Report

Modulation of Cell Adhesion and Motility in the Immune System by Myo1f

See allHide authors and affiliations

Science  06 Oct 2006:
Vol. 314, Issue 5796, pp. 136-139
DOI: 10.1126/science.1131920

Abstract

Although class I myosins are known to play a wide range of roles, the physiological function of long-tailed class I myosins in vertebrates remains elusive. We demonstrated that one of these proteins, Myo1f, is expressed predominantly in the mammalian immune system. Cells from Myo1f-deficient mice exhibited abnormally increased adhesion and reduced motility, resulting from augmented exocytosis of β2 integrin–containing granules. Also, the cortical actin that co-localizes with Myo1f was reduced in Myo1f-deficient cells. In vivo, Myo1f-deficient mice showed increased susceptibility to infection by Listeria monocytogenes and an impaired neutrophil response. Thus, Myo1f directs immune cell motility and innate host defense against infection.

In both mouse and human genomes, 16 genes encode conventional class II muscle and nonmuscle myosins, with 25 “unconventional” myosin genes encoding 11 other classes (1). Natural mutations of various myosin genes result in an array of genetic disorders, including cardiomyopathies, deafness, blindness, glomerular nephritis, and neuropathies (2, 3). The class I myosins are the largest group of unconventional myosins and are evolutionarily ancient, existing in a wide range of species from yeast to vertebrates (1, 4). Mice and humans have a total of eight class I myosin heavy-chain genes, six of which encode short-tailed forms (Myo1a, b, c, d, g, and h) and two of which encode long-tailed (amoeboid) forms (Myo1e and f) (1). All class I myosins consist of an N-terminal motor domain, light-chain–binding IQ motifs, and a basic tail homology 1 (TH1) domain thought to affect interactions with membranes (2). The long-tailed class I myosins have an additional proline-rich TH2 domain and a TH3 domain containing a single Src homology 3 (SH3) domain (2).

The class I myosins in Dictyostelium and yeast are involved in migration, phagocytosis, endocytosis, and actin remodeling (5, 6). Short-tailed class I myosins in vertebrates are involved in more specialized functions, such as the adaptation of hair cells in the ear (7) and the transport of vesicles and organelles (8, 9), as well as the structural maintenance of the enterocyte microvilli (10). However, the function of long-tailed class I myosins in vertebrates is poorly characterized (1114).

Myo1f was first identified in our screen for differentially expressed genes in subsets of murine lymphocytes. In contrast to previous data suggesting the widespread expression of Myo1f in tissues (15), our results, which we obtained using specific probes, showed that Myo1f is selectively expressed in the spleen, mesenteric lymph nodes, thymus, and lung (Fig. 1A). By comparison, specific detection of Myo1e showed a predominant expression pattern in the spleen and mesenteric lymph nodes and moderate expression in the lung, small intestine, and large intestine (Fig. 1B). Within the lymphoid tissues, natural killer (NK) cells, macrophages, and dendritic cells were found to express considerable levels of both Myo1f and Myo1e; neutrophils and B cells showed selective expression of Myo1f and Myo1e, respectively (Fig. 1, C and D).

Fig. 1.

Myo1f is expressed predominantly and differentially in the immune system. (A and B) mRNA levels of Myo1f (A) and Myo1e (B). EtBr, ethidium bromide. Sp, spleen; Ln, lymph node; Mln, mesenteric lymph node; Th, thymus; Lv, liver; Hr, heart; Lg, lung; Kd, kidney; Si, small intestine; Li, large intestine; Br, brain; Ie, inner ear; Ey, eye; Ms, muscle; Sk, skin. Relative mRNA levels of Myo1f (C) and Myo1e (D) in different cell types of the immune system determined by real-time reverse transcription polymerase chain reaction. T, T cells; B, B cells; NK, NK cells; PMN, neutrophils; MΦ, macrophages; DC, dendritic cells. Relative expression levels are displayed as a number on top of each bar. (E) IgG-mediated phagocytosis. Neutrophils were incubated with IgG-opsonized, fluorescein isothiocyanate (FITC)–labeled Escherichia coli at 37°C. Control neutrophils were incubated at 4°C. MCF, mean channel fluorescence; WT, wild type. Error bars represent SD of the mean. (F) The production of reactive oxygen species.

To determine the function of Myo1f in the vertebrate immune system, we generated Myo1f gene–deficient mice. We focused on neutrophils because Myo1f was detected exclusively in neutrophils (Fig. 1, C and D). Immunoglobulin G (IgG)–mediated phagocytosis was similar between wild-type and knockout (KO) neutrophils (Fig. 1E). To evaluate the degree of pathogen killing that follows phagocytosis, we measured the production of reactive oxygen species. Again, no considerable difference was detected between wild-type and KO neutrophils (Fig. 1F). Thus, Myo1f is dispensable for both the phagocytosis of bacteria and their destruction.

Integrin-mediated adhesion to the vascular endothelium is crucial in the process of neutrophil migration to infected tissue, and the dominant integrins involved in this process belong to the β2 integrin (CD18) family (16). Myo1f-deficient neutrophils exhibited stronger adhesion to integrin ligands, including the intercellular adhesion molecule–1 (ICAM-1) (CD54) and fibronectin (Fig. 2, A and B). Activation of neutrophils by the proinflammatory cytokine tumor necrosis factor–α did not compensate for this difference, suggesting that increased adhesion did not result from changes in the activation status of Myo1f-deficient cells. Experiments with a blocking antibody showed that most of the adhesion was mediated by β2 integrin (Fig. 2, A and B). In addition, Myo1f affected only integrin-mediated adhesion, not integrin-independent adhesion to polylysinecoated substrate (Fig. 2C) (17). Spreading of Myo1f-deficient neutrophils on ICAM-1 was also increased as compared to that of wild-type neutrophils (Fig. 2D). Increased spreading was not due to a loss of cortical tension (fig. S2), which acts to maintain the round shape of the cells (Fig. 2E). In contrast, myosin I double mutants in Dictyostelium exhibit abnormalities in surface morphology resulting from an inability to maintain a sufficient level of resting cortical tension (5, 18). Together, these results suggest that, in the absence of Myo1f, neutrophils are more adherent to β2 integrin ligand.

Fig. 2.

Adhesion and spreading of neutrophils through integrins are abnormally increased in the absence of Myo1f. (A and B) Integrin-mediated adhesion to ICAM-1 (1 μg/ml) (A) and to fibronectin (1 μg/ml) (B). *P < 0.02; **P < 0.0003; #P < 0.02; ##P < 0.0005 (Student's t test). Data are representative of at least two independent experiments. TNFα, tumor necrosis factor–α; Ab, antibody. (C) Integrin-independent adhesion to polylysine (1 mg/ml). Error bars in (A) to (C) represent SD of the mean. (D) Spreading of neutrophils to ICAM-1 (1 μg/ml), observed by SEM. (E) Surface morphology of neutrophils attached to polylysine (1 mg/ml), determined by SEM.

Under resting conditions, no considerable difference was seen in the level of β2 integrin on the surface of Myo1f-deficient cells (Fig. 3A). In neutrophils, a large pool of integrin αMβ2 (CD11b/CD18) is stored in cytoplasmic granules. This allows rapid up-regulation of integrin on the surface by granule exocytosis after stimulation (19). Myo1f-deficient neutrophils exhibited hyperinduction of cell-surface β2 integrin upon stimulation with integrin ligands (ICAM-1 and fibronectin) (Fig. 3A), although total cell levels of β2 integrin expression were comparable in wild-type and Myo1f-deficient cells (fig. S3). In addition, this up-regulation was specific for proteins in granules, because no difference was observed in the surface expression of the neutrophil marker Gr-1, which is not known to associate with neutrophil granules (fig. S4). The release of lactoferrin, a marker for the specific granules in which a large portion of integrin αMβ2 is stored (19), is also significantly augmented in Myo1f-deficient cells upon stimulation with ICAM-1 and fibronectin (Fig. 3B). Thus, in the absence of Myo1f, increased granule exocytosis induced by integrin ligands leads to augmented β2 integrin on the cell surface, making neutrophils more adherent.

Fig. 3.

In the absence of Myo1f, granule exocytosis is augmented, resulting in more β2 integrin on the surface and impaired migration on the extracellular matrix. (A) Neutrophils were stimulated with plate-bound ICAM-1 (1 μg/ml) or plate-bound fibronectin (1 μg/ml) or fMLP (10 μM) for the indicated times. Surface levels of β2 integrin (CD18) were represented by MCF, with fluorescence of WT neutrophils at 0 min given a value of 100 units. Data are representative of three independent experiments. (B) Secreted lactoferrin was measured in the supernatant at 30 min after stimulation. *P < 5 × 10–6; **P < 5 × 10–5 (Student's t test). Data are representative of two independent experiments. (C to F) Time-lapse video microscopy was used to examine neutrophil migration (movies S1 to S4) on polylysine (1 mg/ml) [(C) and (E)] or fibronectin (1 μg/ml) [(D) and (F)]. The relative position of each cell is shown as a diamond in the graphs [(E) and (F)] after 10 min of migration under a 10 μM fMLP gradient, assuming that the initial position of the cells was at (0, 0) in an xy plane. Lines with 120° of arc show the area facing the source of fMLP. Data are combined from six independent experiments. (C) and (D) Average migration rate of neutrophil on polylysine (C) and fibronectin (D). #P < 3 × 10–14 (Student's t test). Error bars in (A) to (D) represent SD of the mean.

To examine the migratory properties of neutrophils, N-formyl-Met-Leu-Phe (f MLP)–gradient chemotaxis was measured by time-lapse microscopy. When observed on polylysine-coated glass, migration rates and efficiencies of wild-type and Myo1f-deficient neutrophils were similar (Fig. 3, C and E; fig. S5; and movies S1 and S2), in contrast to Dictyostelium myosin I mutants, which displayed low migration efficiency because of frequent turning (20). However, the migration rate of Myo1f-deficient neutrophils was severely impaired on fibronectin (Fig. 3D), with most neutrophils remaining attached on fibronectin (Fig. 3F and movies S3 and S4). These results suggest that, although the machinery for migration is intact in the absence of Myo1f, the increased integrin level on the surface makes these neutrophils more adherent and less motile (21).

To further explore the molecular basis of augmented granule exocytosis in Myo1f-deficient cells, we examined the intracellular location of Myo1f. Myo1f co-localizes with cortical actin filaments at the rim of the cells (Fig. 4A), and there is little co-localization with lactoferrin, which is a marker for neutrophil granule (fig. S6A). The low levels of diffuse staining in the cytoplasm are probably nonspecific, because this phenomenon is also observed on the KO neutrophils. Class I myosins in Dictyostelium and in yeast interact with the actin remodeling complex. Quantification of F-actin showed that there was a small but significant reduction in the amount of polymerized actin of resting neutrophils in the absence of Myo1f, whereas chemotaxis-related (f MLP-induced) actin polymerization was not affected (Fig. 4B), suggesting differential regulation of these actin structures (22). Cortical actin filaments near the plasma membrane, which constitute most of the F-actin in resting cells, can work as a barrier to granule secretion (23). Thus, it is possible that Myo1f indirectly inhibits granule exocytosis by modulating cortical actin. It is also possible that Myo1f exerts a direct opposing force by associating with granules. If so, this process would presumably occur as granules move into the cortex before exocytosis, because there was little co-localization of Myo1f with cytoplasmic granules even in cells that were stimulated for 7 min (fig. S6B). However, the cortical “rim” staining of Myo1f in activated cells is somewhat discontinuous and more diffuse, suggesting that there is some redistribution of Myo1f upon stimulation (fig. S6B).

Fig. 4.

Myo1f co-localizes with the cortical actin network, and neutrophil-dependent restriction of L. monocytogenes infection in vivo is impaired in the absence of Myo1f. (A) Confocal microscopy of neutrophils stained with antibody to Myo1f (red) and phalloidin for F-actin (green). (B) Quantification of F-actin in neutrophils after stimulation with 10 μM fMLP. MCF was used to quantify F-actin in neutrophils after staining with FITC-phalloidin, with fluorescence of WT neutrophils at 0 min given a value of 100 units. *P < 6 × 10–5 (Student's t test). (C) The number of L. monocytogenes in the spleen of WT (n = 8) and Myo1f KO (n = 9) mice observed 2 days after infection. **P < 0.012 (Student's t test). Red bars indicate mean values. (D and E) Microabscess formation in the liver of WT (D) and Myo1f KO (E) mice at 24 hours after infection. Cells with the strong blue color are neutrophils, and pale pink cells are infected hepatocytes undergoing apoptosis (25). (F) Adoptive transfer of neutrophils into mice infected with Listeria. The number of neutrophils localized in the liver was counted. #P < 5 × 10–9 (Student's t test). Error bars in (B) and (F) represent SD of the mean.

Neutrophils play a crucial role in the early phase of protection against infection by L. monocytogenes (24). To test the effect of the phenotype described above in vivo during infection, wild-type and Myo1f-KO mice were challenged with Listeria. On day 2 after infection, Myo1f-KO mice contained 35 times more colony-forming units (CFUs) of Listeria than did wild-type littermate controls (Fig. 4C). In contrast to the well-formed microabscesses in wide-type mice (25) (Fig. 4D), poorly formed microabscesses in KO mice exposed large areas of infected hepatocytes (pale pink areas in Fig. 4E), indicating that the decreased ability to control Listeria is due to a defect in neutrophil mobilization. To further confirm that this in vivo defect is neutrophil-autonomous, we adoptively transferred wild-type and Myo1f-KO neutrophils into normal mice infected with Listeria. Myo1f-KO neutrophils were found in lower numbers in the recipient livers as compared to the observations in recipients of wild-type neutrophils (Fig. 4F), suggesting that the motility of neutrophils in vivo is compromised in the absence of Myo1f.

In accordance with the reported similarity between Dictyostelium and mammalian neutrophils (26), we observed that long-tailed class I myosin is conserved in cells from two different genera and supports cell motility. However, our results indicate that the ways in which myosin contributes to cell migration in the two models are quite different. Although long-tailed class I myosin in Dictyostelium prevents lateral pseudopod formation during migration (20), Myo1f in neutrophils inhibits exocytosis of integrin-containing granules, thereby preventing excess adhesion. Unlike protozoan cells, metazoan cells use integrins to interact with other cells and with the environment (27). Thus, our results suggest that, in higher organisms, long-tailed myosins have evolved to accommodate and make the best use of integrins to support cell motility. Our work also demonstrates that seemingly closely related myosin motors have different functions. Whereas short-tailed class I myosins facilitate vesicular transport (8, 9), the long-tailed form Myo1f negatively regulates granule exocytosis in neutrophils. This work should lead to a better understanding of human immune defects that are currently of unknown origin and may provide an additional class of therapeutic targets for treating acute inflammation.

Supporting Online Material

www.sciencemag.org/cgi/content/full/314/5796/136/DC1

Materials and Methods

Figs. S1 to S6

References

Movies S1 to S4

References and Notes

Stay Connected to Science

Navigate This Article