, predicted or similar proteins.

The results of top 15 list of serine rate and arginine rate were shown below.

1.Top serine % of proteins in Uniprot database

2. Top arginine % of proteins in Uniprot database

Protein TPRXL is the highest serine rate by 67% and Protamine-B is the highest arginine rate by 70%. According to the unique roles and significant fire retardant effects played by phosphorus (serine) and nitrogen (arginine), serine plus arginine rate was calculated for our needs.


Top 10 of serine plus arginine rates of proteins in Uniprot was shown below.

Top 10 list of serine plus arginine rate showed that Protamine P3 and TPI6 are the top 2 proteins which are rich in arginine but short of serine. Serine/arginine-rich splicing factor 6 is more ideal for our criteria, which contains 33% of arginine and 40% of serine, matching our goal of finding a protein both rich in serine and arginine.


Serine/arginine-rich splicing factor 6 is a member belonging to SR protein family, which is a group of proteins conserved in serine/arginine-rich domain.


In our following experiments, we decided to use proteins in SR family as our target for bio-fire-retardant material.

Fire Retardant Candidate



SR protein family

Reference: Genome Biol. 2009, Biochem J. 2009

The SR protein family comprises a group of structurally related proteins with a conserved domain rich in serine (S) and arginine (R), which is so-called SR domain. They are involved in various mRNA metabolism and play essential roles in diverse cell functions such as constitutive and alternative pre-mRNA splicing, mRNA nuclear export, nonsense-mediated decay, and mRNA translation, etc. SR proteins are important in regulation of gene expression and developmental processes. SR domains are characterized by their ability to interact with other proteins and involved in nuclear translocation.

Which SR protein should we take?

The high serine and arginine contents of protein we predicted by data mining is Serine/arginine-rich splicing factor 6 from the species of Oryctolagus cuniculus (Rabbit). We conducted research on the Internet and take a trip to Academia Sinica, Taiwan to meet expert in SR protein research, Dr. Woan-Yuh Tarn. She did a lot of researches on SR protein and now is transferring interests to mRNA metabolism. She suggested Serine/arginine-rich splicing factor 1 in human with a long stretch of SR domain. And she has materials of SR gene for gene cloning and protein expression as well as SRPK1 gene encoding SR protein kinase, which can phosphorylate SR protein.

Serine/arginine-rich splicing factor 1

Reference: Wikipedia

Serine/arginine-rich splicing factor 1 (SRSF1) is a protein in human encoded by SFRS1 gene, which is located on chromosome 17. SRSF1 is essential for alternative splicing, mRNA nuclear export and translation. The function of SRSF1 is regulated by phosphorylation at the serine residues in SR domain by SR specific protein kinase 1 (SRPK1). And up to 12 phosphoserine sites have been characterized (J Mol Biol. 2008) The phosphorylated SR protein plays crucial roles in heart development, embryogenesis, tissue formation, cell motility, etc.

SRSF1 amino acid sequence with SR domain



SRSF1 has 248 amino acids with a long stretch of SR domain in the C terminus containing a total of 43 arginine (17.3%) and 28 serine (11.2%). The comparison of protein composition between SR protein, wool and casein were made in the following table.

Comparison of phosphorus and nitrogen contents in fire retardant protein candidates, SRSF1, wool and casein

*Nitrogen content calculation : American Journal of Plant Sciences, 2011, 2, 287-296

**Phosphoserine : phosphorylated serine residues


SRSF1 has more nitrogen contents than wool and casein. The serine contents and phosphoserine residues characterized in SRSF1 are also greater than casein. Based on the analysis, the novel fire retardant protein, SRSF1, theoretically, could have better effect than wool and casein.

Gene Cloning



We got the plasmids of SRSF1 gene encoding SR protein and SRPK1 gene encoding protein kinase to phosphorylate SR protein from the lab of Dr. Woan-Yuh Tarn.


First of all, we cloned SR gene and SRPK1 gene onto the standard BioBrick, pSB1C3. Then, SR and SRPK1 genes were transferred to the commercial expression vectors of pGEX-2T and pET-29b, respectively, which have strong expression ability through IPTG induction process.


All the constructs have been confirmed by colony PCR, restriction enzyme check and sequencing. (For more detail, please see here)

SR/pGEX-2T: the GST-fused SR gene is driven by Ptac promoter, which is regulated by LacI repressor and induced by IPTG The vector is ampicillin resistant and has the origin of replication from pBR322.

SRPK1/pET-29b: the SRPK1 gene is fused with 6XHis tag at the C terminus and is driven by Pt7 promoter, which is modified to be regulated by LacI repressor and induced by IPTG. The vector is kanamycin resistant and has the origin of replication from ColE1. The genes carried on pET-29b vector has to be expressed in E. coli BL21 strain which has T7 RNA polymerase.

Protein Expression



We transformed E. coli BL21 with the following groups of vectors and made stocks at -80°C:

1.SR/pGEX-2T plus SRPK1/pET-29b (for phosphorylated SR protein expression)

2.SR/pGEX-2T with pET-29b empty vector (as an unphosphorylated SR control

3.SRPK1/pET-29b with pGEX-2T empty vector (as a SRPK1 only control)

4.pGEX-2T and pET-29b empty vectors (as a vector only control)


To analyze protein expression, we run SDS-PAGE and performed Coomassie blue staining and western blotting with the following standard procedure.


↓ recover E. coli BL21 from the stock at -80°C in 1ml of LB media supplemented with 50ug/ml Ampicillin and 25 ug/ml Kanamycin and culture O/N

↓ 100X dilution in 3ml of LB media with antibiotics and culture for around 2 hr

↓ Until OD600 between 0.6 to 1.0, IPTG was added at the concentration of 1mM and culture for another 4 hr for protein induction

↓ bacteria were collected by centrifuged and lysed in lysis buffer (lysates can stored at -20°C)

↓ After centrifugation, lysates were collected and combined with sample buffer followed by boiling at 95°C for 10 min

↓ Now the samples were ready for running SDS-PAGE

↓ To run PAGE, transfer and western blotting are following the manufacturers’ protocol.


Antibodies we used to blot the protein and predicted molecular weight of indicated proteins are described below.

SDS-PAGE stained with Coomassie Blue

As the figure showed, SPPK was expressed at the predicted 93 kDa (Lane 3) and SR protein was at 53 kD (Lane 2). In addition, the free form of GST (Lane 1 & 3) can also be detected at around 27 kDa


Figure 1. Coomassie Blue staining on SDS-PAGE

Lane 1: Control: lysates of BL21 carrying pGEX-2T and pET-29b

Lane 2: SR:  lysates of BL21 carrying SR/pGEX-2T and pET-29b

Lane 3: SRPK:  lysates of BL21 carrying pGEX-2T and SRPK1/pET-29b

Western blot analysis with anti-His antibody

As Figure 2 showed, SRPK protein fused with 6XHis tag was detected by anti-His antibody and visualized with BCIP/NBT chromogen at the predicted size of 93 kDa in Lanes 3 and 4.


Figure 2. His-SRPK was detected by anti-His antibody.

Lane 1: Control: lysates of BL21 carrying pGEX-2T and pET-29b

Lane 2: SR:  lysates of BL21 carrying SR/pGEX-2T and pET-29b

Lane 3: SRPK:  lysates of BL21 carrying pGEX-2T and SRPK1/pET-29b

Lane 4: phosphorylated SR:  lysates of BL21 carrying SR/pGEX-2T and SRPK1/pET-29b

GST-SR was detected by anti-GST antibody in Western blot analysis

As Figure 3 showed, SR protein fused with GST tag was detected by anti-GST antibody and visualized with DAB chromogen at the predicted size of 53 kDa in Lanes 2 and 4.  Moreover, in Lane 4, the shifted bands indicated the possible of various forms of phosphorylation modified SR proteins by SRPK1 kinase. Free forms of GST can be detected at  27 kDa in Lanes 1 and 3.


Figure 3. GST-SR was detected by anti-GST antibody.

Lane 1: Control: lysates of BL21 carrying pGEX-2T and pET-29b

Lane 2: SR:  lysates of BL21 carrying SR/pGEX-2T and pET-29b

Lane 3: SRPK:  lysates of BL21 carrying pGEX-2T and SRPK1/pET-29b

Lane 4: phosphorylated SR:  lysates of BL21 carrying SR/pGEX-2T and SRPK1/pET-29b

In summary, SR protein and SRPK protein can be expressed in E. coli BL21 carrying the corresponding plasmids, proving the successful protein induction by IPTG and gene expression by pGEX-2T and pET-29b vectors. Furthermore, SR protein modified by SRPK kinase can be clearly observed in the shifted bands, probably demonstrating the phosphorylation process.

Flame Test



In order to clarify the fire retardant properties of phosphorylated SR protein, E. coli BL21 bacteria expressing SR and SRPK1 kinase were subjected to flame test on the samples of cloth and cotton. Protein induction was performed as the procedure in protein expression and analysis. Cloth or cotton were soaked in the culture media of E. coli BL21 for one minute followed by dried completely in the oven at 100°C for 30 minutes. The fire from flame gun was ignited for 3 seconds and last for 30 seconds on soaked cloth samples (Figure 1), and ignited for 2 seconds followed by determining the time period of burning out or self-extinguishing for cotton samples (Figure 2). The burning lengths of both groups of samples were measuring. Groups of samples are listed as follows:
1. Media: samples soaked in the LB culture media only.
2. Vector: samples soaked in E. coli BL21 carrying pGEX-2T and pET-29b vectors.
3. SRPK: samples soaked in SRPK-expressing E. coli BL21 carrying pGEX-2T and SRPK1/pET-29b vectors
4. SR: samples soaked in SR-expressing E. coli BL21 carrying SR/pGEX-2T and pET-29b vectors
5. SR-P: samples soaked in E. coli BL21 producing phosphorylated SR protein modified by SRPK kinase, which carrying SR/pGEX-2T and SRPK1/pET-29b vectors. As figures 1 and 2 showed, both cloth and cotton samples soaked in the culture media containing with E. coli BL21 expressing phosphorylated SR proteins has fewer lengths of burning compared to other groups. These data demonstrated phosphorylated SR protein is able to retard fire. It’s worth to notice that unphosphorylated SR protein has some fire retardant effect because the high nitrogen (arginine) contents of SR protein than SRPK protein or others, especially in cotton samples. SRPK protein has no significant effect in cloth or cotton samples compared to vector only control. Furthermore, it’s more significant that cotton samples soaked in E. coli BL21 expressing phosphorylated SR proteins are really hard to be burned, and the samples were self-extinguishing right after flame ignition (Figure 3).

Figure 1.

Burning lengths of cloth samples soaked with E. coli BL21 expressing proteins.

Figure 2.

Burning lengths of cotton samples soaked with E. coli BL21 expressing proteins.

Figure 3.
The time of burning out or self-extinguishing of cloth samples soaked with E. coli BL21 expressing proteins.

Taken together, the data of flame test strongly proved that phosphorylated SR protein is able to retard the fire propagation and further extinguish the burning. Moreover, the unphosphorylated SR protein has also some effect on fire retardancy. The results are not only demonstrating phosphorylated SR protein is a novel fire retardant protein, but also verifying our theoretical hypothesis of phosphoserine and nitrogen contents that play unique and additive roles on the fire retardant properties.

Summary & Discussion

Casein and wool are natural products with fire retardant properties. Phosphorus and nitrogen play important roles on the fire retardancy. We searched the highest (phospho)serine and highest arginine (nitrogen) contents in the protein databank of Uniprot by data mining. SR proteins are a family of proteins conversed in serine/arginine-rich domain and highly be phosphorylated. We expressed SR protein and SRPK1 kinase protein in E. coli BL21, which were analyzed in SDS-PAGE with coomassie blue staining and Western blotting with specific antibodies. Next, flame tests were performed on cloth and cotton samples soaked in the culture media containing protein-producing bacteria, demonstrating the fire retardant efficacy of phosphorylated SR proteins.


In protein analysis of phosphorylated SR proteins (section “Protein Expression”, Figure 3), total lysates were blotted with anti-GST antibody. The data showed various levels of band shift on SR plus SRPK (Lane 4) compared to SR only (Lane 2) around 53 kDa, indicating modification of SR protein. This result highly suggests SR protein underwent phosphorylation due to the presence of SRPK kinase. On the other hand, we’re trying to blot the lysates of phosphorylated SR protein with anti-phosphoserine antibody. However, we technically failed the experiment and didn’t get a good result with anti-phosphoserine antibody.


Phosphorus and nitrogen are important on the effect of fire retardancy. We tried to synthesize a gene encoding a polypeptide with arginine repeat. However, IDT company, a sponsor of iGEM, replied us a technical difficulty on synthesizing nucleotides with high frequency repeats. As for phosphorus, as we mentioned, proteins don’t carry any phosphate groups until being phosphorylated by specific kinase. Therefore, finding an existing protein and using an identified kinase are ideal to make phosphorylated proteins.


Currently, we don’t compare SR proteins to casein or wool on fire retardant efficiency. Because the cost and time to purify SR proteins are still problems for us so far. We are looking for companies who might have interests in fire retardant bio-materials sponsoring us to solidify our discovery.


In summary, our work is a proof-of-concept study of finding and verifying a novel fire retardant natural protein candidate of SR protein through data mining and flame test. Casein and wool have been proved as natural, non-toxic fire retardant bio-materials. And both of them have been patented. The SR protein we found could be the next-generation biomaterial for studying fire retardant effectiveness.