Oxygen Sensation: Into Thick Air

Abstract

Two recent studies show that a neural globin tunes oxygen responses in the nematode Caenorhabditis elegans. Analysis of wild C. elegans strains suggests that the commonly used Bristol strain may have adapted to life in the laboratory.

Each spring bar-headed geese leave their closely-related greylag geese behind at their winter feeding grounds in the lowlands of India to fly to their nesting grounds in Tibetan highlands. Bar-headed geese are the world’s highest-flying migrants. They have been spotted flying over the top of Mount Everest where the oxygen concentration is one third of that at sea level. The hemoglobin of the bar-headed goose has a much higher affinity for oxygen than that of their lowland relatives allowing the bar-headed geese to cope with life and flight at high altitudes in the Himalayas [1]. Two papers recently published in Neuron [2] and Nature [3] show that globin polymorphisms also shape behavioral responses to oxygen in the nematode Caenorhabditis elegans.

The favorite breeding grounds of C. elegans are found in the soil, compost and rotting fruit, where oxygen concentrations can vary from near hypoxia to atmospheric levels; the worm avoids both hypoxia and hyperoxia and seems to prefer oxygen concentrations ranging from 5 to 12% [4]. Because C. elegans lacks a specialized respiratory system or circulatory organs, it must rely on diffusion for gaseous exchange. Previous work showed that wild C. elegans strains collected from different locations around the world have markedly different responses to oxygen. The C. elegans laboratory strain N2, originally isolated from a compost heap in Bristol England, is relatively indifferent to high oxygen concentrations as long as food is plentiful. The Hawaiian strain CB4856, which was isolated from a pineapple field, avoids high oxygen concentrations in the presence of food, and tends to aggregate and forage at the border of bacterial lawns. Polymorphisms in the neuropeptide receptor, npr-1, are largely responsible for differences in the behavioral responses between Bristol and Hawaiian strains [5]. The Bristol strain has a high activity npr-1(215V) allele resulting in a weak response to oxygen and aversion to carbon dioxide, whereas the Hawaiian strain has a low activity npr-1(215F) allele resulting in an aversion to high oxygen levels and indifference to carbon dioxide.

Changing oxygen concentrations alters locomotion rate and turning behavior of the worm, both of which are likely to allow the animal to find its preferred concentration on an oxygen gradient. McGrath et al. [2] and Persson et al. [3] found that the Hawaiian strain increased their locomotion and turning rate in response to small increases in the oxygen concentration (19/20% to 21%), while the Bristol strain were relatively unresponsive to small shifts in oxygen concentration. Using recombinant inbred strains between the Hawaiian and Bristol relatives, the two groups identified the glb-5 locus as an additional factor required for the different behavioral responses. The glb-5 gene encodes a protein with a globin domain and is a member of a large superfamily (including hemoglobin and myoglobin) of heme-binding proteins that play roles in oxygen storage, transport and sensation. In strains carrying both the Hawaiian npr-1(215F) and glb-5 alleles, oxygen downshifts led to a marked increase in turning rate and a reduction of locomotion rate. Strains carrying the Hawaiian npr-1(215F) allele and the Bristol glb-5 allele showed only minimal changes in turning and locomotion rate.

The Bristol glb-5 locus contains an exon duplication that gives rise to splice variants encoding truncated versions of GLB-5. The Bristol glb-5 allele is recessive to the Hawaiian glb-5 allele, suggesting that the duplication reduces gene function. Absorption spectra of recombinant GLB-5 protein indicate that it can reversibly bind oxygen, like the human neuroglobins. The glb-5 gene is expressed in the URX, AQR/PQR and the BAG sensory neurons, which have been identified as the oxygen and carbon dioxide sensing neurons of the worm. A GFP-tagged version of GLB-5 is highly enriched to the sensory ending of these neurons, suggesting a direct role in sensory transduction. To analyze how glb-5 affects URX responses, both groups [2,3] used genetically encoded calcium indicators to measure intracellular calcium in the URX neurons. These analyses showed that changes in oxygen produced larger responses in the URX neurons of animals carrying the Hawaiian glb-5 allele than in those carrying Bristol glb-5. High GLB-5 activity thus sensitizes the response of the URX neurons to small changes in oxygen concentration, and may change the worm’s preferred oxygen habitat.

Previous work has shown that the URX and BAG neurons also express soluble guanylate cyclases that can bind oxygen [4,6]. Behavioral analysis and calcium imaging showed that the activation of URX neurons by increases in oxygen levels require the soluble guanylate cyclases encoded by the genes gcy-35 and gcy-36 [3,7]. The BAG neurons on the other hand are activated by a decrease in oxygen levels and require the soluble guanylate cyclases encoded by gcy-31 and gcy-33 [7]. Surprisingly, Zimmer et al. [7] found that expression of the URX-specific gcy-35 and gcy-36 in the BAG neurons in a gcy-31; gcy-33 mutant background caused the BAG neurons to respond to increases in oxygen concentration much as do URX neurons. This suggests that soluble guanylate cyclases can act as ‘instructive sensors’ which increase cGMP production in response to downshifts (gcy-31/gcy-33) or upshifts (gcy-35/gcy-36) to allow opening of the cGMP-cation channel TAX-4.

It remains to be established whether GLB-5 sensitizes oxygen responses in the BAG cells, but an intricate picture for oxygen sensation and homeostasis is emerging that requires multiple sensory neurons, soluble guanylate cyclases, neuroglobins and peptidergic modulation. GLB-5 is just one of the 33 globin-domain containing proteins encoded in the C. elegans genome [8]. Most of them have distinct neuronal expression patterns and several are upregulated under anoxic conditions [9]. While the function of mammalian neuroglobins and cytoglobins is largely unknown, the characterization of glb-5 points to a critical role as cellular oxygen sensors.

Much like polymorphisms in the hemoglobin of the bar-headed goose, polymorphism in glb-5 and npr-1 may have allowed wild C. elegans strains to adjust their oxygen preferences and adapt to their respective ecological niches. Closely related Caenorhabditis species and the vast majority of wild C. elegans strains carry both the Hawaiian glb-5 and npr-1 alleles in their genetic arsenal, suggesting that the Bristol glb-5 and npr-1 polymorphisms arose recently [2,3]. Surprisingly, high density SNP analysis using 1454 markers spanning all chromosomes showed that the few ‘wild’ Bristol-like outliers collected from the far corners of the globe have haplotypes that are astonishingly similar to the prevalent Bristol strain, raising the suspicion that these ‘wild’ strains could be lab bench contaminants [2].

What niche might fit the Bristol glb-5 and npr-1 alleles? A persuasive case has been made that the Bristol glb-5 and npr-1 alleles may have originated in the laboratory (Figure 1) [2,10]. The Bristol strain was cultivated on agar plates for almost two decades, which amounts to approximately 1800 generations, before frozen cultures were established — on a human evolutionary time-scale the Neanderthals were making compost heaps in Bristol. Most telling is the existence of a second strain (LSJ1) derived from the laboratory that propagated the Bristol strain before it became famous in the hands of Sydney Brenner. The LSJ1 strain was maintained in liquid cultures (low oxygen) for many generations and is identical to the Bristol strain at 1453 of 1454 SNPs, but harbors the Hawaiian glb-5 and npr-1 alleles. This suggests that the Bristol glb-5 and npr-1 alleles arose after the Bristol and LSJ1 strains separated in the laboratory. The Bristol glb-5 and npr-1 alleles modify behaviors at 21% oxygen and thus may confer a selective advantage to life at high oxygen levels on agar plates compared to the life in the compost heap.

Figure 1. C. elegans ’ journey from the compost heap to the agar plate has subjected it to different selective pressure.

Polymorphisms in the neuropeptide receptor gene npr-1 and the neural globin gene glb-5 may have provided the common laboratory strain Bristol N2 a selective advantage to life at high oxygen levels.

Although genetic selection for behavioral traits may not come as a surprise for any one who eats or pets domesticated animals, this presents a note of caution for those of us who nematomorphize C. elegans behavior. Spending many generations on a slab of agar feeding on copious amounts of E. coli (a bacterial species not usually on its menu), the laboratory Bristol strain undoubtedly has been subjected to different selective pressure. For future neuroethological interpretations of C. elegans behaviors, including foraging strategies, mating and feeding, it would be worthwhile to characterize some of these behaviors in wild strains that are less far removed from the compost heap.