Evolution and development of the Tetrapod Limb

John Merck

Download cheat-sheet

Review: Limb and girdle elements

This should look familiar. Test yourself to make sure you are comfortable identifying elements of the tetrapod forelimb.

But don't forget the hindlimbs:

Eryops sp. right forelimb in anterior view

Forelimb details:

Now some detail:

Manus: the technical term for the "hand."

Tetrapod digits are numbered from I to V (with Roman numerals), with I being the most medial when the manus is in quadruped anatomical position. (E.G.: the human thumb.)

Although there is significant variation over the course of evolution, the tetrapod carpus - the set of carpal elements - ancestrally conforms to a strict pattern where the carpals occupy two distinct rows plus a few:

Eryops sp. right hindlimb in anterior view

Hindlimb details:

Pes: the technical term for the "foot."

The ancestral pattern of the tetrapod tarsus is similar:

Dimetrodon sp. forelimb (above) and hindlimb (below)

Limb regions:

Because the overall patterns of the fore and hindlimbs are so similar ancestrally, it is convenient to refer to their general regions collectively. Thus, each fore and hindlimb has a:

Eusthenopteron foordi forelimb

Development and Evolution

How did this clear and highly conserved pattern evolve? Now we have a problem. In the skeleton of the pectoral fin of Eusthenopteron foordi, an aquatic non-tetrapod we see homologs to some of the elements of the tetrapod forelimb. (See if you can identify.) However the different regions are not strongly differentiated. It gets even worse when we look at more primitive sarcopterygians:

Shubin and Alberch, 1986
chondrification schematic.
Preaxial - left, postaxial - right.

Development and Evolution

Of course, this in no way resembles the forelimb skeleton of actinopterygians. The most we can say is that sarcopterygian pectoral fins seem to have:

The identification of homologs to the non-tetrapod metapterygial axis has been a research priority for over a century, but the breakthrough came with Shubin and Alberch, 1986. They noted that:

Thus, the autopodium encompasses a reversal from preaxial to postaxial segmenting fin rays.

One odd thing: Shubin and Alberch's scheme didn't seem to accommodate digit V, which seems to form separately. How does it relate to the digital arch?

The limb bud
Genomics: Over the last 20 years, identifying the genetic controls over limb element formation has made significant progress but is still preliminary. Our very few research models include:

All gnathostomes surveyed organize their developing limb buds similarly:

The puzzle is by no means assembled. What follows is an inventory of the pieces:

The Zone of Polarizing Activity: Reviewed by Shubin, 2008. (Discuss and review.)

Limb segment identities: Hox genes 9 - 13 are expressed in tetrapod limbs. Remember:

The actions of Hox clusters a and d are best studied.

From Shaking the Tree: Readings from Nature in the History of Life. Right forelimbs in dorsal view.
Step 1: Defining the autopodium: Hoxa gets us started. These are Hox genes that would typically govern segmentation in the posterior trunk, but have "doubled back" to be expressed in the limb buds.

What about Hoxd? In aquatic choanates, each non-terminal segment of the limb axis is followed by two elements, the next axial segment, and a preaxial (in front of the axis) radial branch. The result is:

shown in:
  1. Cladoselache (a primitive chondrichthyan)
  2. Neoceratodus (the living Australian lungfish)
  3. Sauripteris (a primitive sarcopterygian)
  4. Panderichthys (a choanate)
  5. Tulerpeton (a stegocephalian)
In Panderichthys (D) this is reduced to a very simple form. The limb axis is coded by Hoxd11 (light shading in f and g), the radials by Hoxd13 (dark shading in f and g). What is lacking is the hand or foot - the autopodium.

Schematics F and G show the expression of the HOX genes in the embryonic limb buds of zebra-fish (with no autopodium) and land vertebrates (with an autopodium) respectively. Significantly the domains of the Hoxd11 and Hoxd13 genes are reversed at the end of the land vertebrate limb - corresponding to the shift from pre- to postaxial bifurcations. This switch in Hox gene domain apparently defines the developmental identity of the autopodium.

From Woltering et al., 2014
Step two - the digits: But just having an autopodium doesn't automatically give us digits. More recent work by Davis et al. 2007 has shown that the paddlefish Polyodon - a primitive actinopterygian - also has reversed HOX D-11 and HOX D-13 domains. Of course, it lacks digits. Woltering et al., 2014 demonstrated that in land vertebrates the expression of the Hoxd genes is promoted in the proximal and distal ends of the limb bud by promotors that are absent in fish. The combination of Hot genes and promotors results in the elongation of of skeletal elements proximal to the wrist and distal to it.

The result:

This pattern is congruent with the zones of activity of Hoxa. In non-tetrapods, the domain of Hoxa11 overlaps that of Hoxa13 all the way to the end of the fin. In tetrapods, the overlap is reduced to the zone in which carpals and tarsals form.

As some non-tetrapods show, the coordination of Hox domains and promotors may not have evolved lockstep.

Step three - stopping the formation of digits:

Hindlimb of Ichthyostega
The earliest fossils of creatures with a digital arch reveal a developmental exuberance unseen today. Just as nothing prevented the lungfish from having many bifurcations between axial and postaxial elements, nothing limits the number of digital arch bifurcations to five. The number and identity of digits results from interactions of Sonic hedgehog, which: That interaction was fine-tuned over tens of millions of years resulting in the stabilization of digit number at five. (Four fingers for amphibians.)

Final enigmas:

What happened to the fin rays? Good question. Fin rays arise from neural crest ectoderm that condenses in the apical fin fold of the embryonic limb bud in creatures that have them. So far, all fossil stegocephalians have either fin rays or digits on their paired limbs, but never both. Nakamura et al. 2016 have show that fin rays and digits arise from the same precursors, but that the development of fin rays requires the actiona of an autodial HOX13 enhancer.

What about the space between the fingers? The presence of bony digits doesn't guarantee the presence of separate fleshy fingers and toes. To make these requires apoptosis - programmed cell death - of the cells between the digits in the growing limb bud. Where in tetrapod evolution that ability arose is unclear.

Law of similarity? Generally speaking the fore and hindlimbs of tetrapods are similar because their growth is governed by the same developmental mechanisms. Nevertheless, in both evolution and ontogeny, the forelimbs seem to lead the hindlimbs somewhat. Might we eventually encounter a fossil sarcopterygian with digits on its forelimbs and fin rays on its hindlimbs?

What about digit V? Recall that digit V does not seem to be part of the digital arch. Genomics support this inasmuch as experiments in which Hoxa13 and Hoxd13 are made inoperative result in the absence or malformation of the digital arch, but do not prevent that formation of digit V. Carroll 2009 notes the presence of isolated postaxial fin rays in non-tetrapods like Eusthenopteron (above) and Tiktaalik. Could one of these be homologous to digit V? Why would it look like a digit? Maybe because it was near the zone of expression of the protein Fgf8 (fibroblast growth factor 8) that induces segmentation of digits to produce a sequence of phalanges.

Additional reading: