What Sperm Can Teach us About Energy Production


Author’s address (for correspondence): Alexander J. Travis, Cornell Center for Wildlife Conservation, Baker Institute for Animal Health, Cornell University College of Veterinary Medicine, Hungerford Hill Road, Ithaca, NY 14853, USA. E-mail: ajt32@cornell.edu


Mammalian sperm have evolved under strict selection pressures that have resulted in a highly polarized and efficient design. A critical component of that design is the compartmentalization of specific metabolic pathways to specific regions of the cell. Although the restricted localization of mitochondria to the midpiece is the best known example of this design, the organization of the enzymes of glycolysis along the fibrous sheath is the primary focus of this review. Evolution of variants of these metabolic enzymes has allowed them to function when tethered, enabling localized energy production that is essential for sperm motility. We close by exploring how this design might be mimicked to provide an energy-producing platform technology for applications in nanobiotechnology.


Mammalian sperm are well known for their highly polarized structure, which has evolved under strong selective pressures. Sperm competition, occurring in mammals when a female mates with more than one male in a single oestrous cycle, has led to the evolutionary strategy of males having a great many, very small sperm (Parker 1982). To accomplish the goal of producing high numbers of sperm, males invest relatively little in each one, with sperm lacking many organelles common to somatic cells, including the endoplasmic reticulum, golgi and ribosomes, and having scant amounts of cytoplasm. Yet, the competition among sperm from different males in a single female tract exists beyond simple numbers, with sperm motility being a key attribute that distinguishes ‘winners’ from ‘losers’ (Birkhead et al. 1999).

Several species, including both eutherian and American marsupial mammals, have evolved cooperative cellular behaviours to influence this competition. It has long been known that the sperm of didelphid marsupials pair within the epididymis and that these paired sperm have advantages in motility over single spermatozoa (Biggers and Creed 1962). More recently, this phenomenon of sperm aggregation leading to competitive advantages in motility has been shown in rodent sperm that not only can distinguish conspecific from heterospecific sperm, but can also distinguish between sperm of sibling males (Fisher and Hoekstra 2010). This is an important point, because it suggests that the aggregation behaviour is not a generic population phenomenon, but is instead a competitive adaptation.

In this review, we focus not on cellular behaviours, but rather on the underlying cellular machinery that provides energy for sperm motility. This is a narrow focus, in that it includes neither the various regulators of motility such as levels of intracellular calcium, bicarbonate or cAMP, nor the effectors of motility—the microtubule doublets and associated proteins that make up the axoneme. However, the strong selective pressures that have shaped sperm structure have led to unusual metabolic adaptations that are of significance not just to reproductive biology, but to diverse fields such as nanobiotechnology.

Compartmentalized Metabolic Pathways

Sperm flagellar ultrastructure

Mammalian sperm have three distinct regions of the flagellum: the midpiece, the principal piece, and a short terminal endpiece [Please see (Inaba 2011) for a recent review.]. The axoneme runs centrally throughout the length of the flagellum, and surrounding cytoskeletal structures known as the outer dense fibres run varying lengths down the flagellum. The three main flagellar regions in sperm of many species are able to be distinguished morphologically using light microscopy because of the presence/absence of specific structures (Fig. 1). The midpiece is wider because of the presence of helically arranged mitochondria around the outer dense fibres and axoneme. The principal piece, which makes up the majority of the length of the sperm in most species, is characterized by a cytoskeletal element known as the fibrous sheath. This structure lies just beneath the plasma membrane and has two longitudinal columns that are connected by lateral ribs; its termination marks the beginning of the endpiece.

Figure 1.

 A representative mammalian sperm (domestic dog, Canis lupus familiaris) showing the highly polarized structure associated with this cell type. The head and flagellum are the two major regions of the cell, and the flagellum is itself composed of three distinct and easily identifiable regions: the midpiece, principal piece, and endpiece

Oxidative respiration

These regions and structures of the flagellum are important because specific metabolic pathways are compartmentalized to them. Most easily noted is that the midpiece is the only region of sperm that contain mitochondria, the most efficient producers of cellular energy in the form of adenosine triphosphate (ATP). This restricted localization is conserved evolutionarily across species from relatively primitive external fertilizers such as sea urchins to internal fertilizers such as mammals.

The evolution of dependence upon mitochondria for energy production makes sense because of their efficiency and ability to use endogenous substrates as fuel (seawater having no carbohydrate substrates). Because the sperm of internal fertilizers function within a more hospitable and predictable environment, the mitochondria of sperm from different mammalian species have evolved remarkable differences in their enzymatic capabilities. For example, bull sperm mitochondria can utilize the malate/aspartate shuttle as well as the lactate/pyruvate shuttle, enabling them to metabolize endogenous phospholipids (Storey 1980). These capabilities differ from those of rabbit sperm mitochondria, which have maximal oxidative activity with lactate as substrate (Storey and Kayne 1977). It is believed that these differences have evolved in response to widely varying concentrations of oxidative substrates in the oviductal fluid (Storey 1980).

Regardless of the fertilization environment, the dynein ATPases found associated with the axoneme are the major consumers of ATP in sperm (Storey and Kayne 1980; Halangk et al. 1990). This sets up an extremely interesting question of cellular engineering for all sperm: How is energy conveyed from the mitochondria down the length of the flagellum? In many species, a phosphorylcreatine shuttle utilizing creatine kinase is employed to transfer energy to the distal flagellum (Tombes and Shapiro 1985). Yet this system is either poorly developed or completely absent in most mammalian sperm (Kaldis et al. 1997), suggesting some other adaptive response to work around this issue.


For mammalian species, that alternative strategy involves a reliance upon glycolysis to provide less efficient, but very high-throughput production of ATP locally down the length of the principal piece. To achieve this, glycolysable substrates must be able to pass through the plasma membrane of the flagellum. In the flagella of murine sperm, it is believed that GLUT3 is the primary facilitative glucose transporter (Simpson et al. 2008). GLUT3 is noteworthy because it has both a low Km and a high Kcat, giving it higher affinity for glucose than GLUTs 1, 2 or 4, and a fivefold greater transport capability than GLUTs 1 or 4 (Simpson et al. 2008). However, the exact GLUT family members possessed by sperm seem to vary among species, conveying differing abilities to utilize glycolysable substrates such as fructose, the uptake of which is mediated by GLUT5 in bovine sperm (Angulo et al. 1998).

Once glycolysable substrate enters the sperm, it is acted upon immediately by the enzymes of glycolysis (Fig. 2), which are tethered to the fibrous sheath of the principal piece (Storey and Kayne 1975; Westhoff and Kamp 1997; Mori et al. 1998; Travis et al. 1998; Cao et al. 2006; Krisfalusi et al. 2006). Several factors combine to create a high-throughput capability for local production of ATP, including: (i) the large surface-area to volume ratio, (ii) the extremely small cytoplasmic space between the plasma membrane and fibrous sheath, (iii) the high abundance of glycolytic enzymes on the fibrous sheath, and (iv) the high potential usage of ATP by the dynein ATPases, drawing the reactions in the forward direction. Although single enzymes or small groups of glycolytic enzymes are scaffolded in other cells or are localized in processes at a distance from the cell body to provide local energy production (Amberson et al. 1965; Arnold and Pette 1968; Hsu and Molday 1991; Genda et al. 2011), this example of compartmentalized energy production in the sperm fibrous sheath is one of the most elegant designs in cell biology.

Figure 2.

 Glycolysis in murine spermatozoa. The reactions of glycolysis are diagrammed using the forms of the glycolytic enzymes that are believed to be present in the principal piece of murine sperm. Note that although not truly a part of glycolysis itself, lactate dehydrogenase is included in the diagram because its function is essential for regeneration of NAD+

Although the presence of glycolytic enzymes certainly suggests localized ATP production, it is not in and of itself proof that any ATP produced by this machinery plays a significant role in sperm function. Evidence suggesting the functional importance of the ATP produced by glycolysis in the principal piece comes from several sources. First, it was shown that glucose, but not lactate or pyruvate, is required for the protein tyrosine phosphorylation events observed during the process of sperm capacitation (Travis et al. 2001). This work also demonstrated that inhibiting electron transfer or uncoupling oxidative respiration from ATP production did not impact the sperm’s ability to perform these signalling events, which are believed to regulate patterns of motility (Travis et al. 2001).

Next came two important demonstrations that the ATP from glycolysis plays a major role in powering flagellar motility in mouse sperm. Inhibition of mitochondrial function was found not to impair motility and to have little effect on intracellular ATP levels (Mukai and Okuno 2004). In contrast, addition of a non-metabolizable substrate (2-deoxyglucose) inhibited motility and led to depletion of ATP (Mukai and Okuno 2004). Complementing these findings were those obtained from a genetic approach, namely, investigation of the phenotype of a mouse model made null for the germ cell-specific isoform of glyceraldehyde 3-phosphate dehydrogenase (GAPDHS). Sperm from male mice null for GAPDHS were infertile, with pronounced defects in motility (Miki et al. 2004). Moreover, they had significantly less ATP even though their oxygen consumption was unchanged (Miki et al. 2004).

As with any system, production of energy not only requires fuel, but a means to get rid of waste and a means of replenishing any needed co-enzymes. For glycolysis, these additional requirements are dealt with in the same fashion. In addition to producing two molecules of ATP, glycolysis also generates two molecules each of pyruvate (the ‘waste’) and NADH. Because NAD+ is required for GAPDHS activity, this coenzyme must be replenished from the NADH. Sperm accomplish this through the use of a specific lactate dehydrogenase (Goldberg 1975; Burgos et al. 1995). The lactate produced is removed from the sperm by monocarboxylate transporters, to avoid intracellular accumulation of that acid (Garcia et al. 1995).

Special characteristics of sperm glycolytic enzymes

That the enzymes of glycolysis retain function while tethered to the fibrous sheath is not a trivial matter. Tethering might interfere with protein function in a number of ways, such as by blocking a substrate binding site or by interfering with a needed conformational change. This has been seen both when specific sequences are introduced into recombinant proteins (Halliwell et al. 2001; Tachibana et al. 2006), and more generally when proteins are either non-specifically adsorbed to a surface or attached via chemical approaches that lead to random orientations, such as carboxyl-amine chemistry. To avoid these problems, male germ cells have evolved specific variants of several of these enzymes, often having domains that distinguish them from the isoforms found in somatic cells. Thus far, there is molecular and/or biochemical evidence for germ cell-specific modifications of at least 6 of the 10 enzymes of glycolysis, as well as a germ cell-specific variant of lactate dehydrogenase (LDH-C4) (Zinkham 1968; Buehr and McLaren 1981; Gillis and Tamblyn 1984; Boer et al. 1987; Welch et al. 1992; Zhong and Kleene 1999; Auer et al. 2004). At least several of these germ cell-specific domains have been suggested to be involved in protein targeting (Welch et al. 1992; Mori et al. 1998; Travis et al. 1999).

Applications in Nanobiotechnology

The streamlined shape, ‘solid-state’ design of tethering pathways to cytoskeletal elements, and clear connections between cell structure and function all evoke images of sperm as cellular machines (Travis and Kopf 2002). Could findings from the studies reviewed here regarding sperm metabolism be applied to actual nano- or microscale machines?

Energy production in micro- and nanoscale devices

Most plans for nanobiotechnology applications involve biological molecules as the primary components, because these molecules are the most efficient structural and functional agents at this scale. These biological components could be fashioned into hybrid organic–inorganic devices that could be implanted within a patient to carry out some diagnostic or therapeutic function. Examples of hybrid organic–inorganic devices have already been reported on a nanoscale. For example, recombinant F1-ATPase has been produced that is tethered to a column on one end and supports a small nickel rod on the other end (Soong et al. 2000). Upon hydrolysing ATP, this molecule undergoes a rotary motion and the rod can be seen to move (Soong et al. 2000). An ‘on-off switch’ in the form of a reversible binding site has been engineered into this same molecular motor (Liu et al. 2002), showing that such devices can be regulated. Many other investigations of nanoscale machines also involve effectors that utilize ATP. These include RNA helicases, actin-myosin, dyneins and kinesins, among others (Hess et al. 2004; Jankowsky et al. 2005). To be functional, ATP must either be provided exogenously or produced locally on the device itself. A system to generate ATP has been reported, using bacteriorhodopsin and the F0F1-ATP synthase (Luo et al. 2005). However, this system requires exogenous light (photons), which is not practical for in vivo medical applications. The production of cellular energy in a way that is practically suited to drive biological reactions on an implantable device is one of the fundamental challenges in the development of medical applications for nanobiotechnology.

Biomimicry of sperm glycolysis

Humans have often looked to natural models to answer such questions of bioengineering (Ummat et al. 2005; Wendell et al. 2006). Our work on mammalian sperm suggested to us a strategy for how ATP could be produced in vivo by implantable devices, potentially on either a micro- or a nanoscale. We hypothesized that replacing germ cell-specific targeting domains with peptide binding tags (e.g. a histidine repeat to allow binding to surfaces modified with nickel-nitrilotriacetic acid) would allow these enzymes to be tethered to an inorganic support in a specifically oriented fashion and retain a high degree of function. Inclusion of this design on implantable nanodevices might provide a practical method of utilizing freely available, circulating glucose to produce ATP that could power various biological reactions.

To test this hypothesis, we generated recombinant forms of the male germ cell-specific hexokinase (HK: NM 010438; catalyses the conversion of glucose to glucose-6-phosphate, utilizing ATP) and glucose-6-phosphate isomerase (GPI: NM 008155; catalyses the conversion of glucose-6-phosphate to fructose-6-phosphate). We tested their functions individually and in series, in solution and then when tethered. In support of the hypothesis, we found that the specific activity of the GPI tethered via this strategy of oriented immobilization had significantly higher specific activity than GPI that was randomly adsorbed to the surface (Mukai et al. 2009; Fig. 3). After showing the activity of the enzymes tethered individually, we then demonstrated the activity of the coupled reaction with both enzymes tethered to the same support (Fig. 4). To our knowledge, we believe this was the first demonstration of coupled enzyme activities of sequential steps of a pathway on a single surface (Mukai et al. 2009). This finding is an important first step toward a platform technology based on tethered glycolytic enzymes that could provide energy for hybrid organic–inorganic devices (Hess 2009).

Figure 3.

 Oriented immobilization of GPI resulted in improved binding and higher specific activity. (a) Placement of a binding peptide (a repeat of six histidine residues, 6× His tag) enabled oriented immobilization as the tag bound nickel tethered to the gold surface via nitrilotriacetic acid (Ni-NTA). (b) When the His tag was removed, GPI adsorbed to the surface in random orientations. (c) Approximately twice as much GPI bound to the chip surfaces when the His tag was present. (d) Of interest, the specific activity of GPI bound using oriented immobilization was approximately ninefold higher than that adsorbed to the surface randomly. A single asterisk denotes statistical significance at p = 0.012, whereas the double asterisk denotes statistical significance at p = 0.006. Figure modified from Mukai et al. 2009

Figure 4.

 Coupled reaction of tethered His-HK and His-GPI. (a) Both HK and GPI were immobilized on a gold surface in oriented fashion using a 6× His tag binding to Ni-NTA. (b) Using coupled biochemical reactions, the activity in series of HK and GPI was observed as a change in absorbance at 340 nm. Control chips with either HK alone or GPI alone did not show this activity and yielded results identical to a buffer alone control

Future Directions

We are now in the process of generating recombinant forms of the rest of the glycolytic enzymes, in an effort to recapitulate the design of the sperm fibrous sheath on hybrid organic–inorganic devices. However, biomimicry of sperm need not stop with glycolysis. Future implantable medical devices, if encapsulated to minimize release of their cargo of therapeutic drugs or diffusion of reaction intermediates, would run into other engineering needs that sperm have overcome. For example, such devices would need to address questions of substrate uptake, export of waste products, and replenishment of intermediates. Sperm might provide the inspiration for these and other challenges as well.


The authors acknowledge the contributions of the many scientists whose work was described here. We regret our inability to cite all works in sperm metabolism appropriately, though such an undertaking would far exceed the scope of this brief review. In particular, we acknowledge the foundational discoveries of Dr. Bayard Storey, who not only continues to provide an encyclopaedic source of knowledge about sperm biochemistry and metabolism, but also continues to serve as a mentor and friend to many in the field. This work was supported by an NIH Pioneer Award, 5DP1-OD-006431 and a NYSTAR Center for Life Science Enterprise grant (A.J.T.) and by JSPS Grants-in-Aid for Scientific Research (Wakate B) and JSPS Institutional Program for Young Researcher Overseas Visits (CM).

Conflicts of interest

The authors have no financial or personal relationships that could inappropriately bias or influence this work.