Previous studies have clearly shown that Prod1 is likely a membrane-bound protein that acts as a receptor. Naturally, if something is a receptor it must have something that binds to it or acts as ligand. The main purpose of these experiments was to identify exactly what binds to Prod1 in hopes that this molecule plays a role in limb regeneration.
Criteria for candidates
Determining what binds to a receptor can be an arduous process, especially in this case, where the authors had no clear leads. However, Prod1 being a membrane-bound protein involved in limb regeneration does narrow down the possibilities. The authors needed to find a protein that was:
1. Secreted: For that protein to bind to Prod1 it must be secreted to the outside environment or else it would never reach Prod1. Cytoplasmic proteins would not be good candidates.
2. Expressed during limb regeneration: A good candidate is likely expressed in higher levels during limb regeneration as opposed to normal homeostasis.
Although these criteria do make things slightly easier, the number of proteins that could bind to Prod1 was still enormous. To find the ligands or proteins that bind to Prod1 they needed to perform some sort of screen. That is, they needed to look at a large number of potential hits in the simplest fashion possible. The most common technique when dealing with this type of problem is the yeast two-hybrid screen (Y2H).
In short, the Y2H screen allows one to determine what other proteins bind to your protein of interest. In this case, what proteins bind to Prod1. The amazing thing about this technique is that an extremely large number of proteins can be screened. Theoretically, all proteins expressed in the salamander could be screened to see if they interact or bind with Prod1.
In Y2H, yeast are normally transformed with two plasmids. One plasmid expresses the “bait” protein (e.g., Prod1) and the other expresses the “prey” (e.g., whatever binds to Prod1). What makes the system work is that the bait and prey are fused to other proteins. That is, they are combined with another protein to make a longer one with new functions. The bait is fused to a DNA-binding domain and the prey is fused to an activator domain. That means that the bait will now bind to a certain region of DNA when it normally wouldn’t. Similarly, the prey wouldn’t normally activate a gene, but now, with the activator domain fused, it will. The magic happens when the bait and prey bind together. The activator domain-prey protein is not usually within reach of the gene it activates, however when it binds to the bait it comes into close proximity with that gene and is able to activate it. Remember, the bait with its DNA-binding domain is bound to a predetermined region of DNA. The gene activated by this complex is usually some sort of reporter that either provides antibiotic resistance or produces some color in the presence of chemicals. For a diagram of this, please see: http://www.wormbook.org/chapters/www_biochemmolbio/biochemfig1.jpg
DB is the DNA-binding domain.
AD is the activating domain.
X is the bait protein fused to the DNA-binding domain.
Y is the prey protein fused to the activator domain.
Using special molecular cloning techniques, scientists are able to create many plasmids expressing different prey proteins. By these means they are able to screen a large number of proteins to find out which ones have interactions with Prod1 using their Y2H screen.
For an animation overview of Y2H screens, see: http://www.sumanasinc.com/webcontent/animations/content/yeasttwohybrid.html
Y2H screens are usually an enormous amount of work, but do not produce any data that look nice in a scientific publication. As a result, the authors show only the main finding, that the protein nAG appears to bind to Prod1.
They show a picture of a petri dish coated with media (food for the yeast) on which they grow. Three streaks or lines of yeast are shown, each one of a different genotype: nAG plasmid only, Prod1 plasmid only, and both nAG and Prod1 plasmids.
Because both prey and bait are needed to activate the reporter gene, only the yeast with both plasmids are able to do that. In this case, the yeast turn blue in the presence of certain chemicals and activation of the reporter gene.
Now that nAG has been identified as a likely binding partner for Prod1, the authors wanted to see whether other vertebrates also have proteins similar to nAG. You can imagine that findings from this would be highly relevant to human health. If humans have a protein similar to nAG perhaps we could possibly regenerate limbs, too!
The authors produced a phylogenetic tree comparing the relatedness of nAG with proteins found in other organisms and saw that many other organisms appear to have proteins very similar to nAG.
Coimmunoprecipitation and immunoblotting
Although the Y2H screen provides good evidence that Prod1 binds to nAG, it is still important to confirm these results by other means. The Y2H screen is prone to false-positive results. Here the authors check to see whether these proteins come out of solution together (coimmunoprecipitate), which would further suggest that they bind each other.
To do this, the authors expressed nAG or Prod1 in bacterial cells and then “lysed” or broke those cells open. Now, for all proteins, DNA and RNA are free floating in solution because there are no longer cell walls and plasma membranes keeping them in.
This mixture, called lysate, is then mixed with small beads. These beads are unique though in that they are covered in nAG or Prod1 protein. Thus, if nAG in the lysate does bind to Prod1 we would expect nAG to be bound to the Prod1 beads. Next, centrifugal force can then be applied to separate the solution by size. The bead-Prod1-nAG complex will be at the bottom of the tube and the rest of lysate can be thrown away.
Finally, the authors can perform an immunoblot, or western, to separate the proteins by size and to further identify them using antibodies. A video explaining immunoblots can be found at: https://www.youtube.com/watch?v=v-O103PLhm8
An additional video can be found at: https://www.thermofisher.com/us/en/home/life-science/protein-biology/pro…
The authors performed coimmunoprecipitation and immunoblotting as previously described.
Lane 0/Far-left lane: This is the “ladder” for the blot. One of the purposes of immunoblotting is to separate proteins by size, so you need some way to determine the size of your proteins. A ladder contains predetermined proteins of various sizes. Because you know the sizes of the proteins in the ladder ahead of time, you can compare the protein you are studying with the ladder. This is how the authors generate the kilodalton sizes of proteins (25, 37, 50, etc.) next to the blot.
Lane 1: This lane had CTFG beads run by themselves. Because they blotted for antimyc we would expect there to be no nAG present, but because CTGF has a myc tag added to it we see a band between 37 and 50 kilodaltons. Two bands that are likely immunoglobulins are also present. These proteins are a part of the beads and are an artifact of using this method and can be disregarded.
On the whole, this lane acts as a negative control. As expected, no protein of the right size is detected on the blot.
Lane 2: This lane had nAG-bound beads run alone. This lane acts as a positive control because we will expect there to be nAG present in this lane regardless of bead binding. When probing for antimyc (the nAG protein is fused to a “myc” tag), we see a band show up that is indicated by the arrow.
Lane 3: Control beads were mixed with lysate from cells expressing Prod1. These control beads should not pick up any nAG because it is not present in the lysate or bound to the bead. As a result, there is no nAG band present. This acts as an additional negative control.
Lane 4: CTGF beads were incubated with Prod1 lysate. Because nAG is not present in this lysate we still do not see a band correlating nAG. However, there is a return of the CTGF band between 37 and 50 kilodalton that was present in Lane 1.
Lane 5: nAG beads were mixed with Prod1 lysate. In this lane we can see the appearance of a band correlating to nAG as indicated by the arrow. This is expected because the bead is coated in nAG protein. However, if you look at the blot on the right where they performed the same experiment, but used an antibody toward Prod1 instead of nAG (antimyc), you can see there is a band for Prod1 present. This shows that Prod1 did in fact bind to nAG beads.
The authors wanted to determine whether their myc-nAG fusion protein is secreted as with the wild-type nonfusion protein. Results from this study played a large role with how the authors performed experiments for Figure 6.
Left lane: Ladder indicating the size of proteins.
Middle lane: Medium from Cos7 cells expressing myc-nAG. Because nAG is a secreted protein we would expect these cells to release nAG out into the environment. In this instance, the medium these cells are grown in are the outside environment. When this medium is immunoblotted we see reactivity as indicated by the arrow.
Right lane: Medium from Cos7 cells expressing red fluorescent protein. This acts as a negative control.
Mouse PS cells had a plasmid introduced so that they express Prod1 protein on the surface of their plasma membrane. nAG was then added to the medium these cells were growing in and interactions between the two proteins were allowed to happen.
These cells were then washed multiple times such that only nAG bound to Prod1 would be leftover. Any unbound nAG would be removed by the washing steps.
Antibodies toward nAG were then added to the medium where they could then bind to the nAG protein. These antibodies had enzymes fused to themselves, which allows for a blue-purple color to show up in the presence of certain chemicals.
Thus, only areas where there are bound antibodies to nAG will you see color. Any unbound antibodies, like unbound nAG, will be washed away.
All these experiments have shown that nAG is likely a ligand or binding partner for Prod1. Through multiple experiments and techniques the authors have supplied good evidence for follow-up studies of nAG in the living organism.
When and where?
nAG is now a reasonable candidate for binding to Prod1. The next logical experiments would be to determine when and where we see nAG expressed. If nAG is expressed somewhere odd, like in the intestines, but not the regenerating limb, it may not be a good protein to study further.
Very much like immunoblotting, immunohistochemistry relies on antibodies. However, unlike immunoblotting, the secondary antibody instead of being fused to horseradish peroxidase or a similar enzyme, is fused to a fluorophore. This fluorophore, when hit with a certain wavelength of light, emits a fluorescent signal. Through these means one can determine where the antibody is bound when viewed under a microscope by looking for the light emitted.
Also, because tissues from animals are usually very thick, they have to be cut into thin slices or “sectioned.” These sections can then be laid down on microscope slides and stained using the antibodies.
Through these means scientists are able to visualize proteins and their dynamic lives.
For an image of antibody-binding during immunohistochemistry, see: http://sites.lafayette.edu/neur401-sp10/files/2010/04/untitled.jpg
For an explanation of how fluorophores work, see: https://www.thermofisher.com/us/en/home/life-science/protein-biology/pro…
For a video demonstrating this protocol, see: https://www.youtube.com/watch?v=5AcrhhHmQZs
The authors wanted to know what happens to nAG expression when the salamander limb is amputated. Experiments shown in the supplementary figures demonstrated that there is low/no expression of nAG in the intact limb.
Figure 2A clearly shows a dramatic increase in protein levels of nAG as evidenced by the green staining seen in the picture. More nAG protein is present, so a great amount of antibodies bind to it, causing an increase in fluorescence intensity. The results of this experiment show that the nerve sheath is highly stained.
Since the previous figure showed nAG expression in the nerve sheath during amputation the authors wanted to determine what specific cells in the nerve sheath exactly express nAG. To do this they use a Schwann cell marker HNK1. They performed immunohistochemistry using their HNK1 antibody, but the color fluorophore they use is different. Instead of green color, they use red fluorophore. This allows them to visualize both HNK1 and nAG in the same sample. Furthermore, they use a DNA stain that fluoresces blue.
The results from this experiment show that the Schwann cells are stained green in this nerve sheath cross section. However, the axons do not appear to stain. Of note is that this experiment was performed 10 days postamputation.
Experiments were performed at later time points postamputation and the authors noticed expression in the glands of the wound epidermis. This was not seen in earlier time points. As a result, this suggest that expression of nAG starts in the nerve sheath and after time it starts being expressed in the wound epidermis.
Using immunohistochemistry the authors were able to show expression and localization of nAG in the regenerating limb. They initially saw expression only in the Schwann cells of the nerve sheath. Though, after more time passed they saw expression of nAG in gland cells underlying the wound epidermis as well.
These results suggest that nAG likely plays some role in nerve-dependent regeneration. The fact that it is found in cells closely affiliated with nerve and the wound epidermis suggests a role in this process.
As stated before, salamanders are unable to regenerate their limbs when they are denervated. Consequently, the molecules that underlie this are likely not present when the nerve goes away. No nerve, no molecule. Becauase nAG is likely, the authors wanted to determine what exactly happens to nAG expression when you denervate the animal.
Panels A and B
This figure shows immunohistochemistry using an antibody toward nAG. Panel A shows the nerve sheath of an animal with an intact nerve and Figure B shows the other side of the animal, which has been denervated.
Panels C and D
These panels are very similar to Figure 3A and B, but this time they focus on the wound epidermis. The top figure shows the control limb and the bottom panel shows the denervated limb on the same animal. No expression of nAG is seen in the wound epidermis of the denervated limb.
These results show that nAG, itself, is nerve dependent. When the limb is denervated expression dramatically decreases compared with the control limbs. These data further suggest that nAG is the likely molecule underlying nerve-dependent regeneration.
nAG to the rescue
The authors now believe that nAG is probably an important molecule in limb regeneration. However, they still need to test this idea. To do this they, they denervate the salamander, amputate, and then exogenously add back in nAG. Remember, nAG is not present in the denervated limb as shown by Figure 3. In this sense the authors are performing a rescue experiment. In other words, they are trying to see whether adding back nAG will allow the limb to regenerate.
Electroporation is technique by which a scientist can cause a cell to take up DNA found in the external environment. Often times the scientist is trying to overexpress a protein or insert a DNA sequence not normally found in that organism.
This technique is actually very intuitively named. “Electro” is in reference to the electrical force applied to the cells during this technique. “Poration” refers to how this electrical current causes holes in the plasma membrane to form. The holes generated by electroporation are large enough for things like plasmids to enter, which is extremely useful for introducing exogenous DNA. Eventually, the holes of the cell close up and the alien DNA is now trapped inside the cell where it can instruct machinery to make mRNA and proteins.
The authors use a version of this technique called focal-point electroporation, which allows for electroporation of tissues and not individual cells. This technique is especially useful for electroporating tissues such as those that make up the regenerating limb.
For an animation of electroporation as it used in bacteria, see: http://www.auburn.edu/academic/classes/biol/3020/iActivities/ch08/NEanim…
Although this is different from focal-point electroporation, the principle is still the same.
Panels A, B, and C
These experiments show an excellent test for efficiency for the electroporation procedure used in these set of experiments. In figure 4A, the authors demonstrate that 30% to 50% of cells had and expressed their plasmid. Furthermore, they show that the red fluorescent protein expressed from this plasmid is still present 3 weeks later, demonstrating that expression is long-lived.
Figures 4B and C are very similar, but instead of electroporating a red fluorescent plasmid into the salamander, they instead use a nAG plasmid. They then follow up on nAG’s expression using immunohistochemistry to show where nAG is expressed. Taken together, these experiments demonstrate that focal-point electroporation is an efficient way to ectopically express proteins in salamanders.
This shows the setup for the experiments performed for Figure 4E.
Using the experimental setup from 4E, the authors show that electroporation of a nAG plasmid rescues the phenotype of denervated animal (i.e., they regenerate limbs).
Left salamander: Both limbs are amputated, but only the right limb is denervated. The right limb is electroporated with a control plasmid. As expected, this animal is able to regenerate its left limb, but not its denervated right limb.
Right salamander: Both limbs are amputated, but only the right limb is denervated. The right limb is electroporated with a nAG plasmid. The only difference between the right salamander and the left is what is electroporated. In this case, electroporation of nAG is able to rescue the phenotype and one can clearly see the regenerated limb on the right side of the animal.
The authors have demonstrated that expression of nAG in denervated limbs is able to allow for limb regeneration. This is an important finding that continues to hammer away at the idea that nAG underlies the molecular mechanisms underpinning nerve-dependent regeneration. Here the authors show that nAG is necessary for proper limb regeneration.
Necessary vs. sufficient
Although the authors have shown that nAG is necessary for proper limb regeneration, they haven’t shown whether nAG by itself is sufficient to regenerate it properly.
That is, they want to know whether there are any abnormalities in the denervated limb rescued with nAG.
Panels A and B
Immunohistochemistry for acetylated-tubulin is performed on normal (A) and rescued (B) limbs.
In the limbs rescued with nAG electroporation, there appears to be much less acetylated tubulin staining, which shows that there is much less peripheral nervous system tissue in the nAG rescued limb.
Panels C and D
A similar experiment to Figure 5A and B was performed but with an antibody toward myosin. This stains skeletal muscle in the samples.
Like Figure 5B, much less staining is found in the rescued limbs.
This demonstrates that skeletal muscle was not regenerated to as great a level as in a normal regenerated limb.
Clearly, there are some major differences between regeneration in the normal limb versus the nAG rescued limb. At the least, in regards to nervous and muscle tissue.
Of note, results from these experiments also show that nAG works by some means other than simply regenerating the nervous tissue.
If it did regenerate the nerves, those nerves could potentially then send the signals for a correct regenerative response.
The muscle follows along with this because it, too, is nerve-dependent in regard to regeneration.
If the nerve is severed or destroyed, the skeletal muscle will be unable to regenerate.
Likewise, the results we see in this experiment point to lack of proper nerve formation, which leads to lack of skeletal muscle regeneration.
nAG has now been shown to be an extremely important molecule in the regenerating limb.
Addition of nAG is able to rescue the phenotype of denervated limb and allow for it to regenerate.
Now that the role of nAG in limb regeneration has been well-studied, what is left is finding out the mechanism behind how nAG works.
Primary cell culture
To determine the mechanism of nAG, the authors use a technique called primary cell culture.
This technique takes cells directly from the animal they are studying where they can then be grown in a petri dish or flask.
Unlike normal cell culture, though, these cells are usually not immortal. In other words, they will eventually cease to grow or die off.
The benefit of this though is that these cells are usually much more biologically relevant to what is going on in the living organism compared with an immortal cell line.
This picture simply shows the cells growing in culture. Notice that they are for the most part separated from each other.
These cells were taken from the limb blastema and chemically separated or “dispersed.”
Whenever new DNA is synthesized, new nucleotides are used to synthesize the strands.
BrdU is an artificial nucleotide similar to thymidine. It is useful because it can be artificially added to cells to measure how many cells are making new strands of DNA.
Usually fluorescent antibodies, as in immunohistochemistry, are used to visualize BrdU nucleotides incorporated into the strands of DNA. Through these means one is able to determine how many cells are getting ready to divide. A nondividing cell is not synthesizing new DNA and would not incorporate the BrdU nucleotides.
Panels B and C
Primary limb blastemal cells were cultured in either control medium or nAG containing medium and pulsed with BrdU. Cells with the nAG containing medium were significantly brighter in fluorescent signal compared with the control, demonstrating that these cells were actively preparing to divide.
With much more BrdU taken up in the nAG exposed cells, it appears as though the authors have a mechanism behind nAG’s function.
It appears as though nAG stimulates cells to grow and divide. Imagining this in the living organism, one can imagine that nAG signals cells of the amputated limb to start growing and dividing to reproduce the amputated limb.
Although these data will have to followed up in the actual salamander, it does leave a path down which the authors can continue their studies.
- J. P. Brockes, A. Kumar, Science 310, 1919 (2005).
- J. P. Brockes, Science 276, 81 (1997).
- M. Maden, Nature 295, 672 (1982).
- S. D. Thoms, D. L. Stocum, Dev. Biol. 103, 319 (1984).
- S. Morais da Silva, P. B. Gates, J. P. Brockes, Dev. Cell 3, 547 (2002).
- Kumar, P. B. Gates, J. P. Brockes, C.R. Biol. 330, 485 (2007).
- K. Echeverri, E. M. Tanaka, Dev. Biol. 279, 391 (2005).
- D. L. Stocum, Curr. Top. Microbiol. Immunol. 280, 1 (2004).
- M. Singer, Q. Rev. Biol. 27, 169 (1952).
- P. Brockes, Science 225, 1280 (1984).
- D. B. Drachman, M. Singer, Exp. Neurol. 32, 1 (1971).
- R. L. Sidman, M. Singer, Am. J. Physiol. 165, 257 (1951).
- W. Egar, Anat. Rec. 221, 550 (1988).
- T. Endo, S. V. Bryant, D. M. Gardiner, Dev. Biol. 270, 135 (2004).
- Singer, L. Craven, J. Exp. Zool. 108, 279 (1948).
- L. Mescher, J. Exp. Zool. 195, 117 (1976).
- J. P. Brockes, C. R. Kintner, Cell 45, 301 (1986).
- L. Wang, M. A. Marchionni, R. A. Tassava, J. Neurobiol. 43, 150 (2000).
- L. M. Mullen, S. V. Bryant, M. A. Torok, B. Blumberg, D. M. Gardiner, Development 122, 3487 (1996).
- L. Mescher, E. Connell, C. Hsu, C. Patel, B. Overton, Dev. Growth Differ. 39, 677 (1997).
- M. Globus, M. J. Smith, S. Vethamany-Globus, Ann. N. Y. Acad. Sci. 632, 396 (1991).
- F. Aberger, G. Weidinger, H. Grunz, K. Richter, Mech. Dev. 72, 115 (1998).
- Persson et al., Mol. Phylogenet. Evol. 36, 734 (2005).
- P. J. Adam et al., J. Biol. Chem. 278, 6482 (2003).
- D. A. Thompson, R. J. Weigel, Biochem. Biophys. Res. Commun. 251, 111 (1998).
- L. Iten, S. Bryant, Willhem Roux, Arch. Dev. Biol. 173, 263 (1973).
- J. D. Salley, R. A. Tassava, J. Exp. Zool. 215, 183 (1981).
- M. Carlson, Principles of Regenerative Biology (Elsevier Inc., London, 2007).
- F. V. Mariani, G. R. Martin, Nature 423, 319 (2003).
- L. Niswander, Nat. Rev. Genet. 4, 133 (2003).
- K. Crawford, D. L. Stocum, Development 102, 687 (1988).
- L. T. Pecorino, A. Entwistle, J. P. Brockes, Curr. Biol. 6, 563 (1996).
- G. C. Fletcher et al., Br. J. Cancer 88, 579 (2003).
- M. Rosel, C. Claas, S. Seiter, M. Herlevsen, M. Zoller, Oncogene 17, 1989 (1998).
- G. Lemke, Sci. STKE 2006, pe11 (2006).
- L. Yntema, J. Exp. Zool. 142, 423 (1959).
- S. Thornton, M. T. Thornton, J. Exp. Zool. 173, 293 (1970).
- R. J. Goss, Principles of Regeneration (Academic Press, New York, 1969).
- M. Singer, K. Rzehak, C. S. Maier, J. Exp. Zool. 166, 89 (1967). CrossRefMedlineWeb of Science
- We thank P. Driscoll, I. Gout, and P. Martin for their help and comments, and M. Larkum for fabrication of the electroporation electrodes. This work was supported by a Medical Research Council (UK) Research Professorship and Programme Grant to J.P.B.