Difference between revisions of "Part:BBa K2423007"

 
(8 intermediate revisions by 2 users not shown)
Line 8: Line 8:
  
 
Due the enzyme's previously poor charactarization, a homology model has been created for this part. That model was then used in production molecular dynamics simulations and steered molecular dynamics simulations to verify that CsADH2946 has specificity towards crocetin dialdehyde.
 
Due the enzyme's previously poor charactarization, a homology model has been created for this part. That model was then used in production molecular dynamics simulations and steered molecular dynamics simulations to verify that CsADH2946 has specificity towards crocetin dialdehyde.
 +
 +
For more information check out our wiki here: http://2017.igem.org/Team:Uppsala.
 +
 +
__TOC__
  
 
===Usage and Biology===
 
===Usage and Biology===
Line 23: Line 27:
  
 
For purification CsADH2946 was transformed and expressed in ''E. coli'' strain BL21(DE3*) and purified using IMAC on an ÄKTA protein purification system. A gradient of imidazole concentration from 20–500 mM was used. The peak pointed at by the arrow in the chromatogram (figure 1) indicates protein that elutes at high imidazole concentration, i.e the desired his-tagged CsADH2946. The purification was followed by SDS-PAGE to analyse the fractions, control purity and verify the protein product. In figure 2a the band at around 60 kDa in the crude pellet indicate an overexpression of a protein in that size range. In the SDS gel of fractions 16–26 collected between 115–145 mL elution volume (figure 2b) there is a strong band at 60 kDa corresponding to the molecular weight of CsADH2946, indicating that the protein was successfully overexpressed and well-separated.
 
For purification CsADH2946 was transformed and expressed in ''E. coli'' strain BL21(DE3*) and purified using IMAC on an ÄKTA protein purification system. A gradient of imidazole concentration from 20–500 mM was used. The peak pointed at by the arrow in the chromatogram (figure 1) indicates protein that elutes at high imidazole concentration, i.e the desired his-tagged CsADH2946. The purification was followed by SDS-PAGE to analyse the fractions, control purity and verify the protein product. In figure 2a the band at around 60 kDa in the crude pellet indicate an overexpression of a protein in that size range. In the SDS gel of fractions 16–26 collected between 115–145 mL elution volume (figure 2b) there is a strong band at 60 kDa corresponding to the molecular weight of CsADH2946, indicating that the protein was successfully overexpressed and well-separated.
 
+
<html>
https://static.igem.org/mediawiki/2017/b/bf/CraftingCrocinElutionStep2.png
+
<div class="col-xs-10">
 
+
      <figure class="figure">
Figure 1. Chromatogram from IMAC-purification of CsADH2946.
+
        <img src="https://static.igem.org/mediawiki/2017/b/bf/CraftingCrocinElutionStep2.png" class="figure-img img-fluid" style="display: block; margin: auto; width: 65%; height: auto; padding-top: 5%;">
 
+
        <figcaption class="figure-caption figtext" style="padding-left: 20%; padding-right: 20%; padding-bottom: 3%;"> Figure 1. Chromatogram from IMAC-purification of CsADH2946.
https://static.igem.org/mediawiki/2017/9/9e/CraftingCrocinSDS-PAGE1.png
+
</figcaption>
https://static.igem.org/mediawiki/2017/c/cd/CraftingCrocinSDS-PAGE2.png
+
      </figure>
 
+
    </div>
Figure 2. a) SDS-PAGE gel of from IMAC purification. 1) Crude pellet. 2) Pellet after lysis. 3) Supernatant after lysis. 4) Flow through. 5) Wash with buffer A. 6) PageRuler protein prestained ladder. b) SDS-PAGE gel from IMAC purification. Fractions 16-26 were collected between 115 and 145 ml elution volume. A band at about 60 kDa is clearly visible.
+
</html>
 +
<html>
 +
<div class="col-xs-10>
 +
      <figure class="figure">
 +
        <img src="https://static.igem.org/mediawiki/2017/9/9e/CraftingCrocinSDS-PAGE1.png" class="figure-img img-fluid picturerow" style="display: block; margin: auto; width: 65%; height: auto; padding-top: 5%;">
 +
        <img src="https://static.igem.org/mediawiki/2017/c/cd/CraftingCrocinSDS-PAGE2.png" class="figure-img img-fluid picturerow" style="display: block; margin: auto; width: 65%; height: auto; padding-top: 5%;">
 +
        <figcaption class="figure-caption figtext" style="padding-bottom: 3%; padding-left:8%; padding-right:10%;"> Figure 2. a) SDS-PAGE gel of from IMAC purification. 1) Crude pellet. 2) Pellet after lysis. 3) Supernatant after lysis. 4) Flow through. 5) Wash with buffer A. 6) PageRuler protein prestained ladder. b) SDS-PAGE gel from IMAC purification. Fractions 16-26 were collected between 115 and 145 ml elution volume. A band at about 60 kDa is clearly visible. </figcaption>
 +
      </figure></div>
 +
</html>
  
 
The purified CsADH2946 was activity verified in converting crocetin dialdehyde to crocetin. An activity measurement assay was performed on a plate reader measuring absorbance of the substrate and product of the reaction. For the experiment we used a 96-well plate in which we included wells with enzyme from pooled fractions + substrate, as well as positive and negative controls, see table 1 for the specifics.
 
The purified CsADH2946 was activity verified in converting crocetin dialdehyde to crocetin. An activity measurement assay was performed on a plate reader measuring absorbance of the substrate and product of the reaction. For the experiment we used a 96-well plate in which we included wells with enzyme from pooled fractions + substrate, as well as positive and negative controls, see table 1 for the specifics.
  
Table 1. Content of wells used for activity measurement of CsADH2946.
+
<html>
 +
<div class="col-xs-10">
 +
      <figure class="figure">
 +
<figcaption class="figure-caption figtext" style="padding-left: 20%; padding-right: 20%;"> Table 1. Content of wells used for activity measurement of CsADH2946.
 +
</figcaption>
 +
        <img src="https://static.igem.org/mediawiki/2017/7/7c/CraftingCrocinTableStep2.png" class="figure-img img-fluid" style="display: block; margin: auto; width: 65%; height: auto; padding-top: 5%;">
 +
      </figure>
 +
    </div>
 +
</html>
  
https://static.igem.org/mediawiki/2017/7/7c/CraftingCrocinTableStep2.png
+
As can be seen in the activity spectra (figure 3), the absorbance of the product crocetin increases over time in well 2 containing enzyme and the substrate crocetin dialdehyde. After 9 hours of reaction, the blue curve corresponding to the enzyme + substrate mixture has increased its absorbance in the exact range of the product. The negative and positive control curves look similar to time point zero, apart from some precipitation of product and substrate indicated by the decreased curves. This is an evidence that the enzyme is functional. Using this data, K<sub>M</sub> = 20.7842 µM ± 3.5264 was estimated. This estimation was done using a Bayesian inference algorithm, see figure 4 for the fitted curve and eperimental data.
  
As can be seen in the activity spectra (figure 3), the absorbance of the product crocetin increases over time in well 2 containing enzyme and the substrate crocetin dialdehyde. After 9 hours of reaction, the blue curve corresponding to the enzyme + substrate mixture has increased its absorbance in the exact range of the product. The negative and positive control curves look similar to time point zero, apart from some precipitation of product and substrate indicated by the decreased curves. This is an evidence that the enzyme is functional. Using this data, K<sub>M</sub> = 20.7842 µM ± 3.5264 was estimated. This estimation was done using a Bayesian inference algorithm.
+
In addition, in figure 5 we can see that well 2 containing enzyme and crocetin dialdehyde has changed color compared to the negative control, to become more yellow like the product crocetin in well 8. This also shows that CsADH2946 was produced and that it converts crocetin dialdehyde into crocetin.
 
+
In addition, in figure 4 we can see that well 2 containing enzyme and crocetin dialdehyde has changed color compared to the negative control, to become more yellow like the product crocetin in well 8. This also shows that CsADH2946 was produced and that it converts crocetin dialdehyde into crocetin.
+
 
<html>
 
<html>
 
<div class="col-xs-10">
 
<div class="col-xs-10">
 
       <figure class="figure">
 
       <figure class="figure">
         <img src="https://static.igem.org/mediawiki/2017/d/d3/CraftingCrocinActivityStep2.png" class="figure-img img-fluid" style="display: block; margin: auto; width: 65%; height: auto; padding-top: 5%;">
+
         <img src="https://static.igem.org/mediawiki/2017/d/d3/CraftingCrocinActivityStep2.png" class="figure-img img-fluid" style="display: block; margin: auto; width: 65%; height: auto; padding-top: 5%;"><br>
 
         <figcaption class="figure-caption figtext" style="padding-left: 20%; padding-right: 20%; padding-bottom: 3%;"> Figure 3. Activity measurement curve. The dotted lines describe wells at the starting time and the fully drawn lines describe the absorbance after 9 hours. The blue lines indicate wells containing protein and crocetin dialdehyde, red lines describe positive control with only crocetin dialdehyde and black lines correspond to the negative control only containing the desired product crocetin.
 
         <figcaption class="figure-caption figtext" style="padding-left: 20%; padding-right: 20%; padding-bottom: 3%;"> Figure 3. Activity measurement curve. The dotted lines describe wells at the starting time and the fully drawn lines describe the absorbance after 9 hours. The blue lines indicate wells containing protein and crocetin dialdehyde, red lines describe positive control with only crocetin dialdehyde and black lines correspond to the negative control only containing the desired product crocetin.
 
</figcaption>
 
</figcaption>
Line 51: Line 69:
 
     </div>
 
     </div>
 
</html>
 
</html>
 
+
<html>
https://static.igem.org/mediawiki/2017/b/bf/CraftingCrocinWellStep2.png
+
<figure class="figure">
 
+
        <img src="https://static.igem.org/mediawiki/2017/8/89/Uppsala-Modeling-Pic13.png" class="figure-img img-fluid" style="display: block; margin: auto; width: 60%; height: auto;">
Figure 4. 96-well plate for activity measurements post 24 hours reaction and plate reading. The plate include pooled enzyme fractions 10–15 + substrate crocetin dialdehyde (well 1), pooled enzyme fractions 16–23 + crocetin dialdehyde (well 2), flow through + crocetin dialdehyde (well 3), negative control with only crocetin dialdehyde (well 4), pooled enzyme fractions 10-15 + product crocetin (well 5) pooled enzyme fractions 16–23 + crocetin (well 6), flow through + crocetin (well 7) and positive control with only crocetin (well 8).
+
        <figcaption class="figure-caption figtext" style="text-align: center; padding-bottom: 2%; padding-left:20%;padding-right:20%"> Figure 4. Plot of experimental product concentration Ĉ against time (red) and a modelled curve using estimated Michaelis-Menten kinetic parameters to best fit the experimental data (blue). The values of the estimated kinetic parameters obtained by the algorithm are also displayed in the plot. K<sub>M</sub> = 20.7842 µM &#177; 3.5264 with standard diviation.</figcaption>
 +
      </figure>
 +
</html>
 +
<html>
 +
<div class="col-xs-10">
 +
      <figure class="figure">
 +
        <img src="https://static.igem.org/mediawiki/2017/b/bf/CraftingCrocinWellStep2.png" class="figure-img img-fluid" style="display: block; margin: auto; width: 65%; height: auto; padding-top: 5%;">
 +
        <figcaption class="figure-caption figtext" style="padding-left: 20%; padding-right: 20%; padding-bottom: 3%;"> Figure 5. 96-well plate for activity measurements post 24 hours reaction and plate reading. The plate include pooled enzyme fractions 10–15 + substrate crocetin dialdehyde (well 1), pooled enzyme fractions 16–23 + crocetin dialdehyde (well 2), flow through + crocetin dialdehyde (well 3), negative control with only crocetin dialdehyde (well 4), pooled enzyme fractions 10-15 + product crocetin (well 5) pooled enzyme fractions 16–23 + crocetin (well 6), flow through + crocetin (well 7) and positive control with only crocetin (well 8).
 +
</figcaption>
 +
      </figure>
 +
    </div>
 +
</html>
  
 
===Modeling of CsADH2946===
 
===Modeling of CsADH2946===
Since the enzyme was previously poorly characterized, an homology model for CsADH2946 was created using the model PDB entry 4fqf and a stability simulation was performed to verify that our model was good. The homology modeling showed that CsADH2946 is homo-tetrameric. Steered molecular dynamics (pulling) was performed between the enzyme active site and its substrate crocetin dialdehyde in order to estimate binding energy and calculate a theoretical Kd (=4.9321 µM). The resulting structure of the homology modeling and a plot of the pulling simulation can be seen in figure 5.
+
Since the enzyme was previously poorly characterized, an homology model for CsADH2946 was created using the model PDB entry 4fqf (figure 7) and a stability simulation was performed to verify that our model was good (figure 6). For the template used (PDB entry 4fqf) the GMQE was 0.77 and the QMEAN was -0.24 for the model. The homology modeling showed that CsADH2946 is homo-tetrameric. Steered molecular dynamics (pulling) was performed between the enzyme active site and its substrate crocetin dialdehyde in order to estimate binding energy and calculate a theoretical K<sub>d</sub> (=4.9321 µM). The pulling simulation was performed on with both crocetin dialdehyde and acetaldehyde (figure 8). For crocetin dialdehyde there are two peaks in the pulling force in figure 6 which indicates energy barriers where more force is required to pull crocetin dialdehyde out. For acetaldehyde there is no clear peak in the same figure, indicating that our model is not able to bind acetaldehyde as strongly as crocetin dialdehyde.
 
<html>
 
<html>
 
<div class="col-xs-10">
 
<div class="col-xs-10">
 
       <figure class="figure">
 
       <figure class="figure">
 
         <img src="https://static.igem.org/mediawiki/2017/d/df/CraftingCrocinModelingStep2.png" class="figure-img img-fluid" style="display: block; margin: auto; width: 65%; height: auto; padding-top: 5%;">
 
         <img src="https://static.igem.org/mediawiki/2017/d/df/CraftingCrocinModelingStep2.png" class="figure-img img-fluid" style="display: block; margin: auto; width: 65%; height: auto; padding-top: 5%;">
         <figcaption class="figure-caption figtext" style="padding-left: 20%; padding-right: 20%; padding-bottom: 3%;"> Figure 5. Homology model of CsADH2946 and a plot demonstrating the pulling of the substrate crocetin dialdehyde from the active site of CsADH2946.
+
         <figcaption class="figure-caption figtext" style="padding-left: 20%; padding-right: 20%; padding-bottom: 3%;"> Figure 6. Homology model of CsADH2946 and a plot demonstrating the pulling of the substrate crocetin dialdehyde from the active site of CsADH2946.
 
</figcaption>
 
</figcaption>
 
       </figure>
 
       </figure>
Line 68: Line 97:
 
</html>
 
</html>
  
 +
 +
<html>
 +
<video width="auto" height="345vh" autoplay loop class='center'>
 +
<source src="https://static.igem.org/mediawiki/2017/4/40/Uppsala-CsADH2946.mp4" type="video/mp4"></source>
 +
Your browser does ot support video.
 +
</video>
 +
<div class="figtext" style="padding-bottom: 2%; padding-left: 20%; padding-right: 20%;"> Figure 7. This video is the result of simulating CsADH2946 in saline water for 100 ns using GROMACS.</div>
 +
</html>
 +
 +
<html>
 +
<video width="auto" height="345vh" autoplay loop style="padding-left:18%">
 +
<source src="https://static.igem.org/mediawiki/2017/c/ce/Uppsala-Pull_crocetin_dialdehyde.mp4" type="video/mp4"></source>
 +
Your browser does ot support video.
 +
</video>
 +
<video width="auto" height="345vh" autoplay loop>
 +
<source src="https://static.igem.org/mediawiki/2017/5/58/Uppsala-Pull_Acetaldehyde.mp4" type="video/mp4"></source>
 +
Your browser does ot support video.
 +
</video>
 +
<div class="figtext" style="text-align: center; padding-bottom:2%; padding-left:20%;padding-right:20%"> Figure 8. Results of the pulling simulations of one subunit of CsADH2946 with crocetin dialdehyde (leftmost) and acetaldehyde (rightmost). The molecules were pulled with a velocity of 0.25 nm/ns.</div>
 +
</html>
  
  

Latest revision as of 02:40, 2 November 2017


CsADH2946 under control with BBa_J04500

This BioBrick contains the gene coding for the aldehyde dehydrogenase called CsADH2946, which is under regulation with BBa_J04500. This enzyme is a part of the second step in the zeaxanthin to crocin pathway. More specifically it catalyzes the reaction from crocetin dialdehyde to crocetin using NAD+ as a cofactor. The enzyme can be found naturally in Crocus Sativus (the plant that saffron is harvested from). This BioBrick is confirmed with sequencing, purification and activity.

Since the gene of interest is under regulation BBa_J04500 it has to be induced with lactose or any closely related derivate such as IPTG. In our project we used IPTG to induce overexpression of CsADH2946.

Due the enzyme's previously poor charactarization, a homology model has been created for this part. That model was then used in production molecular dynamics simulations and steered molecular dynamics simulations to verify that CsADH2946 has specificity towards crocetin dialdehyde.

For more information check out our wiki here: http://2017.igem.org/Team:Uppsala.

Usage and Biology

Saffron, a well recognized but expensive spice, has not only uses in terms of cooking but compounds found in saffron have been shown to help with inflammation (1), neurodegenerative diseases (2) and more. Some of those compounds namely zeaxanthin, crocetin dialdehyde, crocetin and crocin are all a part of the same metabolic pathway in the plant specie Crocus Sativus. Not only are these compounds in saffron helpful in terms of their potential medicinal properties, but also the fact that they are very colorful makes them interesting as organic dyes for industrial purposes. These aspects are what drew us at iGEM Uppsala 2017 to work with the pathway from zeaxanthin to crocin in the BioBrick format, but also to integrate the metabolic steps in the pathway from farnesyl pyrophospate (FPP) to zeaxanthin on the chromosome of Escherichia Coli. The enzyme presented on this page, CsADH2946 catalyzes the second reaction in the zeaxanthin to crocin pathway.

In more detail CsADH2946 is an aldehyde dehydrogenase (ALDH) that oxidizes the two aldehyde groups at each end of crocetin dialdehyde to carboxylic acids using NAD+ as a cofactor. The resulting molecule from this reaction is crocetin. CsADH2946 was discovered through transcriptomic analysis of the chromoplasts of Crocus Sativus (3). The active site of CsADH2946 can be found around a loop containing three cystenin residues in a row (C337, C338, C339; positions were determined from the homology model). The residues that are conserved were found by looking at the template (PDB: 4fqf) (4) used in the homology modeling.


Sequence and Features


Assembly Compatibility:
  • 10
    COMPATIBLE WITH RFC[10]
  • 12
    COMPATIBLE WITH RFC[12]
  • 21
    INCOMPATIBLE WITH RFC[21]
    Illegal BamHI site found at 360
  • 23
    COMPATIBLE WITH RFC[23]
  • 25
    INCOMPATIBLE WITH RFC[25]
    Illegal NgoMIV site found at 767
  • 1000
    INCOMPATIBLE WITH RFC[1000]
    Illegal BsaI site found at 1284
    Illegal SapI.rc site found at 392

Characterization

iGEM Uppsala 2017 are the first to express and characterize CsADH2946 (Crocus Sativus aldehyde dehydrogenase 2946).

For purification CsADH2946 was transformed and expressed in E. coli strain BL21(DE3*) and purified using IMAC on an ÄKTA protein purification system. A gradient of imidazole concentration from 20–500 mM was used. The peak pointed at by the arrow in the chromatogram (figure 1) indicates protein that elutes at high imidazole concentration, i.e the desired his-tagged CsADH2946. The purification was followed by SDS-PAGE to analyse the fractions, control purity and verify the protein product. In figure 2a the band at around 60 kDa in the crude pellet indicate an overexpression of a protein in that size range. In the SDS gel of fractions 16–26 collected between 115–145 mL elution volume (figure 2b) there is a strong band at 60 kDa corresponding to the molecular weight of CsADH2946, indicating that the protein was successfully overexpressed and well-separated.

Figure 1. Chromatogram from IMAC-purification of CsADH2946.
Figure 2. a) SDS-PAGE gel of from IMAC purification. 1) Crude pellet. 2) Pellet after lysis. 3) Supernatant after lysis. 4) Flow through. 5) Wash with buffer A. 6) PageRuler protein prestained ladder. b) SDS-PAGE gel from IMAC purification. Fractions 16-26 were collected between 115 and 145 ml elution volume. A band at about 60 kDa is clearly visible.

The purified CsADH2946 was activity verified in converting crocetin dialdehyde to crocetin. An activity measurement assay was performed on a plate reader measuring absorbance of the substrate and product of the reaction. For the experiment we used a 96-well plate in which we included wells with enzyme from pooled fractions + substrate, as well as positive and negative controls, see table 1 for the specifics.

Table 1. Content of wells used for activity measurement of CsADH2946.

As can be seen in the activity spectra (figure 3), the absorbance of the product crocetin increases over time in well 2 containing enzyme and the substrate crocetin dialdehyde. After 9 hours of reaction, the blue curve corresponding to the enzyme + substrate mixture has increased its absorbance in the exact range of the product. The negative and positive control curves look similar to time point zero, apart from some precipitation of product and substrate indicated by the decreased curves. This is an evidence that the enzyme is functional. Using this data, KM = 20.7842 µM ± 3.5264 was estimated. This estimation was done using a Bayesian inference algorithm, see figure 4 for the fitted curve and eperimental data.

In addition, in figure 5 we can see that well 2 containing enzyme and crocetin dialdehyde has changed color compared to the negative control, to become more yellow like the product crocetin in well 8. This also shows that CsADH2946 was produced and that it converts crocetin dialdehyde into crocetin.


Figure 3. Activity measurement curve. The dotted lines describe wells at the starting time and the fully drawn lines describe the absorbance after 9 hours. The blue lines indicate wells containing protein and crocetin dialdehyde, red lines describe positive control with only crocetin dialdehyde and black lines correspond to the negative control only containing the desired product crocetin.
Figure 4. Plot of experimental product concentration Ĉ against time (red) and a modelled curve using estimated Michaelis-Menten kinetic parameters to best fit the experimental data (blue). The values of the estimated kinetic parameters obtained by the algorithm are also displayed in the plot. KM = 20.7842 µM ± 3.5264 with standard diviation.
Figure 5. 96-well plate for activity measurements post 24 hours reaction and plate reading. The plate include pooled enzyme fractions 10–15 + substrate crocetin dialdehyde (well 1), pooled enzyme fractions 16–23 + crocetin dialdehyde (well 2), flow through + crocetin dialdehyde (well 3), negative control with only crocetin dialdehyde (well 4), pooled enzyme fractions 10-15 + product crocetin (well 5) pooled enzyme fractions 16–23 + crocetin (well 6), flow through + crocetin (well 7) and positive control with only crocetin (well 8).

Modeling of CsADH2946

Since the enzyme was previously poorly characterized, an homology model for CsADH2946 was created using the model PDB entry 4fqf (figure 7) and a stability simulation was performed to verify that our model was good (figure 6). For the template used (PDB entry 4fqf) the GMQE was 0.77 and the QMEAN was -0.24 for the model. The homology modeling showed that CsADH2946 is homo-tetrameric. Steered molecular dynamics (pulling) was performed between the enzyme active site and its substrate crocetin dialdehyde in order to estimate binding energy and calculate a theoretical Kd (=4.9321 µM). The pulling simulation was performed on with both crocetin dialdehyde and acetaldehyde (figure 8). For crocetin dialdehyde there are two peaks in the pulling force in figure 6 which indicates energy barriers where more force is required to pull crocetin dialdehyde out. For acetaldehyde there is no clear peak in the same figure, indicating that our model is not able to bind acetaldehyde as strongly as crocetin dialdehyde.

Figure 6. Homology model of CsADH2946 and a plot demonstrating the pulling of the substrate crocetin dialdehyde from the active site of CsADH2946.


Figure 7. This video is the result of simulating CsADH2946 in saline water for 100 ns using GROMACS.

Figure 8. Results of the pulling simulations of one subunit of CsADH2946 with crocetin dialdehyde (leftmost) and acetaldehyde (rightmost). The molecules were pulled with a velocity of 0.25 nm/ns.


References

1. Papandreou MA, Kanakis CD, Polissiou MG, Efthimiopoulos S, Cordopatis P, Margarity M, et al. Inhibitory Activity on Amyloid-β Aggregation and Antioxidant Properties of Crocus sativus Stigmas Extract and Its Crocin Constituents. J Agric Food Chem. 2006 Nov 1;54(23):8762–8.

2. Chen L, Qi Y, Yang X. Neuroprotective effects of crocin against oxidative stress induced by ischemia/reperfusion injury in rat retina. Ophthalmic Res. 2015;54(3):157–68.

3. Gómez-Gómez L, Parra-Vega V, Rivas-Sendra A, Seguí-Simarro JM, Molina RV, Pallotti C, et al. Unraveling Massive Crocins Transport and Accumulation through Proteome and Microscopy Tools during the Development of Saffron Stigma. Int J Mol Sci [Internet]. 2017 Jan 1;18(1). Available from: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5297711/

4. Lang BS, Gorren ACF, Oberdorfer G, Wenzl MV, Furdui CM, Poole LB, et al. Vascular Bioactivation of Nitroglycerin by Aldehyde Dehydrogenase-2. J Biol Chem. 2012 Nov 2;287(45):38124–34.